Three-dimensional tuberculoma bioplatform and uses thereof

Abstract

Provided herein is a 3-dimensional bioplatform comprising immune cells and mycobacteria, which form mycobacteria-in-spheroids and develop into tuberculoma with granuloma lesions. Also provided are methods of making the bioplatform and uses of the bioplatform in screening methods.

Description

THREE-DIMENSIONAL TUBERCULOMA BIOPLATFORM AND USES THEREOFSTATEMENT REGARDING FEDERALLY SPONSORED RESEARCH OR DEVELOPMENT

[0001] This invention was made in part with Government support under Intramural Funding [Combating Antibiotic-Resistant Bacteria, Laboratory Safety Science and Innovation, and Division of Tuberculosis Elimination Intramural Research Funds] awarded by the Centers for Disease Control and Prevention (CDC). The Government has certain rights in this invention.FIELD OF THE INVENTION

[0002] The present invention relates to a 3-dimensional bioplatform comprising immune cells and mycobacteria, which form ‘mycobacteria-in-spheroid’ co-cultures and develop into tuberculomas with a collection of granuloma lesions. The invention also relates to the methods of making a fresh or cryo-stable bioplatform and the uses of the bioplatform in screening methods.BACKGROUND OF THE INVENTION

[0003] Tuberculosis (TB) is among the leading infectious killers of humankind worldwide, and the current strategies for the prevention and treatment of TB are not enough to achieve elimination in this century. One key strategy to meet this goal is the development of shorter treatment regimens for both drug-susceptible and resistant TB. The standard treatment for drug- susceptible TB, although effective, is lengthy, needing 4 to 9 months of treatment, and the treatment for drug-resistant TB increases to 18-24 months. The prolonged treatment duration combined with drug toxicity can complicate treatment completion, thereby contributing to the development of drug resistance. Extensive efforts are underway to develop new antibiotics for the treatment of TB, but conventional pathogen-targeted strategies suffer from the serious disadvantage of fostering microbial resistance. To circumvent these problems, a new treatment paradigm that involves therapeutic modulation of the host immune response to improve pathogen eradication, known as host-directed therapy (HDT), has emerged. HDT has the potential to not only shorten treatment but decrease the probability of developing resistance. Besides, HDT can reduce immunopathology in target organs, promote immunological memory, and prevent disease relapse.

[0004] Most new therapeutics are initially screened in vitro against a specific pathogen without any host immune cells, for example, using agar or liquid cultures. However, this is not the best approach against intracellular pathogens such as Mycobacterium tuberculosis (Mtb). Animal models can be used to identify new therapeutics, but these models are very costly and limit the number of compounds that can be screened. Furthermore, animal models may not accurately represent the human latent TB infection (LTBI) and TB disease. Using animal models of TB for such screens also presents significant challenges due to the requirement of an animal biological safety level-3 (ABSL-3) facility to conduct these studies.

[0005] Macrophages are natural target host cells for Mtb infection and serve as the major niche for infection and as front-line enforcers of protective immunity, both innate and acquired immunity, to control or eliminate the pathogen. Mtb, however, presents numerous challenges by manipulating host-signaling pathways to evade, tolerate, and subvert innate and adaptive immune responses. This, nevertheless, also creates opportunities for novel treatment strategies like HDT. Recently, there has been considerable progress in demonstrating the feasibility of the HDT approach against TB. Yet little is known about the potential targets of HDT in humans. To progress HDT toward clinical application, a greater understanding of interactions between the host and pathogen is required, especially at the level of TB lesions called granulomas. While granuloma structure is traditionally thought to benefit the host by limiting infection, recent studies using animal models have revealed that it facilitates mycobacterial growth and dissemination and can be detrimental to the host. One of TB’s more severe clinical manifestations is the consolidation of multiple adjacent tubercular granulomas into structurally organized 3-dimensional (3D) mass termed a ‘tuberculoma’ in the infected organs, most often in the lungs and brain.

[0006] The field needs to overcome numerous challenges to accelerate studies to identify host pathways manipulated by Mtb that an HDT or immunotherapy can target. Most importantly, we lack a fast, robust, and widely applicable bioplatform for rapidly screening host-directed compounds and other therapeutics against TB.

[0007] Therapeutics that modulate host-A / rf> interactions can be identified using traditional 2- dimensional (2D) cell cultures, also called macrophage “monolayers.” However, this approach is limited by poor predictability of the drug efficacy and toxicity, the absence of accurate insights into the mode of action of selected compounds, and the failure of the system to mirror themicroenvironments within the TB granuloma lesions. Certain drugs that work against TB in the granuloma milieus cannot be identified using the 2D systems, as they are ineffective in the in vitro 2D environment. Yet many published studies have used 2D cell culture of human or mouse monocyte / macrophage cell lines, often after mitogen stimulation and differentiation, which result in alterations in cell signaling and other host cellular processes. Primary human macrophages are generally preferred over monocyte / macrophage cell lines to investigate host-pathogen interactions in vitro. However, even bronchoalveolar lavage (BAL) or leukapheresis does not usually generate enough lung or blood monocyte-derived macrophages from human donors for high throughput screens. To circumvent many of these issues and mimic the microenvironment in human TB granulomas, the in vitro cell culture models of granuloma formation have been developed using A-f / b-infected human peripheral blood mononuclear cells (PBMCs). Many of these granuloma models are developed in the 2D system, such as flat-bottomed cell culture microplates. These 2D models manifest some traits of the granuloma, such as macrophage aggregation, multinucleated giant cell formation, and necrosis; however, these structures are poorly organized, microscopic in size, easily disturbed during media replenishment, and lack key granuloma attributes due to their failure to undergo the complete gamut of changes associated with the human 3D tuberculous granuloma structure. These 2D granuloma models are usually developed using granulocyte-macrophage colony-stimulating factor (GM-CSF) or macrophage colony-stimulating factor (M-CSF) differentiated human PBMCs. However, because of such in vitro differentiation and activation processes, they are not suitable for uncovering the accurate mechanisms of action of HDT compounds and immuno-therapeutics.

[0008] A human TB granuloma lesion is an organized 3D collection of immune cells, primarily consisting of a central core of tissue-resident and recruited blood monocyte-derived macrophages surrounded by lymphocytes in the infected lungs. Therefore, the in vitro granuloma models using human PBMCs and / or lung cell lines embedded in the collagen matrix in flat-bottomed microplates to form 3D structures have also been developed. In these 3D models, loose miniature granulomatous aggregates of macrophages are formed in response oMtb infection, and bacteria develop features of dormancy. Despite its usefulness in studying host-pathogen interactions during latency and resuscitation, this model is typically laborious and suffers from low throughput. Moreover, there are technical challenges to adding more host cells to maintain thedynamic granuloma structures over an extended period, and enzymatic treatment with collagenase is required to release cells from the matrix for downstream applications.

[0009] Another recently developed in vitro model uses a 3D system based on AAA-infected PBMCs and collagen matrix encapsulation within the alginate microspheres by bioelectrospraying. This encapsulated system is then imaged and used in the 2D microplates to study host-pathogen interactions or effects of anti-TB drugs. Although applicable and helpful in investigating diverse aspects of host-pathogen interactions and the effects of antimycobacterial agents, the cellular aggregates formed in this model are also miniature and lack well-organized solid or cavitary granuloma lesion formation and associated microenvironments, including central necrosis, cavity formation, and the gradient of oxygen, nutrients, and drugs due to porous nature of microspheres. The effect of the electric voltage applied during bio-electrospraying on host cell response is also not completely understood. This 3D system development requires technical expertise and instrumentation for bioelectrospraying in the BSL-3 facility. Further addition of cells simulating cell recruitment within the hardened microspheres is not possible, and releasing cells from the microspheres for downstream assays like flow cytometry or RNA sequencing is required, which can affect the transcriptomic and immunological signatures.Recently, a new ex vivo 3D culture system has been described. In this technique, fully organized granulomas from the AT. rzrznw -infected adult zebrafish model are micro-dissected and maintained in the 3D cell culture. Although this technique is helpful for high-resolution imaging and manipulation of granulomas, only a small number of granulomas (20-100) per animal can be micro-dissected, which lacks efficiency for high throughput compound screens, and it also requires a special facility to house infected animals, specialized instrumentation, and technical expertise.

[0010] As described above, available 2D and 3D cell culture systems and in vitro granuloma models lack key features and microenvironments in human TB lesions, such as the formation of tuberculoma structure, well-organized granuloma lesions, biochemical and physicochemical gradients, and development of enhanced hypoxia, necrosis, acidosis, and cavity formation. The confined 3D bioplatforms, such as the collagen embedded human PBMC model or the alginate microsphere model encapsulating human PBMCs admixed with the extracellular matrices, are complex, lack pliability and throughput for drug screening, and releasing viable cells from these bioplatforms for additional downstream investigations is challenging. The size and density ofmicroscopic cell aggregates formed in these models vary greatly between blood donors, and these aggregates are highly unstable. In addition to reducing bacterial burden, one of the desirable aspects of human TB treatment is the resolution of the number and size of granuloma lesions (immunopathology) as a measure of treatment success, cure, and preservation of tissue architecture. The currently available in vitro models and bioplatforms lack the ability to provide a multidimensional assessment of treatment efficacy, such as penetration of the drug in the granulomatous lesions and resolution of the number and size of these lesions, in addition to the reduction in bacterial burden. Furthermore, successful preservation, maintenance, and long-term storage of screening bioplatform is highly desirable for commercial and research applications, and the shelf-stability of currently available granuloma platforms is unknown. Accordingly, there is a need in the art for improved, robust, and shelf-stable 3D bioplatforms for studying hostpathogen interactions involving TB and screening for TB therapeutics.SUMMARY OF THE INVENTION

[0011] Provided herein is a three-dimensional (3D) cell co-culture comprising a plurality of human immune cells and a plurality of mycobacteria. The 3D co-culture may be made by (a) suspending the human immune cells in a Roswell Park Memorial Institute cell culture medium, further comprising one or more of L-glutamine, FBS, sodium pyruvate, and HEPES buffer; (b) mixing the mycobacteria with the human immune cells at a MOI of 0.005-0.5; wherein the mixture is contained in a polystyrene U-bottom shaped ultra-low attachment well of a microplate, wherein the well comprises a hydrogel coating, and wherein the hydrogel is hydrophilic, neutrally charged, and biologically inert; and (c) incubating the mixture at 37°C for a time sufficient for the human immune cells and mycobacteria to form mycobacteria-in- spheroid structures and granuloma lesions.

[0012] Also provided herein is a method of making a 3D cell co-culture comprising a plurality of human cells and a plurality of mycobacteria. The method may comprise (a) suspending the human immune cells in a cell culture medium, further comprising one or more of L-glutamine, FBS, sodium pyruvate, and HEPES buffer; (b) mixing the mycobacteria with the human immune cells at a MOI of 0.005-0.5; wherein the mixture is contained in a well of a microplate; and (c) incubating the mixture at 37°C for a time sufficient for the human immune cells and mycobacteria to form mycobacteria-in-spheroid structures and granuloma lesions.

[0013] The mycobacteria-in-spheroid structures disclosed herein may be formed by the human immune cells and the plurality of mycobacteria. The human immune cells may comprise cells selected from the group consisting of monocytes, macrophages, and peripheral blood mononuclear cells (PBMCs). The monocytes may comprise THP-1 cells, U937 histiocytes, purified CD14+monocytes, or monocyte cell subsets in PBMCs. The mycobacteria may comprise Mycobacterium marinum (Mm), Mycobacterium tuberculosis (Mtb), or a Mycobacterium tuberculosis complex species. The Mm may comprise strain 1218 or strain M. The Mtb may comprise strain H37Rv, strain Erdman, strain CDC1551, strain Beijing F2, or a clinical isolate of a drug-susceptible or drug-resistant Mtb.

[0014] The 3D co-culture may comprise 500-5000 colony forming units (cfu) of mycobacteria per 105human immune cells, resulting in a multiplicity of infection (MOI) of 0.005-0.05. The 3D co-culture may comprise Mm strain 1218 or strain M, and the MOI may be about 0.006- 0.012. The 3D co-culture may comprise Mtb, and the MOI may be about 0.012-0.05, or about 0.025 in particular. The 3D co-culture may comprise Mtb strain H37Rv, and the MOI may be about 0.015-0.05. The MOI may be about 0.025. The 3D co-culture may comprise Mtb Erdman or Beijing F2, and the MOI may be about 012-0.05. The 3D co-culture may comprise Mtb Erdman and the MOI may be about 0.05. In the methods disclosed herein, the mycobacteria may be mixed with human immune cells at the MOI.

[0015] The mycobacteria may express Green Fluorescent Protein (GFP), Red Fluorescent Protein (RFP), a fluorescent marker, or a luminescent marker. The RFP may comprise tdTomato or a far-red fluorescent protein. The 3D cell co-culture may be contained in a U-bottom shaped well, which may comprise a hydrogel coating. The hydrogel may be one or more of hydrophilic, neutrally charged, and biologically inert. The hydrogel may comprise a perfluorinated polymer, olefin, or a combination thereof. The hydrogel may comprise a CORNING® PURAMATRIX™ Peptide Hydrogel. The hydrogel may comprise one or more of collagen, fibrin, and alginate. The well may be of a CORNING® Ultra-Low Attachment Spheroid Microplate or an S-BIO PRIMESURFACE® 3D Culture Spheroid Plate. The 3D cell co-culture may further comprise one or more additional cell type selected from the group consisting of A549 human lung epithelial cells, HUVEC-1 human fetal endothelial cells, HULEC human lung endothelial cells, and MRC-5 human lung fibroblasts or primary human epithelial, endothelial, and fibroblast cells.

[0016] In the methods disclosed herein, the mycobacteria-in-spheroid structures may be contacted with an extracellular matrix (ECM). The ECM may be added to the mycobacteria-in- spheroid structures 3 days after the human immune cells and the mycobacteria are mixed. The methods may further comprise placing and storing the mixture of human immune cells and mycobacteria at a temperature of -80°C, and optionally at a temperature of < -160°C, within 30 minutes of mixing the human immune cells and the mycobacteria. The methods may further comprise after mixing the human cells and the mycobacteria, incubating the mixture at 37°C for 16-72 hours; removing the cell culture medium; adding a cryopreservative to the co-culture of human cells and mycobacteria; and freezing the spheroid co-culture. The cryopreservative may comprise 5% (v / v) dimethyl sulfoxide (DMSO) in RPMI-1640 or 3D cell culture medium. The cryopreservative may further comprise heat-inactivated fetal bovine serum.

[0017] Provided herein is a method of screening for a molecule capable of preventing, treating or reducing a mycobacterial infection. The method may comprise (a) contacting the 3D bioplatform with the molecule; (b) measuring one or more characteristics of the mycobacteria-in-spheroid structures; and (c) comparing the one of the one or more characteristics to a control, wherein a change in at least one of the one or more characteristics is indicative that the molecule is capable of treating or reducing the mycobacterial infection or one or more pathology features thereof. The one or more characteristics may comprise one or more of the following: (a) the amount of fluorescence produced by the mycobacteria, wherein the mycobacteria express a fluorescent molecule; (b) integrity of granulomas; (c) mycobacterial growth; (d) granuloma formation;(e) granuloma growth; (f) granuloma number and size; (g) autophagolysosome formation;(h) inflammasomes and pyroptosis induction; (i) hypoxia induction or inhibition; (j) lysosomal acidification; (k) host cellular toxicity; and (1) an expression level of one or more genes or protein biomarkers expressed by the cells of the 3D bioplatform.

[0018] The one or more characteristics may comprise the amount of fluorescence produced by the mycobacteria. The amount of fluorescence may be indicative of mycobacterial burden. A reduction in mycobacterial burden for the molecule by a threshold value in comparison to the control may be indicative that the molecule is capable of treating or reducing a mycobacterial infection or one or more pathology features thereof. The threshold value may be a reduction of at least 25%. The threshold value may be a z score less than -2 or less than -4.

[0019] The one or more characteristics may comprise the expression level of one or more genes expressed by the cells of the 3D bioplatform. A change in the expression level of the one or more genes, proteins or immune processes as compared to the control may be indicative that the molecule is capable of treating or reducing a mycobacterial infection or one or more pathology features thereof The control may comprise uninfected human immune cells or infected but untreated human immune cells. The molecule may be added about 0-6 days after the mycobacteria and the human immune cells were co-incubated. The one or more characteristics may be observed one or more times over 1-14 days, optionally once on days 7, 9, and 12, after the 3D bioplatform is contacted with the molecule.

[0020] Provided herein is a method of treating or reducing tuberculoma lesions in a subject in need thereof. The method may comprise administering to the subject an anti-mycobacterial agent. Also provided are the anti-mycobacterial agent for treating or reducing tuberculoma lesions, and use of the anti-mycobacterial agent in the manufacture of a medicament for treating or reducing tuberculoma lesions. The anti-mycobacterial agent may be selected from the group consisting of an anti-CDl la antibody, an anti-a4p7 integrin antibody, an anti-CD30 antibody, an anti-IGFIR antibody, an anti-IL-6R antibody, AT9283, Tizoxanide (active metabolite of nitazoxanide), Dasatinib, Quinacrine diHCL, All-trans Retinoic acid (ATRA), Vorinostat, Sitagliptin, H89, and Lansoprazole. The anti-mycobacterial agent may be selected from the group consisting of an anti-CDl la antibody, an anti-a4 7 integrin antibody, an anti-CD30 antibody, an anti-IGFIR antibody, an anti-IL-6R antibody, AT9283, Tizoxanide (active metabolite of nitazoxanide), Dasatinib, Quinacrine diHCL, All-trans Retinoic acid (ATRA), Vorinostat, Sitagliptin, H89, and Lansoprazole. The subject may have a mycobacterial infection.

[0021] Provided herein is a method of screening for a candidate agent that induces trained immunity in a human immune cell. The method may comprise measuring the growth of the pathogenic mycobacteria in the 3D platform after co-culture of the human immune cells and the pathogenic mycobacteria. The human immune cells may have been contacted with the candidate agent. A decrease in growth of the pathogenic mycobacteria or reduced granuloma lesion development by the co-culture may be indicative that the candidate agent induces trained immunity in the human immune cell. The candidate agent may be a vaccine, immunotherapeutic, biologic, or host-directed therapeutic.BRIEF DESCRIPTION OF THE DRAWINGS

[0022] The patent or application file contains at least one drawing executed in color. Copies of this patent or patent application publication with color drawing(s) will be provided by the Office upon request and payment of the necessary fee.

[0023] FIGS. 1 A-D show schematics of ‘mycobacteria-in-spheroid’ co-culture workflow types in a 96-well 3D cell culture microplate to generate human 3D tuberculoma bioplatform. FIG. 1 A. A workflow using freshly cultured immortalized human THP-1 monocytes and pathogenic mycobacteria to generate innate, solid 3D tuberculoma-like structures. FIG. IB. A modified workflow of THP-1 monocytes and pathogenic mycobacteria incorporating human extracellular matrix to generate 3D tuberculoma-like structures with cavitary features. FIG. 1C. A workflow of purified primary human blood CD14+monocytes and pathogenic mycobacteria that subsequently includes lymphocyte-rich autologous PBMC subsets to generate human donorspecific nascent 3D in vitro tuberculomas with innate and adaptive immune cells. FIG. ID. A workflow for cryo-stable 3D tuberculoma bioplatform using human THP-1 monocytes and pathogenic mycobacteria that can be frozen for future use and revived on demand to generate solid 3D tuberculomas. 3D, three-dimensional; THP-1, a monocytic cell line isolated from the peripheral blood of human leukemia patient; ECM, extracellular matrix; and PBMC, peripheral blood mononuclear cells.

[0024] FIG. 2A-D show results from investigating different multiplicity-of-infection doses of two fluorescent M. marinum strains in the ‘mycobacteria-in-spheroid’ co-culture workflow to optimize 3D tuberculoma bioplatform. FIG. 2A-B show kinetics of mycobacteria proliferation, granuloma lesion formation, and the ensuing pathogenesis in ‘mycobacteria-in-spheroid’ cocultures of THP-1 monocytes (l>< 105) generated with eight different MOI doses (1 to 6000 CFU) of (FIG. 2A) M marinum 1218 (GFP) and (FIG. 2B)Af. marinum ‘M’ (tdTomato) strain. Representative images are of one of the six 3D spheroids generated using each of the eight MOI doses or no infection controls and captured on five different time points over three weeks of growth from one of the two experiments performed for each strain. Scale 2000 pm. FIG. 2C-D show the percent area affected by mycobacterial proliferation and granulomatous lesions in the 3D ‘mycobacteria-in-spheroid’ co-cultures generated using THP-1 monocytes (I x fO5) and nine different increasing MOI doses (1 to 12,000 CFU) of (FIG. 2C) M. marinum 1218 (GFP) and (FIG. 2D) M. marinum ‘M’ (tdTomato) strain. The circle represents the degree of pathogenesisin one 3D spheroid. The bar represents the average percent area affected by pathogenesis. The data are mean ± SD (n = 6 spheroids) from one of the two experiments. The co-culture using low MOI 0.006 (600 CFU) of M. marinum strains results in a well-organized granuloma lesion formation with < 50% and 75% of the 3D tuberculoma area affected by mycobacterial growth by days 12 and 21, respectively. The percentage of area affected, and pathogenesis induced following low MOI 0.006 is compared with that caused by standard MOI 0.12 and with uninfected controls. * p < 0.05 and **** p < 0.0001 using Kruskal-Wallis with Dunn's post-hoc test. NS, nonsignificant.

[0025] FIG. 3A-D show that tuberculoma forms developed using 3D ‘mycobacteria-in-spheroid’ co-culture workflows of freshly cultured human immortalized THP-1 or purified primary blood monocytes and pathogenic mycobacteria. FIG 3A. Solid tuberculomas generated following 3D co-cultures of freshly cultured THP-1 monocytes and pathogenic, M. marinum ‘M,’ M. tuberculosis H37Rv, Beijing F2, CDC1551 or Erdman strain expressing tdTomato fluorescent protein and imaged on day 12. Scale 2000 pm. FIG. 3B. Representative images of solid tuberculomas generated following 3D co-cultures of THP-1 monocytes and wild-type M. marinum ‘M’ or uninfected control spheroids and stained for necrosis (green), hypoxia (red), and acidosis (red) using commercially available staining kits, as per the manufacturers’ instructions. FIG. 3C. shows 3D tuberculomas generated following a modified 3D co-culture workflow of freshly cultured THP-1 monocytes and pathogenic M. marinum 1218 strain expressing GFP and incorporating human ECM comprised of type 1 human collagen and fibronectin (0 pl, 10 pl, or 30 pl). Scale 2000 pm. Cavitary features developed in the 3D tuberculomas but not in control 3D spheroids with 30 pl of ECM. FIG. 3D. Solid tuberculomas generated following 3D co-cultures of total human PBMCs and M. marinum 1218 (GFP) or M. marinum ‘M’ (tdTomato) or following 3D co-cultures of purified CD14+monocytes from human blood and M. tuberculosis Erdman supplemented with autologous lymphocyte rich PBMC fraction. While 3D co-cultures of total PBMCs and M. marinum developed small cellular granulomatous aggregates, 3D cocultures of purified CD14+monocytes and M. tuberculosis Erdman developed well-organized granuloma lesions. Representative images in FIG. 3A-D are from two or more independent experiments with at least three 3D spheroids.

[0026] FIG. 4A-F show that solid tuberculomas formed following thawing and revival of cryopreserved 3D ‘mycobacteria-in-spheroid’ co-cultures generated using THP-1 monocytes andM. marinum ‘M’ tdTomato. FIG. 4A-B show the effects of three different freezing mediums on the granuloma formation in the 3D co-cultures. Significantly more well-organized granuloma lesions developed in 3D co-culture frozen 16 hours post-co-culture in the RPMI 1640-based freezing medium than in FBS-based or L15 medium -based freezing medium and revived 6 weeks or 16 months after freezing. FIG. 4C shows the kinetics of granuloma lesion formation after thawing frozen 3D co-cultures. A comparable trend in granuloma lesion formation was found in a 3D co-culture frozen 30 minutes or 16 hours post-co-culture in the RPMI 1640-based freezing medium and revived 6 weeks after freezing. FIG. 4D-E show the effects of duration of freezing on the granuloma lesion formation. No significant difference in the granuloma count was found in the 3D co-culture frozen in the RPMI 1640-based freezing medium 30 minutes post-co-culture and revived after 3 days, 6 weeks, or 16 months of freezing. FIG. 4F shows the effects of time of freezing after 3D co-culture on organized granuloma formation. Numerous well-organized granuloma lesions developed in the 3D co-culture that was frozen 30 minutes or 16 hours post-co-culture (in FIG. 4A and C) compared to 72 hours post-co-culture in the RPMI 1640-based freezing medium and revived after 3 days of freezing. Scale 2000 pm. Representative images and data (mean ± SD) are from one to three independent experiments (n = 6-16 3D spheroids per group). Circles in histograms represent granuloma lesion count per 3D co-culture. *** and **** indicate p < 0. 001 and < 0.0001 using the Kruskal-Wallis with Dunn's post-test. NS, nonsignificant.

[0027] FIG. 5A-E show the uniformity and reproducibility of 3D tuberculomas and bioplatform. FIG. 5A-D show the consistency in the diameter (FIG. 5A), area (FIG. 5B), granuloma lesion count (FIG. 5C), and bacterial burden (FIG. 5D) in the solid 3D tuberculomas developed in 96- well microplates by two separate assay performers using freshly cultured THP-1 monocytes and M. marinum ‘M’ tdTomato. FIG. 5E shows the separation of bacterial burden in nitazoxanide or rifampicin-treated (positive controls) and untreated or DMSO-treated (negative controls) 3D tuberculomas. Data in box plots (whiskers: 10-90 percentile) in A-D are from 72-144 tuberculomas per treatment from independent experiments. Circles represent a response per 3D tuberculoma. Parameters in the 3D tuberculomas developed by the two performers were not significantly different using the Mann-Whitney test. ND, not done. Bacterial burden data in (FIG. 5E) are from 3-6 tuberculomas per treatment from a total of 48 3D cell culture plates (24 plates per performer) developed for screening potential pathogen-targeted and host-directedcompounds using 3D tuberculoma bioplatform. Rifampicin was not tested in plates 25-48 developed by performer 2.

[0028] FIG. 6A-D show assay quality and high-throughput screening compatibility of 3D tuberculoma bioplatform determined by Z’-factor. FIG. 6A-C show the effects of infection dose (MOI) on the Z’-factor of the drug screening assay in the 3D tuberculoma bioplatform developed using THP-1 monocytes and varying infection doses of M. marinum ‘M’ (tdTomato) in the 96- well microplate. Three different MOI doses, 0.0059 (FIG. 6A), 0.0068 (FIG. 6B), and 0.0091 (FIG. 6C), were investigated. The upper panels show the Z’-factor in the assay using 8-10 microplates per MOI experiment, including negative and positive controls in each plate. Relative fluorescence unit (RFU) data indicating bacterial burdens are presented as box plots with whiskers (minimum to maximum) showing all data points, where the individual circle represents the RFU level in one solid 3D spheroid. Fluorescence intensity was measured using auto gain, and the range of auto gain values for microplates used in an experiment is shown. The lower panels show Z’ -factors in individual plates of the bioplatform assay, where an individual symbol represents Z’-factor in one microplate. The assay using MOI 0.0091 indicated an excellent to good assay quality (average Z’ factor > 0.5 shown by dotted red line). FIG. 6D shows a plate map showing the typical location of control and test wells on a 96-well plate. In a primary screen of the compound library, a total of 16 compounds can be screened in triplicate per plate.FIG 7A-J show that the 3D cell culture of human monocytes and pathogenic mycobacteria establishes a facile in vitro tuberculoma model. FIG 7A shows a schematic overview of the workflow of ‘mycobacteria-in-spheroid’ co-culture using pathogenic fluorescent mycobacteria and human THP-1 or primary CD14+blood monocytes in the Coming 3D cell culture microplates for 3D tuberculoma model generation and therapeutic screening and characterization. Magnetic-associated cell separation (MACS)-purified human primary blood CD14+monocytes or THP-1 monocytes are co-cultured with fluorescent pathogenic mycobacteria in the rationally selected 3D cell culture microplates. On day 3, CD14+monocyte and mycobacteria co-cultures are supplemented with autologous human CD14 PBMC subsets. Alternatively, the ‘mycobacteria-in-spheroid’ co-cultures are generated using the reporter THP-1 cells and mycobacteria to investigate microenvironments and immune mechanisms. 3D tuberculoma-like structures generated with conglomerating granulomas are used for fluorescence-based serial high-content imaging and investigating host-pathogen interactions. Theresultant HTS-compatible bioplatform is utilized to rapidly screen biologies and chemical compounds for quantitative assessment of drug efficacy in terms of bacterial growth inhibition and granuloma resolution, cytotoxicity, and host-directed immune mechanisms of drug actions in situ. The animal model can subsequently validate the identified ‘top hit’ compounds.

[0029] FIG 7B shows the microscopy of co-cultures generated using human THP-1 monocytes (I xlO5) and Mm 1218 GFP (MOI 0.006) demonstrating the 3D spheroid formation (day 2) and mature structure (day 16 and 28) comprising cellular aggregates resembling well-organized, florid granulomatous foci (average diameter 375 pm) in a solid tuberculoma (average diameter 2024 pm). FIG. 7C shows serial imaging over three weeks in the co-cultures generated using Mm strain ‘M’ (tdTomato) or 1218 (GFP) (MOI 0.006). Representative images are from one of the two experiments investigating the growth kinetics of co-cultures generated in 200 pl medium and later supplemented with fresh 50 pl medium (days 6 and 14). FIG. 7D shows Mm ‘M’ (tdTomato) growth in 3D co-cultures measured using relative fluorescence unit (RFU) and colony-forming unit (CFU) counts. FIG 7E shows Host-cell cytotoxicity measured by CytoTox Gio assay in AT / ?? ‘M’ (tdTomato)-infected or uninfected 3D cell cultures. The data in FIG. 7D and 7E are from one of the two experiments (n = 204-282 infected and 18 uninfected spheroids for RFU, n = 6 infected spheroids per time point for CFU, and n = 12 infected and 12 uninfected spheroids per time point for cytotoxicity). Error bars indicate SD. ****p < 0.0001 by 2-way ANOVA with Dunnett’ s test (FIG 7D) and Welch’s t-test (FIG. 7E) compared to uninfected controls. FIG. 7F shows 3D co-cultures of THP-1 monocytes and virulent Mtb strains (tdTomato) using MOI 0.01 (a-d) or attenuated M. bovis BCG strains (WT or GFP) using MOI 0.01 (i and iii) or ten-fold higher MOI 0.1 (ii and iv). FIG. 7G and H show 3D co-cultures of THP-1 monocytes and WT Mm ‘M’ (MOI 0.004), immunostained (red) for Pan-cadherin, E- cadherin (FIG. 7G), or ICAM-1 (FIG. 7H) on day 9. FIG. 71 shows the 3D spheroid culture of human whole PBMCs with or without mycobacterial infection. Representative co-culture image is using Mm 1218 GFP (MOI 0.01). FIG. 7J shows 3D co-culture using MACS-purified primary CD14+blood monocytes (2 x 105 / well) andMtb Erdman (tdTomato) (MOI 0.05) and later supplemented with autologous CD14 PBMC subsets (4 xlO5 / well) on day 3.

[0030] FIG. 8A-D show the ‘mycobacteria-in-spheroid’ 3D co-cultures using human or mouse macrophage cell lines and fluorescent M. marinum strains. FIG. 8 A shows the growth oiMm ‘M’ (tdTomato) and 1218 (GFP) strains in the Middlebrook 7H9 or RPMI-1640 medium at 30 or 37°C measured by spectrophotometry. Data are optical density (O.D.) at 600 nm of axenic cultures(n = 3) grown in 5 ml medium in culture tubes and initiated using 100 pl inoculum (1 x 107CFU / ml) from frozen stocks. Error bars indicate SD. **p < 0.01, ***p < 0.001, and ****p < 0.0001 comparing strain growths at two temperatures by 2-way ANOVA with Tukey’s test for three-week growth curves (black asterisks) or Welch’s t-test for growths at different time points (colored asterisks). FIG. 8B and 8C show growth of Mm ‘M’ (tdTomato) and Mm 1218 (GFP) in 3D co-cultures of monocyte-macrophage cell lines generated at 37 °C in RPMI-1640 medium in the 96-well Coming 3D cell culture microplates. Five different cell lines, THP-1 monocytes (human leukemia), U937 pleural-fluid macrophages (human histiocytic-lymphoma), J744 ascites macrophages (BALB / c mouse reticulum-cell-sarcoma), RAW -264.7 macrophages (BALB / c mouse leukemia virus-transformed), and AMJ2 alveolar macrophages (C57BL / 6J mouse), were investigated. FIG. 8B shows representative images of w-infected or uninfected control 3D spheroids at day 12. FIG. 8C shows fluorescence intensity measured as relative fluorescent units (RFU) demonstrated the background autofluorescence in uninfected 3D spheroids in the GFP channel despite exchanging phenol red RPMI-1640 medium with PBS (pH 7.2). Data are of 5-8 infected and 4 control spheroids per cell line. Error bars indicate SD. *p < 0.05, **p < 0.01, and **** / ? < 0.0001 by Brown-Forsythe and Welch ANOVA multiple comparison tests. FIG 8D shows a linear relationship between the number of red-fluorescent bacteria and RFUs in the 3D spheroids, 7H9 medium, or RPMI 1640 medium. Error bars indicate SD. Pearson correlation coefficients R2and two-tailed P values are shown.

[0031] FIG. 9A-D show that the generation of tuberculoma-like structure with well-organized lesions in the co-culture depends on the cell culture ware and growth conditions. FIG. 9A shows a co-culture of THP-1 monocytes andAT / w ‘M’ (tdTomato) in 96-well 3D and 2D cell culture plates with distinct microwell geometry and cell-culture surface. FIG. 9B shows a co-culture of THP-1 monocytes and Mm ‘M’ (tdTomato) in a polymer-encapsulated 3D ‘Cell-in-a-Box’ system. Miniature granulomatous aggregates rather than well-organized lesions developed in the encapsulated 3D system, regardless of whether THP-1 cells were admixed or infected (overnight) with bacteria before encapsulation in natural cellulose polymer microparticles. FIG 9C shows that the lowest background noise and autofluorescence in the RPMI-1640 alone and uninfected controls were detected in the Coming 3D ultra-low-attachment (ULA) plate, even though a significant reduction in fluorescence was found in rifampicin (lpg / ml)-treated co-cultures relative to untreated or DMSO-treated controls in all plate types. Data are of 3-21 replicates per treatment or control. RFU, relative fluorescence units. FIG. 9D shows that the gradual reduction in bacterial burden with increasing rifampicin dose was best found in the Corning 3D ULA plate. Data are of 21 untreated (0 pg / ml) and 3 treated spheroids per dose. Error bars in FIG 9C and D indicate SD. *p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001 by Brown-Forsythe and Welch ANOVA multiple comparison tests. The low MOI (0.006) was used in all experiments, and representative images (FIG. 9A and B) and fluorescence intensity readings (FIG. 9C and D) were captured on day 12 post-co-culture.

[0032] FIG. 10A-C show that the dynamic granuloma lesions develop the 3D cell culture model and express characteristic markers. FIG. 10A shows spatiotemporal host and bacterial -growth dynamics and granulomatous lesion development in the 3D spheroid co-cultures generated using red-fluorescing THP-1 monocytes and green-fluorescing mycobacteria or green-fluorescing THP-1 cells and red-fluorescing mycobacteria. Examples of coalescing granulomatous lesions (white arrows), lesions that progressed quickly (blue arrows), and those that remained relatively stable (green arrows) are indicated using arrowheads. FIG. 10B shows the macrophage epithelization and adherence junction markers, Pan-cadherin, E-cadherin, N-cadherin, ICAM-1 (all red), and y-catenin (green), detected in nascent granulomatous aggregates by immunostaining on day 10 in the 3D co-cultures of THP-1 cells and WT Mm ‘M’ (MOI 0.004). FIG. 10C shows the 3D co-cultures immunostained with isotype control antibodies for immune target-specific antibodies investigated in the study demonstrated the absence of non-specific staining in the developing granuloma lesions.

[0033] FIG. 11 A-C show that the 3D model expresses classical features and milieus in human tuberculomas and indicates structural organization. FIG. 11 A shows increased expression of angiogenesis marker VEGFR-2 (a), microtubule marker tubulin suggesting cytoskeleton reorganization (b), and immunoregulatory and endomembrane damage marker galectin 9 (c) in the granuloma macrophages as detected by immunostaining (red) in the 3D model of THP-1 monocytes and WT / M ‘M.’ (d) 3D co-cultures of autophagy reporter THP-1 monocytes indicating a block in the autophagy flux (yellow-green puncta) with inhibition of autophagolysosome formation in the permissive macrophages of granulomatous foci, (e) 3D cocultures of inflammasome reporter THP-1 monocytes showing inflammasome activation (green specks) in the framework macrophages, (f) Enlargement of dotted boxes in (e) showing increasedinflammasome activation in epithelioid macrophages of granuloma attempting Mtb (yellow) restraint (box i) or signs of inflammasome and pyroptosis inhibition in the lesion allowing Mtb (red) growth (box ii). (g) Increased staining (red) in the infected spheroid core but not in the zone of granulomatous foci following Lyso-ID dye staining, which stains acidic organelles or environment resulting from cell death. Staining with image-iT hypoxia reagent (h), apoptosis / necrosis detection reagents (i), and live / dead staining reagent (intracellular cell amine sensor dye detecting dead cells) (j) showing increased hypoxia, necrosis, and cell death in the infected spheroid center. 3D co-cultures investigated were developed using THP-1 monocytes and WT Mm ‘M’ (a-c and h-j), WT Mtb Erdman (d and g), or Mtb Erdman tdTomato (e and f), and uninfected spheroids were used as controls. FIG. 1 IB shows the development of cavity-like features in the Mm 1218 (GFP)-infected 3D model of THP-1 cells following deposition of 30 pl but not with 5 pl ECM. FIG. 11C shows treatment on day 6 with anti-TNF-a and CD1 la biologies or mAb biosimilars (62.5 ng) exhibits disruption of organization integrity (blue arrows) in the nascent granulomas on day 10 in theMZ> Erdman (tdTomato) infected 3D model.

[0034] FIG. 12A-B shows that the 3D model expresses key features and microenvironments in human tuberculomas, including mycobacterial biofilm formation. FIG. 12A shows the expression of tubulin (green), VEGFR-2, collagen type-3, and gal ectin-9 (red) determined by immunostaining in the uninfected control 3D spheroids and 3D tuberculomas of THP-1 monocytes and WT Mm M (MOI 0.004). Nuclei are stained blue using Hoechst 33342 dye. The increased expression of macrophage differentiation marker CD68 (red) in the granulomas and the core of the 3D tuberculomas is also shown. FIG. 12B shows AT / A-biofilm formation in the 3D tuberculomas of THP-1 monocytes and / Erdman (WT) as detected by Fun-1 and Cal cofluor white (CW) staining of biofilm-matrix component cellulose. Fun-1, a unique two-color fluorescent probe that readily diffuses into cells, fluoresces in red (Texas red) and green (GFP) channels, and CW labels chitin and cellulose (0 (l,4)-D-glucopyranosyl units) with bluefluorescence. Composite images from red, green (Fun-1), and blue (CW) fluorescence channels overlaid with or without one from the brightfield channel are shown. Except for a diffuse green (background) fluorescence, no fluorescence was detected in the uninfected 3D spheroids.

[0035] FIG. 13A-B show the autophagy and inflammasome activation in the 3D model of THP- 1 reporter cells infected with Mtb Erdman. FIG. 13 A shows autophagy induction in the 3D tuberculomas of THP-1 -Difluo hLC3 reporter cells infected with AM Erdman (WT). The leftpanel shows the basal expression of hLC3-GFP-RFP in the reporter cells grown in the Coming ULA flat-bottom microwell as a 2D cell culture (uninfected). Uninfected 3D cell cultures and tuberculoma co-cultures treated with DMSO (drug carrier), rapamycin (mTOR inhibitor and autophagy inducer), chloroquine (lysosomal inhibitor), and rapamycin plus chloroquine are shown on the right. Untreated or DMSO-treated 3D co-cultures exhibit accumulation of yellowgreen puncta in the granuloma-zone macrophages and numerous cellular clusters on day one and day six post-treatment i.e., day 7 and 12 post-co-culture, respectively, indicating a block in autophagolysosome formation and incomplete autophagy flux. Increased red fluorescence on day six post-treatment indicates delayed autophagy induction (Mtb and starvation-induced) in these 3D co-cultures. On the contrary, rapamycin-treated 3D co-cultures exhibit relatively rapid induction of autophagy (red fluorescence) throughout the 3D co-culture (< day one posttreatment). Chloroquine-treated co-cultures show increased autophagy on day six than on day 1. Chloroquine has been shown to inhibit autophagy by blocking autophagosome fusion with lysosome and slowing down lysosomal acidification. However, chloroquine-induced lysosomal inhibition can inhibit mTORCl and secondarily induce autophagy. It can also induce LC3-II formation independently of autophagy. FIG. 13B shows inflammasome and pyroptosis induction in the 3D tuberculomas of THP-l-ASC-GFP reporter cells infected with Mtb Erdman (tdTomato). The left panel shows the basal expression of ASC-GFP in the reporter cells grown in Corning ULA flat-bottom microwell as an uninfected 2D cell culture. The ASC-GFP expression is driven by NF -KB -inducible promoter. Hence, little GFP signal is detected in resting cells. Uninfected 3D cultures and 3D co-cultures treated with DMSO (drug carrier), nigericin (NLRP3 inflammasome inducer), and MCC950 (NLRP3 inhibitor) are shown on the right. Untreated or DMSO-treated 3D co-cultures show that Mtb (red) induces the assembly of ASC-dependent inflammasomes (green) in THP-1 cells and, eventually, cell death. Nigericin-treated 3D cocultures exhibit ASC-speck (green) formation, relatively rapid macrophage death (pyroptosis), and inhibition o£Mtb (red) growth. Conversely, MCC950 treatment, known to inhibit inflammasomes and ASC speck formation, fails to reduce the Mtb burden in 3D co-cultures by day six post-treatment relative to DMSO. Mtb MOI 0.05 (FIG. 13A-B).

[0036] FIG 14A-C show that the hypoxia and necrosis develop in the center of the 3D tuberculomas. FIG. 14A shows increased hypoxia red staining in the center of 3D tuberculomas of THP-1 monocytes and WT Mtb Erdman (MOI 0.05). 3D tuberculomas or control spheroidswere exposed to pre-titrated hypoxia red fluorogenic probe (ROS-ID Hypoxia / Oxidative Stress detection kit, Enzo Life Sciences) in 100 pl RPML1640 medium and cultured for 24 or 48 hr in the Co2 incubator. After incubation, the RPML1640 medium in the well was carefully exchanged with sterile PBS twice to remove the hypoxia red probe after incubation. The stained 3D spheroids were imaged using the Texas red channel in the Cytation-5 cell imager. Red Hypoxia Detection Reagent (probe) is a non-fluorescent or weakly fluorescent aromatic compound containing a nitro (NO2) moiety. Due to a nitro-reductase activity present in hypoxic cells, the nitro group is converted in a series of chemical steps to hydroxylamine (NHOH) and amino (NH2) group, the original molecule then degrades, releasing the fluorescent probe, which stains hypoxic cells red. Hypoxia inducer deferoxamine (200 pM for 3 or 6 hr) and hypoxia inhibitor nitazoxanide (20 pM for 16 hr) were used as positive and negative controls, respectively. FIG. 14B shows increased cell death (red) in the center of 3D tuberculomas compared to control 3D spheroids. Nuclei are stained blue using Hoechst 33342 dye. 3D spheroids were stained with a Fixable Red Dead Cell Stain kit (Thermofisher) as per the manufacturer’s instruction for 30 min in the RPML1640 medium. To confirm results, live / dead cell staining was also performed using a live / dead cell imaging kit containing Esterase and BOBO-3 iodide. Esterase staining (green) indicates live cells, and BOBO-3 iodide (red) staining indicates free DNA or dead / damaged cells. Apoptotic and necrotic cell staining indicated that the increased cell death in the center of 3D tuberculosis compared to control spheroids is due to increased monocyte / macrophage necrosis (green) rather than apoptosis (red). The staining was performed using the Abeam Apoptosis / Necrosis kit, which includes Apoxin, a phosphatidyl serine (PS) sensor for apoptosis (red), Nuclear Green DCS1, a membrane- impermeable dye, that stains the nucleus of damaged cells or those undergoing necrosis (green) and CytoCalcein dye that stains live cells (blue / violet). FIG. 14C shows increased cell necrosis in the 3D tuberculomas over 12 days of co-culture.

[0037] FIG. 15A-D show that the cavity formation occurs following increased extracellular matrix deposition in the 3D tuberculomas. FIG. 15A shows cavity -like features on day 21 in the 3D co-cultures of THP-1 monocytes and Mm 1218 GFP (MOI 0.005) following the deposition of an increasing amount (5-50 pl) of the extracellular matrix (ECM) mixture containing human collagen type-1 and fibronectin. FIG. 15B shows the kinetics of cavity formation in the 3D cocultures of THP-1 monocytes andA / w 1218 GFP (MOI 0.005). 3D co-cultures deposited with 10or 50 pl of ECM are shown. Cavity formation starts around day 12 and grows over time in the 3D co-cultures deposited with 50 pl of ECM. FIG. 15C shows that the cavitary transformation does not occur in the uninfected control 3D spheroids despite the deposition of a high dose of ECM. Representative 3D cell cultures deposited with 30 pl of ECM are shown. FIG. 15D shows that the cavitary transformation does not occur in 3D co-cultures of THP-1 monocytes generated using 10-fold lower MOI (0.0005, despite the deposition of a high dose (> 30 pl) of ECM.

[0038] FIG. 16A-F show that the screenings of biologies and biosimilars in the 3D co-culture identify immunotherapeutics impacting the structural organization and integrity of developing granulomas. FIG. 16A shows a 3D co-culture of THP-1 and > Erdman (tdTtomato) (MOI 0.01) exhibiting a loss of structural integrity in nascent granulomas by day 10, following anti- CD1 la or anti-TNF but not anti-CD52 or isotype Ab treatment, regardless of the dose used or timing of treatment. The names of biologies or biosimilars investigated are described in the bracket. FIG. 16B shows a 3D co-culture of THP-1 and Mm M (tdTomato) (MOI 0.006) exhibiting the formation of numerous miniature (miliary) granulomas rather than large, well- organized, and compact granulomas by day 12 following anti-TNF adalimumab biosimilar treatment. FIG. 16C-D show that anti-CDl la or anti-TNF treatment in the 3D co-culture of THP-1 and Mtb Erdman (tdTomato) results in the formation of mature granulomas (FIG 16C) and development of bacterial loads (FIG 16D) comparable to isotype Ab treatment or medium alone (no Ab) controls by day 14, despite early interference with structural integrity.Nonsignificant using ANOVA when bacterial burdens (RFU) in anti-CDl la or anti-TNF -treated co-cultures were compared with isotype Ab-treated or medium alone (no Ab) controls. FIG. 16E shows the effect of biologies and biosimilars (n = 15) other than anti-CDl la or anti-TNF on the structural organization of developing granulomas in the 3D co-culture of THP-1 and Mtb Erdman (tdTomato) by day 10. FIG. 16 F shows the effect of biologies and biosimilars (n=l 8) on the bacterial burdens in the 3D co-culture of THP-1 and Mtb Erdman (tdTomato) as measured on day 14. * / ? < 0.05, **p < 0.01, and ***p < 0.001 by one-way ANOVA with Sidak’s post-hoc test.

[0039] FIG. 17A-C show the differentially expressed immunological proteins in the host cell lysates and cultural supernatants of 3D tuberculomas quantified using the Human Kiloplex Proteomics Assay. The culture supernatants and cell lysates from the 3D co-cultures of THP-1 monocytes and Mm 1218 GFP (MOI 0.005) generated with or without the addition of extracellular matrix (ECM) solution (30 pl) were isolated on day 16 and subjected to QuantibodyHumcm Kiloplex Array (QAH-CAA-X00-1) containing human immunological panels with 1000 protein biomarkers. Each sample was analyzed in quadruplicates, and the protein expression levels (pg / ml) in the cell lysate samples were determined following the normalization of amounts to the total protein concentrations present in the lysate. Relative fold change in protein expression level was calculated considering the limit of detection (LOD) and the expression amount in the control and infected samples or samples with or without the ECM addition. The relative fold change of selected protein biomarkers was plotted by dot plot using the ggplot2 package in R. The size of the dots indicates the relative fold change of protein, and the fold change values are shown at the bottom of each dot. FIG. 17A shows the relative fold change in the expression of the select cell lysate proteins in the m-infected 3D co-culture compared to the uninfected control 3D culture. FIG. 17B shows the relative protein expression in the select culture supernatant proteins of the A / m-infected 3D co-cultures compared to the uninfected control 3D cultures. The top panels (in FIG. 17A and B) show proteins with a relative expression fold-change between the range of 50-500-fold, and the bottom panels show the proteins with a relative expression fold-change between the range of 5-50-fold. ECM+indicates 3D co-cultures with the addition of ECM, and ECM indicates 3D co-cultures without ECM. FIG. 17C shows the effect of ECM addition on the protein expression level change in the cell lysates (the left panel) or cultural supernatants (the right panel). In each panel, the left column (titled ‘Naive’) shows the protein expression level fold-change between the uninfected control 3D culture with ECM and the uninfected 3D culture without ECM. The right column (titled ‘Infected’) shows the protein expression level fold-change between the infected 3D co-culture with ECM and the infected 3D co-culture without ECM. In FIG 17 A-C, red dots denote up-regulation, and blue dots denote down-regulation in protein expression.

[0040] FIG. 18 shows the confirmation of the secretion of five selected proteins in the 3D tuberculomas of the Mm 1218-infected THP-1 monocytes quantified using the enzyme-linked immuno-spot (ELISPOT) assay. The spot forming units (SFUs) of 5 key chemokines, cytokine, or MMPs secreted per 105THP-1 cells are plotted, and the representative images of triplicate wells in ELISPOT assay performed on day 16 post-3D culture are shown.

[0041] FIG.19 A-E show that transcriptomic analysis of the 3D co-culture provides insights into the macrophage-mycobacterium interactions leading to granuloma formation. RNAseq was performed on the THP-1 cells isolated from the Mm ‘M’ tdTomato-infected (MOI 0.005) 3D co-cultures or uninfected control 3D cultures on days 3, 6, 9, and 12 post-3D culture. FIG. 19A shows the Principal Component Analysis (PC A) plot for all RNAseq samples using the DESeq2 tools in R. FIG. 19B shows the volcano plots of significantly up- and down-regulated genes in infected 3D co-cultures compared to the uninfected 3D cultures on days 3, 6, 9, and 12 post-3D culture. DEGs: differentially expressed genes. Red dots denote up-regulated DEGs with log2 (foldchange) > 1 and adjusted p- value < 0.05, blue dots denote down-regulated DEGs with log2 (foldchange) < -1 and adjusted / ?- value < 0.05, and dark grey dots denote less significant genes (not DEGs) with adjusted / ?- value > 0.05. The plots were generated using the DESeq2 package in R. FIG. 19C shows the Venn diagrams of the up-regulated and down-regulated DEGs shared on days 3, 6, 9, and 12 post-3D culture. The diagrams were plotted using the online program (bioinformatics.psb.ugent.be / webtools / Venn / ). The shared DEGs (n = 539 in the central of the Venn diagram constituting both up and down-regulated DEGs) were subjected to Gene Set Enrichment Analysis (GSEA) using the ShinyGo (v.0.77, ShinyGO 0.77 (sdstate.edu) . The top select enriched KEGG pathways are presented as a lollipop plot using the ggplot2 package in R. The line or dot’s color indicates the -log2(FDR), the length of the bar indicates the enrichment score, and the size of the dot indicates the number of enriched genes. FIG 19D shows the enriched canonical pathways altered before, during, and after the granuloma formation in the infected 3D co-cultures compared to the 3D controls. The DEGs with absolute |log2(foldchange)| > 1 and adjusted p-value < 0.05 were subjected to gene-set enrichment analysis using the Qiagen IPA canonical pathway analysis tool. The length of the bar in plots represents the Z scores of the most enriched canonical pathways (n=45), ranked based on the score of day 3 post-culture. Bars facing the left denote pathway downregulation, and bars facing the right denote pathway activation. Red circles denote neurological signaling pathways, purple triangles denote cholesterol-related pathways, and blue asterisks denote pathways mainly activated at the early culturing day.

[0042] FIG. 19E shows the enriched KEGG pathways altered before, during, and after granuloma development in the infected 3D co-cultures compared to the 3D controls. The same sets of the DEGs from FIG. 19D were subjected to gene set enrichment analysis (GSEA) using the clusterprofiler package in R. The top 40 most enriched KEGG pathways are presented as bar plots. The length of the bar indicates the normalized enrichment score (NES). Bars facing the left denote pathway inhibition, and bars facing the right denote pathway activation. Red dotsrepresent pathways shared with IPA analysis (FIG. 19D). Blue asterisks denote interesting pathways identified by the GSEA.

[0043] FIG. 20A-G show the schematic workflow of the transcriptomic characterization of the 3D tuberculoma model that was used to generate the data shown in FIG. 19. FIG. 20A shows that the 3D cultures of THP-1 cells were developed with or without Mm ‘M’ tdTomato infection (MOI 0.005), and the total RNA was isolated from THP-1 cells on 0, 3-, 6-, 9-, and 12-days postculture using the Qiagen RNeasy Plus kit following the manufacturer’s instructions. FIG. 20B shows that the cDNA libraries were prepared using the Illumina Truseq stranded mRNA library kit, followed by the Illumina Nextseq sequencing (75><2 cycles). FIG. 20C shows that the RNA sequencing data was quality-checked by the FastQC (v.0.11.5) and trimmed by the Trimmomatic (v.0.39) tools. FIG. 20D shows that the high-quality reads were mapped to the human reference genome (GRCh38pl3) using the STAR aligner software (v.2.5.2b), annotated by the GRCh38.gif file. FIG. 20E shows that the reads mapped to the gene were counted by the featureCounts function in the RSubread package in R. The read count table was generated and normalized by the sequencing depth using the DESeq2 package. Principal component analysis (PCA) plot and sample distance matrix were generated to visualize the clustering of samples. FIG. 20F shows that the differentially expressed genes (DEGs) were identified using the DESeq2 package and visualized by MA and volcano plots. Core DEGs shared by different infection days were determined using a Venn diagram. FIG. 20G shows that the Gene Set Enrichment Analysis (GSEA) was performed using the IPA software (Qiagen. Inc.) or using the clusterprofiler package with the KEGG database in R. Expression changes (log2Foldchange) of core genes in selected pathways are then presented as a heatmap.

[0044] FIG. 21 A and B show additional RNA sequencing data analysis of the 3D cell culture samples. FIG. 21 A shows the heatmap and hierarchical clustering of the 3D co-cultures and uninfected 3D controls on days 0, 3, 6, 9, and 12 post-culture. FIG. 21B shows the MA plot representing the log2 fold-change over the base mean of each gene in the 3D co-cultures compared to the 3D controls. Red dots show up-regulated genes with log2FC >1 and adjusted p- value < 0.05, blue dots indicate down-regulated genes with log2FC < -1 and adjusted p-value < 0.05, and grey dots indicate genes with insignificant change in the expression with adjusted p- value > 0.05.

[0045] FIG. 22A-G show the expression of differentially expressed core genes in the selected pathways. The log2(fold change) in the expression of core genes involved in the pathway is plotted as a heatmap using the ggplot2 package in R. The horizontal four blocks denote the gene expression levels at 3-, 6-, 9-, and 12-days post-culture. The color of the block indicates log2(fold change). Cell adhesion molecule (FIG. 22A), angiogenesis (FIG. 22B), necroptosis (FIG. 22C), autophagy and phagosome (FIG. 22D), inflammasome and pyroptosis (FIG. 2 IE), proteases, cathepsins and metal -metalloproteases (FIG. 2 IF), neuroactive ligand-receptor interaction (FIG. 22G), and chemokine and cytokine (FIG. 22H) pathways are shown.

[0046] FIG. 23 shows that the 3D tuberculoma model can be used to investigate trained immunity induced by a vaccine against secondary pathogenic mycobacterial infections. THP-1 monocytes (1 x 106cells / ml of 3D cell culture medium) were infected with live attenuated BCG Danish vaccine strain (5000 CFU) in the cell culture tubes at 37 °C in a Co2 incubator for 16 hr. The next day, free BCG bacilli were removed from the THP-1 cell suspension by low-speed centrifugation (50 g for 5 min). Infected THP-1 monocytes can be alternatively physically separated from free BCG by MACS column separation using anti-CD32-biotin antibody and anti-biotin Microbeads. BCG-infected THP-1 monocytes (l *107cells / 50 ml of 3D cell culture medium) were cultured in a tissue culture flask for 5 days to induce trained immunity. Control THP-1 monocytes ( I MO7cells) were simultaneously trained with 3D cell culture medium i.e., RPMI-1640 only. BCG-trained or RPMI medium-trained THP-1 cells were next used to develop 3D tuberculomas in a 96-well plate (Corning 3D) with or without infection with M 1218-GFP (MOI 0.005). BCG-trained but not RPMI-1640-trained THP-1 monocytes / macrophages prevented the growth of pathogenic m 1218-GFP and inhibited the development of granuloma lesions (green) as imaged on day 15 post-3D cell culture.

[0047] FIG 24A-D show the identification of HDT compounds inhibiting pathogenic mycobacteria in the 3D tuberculoma bioplatform. FIG. 24A shows potential HDT compounds in the custom-made library were screened at 20 pM in the THP-l-Mm ‘M’ (tdTomato) 3D tuberculomas, and results are expressed as a normalized bacterial burden (%). Data are from three independent experiments (n = 9 tuberculomas / compound and 36-72 tuberculomas / control), and filled circles display bacterial burden in individual tuberculomas. FIG 24B shows the ‘top hits’ identified in chemical screens to reduce bacterial burdens by > 50% in the 3D tuberculomas of four individual Mm and Mtb strains (tdTomato) compared to untreated controls. FIG. 24Cshows the results of screenings in 3D tuberculoma bioplatform investigating inhibitory effects of compounds on the four different mycobacterial strains. Rifampin was used as a positive control for bacterial burden reduction. The average z-score and standard error for each compound and controls screened in 2-4 experiments per mycobacterial strain are displayed by four symbols. The dashed line depicts a cutoff at a z-score of -4. FIG 24D shows the identification of HDT compounds inhibiting Mtb Erdman (tdTomato) in the 3D tuberculomas of human PBMCs (CD14+monocytes supplemented with CD 14 subsets) or THP-1 monocytes. Data are from two THP-1 experiments (n = 6 tuberculomas / compound) and four human donor PBMC experiments (n = 12 tuberculomas / compound). (FIG. 24 A, C, and D) Mean ± SEM is shown. *p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001 compared to DMSO control by Kruskal -Wallis with Dunn’s post-hoc test. The green and black asterisks in FIG. 24D denotes significance levels in THP-1 and PBMC co-cultures, respectively.

[0048] FIG. 25A-C show that the compound screening assay using 3D tuberculoma bioplatform exhibits excellent quality and performance. FIG. 25A shows the impact of Mm ‘M’ (tdTomato) MOI on the screening assay quality in the THP-1 tuberculoma bioplatform as assessed by the Z’- factor statistic. The Z’-factor describes how well separated the positive and negative controls are and indicates the probability of false positives or negatives. It was investigated using assay controls without the intervention of test compounds in the 96-well plate format. One infection dose per plate (Coming 3D) and five doses (MOI range 0.001 to 0.012) were investigated. Nitazoxanide (20 pM / well) and rifampicin (1 pg / ml) were used as positive controls for bacterial reduction, and DMSO and medium only (no drug or untreated) were negative controls in the assay. FI was measured on day 12, and the data shown are from one of the two experiments performed (n = 12 tuberculomas per positive control and 12-24 tuberculomas per negative control per assay plate). The dashed line denotes a cutoff at a Z’-factor of 0.5. The Z’-factor value of 1.0 indicates an ideal assay, values between 0.5 and 1.0 indicate an excellent assay, and values between 0 and 0.5 indicate an acceptable assay. Values below 0 indicate that the assay conditions have not been optimized, and the assay is unlikely to generate useful data. The optimal MOI range identified is 0.007 to 0.012 for the assay using Mm ‘M’ to exhibit an excellent Z’-factor and quality. FIG. 25B shows the investigation of the optimal MOI of Mtb (tdTomato) strains in the THP-1 bioplatform for an excellent screening assay quality. FI measurements on day 14 were considered for slow-growing Mtb strains. The optimal MOI rangeidentified for Mtb strains H37Rv, Erdman, and Beijing is 0.025 to 0.05. Excellent assay quality could not be reached for the M / A CDC1551 strain, even with the higher MOI of 0.05 tested. Increased M / A-induced THP-1 cell death was observed over 14 days in an assay investigating the MOI of 0.1 for virulent Mtb strains; hence, the MOIs > 0.05 were not considered. FIG. 25C shows the performance and quality of the screening assay using 3D tuberculomas of THP-1 cells infected with M / A H37Rv (tdTomato) in several 96-well plates. The assay was performed in 10 plates using an optimal MOI of 0.025, and FI was measured on day 14. Filled circles represent fluorescence in individual 3D tuberculomas or uninfected spheroids. RFU, relative fluorescent units. A total of 126 untreated, 180 DMSO-treated, and 120 each nitazoxanide or rifampicin- treated tuberculomas were investigated. To determine background fluorescence, 54 uninfected spheroids were included. The assay demonstrated excellent quality and performance (Z’ -factor > 0.5). The performance in the individual plates was also assessed, and the Z’-factor > 0.5 in all ten plates further confirms the suitability of an assay for HTS applications.

[0049] FIG. 26A-D show the inhibition of granuloma lesions following HDT compound treatments in the 3D tuberculoma bioplatform of THP-1 infected with Mm or Mtb strains. Sixty- five potential HDT compounds were screened in the 3D tuberculoma bioplatform of THP-1 cells individually infected with M M (FIG. 26A), Mtb H37Rv (FIG. 26B), Mtb Beijing F2 (FIG. 26C) and Mtb Erdman expressing tdTomato as described in FIG. 24 and granuloma numbers in the 3D co-cultures were counted and expressed as normalized lesion counts (%) relative to untreated controls. For granuloma lesion counts, Z-projected images of 3D spheroids were subjected to cellular analysis and granuloma lesion counts using Gen 5 software, followed by manual lesion counts by three blinded readers for quality control, and the average of three readers’ counts was considered. The data are displayed as box plots with whiskers (minimum to maximum), showing all data points (replicates) as circles. The dotted line indicates a 25% inhibition of granuloma lesions. The data are from four (FIG. 26B), three (FIG. 26A and C), and two (FIG 26D) independent experiments performed per strain. Each assay plate contained tuberculomas treated with no drug (untreated, n = 6), DMSO (n = 6), rifampicin (n = 3-6), and test compounds (n = 3 per compound). *p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001 compared to DMSO control by Kruskal -Wallis with Dunn’s post-hoc test.

[0050] FIG. 27 shows representative images of inhibition of granuloma lesions following potential HDT compound treatments in the 3D tuberculomas of THP-1 individually infected withMm o Mtb strains. The 3D tuberculomas were treated with individual compounds on day 6 post- co-culture, and images were captured on day 6 (Mm 3D) or 8 (Mtb 3D) post-treatment using Cytation-5 cell imager.

[0051] FIG 28A-D show that the ‘top hit’ HDT compounds induce crucial innate immune mechanisms in the 3D tuberculomas. Hypoxia, autophagy, and lysosomal acidification induced by (FIG. 28A) non-cytotoxic ‘top hit’ compounds, (FIG. 28B) representative non-cytotoxic ‘nonhit’ compounds, (FIG. 28C) cytotoxic ‘top hit’ compounds, and (FIG. 28D) cytotoxic compounds exhibiting Mycobacterium species-specific heterogeneity in growth inhibition but 25-50% Mtb Erdman growth inhibition in the 3D tuberculomas at 20 pM on day six posttreatment. Hypoxia and lysosomal acidification in the WT Erdman -infected THP-1 tuberculomas were investigated using Hypoxia Red and Lyso-ID Red Reagents. In contrast, autophagy was investigated in the tuberculomas of WT Erdman-infected THP-1 RFP-GFP-LC3 reporter cells. Blue arrows indicate Hypoxia-Red-stained live cellular aggregates distinct from unstained Mtb- permissive lesions (white) in the granuloma zone or the core. Yellow arrows indicate cellular aggregates lacking autophagosome maturation (yellow-green puncta) and lysosomal acidification (no Lyso-ID Red staining).

[0052] FIG. 29 shows that HDT compounds can be classified into functional clusters based on immune mechanisms and therapeutic efficacy in the 3D bioplatform. Compounds belonging to Cluster 1 (cytotoxic ‘top hit’), Cluster 2 (non-cytotoxic ‘top hit’), Cluster 3 (cytotoxic potential ‘hit’), Cluster 4 (non-cytotoxic potential ‘hit’), Cluster 5 (cytotoxic with unclear effect), Cluster 6 (noncytotoxic with unclear effect) and Cluster 7 (insignificant effect) are listed, and the key innate immune mechanisms such as hypoxia, autophagy, and lysosomal acidification induced by the representative compounds are shown.

[0053] FIG. 30A and B show that AT9283, a small molecule and aurora kinase inhibitor, significantly inhibits bacterial burdens and granuloma lesions in the lungs of Mtb Erdman aerosol-infected mice. Seven days after the aerosol Mtb inoculation, treatment was initiated, and mice received a 35 mg / kg dose of AT9283 (test drug) or 10 mg / kg dose of rifampicin (positive control antibiotic) in 150 pl volume of sterile water daily, 5 days per week, for three weeks. Mice thus received a total of 16 doses of AT9283 or rifampicin. Drugs were initially dissolved in DMSO at 200 mg / ml to prepare frozen stocks, and fresh dilutions in distilled water were prepared daily for oral gavage. A group of mice received a drug carrier DMSO diluted in sterilewater and served as a negative control. On day 28, after aerosol infection, mice were euthanized, and left lung homogenates were plated for CFU. The right middle lung lobes were paraformaldehyde-fixed, paraffin-embedded, and tissue sections were subjected to hematoxylin and eosin (H and E) staining and immunostaining forbacteria. FIG 30A shows the reduction in Mtb Erdman CFU burden in the lungs of AT9283 or rifampicin-treated mice compared to untreated and DMSO controls. Data from two independent experiments is shown. *p < 0.05, **p < 0.01 compared to DMSO control by Kruskal-Wallis with Dunn’s post-hoc test. Error bars are standard deviation, and the horizontal line indicates the geometric mean. FIG 3 OB shows decreased granuloma lesion burden in the tissue sections of the right middle lung lobes of AT9283 and rifampicin-treated mice compared to DMSO controls. Blue stars denote granuloma lesions.

[0054] FIG. 31A-D show the workflow for evaluating the therapeutic efficacy of selected HDT compounds in C3Heb / Fej mice in terms of reducing bacterial burdens and tuberculosis granuloma lesions. FIG. 31 A shows the workflow schematic of the evaluation of therapeutic efficacy of selected HDT compounds. FIG. 3 IB shows a Glas-Col inhalation exposure system. Female, 6-8-week-old C3Heb / Fej mice were aerosol -infected with 10-15 Mtb Erdman CFU using the Glas-Col inhalation exposure chamber in the ABSL-3 facility. Seven days postinfection, groups of mice received potential HDT compound (35 mg / kg / day), rifampicin (10 mg / kg / day), no drug, or drug-carrier DMSO (< 5% v / v) in 150 pl of sterile water five days per week for three weeks. Fresh dilutions of drugs prepared daily in distilled water from the frozen stocks (200 mg compound / ml dissolved in DMSO) of drugs were used for oral gavage, and mice received 16 doses of treatment. FIG. 31C shows screening of HDT compounds (n = at least 5-6 mice / group). On day 28 post-infection, mice were euthanized, and left lung and spleen homogenates were plated on Middlebrook 7H10 agar plates. Mtb CFU were enumerated 4 and 8 weeks after plating and incubation. The paraformaldehyde-fixed middle or inferior right lung lobe sections underwent hematoxylin and eosin staining and immunostaining for > bacteria. Serum was isolated from the blood to determine the drug concentrations achieved in the circulation. FIG. 3 ID shows bacterial burden determination in the organ homogenates oiMtb Erdman-infected but no drug-treated control mice. Total lung (right and left lung) homogenates plated from the mice euthanized 24 hr after aerosol infection revealed that an average of 12 CFU were deposited in the lungs of individual mice. The M / 7? Erdman burden in the left and rightlungs of mice euthanized 4 weeks after aerosol infection was comparable (around 107CFU), confirming that both lungs are equally infected after low-dose mycobacterial infection and harbor similar levels of bacterial load at the peak of infection.

[0055] FIG. 32A-B show validation of ‘hit’ HDT compounds identified from the 3D tuberculoma bioplatform screens in the C3Heb / FeJ mouse model oiMtb aerosol infection in terms of reduction of bacterial burdens. FIG. 32A shows Mtb Erdman bacterial burden reduction in the 3D tuberculoma bioplatform after treatment with individual selected potential HDT compounds. The 3D tuberculomas of Mtb Erdman (tdTomato)-infected THP-1 cells were treated on day 6 using a 20 pM dose, and fluorescence intensity (FI) was measured 6 days posttreatment. Results are expressed as a normalized bacterial burden (%) compared to no drug (i.e., medium alone)-treated controls. Data are from two independent experiments (n = 6 tuberculomas per compound, 36 tuberculomas per DMSO, and 72 tuberculomas per no-drug control (not shown)). Filled circles display bacterial burden in individual 3D tuberculomas. *p < 0.05 and 0.0001 compared to DMSO control by one-way ANOVA and Holm-Sidak’s multiple comparison test, ns, nonsignificant. FIG. 32B shows Mtb Erdman (WT) bacterial burden reduction in the left lungs of aerosol-infected C3Heb / Fej mice after treatment with individual potential HDT compounds. The workflow for evaluating the therapeutic efficacy of selected HDT compounds is described above in Fig. 31. Filled circles display bacterial burden in the left lung of individual mice. Mice treated with antibiotic rifampicin were used as a positive control. *p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001 compared to DMSO control by Kruskal -Wallis with Dunn’s post-hoc test.

[0056] FIG. 33 shows the resolution of tuberculous granuloma burden in M. tuberculosis- infected C3Heb / Fej mice lungs following treatment with selected HDT compounds. Reduction in tuberculous granuloma lesion numbers and size in the lungs oiMtb Erdman-infected and individual HDT compound-treated mice compared to DMSO-treated control mice. Middle or inferior right lung lobes of individual mice from each treatment or control group were randomly collected (n = 5-7 lung lobes / group) and fixed in 4% paraformaldehyde in PBS, followed by 80% ethanol. Tissues were processed, embedded in paraffin, sectioned, and stained with hematoxylin and eosin or anti-mycobacterial antibodies. Representative lung lobe images are shown (n = 5-6 mice / group). Lung lobes collected from naive uninfected mice, A-infected butDMSO-treated, and rifampicin-treated AT / A-infected mice served as negative and positive controls.DETAILED DESCRIPTION

[0057] The inventors have developed an in vitro bioplatform that includes ‘mycobacteria-in- spheroid’ 3D co-cultures, analogous to a human tuberculoma with aggregation of granulomas, using Mycobacterium species and human immune cells that can be used in a BSL-2 or BSL-3 level laboratory for screening of new anti-TB modalities. The bioplatform has several surprising advantages over traditional methods, including reduced animal usage, costs, and time and increased throughput and efficiency of screening for potential host-directed therapy (HDT) compounds, antimicrobials (including antibiotics), and biologies in the tuberculoma microenvironments. In addition, the BSL-2 platform reduces the safety concerns for researchers. To accomplish this goal, the inventors implemented a 3D cell culture approach.

[0058] Three-dimensional cell cultures have numerous advantages over traditional 2D cell cultures, in that 3D cell cultures allow for physiologic cell-to-cell contact, the cells interact with extracellular matrix (ECM), there is a diffusion gradient with an increased carbon dioxide and waste and decreased oxygen and nutrients toward the center of the 3D spheroid. All these features are absent in the traditional 2D culture. Since the human TB granuloma is primarily an organized 3D collection of infected macrophages in the lung, the core of the research conducted was to answer the question of whether it would be possible to develop a 3D spheroid using human immune cells infected with Mycobacterium species, which is a BSL-2 organism and that expresses fluorescent protein to aid in the investigations of growth dynamics of pathogen and host cells by imaging in situ.

[0059] The 3D cell cultures of human immune cells and Mycobacterium species disclosed herein form a “tubercle or spheroid” within 3 days of infection. In 2D cell culture, researchers only observed a monolayer of infected immune cells. Researchers observed the green or red-colored, well-organized, compact granulomatous foci formed by the Mycobacterium strain expressing green or red fluorescent protein (i.e., GFP or tdTomato) in the bioplatform 10 to 12 days postinfection of the immune cells. In 2D cell culture, they observed that infected cells formed only loose aggregates and that virulent but not attenuated Mycobacterium strains could form organized granulomatous foci. Like in human TB granulomas, the inventors also observedincreased hypoxia, necrosis, and cavity formation in the 3D bioplatform. They also observed that the increased cell death in the center of the infected 3D spheroid compared to the uninfected control spheroid was due to necrosis rather than apoptosis. Thus, the key features of human TB granuloma were also seen in the 3D in vitro tuberculomas in the bioplatform. Other key features of human granulomas developed in the bioplatform include epithelioid macrophage transformation, upregulation of vascularization and angiogenesis markers, collagen secretion, biofilm formation, matrix metalloproteinase (MMP) activity, and acidosis.

[0060] The 3D bioplatform disclosed herein can be used to screen potential HDTs, biologies, therapeutic vaccines, antimicrobials, and antibiotics. Results of tests conducted show that the 3D bioplatform can indeed be used to screen both host and pathogen-directed therapeutics, considering significantly reduced bacterial burdens and resolution of granulomatous lesions in the 3D spheroids. The available 2D and 3D granuloma models lack the ability to investigate the effects of therapeutics on the resolution of granuloma lesions and cavitary transformations since discrete, well-organized lesions and cavities are not formed in the available models due to their miniature size.

[0061] Screening of an array of FDA-approved drugs in the 3D bioplatforms showed that 9% of drugs reduced the bacterial burden by over 75%, 12% reduced the bacterial burden by 25-75%, and 62% of the drugs did not have any significant effect on the bacterial burden, but 8% increased the bacterial burden by over 25%. The FDA-approved drugs that significantly reduce the bacterial burden and granuloma lesions can be potentially repurposed for the treatment of TB. This bioplatform could potentially be used to screen HDT compounds and other small molecules at a much lower cost, higher efficiency, and faster than current methods.

[0062] Potential uses of the bioplatform beyond HDTs could span the development of treatment regimens against drug-susceptible and resistant Mtb, evaluation of vaccine efficacy in vitro in clinical trials using immune cells of vaccinated volunteers, study of immune mechanisms of protection, study of TB co-infections and co-morbidities in vitro and could also potentially serve as a platform for wide range of granulomatous and other human diseases where cell-to-cell and cell-matrix interactions are important.1. Definitions.

[0063] The terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting. As used in the specification and the appended claims, thesingular forms “a,” “an” and “the” include plural referents unless the context clearly dictates otherwise.

[0064] For recitation of numeric ranges herein, each intervening number there between with the same degree of precision is explicitly contemplated. For example, for the range of 6-9, the numbers 7 and 8 are contemplated in addition to 6 and 9, and for the range 6.0-7.0, the numbers 6.0, 6.1, 6.2, 6.3, 6.4, 6.5, 6.6, 6.7, 6.8, 6,9, and 7.0 are explicitly contemplated.2. Three-dimensional bioplatform

[0065] Provided herein is a 3D bioplatform. The 3D bioplatform may comprise a container, which may comprise at least one well suitable for culturing human cells. The well may comprise immune cells and mycobacteria. In one example, the well comprises a mycobacteria-in-spheroid co-culture comprising the immune cells and the mycobacteria. In another example, the well comprises tuberculoma with granuloma lesions. The immune cells may be human immune cells, which may comprise one or more of monocytes and macrophages. The human immune cells may comprise THP-1 cells, U937 histiocytes, or CD14+monocytes. The human immune cells may comprise peripheral blood mononuclear cells (PBMC), purified immune cell subsets of myeloid and lymphoid origin from PBMCs, BAL, or biopsies. The 3D bioplatform may also comprise supplementary cells, which may comprise one or more of epithelial cells, endothelial cells, and fibroblasts, which may be of either human primary or cell line origin. The immune cells may be obtained from a healthy human. The immune cells may be obtained from a AF / 6-infected, TB- diseased, TB-treated, or TB-vaccinated individual. The immune cells may be obtained from a co- morbid (for example, diabetic) individual or a co-infected (for example, HIV) individual.

[0066] The mycobacteria may comprise A / , marinum (also referred to herein as Mm), M. tuberculosis (also referred to herein as Mtb), or a Mtb complex species. The mycobacteria may express a fluorescent marker, which may be a green fluorescent protein (GFP) or a red fluorescent protein (RFP). The RFP may comprise tdTomato. In one example, the M. marinum is of strain 1218 or strain M. The Mm 1218 may express GFP, and the Mm strain M may express RFP (for example, tdTomato), vice versa, or other fluorescent proteins. The mycobacteria may comprise Mtb, which may comprise one or more of strain H37Rv, strain Erdman, strain CDC1551, strain Beijing F2, and a clinical isolate of a drug-susceptible or drug-resistant Mtb. M. tuberculosis may express a luminescent marker or a fluorescent marker, which may be one or more of GFP, RFP, a far-red fluorescent protein, and another fluorescent protein.

[0067] The mycobacteria may be combined with the immune cells in amount of colony-forming units (cfu). In one example, the mycobacteria comprise M. marinum and the amount for infection is about 600-1200 cfu. The multiplicity of infection (MOI) may be about 0.006-0.012, as measured in cfu per 105live immune cells. The amount may be about 800 cfu (or an MOI of about 0.008). In another example, the mycobacteria comprise M. tuberculosis H37Rv, and the amount for infection is about 1500-5000 cfu (or an MOI of about 0.015-0.05). The amount for infection may be about 2500 cfu (or a MOI of about 0.025). In a further example, the mycobacteria comprise A / , tuberculosis Erdman or Beijing F2, and the MOI is about 0.012-0.05. In one example, the mycobacteria comprise M. tuberculosis Erdman, and the MOI is about 0.05. The ideal MOI may vary with the Mycobacterium species or strains used in 3D co-culture and may be experimentally determined by monitoring the dynamics of bacterial growth and granulomatous lesion formation within the 3D spheroid structures and ascertaining the Z’-factor of a high-throughput screening assay to be performed using the bioplatform.

[0068] The 3D bioplatform container may comprise a microplate, which may be a 3D spheroid microplate. The microplate may comprise at least one well. The well may have a U-bottom. The well may comprise one or more walls comprising polystyrene, which may be virgin polystyrene. The well may be optically clear. The well may comprise a coating, which may comprise a hydrogel. The hydrogel may be hydrophilic, neutrally charged, and biologically inert. The hydrogel may be covalently bonded to at least one surface of the well. The hydrogel may comprise one or more of naturally derived, synthetically derived, and hybrid materials. The hydrogel may comprise one or more of collagen, fibrin and alginate. The hydrogel may be made synthetically using polyacrylamide and polyethylene glycol. The hydrogel may comprise one or more hybrid materials, which may be one or more of hyaluronic acid, a polypeptide, and a polymer. The hydrogel may comprise a perfluorinated polymer, olefin, or a combination thereof. The perfluorinated polymer may comprise poly 4-methylpentene. The hydrogel may be CORNING® PURAMATRIX™ Peptide Hydrogel of synthetic matrix. The PURAMATRIX™ Peptide Hydrogel may contain growth factors, extracellular matrix proteins, and / or other bioactive molecules required for optimal cell culture. CORNING PURAMATRIX Peptide Hydrogel may contain standard amino acids (1% w / v) and 99% water. Under physiological conditions, the peptide component may self-assemble into a 3D hydrogel with a nanometer-scale fibrous structure. In one example, the microplate is a CORNING® Ultra-Low AttachmentSpheroid Microplate. In another example, the microplate is an S-BIO PRIMESURFACE® 3D Culture Spheroid Plate. The microplate may lack a V-shape bottom, M-shape bottom, flat bottom, and may lack microcavities.

[0069] The well may comprise a cell culture medium. The cell culture medium may comprise the Roswell Park Memorial Institute medium (RPMI 1640). RPMI 1640 may comprise the following ingredients.Table 1

[0070] The cell culture medium may further comprise one or more of L-glutamine, heat- activated fetal bovine serum (FBS), sodium pyruvate, and HEPES buffer, and may comprise all of the foregoing. The L-glutamine may be at a concentration of about l-2mM. The FBS may be at a concentration of about 9.63-10% (v / v). The sodium pyruvate may be at a concentration of 0.87-1.0% (v / v). The HEPES buffer may be at a concentration of about 0.87-1.0% (v / v). The cell culture medium may exclude a mitogen or all mitogens. In one example, the excluded mitogen is PMA.

[0071] Each well may comprise a cryopreservative. The cryopreservative may comprise the cell culture medium and dimethyl sulfoxide (DMSO). The DMSO may be at a concentration of about 5% (v / v). The cryopreservative may comprise heat-inactivated FBS and DMSO. The DMSO may be at a concentration of about 5% (v / v).

[0072] In one example, one or more wells of the 3D bioplatform may additionally comprise an extracellular matrix (ECM). The ECM solution mixture may comprise human fibronectin and collagen, which may be Type 1 human collagen. The ECM solution mixture may be added to mycobacteria-in-spheroid co-cultures. The ECM solution mixture added may be 50 pl to 200 pl cell culture medium of the co-culture per well. The human type I collagen may be from VITROCOL®, which may comprise about 97% Type I human collagen, with the remainder being Type III human collagen. The collagen may be at a final concentration of about 0.048 mg / ml or 0.048%. The fibronectin may be at a final concentration of about 0.0008 mg / ml (i.e., 0.8 pg / ml) or 0.00008%.

[0073] The microplate may comprise a body. The microplate body may be opaque and may shield each well from well-to-well crosstalk.

[0074] The 3D bioplatform may comprise one or more additional cell types, which may comprise A549 human lung epithelial cells, HUVEC-1 human fetal endothelial cells, HULEC human lung endothelial cells, and MRC-5 human lung fibroblasts.

[0075] As an alternative or in addition to the mycobacteria, the 3D bioplatform may comprise a virus and the human immune cells. The virus may be one or more of human immunodeficiency virus (HIV), an influenza virus, and a coronavirus. The coronavirus may be severe acute respiratory syndrome (SARS)-CoV-2.3. Methods of making 3D bioplatforms

[0076] Provided herein is a method of making a 3D bioplatform described herein. The method may comprise co-culturing the immune cells and the mycobacteria. Before co-culturing, the immune cells may be suspended in a 3D cell culture medium. The 3D cell culture medium may comprise the RPMI 1640 medium as described herein, which may be supplemented with one or more of L-glutamine, FBS, sodium pyruvate, and HEPES buffer as described herein. The 3D cell culture medium may comprise an antibiotic, which may be one or more of penicillin and streptomycin, during growth but may not contain antibiotics during co-culture with mycobacteria. The immune cells may have a threshold viability of at least 90, 91, 92, 93, 94, 95, 96, 97, 98, or 99% viability. Cell viability may be measured by using a trypan blue dye exclusion method. In one example, the optimal threshold viability is at least 95%. The immune cells may have a final concentration of about 1x106live cells / mL. The immune cells may be washed 1, 2, 3, 4, or more times, thereby removing the antibiotic, which may be 4 times. The washing may comprise centrifuging the immune cells and resuspending them in a cell culture medium. The washed immune cells may have the threshold viability.

[0077] If the immune cells are PBMCs, which may comprise CD14+blood monocytes, then the PBMC monocyte subset may be purified and enriched before being combined with the mycobacteria, as described below. In one example, the PBMCs are isolated from human blood and purified by positive selection using a magnetic-activated cell sorting (MACS) method. The MACS method may comprise contacting a cell suspension containing PBMCs with magnetic nanoparticles. The nanoparticles may comprise anti-CD14 antibodies. The CD14+monocytes (bound to nanoparticles) may be separated from a lymphocyte-rich cell fraction comprising lymphocytes and CD14’ mononuclear cells, by running the cell suspension over a column placed in a magnetic field. The magnetic field may capture the nanoparticles and allow the lymphocyterich fraction to pass through. The CD14+monocytes can be subsequently purified by removing the washed column from the magnetic field. In a further example, if the immune cells are PBMC-purified subsets, then the mycobacteria may comprise Mtb. he Mtb may be A / / / ? Erdman and may be combined with the PBMC at about 5000 cfu or an MOI of about 0.05. In one example, the lymphocyte-rich cell fraction purified from the autologous PBMCs may be added to mycobacteria-in-spheroids comprising CD14+blood monocytes and Mtb, and the resulting 3D cell culture may be used in a screening method disclosed herein.

[0078] The method may further comprise adding the mycobacteria to the immune cells in a cell culture medium. The mycobacteria may be added in an amount or MOI disclosed herein. The mycobacteria may be suspended in the 3D cell culture medium. Before being suspended, the recombinant mycobacteria may be grown in a bacterial growth medium containing an agent that put selection pressure on the plasmid expressing the fluorescent marker. The agent may be an antibiotic, which may be hygromycin or kanamycin. The bacterial growth medium may be Middlebrook 7H9 broth, which may comprise 7H9 broth and Tween-80, which may be at a concentration of 0.05% (v / v). The bacterial growth medium may also have a pH of 6.8-7. The bacterial growth medium may also be supplemented with ADC, which may be at a concentration of 10% (v / v), and glycerol, which may be at a concentration of 0.4% (v / v). ADC may comprise 2.5 g bovine albumin fraction V, 1 g dextrose, and 0.0015 g catalase per 50 mL of ADC. After being suspended in the 3D cell culture medium, the suspension may be passed through a 25-27G needle 10-15 times to generate a single-bacterial-cell suspension. The mycobacterial suspension may be diluted with the 3D cell culture medium to the amount or MOI disclosed herein.

[0079] The immune cell suspension may be added to one or more wells of a container disclosed herein. About 105live cells may be added to each well. The volume of cell suspension may be a minimum of 100 pl. The mycobacterial suspension may be added to one or more wells of the container, which may be in an amount or MOI disclosed herein. The volume of mycobacterial suspension may be a minimum of 100 pl. The immune cell suspension and mycobacterial suspensions may be mixed after adding to the well, and the sterile 3D cell culture media or cell culture grade water may be added to periphery wells to avoid drying or boundary effect. The container may be incubated at 37°C with 5% CO2 and 100% humidity. The incubation may be for enough time for mycobacteria-in-spheroid structures to form. The incubation may be between 6-14 days for tuberculoma-like feature formation in the mycobacteria-in-spheroid co-cultures. If the mycobacteria-in-spheroid culture is intended to be incubated for more than 6 days, the addition of fresh RPMI medium in a minimum 50 pl volume is required once per week. The cfu for each well may be measured, which may be by plating dilutions of the mycobacteria inoculum on agar plates, which may be Middlebrook 7H10 agar plates, at 30°C for 12-14 days if the mycobacteria are Mm or at 37°C for 3-4 weeks if the mycobacteria x Mtb.

[0080] In one example, the ECM is added to the mycobacteria-in-spheroids. The ECM may be in solution as described herein and may be added about 3 days after the immune cells and the mycobacteria are co-cultured.

[0081] The 3D bioplatform comprising mycobacteria-in-spheroids may be frozen using the freezing medium containing cryopreservative. In one example, within 30 minutes of co-culturing mycobacteria and immune cells as described herein, the microplate is frozen, which may be by placing it at -80°C. If the microplate is intended to be frozen for longer than 72 h, then the microplate may be placed at -160 to -196°C, which may be in liquid nitrogen. In another example, after combining immune cells and mycobacteria, the co-culture may be incubated at 37°C with 5% CO2 and 100% humidity for 16-72 h, 3D cell culture medium may be removed and replaced with a freezing medium containing cryopreservative as described herein, and then the microplate may be frozen as described herein. The frozen 3D bioplatform may be thawed, the cryopreservative removed, the freezing medium replaced with fresh 3D cell culture medium at least 3 times and incubated in fresh 3D cell culture medium as described herein. The 3D bioplatform may then be used in a screening method described herein.4. Methods of screening

[0082] Provided herein is a method of screening for a candidate molecule, which may be capable of treating or reducing a mycobacterial infection or one or more pathology features thereof. The method may comprise contacting or exposing a 3D bioplatform disclosed herein with one or more candidate molecules or serial dilutions thereof. The method may comprise measuring one or more characteristics of mycobacteria-in-spheroids formed by a mixture of immune cells and mycobacteria of the 3D bioplatform. The characteristic may be the amount of fluorescence produced by the mycobacteria, which may be fluorescence intensity. The amount of fluorescence may be indicative of bacterial burden. In one example, the characteristics comprise loss of integrity of granulomas. In another example, the characteristics comprise one or more of reduced bacterial growth and granuloma formation. In one example, the characteristics comprise a reduction of granuloma number and size. In one example, the characteristics comprise autophagolysosome formation or autophagy flux. A change in the one or more characteristics compared to a control may be indicative that the molecule is capable of treating or reducing a mycobacterial infection or one or more pathology features thereof. In one example, the control is a negative control, which may comprise a drug carrier alone. In another, the control is a positivecontrol, which may comprise an agent known to treat or reduce a mycobacterial infection, or one or more pathology features thereof.

[0083] A decrease in mycobacterial burden in the 3D bioplatform by a threshold value may be indicative of a molecule capable of treating or reducing mycobacterial infection. The threshold value may be greater than 20, 25, 30, 35, 40, 45, 50, 55, 60, 65, 70, 75, 80, 85, 90, or 95%. In one example, the threshold is >25%. In one example, the threshold value may be z score less than -2. In one example, the threshold of z score is <-4. The MOI for the mycobacteria in the 3D bioplatform used to screen may be 0.007-0.012, which may be used when the 3D bioplatform comprises Mm and THP-1 cells; or 0.012-0.05, which may be used when the 3D platform comprises Mtb H37Rv, Erdman, or Beijing F2 and THP-1 cells. In one example, the MOI is 0.025, which may be used when the 3D bioplatform comprises Mtb H37Rv. In one example, the MOI is 0.05, which may be used when the 3D bioplatform comprises Erdman or Beijing F2.

[0084] In another example, the 3D bioplatform may be contacted with or exposed to one or more candidate molecules or serial dilutions thereof, and the number and amount of one or more genes upregulated or downregulated or one or more proteins expressed by the cells of the 3D bioplatform may be measured, as compared to a control. Each of the one or more genes or proteins may be related to cell activation and differentiation, signal transduction, cell adhesion, granuloma and cavity formation, angiogenesis, hypoxia, cytokines and chemokines, immune response, autophagy, necrosis, necroptosis, pyroptosis, and other cell death pathways. The control may comprise uninfected control human immune cells. The 3D bioplatform cells may or may not have been contacted with ECM. The protein expression levels may be measured from a cell lysate or culture supernatant. The gene expression levels may be measured using cell- purified RNA.

[0085] In a further example, expression levels of one or more genes of the 3D bioplatform may be quantified after having been contacted by a candidate molecule, as compared to a control.

[0086] Each candidate molecule may be diluted in the 3D cell culture medium and may be added to mycobacteria-in-spheroids. Each dilution of the candidate molecule may be added about 0-6 days after the mycobacteria and the immune cells have been co-incubated. In one example, the molecule is added at 0 days. In another example, the molecule is added at 6 days.

[0087] One or more characteristics of the mycobacteria-in-spheroids may be observed or measured about 1-14 days after each candidate molecule is added. In one example, one or morecharacteristics are observed 6 days after each candidate molecule is added. In another example, observations are made on days 7, 9, and 12. Granulomatous lesion formation may be observed, which may be by using a cell imaging reader or an inverted microscope. The cell imaging reader may be capable of automated microscopy and microplate detection. The cell imaging reader may be a Cytation 5 imager. Fluorescence intensity produced by the mycobacteria may be measured, which may be by using the cell imaging reader. A difference in the one or more characteristics of the mycobacteria-in-spheroids as compared to a control may be indicative that a candidate molecule has a therapeutic effect against the mycobacteria. The control may be a negative control, such as DMSO, or may be a positive control, such as a molecule known to treat mycobacteria. a. Methods of identifying the mechanism of action of host-directed drugs, biologies, and immunotherapeutics in 3D tuberculomas in situ(1) Methods of screening for the effects of a candidate molecule on autophagy in 3D tuberculomas

[0088] Provided herein are methods of screening for autophagy induction or inhibition of 3D tuberculomas by a candidate molecule.(a) First exemplary method

[0089] The method may comprise a 3D bioplatform comprising a co-culture of THP-1 autophagy reporter cells, which may comprise THP-1 hLC3-GFP-RFP cells, and wild-type Mtb or Mm (nonfluorescent) mycobacteria. The method may comprise contacting the 3D bioplatform on day 6 after co-incubation with the candidate molecule, and imaging (which may be using Cytation 5) on days 7, 9, and 12 after the contacting with the candidate molecule to determine the autophagy flux. The reporter cells may express human autophagy protein LC3B fused with two fluorescent proteins, GFP and RFP. The GFP (acid-sensitive) is quenched or degraded in the acidic environment of autophagolysosome, whereas RFP remains stable. A block in the autophagy process results in higher yellow signals (red plus green co-localization) compared to autophagy induction (red signals) in 3D tuberculomas. Thus, the red signal indicates autophagolysosome formation or induction of autophagy, while the yellow signal indicates incomplete autophagy with autophagosome formation without fusion with acidic lysosomes. Both yellow and red puncta are reduced during the inhibition of autophagy.(b) Second exemplary method

[0090] The method may comprise a 3D bioplatform comprising a co-culture of THP-1 cells (or human monocytes) transfected with an autophagy sensor LC3B-RFP (BacMam 2.0). The 3D bioplatform may comprise the transfected monocytes and WT Mtb o Mm (nonfluorescent). The 3D bioplatform may be contacted with a candidate molecule, or serial dilutions of the candidate molecule, on day 6 after co-culture. The contacted 3D bioplatform may be imaged using Cytation 5 on days 7, 9, and 12 after the contacting to determine the autophagy flux. The red signal indicates autophagolysosome formation or induction of autophagy.

[0091] The commercially available BacMam technology is based on an insect virus (baculovirus) to help efficiently deliver and express genes in mammalian cells. The baculovirus has been modified to include an expression cassette for transgene expression in mammalian cells. BacMam 2.0 incorporates elements that help greatly enhance transduction efficiency and expression levels: a pseudotyped capsid protein for more efficient cell entry and genetic elements (enhanced CMV promoter and Woodchuck Post-transcriptional Regulatory Element), that boost expression levels. Baculoviruses do not replicate in mammalian cells and thus have an excellent safety profile and lack cytopathic effects on cells.(c) Third exemplary method

[0092] The method may comprise a 3D bioplatform comprising a co-culture of THP-1 monocytes and Mtb o Mm (nonfluorescent or red fluorescent). The 3D bioplatform may be contacted with a candidate molecule or serial dilutions thereof, on day 6 after co-culture. Autophagy may be detected using the Enzo Life Sciences CYTO-ID® Autophagy Detection Kit. The kit measures autophagic vacuoles and monitors autophagic flux in lysosomally inhibited live cells using a novel dye that selectively labels accumulated autophagic vacuoles. The 488nm- excitable green dye has been optimized through the identification of titratable functional moieties that allow for minimal staining of lysosomes while exhibiting bright fluorescence (green) upon incorporation into pre-autophagosomes, autophagosomes, and autolysosomes (autophagolysosomes). The kit also includes the Hoechst 33342 dye for the nuclear staining, an Autophagy Inducer (Rapamycin), and a Lysosomal Inhibitor (Chloroquine).(2) Methods of screening the effects of candidate molecules on inflammasomes and pyroptosis induction or inhibition in 3D tuberculomas

[0093] Provided herein is a method of screening for the effects of a candidate molecule on inflammasomes and pyroptosis induction or inhibition in 3D tuberculomas. The method may comprise using a 3D bioplatform comprising a co-culture of THP-1 inflammasome reporter cells (THP-1 LC3-GFP cells) and WT Mtb (nonfluorescent). The 3D bioplatform may be contacted with the candidate molecule on day 6 after co-culture and may be imaged using Cytation 5 on days 7, 9, and 12 after the contacting to determine the effect on AT / A-induced inflammasomes, ASC-GFP speck formation, and pyroptosis in 3D tuberculomas. In these reporter cells, Mtb infection or antigen exposure leads to ASC-GFP expression and ASC-speck formation following inflammasome activation. Exposure to test molecule can either increase the inflammasome activation, speck formation, and pyroptosis (cell death) or reduce the inflammasome activation and pyroptosis, leading to increased cell survival. Increased inflammasome activation in macrophages is thought to be associated with reduced bacterial growth.(3) Methods of screening the effects of candidate molecules on hypoxia induction or inhibition in 3D tuberculomas

[0094] Provided herein is a method of screening for the effects of a candidate molecule on hypoxia induction or inhibition in 3D tuberculomas. The method may comprise using a 3D bioplatform comprising a co-culture of THP-1 monocytes and WT Mtb o Mm (nonfluorescent). The 3D bioplatform may be contacted with a candidate molecule or serial dilutions thereof, on day 6 after co-culture. Hypoxia may be detected using the Enzo Life Sciences ROS-ID Hypoxia / Oxidative Stress detection kit. 3D tuberculomas or control spheroids may be exposed to pre-titrated hypoxia red fluorogenic in 100 pl RPML1640 medium on day 12 and cultured for 24 hr in a CO2 incubator. After incubation, the RPML1640 medium in the well may be exchanged with sterile PBS twice to remove the hypoxia red probe after incubation. The stained 3D spheroids may be imaged using the Texas red channel in the Cytation-5 cell imager. Red Hypoxia Detection Reagent (probe) is a non-fluorescent or weakly fluorescent aromatic compound containing a nitro (NO2) moiety. Due to a nitro-reductase activity present in hypoxic cells, the nitro group is converted in a series of chemical steps to hydroxylamine (NHOH) and amino (NH2) group, the original molecule then degrades, releasing the fluorescent probe, which stains hypoxic cells red in 3D tuberculomas. Induction of hypoxia and HIF-1 in infectedmacrophages by test molecule can activate an anti-mycobacterial response, but increased hypoxia can result in increased cell death (by necrosis).(4) Methods of screening for the effects of a candidate molecule on lysosomal acidification

[0095] Provided herein is a method of screening for the effects of a candidate molecule on lysosomal acidification. The method may comprise a 3D bioplatform comprising a co-culture of THP-1 monocytes and WT Mtb ox Mm (nonfluorescent). The 3D bioplatform may be contacted with a candidate molecule or serial dilutions thereof, on day 6 after co-culture. Lysosomal acidification on day 12 after the contacting may be detected using the Enzo Life Sciences Lyso- ID Red cytotoxicity kit. The cell culture medium may be aspirated from 3D tuberculomas or control spheroids. They may be washed once by exchanging medium with the assay buffer provided in the kit. The 3D tuberculomas and control spheroids may be exposed to 50 pl dual color detection reagent for 2 hr. Detection reagent may be aspirated, and 3D tuberculomas or control spheroids may be washed with 100 pl assay buffer. Spheroids may be suspended in 50 pl assay buffer and imaged. The red lysosome stain can be read with a Texas Red filter (Excitation 540 nm, Emission 680), and the blue nuclear counterstain can be read with a DAPI filter (Excitation 340, Emission 480). The detection reagent includes a unique drug-like dye that rapidly partitions into cells and labels acidic organelles (red) and is suitable for monitoring the accumulation of lysosomes and lysosome-like structures in live cells.

[0096] The dye also fluoresces in the acid environment in the center of 3D tuberculomas generated by increased cell death (necrosis) and release of acidic organelles. b. Methods of identifying host cellular toxicity caused by a candidate molecule in 3D tuberculomas

[0097] Provided herein is a method of screening the effects of a candidate molecule on cytotoxicity in 3D tuberculomas.(1) Method one: CytoTox-Glo Cytotoxicity Assay

[0098] The method may comprise using a 3D bioplatform comprising a co-culture of THP-1 cells and assessing cytotoxicity using a CYTOTOX-GLO assay (G9291, Promega, Mannheim, Germany). This assay measures the dead-cell protease activity released from cells that have lost membrane integrity. The method may comprise a 3D bioplatform comprising a human immune cell-mycobacterium co-culture in a U-shape, clear bottom, 96 well plate (Coming). The 3D bioplatform may comprise a 3D co-culture of THP-1 monocytes or human CD14+monocytes incombination with other PBMC subsets, and Mtb o Mm (fluorescent or nonfluorescent). The 3D bioplatform may be contacted with a candidate molecule or serial dilutions thereof on day 6 after co-culture. The cytotoxicity may be measured on days 6, 9, 12, and 14. During the assay, each 3D spheroid (200 pl volume) may be dissociated and re-suspended by pipetting, split into 2 equal volumes (100 pl each), and transferred into two separate wells (well A and B) in a new white bottom luminescence plate. The 50 pl of Lysis Reagent (digitonin 30 pg / ml) may be added to well A, and 50 pl of RPMI medium may be added to well B. 50 pl of prewarmed CYTOTOX- GLO reagent may be added to all of the wells in the plates (1 :4 dilution), and cells may be incubated for 15 min at room temperature with orbital shaking at 700-900 rpm. The luminescence may be measured using BioTek Cytation 5 Cell Imagining Multimode Reader. Control samples were included in every plate, with a 2-fold serial dilution of freshly grown THP1 cells (starting from 5 x 106cells per well) to obtain the standard curve. The total number of cells may be calculated from well A, and the total number of dead cells may be calculated from well B, using a standard curve obtained from control wells. Data was analyzed by Microsoft Excel.

[0099] Cell cytotoxicity(2) Use of THP-l-RFP cells

[0100] The method may comprise measuring the cytotoxicity of a candidate molecule or serial dilutions thereof, by determining the effects of the candidate molecule on the host cell viability and finding the cell cytotoxicity 50 (CC50) and effective concentration 50 (EC50) values of the candidate molecule. The method may comprise a 3D bioplatform comprising a co-culture of THP-1 cells that express RFP, and WT Mtb o Mm (nonfluorescent). The 3D bioplatform may be contacted with the candidate molecule or serial dilutions thereof on day 6 after co-culture. The THP-1 red fluorescence (excitation 588 and emission 633nm) may be measured on days 6, 9, 12, and 14 after the contacting. Untreated or DMSO-treated 3D tuberculomas may be used as controls. The amount of red fluorescence indicates live THP-1 cells, and loss of red fluorescence indicates dead cells. Cytotoxicity (%) caused by test molecules in 3D tuberculomas relative to no treatment or DMSO treatment was calculated.5. Methods of treatment

[0101] Provided herein is a method of treating or reducing a mycobacterial infection, which may be tuberculosis. The method may comprise treating or reducing tuberculoma lesions. The method may comprise administering an anti -mycobacterial agent to a subject in need thereof. Also provided herein are the anti-mycobacterial agent for use in treating or reducing the tuberculoma lesions, or use of the anti-mycobacterial agent in the manufacture of a medicament for treating or reducing the tuberculoma lesions. The anti-mycobacterial agent may comprise one or more of a tumor necrosis factor (TNF) blocker, such as an anti-TNF antibody; a CD1 la blocker, such as an anti-CDl la antibody; an anti-vascular endothelial growth factor (VEGF) blocker, such as an anti-VEGF antibody; an integrin a4p7 blocker, such as an anti-integrin a4[37 antibody; a CD30 blocker, such as an anti-CD30 antibody; an insulin growth factor-1 receptor (IGF1R) blocker, such as an anti-IGFIR antibody; an IL-6Ra blocker, such as an ani-IL-6Ra antibody; an IL- 10 blocker, such as an anti-IL-ip antibody; an IL-1R blocker, such as an anti-IL-lR antibody; compound AT9283; auranofin; clemastine fumarate; chlorpromazine hydrochloride; 3'4' dichlorobenzamil HCL; dovitinib; doxycycline; H89; lansoprazole; metixene hydrochloride; nitazoxanide; zuclopenthixol dihydrochloride; GW5074; simvastatin; vorinostat; all-trans- retinoic acid; pazopanib; tin protoporphyrin IX; fluoxetine hydrochloride; dasatinib; GW 5074; SRT 1720; gefitinib; loperamide; ezetimibe; and active metabolites of the above compounds. In one example, the anti-mycobacterial agent comprises AT9283.

[0102] In one example, the anti-mycobacterial agent comprises an antibody. The antibody may be an anti-CDl la, anti-a4 7 integrin, anti-CD30, anti-IGFIR, or anti-IL-6R antibody. The antibody may be monoclonal and may be humanized. The anti-mycobacterial agent may comprise a biosimilar to the antibody. In another example, the anti-mycobacterial agent comprises a compound. The compound may comprise AT9283, Tizoxanide (an active metabolite of nitazoxanide), Dasatinib, Quinacrine diHCL, All-trans-Retinoic acid (ATRA), Vorinostat, Sitagliptin, H89, or Lansoprazole. The antibody, biosimilar thereof, or compound may kill mycobacteria and reduce granuloma lesions by activating host cellular mechanisms in 3D tuberculomas. The anti-mycobacterial agent may be used to treat mycobacterial infection or tuberculosis disease.

[0103] The anti -mycobacterial agent may comprise one or more of host-directed molecules or biologies acting via host cell together with one or more antibiotics or other antimicrobials acting directly on mycobacteria.

[0104] The present invention has multiple aspects, illustrated by the following non-limiting examples.Example 1 Three-Dimensional Bioplatform

[0105] This example demonstrates a method of making a 3D bioplatform disclosed herein, and methods of its use in screening.Abstract

[0106] Tuberculomas are the conglomeration of tuberculous granulomas into structurally organized 3 -dimensional (3D) masses that result from Mycobacterium tuberculosis infection and are one of the more severe forms of tuberculosis (TB). Several in vitro models that mimic human TB granulomas have been reported to date to decipher complex host— A 7 / ? biology and discover new prophylactic and therapeutic interventions. However, these models lack well-organized granuloma lesions and a classic tuberculoma structure. Furthermore, they are impractical for screening large compound libraries owing to their low throughput, limited scalability, batch-to- batch variability, and high cost. Here, we describe a ‘mycobacteria-in-spheroid’ co-culture workflow in a standard 96-well -plate format that generates a robust 3D cell culture model. This model reproduces key attributes and microenvironments in human tuberculomas and can be scaled up as a high-throughput screening bioplatform. The tuberculoma-like structures generated encompass well-organized, florid granulomatous lesions and exhibit solid, necrotic, and cavitary morphologies not previously described in 3D cell culture models of TB. This model can be developed using freshly isolated primary human blood monocytes or a monocytic cell line with virulent mycobacteria. The platform combines the entire workflow from generation to imaging of tuberculoma-like structures in situ. It permits serial quantitation of drug efficacy and observations of resolution of lesion size and numbers following a single treatment over the course of several days to weeks. A methodology to adopt this workflow as a cryo-preserved bioplatform for potential commercial use is also described. The ease of generation, reproducibility, pliability, and cryo-shelf stability of the bioplatform make it ideal for HTSapplications and suitable for easy scale-up and implementation in the discovery programs of tuberculosis and other granulomatous diseases.Background

[0107] Tuberculosis (TB), an airborne disease caused by Mycobacterium tuberculosis Mtb is among the leading infectious killers worldwide. In 2021, an estimated 10.6 million people fell ill, and 1.6 million died from TB (WHO Global Tuberculosis Report, 2022). TB is now projected to kill almost twice as many people as COVID-19 every day (Bagcchi, 2023). A histopathological hallmark of TB is the formation of dynamic, spatially organized, multicellular clusters called granulomas. These macrophage-rich structures serve as the niche ioxMtb growth and dissemination while providing the environment where infected macrophages interact with other recruited cells to “wall off’ and fight the offending pathogen (Ramakrishnan, 2012; Pagan and Ramakrishnan, 2018; P. Elkington et al., 2022). One of the more severe clinical manifestations of TB is the formation of tuberculomas, which conglomerate tuberculous granulomas into well- circumscribed masses, most often in the lungs and brain, resembling cancer tumors in these organs (Culver etal., 1950; Moyes, 1951; Sochocky, 1958; Nicolls etal., 2005; Monteiro etal., 2013). Tuberculomas exhibiting a severe morphological form are present in around 5-10% of pulmonary TB patients (Lee et al., 2004; Wetscherek etal., 2022). In pulmonary TB patients, granulomas are highly polymorphic and exhibit a spectrum of structures, including solid, hypoxic, necrotic, and cavitary transformations, and display complex morphologies in the lung sections (Cadena et al., 2017; Wells et al., 2021; Sawyer etal., 2023). Although 2-5 mm in diameter spheroid granuloma nodules are common in the lungs, the tuberculoma structure varies between a few mm to > 10 cm in size (Lee et al., 2004; Wells et al, 2021). The cavitary transformation in tuberculomas increases the risk of person-to-person transmission. In addition, it is associated with poor treatment outcomes, relapse, and the likelihood of developing drug resistance in pulmonary TB (Urbanowski et al., 2020).

[0108] The recent Covid- 19 pandemic has dramatically interrupted and reversed the positive trends of declining global TB incidence and mortality attained during the past decade. It is still negatively affecting TB diagnosis and care. This setback in TB control efforts is predicted to translate into long-term increases in TB-related deaths and delay the United Nations Sustainable Development Goal target of ending the TB endemic by 2030 (WHO Global Tuberculosis Report, 2022) (Pai et al., 2022; Bagcchi, 2023). To meet the goal of TB elimination, improved andshorter treatment regimens for drug-resistant and susceptible forms of TB are urgently required. Such regimens may arise from a better understanding of the efficacy of new therapeutics within the range of granuloma forms and microenvironments. A growing number of preclinical and clinical studies highlight the importance of designing novel pathogen-targeted and host-directed therapies (HDTs) that readily penetrate and act within the granuloma environment (Prideaux et al., 2015; Strydom etal., 2019; Larkins-Ford etal., 2021; Dartois and Rubin, 2022; Wallis et al., 2022). Therapeutics that modulate host-A / interactions in granulomas can be identified using animal models and in vitro cell culture systems. Several 2-dimensional (2D) and 3-dimensional (3D) in vitro cell-culture models that mimic nascent TB granulomas have been described in recent years using Mtb infection of primary human cells (Puissegur et al., 2004; Birkness et al., 2007; Guirado et al, 2015; Kapoor etal., 2013; Tezera etal., 2017a; Berry et al., 2020; Kotze et al., 2021). Although helpful in screening a limited number of compounds and probing b-host interactions (Bielecka et al., 2017; Tezera et al., 2017b), these models suffer from low throughput, limited malleability, and restricted scalability (P. Elkington et al., 2019; Arbues et al., 2020). They form small granulomatous aggregates of macrophages but lack organized granuloma lesions and relevant tuberculoma size, structures, and forms.

[0109] Consequently, physiological gradients of nutrients, oxygen, pH, and pertinent microenvironments present in heterogeneous TB lesions are either absent in these models or not comprehensively reproduced. Critical features like hypoxia and / A dormancy observed in some of these 3D models with miniature granulomas primarily result from the encapsulation of macrophages in microspheres and embedding in the extracellular matrix (ECM) rather than the granuloma structures themselves (Colom et al., 2014; Figueiredo et al., 2018; Arbues et al., 2021). The lack of a continuous influx of immune cells in these models makes maintaining dynamic structures and the extended prolongation of experiments challenging. Using animal models for such screening is costly, time-consuming, and limits the number of compounds screened in a high-containment facility. A high-throughput screening (HTS)-compatible and widely applicable bioplatform that reproduces crucial features and microenvironments in TB lesions and allows serial multiparametric readouts of host and pathogen physiology and spatiotemporal existence is highly desirable.

[0110] The protocol described here provides simple workflows (Figure. 1) to develop an HTS- compatible bioplatform using human monocytes and virulent Mycobacterium strains expressingbright-red fluorescent protein (tdTomato) for the fluorescence intensity and image-based assessment of drug efficacy in the 3D cell-culture microplates. The ‘mycobacteria-in-spheroid’ co-culture generated in microwells consistently mimics 3D human tuberculoma structure and size. Since macrophages are cardinal to the core-scaffold formation, which shapes host immune responses and therapeutic access in tuberculous granulomas (Cronan, 2022), we employed human THP-1 monocytes or peripheral blood mononuclear cells (PBMCs) as a source of macrophages. Methodologies for three different versions of the bioplatform using freshly cultured THP-1 monocytes or purified CD14+monocytes from the PBMCs are described (Figure 1A-C). To our surprise, human THP-1 monocytes alone, with a self-renewal and recruiting capability, could develop structurally organized granulomatous lesions without other myeloid, lymphoid, and nonhematopoietic cells in the 3D co-culture.[OHl] Furthermore, this 3D spheroid co-culture developed a range of morphological forms imitating solid, necrotic, and cavitary tuberculomas not described by the available in vitro 3D granuloma models. Other advantages of this platform include ease of development, increased throughput, scalability, pliability, and real-time monitoring of bacterial burden and lesions in situ. In addition, we successfully used risk group-2 pathogen M. marinum as an alternative for Mtb to develop this bioplatform in the BSL-2 laboratory, which resulted in a comparable bioplatform as the one created using Mtb. Mm can cause granulomatous skin infection in humans and develop human TB-like disease with macrophage-epithelization and tuberculous granulomas in zebrafish (Cosma etal., 2003; Cronan etal., 2021). The Mm- -zebrafish infection model has emerged as a useful in vivo platform to study the dynamics of tuberculous granuloma formation and host-pathogen interactions (Davis et al., 2002; Cronan, 2022). This workflow with Mm can significantly decrease staff hours required during extensive screening efforts in the BSL-3 laboratory while potentially reducing safety concerns and costs.

[0112] Since successful preservation, maintenance, and long-term storage of screening bioplatform is highly desirable, we adapted the workflow to develop a cryo-stable version for potential commercialization that can be frozen for future use and revived on demand (Figure ID). The bioplatform and techniques described have the potential to minimize the amount of animal testing in the high-containment facility by improving screening capacity and efficiency while saving time and resources. Even though the system was developed for Mtb, we have provided the foundation to allow others to modify this system for other granulomatousinflammatory and infectious diseases, including co-morbidities and co-infections such asMtb infection with HIV, influenza, or SARS coronavirus 2. The bio-protocol described here advances the “human in vitro 3D granuloma model” technology.Materials and Reagents

[0113] RPMI 1640 medium with L-glutamine and with or without phenol red (Gibco™, catalog numbers: 11875-093 and 11835-030)

[0114] Penicillin-Streptomycin solution, 10,000 units / ml of each antibiotic (Gibco™, catalog number: 15140-122)

[0115] Sodium pyruvate solution, 100 mM (Gibco™, catalog number: 11360-070)

[0116] HEPES buffer, IM (Gibco™, catalog number: 15630-080)

[0117] Fetal bovine serum (FBS), endotoxin-level tested and heat-inactivated, U.S. origin (Atlas Biologicals, catalog number: F-0500-A)

[0118] Nalgene® Rapid-Flow™ single-use vacuum filter units, sterile (500 ml, with 0.2 pm membrane) (e.g., Thermo Fisher Scientific, catalog number: 5660020)

[0119] Reagent reservoirs, sterile, disposable (e.g., Aquafill, catalog number: S-5501080)

[0120] Micropipette tips with filters (various volumes), sterile, disposable (e.g., Rainin™, catalog numbers: 30389257, 30389272, 30389276, and 30389274)

[0121] Serological pipettes (various volumes) (e.g., Pyrex, catalog numbers: 7077-5N and 7077- 10N)

[0122] Cell culture flasks, 75 cm2, 150 cm2and 175 cm2, sterile (Coming™, catalog numbers: 430720U, 431465, and 431080)

[0123] Falcon™ conical centrifuge tubes, 15 and 50 ml, sterile (Corning™, Falcon®, catalog numbers: 352097 and 352098)

[0124] Microcentrifuge tubes (various volumes, sterile) (e.g., GreenTree scientific, catalog number: T5040G and LabSource, catalog number T56-950)

[0125] Tube holder or rack (e.g., Fisher Scientific, catalog numbers: 21-200-285, 03-448-17, 21- 402-18)

[0126] Human THP-1 cells (American Type Culture Collection, catalog number: TIB-202)

[0127] Mtb strain H37Rv or Erdman expressing deep-red fluorescent protein tdTomato, Mm strain M expressing tdTomato, and Mm strain 1218 expressing green fluorescent protein (GFP)

[0128] Note: The co-culture requires pathogenic mycobacteria (risk group 2 or 3). The attenuated M. bovis BCG strains lack a genomic virulence locus RD1 and are unsuitable.

[0129] Syringe fitted with a needle (25 or 27G), sterile (e.g, Becton Dickinson (BD), catalog number: 309626)

[0130] Cell culture grade water, sterile (Coming™, catalog number:25-055-CM)

[0131] Dulbecco’s Phosphate Buffered Saline (PBS), Calcium and Magnesium free, sterile, pH 7.4 (Gibco™, catalog number: 14190-136)

[0132] 3D cell culture plates (Corning® Spheroid Microplates, catalog number: 4515)

[0133] We tested microplates from several different vendors. Corning® Ultra-Low Attachment (ULA) Spheroid Microplates were optimal for generating 3D tuberculoma structures described here. This 96-well microplate has an opaque black body that shields each optically clear microwell from well-to-well crosstalk. The round-bottom microwells have a hydrophilic, biologically inert surface. They are coated with a covalently bonded, non-ionic, neutrally charged hydrogel that minimizes activation and cell adhesion to the well surface and enables uniform and reproducible 3D tuberculoma-like structure formation. Another helpful alternative is S-BIO PrimeSurface® 3D Culture Spheroid Plates (S-BIO, catalog number: MS-9096UZ)

[0134] Trypan blue stain solution (e.g, Invitrogen™, catalog number: T10282)

[0135] Countess™ Cell Counting Chamber Slides (Invitrogen™, catalog number: C- 10283) or Neubauer cell counting chamber (e.g., Millipore Sigma, Bright-Line hemocytometer, catalog number: Z359629)

[0136] Mycobacterial culture bottle (e.g., 490 cm2sterile roller bottle, Corning®, catalog number: 430195)

[0137] Mycobacterial liquid growth medium, Middlebrook 7H9 broth with supplements (e.g, Prepared in-house, see Recipes. Middlebrook 7H9 broth dehydrated base, BD, catalog number: 271310; Tween 80, Fisher, catalog number: BP338-500; Middlebrook bovine albumin fraction V, dextrose, catalase (ADC) enrichment, BD, catalog number: 212352; and glycerol, Sigma- Aldrich, catalog number: G7893)

[0138] Mycobacterial culture media plates (e.g., Middlebrook 7H10 agar plates without antibiotics and supplemented with oleic acid, albumin, dextrose, and catalase (OADC) enrichment 10% (vol / vol) and glycerol 0.5% (vol / vol) prepared in-house. Middlebrook 7H10Agar, BD Difco™, catalog number: 262710; Middlebrook OADC, BD, catalog number 212351; and Glycerol, Sigma-Aldrich, catalog number: G7893)

[0139] Bacterial cell spreaders (e.g., Fisher Scientific, catalog number: 14-665-230)

[0140] Triton X-100 (e.g, Sigma-Aldrich, catalog number: X-100)

[0141] Human PBMCs from whole blood of healthy, tuberculin skin test-negative donors collected in BD Vacutainer® Cell Preparation Tubes (CPT™) (catalog number: 362761).

[0142] VitroCol Type 1 human collagen solution (Advanced BioMatrix, catalog number: 5007-20ml)

[0143] Fibronectin from human plasma, 0.1% solution, 1 mg / ml (Sigma-Aldrich, catalog number: F0895)

[0144] Sodium hydroxide solution (NaOH) 0.1 M for cell culture (Advanced Biomatrix, catalog number: 5078)

[0145] Phosphate buffer saline (PBS) 10X without Calcium and Magnesium (AdvancedBioMatrix, catalog number: 5076-B)

[0146] Red blood cell (RBC) lysis buffer, IX eBioscience™ (Invitrogen, catalog numberOO- 4333-57)

[0147] Ethylenediaminetetraacetic acid (EDTA) solution, 0.5 M, sterile (Amresco, catalog number: C-177)

[0148] Magnetic-assisted cell sorting (MACS) LS columns (Miltenyi Biotec, catalog number: 130-042-401)

[0149] MACS® CD14 MicroBeads, human CD14+monocyte separation reagent (MiltenyiBiotec, catalog number: 130-050-201)

[0150] Cell scraper (e.g., Costar™, catalog number: 3010)

[0151] Petri dish (e.g, Falcon™, catalog: 351029)

[0152] Dimethyl sulfoxide (DMSO), cell-culture grade, and endotoxin level tested (Sigma-Aldrich, catalog numbers: D2650-5 x 10ML and D5879)

[0153] Assay block 96 well, 2 ml capacity, sterile (Corning™, catalog: 3960)

[0154] Leibovitz’s L-15 medium (Gibco™, catalog number: 11415-064)

[0155] Polyvinylpyrrolidone (PVP) (MP Biomedicals, catalog number: 102786)

[0156] Cell growth medium (complete RPMI-1640 medium with antibiotics) (see Recipes)

[0157] 3D cell culture medium (complete RPMI-1640 medium without antibiotics) (see Recipes)

[0158] Middlebrook 7H9 broth (see Recipes)

[0159] Extracellular matrix (ECM) solution (see Recipes)

[0160] PBMC wash buffer (see Recipes)

[0161] MACS buffer (see Recipes)

[0162] 3D cell culture freezing mediums (see Recipes)Equipment

[0163] Class II Type A2 biological safety cabinet (BSC) (e.g., Nuaire, Model: NU-543-600)

[0164] Micropipettes (various volumes) (e.g., Rainin™, catalog numbers: Pipet-Lite LTSPipettes L-20XLS, L-200XLS, L-1000XLS, L-5000XLS, and L8-200XLS)

[0165] Novaspec II spectrophotometer (Amersham Pharmacia Biotech) or equivalent

[0166] Aerosolve® canisters or equivalent

[0167] Cell culture incubator, set at 37 °C with 5% CO2 and 100% humidity (Forma Scientific,Model: 3110)

[0168] Microcentrifuge for microtubes (Eppendorf, Model: 5430 R)

[0169] Centrifuge for 15- or 30-ml conical tubes (Eppendorf, Model: 5810 R)

[0170] Cell counting equipment (Thermo Fisher Scientific, Model: AMQAX1600 Countess II orAMQAX2000 Countess III)

[0171] pH meter or pH paper (e.g., Fisher brand, catalog number: 13-640-510)

[0172] Cytation 5 microplate fluorescence reader and cell imager or equivalent with optionalCO2 source (i.e., CO2 controller) (Agilent / BioTek, Model: CYT5MPV with GEN5PRIME andGEN5SPOT)

[0173] Note: Fluorescence reader and cell imager other than Cytation 5 can be used.

[0174] MidiMACS™ Separator (Miltenyi Biotec, catalog number: 130-042-302) and MACSMulti-Stand (Miltenyi Biotec, catalog number: 130-042-303)

[0175] Water bath (Precision 180 series)

[0176] Refrigerator, set at 2 to 8 °C

[0177] Freezer, set at -80 °C

[0178] Liquid nitrogen storage for cell culture (temperature -196 °C)

[0179] AutoclaveSoftware

[0180] Gen5 V3.08 with Spot Counting add-on module (Agilent / BioTek)

[0181] Gen5 Image+ software controls the operation of the Cytation 5 for both the photomultiplier tube (PMT)-based microplate reading and automated digital microscopy.

[0182] GraphPad Prism V9.3 (GraphPad by Dotmatics)ProcedureWorkflow for 3D tuberculoma bioplatform using freshly cultured THP-1 monocytes

[0183] This optimized workflow describes the steps to develop a tuberculoma bioplatform in a 96-well 3D cell culture microplate using a simple co-culture of THP-1 monocytic cells and fluorescent pathogenic mycobacteria for high-content screening of potential therapeutics. All work involving the handling of virulent Mtb strains (risk-group 3 organisms) should be performed in a BSL-3 laboratory under a class II BSC wearing appropriate personal protective equipment (PPE). If the organism of interest is Mm (risk-group 2), the experiment is conducted in a BSL-2 laboratory under a class II BSC wearing appropriate PPE. The typical workflow for screening test compounds, small molecules, or biologies spans 14 days. Depending on the experiment’s goal, the workflow can be extended beyond two weeks if the microwells are replenished with fresh culture medium and THP-1 cells. Using Mm 1218 strain and carefully exchanging a 50% medium in microwells with a new medium once every seven days, we successfully cultured the ‘mycobacteria-in-spheroid’ model for up to 50 days.Day 1. Development of 3D ‘mycobacteria-in-spheroid’ co-culture

[0184] Preparing THP-1 cell suspension for 3D cell culture.

[0185] Aliquot 3D cell culture medium (see Recipe 1) in 50 ml conical tubes.

[0186] Note: Check the sterility of the medium in advance before use.

[0187] Prewarm 3D cell culture medium in the cell culture incubator at 37 °C for 1 hr.

[0188] Count THP-1 cells cultured in a growth medium containing antibiotics (see Recipe 2) and assess cell viability by trypan blue dye exclusion method.

[0189] Cell viability must be > 95%, and the final concentration of 1 x 106live cells / ml is required. Use fully dissolved trypan blue dye to avoid interference with the automated cell counting caused by the dye precipitates. We used automatic cell counter Countess II or III for cell counting. Cells may be pooled from two or more 150 cm2cell culture flasks cultured in the same batch and subjected to cell count and viability assessment. THP-1 cells cultured in two or more flasks will be required for extensive experiments requiring > 4 microplates.

[0190] Centrifuge cells for 6 min at 200-250 x g at room temperature (15-20 °C) in 50 ml conical tubes and discard the supernatant without disturbing the cell pellet.

[0191] Resuspend the cell pellet in 15 ml of prewarmed (37 °C) 3D cell culture medium (see Recipe 1) and gently mix using a pipette. Centrifuge cells again for 6 min at room temperature at 200-250 x g and discard the supernatant.

[0192] Repeat this washing step two more times.

[0193] After the third wash, resuspend the cell pellet in 15 ml of prewarmed (37 °C) 3D cell culture medium, mix gently using a pipette, and let the cell suspension rest for 30 min at room temperature.

[0194] This step is incorporated to exchange and remove antibiotics pinocytosed by THP-1 cells during a monolayer culture in an antibiotic-containing growth medium.

[0195] Centrifuge cells for 6 min at room temperature at 200-250 x g and discard the supernatant. Resuspend the cell pellet in 10 ml of prewarmed (37 °C) 3D cell culture medium and gently mix using a pipette.

[0196] Perform final cell count and assess cell viability by trypan blue dye exclusion method.

[0197] Adjust the cell concentration to 1 x 106live cells / ml using a 3D cell culture medium.

[0198] In the microplate format tested, each microplate precisely requires 6 ml of THP-1 cell suspension. Cells are not dispensed in the microwells on the periphery of the plate to avoid an “edge effect.” The outer wells are filled with sterile 3D cell culture medium or cell culture grade water to mitigate evaporation.

[0199] Preparing / / / 7> H37Rv (tdTomato) or Mm M (tdTomato) suspension for 3D cell culture

[0200] We tested Mtb (H37Rv, Erdman, Beijing F2, and CDC 1551) and Mm (M) strains that expressed red fluorescent protein tdTomato for this workflow. The fluorescent protein tdTomato was cloned into the constitutively expressed mycobacterial pmspl2 vector using pTEC27 plasmid ((Takaki et al., 2013) and also see example 2 described below This red fluor provides optimal brightness and low background autofluorescence in the spheroid co-culture. To prepare frozen mycobacterial stocks, fluorescent Mtb and Mm strains were grown in Middlebrook 7H9 broth (see Recipe 3) supplemented with appropriate antibiotic (hygromycin 50 pg / ml for strains expressing tdTomato or kanamycin 25 pg / ml for strain expressing GFP). We inoculated 5 ml of 7H9 broth with 100-150 pl from a strain collection stored at -80 °C in a 30 ml-capacity 25 * 150 mm glass culture tube and incubated at 30 °C (Mm) or 37 °C Mtb) for 5-10 days. Inexperiments comparing multiple strains, stocks were prepared simultaneously using the same growth medium batch. The optical density of the inoculum was measured at 600 nm (ODeoo). When ODeoo of inoculum reached 0.6- 1.0, the 5 ml inoculum was added to 150 ml of 7H9 broth in a 490 cm2sterile culture bottle. The culture was incubated at 30°C (Mm) or 37°C (Mtb) for 10-12 days, with once a daily manual shaking, until the ODeoo reached 1.0-1.5, as measured by Novaspec II spectrophotometer. Cultures were pelleted by centrifugation at 4000 x g for 30 min at room temperature. The supernatant was removed by decanting or pipetting, and the pellet was resuspended in 10 ml of fresh 7H9 broth (see Recipe 3), aliquoted at 1.0 ml per cryovial, and stored at -80 °C. The 7H9 broth supplemented with appropriate antibiotics was used to prepare stocks of recombinant fluorescent strains. An aliquot was thawed seven days after freezing and sub-aliquoted in the 100 pl working stocks. The viable count was determined on 7H10 agar without antibiotics by incubating at 30 °C for 12-15 days (Mm) or 37 °C for 28-30 days (Mtb). Frozen stocks can be stored for one year after preparation; new stocks must be prepared using the original strain collection after one year.

[0201] Transport the cryovial containing frozen stock of Mtb H37Rv (tdTomato) or Mm M (tdTomato) in an AEROSOLVE® (transport) canister from a storage freezer to the BSC. Place the cryovial in a tube holder in the BSC.

[0202] Transfer 900 pl of 3D cell culture medium into a microtube containing 100 pl of mycobacterial working stock and mix well.

[0203] Centrifuge at 5000 x g for 30 min at 8 to 10 °C. When finished, carefully discard the supernatant using a pipette without disturbing the mycobacterial pellet.

[0204] Repeat this wash step using a new 1 ml 3D cell culture medium and resuspend the pellet in 1 ml of pre-warmed (37 °C) 3D cell culture medium.

[0205] Pass the mycobacterial suspension through a 25-27G needle fitted to a 1 ml syringe 10 to 20 times to make a single-bacterial-cell suspension.

[0206] Perform this step immediately before THP-1 cell infection to avoid aggregation of bacteria. CAUTION: DO NOT CAP THE NEEDLE ON THE SYRINGE. Place the syringe with a needle in the sharps-disposal container carefully.

[0207] Transfer the required amount of mycobacterial single-cell suspension into a sterile reservoir. Dilute mycobacterial suspension using a volume of prewarmed 3D cell culturemedium needed to reach the intended multiplicity of infection (MOI) and to make a necessary volume of bacterial inoculum for THP-1 cell infection in microplates.

[0208] Our experiments are routinely performed using five microplates, and we typically prepare 35 ml each of THP-1 cell-suspension and mycobacterial-suspension (infection inoculum) in sterile reservoirs. During the transfer process into microwells, we frequently mixed the inoculum by pipette up and down to maintain the uniform single-cell suspension and prevent mycobacteria from settling in the reservoir and clumping together. We determined the optimal range of bacterial number required for the infection of monocytes in the 3D co-culture without overcolonization in 3D spheroids formed over 2-3 weeks of incubation in a pilot experiment (for example, using fluorescent Mm strains see Figure 2A-D). MOI was calculated as input CFU / input number of monocytes added per well. We did not wash monocytes following infection and avoided using aminoglycoside antibiotics to kill extracellular mycobacteria, if any. Pinocytosed aminoglycosides can reach macrophage phagosomes and contribute substantially to cells’ antimicrobial activity (Vogt and Nathan, 2011). We determined that the optimal MOI using Mm M (tdTomato) is 0.008 (i.e., 800 CFU per 105live THP-1 cells), and the optimal MOI range is 0.006-0.012. Our optimized MOI for Mtb H37Rv (tdTomato) is 0.02 (MOI range 0.015-0.025). The ideal MOI varies with the Mycobacterium species or strains used in 3D coculture and can be experimentally determined by monitoring the dynamics of bacterial growth and granulomatous lesion formation within the 3D spheroid structures and ascertaining the Z’- factor of the HTS assay performed using the bioplatform (refer to data analysis section below). Low physiological MOI ensures the complete gathering of bacilli by aggregating THP-1 cells during the process of 3D spheroid formation and avoids unwanted colonization outside the spheroid structures over two weeks without needing the use of aminoglycoside antibiotics that are believed to kill only extracellular bacteria.

[0209] Preparing 3D microplates for 3D cell culture

[0210] Label CORNING® 3D spheroid microplates with plate number, date, experiment, and performer’s name. Mark the peripheral boundary wells. See the data analysis section for a schematic of a plate used in the compound screening assay.

[0211] Add 100 pl of THP-1 cell suspension (1 x 105live cells) into each microwell except those on the periphery.

[0212] Use six micropipette tips fitted on a multichannel pipette. Micropipette tips are slanted against the wall of wells while dispensing cell suspension. Therefore, avoid touching the micropipette tips to the bottom of the wells.

[0213] Add 100 pl of mycobacterial suspension into each microwell except those on the periphery.

[0214] Control wells with THP-1 cells can be kept without mycobacterial infection, depending on the experiment's goal. Add 100 pl of 3D cell culture medium in these wells in place of the bacterial suspension.

[0215] Mix thoroughly up and down at least three times without touching the microwell bottom to obtain a homogeneous suspension of THP-1 and mycobacteria. Avoid bubble formation during mixing.

[0216] Fill the peripheral microwells with 250-300 pl of sterile 3D cell culture media or cell culture grade water using a multichannel micropipette.

[0217] Place the microplates in a cell culture incubator set at 37 °C with 5% CO2 and 100% humidity to continue the co-culture experiment and generate 3D ‘mycobacteria-in-spheroid’ structures in ULA round-bottom microwells.

[0218] Determine the actual CFU count in the 100 pl of mycobacterial suspension by plating dilutions onto Middlebrook 7H10 agar plates. Incubate agar plates at 30 °C for Mm M (tdTomato) for 12-14 days and 37 °C Mtb H37Rv (tdTomato) for 3-4 weeks.

[0219] Clean the BSC with tuberculocidal disinfectant (e.g., Cavicide).

[0220] Fill out the experiment worksheet.Day 6. Addition of test compounds, small molecules, or other therapeutics

[0221] Preparing and adding drug dilutions for screening in the 3D bioplatform

[0222] Prepare a dilution of test drugs in a pre-warmed (37 °C) 3D cell culture medium.

[0223] Optimal drug concentration can be experimentally assessed. We routinely performed screening experiments using a standard concentration of 20 pM of drug per microwell. In addition, we usually tested six different concentrations ranging from 20 pM to 0.625 pM for selected drugs (using 2-fold serial dilution). We performed drug dilutions in 96-well, 2-ml capacity, sterile assay blocks. For screening of large compound libraries, stock compound solutions (10 mM) in DMSO were obtained from the commercial vendors, aliquoted, and stored at -20 °C or the recommended temperature. To get a final concentration of 20 pM of testcompound per microwell, 10 pl of stock compound (10 mM) was diluted in 990 pl of prewarmed (37 °C) 3D cell culture medium, and 50 pl of this diluted compound solution was added to the microwells containing spheroids in 200 pl of the 3D cell culture medium.

[0224] Observe the granulomatous lesion formation in the 3D ‘mycobacteria-in-spheroid’ coculture using a manual mode in Cytation 5 imager or an inverted microscope (optional).

[0225] Florid granuloma lesions begin to develop between days 5 and 6 in the 3D spheroid infected with the optimized low MOI of Mm M and Mtb H37Rv and between days 8 and 10 for Mm 1218. The granulomatous lesions grow over time and become structurally organized.

[0226] Add diluted drug in 50 pl volume to the microwells slowly and carefully without disturbing the 3D spheroids. Keep a minimum of three technical replicates for individual test drugs and include appropriate positive and negative control wells in each microplate.

[0227] The final volume of the medium in the microwell after drug treatment will be about 250 pl. Negative control wells will receive 50 pl of 3D cell culture medium. Pipette tips are slanted against the wall of microwells while dispensing drug solution slowly to avoid disturbing the formed 3D spheroids and granuloma lesions. Please refer to the data analysis section for more information about positive and negative controls kept in our drug screening assay.Day 12. Fluorescence intensity reading and imaging to assess antitubercular drug efficacy

[0228] 1. Reading microplates to detect fluorescence intensity as a measure of bacterial burden

[0229] Place the microplate with lid (without bottom plate-stand) into Cytation 5 multi-mode plate reader, previously set to 30 °C or 37 °C and 5% CO2.

[0230] Cytation 5 combines automated digital widefield microscopy with conventional multimode microplate detection to provide phenotypic cellular information and well-based quantitative data. In Gen5 software, ensure that the appropriate plate type (e.g., Corning ULA round well bottom) is selected, and the vessel’s bottom elevation is defined. The fluorescence intensity in the co-culture can be measured in the absence of CO2 in the Cytation 5 plate reader. For / z co-culture, we performed in situ fluorescence reading and imaging at 30°C without a CO2 source and controller. We tested auto gain and several fixed gain readings in the microplate with 3D spheroids and found the autogain reading suitable for compound screening assay. For fluorescence intensity readings in the 3D spheroids generated, the ‘top’ reading of the microplate is more optimal than the ‘bottom’ reading. A microplate fluorimeter can be a low-cost alternativeto a multimode fluorescence reader. The optimal time to read fluorescence intensity or imaging after drug treatment should be experimentally determined.

[0231] Read the tdTomato fluorescence intensity using the fluor-specific fluorescence intensity reading protocol developed, validated, and stored on the computer attached to Cytation 5.Table 2 Fluorescence intensity reading parameters.

[0232] To read fluorescence intensity and growth of mycobacteria expressing tdTomato in the Corning 96-well 3D cell culture plate, follow the below four steps.

[0233] i. Open the Gen5 software. Next, select the ‘Experiments’ tab on the left panel in the ‘Task Manager' screen.

[0234] ii. On the right-hand side of the screen, choose ‘Create using an existing protocol’ and select the pre-developed and saved tdTomato fluorescence intensity reading protocol. The ‘Task Manager’ screen will be closed. The parameters of the tdTomato fluorescence reading protocol are listed in Table 2.

[0235] iii. On the top panel of the Gen5 software, click on the green button ‘Read New.’ The software will ask to save the experiment name and other details.

[0236] iv. After saving the experiment file, the software will proceed to read fluorescence intensity in the pre-selected microwells of the plate.

[0237] Export the fluorescence intensity data as an Excel file generated by Gen5 and save it on the computer for data analysis.

[0238] 2. Automated plate imaging using Cytation 5

[0239] Place the microplate with lid into Cytation 5 set at 30 °C or 37 °C and 5% CO2.

[0240] Capture images using the optimized protocol and the parameters listed in Table 3.

[0241] The ‘mycobacteria-in-spheroid’ co-culture imaging was carried out to generate quality images of 3D tuberculoma-imitative structures and assess drug efficacy in reducing fluorescent mycobacterial growth and resolution of granuloma lesions. We used a 2.5x objective and 2 ^ 2 image montage to image the entire well. A brightfield imaging channel was used to capture total spheroid images, and a red fluorescence channel (Texas red) was used to capture mycobacteria expressing tdTomato.Table 3 Automated fluorescence imaging parameters

[0242] To image a 3D cell culture microplate, follow the steps below.

[0243] i. Open the Gen5 software. Select the ‘Experiments’ option in the ‘Task Manager’ window.

[0244] ii. On the right-hand side of the screen, choose ‘Create using an existing protocol’ and select the pre-developed 3D spheroid imaging protocol. The ‘Task Manager’ screen will be closed. The complete parameters of the imaging and analysis protocol are listed in Tables 3-7.

[0245] iii. Click the ‘Read New’ button on the software's top panel. The software will ask to save the experiment name and details in a specific location (directory) on the computer.

[0246] iv. After saving the experiment file, the software will automatically start imaging the preselected microwells in the plate.

[0247] We used montage imaging, image stitching, and Z-stacking in experiment mode in Gen5 to capture 3D tuberculoma images. To capture the entire 3D tuberculoma structure (> 2000 pM in diameter) that span outside the field-of-view of the objective, montage image capture mode was used to capture four image segments (tiles) in the microwell. Individual tiles in the montage are stitched together to make a whole single image. The ‘auto for stitching’ method was used to capture the montage. The 3D tuberculoma structure exists within a range of Z-planes, and multiple images (slices) must be acquired by moving the objective in the z-axis (or focal) planes. Therefore, we used a Z-stacking imaging procedure within Gen5 by selecting “Image Z-stack.” Multiple automated image slices were captured below and above the focal plane to ensure that the 3D tuberculomas, cells, and granuloma lesions were imaged at the proper Z-height. After the Z-stacked images are captured, a projection of the pictures of the Z-stack is completed by performing a “Z Projection” step and using the focus stacking algorithm to create a final composite image. For image stitching and Z-projection, Gen5 Image+ software is required. Each well takes about 2 min to complete the imaging process using the parameters described.

[0248] For stitching of the individual tiles to generate one single stitched image, use the parameters listed in Table 4.Table 4 Image stitching parameters

[0249] For Z-stacking of the individual slices to generate one stacked image, use the parameters listed in Table 5.Table 5 Z-stacking parameters for 3D imaging

[0250] 3. Cellular analysis of 3D tuberculoma images in Gen5

[0251] Perform the cellular analysis of projected images using the parameters listed in Table 6 and 7

[0252] We used cellular analysis to count the diameter and size of 3D tuberculomas, the number of encompassing granuloma lesions, and the area affected by bacterial growth and lesions in the tuberculoma structures.Table 6 3D Spheroid Cellular Analysis Parameters [For Spheroid analysis]Table 7 3D Spheroid Cellular Analysis Parameters [For Lesion analysis]Day 13-14. Additional fluorescence intensity reading of microplates with 3D culture

[0253] For relatively slow-growing virulent Mtb strains, for example, Mtb H37Rv or Erdman (tdTomato), we performed additional fluorescence intensity readings of the co-cultures on day 14. We routinely performed drug screening experiments using five microplates but used up to 12 microplates in some experiments. For experiments employing more than five microplates, imaging of the remaining plates can be continued the following day.

[0254] Read the tdTomato fluorescence intensity using the above-described parameters.

[0255] The workflow described above generates a solid tuberculoma-like structure in each microwell with 3D co-culture (Figure 3A). These 3D tuberculoma-imitative structures encompass well-organized granulomatous lesions and develop critical attributes including hypoxia, necrosis, acidosis (Figure 3B), and biofilm formation in the central core (see example-2 described below). The resulting 3D tuberculoma structures are highly uniform in size distribution, bacterial growth, and granuloma lesions encompassed. The bioplatform in 96-well format can be employed as a high-content screening and imaging assay platform for anti-TB drug discovery (see data processing and analysis section).Modified workflow for 3D tuberculoma bioplatform incorporating human extracellular matrix

[0256] This workflow details the steps required to develop a bioplatform using a co-culture of THP-1 cells and pathogenic mycobacterial strain in the presence of physiologically relevant human ECM components. ECM contributes to the architecture of tuberculous granulomas in the human lungs, and its immunopathological destruction leads to cavity formation in TB patients (P. T. Elkington et al., 2011; Urbanowski et al., 2020; P. Elkington et al., 2022). In addition, collagen-rich ECM is known to regulate the survival of macrophages in 3D in vitro granuloma models (Tezera et al., 2017a). Still, the human in vitro granuloma model with cavitary transformations has yet to be described. Advances in 3D in vitro granuloma models that develop cavitary features will allow investigations of drivers of cavitation and pharmacological interventions to identify HDT drugs that can prevent or treat / b-induced tissue destruction and immunopathology. The 3D tuberculoma model described here develops cavity-like features in spheroid structures.Day 1. Development of 3D ‘mycobacteria-in-spheroid’ co-cultures

[0257] Prepare THP-1 cell suspension as described above in workflow A.

[0258] Prepare mycobacterial suspension as described above in workflow A.

[0259] We used Mm 1218 strain expressing GFP to infect freshly grown THP-1. This strain permitted 3D co-culture for a relatively longer duration when 50% of culture medium in microwells is exchanged with fresh medium once every seven days. We used the low MOI of 0.005 (optimal range 0.005 to 0.008) to develop tuberculoma-imitative structures with cavity-like features. Optimization of the MOI and the quantity of extracellular matrix added is recommended when using different mycobacterial species or strains. We have yet to investigate the ability of additional Mm oxMtb strains in our collection to develop this feature in 3D cell culture.

[0260] Prepare 3D spheroid microplates for 3D cell co-cultures as described above in workflow A.Day 3. Incorporation of the ECM solution in 3D ‘mycobacteria-in-spheroid’ co-cultures

[0261] Preparing the ECM solution

[0262] Add the required quantity of human collagen solution (3mg / ml) into a 15 ml sterile conical tube.

[0263] Prepare the ECM solution using human collagen and fibronectin solutions as described in Recipe 4.

[0264] We have used a purified human collagen solution VITROCOL®. VITROCOL collagen is naturally secreted from human neo-natal fibroblast cells in the in vitro cell culture. The processed and pure form provided by the supplier was used. VITROCOL is approximately 97% Type I human collagen, with the remainder comprising Type III collagen. Storage of VITROCOL collagen at 2-8 °C is essential. Do not freeze. While preparing collagen mixture, gentle but thorough mixing and careful pH monitoring are critical. Keep the mixture at 4 °C in the refrigerator or ice to prevent gelation. The mixture can be prepared a day in advance and stored at 2-8 °C overnight in the fridge to allow homogeneous pH adjustment of the solution. The fibronectin solution is added at the end of the mixture.

[0265] Adding the ECM solution

[0266] Add 30-50 pl of the ECM solution to each microwell containing the 3D co-culture.

[0267] Dispense the ECM solution slowly by slanting the micropipette tips against the side wall of the microwells. Control microwells without ECM addition can be kept depending on the experiment's objectives. We have investigated incorporating different volumes of ECM solution ranging from 5 to 50 pl in the 3D co-cultures. A volume of ECM used for incorporation and MOI of Mm 1218 used in the co-culture influence the formation of cavity-like features in 3D tuberculoma structures. A 30-50 pl of ECM incorporation led to the cavity -like feature formation in our hand (Figure 3C). Very low MOI (50 CFU) or lesser volumes of ECM incorporation (< 25 pl) did not lead to cavity formation.

[0268] Place the microplates back into the CO2 incubator set at 37 °C with 5% CO2 and 100% humidity and continue the spheroid co-culture by incubating the microplates.Day 7 and Day 14. Addition of fresh medium

[0269] Adding a fresh 3D cell culture medium (perform as described above)

[0270] Carefully exchange 50% of the cell culture medium in microwells with a fresh prewarmed (37 °C) medium.

[0271] Exchange with a new medium is performed for co-cultures grown greater than 2 weeks.

[0272] Observe the growth of spheroid co-culture using a Cytation 5 imager or an inverted microscope (optional).

[0273] Continue the co-culture by incubating the microplates in the CO2 incubator.Day 16 and 21. Imaging and fluorescence intensity reading of microplates with 3D coculture

[0274] Place the microplate with lid into Cytation 5 set at 30 or 37 °C, with or without 5% CO2.

[0275] Read the fluorescence intensity using optimized parameters as described above in Table 1, except that the excitation and emission wavelengths for GFP are 488 nm and 510 nm, respectively.

[0276] Capture images using the optimized protocol and parameters for imaging with brightfi eld filter and fluorescence filter as described above in Table 2, except for using a GFP filter instead of a Texas red.

[0277] Perform 3D image processing using the 3D montage imaging, stitching, and Z-stacking parameters described above (Tables 3 and 4).

[0278] Following the capture of images, a Z-proj ection of the Z-stack images was performed using the focus stacking algorithm described above in workflow A to create the final composite images.

[0279] Perform cellular analysis of 3D projected images using the parameters described above in Table 5Workflow for 3D bioplatform using purified human PBMC subsets

[0280] This optimized workflow details the sequence of steps required to develop a bioplatform using co-cultures of purified human CD14+blood monocytes and virulent Mtb Erdman (tdTomato) in the 3D cell culture microplate. Profiling of granulomas in / V / / / ?-infected animal models and humans revealed that macrophages constitute a significant portion (> 40-50%) of granuloma cells. However, the cellular composition can vary with early and late granulomas and granuloma forms (Gideon et al., 2022). Primary monocytes that differentiate into macrophages constitute a small portion, approximately 10% of cells in human PBMC samples. Because of the insufficient numbers of monocytes and macrophages in the PBMC samples for the continuous recruitment in and maintenance of the 3D in vitro granulomas, whole PBMCs lead to the generation of small cellular aggregates rather than well-organized, dynamic granulomas. Therefore, purification and enrichment of CD14+monocytes from the PBMCs are critical steps in this workflow to develop the 3D tuberculoma model. In this workflow, 3D ‘mycobacteria-in- spheroid’ co-cultures are developed using MACS column-purified human primary CD14+monocytes that differentiate into macrophages to simulate the formation of the core scaffold inhuman tuberculomas. Use of virulent Mtb strains is required. A flow-through unbound fraction generated during MACS column CD14+monocyte purification contains CD14 microbead- unlabeled lymphocytes and CD14' cells from the PBMCs. This ‘lymphocyte-rich’ fraction is cultured separately for two days, washed, and added to 3D spheroids generated using purified autologous CD14+monocytes in the 96-well microplates to simulate lymphocytes accumulated on the periphery of human tuberculomas with macrophage-rich centers. Granuloma lesions are developed in ‘mycobacteria-in-spheroid’ co-cultures even without adding this ‘lymphocyte-rich’ fraction. This bioplatform employing innate and adaptive cell subsets purified from the PBMCs can be used for a confirmatory screening of ‘top hits’ identified from a primary screen in the THP-1 platform.Day 1. Development of 3D ‘mycobacteria-in-spheroid’ co-culture

[0281] Separation of the PBMCs from whole human blood

[0282] PBMCs can be isolated from anti coagulated human blood or “buffy coat” by density gradient centrifugation, for example, using Ficoll-Paque™. We used BD Vacutainer® Cell Preparation Tubes (CPT™) for blood collection and PBMC separation. PBMCs were isolated following the manufacturer’s protocol (BD CPT Manual VDP 40104-05 pg l-2 We used 200 ml of blood per human donor to isolate PBMCs and purify CD14+monocytes. The PBMCs isolated and CD14+monocytes purified were enough to develop five microplates.

[0283] Collect blood into BD Vacutainer® tubes. Gently mix the blood by inverting the tubes 5 to 10 times before centrifugation.

[0284] BD VACUTAINER® tubes should be stored at room temperature (18-25 °C) and labeled adequately for human donor identification.

[0285] Centrifuge tubes with a blood sample at room temperature in a horizontal rotor (swing- out head) for 30 min at 1500-1800 xg.

[0286] Carefully pipette out the plasma without disturbing the below white blood cell layer immediately after centrifugation using a pipette. Collect the cell layer that contains mononuclear cells into a 50 ml conical tube. The PBMCs from several CPT tubes can be pooled into one conical tube. Centrifuge the tubes at 1200 xg for 15 min. Remove the supernatant without disturbing the cell pellet.

[0287] Lyse the RBCs by resuspending the pellet in 10 ml of lx RBC lysis buffer for 5 min at room temperature.

[0288] Stop the lysis reaction by adding 25 to 30 ml of PBMC wash buffer (see Recipe 5) and gently mixing the cell suspension using a pipette. Centrifuge cells at room temperature at 275-300 x g for 10 min. Carefully discard the supernatant.

[0289] Resuspend the cell pellet in the 20 ml PBMC wash buffer. Repeat the washing step thrice by centrifugation at 200 x for 10 min at 15-20 °C.

[0290] These washing steps at low-speed centrifugation are required to remove platelets.

[0291] Resuspend the cell pellet in the PBMC wash buffer. Perform cell count and assess cell viability by trypan blue dye exclusion method.

[0292] Cell viability should be > 95%. Dead cells may bind nonspecifically to MACS MicroBeads during the purification of CD I4 monocytes.

[0293] Centrifuge cell suspension at 300 xg for 10 min. Remove the supernatant completely.Adjust the final PBMC concentration to 1x107live cells in 80 pl of MACS buffer (see Recipe 6) to purify CD14+monocytes.

[0294] Buffer containing Ca2+or Mg2+is not recommended. Use a cold buffer (2-8 °C) to minimize nonspecific cell labeling.

[0295] Purification of CD14+monocytes from PBMCs by positive selection.

[0296] We purified CD14+monocytes from the PBMCs by using the MACS technique and following the manufacturer’s protocol.

[0297] Add 20 pl of human CD 14 MicroBeads to 107live PBMCs in 80 pl of MACS buffer and mix well by pipetting to label the CD14+monocytes in PBMCs magnetically. For higher cell numbers, scale up all the reagent volumes and total volumes accordingly.

[0298] MicroBeads are conjugated to monoclonal anti-human CD14 antibodies. Since CD14 lacks a cytoplasmic domain, it is considered that binding of the antibody to CD14 does not trigger signal transduction in monocytes. Therefore, it does not affect the phagocytosis o Mtb bacilli.

[0299] Incubate for 15 min at 2-8 °C.

[0300] Wash cells by adding 1-2 ml of the MACS buffer (Recipe 6) per 107cells and centrifuge at 300 x g for 10 min. Remove the supernatant completely.

[0301] Resuspend up to 1 x 108cells in 500 pl of the MACS buffer.

[0302] Proceed to magnetic separation using LS columns.

[0303] Choose an appropriate MACS Column and MACS Separator according to the number of total cells and CD14+cells. For LS columns, the recommended sample size for leukocytes is 105- 108labeled cells in 1 x 107to 2 x 109total cells.

[0304] Place the LS column in the magnetic field of a MidiMACS separator attached to the MACS MultiStand.

[0305] Prepare the column by rinsing it with 3 ml of the MACS buffer (Recipe 6). Label collected effluent as ‘wash.’

[0306] Apply cell suspension onto the column.

[0307] Collect unlabeled cells that pass through and wash the column with 3 ml of the MACS buffer and collect total effluent in a 50 ml conical tube. This unlabeled cell fraction contains lymphocytes and mononuclear cells other than CD14+cells. Perform washing steps by adding 3 ml of MACS buffer three times. Label the tube as a ‘lymphocyte-rich’ fraction.

[0308] Only add a new buffer during washing steps when the column reservoir is empty.

[0309] Remove the column from the separator and place it in a new 50 ml conical tube.

[0310] Pipette the 5 ml of the MACS buffer onto the column. Immediately flush out the magnetically labeled cells by firmly pushing the plunger into the column. Label the fraction as CD14+monocytes.

[0311] Perform purification of CD 14 monocytes from all isolated PBMCs from a donor.Centrifuge conical tubes containing purified CD14+monocytes or ‘lymphocyte-rich’ fraction at 300 xg for 10 min.

[0312] Resuspend ‘lymphocyte-rich’ fraction in 50 ml of PBMC growth medium (Recipe 2). Transfer cells to a 150 cm2cell culture flask. Culture cells in a cell culture incubator set at 37 °C with 5% CO2 and 100% humidity for two days.

[0313] Resuspend the pellet of purified CD14+monocytes in 10 ml of PBMC growth medium (Recipe 2) and centrifuge at 300 x for 10 min. Remove the supernatant completely.

[0314] Wash purified CD14+monocytes three times by adding 10-15 ml of prewarmed (37°C) 3D cell culture medium (Recipe 1) and centrifugation at 300 for 10 min. Remove the supernatant completely. Resuspend purified monocytes in 10 ml of 3D cell culture medium.

[0315] Perform cell count and assess cell viability by trypan blue dye exclusion method.

[0316] Adjust the cell concentration of CD14+monocytes to 2 x 106live cells / ml using a 3D cell culture medium.Development of 3D ‘mycobacteria-in-spheroid’ co-culture using CD14+monocytes

[0317] Develop ‘mycobacteria-in-spheroid’ co-cultures using CD14+monocytes andMZ> Erdman (tdTomato) in the 96-well Corning 3D spheroid microplates as described above in workflow A.

[0318] Our experiments are routinely performed using 5000 CFU oiMtb Erdman for infection of 2 x 105live monocytes per microwell (MOI 0.025).Day 3. Addition of lymphocytes and other CD14 negative mononuclear cells to the coculture

[0319] Preparation and addition of ‘lymphocyte-rich’ cell subsets in the 3D microplate

[0320] Collect ‘lymphocyte-rich’ cell fraction from the cell culture flask into a 50 ml conical tube.

[0321] Use the cell scraper gently to release any adherent cells attached to the surface of the cell culture flask during the 2-day culture period.

[0322] Centrifuge the cell suspension at room temperature at 300 xg for 10 min. Remove the supernatant completely. Resuspend cell pellet in 15 ml of prewarmed (37 °C) 3D cell culture medium.

[0323] Wash cells four times by adding 15 ml of prewarmed (37°C) 3D cell culture medium. Centrifuge at 300 x for 10 min. Allow the cell suspension to rest for 30 min at room temperature following the third wash before final centrifugation. After the last wash, resuspend cells in 10 ml of prewarmed (37 °C) 3D cell culture medium.

[0324] Perform cell count and assess cell viability by trypan blue dye exclusion method. Then, adjust the cell concentration to 4 * 106live cells / ml using a 3D cell culture medium.

[0325] Add 100 pl of cell suspension (4 x 105live cells) into each microwell without disturbing the spheroids formed by CD14+monocytes. Periphery wells in the microplate are not used as described above.

[0326] Adding the ‘lymphocyte-rich’ cell fraction facilitates the recruitment of additional immune cell types, including those involved in the adaptive immunity in the monocyte and macrophage-dominant spheroid core.Day 6. Addition of test compounds or therapeutics

[0327] Carefully pipette 100 pl of the supernatant culture medium from the microwells without disturbing the 3D spheroids structures. Next, add test compounds in the 50 pl solution described above in workflow A.Day 12. Fluorescence intensity reading and imaging to assess antitubercular drug efficacy

[0328] Perform fluorescence intensity reading and automated plate imaging as described above in workflow A for the THP-1 bioplatform.

[0329] The workflow generates a complex, solid 3D tuberculoma-like structure incorporating multiple adaptive and immune cell subsets in microwells with co-cultures (Figure 3D). A donor- to-donor variability is observed regarding the number of granulomatous lesions formed. Since primary human blood monocyte-derived macrophages lack the self-renewing capability, granulomatous lesions might disintegrate relatively quickly after day 12 for some donors.Day 14. Additional fluorescence intensity reading for microplates with 3 / t / ?-inlecled 3D culture

[0330] Perform additional fluorescence intensity reading on day 14 as described in workflow A.Workflow for the cryo-shelf-stable 3D bioplatform

[0331] Cry opreservation is one of the most promising methods of long-term storage of cells and tissues. Therefore, we developed a cryo-stable bioplatform by freezing 3D co-cultures in situ in microplates at a cryogenic temperature. Three different time points, i.e., 30 min, 16 hr, and 72 hr post- w ‘M’ (tdTomato) infection of THP-1 cells, were tested to determine the optimal time to freeze the microplates with 3D co-cultures. We also investigated three different freezing mediums, i.e., 3D cell-culture medium containing 5% DMSO, heat-inactivated FBS containing 5% DMSO, and Lebovitz’s L-15 medium containing cryoprotective agents (see Recipes 7a, b, and c). Best results in terms of organized granulomatous lesion formation following thawing and revival were obtained when 3D co-culture was frozen within 30 min or after 16 hr of Mm infection and 3D cell culture medium containing 5% DMSO was used for cry opreservation (Figure 4A-F). However, cryopreservation of 3D co-cultures was not optimal following more extended co-culture periods, i.e., over 72 hr, when solid 3D structure, cell-to-cell junctions, and nascent lesions are formed. The optimized workflow described below details the steps required to develop a shelf-stable bioplatform in the 96-well 3D spheroid microplate.Day 1. Development and cryopreservation of 3D cell cultures

[0332] Preparing 3D spheroid microplates for co-culture and cryopreservation

[0333] Prepare THP-1 cell suspension and Mm suspension as described above in workflow A.

[0334] Resuspend the THP-1 cell pellet in the freezing medium (see Recipe 7a) to prepare a 2 x 106live cells / ml cell suspension.

[0335] A greater number of THP-1 cells than workflow A are required to compensate for increased cell death during freezing and thawing.

[0336] Resuspend the Mm M (tdTomato) pellet in the freezing medium (see Recipe 7a) to prepare an 800-1000 CFU / ml suspension.

[0337] Dispense 100 pl of THP-1 cell suspension (1 x 105live cells) into each microwell except those on the periphery.

[0338] Add 100 pl of mycobacterial suspension into each microwell except those on the periphery.

[0339] Mix thoroughly up and down three times without touching the well bottom to obtain a homogeneous suspension of THP-1 cells and mycobacteria.

[0340] Fill the periphery microwells with 250-300 pl of 3D cell culture media or sterile cell culture grade water using a multichannel pipette to avoid boundary effect.

[0341] Keep the microplates for up to 30 min at room temperature if freezing straightaway.

[0342] For ‘mycobacteria-in-spheroid’ co-cultures that were intended to be frozen after 16 or 72 hr, we incubated plates at 37 °C with 5% CO2 and 100% humidity for 16 or 72 hr (see below).

[0343] Place the microplates in cryo-boxes within 30 min post-infection or co-culture and transfer the 3D microplates into a -80 °C freezer. Keep microplates in a -80 °C freezer for 12-72 hr.

[0344] Transfer the microplates to liquid nitrogen at -160 °C to -196 °C for long-term storage. Then, transport the microplates to the liquid nitrogen storage facility on dry ice.

[0345] For ‘mycobacteria-in-spheroid’ co-cultures intended to be frozen after 16 or 72 hr, we performed the co-culture in the 3D cell culture medium in microplates as described in workflow A. After 16 or 72 hr, the cell culture medium was carefully removed using a micropipette and replaced with a freezing medium. The mycobacterial infection process in the spheroid is far advanced by 72 hr with the formation of cell-to-cell junctions and adhesions. We avoided disturbing early 3D spheroids formed while adding the freezing medium.Thawing of 3D co-culture and further culture on-demand

[0346] Thawing of 3D cell co-culture

[0347] Transport the microplates on dry ice to the cell culture laboratory from the liquid nitrogen storage facility.

[0348] Thaw frozen co-cultures in a CO2 incubator set at 37 °C with 5% CO2 and 100% humidity for 10 to 15 min.

[0349] We removed microplate lids in a sterile incubator to facilitate quick thawing. Alternatively, microplates with co-culture can be carefully thawed on a platform in a water bath set at 37 °C.

[0350] Add 50 pl of prewarmed (37 °C) 3D cell culture medium per microwell when at least half of the freezing medium ice is melted visibly.

[0351] Following complete thawing at 37 °C after 30 to 35 min, centrifuge the microplates at 430 xg for 6 min.

[0352] Carefully pipette out the supernatant freezing medium without disturbing the cell pellet.

[0353] Add fresh, 37 °C prewarmed medium (200 pl per microwell) using a micropipette and resuspend the cell pellet by pipetting up and down five times. Centrifuge the microplates at 430-450 x g for 6 min. Perform this washing step 3 times to remove the freezing medium completely.

[0354] For ‘mycobacteria-in-spheroid’ co-cultures frozen 16 or 72 hr post-infection in microplates, where compact 3D structures (and cell-to-cell junctions) are already formed, we did not centrifuge the microplates. Washing steps were performed by carefully adding pre-warmed medium (37 °C) without disturbing the 3D spheroid structures in microwells.

[0355] Resuspend the pellet in 200 pl of prewarmed (37 °C) 3D cell culture medium per microwell after the final wash.

[0356] Place the microplates in a CO2 incubator set at 37 °C with 5% CO2 and 100% humidity for further incubation and revival of 3D co-culture and the formation of granuloma lesions.Day 6 and Day 12 post-thawing. Addition of fresh 3D cell culture medium

[0357] Add prewarmed (37 °C) 3D cell culture medium (50 pl) without disturbing the 3D spheroids.

[0358] Return the microplates to the CO2 incubator to continue the growth of 3D cell co-culture.Day 18 and Day 21. Fluorescence intensity reading and imaging

[0359] 1. Perform the fluorescence intensity reading as described in workflow A.

[0360] 2. Perform imaging and lesion count as described in workflow A.

[0361] Well-organized lesions in a cryopreserved and revived ‘mycobacteria-in-spheroid’ coculture develops relatively late compared to the freshly developed co-culture system. The optimal time to treat with investigational compounds in this workflow is on day 12 or 15 postrevival to study the drug efficacy.Data analysisData processing and analysis

[0362] In previous approaches (Sable and Li et al.), the 3D cell culture workflows used a coculture of human THP-1 cells or primary CD14+monocytes with pathogenic mycobacteria. These workflows generated 3D structures analogous to solid tuberculomas without employing ECM embedding, artificial scaffolds, or magnetic levitation. The resultant solid tuberculoma bioplatform in a 96-well format was used to screen a custom library of known potential HDT drugs in proof-of-concept experiments. The efficacy of the HDT drug was measured in terms of inhibition of bacterial burdens and granuloma lesions in 3D spheroid structures. The mycobacterial load was measured by using red fluorescent Mm o Mib strains that express tdTomato or by counting CFU on Middlebrook 7H10 agar after disrupting the spheroids, lysing cells using Triton X-100 (0.1%) for 10 min and plating the lysate. The number of granuloma lesions in the 3D tuberculoma structures was measured by automated imaging and cellular analysis of stitched and Z-stacked images. The quality control of granuloma lesion count was performed by visually counting the lesions per 3D tuberculoma by three individual readers and comparing those with the numbers calculated by Gen5. We determined the homogeneity and reproducibility of 3D tuberculoma-like structures formed in the 96-well platform in several microplates and different independent experiments.

[0363] Uniformity and reproducibility of bioplatform: The diameters and areas of 3D tuberculoma-like structures generated in different microplates and batches developed by two performers were measured to determine uniformity and reproducibility. We also analyzed the bacterial burden, the number of granulomatous lesions developed in the 3D tuberculomas, and the percent area of the tuberculoma structures affected by mycobacterial growth and lesions. The resulting 3D structures showed little within and between-batch variability in size distribution, mycobacterial growth, and granulomas encompassed by individual tuberculoma structures (Figure 5). Their uniformity, reproducibility, and ease of development render them ideal for high-throughput screening applications.

[0364] Application in drug screening assay: The bioplatform allows serial quantitation of drug efficacy in situ in terms of reduction in bacterial burden using a fluorescence plate reader and the resolution of granulomatous lesions using an automated cell imaging system as described in the workflows above. The efficacy of test compounds in terms of reduction or increase in bacterial burdens was determined by treating 3D co-cultures with individual test compounds in at least triplicate wells (termed test wells). Wells treated with drug diluent or carrier (i.e., DMSO) and cell culture medium alone (negative controls) and the antibiotic rifampicin or the experimental HDT drug nitazoxanide (positive controls) were kept in each microplate (see plate map, Figure 6). A pre-defined plate map was used to identify drugs added to microwells. First, the average fluorescence intensity reading of untreated wells (n=6) was subtracted from the individual readings of test compound-treated wells (n=3). Then, the value was divided by the average fluorescence readings of untreated wells to obtain the normalized percent reduction of bacterial burden in test compound-treated wells.

[0365] Quality and suitability of assay for use in high-throughput screens: The bioplatform assay reliably detected an increase or decrease in mycobacterial burden, lesion numbers, and lesion size across a large number of microwells following treatment with test compounds. With rifampicin and nitazoxanide, we obtained consistent results of reduction in bacterial burdens with Z’ and Z-factors of >0.5 when optimized MOI of 0.008 (range 0.006 to 0.012) of M (tdTomato) was used in the microplates, suggesting the robustness of this assay for high- throughput screening (Figure 6). Please refer to example-2 described below for detailed information regarding other data processing and analysis aspects when screening potential therapeutics.

[0366] Other applications: Beyond drug screening applications, we demonstrated the utility of the tuberculoma bioplatform to simultaneously assess the effects of the test compounds on the host-cell viability, the type of cell-death induced, and the potential innate immune mechanisms of actions of the test compounds in the 3D microenvironments in situ (see example-2 described below). We also demonstrated the utility of this 3D model system to investigate the effects of biologies and biosimilars on the formation and structural integrity of granulomas and to decipher early M / / ? host interactions in solid or cavitary tuberculoma milieus.

[0367] Statistical analysis

[0368] The normalized reduction of bacterial burden (%) in test compound-treated wells was calculated using the formula below, where Ftis the fluorescence intensity of test compound- treated wells, and Mean Fuis the average fluorescence intensity of untreated tuberculoma wells (n= 6).

[0369] Percent bacterial burden reduction = (Fr-Mean Fu) / Mean Fu

[0370] The statistical differences between two groups were assessed by the Mann-Whitney test, and those between three or more groups were measured by the 1-way ANOVA and the nonparametric Kruskal-Wallis-test followed by Dunn's post-test (GraphPad Prism V9.3). A / ? < 0.05 was considered statistically significant, and *, **,*** and **** in the figures indicate p < 0.05, < 0.01, < 0.001, and < 0.0001, respectively. The quality of the high-content drug screening assay in the 3D tuberculoma bioplatform using a co-culture of THP-1 monocytes and A- / / ?? was determined using the Z’ statistic. The Z’-factor describes how well-separated the positive and negative controls are in the HTS assay without the intervention of test compounds. The Z’-factor was calculated by performing the assay in a batch of 8-10 different 3D cell culture microplates as follows: z ~ .1.where oc. is the standard deviation of the untreated or DMSO control wells, oc+ is the standard deviation of the rifampicin or nitazoxanide control wells, pG- is the mean of the untreated or DMSO control wells, and pc+ is the mean of the rifampicin or nitazoxanide control wells. Z’ value between 0.5 and 1 is considered excellent, a value between 0 and 0.5 is acceptable, and a value less than 0 indicates that the assay is unlikely to be suitable for HTS applications.Perspectives

[0371] The 3D cell cultures of pathogenic mycobacteria and human immune cells and the in vitro human TB granuloma models with high-content screening or HTS capabilities are highly informative systems. However, the available systems still need to improve simplicity, robustness, reproducibility, scalability, high throughput, and efficiency. A simplistic, high-fidelity bioplatform is highly desirable. However, the minimum composition of such a robust model to investigate fundamental disease mechanisms and the effectiveness of potential therapeutics without compromising the development of crucial tuberculoma structures and attributes has remained enigmatic. The observations using 3D cell culture of human monocytes, eitherimmortalized THP-1 or primary CD14+monocytes, with pathogenic mycobacteria, as minimum essential components in the workflows described here are very striking. This simple 3D ‘mycobacteria-in-spheroid’ 3D co-culture led to the differentiation of monocytes into epithelioid macrophages and the formation of a range of tuberculoma structures with well-organized granuloma lesions. It not only validated the findings in animal models that the construction of tuberculous granulomas occurs in the sole context of innate immunity and without the contribution of adaptive immunity (Davis et al. , 2002; Grant et al. , 2022) but provided the HTS- compatible platform system with a range of tuberculoma microenvironments for investigating the host-pathogen interactions and for TB therapeutic discoveries. This 3D tuberculoma model is developed without artificial scaffold, matrix-encapsulation, or collagen embedding of human monocytes. It does not require the incorporation of ECM in the 3D co-culture for the formation of early solid tuberculoma-like structures with hypoxic and necrotic cores. This reduces complexity in the system and helps preserve the tractability, scalability, and reproducibility of the bioplatform. Furthermore, observation in the workflow using ECM suggests that the cavity formation might also occur in the sole context of innate immunity during impaired lymphocyte access, contingent upon the amount of ECM deposited and remodeled by the innate immune cells, including macrophages and stromal cells in the tuberculomas. This observation, however, needs further conformation using in vivo models of TB cavitation.

[0372] The use of primary human cells such as PBMCs and bronchoalveolar lavage cells is preferred over cell lines in the ideal 3D model system (P. Elkington etal., 2019), but their scarcity, donor-to-donor variability, and high procurement costs make their use impractical in high-throughput screening systems. The use of human stem-cell-derived monocytes can circumvent some of these limitations and generate macrophages in large quantities (Hong etal., 2018; Han et al., 2019). However, owing to the inability of differentiated macrophages to proliferate and self-renew, their repeated incorporation in the bioplatform will be required to facilitate influx in the macrophage aggregation for the formation and maintenance of organized granuloma lesions in experiments lasting over 10 days. Immortalized macrophage or monocyte cell lines including THP-1 that originate from cancer patients are not generally considered owing to concerns that they might harbor mutations and produce immunoregulatory microenvironments that can influence Mtb interactions (P. Elkington et al., 2019). However, emerging evidence suggests striking similarities in immunoregulatory traits and microenvironments in cancer tumorsand TB granulomas (Gern et al., 2021; McCaffrey et al., 2022), which could open a new realm of HDTs and immunotherapeutic possibilities for treating TB that have been developed for cancer treatments. Larger-scale production and uniform infection of THP-1 monocytes with selfrenewal and recruitment capability to form well-organized granulomatous lesions and key tuberculoma attribute therefore might represent a simple strategy to improve throughput, reproducibility, and assay robustness in primary screenings.

[0373] Our investigations suggest that simply co-culturing human THP-1 cells or CD14+monocytes with mycobacterial strain in an arbitrary 3D cell culture ware might not lead to the formation of well-organized lesions and tuberculoma forms. Despite the use of 3D spheroid technology, primary macrophages or cell lines, and mycobacteria for co-culture (Mukundan et al., 2021a; Mukundan etal., 2021b; Kotze etal., 2021), well-organized granuloma lesions or cavitary features were not developed in recent studies, despite the generation of infected 3D spheroid structures with features common to control uninfected spheroids like central hypoxia and necrosis. We found that the spatiotemporal factors including the use of suitable 3D cell culture ware, optimal microwell geometry, co-culture surface chemistry, culture conditions, and culture timeline are critical for the tuberculoma-imitative structures and organized granulomatous lesions formations (see example 2 described below). This invention describes these factors and culture conditions along with minimum essential elements (i.e., macrophages and pathogenic mycobacteria) required to form tuberculomas in vitro.

[0374] As described above, our 3D tuberculoma bioplatform has several advantages over existing systems. These include ease of development and use, robustness, increased throughput and efficiency, reduced costs, real-time monitoring of bacterial burden, granuloma features, host cytotoxicity, and tracking of other attributes of bacterial and host cell physiology in situ. However, the limitation of our 3D tuberculoma model and workflows described herein is the absence of relevant ancillary cells in human tuberculous granulomas, such as the neutrophil subsets and non-hematopoietic cells like fibroblast, epithelial cells, and endothelial cell. Therefore, our model system could not investigate the contributions of neutrophils and stromal cells in forming necrotizing and caseous tuberculomas and cavitary transformation. To further improve our model, we have nonetheless successfully attempted incorporating these cell types in the form of human cell lines in the THP-1 bioplatform to demonstrate the desired flexibility ofour system and to investigate the effects of the addition of these cell types on granuloma organization in preliminary experiments (data not shown).

[0375] Compared with other available 3D in vitro granuloma models, such as those using matrix-encapsulation or embedment, bio-electro-spraying, and magnetic levitation, the methodology described here is simple without the need for specialized instruments or complex materials. The bio-protocol described here advances the human 3D in vitro granuloma technology and the tool for investigating heterogeneous human granuloma responses to vaccines, immunotherapies, and chemotherapies. Beyond high-throughput screening of potential HDT compounds and pathogen-targeting antimicrobials, the bioplatform might also allow the possible identification of personalized medicine using engineered primary human cells, for example, the chimeric antigen receptor (CAR)-macrophages and T cells of patients with difficult-to-treat mycobacterial infections or granulomatous diseases other than TB.

[0376] Recipes

[0377] 1. 3D cell culture medium

[0378] To prepare a 3D cell culture medium, supplement RPMI 1640 containing L-glutamine (2 mM) with 9.73-10% (vol / vol) heat-inactivated FBS, 0.88-1% (vol / vol) sodium pyruvate solution, 0.88-1% (vol / vol) HEPES buffer and sterile filter using 0.2 pm Nalgene filter assembly.

[0379] Cell growth medium

[0380] To prepare a complete cell growth medium, supplement RPMI 1640 containing L- glutamine (2 mM) with 9.63-10% (vol / vol) heat-inactivated FBS, 0.87-1% (vol / vol) sodium pyruvate solution, 0.87-1% (vol / vol) HEPES buffer, and 1% (vol / vol) Penicillin- Streptomycin solution and sterile filter using 0.2 pm Nalgene filter assembly.

[0381] Middlebrook 7H9 broth

[0382] To prepare 1 liter of Middlebrook 7H9 broth, dissolve 4.7 gm of 7H9 broth in 896 ml of distilled or deionized water. Add 0.05% (vol / vol) Tween-80 and 0.4% glycerol, mix well, adjust pH to 6.8-7.0, and filter-sterilize using 0.2 pm filter assembly. Supplement with 10% (vol / vol) ADC aseptically. Store at 2-8 °C until use. Determine sterility of supplemented 7H9 broth by incubating a small volume before use.

[0383] Extracellular matrix (ECM) solution

[0384] To prepare an ECM solution

[0385] Slowly add 1 part of chilled 1 OX PBS to 8 parts of chilled VitroCol Type 1 human collagen solution (3 mg / ml) with gentle swirling.

[0386] Adjust the pH of the mixture to 7.2-7.6 using sterile 0.1M NaOH. Mix gently by pipetting up and down. Monitor pH adjustment carefully using pH paper or a pH meter.

[0387] Adjust the final volume to 10 parts with sterile cell culture-grade water.

[0388] Maintain the mixture’s temperature at 2-10 °C to prevent gelation.

[0389] Add 20 pl of human fibronectin (0.1%) solution to 5 ml collagen mixture. Mix gently by pipetting up and down.

[0390] PBMC wash buffer

[0391] To prepare the buffer required to wash PBMCs after RBC lysis, supplement PBS without calcium and magnesium, pH 7.2-7.4, with 10% (vol / vol) heat-inactivated FBS and 2 mM EDTA. Filter buffer using sterile vacuum filter, 0.2 pM membrane units. Keep buffer cold (2-8 °C).

[0392] MACS magnetic labeling and column elution buffer

[0393] To prepare MACS buffer, supplement PBS without calcium and magnesium, pH 7.2-7.4, with 0.5% (vol / vol) heat-inactivated FBS and 2 mM EDTA.

[0394] 3D cell culture freezing mediums

[0395] 3D cell culture medium with DMSO

[0396] To prepare the freezing medium, supplement RPMI 1640-based complete 3D cell culture medium with 5% (vol / vol) DMSO. Prepare before use and filter sterilize using a 0.2 pm filter.

[0397] FBS with DMSO

[0398] To prepare the freezing medium, supplement a heat-inactivated FBS with 5% (vol / vol) DMSO. Prepare before use and filter sterilize using a 0.2 pm filter.

[0399] L15 medium with cryoprotectants

[0400] This freezing medium contains equal volumes of serum freezing medium “A” 2X and DMSO freezing medium “D” 2X. Add 100 pl of freezing medium “A” to microwells containing 3D cell culture first and then add 100 pl of freezing medium “D” before transferring the microplates to a -80 °C freezer.

[0401] To make 1000 ml of serum freezing medium “A” 2X, supplement 484 ml of L15 medium with 16 ml of IM HEPES buffer, 300 ml of heat-inactivated FBS, and 200 ml of PVP-10X stock and filter sterilize using 0.2 pm filter. To prepare PVP-10X stock, add 10% PVP (wt / vol) to IXHEPES buffered saline. Stock 10X HEPES buffered saline contains 70.7 gm of sodium chloride, 17.0 gm of glucose (dextrose), 2.0 gm of potassium chloride, 19.4 gm of sodium phosphate monobasic 2H2O, 47.6 gm of HEPES buffer, and 0.01 gm of phenol red in up to 1000 ml of cell culture grade of water and IX HEPES buffered saline is prepared by adding 900 ml of cell culture grade water in 100 ml of 10X HEPES buffered saline.

[0402] To make 1000 ml of DMSO freezing medium “D,” add 16.02 ml of IM HEPES buffer and 150.63 ml of DMSO in 833.3 ml of L15 medium and filter sterilize using a 0.2 pM filter.The shelf-life of serum freezing medium “A” and DMSO freezing medium “D” stored in glass bottles at -20°C is one year.References

[0403] Arbues, A., Kammuller, M. and Portevin, D. (2020). Generating Three-dimensional Human Granulomas in vitro to Study Mycobacterium tuberculosis-Host Interaction. Bio Protoc 10(22): e3820. ncbi.nlm.nih.gov / pubmed / 33659472

[0404] Arbues, A., Schmidiger, S., Kammuller, M. and Portevin, D. (2021). Extracellular Matrix-Induced GM-CSF and Hypoxia Promote Immune Control of Mycobacterium tuberculosis in Human In Vitro Granulomas. Front Immunol 12: 727508. ncbi.nlm.nih.gov / pubmed / 34603299

[0405] Bagcchi, S. (2023). WHO's Global Tuberculosis Report 2022. Lancet Microbe 4(1): e20. ncbi.nlm.nih.gov / pubmed / 36521512

[0406] Berry, S. B., Gower, M. S., Su, X , Seshadri, C. and Theberge, A. B. (2020). A Modular Microscale Granuloma Model for Immune-Microenvironment Signaling Studies in vitro. Front Bioeng Biotechnol 8: 931. ncbi.nlm.nih.gov / pubmed / 32974300

[0407] Bielecka, M. K., Tezera, L. B., Zmijan, R., Drobniewski, F., Zhang, X., Jayasinghe, S. and Elkington, P. (2017). A Bioengineered Three-Dimensional Cell Culture Platform Integrated with Microfluidics To Address Antimicrobial Resistance in Tuberculosis. mBio 8(1). ncbi.nlm.nih.gov / pubmed / 28174307

[0408] Birkness, K. A., Guarner, J., Sable, S. B., Tripp, R. A., Kellar, K. L., Bartlett, J. and Quinn, F. D. (2007). An in vitro model of the leukocyte interactions associated with granuloma formation in Mycobacterium tuberculosis infection. Immunol Cell Biol 85(2): 160-168. ncbi.nlm.nih.gov / pubmed / 17199112

[0409] Cadena, A. M., Fortune, S. M. and Flynn, J. L. (2017). Heterogeneity in tuberculosis. Nat Rev Immunol 17(11): 691-702. ncbi.nlm.nih.gov / pubmed / 28736436

[0410] Colom, A., Galgoczy, R., Almendros, I., Xaubet, A., Farre, R. and Alcaraz, J. (2014).Oxygen diffusion and consumption in extracellular matrix gels: implications for designing three- dimensional cultures. J Biomed Mater Res A 102(8): 2776-2784. ncbi.nlm.nih.gov / pubmed / 24027235

[0411] Cosma, C. L., Sherman, D. R. and Ramakrishnan, L. (2003). The secret lives of the pathogenic mycobacteria. Annu Rev Microbiol 57: 641-676. ncbi.nlm.nih.gov / pubmed / 14527294

[0412] Cronan, M. R. (2022). In the Thick of It: Formation of the Tuberculous Granuloma and Its Effects on Host and Therapeutic Responses. Front Immunol 13: 820134. ncbi.nhri.nih.gov / pubmed / 35320930

[0413] Cronan, M. R., Hughes, E. J., Brewer, W. J., Viswanathan, G., Hunt, E. G., Singh, B., Mehra, S., Oehlers, S. H., Gregory, S. G., Kaushal, D. and Tobin, D. M. (2021). A non-canonical type 2 immune response coordinates tuberculous granuloma formation and epithelialization. Cell 184(7): 1757-1774 el714. ncbi.nlm.nih.gov / pubmed / 33761328

[0414] Culver, G. J., Concannon, J. P. and Mac, M. J. (1950). Pulmonary tuberculomas; pathogenesis, diagnosis, and management. J Thorac Surg 20(5): 798-822. ncbi.nlm.nih.gov / pubmed / 14785019

[0415] Dartois, V. A. and Rubin, E. J. (2022). Anti -tuberculosis treatment strategies and drug development: challenges and priorities. Nat Rev Microbiol, ncbi.nlm.nih.gov / pubmed / 35478222

[0416] Davis, J. M., Clay, H., Lewis, J. L., Ghori, N., Herbomel, P. and Ramakrishnan, L. (2002). Real-time visualization of mycobacterium-macrophage interactions leading to initiation of granuloma formation in zebrafish embryos. Immunity 17(6): 693-702. ncbi.nlm.nih.gov / pubmed / 12479816

[0417] Elkington, P., Lerm, M., Kapoor, N., Mahon, R., Pienaar, E., Huh, D., Kaushal, D. and Schlesinger, L. S. (2019). In Vitro Granuloma Models of Tuberculosis: Potential and Challenges. J Infect Dis 219(12): 1858-1866. ncbi.nlm.nih.gov / pubmed / 30929010

[0418] Elkington, P., Polak, M. E., Reichmann, M. T. and Leslie, A. (2022). Understanding the tuberculosis granuloma: the matrix revolutions. Trends Mol Med 28(2): 143-154. ncbi.nlm.nih.gov / pubmed / 34922835

[0419] Elkington, P. T., D'Armiento, J. M. and Friedland, J. S. (2011). Tuberculosis immunopathology: the neglected role of extracellular matrix destruction. Sci Transl Med 3(71): 71ps76. ncbi.nlm.nih.gov / pubmed / 21346167

[0420] Figueiredo, L., Pace, R., D'Arros, C., Rethore, G., Guicheux, J., Le Visage, C. and Weiss, P. (2018). Assessing glucose and oxygen diffusion in hydrogels for the rational design of 3D stem cell scaffolds in regenerative medicine. J Tissue Eng Regen Med 12(5): 1238-1246. ncbi.nlm.nih.gov / pubmed / 29489057

[0421] Gem, B. H., Adams, K. N., Plumlee, C. R., Stoltzfus, C. R , Shehata, L., Moguche, A. O., Busman-Sahay, K., Hansen, S. G., Axthelm, M. K., Picker, L. J., Estes, J. D., Urdahl, K. B. and Gemer, M. Y. (2021). TGFbeta restricts expansion, survival, and function of T cells within the tuberculous granuloma. Cell Host Microbe 29(4): 594-606 e596. https: / / www.ncbi.nlm.nih.gov / pubmed / 33711270

[0422] Gideon, H. P., Hughes, T. K., Tzouanas, C. N., Wadsworth, M. H., 2nd, Tu, A. A., Gierahn, T. M., Peters, J. M., Hopkins, F. F., Wei, J. R., Kummerlowe, C., Grant, N. L., Nargan, K., Phuah, J. Y., Borish, H. J., Maiello, P., White, A. G., Winchell, C. G., Nyquist, S. K., Ganchua, S. K. C , Myers, A., Patel, K. V., Ameel, C. L., Cochran, C. T , Ibrahim, S., Tomko, J. A., Frye, L. J., Rosenberg, J. M., Shih, A., Chao, M., Klein, E., Scanga, C. A., Ordovas- Montanes, J., Berger, B., Mattila, J. T., Madansein, R., Love, J. C., Lin, P. L., Leslie, A., Behar, S. M., Bryson, B., Flynn, J. L., Fortune, S. M. and Shalek, A. K. (2022). Multimodal profiling of lung granulomas in macaques reveals cellular correlates of tuberculosis control. Immunity 55(5): 827-846 e810. ncbi.nlm.nih.gov / pubmed / 35483355

[0423] Grant, N. L., Maiello, P., Klein, E., Lin, P. L., Borish, H. J., Tomko, J., Frye, L. J., White, A. G., Kirschner, D. E., Mattila, J. T. and Flynn, J. L. (2022). T cell transcription factor expression evolves over time in granulomas from Mycobacterium tuberculosis-infected cynomolgus macaques. Cell Rep 39(7): 110826. ncbi.nlm.nih.gov / pubmed / 35584684

[0424] Guirado, E., Mbawuike, U., Keiser, T. L., Arcos, J., Azad, A. K., Wang, S. H. and Schlesinger, L. S. (2015). Characterization of host and microbial determinants in individuals with latent tuberculosis infection using a human granuloma model. mBio 6(1): e02537-02514. ncbi.nlm.nih.gov / pubmed / 25691598

[0425] Han, H. W., Seo, H. H., Jo, H. Y., Han, H. I, Falcao, V. C. A., Delorme, V., Heo, J., Shum, D., Choi, J. H., Lee, J. M., Lee, S. H., Heo, H. R., Hong, S. H., Park, M. H., Thimmulappa, R. K. and Kim, J. H. (2019). Drug Discovery Platform Targeting M. tuberculosiswith Human Embryonic Stem Cell-Derived Macrophages. Stem Cell Reports 13(6): 980-991. www.ncbi.nlm. nih.gov / pubmed / 31680058

[0426] Hong, D., Ding, I, Li, 0., He, Q., Ke, M , Zhu, M., Liu, L., Ou, W. B., He, Y. and Wu, Y. (2018). Human-induced pluripotent stem cell-derived macrophages and their immunological function in response to tuberculosis infection. Stem Cell Res Ther 9(1): 49. www.ncbi.nlm.nih.gov / pubmed / 29482598

[0427] Kapoor, N., Pawar, S., Sirakova, T. D., Deb, C., Warren, W. L. and Kolattukudy, P. E. (2013). Human granuloma in vitro model, for TB dormancy and resuscitation. PLoS One 8(1): e53657. ncbi.nlm.nih.gov / pubmed / 23308269

[0428] Kotze, L. A., Beltran, C. G. G., Lang, D., Loxton, A. G., Cooper, S., Meiring, M., Koegelenberg, C. F. N., Allwood, B. W ., Malherbe, S. T., Hiemstra, A. M., Glanzmann, B., Kinnear, C., Walzl, G. and du Plessis, N. (2021). Establishment of a Patient-Derived, Magnetic Levitation-Based, Three-Dimensional Spheroid Granuloma Model for Human Tuberculosis. mSphere 6(4): e0055221. ncbi.nlm.nih.gov / pubmed / 34287004

[0429] Larkins-Ford, J., Greenstein, T., Van, N., Degefu, Y. N., Olson, M. C., Sokolov, A. and Aldridge, B. B. (2021). Systematic measurement of combination-drug landscapes to predict in vivo treatment outcomes for tuberculosis. Cell Syst 12(11): 1046-1063 el047. ncbi.nlm.nih.gov / pubmed / 34469743Lee, H. S., Oh, J. Y., Lee, J. H., Yoo, C. G., Lee, C. T., Kim, Y. W., Han, S. K., Shim, Y. S. and Yim, J. J. (2004). Response of pulmonary tuberculomas to anti-tuberculous treatment. Eur Respir J 23(3): 452-455. ncbi.nlm.nih.gov / pubmed / 15065838

[0430] McCaffrey, E. F., Donato, M., Keren, L., Chen, Z., Delmastro, A., Fitzpatrick, M. B., Gupta, S., Greenwald, N. F., Baranski, A., Graf, W., Kumar, R., Bosse, M., Fullaway, C. C., Ramdial, P. K., Forgo, E., Jojic, V., Van Valen, D., Mehra, S., Khader, S. A., Bendall, S. C., van de Rijn, M., Kalman, D., Kaushal, D., Hunter, R. L., Banaei, N., Steyn, A. J. C., Khatri, P. and Angelo, M. (2022). The immunoregulatory landscape of human tuberculosis granulomas. Nat Immunol 23(2): 318-329. www.ncbi.nlm.nih.gov / pubmed / 35058616

[0431] Mukundan, S., Bhatt, R., Lucas, J., Tereyek, M., Chang, T. L., Subbian, S. and Parekkadan, B. (2021a). 3D host cell and pathogen-based bioassay development for testing antituberculosis (TB) drug response and modeling immunodeficiency. Biomol Concepts 12(1): 117- 128. www.ncbi.nlm.nih.gov / pubmed / 34473918

[0432] Mukundan, S., Singh, P., Shah, A., Kumar, R., O'Neill, K. C., Carter, C. L., Russell, D. G., Subbian, S. and Parekkadan, B. (2021b). In Vitro Miniaturized Tuberculosis Spheroid Model. Biomedicines 9(9). www.ncbi.nlm.nih.gov / pubmed / 34572395

[0433] Monteiro, R., Cameiro, J. C., Costa, C. and Duarte, R. (2013). Cerebral tuberculomas - A clinical challenge. Respir Med Case Rep 9: 34-37. ncbi.nlm.nih.gov / pubmed / 26029627

[0434] Moyes, E. N. (1951). Tuberculoma of the lung. Thorax 6(3): 238-249. ncbi.nlm.nih.gov / pubmed / 14884137

[0435] Nicolls, D. J., King, M., Holland, D., Bala, J. and del Rio, C. (2005). Intracranial tuberculomas developing while on therapy for pulmonary tuberculosis. Lancet Infect Dis 5(12): 795-801. ncbi.nlm.nih.gov / pubmed / 16310151

[0436] Pagan, A. J. and Ramakrishnan, L. (2018). The Formation and Function of Granulomas. Annu Rev Immunol 36: 639-665. ncbi.nlm.nih.gov / pubmed / 29400999

[0437] Pai, M., Kasaeva, T. and Swaminathan, S. (2022). Covid-19's Devastating Effect on Tuberculosis Care - A Path to Recovery. N Engl J Med 386(16): 1490-1493. ncbi.nlm.nih.gov / pubmed / 34986295

[0438] Prideaux, B , Via, L. E., Zimmerman, M. D., Eum, S., Sarathy, J., O'Brien, P., Chen, C., Kaya, F., Weiner, D. M., Chen, P. Y., Song, T., Lee, M., Shim, T. S., Cho, J. S., Kim, W., Cho, S. N., Olivier, K. N., Barry, C. E., 3rd and Dartois, V. (2015). The association between sterilizing activity and drug distribution into tuberculosis lesions. Nat Med 21(10): 1223-1227. ncbi.nhri.nih.gov / pubmed / 26343800

[0439] Puissegur, M. P., Botanch, C., Duteyrat, J. L., Delsol, G., Caratero, C. and Altare, F. (2004). An in vitro dual model of mycobacterial granulomas to investigate the molecular interactions between mycobacteria and human host cells. Cell Microbiol 6(5): 423-433. ncbi . nlm . nih . gov / pubmed / 15056213

[0440] Ramakrishnan, L. (2012). Revisiting the role of the granuloma in tuberculosis. Nat Rev Immunol 12(5): 352-366. ncbi.nlm.nih.gov / pubmed / 22517424

[0441] Sawyer, A. J., Patrick, E., Edwards, J., Wilmott, J. S., Fielder, T., Yang, Q., Barber, D. L., Ernst, J. D., Britton, W. J., Palendira, U., Chen, X. and Feng, C. G. (2023). Spatial mapping reveals granuloma diversity and histopathological superstructure in human tuberculosis. J Exp Med 220(6). ncbi. nlm. nih. gov / pubmed / 36920308

[0442] Sochocky, S. (1958). Tuberculoma of the lung. Am Rev Tuberc 78(3): 403-410. ncbi.nlm.nih.gov / pubmed / 13571599

[0443] Strydom, N., Gupta, S. V., Fox, W. S., Via, L. E., Bang, H., Lee, M., Eum, S., Shim, T., Barry, C. E., 3rd, Zimmerman, M., Dartois, V. and Savic, R. M. (2019). Tuberculosis drugs' distribution and emergence of resistance in patient's lung lesions: A mechanistic model and tool for regimen and dose optimization. PLoS Med 16(4): el002773. ncbi.nlm.nih.gov / pubmed / 30939136

[0444] Takaki, K., Davis, J. M., Winglee, K. and Ramakrishnan, L. (2013). Evaluation of the pathogenesis and treatment of Mycobacterium marinum infection in zebrafish. Nat Protoc 8(6): 1114-1124. ncbi.nlm.nih.gov / pubmed / 23680983

[0445] Tezera, L. B., Bielecka, M. K., Chancellor, A., Reichmann, M. T., Shammari, B. A., Brace, P., Batty, A., Tocheva, A., Jogai, S., Marshall, B. G., Tebruegge, M., Jayasinghe, S. N., Mansour, S. and Elkington, P. T. (2017a). Dissection of the host-pathogen interaction in human tuberculosis using a bioengineered 3-dimensional model. Elife 6. ncbi.nlm.nih.gov / pubmed / 28063256

[0446] Tezera, L. B., Bielecka, M. K. and Elkington, P. T. (2017b). BioelectrosprayMethodology for Dissection of the Host-pathogen Interaction in Human Tuberculosis. Bio Protoc 7(14). ncbi.nlm.nih.gov / pubmed / 28904991

[0447] Urbanowski, M. E., Ordonez, A. A., Ruiz-Bedoya, C. A., Jain, S. K. and Bishai, W. R. (2020). Cavitary tuberculosis: the gateway of disease transmission. Lancet Infect Dis 20(6): el 17-el 28. ncbi.nlm.nih.gov / pubmed / 32482293

[0448] Vogt, G. and Nathan, C. (2011). In vitro differentiation of human macrophages with enhanced antimycobacterial activity. J Clin Invest 121(10): 3889-3901. ncbi.nlm.nih.gov / pubmed / 21911939

[0449] Wallis, R. S., O'Garra, A., Sher, A. and Wack, A. (2022). Host-directed immunotherapy of viral and bacterial infections: past, present and future. Nat Rev Immunol. ncbi.nlm.nih.gov / pubmed / 35672482

[0450] Wells, G., Glasgow, J. N., Nargan, K., Lumamba, K., Madansein, R., Maharaj, K., Hunter, R. L., Naidoo, T., Coetzer, L., le Roux, S., du Plessis, A. and Steyn, A. J. C. (2021). Micro-Computed Tomography Analysis of the Human Tuberculous Lung Reveals RemarkableHeterogeneity in Three-dimensional Granuloma Morphology. Am J Respir Crit Care Med 204(5): 583-595. ncbi.nlm.nih.gov / pubmed / 34015247

[0451] Wetscherek, M. T. A., Sadler, T. J., Lee, J. Y. J., Karia, S. and Babar, J. L. (2022). Active pulmonary tuberculosis: something old, something new, something borrowed, something blue. Insights Imaging 13(1): 3. ncbi.nlm.nih.gov / pubmed / 35001143.Example 2High-Throughput Screening-Compatible 3D Tuberculoma Bioplatform Enables EfficientIdentification and Characterization of Host-Directed Therapeutics for TuberculosisIntroduction

[0452] Researchers employ cell culture systems to discover the basic tenets of life, hostpathogen interactions, and disease interventions. Tuberculosis (TB), an age-old infectious disease caused by an obligate human intracellular pathogen Mycobacterium tuberculosis (Mtb would greatly benefit from the in vitro cell-culture model that recapitulates critical features of tubercular lesions in a high-throughput screening (HTS)-compatible format. Despite extensive prophylactic and therapeutic interventions, TB remains among the leading infectious killers worldwide, with - 1.6 million deaths annually, emphasizing the need for improved control measures and tools. The incomplete understanding of host-M / > interactions and the pathogen elimination within the hallmark tubercular lesions has impeded the efforts to develop effective control measures. TB exhibits a complex spectrum of pathology and diverse clinical presentations in humans. The defining feature of TB pathology is the formation of granulomas, tightly organized macrophage-rich aggregates in the infected tissues, most often in the lungs. A pulmonary granuloma begins with the phagocytosis of inhaled Mtb bacilli by macrophages, followed by the extravasation and recruitment of additional leukocytes at the site of infection from blood. Beyond the macrophage-rich scaffold, the granuloma structure is shaped by other myeloid, lymphoid, and ancillary cells. While TB granuloma is developed in response to the offending pathogen by the host immune cells to isolate and control infection, the structure can also promote bacterial growth and dissemination. Granuloma is therefore considered the epicenter of the host-pathogen battle that determines the infection outcome.

[0453] TB granulomas are highly dynamic structures and can exhibit heterogeneous morphologies, bacillary burdens, and trajectories, even within the lungs of a single individual. Over time, they undergo contraction, expansion, or coalescence with nearby granulomas. In severe disease, consolidation of granulomas into a 3 -dimensional (3D) firm mass, termed a ‘tuberculoma,’ can occur, resembling solid cancers. It develops a hypoxic and necrotic center over time with a surrounding wall of a florid granulomatous response containing macrophages in different stages of transformations and the lymphocyte-rich cuff on the periphery. Pulmonary tuberculomas can become mineralized or cavitary and breach into airways, releasing and transmitting Mtb. Activated T lymphocytes entering granulomas contribute to the control of infection. However, the adaptive responses in the lungs of / W?-infected human hosts and animal models, particularly within the pulmonary granulomas, are delayed relative to other respiratory pathogens. Furthermore, the infiltration of lymphocytes into the granuloma core and their interaction with infected macrophages is restricted by numerous barriers, including epithelioid and foamy transformations of macrophages.

[0454] While CD4+T helper-1 lymphocyte response is necessary for protection, a paradox is that this response is insufficient, and the aberrant lymphocyte activity can promote immunopathology. Moreover, recruited CD4+T lymphocytes are rapidly depleted from TB granulomas in acute HIV / SIV coinfections, leaving macrophage and other phagocytes as the focal point of much of host-M / > interactions. Growing evidence suggests that macrophage-rich granulomas can contain Mtb infection in some hosts, despite having few or no CD4+T lymphocytes. Nevertheless, Mtb employs multiple immune evasion strategies in macrophages and granulomas and frequently succeeds in establishing long-term infection. Emerging understandings in TB granuloma biology highlight the crucial role of cell states and interactions in granulomas in shaping infection outcomes. The temporospatial segregation of immune cell subsets and signaling in the granulomas implies that the emerging therapeutics must penetrate and act within the macrophage-rich center. In diverse granuloma forms, bacteria are present in disparate microenvironments and physiological states, which influence their susceptibility to antibiotics, thereby making TB drug discovery and development a challenging endeavor.

[0455] One key strategy to meet the goal of TB elimination is the development of improved and shorter treatment regimens. The standard antibiotic treatment for TB is lengthy and can range from 6 to 24 months, depending on the drug-susceptible or resistant disease. This pathogen-targeted approach has the potential disadvantage of selecting for microbial resistance. An alternative approach termed host-directed therapy (HDT) has received much attention lately. It involves therapeutic modulation of host immune responses to augment immunity or targeting the host cellular processes the pathogen exploits for survival. HDT is a transformative approach that has the potential to shorten the treatment and decrease the probability of developing microbial resistance, ameliorate immunopathology, and prevent relapse. The past decade has witnessed considerable progress in preclinical investigations of potential HDTs, and a handful of therapeutics have progressed to clinical trials. However, the need for a physiologically relevant, high-fidelity in vitro platform to rapidly identify potential HDTs and predict their therapeutic efficacy has hindered significant discoveries and translational advances.

[0456] Although animal models are the preclinical tool of choice to understand the effect of potential therapeutics, they fail to reproduce the complete spectrum of human TB disease. Furthermore, low throughput and high costs of animal-based screening regimes in the high- containment facility create a significant bottleneck for drug discovery. To replace animal testing, multiple tools, and methodologies have been developed. Yet, the complexity of human granulomas means that the traditional broth or monolayer cell-culture systems seldom mimic the full spectrum of granuloma microenvironments.

[0457] Over the past decade, due to their physiological relevance, human in vitro granuloma models bioengineered using primary cells have found a rapid following. Early granuloma models developed using traditional monolayer culture of peripheral blood mononuclear cells (PBMCs) develop some traits in human granulomas, such as macrophage aggregation and multinucleated giant-cell formation. However, the available models lack well-organized granuloma structures, diverse forms, and microenvironments. As a result, some drugs (for example, pyrazinamide) that are effective in human TB granulomas cannot be identified using these 2-dimensional (2D) systems. Recently developed 3D cell culture models have raised the hopes of better predicting drug efficacy and providing new insights in early host-M > interactions, as they mimic the granuloma microenvironment more closely than 2D systems. However, these 3D models, despite generating granulomatous aggregation of macrophages, fail to develop a classic tuberculoma structure and features such as well-organized lesions and a spectrum of solid, necrotic, and cavitary transformations. Current 3D granuloma models bioengineered using extracellular matrices (ECMs), microencapsulation, or magnetic levitation are complex and need specializedinstrumentation for development. Their use in HTS applications is fettered by low throughput, restricted tractability, and limited scalability. In addition, releasing viable cells for downstream investigations from these systems requires extra manipulations like enzymatic treatment, which affect their accurate characterizations. A new explant 3D culture technique of micro-dissected granulomas from the infected animals can overcome these limitations of in vitro models. However, this system is technically demanding, and very few granulomas can be micro-dissected from infected animals. Since biopsies, leukapheresis, or bronchoalveolar lavage do not generate enough macrophages from a single human donor, available 3D models employing primary cells, although useful in late-stage discovery, are impractical for primary lead generation efforts. A 3D model, analogous to classical TB lesions, is required, which can be scaled up as an HTS- compatible platform to test thousands of potential therapeutics rapidly, efficiently, and economically.

[0458] Here, we report the development of a robust bioplatform that can be used for wide- ranging applications, from decoding the host-M / > biology and identifying associated transcriptional changes to screening multiple compounds and biologies in vitro (Fig. 7 A). This physiologic system uses a 3D co-culture of primary human cells or monocytic cell lines with virulent mycobacteria to mimic the natural microenvironment. The advantage of this system is the generation of 3D structures that replicate critical traits and conditions Mtb encounters within classic solid, necrotic, or cavitary tuberculomas in TB patients that are not previously demonstrated using available in vitro granuloma models. The system employs the workflow in a standard 96-well format that combines generation, maintenance, fluorescence readings, immunostaining, and high-content imaging of the resulting tuberculoma-like structures in situ. This platform allows the screening and characterization of immune mechanisms induced by putative HDT compounds that preferentially act on the host cells and inhibit the pathogen growth in 3D tuberculomas. It also allows the screening of antimicrobials or compounds with both pathogen-targeting and host-directed activity. We show that a biosafety level (BSL)-2 surrogate pathogen, M. marinum, instead o Mtb, can be used in this model to prescreen compounds, which saves hundreds of staff hours in the BSL-3 laboratory while potentially reducing safety concerns. In proof-of-concept experiments, we used this platform to screen a custom library of potential HDT compounds identified in the published studies using 2D cell cultures infected with mycobacteria. We uncovered select ‘top hit’ compounds substantially reducing the granulomalesions and bacterial burdens in 3D in vitro tuberculomas. Although the antibacterial effects of compounds varied across the mycobacterial strains, identified ‘top hit’ compounds with diverse annotated pathway targets consistently reduced bacterial loads in 3D tuberculomas, regardless of theM / > strain employed. Investigations of pharmacological activation of innate defense mechanisms in situ in the 3D tuberculomas revealed that several hit compounds induced rapid autophagy flux in the macrophages. Most of these drugs are approved by the FDA or have been studied in clinical trials for other indications and hence can be repurposed for TB treatment. Notably, we showed that one of the ‘top hit’ drugs and clinically used multi-kinase inhibitors with potent aurora kinase inhibitory activity, AT9283, has significant efficacy for inhibiting Mtb and lesion burdens in 3D in vitro tuberculomas and mice. The platform described here advances “3D human in vitro granuloma technology” and provides the HTS-compatible physiological tool for discovery programs of tuberculosis and other granulomatous diseases.ResultsA ‘mycobacteria-in-spheroid’ co-culture establishes a facile human 3D tuberculoma-like model.

[0459] To enable the identification of host-directed therapeutics and their mechanisms of action controlling intracellular mycobacterial survival, we sought to develop a simple 3D in vitro granuloma model that can emulate the classic tubercular lesions and microenvironment in a standard 96-well microplate. Yet, the minimal composition of such a simplistic model that can permit long-term experiments and develop dynamic granuloma milieus, simulating prolonged host-pathogen interactions, is still awaiting definition. 3D cell culture in the form of spheroids and organoids has long been used to model complex microenvironments in human cancers and found utility in HTS applications. Given the significant overlap between the structures and microenvironments in cancer tumors and tuberculous granulomas, we generated a selfassembling ‘mycobacteria-in-spheroid’ co-culture in 96-well 3D cell culture plate using pathogenic mycobacteria and human macrophages, which are central to the granuloma formation. Because increasing complexity in 3D cell culture can undermine reproducibility, pliability, and throughput, we omitted ECM embedding in early iterations to reduce batch-to- batch variability. Since macrophages can secrete ECM naturally, we generated co-culture exempt from scaffolding elements or exogenous biomaterial. Furthermore, incorporating external ECMin the 3D granuloma model can reduce the size of granulomatous aggregates formed and Mtb proliferation despite prolonging macrophage survival compared to the 2D monolayer culture.

[0460] To simplify workflow, increase efficiency, and reduce potential safety concerns associated with the experimentation in the BSL-3 laboratory, we utilized M. marinum (Mm), a risk group-2 pathogen and genetically close relative oiMtb, in the BSL-2 laboratory. Mm has been extensively studied in the zebrafish model to understand tuberculous granuloma biology as a surrogate for Mtb. We first assessed two different Mm strains, strain 1218 and ‘M,’ in the 3D co-culture. Mm 1218 and ‘M’ constitutively expressed widely used fluors GFP (green) and tdTomato (bright red) to track bacterial -growth dynamics and were initially isolated from the infected fish and human skin lesions, respectively. They have been successfully used to investigate host- w interactions in 2D macrophage culture and zebrafish models. Since Mm grows well naturally at 30 °C in poikilothermic animals, we first compared the ability of two strains to grow at 37 °C in vitro. We found that these two strains grew well at 37 °C in the Middlebrook 7H9 medium, although relatively slower than at 30 °C, and showed robust and comparable growth kinetics over three weeks (FIG. 8A). However, unlike the 7H9 growth medium, both strains grew very slowly in the RPMI-1640, a 3D cell culture medium, without macrophages.

[0461] To overcome the barriers to HTS applications, particularly concerning the paucity of adequate numbers of primary macrophages from human donors, we tested the growth of two Mm strains in several human and mouse monocyte-macrophage cell lines, including THP-1 cells, U937, J774A.1, RAW-264.7, and AMJ2 cells. We used a low multiplicity of infection (MOI) of 0.008 (range 0.005-0.01) and did not wash co-cultures after mixing cells and bacteria in the microwells. Since pinocytosed aminoglycosides can reach macrophage phagosomes and contribute to macrophages’ apparent antibacterial activity, we avoided using aminoglycoside antibiotics believed to kill only extracellular bacteria in the cell culture. The workflow also omitted using mitogens (for example, PMA) for macrophage differentiation, which affects cellular signaling and other cellular processes and could alter host-pathogen interactions. We used an automated multimode plate reader and confocal cell imager for fluorescence monitoring, microscopy, and 3D co-culture imaging in situ to investigate host and bacterial cell dynamics over time. Among cell lines screened, human THP-1 monocytes and U937 histiocytes (tissue macrophages) allowed for the most consistent and robust bacterial growth in 3D co-culture (FIG.8B and 8C). Conversely, mouse macrophage cell lines J774A.1, RAW-264.7 and AMJ2 allowed relatively poor growth oiMm strains over 12 days. Interestingly, cellular aggregates resembling well-organized granulomatous collections of macrophages that support mycobacterial growth were observed only in the infected 3D cultures of THP-1 cells. No aggregates were formed in the uninfected control 3D cell cultures.

[0462] Given these exciting findings, we next carried out a detailed investigation of the spatiotemporal dynamics of THP-1 monocyte and mycobacterial proliferation in the 3D coculture. Real-time visualization and serial imaging of events following co-culture revealed that by day 2, solitary spheroids began to form in the microwells (Fig. 7B and 7C). Since this coculture used a low MOI of Mm, inoculated bacteria are completely gathered by amassing monocytes in the 3D spheroid formation process, and the bacteria subsequently proliferate primarily within the spheroids. After three days of incubation, compact spheroids are formed, which increase in size by day 6. Cellular aggregates develop around days 6 and 8 in the spheroids infected with strain ‘M’ and 1218, respectively. These cellular aggregates grow over time, and large granulomatous aggregates supporting robust bacterial growth are formed by day 12 and day 16. The cellular aggregates are primarily developed in the oxygen-rich zone, considering the maximum Ch diffusion distance through tissues is around 250 pm, beyond which tissues become hypoxic. At this stage, the 3D ‘mycobacteria-in-spheroid’ structure resembles a conglomeration of well-organized, florid granulomatous foci in the solid tuberculoma (Fig. 7B and 7C)

[0463] Corresponding with the growth of cellular aggregates, bacteria fluorescence, which was observed to increase gradually until day 6, increased dramatically once granulomatous aggregates started to expand (Fig. 7D). This observation, in agreement with previously reported finding in the zebrafish model, indicated that the cellular aggregates formed are likely genuine granulomas and mycobacterial expansion coincide with the granuloma formation and growth in the 3D tuberculoma-like model. Plating THP-1 lysate on the 7H10 agar plates and quantifying colony-forming units confirmed this increased bacterial burden over three weeks. The advantage of the model developed without ECM embedding is that the mature 3D spheroids can be readily dissociated by pipetting, and cells can be released for downstream applications without additional treatment and potential changes in their established characteristics. Contrary to the increased bacterial burden, the survival of THP-1 cells in the .Mw-infected 3D spheroids wasfound to decrease more compared to the uninfected spheroids progressively over three weeks, as investigated by the cytotoxicity assay (Fig. 7E). Taken together, these results suggest that the in vitro model can permit relatively long-term and quantifiable characterization of the host and pathogen growth dynamics within a physiologically relevant 3D microenvironment.

[0464] While investigating the ‘mycobacteria-in-spheroid’ co-cultures using Mm strains, we observed that among the two fluors tested, tdTomato provided a brighter signal and less autofluorescence noise, consistent with the established observation of less biological autofluorescence at longer wavelengths in diverse cell types and tissues. The fluorescence intensity (FI) readings further revealed more background fluorescence in the uninfected control spheroids in the GFP channel compared to the tdTomato channel (Fig. 8C). Therefore, we selected the Mm strain expressing tdTomato for further FI readings experiments. Importantly, we found a linear relationship between FI and the number of bacteria in the 3D spheroids in 96-well plates or liquid mediums. This confirmed that the fluorescence reading in our model is a dependable indicator of relative bacterial number (Fig. 8D). Co-culturing THP-1 monocytes and Mm in an arbitrary 3D cell-culture ware did not always generate a tuberculoma-like structure with well-organized lesions. It required unique spatial features in a 3D cell-culture ware. When we screened several vendor-provided 96-well plates with distinct microwell-surface chemistry and bottom geometry, spheroids with large granulomatous lesions were formed only in the 3D spheroid microplates with ultralow attachment (ULA)-surface and U-bottom but not in the microplates with V-bottom or surface with microcavities (Fig. 9). Notably, the granulomatous aggregates formed in the 2D microplates or 3D microparticles encapsulating w-infected monocytes were miniature and loosely organized. Among the microplates tested, well-organized, compact, and relatively stable lesions were exclusively formed in Coming ULA 3D spheroids plates with U-bottom and neutrally charged hydrophilic surfaces with covalently bonded inert hydrogel coating. Therefore, we selected Corning 3D spheroid plates for future experiments.

[0465] Human TB granulomas exhibit divergent trajectories and fates. So, we sought to investigate the spatiotemporal evolution of granulomas in our model. We used green-fluorescing THP-1 (THP-l-GFP) cells with Mm ‘M’ tdTomato or red-fluorescing THP-1 (THP-l-RFP) with Mm 1218 GFP for 3D co-culture. We observed dynamic lesion formation and coalescence with neighboring lesions in the 3D model (Fig. 10A). While some lesions remained relatively stable, others progressed and contained higher bacterial burdens. On day 16 post-infection, viable(fluorescing) monocytes supporting bacterial expansion were observed almost exclusively in the granulomatous aggregates, withstanding a marked decrease in survival in the remainder of the spheroid, indicating that mycobacteria either recruit new viable monocytes or promote permissible monocyte survival in these granulomatous aggregates for their benefit. Comparable well-organized lesions were developed withMZi strains, H37Rv, Erdman, Beijing F2, and CDC 1551 (tdTomato), investigated using low MOI (0.01). However, no well-organized lesions were formed with live attenuated Bacillus Calmette-Guerin (BCG) vaccine strains of bovis, although 0 to 5 loose cellular aggregates were observed by day 12 (Fig. 7F). Of note, the virulence locus ESX-1 / RD1 is present in Mm and Mtb strains investigated, but the BCG strains, Danish (WT) and Pasteur (GFP), had lost the RD1 locus, which explains their reduced virulence. When a 10-fold higher MOI of BCG strains was used, more cellular aggregates were formed, but they were miniaturized and loosely organized. Thus, this model could independently validate the established observations in the zebrafish model that mycobacterial ESX-1 / RD1 is required to accelerate macrophage recruitment in developing granulomas, and ARD1 infection results in defects in granuloma formation and expansion. A central feature of granulomas in TB patients is the epithelioid transformations of macrophages with aggregation, interdigitation, and development of E-cadherin-positive junctions. So, we asked whether the granulomatous aggregates formed were mere macrophage aggregations or bona fide granuloma lesions. Like in tuberculous granulomas, we detected expression of the canonical epithelial markers and celladhesion molecules, stained for the cadherins and intercellular cell adhesion molecule-1 (ICAM- 1), primarily in the developing aggregates in the Mtb Erdman (WT)-infected spheroids (Fig. 7G and H). Importantly, cells within the nascent granulomas displayed epithelioid-mesenchymal transformation, with greater expression of cell-adhesion and junction markers, E-cadherin, N- cadherin, ICAM-1, and plakoglobin (y-catenin), which validates that the cellular aggregates formed in the model are indeed granulomas (Fig. 10B and C).

[0466] We next developed 3D co-cultures using human PBMCs and pathogenic mycobacteria, but only small granulomatous aggregates were formed in this model (Fig. 71). In contrast, well- organized granulomas developed when primary CD14+monocytes purified from the PBMCs were used for co-culture withMZ) Erdman (tdTomato) and later supplemented with lymphocyterich CD14" PBMC-fraction following 3D spheroid formation (Fig. 7J). However, the granulomas developed in this model using primary monocytes, lacking self-renewal and replenishmentcapability, progressed quickly and disintegrated by day 14. This observation is analogous to the finding in the zebrafish granulomas that exhibited accelerated macrophage death by necrosis following a reduction in macrophage supply below a critical threshold and curtailed replenishment in granulomas. Nevertheless, these results demonstrate that a facile 3D co-culture of human monocytes and pathogenic mycobacteria in a rationally selected culture ware is enough to produce a tuberculoma-like structure with a conglomeration of well-organized granulomas and validate the long-standing observation that tuberculous granulomas are formed in the sole context of innate immunity. The ‘mycobacteria-in-spheroid’ co-culture described here establishes a robust cellular model that mimics human tuberculoma, and the events in the natural hosts have numerous parallels in this model.Crucial attributes and microenvironments in human tuberculomas develop in the 3D model.

[0467] Like in TB patients and animal models, cells in the granulomas and core of 3D tuberculomas generated using Mm (WT)-infected THP-1 monocytes matured into macrophages with greater expression of CD68, cytoskeleton marker tubulin, and an angiogenesis marker vascular endothelial growth factor receptor-2 (VEGFR-2) (Fig. 11A and Fig. 12A). Furthermore, cells within the infected spheroids exhibited more expression of activation or immunoregulatory marker galactin-9 compared to control spheroids. More deposition of collagen type 3, and .W / A- biofilm formation, stained for biofilm-matrix cellulose, was found in the core of infected spheroids (Fig. 12 A and 12B). These results show that monocytic cell maturation, differentiation, activation, and immunoregulation occur in the 3D model.

[0468] Mtb evades killing by inhibiting macrophage innate defense mechanisms, including phagosome-lysosome fusion, phagosome acidification, autophagy, and inflammasome activation. Therefore, to monitor autophagy induction in 3D in vitro tuberculomas, we generated co-cultures of THP-1 -Difluo hLC3 reporter monocytes and Mtb Erdman (WT). These reporter cells express human autophagy protein LC3B fused with two fluorescent proteins, GFP and RFP. Since GFP (acid-sensitive) is quenched or degraded in the acidic environment of autophagolysosome, whereas RFP remains stable, a block in the autophagy process results in higher yellow signals (red plus green co-localization) compared to autophagy induction (red signals). On day 7, we observed the accumulation of more yellow-green than red puncta in the outer zone where granulomas are formed compared to the spheroid core, indicating that the autophagy induction isspecifically inhibited or blocked in the AAA-permissive macrophages of granulomas in the model (Fig. 11A and Fig. 13A). Consistent with these results, we also observed a relatively more acidic environment in the spheroid core than in the zone of granuloma development. Since Mtb induces assembly of the apoptosis-associated speck-like protein containing a CARD domain (ASC)- dependent inflammasomes, NLRP3 or AIM2, in the infected macrophages, we used 3D coculture oiMtb Erdman (tdTomato) and THP-l-ASC-GFP inflammasome reporter monocytes that stably express a gene encoding ASC-GFP fusion protein to monitor ASC-dependent inflammasome formation. On days 7 and 9, we observed more ASC-GFP speck formation and inflammasome activation, specifically in the spheroid-scaffold cells without pronounced Mtb growth and the epithelioid macrophages in those granulomas that appear to attempt Mtb containment compared to granulomas permitting unfettered Mtb growth (Fig. 11A and Fig. 13B). These results indicate functional diversity in granuloma lesions in the 3D model and suggest autophagy and inflammasome inhibition in the AAA-permissive granulomas.

[0469] Mature tuberculomas can develop hypoxic, necrotic centers and cavitary morphology over time. So, we sought to investigate these attributes and microenvironments in the 3D model. Using different hypoxia probes that fluoresce in the hypoxic milieu, we found more hypoxia in the centers compared to the periphery of the spheroids infected with WT Mm ox Mtb, and the hypoxia increased more in the centers of infected compared to the control spheroids gradually over time as they matured (Fig. 11A and Fig. 14). These results show that granuloma lesions supporting bacterial proliferation preferably develop in the oxygen- sufficient region. Although a hypoxic milieu is generated inherently in both control and infected spheroids, owing to the sizable 3D cellular structure with >2000 pM diameter, macrophage-transformations with granuloma conglomeration likely contribute to hypoxia increase in the solid tuberculoma form. The induction of hypoxia in the AAA-infected macrophages plays a protective role by stimulating the microbicidal capacity of macrophages. Still, excess hypoxia and responsive HIF-1 can cause pathogenic necrosis and matrix destruction. Likewise, we found more cell death, characterized mainly by necrosis than apoptosis in the central region using cell-death probes, which increased over time in the infected compared to the control spheroids (Fig. 14).

[0470] Central necrosis and ECM destruction are considered the crucial pathological drivers of cavitation in human tuberculomas. Despite increased matrix metalloproteinase (MMP) activity and matrix-remodeling in the 3D in vitro granuloma models with exogenous collagen-rich ECMaddition, cavitation has not been described. Hence, we next investigated the effect of ECM inclusion in the 3D tuberculoma model. We used the Mm 1218 (GFP) strain, which permitted relatively longer (> 50 days) co-culture with THP-1 monocytes when low MOI (0.005) was used, and 50% of the 3D culture medium was exchanged with a fresh medium once per week. To simulate increased ECM deposition in human tuberculomas, we tested increasing quantities (from 5 to 50 pl) of ECM mixture comprised of human collagen type 1 and fibronectin. The mixture was added on day three post-coculture once spheroid structures were formed. Interestingly, cavity-like features developed in the infected spheroids with higher quantities (> 30 pl) of ECM after 21 days of co-culture. Despite the inclusion of > 30 pl of ECM mixture, cavitary morphology was not formed in the uninfected spheroids (Fig. 11C and Fig. 15) or infected 3D cultures generated using a 10-fold lower MOI, which subsequently developed relatively lower bacterial and granuloma burdens. These results warrant further in vivo investigations and support the premise that enhanced deposition of ECM beyond the threshold limit with concurrently increased mycobacterial and coalescing granuloma loads likely contribute to cavity formation in the tuberculomas. The results demonstrate solid, necrotic, and cavitary transformations in the tuberculoma model.

[0471] Tumor necrosis factor (TNF) has long been known as a critical cytokine required to maintain granuloma integrity by restricting bacterial growth and preventing macrophage necrosis. Yet, excess TNF can trigger programmed necrosis in infected macrophages. Therefore, we next investigated the effect of human TNF blockers, etanercept, a fusion protein of TNF receptor type II (TNFR2) fused to Fc fragment, and a human anti-TNF monoclonal antibody (mAb) adalimumab biosimilar on the bacterial burden and granuloma architecture (Fig. 11D). To investigate the high-content screening (HCS) potential of the 3D model in the 96-well format, we simultaneously screened 16 additional human recombinant mAb or antagonist protein biosimilars (Fig. 16), equivalent to those approved to treat cancers, inflammatory disorders, autoimmune diseases, and other conditions. To account for differences in the bioavailability, half-life, and pharmacologic properties of immunotherapeutics, 3D co-cultures of THP-1 cells and pathogenic mycobacteria were exposed to these biologies or biosimilars immediately after the co-culture, or on day 6, as a single treatment. We chose a high (500 ng) and low (62.5 ng) dose for the treatment after a dose titration in the preliminary experiment. Treatment with TNF blockers but not with the isotype or polyclonal control antibodies exhibited a loss of organization integrity(blue arrows) in the developing granulomas on day 10 in the 3D co-cultures using Mtb Erdman (tdTomato) (Fig. 11D). This loss of integrity occurred regardless of apparently comparable bacterial burdens in TNF blocker-treated and control co-cultures (white arrows; Fig. 11D), treatment dose used, or timing of treatment exposure (Fig. 16A). In the spheroids infected with Mm M (tdTomato), exposure to anti-TNF mAbs interestingly resulted in multiple miniature (miliary) granulomas (Fig. 16B).

[0472] The mAb biosimilar targeting integrin CD1 la, a cell-surface molecule involved in cellular adhesion, but not the CD52, a molecule with anti-adhesion function, caused the early loss of granuloma organization and integrity and reduction in Mtb burden by day 10 (Fig. 16A). Yet, relatively normal, mature granuloma formations with exuberant and comparable Mtb loads later progressed by day 14 in the 3D spheroids exposed to TNF antagonists, anti-CDl la mAbs, or control antibodies (Fig. 16C and D), which independently reaffirm previous findings using TNF antagonists in the animal models. These results, which could not be reproduced in the ECM-embedded 3D granuloma model, validate that TNF is required for nascent granuloma integrity rather than formation. The finding that integrin CD1 la expression on monocytes / macrophages contributes to the integrity of nascent granulomas and early bacterial containment is novel and needs further in vivo investigation. Additionally, exposure on day 6 with several biosimilars, such as those targeting vascular endothelial growth factor (VEGF), integrin a.4p7, CD30 (TNFRSF8), insulin growth factor-1 receptor (IGF1R), IL-6Ra, IL-1 p, or IL-1R, appeared to slow the expansion of granuloma lesions as evident by their smaller size on day 10 (Fig. 16E). In contrast, other biosimilars targeting CTLA-4 and CD52 seemed to aid in expanding granulomas compared to isotype control mAbs. However, after exposure on day 6, only anti-VEGF and anti-a4p7 (500 ng) mAbs significantly inhibited, whereas anti-CD52 (62.5 ng) mAbs increased the Mtb loads relative to controls on day 14 (Fig. 16F). These results confirm previously published observations in animal models that targeting the pro-angiogenic molecule VEGF reduces mycobacterial burden in granulomas. They also indicate that the interception of angiogenesis signaling via VEGF and pro-adhesion activity via integrin a407, in the exclusive context of myeloid cells without the presence of endothelial cells or adaptive immunity, can limit granuloma expansion and mycobacterial growth, which warrant further investigations. Nevertheless, these results collectively demonstrate that human tuberculoma-likefeatures and microenvironments develop in the 3D in vitro tuberculomas and indicate the HCS potential of the model.Protein biomarker profiling by multiplex arrays provides insight into the host-pathogen interactions leading to cavitary transformation

[0473] To identify proteins involved in the host-pathogen interactions leading to granuloma formation and cavitation in the 3D model, cell lysates and culture supernatants from the 3D cell cultures were subjected to Quantibody Human Kiloplex Proteomics Assay (Fig. 17). A total of 1,000 key proteins related to cell activation and differentiation, signal transduction, cell adhesion, angiogenesis, cytokines and chemokines, immune response, apoptosis, and cell death were investigated. Relative fold-change (FC) in protein expression of infected 3D co-culture samples compared to uninfected controls in a cell lysate or cultural supernatant, with or without adding an ECM, was determined. This investigation aimed to identify the modulation of protein expression levels and discover the effect of ECM on the change in protein profde leading to cavitary transformation in our model. The raw values of 1,000 protein markers were determined, and the ‘top’ up-regulated or down-regulated proteins in the cell lysates and cultural supernatants are presented as dot plots in Fig. 17A and B. The size of the dots in the figure represents the relative fold-change in protein expression, while the red and blue colors indicate up- and downregulation, respectively. Among the top six up-regulated markers in the cell lysates of 3D coculture without ECM addition (Fig. 17A), MCP-2 showed the highest expression levels (foldchange 250.2) compared to controls. Other top-up-regulated proteins in the cell lysates include chemokines MCP-2, MIG, MIP-3a, LTAC, and GROa. These are previously well-recognized biomarkers of the early innate immune responses following human mycobacterial infections.Another top-up-regulated protein nucleoporin 85 (NUP85), with a potential role in CCR-2 mediated chemotaxis and influenza and HIV infections, has not been previously identified during mycobacterial infections. The MCP-2 and MIP-3a were also over-expressed in the supernatants (Fig. 17B), along with IP- 10 and MCP-4. Interestingly, albumin and RBP4 were only significantly upregulated in the co-culture with ECM. These two proteins are recently identified as potential serum biomarkers of active TB, indicating their upregulation during advanced infection. The intermediate up and down-regulated proteins (FC: 5-50) are shown in the bottom panels of Fig. 17A (cell lysate) and 17B (supernatant). Most of the proteins identified using this proteomics approach correlated well with biomarkers identified in other studies using samplesfrom TB patients. The expression of 5 protein biomarkers identified using this approach in the 3D co-culture model was confirmed using ELISPOT (Fig. 18). The ECM plays a critical role in regulating the host-pathogen interactions and granuloma necrosis, and it has shown a significant impact on the expression of select cytokines in our 3D co-culture infected with Mm, compared to the uninfected control 3D culture. The top 25 most affected chemokines due to the addition of ECM in the AT / ?? -infected or control 3D culture are shown in Figure 17C. In the control 3D culture, ECM significantly up-regulated CD7 (117), galectin-7 (39.6), and transferrin (13.6) levels, and downregulated RBP4 (-17.6) and CD163 (-11.1), as some of these proteins interacts with the ECM. In the w-infected 3D co-culture, ECM significantly upregulated CD7 (189), vWF-A2 (72.6), RBP4 (63.3), and albumin (13.6), and down-regulated CEACAM-3 (155.7), DCTN1 (56.1), Flt-3 (41.4), lumican (25.5), and PRX2 (21) levels. By comparing the infected versus control 3D cultures, the most dramatic change, due to the addition of ECM, was the downregulation of CEACAM-3, DCTN1, Fit-3, lumican, FABP4, ADAM22, and I-TAC expression and upregulation of vWF-A2, RBP4, albumin, and PCK1 in the AT / ??-infected coculture, suggesting that their levels are modulated by ECM addition leading to cavitation. Further studies are required to understand better the role of ECM deposition and modulation of these proteins on cavity formation in tuberculomas.RNA sequencing offers insight into the transcriptional changes associated with granuloma formation and supports the development of human tuberculoma-like fate in the model

[0474] To identify the transcriptomic dynamics of host-pathogen interactions over time in the 3D model, THP-1 monocytes were co-cultured with Mm at low MOI. On day 0 (infection day), day 3 (formation of spheroids), day 6 (start of granuloma development), day 9 (formation of well- organized granulomas), and day 12 (late-window in granuloma formation), 3D spheroids were collected and subjected to total RNA extraction, cDNA library preparation and RNA sequencing (Fig. 20). Uninfected 3D cultures collected at the corresponding days of incubation were prepared using the same procedures.

[0475] We sequenced a total of 36 cDNA libraries. Each library generated 26.4 ~ 94.4 million paired end reads, averaging 28.1 million reads per sample. Initial quality assessment of the 36 raw FASTQ files confirmed that all libraries passed the quality criteria with a Phred score > 30. An average of 85.1% of the reads were aligned to the reference genome’s unique location for each sample. The total read length, GC%, aligned reads, and mapping quality of each samplewere determined. Principle component analysis (PCA) in Fig. 19A showed good clustering of biological replicates of the same sample, except for one outlier in the uninfected samples at day 3 (the outlier was included in the downstream analysis). Samples infected with Mm were located in the plot's right area, distinctly separated from the control samples (left area). Early infected samples (days 3 and 6) were mainly found at the bottom of the plot, while the late infected samples (days 9 and 12) were clustered at the top area. The hierarchical clustering heatmap (Fig. 21A) also demonstrated a clear separation of naive samples with Mm infected samples and good clustering of biological replicates within each sample.

[0476] We analyzed differentially expressed genes (DEGs) between naive and infected samples using DESeq2 tools. DEGs were filtered using a adjusted p-value < 0.05 and log2(fold change) > 1 or < -1 for biological significance. Volcano plots in Fig. 19B represent dynamic changes in significant DEGs of infected samples compared to controls. 988, 3,352, 5,193, and 5,618 DEGs (both up- and down-regulated) were identified in samples infected for 3-, 6-, 9-, and 12 days compared to their corresponding controls, respectively. The longer the infection days, the more significant DEGs were found. The peak of up-regulated DEGs was found on day 9 (n=2,868), and the peak of down-regulated DEGs was found on day 12 (n=3,062). Similar results are shown in the MA plot in Fig. 21B.

[0477] To further explore the core DEGs shared by four infection time points, red and blue dots from volcano plots were plotted by Venn diagram. As shown in Fig. 19C, 465 up-regulated DEGs (bottom left) and 71 down-regulated DEGs (bottom right) were found in the core of the diagrams, representing 47.1% and 31.7% DEGs of day 3 post-infection samples. ShinyGO gene ontology enrichment analysis of the core shared genes (both up- and down- regulated DEGs, n=539) confirms a significant hit of several immune response pathways, such as chemokine signaling pathway, cytokine-cytokine receptor interaction, cell adhesion molecules, necroptosis, TNF signaling pathway, and IL- 17 signaling. Raw expression values of the shared up and down- regulated DEGs are shown in Table S3.

[0478] Within the top-up and down-regulated DEGs, CCL8, ADAM7-AS1, CXCL10 / 11, and VCAM1 showed a rapid response to the early infection, with a significant (>100) increase in gene expression at day three post-infection. CCL8 reached a 10,000-fold increase in gene expression at day 12 post-infection, indicating its critical role in forming granuloma and immune response to mycobacterial infection. Some genes showed a trend of increased expression as theinfection progressed, such as TNFAIP6, IGFBP3, IFI27 / 44L, IFIT1, CCL4, IL6, CXCL2 / 8 / 14, MMP8, SUGCT-AS1, AIM2, MET, IER3, and reached a maximum expression level on day 12 post-infection. IDO1, MMP1 / 12, ITGB8, ILIA, CCR7, CCL20, CXCL1 / 3, and IL1B showed increased or peak expression at around day 9 post-infection, but their expression dropped on day 12 post-infection. Our study demonstrated the dynamic changes in the immune response-related transcripts abundance during the development of granuloma lesions.

[0479] To identify the DEGs most biologically relevant to the immune response against pathogenic infection, all the DEGs (red and blue dots) from the volcano plots were subjected to the Gene Sets Enrichment Analysis (GSEA). Some studies have shown that the same pathway from different databases might contain other represented genes. The content variations in each database significantly impact the enrichment result and predictive modeling

[0025] , Since our study is most interested in identifying the disease-related immune response, we performed enrichment analysis using two approaches: Integrity pathway analysis (IP A) and GSEA by gseKEGG from Clusterprofller package (R). IPA is a robust web-based application that provides information about pathways, diseases, and other signatures that may be significantly altered across different samples

[0026] , We performed canonical pathway analysis with the z-score representing either activation (positive) or inhibition (negative). gseKEGG uses a computational method to determine whether a set of genes showed statistically significant modulation in a specific pathway (KEGG pathway in this study), with an enrichment score representing the activation or inhibition of the pathway.

[0480] The bar plot in Fig. 19D shows the dynamic changes in the activation or inhibition of the top 45 most enriched canonical pathways identified by IPA. Overall, 25 pathways were constantly activated during the entire infection period (12 days) and were primarily related to innate immune responses of monocytes and macrophages against infection (i.e., phagosome formation, IL 17 signaling, pulmonary fibrosis idiopathic signaling, pyroptosis signaling, Tolllike receptor, and chemokine signaling). Some novel pathways were identified. For example, two neuro-related signaling pathways (neuroinflammation signaling pathway and CREB signaling in neurons) were significantly activated during the entire infection period (red dot). Five cholesterol-related pathways (cholesterol biosynthesis I, II, and III, glycolysis I, and super pathway of cholesterol biosynthesis (purple triangles)) were not activated at the early stage but significantly upregulated when granulomas were formed. These results suggest that lipid andcholesterol metabolic changes may play a critical role in forming granulomas in the 3D model. Two pathways (CREB signaling in neurons and G-protein coupled receptor signaling (blue asteroid)) were activated early when the infection started. Still, they were diminished as the infection progressed, suggesting their roles in the early innate immunity response against the pathogen. Eight inhibited pathways during infection were shown at the bottom of the plot.

[0481] The same sets of DEGs were subjected to the gseKEGG analysis in R using the KEGG database in parallel. As shown in Fig. 19E, dynamic changes in the top 40 enriched KEGG pathways are presented using normalized enrichment scores (NES). Ten pathways were shared by both IPA and gseKEGG analysis (red dot). Some essential signaling pathways for the innate immune response against infections were solely found in the gseKEGG analysis (blue asteroid), such as TNF signaling, cytokine-cytokine receptor interaction, cell adhesion molecules, apoptosis, and focal adhesion pathways. Interestingly, the inhibited pathways in gseKEGG analysis were very different from the result of IPA analysis, indicating that the selection of databases plays a critical role in the GSEA results.

[0482] The transcripts’ abundance in selected pathways from both IPA and gseKEGG are presented in the heatmaps (Fig. 22 A-H). DEGs in these selected pathways are mainly involved in the induction of innate immune responses against mycobacterial infection; some key genes are cell adhesion molecules (ICAM1 / 2 / 3, VCAM1, SIGLEC1, CADM1, SELL), angiogenesis (ANGPTL4, ANGPT1, ANGPT2, ANGPT4), necroptosis (TNF, IFNB1, TLR3, TNFSF10), inflammasome and pyroptosis (IFI27, AIM2, GBP1 / 2 / 3 / 5, CASP 1 / 3 / 4 / 6 / 8), MMPs (MMP1 / 2 / 3 / 8 / 9 / 10 / 12 / 13 / 14 / 19), neuroactive ligand-receptor interaction (ADM, SSTR2 / 3), chemokines and cytokines (CD40, CCL 1 / 2 / 3 / 7 / 14 / 17 / 20 / 22 / 24, IL2 / 5 / 7 / 10 / 23 / 36, CXCL1 / 3 / 8 / 12, TGFB1). Gene expression changes in the selected pathway were visualized using the R path view tools (data not shown).

[0483] Our transcriptomic results in the THP- l-ATw 3D model showed a good agreement with other in vitro and in vivo TB studies. By using the same filtering parameters (adjusted P-value < 0.05, |log2fold-change >11), our model shared 529 DEGs with the polymer-encapsulated PBMC 3D model incorporating ECM presented by Elkington et al. (data not shown) and shared 728 DEGs with the transcriptomic studies of TB clinical samples (lymph nodes with granulomas) studied by the same group (data not shown). THP1 3D model also shared 288 DEGs with the mouse and non-human primate (NHP) samples infected with TB (data not shown). Among the16 blood RNA signature genes for tuberculosis identified by Zak et al., our RNAseq data showed up-regulation of 11 genes (68.8%) (data not shown). Within the 70 signature genes consistently upregulated in active TB and LTBI from UK London and Leicester cohort datasets investigated by Berry and colleagues, THP1 3D model showed upregulation of about 90% of genes (data not shown). These results indicate that RNA sequencing in the model offers insight into the transcriptional changes associated with granuloma formation.

[0484] The 3D in vitro tuberculoma model can be employed to investigate trained immunity against TB and other mycobacterial infections.

[0485] Trained immunity is a heterologous immunity which is induced by epigenetic and metabolic imprinting and reprogramming in monocytes, macrophages, NK and other innate immune cells after exposure to live attenuated vaccine, vectored vaccine, microorganism(s), specific cytokines, or therapeutics resulting in a long-term innate immune memory and potent immune response. This concept was first fully demonstrated in 2012, and BCG vaccination against TB has been shown to confer cross-protection via innate immunity and reduce mortality caused by a lethal Candida albicans infection or lung cancer (melanoma) in severe combined immunodeficient mice, which lack T and B lymphocytes and therefore cannot mount an adaptive immune response. In addition to pathogenic stimuli, self-derived molecules, such as DAMPs (for example, vimentin) and cytokines (for example, granulocyte-macrophage colony-stimulating factor), can also induce trained immunity. We found that live attenuated and avirulent BCG vaccine-trained THP-1 monocyte derived macrophages significantly prevented the growth of secondary infection caused by virulent mycobacteria i.e., Mm 1218 expressing GFP relative to RPMI alone-trained THP-1 cells in a 3D cell culture and also prevented granuloma lesion development in 3D tuberculoma model (Fig 23). These results demonstrate that 3D in vitro tuberculoma model can be employed to investigate trained immunity induced by live attenuated and subunit TB vaccines, immunotherapeutics, biologies, and host-directed therapeutics in development against TB and other mycobacterial diseases.The 3D in vitro tuberculoma model can be employed as an HTS-compatible platform to identify potential HDT compounds.

[0486] Screening of biosimilars provided feasibility indications that the 3D model could be used as an HTS bioplatform for identifying probable anti-TB therapeutics, with a potential advantage over available in-vitro granuloma models in discovering their effects on both the mycobacterialburdens and granuloma lesions. Given the macrophage-rich framework in natural granulomas and the difficulty in targeting bacilli in tuberculomas, we asked whether this platform could identify and characterize HDTs modulating macrophage responses in the physiological tubercular milieu. Such HDTs modulating the perturbed granuloma-macrophage defense pathways likely have the maximum impact in controlling the early mycobacterial infection, considering the delayed onset and entry of adaptive immune cells in the granulomas. Here, we used solid tuberculomas developed using 3D co-cultures of THP-1 monocytes and Mm M (tdTomato) to screen a customized library of 65 known potential HDT compounds (Fig. 24A), which includes 44 FDA-approved drugs, 10 compounds in clinical trials or human use for diverse indications, and 11 investigational compounds not yet studied in humans. These compounds were selected based on their known capability to reduce the mycobacterial burdens in the 2D cell cultures or the potential to induce host-directed anti -mycobacterial immune responses in humans or animal models after examining literature published between 2000-2020. However, their anti-mycobacterial effects in the relevant 3D tuberculoma milieu have yet to be investigated in a head-to-head comparison. Based on the known mechanisms of action, many of these compounds exhibit diverse and numerous host-directed effects and fall into broad categories of pharmaceuticals inducing anti-cancer, anti-inflammatory, kinase modulatory, ionchannel blocking, neuroleptic, antioxidant, epigenetic modulatory, anti-diabetic, and other effects. In addition, at least 13 compounds have previously reported some direct mycobacteriatargeting activity, although not necessarily bactericidal, in axenic cultures. We assessed the efficacy of each of these 65 compounds in reducing bacterial growth and granuloma lesions following a single treatment on day 6 at a concentration of 20 pM by performing FI readings and imaging of 3D co-cultures 6-8 days after treatment.

[0487] From the four independent screens, we found that only 31 of these compounds decreased the average Wm-burdens by > 25% in the 3D milieu compared to untreated controls, and just 18 of those significantly (p<0.05) reduced the A / m-burdens relative to drug-carrier DMSO-treated negative controls (Fig. 24A). Seven compounds interestingly increased A / -burdens compared to untreated controls. Importantly, we detected the significant anti-bacterial activities of the known pathogen- and host-directed drug nitazoxanide (20 pM / well) and antibiotic rifampicin (Ipg / ml) in this cellular model (Fig. 25A), which served as positive controls and informed us the reliability of these results. When we evaluated a widely used statistical measure of assay qualitythat describes how well separated the positive and negative controls are, the assays performed in the Mm infection model using an MOI range of 0.007 to 0.012, exhibited Z’-factors > 0.5, which is considered excellent in robustness for HTS assays (Fig. 25A). When we investigated optimal MOI for four b (tdTomato) strains that display diverse growth rates in THP-1 cells, we found that a relatively higher MOI (0.012-0.05) than Mm was needed using slow-growing Mtb strains H37Rv, Erdman, and Beijing F2 for the assay to exhibit an excellent Z’ -factor. For the CDC1551 strain, which proliferates in small clusters outside the 3D structure and appears to evade the containment in the spheroid, the assay quality could not reach an excellent level even with the MOI of 0.05 and despite well-organized lesion formation in the spheroids (Fig. 25B). Next, we confirmed the assay performance using Mtb H37Rv at an optimum MOI of 0.025 in a large number (n = 10) of 3D spheroid plates. Reproducible b-growth inhibition using nitazoxanide and rifampicin was detected across large numbers of wells, and results in each one of these plates exhibited excellent assay quality (Fig. 25C). These results collectively informed the tuberculoma model’s suitability as an HTS-compatible bioplatform. They further allowed us to extend our investigations of compounds in the BSL-3 laboratory using a panel of four Mtb strains representing a heterogeneous group of genotypes.

[0488] Using Mtb strains H37Rv, Erdman, and Beijing F2, which exhibited excellent assay quality, and from 2-4 independent screens per strain, we identified 29-31 compounds that reduced the average A7 / / ?-burdens by > 25% compared to untreated controls (data not shown), despite strain-specific differences in bacterial growth inhibition by HDT compounds. Using the Mtb CDC1551 infection model and the same parameters, we identified 39 compounds. Since this platform did not exhibit excellent assay quality, we did not consider these results and the strain for further investigations. A comparison of the screening results in the BSL-2 (Mm M) and BSL- 3 (three Mtb strains) platforms notably revealed considerable overlap between the ‘top-hits’ that reduced bacterial burdens by > 50%, notwithstanding strain-specific variations in dependence on the host cellular factors for intracellular survival. These top-hits include 12 drugs with diverse known host-directed effects, AT9283, auranofin, clemastine fumarate, chlorpromazine hydrochloride, 3'4' dichlorobenzamil HCL, dovitinib, doxycycline, H89, lansoprazole, metixene hydrochloride, nitazoxanide, and zuclopenthixol dihydrochloride, that were effective in all four infection models (Fig. 24B). We compared the list of hits identified using > 50% bacterial inhibition with the hits discovered using statistical - value < 0.05 and composite z-scores using arobust cutoff < -4 in all infection models and encouraged to find that the list of ‘top-hits’ identified using both methods was effectively the same (Fig. 24C). A high degree of agreement (> 90%) in the top hits identified in Mm and Mtb infection models reassured us about the performance and validity of results in our BSL-2 and -3 platforms. We next considered the capability of 65 compounds to mitigate immunopathology by reducing granuloma lesions (Fig 26 and Fig 27). The 12 ‘top hit’ compounds also effectively alleviated lesion burdens, except clemastine fumarate and metixene hydrochloride, which did not efficiently reduce lesion burdens in the Mtb Erdman infection model. Despite consistently reducing the average lesion burdens by > 25% in all infection models, the effect was not also statistically significant for lansoprazole, dovitinib, 3'4' dichlorobenzamil HCL, and AT9283 in the Erdman model.

[0489] Lymphocytes can influence antimycobacterial response in tuberculous granulomas, and the performance of HDTs in the context of innate and adaptive primary human immune cells is paramount. Therefore, we further investigated 65 compounds in the 3D tuberculoma model generated using Mtb Erdman (tdTomato)-infected primary human CD14+blood monocytes and later supplemented with autologous PBMC subsets comprising lymphocytes and CD14' myeloid cells (Fig. 24D). We found a considerable agreement between the compound-screening results in the TEIP-1 monocyte and PBMC platforms. Of the 65 compounds screened, 29 inhibited Mtb Erdman burdens by > 25% in the THP-1 model, and 23 were also effective in the PBMC model. However, on account of variations in the individual donor responses, only 14 of these compounds caused a significant decrease (p < 0.05) in Mtb burdens relative to DMSO-treated controls in the PBMC model, as opposed to 23 compounds in the THP-1 model. Nonetheless, when the percent normalized Mtb burdens in the two cellular models for individual compounds were compared, no significant difference in efficacy was observed for 64 compounds. One drug, metixene hydrochloride, reduced significantly more Mtb burdens (p < 0.05) in the THP-1 compared to the PBMC model, suggesting that this drug might not be effective in the context of primary human immunity.

[0490] Next, we considered the consistency of effectiveness of compounds in the above described 5 infection models using THP-1 monocytes or PBMCs (Fig. 24). Interestingly, none of the FDA-approved anti-diabetics (metformin, glyburide, sitagliptin) and non-steroidal or steroidal anti-inflammatory drugs (dexamethasone, prednisone, aspirin, ibuprofen, indomethacin, diclofenac sodium) tested, decreased > 25% bacterial burdens in these models. Further, 1-methyl-D tryptophan (IDO inhibitor), zileuton (5 -lipoxygenase inhibitor), BzATP (P2X7 receptor agonist), verapamil (calcium channel blocker), zoledronic acid (y8 T-cell response inducer), N-Acetyl-L-cysteine (anti-oxidant), sodium valproate (anti-convulsant and sodium channel blocker), L-citrulline (nitric oxide booster), sodium phenylbutyrate (histone deacetylase inhibitor), maraviroc (CCR5 receptor antagonist), and irbesartan (anti-hypertensive and PPARy activator), previously identified as potential HDTs (Fig 24), were also not effective as standalone therapeutics, except indomethacin and L-citrulline, which modestly inhibited bacterial burdens in single THP-1 model using Mtb Beijing-F2 or Erdman.

[0491] Interestingly, not all hit compounds identified in previous studies using mycobacterium- infected 2D cell-culture screens were efficacious in our 3D model. While these prior studies had essentially treated cell cultures with compounds either during or soon after mycobacterial uptake by host cells, our screens investigated a later time window in infection involving the establishment of a granuloma milieu. For example, nine compounds in our custom library were previously identified as among the top hits in a library of 1260 pharmacologically active compounds (LOPAC)-based drug-repurposing screens in the 2D cell culture of Mtb H37Rv- infected melanoma cell line MelJuSo or primary human monocyte-derived macrophages exposed to compounds immediately after bacterial uptake. However, only four compounds, dovitinib, H89, 3'4' dichlorobenzamil HCL, and GW5074, consistently inhibited 25-90% bacterial loads in all our infection models (Fig. 24). Three compounds, quinacrine diHCL, tyrphostin AG459, and haloperidol, reduced > 25% bacterial burdens in one to three but not all infection models, and SU 6656 and SB 216763 were ineffective in our models. Since the efficacy of these compounds in the animal models of TB is still being ascertained, these results need further in vivo validation. Of particular importance, the hit compounds previously reported to inhibit mycobacterial burdens in animal models as standalone therapeutics were also efficacious in our infection models. For example, a gastric proton-pump inhibitor (PPI) lansoprazole, previously identified as the potent hit after screening FDA-approved 1280 drugs of the Prestwick Chemical Library in the / i-infected 2D lung-fibroblast (MRC-5) culture and with partial inhibitory activity against Mtb in mice, was also highly effective in all our infection models. Similarly, chlorpromazine hydrochloride, metixene hydrochloride, clemastine fumarate, and zuclopenthixol dihydrochloride were identified among the lead HDTs in the FDA-approved Prestwick library -based screens in the M«-infected zebrafish larvae model. These drugs were also identified as top hits in all ourinfection models, except metixene hydrochloride, which was ineffective in the PBMC model, as described above. Furthermore, antimicrobial drugs nitazoxanide and doxycycline, each previously reported to possess host-directed effects and inhibit mycobacterial burdens in the animal models, exhibited potent anti-mycobacterial activity in our infection models. These results support the capability of our platform to identify in vivo effects of known host-directed and pathogen-targeting drugs in the 3D in vitro tuberculoma milieu. To further explore compounds that better inhibit bacterial growth in a 3D microenvironment than in a 2D milieu, we screened compounds in the / W? H37Rv-infected 2D THP-1 culture during a similar later time window involving the establishment of infection (day six post-infection) as in 3D co-culture (data not shown). While AT9283 and simvastatin reduced bacterial burden better in the 3D tuberculoma model, imatinib and quinacrine dihydrochloride were more effective in the 2D cell culture. Interestingly, several compounds identified in published studies involving treatment during an early time window in infection in 2D cell culture were ineffective once the pathogen established a sustainable equilibrium with the host intracellular milieu.

[0492] To differentiate between drugs with host-directed effects and those that directly target bacteria in the 3D milieu, we treated Mm M oxMtb Erdman (tdTomato) broth cultures with compounds in the 3D ULA microplates. The bacterial growth was monitored as axenic 3D culture, and FI was measured on days 6 and 8 post-compound treatment (data not shown). Thirteen compounds from the customized library had previously described anti-mycobacterial activity in broth cultures besides host-directed effects, including an antibiotic and MMP inhibitor, doxycycline (data not shown). As expected, antibiotics doxycycline and rifampicin exhibited significant inhibitory activity for Mm ‘M’ and Mtb Erdman in 3D axenic cultures compared to untreated controls, validating the capability of the assay to reproduce the effects of known anti-TB drugs in 3D broth cultures. Interestingly, auranofin (anti-rheumatic), nitazoxanide (anti-parasitic), and lansoprazole (PPI), drugs with known antimicrobial activity, exhibited significant inhibitory activity ioxMtb Erdman but were ineffective against Mm at 20pM concentration tested (data now shown). Yet, these drugs significantly inhibited both species in the 3D tuberculoma bioplatform. Chlorpromazine hydrochloride (anti-psychotic and acid sphingomyelinase inhibitor) and fingolimod hydrochloride (acid sphingosine- 1 -phosphate receptor modulator) have reported direct anti-mycobacterial activity and also showed strong intracellular mycobacterial-growth inhibitory activity in 3D tuberculoma bioplatform; however,-I l l-these drugs were ineffective in 3D axenic cultures. Among other compounds with known direct anti-mycobacterial activity, N-acetyl-L-cysteine, verapamil, sodium valproate, and irbesartan were ineffective in both 3D broth culture and tuberculoma bioplatform. Interestingly, five compounds previously identified as potential HDT drugs, clementine fumarate, tin protoporphyrin IX, tyrphostin AG 494, metixene hydrochloride, and fluoxetine hydrochloride, exhibited significant growth inhibition compared to controls for either one or both species in the 3D axenic cultures, which suggests that these HDT compounds possess some pathogen-targeting effects and need further investigation. By excluding any direct microbicidal activity in axenic culture against mycobacteria, these results indicate that among 12 ‘top hit’ compounds (Fig. 24), AT9283, zuclopenthixol dihydrochloride, dovitinib, 3'4' dichlorobenzamil HCL, and H89 exert their intracellular mycobacteria inhibitory effects mainly through the host macrophages.

[0493] Given the high attrition of lead compounds due to host toxicity and to verify whether any observed effects are derived from an influence on host-cell viability, we treated uninfected 3D cultures of THP-1 cells or PBMCs with individual compounds and measured cytotoxicity using CytoTox Gio assay. The cytotoxicity screening results exhibited high concordance in these two cell cultures, regardless of the monocytic cell line or primary host cells used. Among the 15 compounds that caused > 50% cytotoxicity in these cell cultures, 13 were cytotoxic for both cell types at 20 pM concentration (data not shown). Interestingly, 11 of these compounds are FDA- approved or evaluated in clinical trials and found safe for human use, indicating discordance between in vitro toxicity and in vivo testing results. Although toxicity results are an indispensable part of the therapeutic evaluation process, these results also imply that in vitro cytotoxicity methods and doses cannot capture all the complexities in the human body encountered during in vivo testing. Furthermore, despite a positive correlation between the compound-induced cytotoxicity and the normalized bacterial loads in the Mw-infected THP-1 model (data not shown), among 17 hit compounds that decreased average Mm burdens by > 50%, only nine compounds caused a > 50% reduction in THP-1 viability. Conversely, four compounds that increased Mm burdens also caused a > 50% cytotoxicity. To further determine the cytotoxic concentration leading to a 50% reduction in THP-1 cell viability (CC50), we tested a range of concentrations of each cytotoxic compound in 3D cultures of THP-1 -RFP cells (data not shown). Out of 12 ‘top hit’ compounds, six were cytotoxic at 20 pM concentration, and the effective concentration (EQso values could be obtained for nitazoxanide, doxycycline, anddovitinib. At nontoxic doses, auranofin and 3'4' dichlorobenzamil HCL reduced > 35% of Mm burdens, but AT9283 could not inhibit Mm growth. Considering Mm and Mtb can grow intracellularly after macrophage death and Mm can quickly overtake the spheroid structure by extracellular proliferation, we inferred after considering all results that compound-induced cytotoxicity may not always associate with a reduction in intracellular mycobacterial load in in- vitro assays. We further deliberated that cell death accompanying innate immune mechanisms induced by host-directed therapeutics, for example, apoptosis, autophagy, and pyroptosis, could reduce cellular viability while inhibiting intracellular pathogen survival. Since the ideal bioplatform should be able to characterize such immune mechanisms of action while simultaneously identifying therapeutic efficacy within 3D granulomas, we next investigated the innate macrophage defenses activated by the HDT compounds.The readouts of immune mechanisms of HDTs can be incorporated in situ in the 3D tuberculoma bioplatform

[0494] Physiological or pharmacological activation of hypoxia, phagolysosome fusion, lysosomal acidification, autophagy, and inflammasome pathways in infected human macrophages and animal models can severely decrease the survival of intracellular Mtb. Therefore, we investigated whether the readouts of such innate immune mechanisms can be included in our model. We generated 3D co-cultures of THP-1 monocytes and WT Mtb Erdman and incorporated fluorogenic probes that emit fluorescence in positive microenvironments. While the exposure of 3D tuberculomas to hypoxia-probe emits red fluorescence when the oxygen level reaches < 4% in the multicellular core, exposure to lysosomal-acidification-probe produces red fluorescence when the lysosomal number or acidification increases in the cellular scaffold. We generated 3D co-cultures of Mtb Erdman (WT) and THP-1 -Difluo hLC3 reporter cells to monitor autophagic flux. In these reporter cells, the RFP-GFP tandem fluorescent-tagged LC3 probe emits red plus green (yellow) signals in the cytosol and autophagosomes but only red fluorescence in autophagolysosomes since acid-sensitive GFP is more easily quenched or degraded in an acidic environment than the acid-stable RFP. Following the failure in autophagosome and lysosome fusion, yellow puncta accumulate, whereas yellow and red puncta are reduced in the case of autophagy inhibition. In addition, we used 3D co-cultures of Mtb Erdman (tdTomato) and THP-1 -ASC-GFP reporter cells to monitor ASC-dependent inflammasome formation. In these reporter cells, Mtb infection or antigen exposure leads toASC-GFP expression and ASC-speck formation following inflammasome activation. We used high-content imaging to spatiotemporally resolve the components of the innate immune pathways in 3D tuberculomas and their modulation by HDT compounds in a 96-well format. Despite reducing cell viability following treatment and potentially inhibiting reporter-probe activity, we did not exclude cytotoxic drugs in this screen.

[0495] Following treatment with individual 6 ‘top-hit,’ non-cytotoxic compounds at 20 pM dose, 3D tuberculomas exhibited increased hypoxia in cores or cellular aggregates (blue arrows) distinct from bacteria-permissive necrotizing lesions (white arrows), on day 6 post-treatment, compared to DMSO or no treatment (Fig. 28A and B). These compound-treated tuberculomas also showed cells with increased autophagolysosome formation and acidification, as revealed by increased red over green or yellow fluorescence. Rifampicin-treated tuberculomas exhibited similar hypoxia and autophagy activation features, but their scaffold displayed a less acidic microenvironment, likely due to a reduction in AAA-induced cell death. Comparatively, tuberculomas treated with six non-cytotoxic, representative ‘non-hit’ compounds showed more cellular aggregates with autophagosome accumulation (yellow arrows) or without acidification despite the acidification of the tuberculoma-cores, indicating blockade of downstream steps in autophagy. They also showed less hypoxia-staining, like in DMSO-treated or untreated controls, on day 6 post-treatment (Fig. 28B). Since autophagy induction is a dynamic multi-step process, we performed serial imaging in tuberculomas over six days to monitor autophagic flux. The autophagy flux progressed relatively swiftly within three days in tuberculomas treated with ‘tophit’ compounds (data not shown), especially with lansoprazole and H89, and more ASC-GFP- expressing cells persisted by day six post-treatment, with reduced AAA growth compared to negative controls or ‘non-hit’ compound-treated tuberculomas, suggesting increased inhibition of AAA-induced cell death. In ‘non-hit’ compounds-treated tuberculomas, AAA growth occurred unfettered and autophagy flux developed gradually between 3-6 days, with several cell clusters lacking autophagosome maturation. In rifampicin-treated tuberculomas, autophagy developed slowly by day 6, likely on nutrient starvation, with some cellular aggregates lacking autophagosome maturation or permitting AAA persistence with strong inflammasome activation, suggesting delayed autophagy induction and incomplete sterilization. Taken together, screening results of noncytotoxic ‘top-hit’ HDT compounds extend the earlier findings in AAA-infected 2Dcell cultures by establishing the susceptibility oiMtb to pharmacological induction of autophagy and lysosomal acidification in the 3D tuberculoma model.

[0496] Tuberculomas treated with 6 ‘top hit’ cytotoxic compounds showed relatively less hypoxia. They still exhibited autophagy induction and acidification in the centers on day six post-treatment (Fig. 28C and D), potentially due to decreased cell viability. In these tuberculomas, autophagy flux seemed to progress between 1-3 days, and ASC-GFP-expressing cells decreased with concurrent control o Mtb growth on day 6, relative to DMSO-treated controls (data not shown). We also investigated 6 remaining cytotoxic compounds, which exhibited species-specific heterogeneity in mycobacterial growth inhibition but caused a 25-50% reduction in Mtb Erdman burdens. Tuberculomas treated with simvastatin or fingolimod HCL showed autophagy induction between 3-6 days. In contrast, histone deacetylase inhibitors resveratrol and vorinostat, and previously described autophagy inhibitors, quinacrine dihydrochloride, and sunitinib malate, which emit autofluorescence in the GFP channel, induce incomplete autophagic flux in 3D tuberculomas.

[0497] Since excessive hypoxia can cause pathogenic necrosis of mycobacterium-infected macrophages, and autophagic and inflammasome activity can cause non-apoptotic programmed cell death and pyroptosis, we carefully examined the results of all 65 compounds to characterize further the association between these processes with Mtb control in the 3D tuberculomas. When we considered these mechanisms individually, no direct association was found with the inhibition of Mtb Erdman burdens on day six post-treatment. For example, KN62, loperamide, and glibenclamide increased hypoxia compared to DMSO controls (data not shown), and imatinib, KN62, and loperamide induced autophagy with autophagolysosome formation or acidification of core but did not significantly inhibit A / / / ? -burdens. Likewise, among nine compounds previously reported to activate autophagy in theM / >-infected 2D cultures (data not shown), only gefitinib and everolimus induced autophagy with autolysosome formation in the tuberculoma cores by day 6. Incomplete autophagy flux in numerous cellular clusters was observed with the remaining seven drugs, indicating discordance in their ability to induce autophagy in the 3D tuberculoma milieu. None of these nine compounds, except SRT 1720, significantly inhibited Mtb Erdman loads. Notably, the autophagy flux was delayed in 3D tuberculomas treated with these HDT compounds (data not shown). These results collectively indicate that rapid pharmacological induction of complete autophagy flux rather than mereautophagosome initiation correlates with the inhibition ot' Mtb growth in 3D tuberculomas. They also demonstrate the versatility of the bioplatform to incorporate the readouts of innate immune mechanisms induced by therapeutics in situ.HDT compounds can be classified into functional clusters based on immune mechanisms and therapeutic efficacy in 3D bioplatform

[0498] Several compounds reduced bacterial burdens and lesions in more than one of the five mycobacterial-infection models investigated (data not shown). Six ‘top-hit’ compounds, nitazoxanide, auranofin, doxycycline, dovitinib, 3'4' dichlorobenzamil HCL, and AT9283, which caused significant cytotoxicity in both THP-1 and PBMC cultures, inhibited > 50% and significant bacterial burdens compared to untreated controls in all infection models, except for AT9283, which reduced an average of 42.2% but significant Mtb Erdman-burdens in the PBMC model. These 6 cluster 1 compounds also reduced > 50% of lesion-burdens in 3 of the four infection models using THP-1 cells. They induced rapid autophagy flux with lysosomal acidification but did not induce hypoxia at a cytotoxic 20pM dose (Fig. 29). Using a range of concentrations, we confirmed that these compounds induce autophagy at the CC50 or EC50 concentrations in Mm (WT)-infected THP-1 monocytes transfected with autophagy sensor LC3B-RFP (BacMam 2.0) (data not shown). With the decreasing concentrations of auranofin and AT9283, autophagy flux decreased, and autophagosome accumulation increased in the 3D tuberculomas of THP-1 -Difluo autophagy reporter cells (data not shown). The decreased autophagy flux broadly corresponded with the reduced inhibition of Mtb Erdman-burden in the 3D tuberculomas of ASC-GFP inflammasome reporter cells. Despite the increased cellular viability with decreasing compound concentrations, the corresponding increase in hypoxia induction was not observed, confirming that the efficacy of these top-hit compounds correlates with autophagy induction but not hypoxia in the 3D tuberculoma model.

[0499] Six other ‘top-hit’ compounds, metixene HCL, H89, chlorpromazine HCL, lansoprazole, zuclopenthixol diHCL, and clemastine fumarate, which caused < 50% toxicity in THP-1 cultures, also reduced > 50% of bacterial burdens compared to untreated controls in all mycobacterial-infection models, except for metixene HCL, zuclopenthixol diHCL, and clemastine fumarate, which did not significantly inhibit Mtb Erdman-burdens in the PBMC model. Interestingly, lansoprazole, an FDA-approved drug, induced > 50% toxicity in the 3D PBMC cultures. These six cluster-2 compounds inhibited > 50% of lesion-burdens in 3 infectionmodels using THP-1 and induced hypoxia, rapid autophagy flux, and lysosomal acidification (Fig. 29). Notably, with the decreasing concentrations of lansoprazole and H89, hypoxia, autophagy, and ASC-speck formation decreased on day 6, and / ? Erdman-burdens increased, suggesting that these pathways contribute to the efficacy of these compounds in the 3D tuberculoma model (data not shown).

[0500] Five potentially ‘hit’ cytotoxic compounds, sunitinib malate, fmgolimod HCL, dasatinib, GW 5074, and SRT 1720, formed cluster 3. They reduced > 25% of bacterial burdens compared to untreated controls in all mycobacterial-infection models, except for dasatinib, which failed to inhibit Mtb Erdman-burdens in the THP-1 and PBMC models. They also inhibited > 25% of lesion-burdens in at least three infection models using THP-1 and induced autophagy flux, although slowly between 3-6 days post-treatment, with lysosomal acidification but not hypoxia (Fig. 29). Five potentially ‘hit’ noncytotoxic compounds, fluoxetine HCL, gefitinib, loperamide, haloperidol, tin protoporphyrin IX, and ezetimibe, formed cluster 4. They inhibited > 25% of bacterial burdens compared to untreated controls in at least three mycobacterial-infection models using THP-1 cells. However, except for fluoxetine HCL and tin protoporphyrin IX, the other three compounds did not inhibit > 25% Mtb Erdman-burdens in the PBMC model. Furthermore, they showed heterogenous efficacy in reducing lesion burdens and induced autophagy flux slowly between 3-6 days, with hypoxia and lysosomal acidification.

[0501] The comparison also revealed the species-specific heterogeneity in bacterial growth inhibition by HDT compounds (Fig. 29). The four cytotoxic compounds, simvastatin, vorinostat, resveratrol, and quinacrine diHCL, exhibited species-specific heterogeneity in mycobacterial growth inhibition in 3D tuberculomas and formed cluster 5. They did not decrease Mw-burdens but inhibited A7 / A-burdens by > 25% in all four or at least one of the four AT / A-infection models of THP1 and PBMCs. Conversely, pazopanib, all-trans retinoic acid (ATRA), imatinib, omeprazole, and sildenafil were effective only in the Mm infection model using this criterion and formed cluster 6. Compounds in these two clusters exhibit heterogeneity in inhibiting lesion burdens and induction of innate defenses in the THP-1 infection models. These observations agree with the reports of subtle differences in human macrophage responses to infection with Mm and Mtb strains.

[0502] In summary, our screenings demonstrate that kinase inhibitors, dovitinib, H89, AT9283, sunitinib malate, and GW5074, and antipsychotics and antidepressants, chlorpromazinehydrochloride, metixene hydrochloride, zuclopenthixol dihydrochloride, haloperidol, and fluoxetine hydrochloride, and antimicrobials with HDT effects, nitazoxanide, doxycycline, auranofm, consistently inhibit bacterial burdens between 25-90% in all or > three infection models.AT9283 effectively inhibits Mtb burdens and granuloma lesions in vivo in mice.

[0503] One of the ‘top hit’ compounds and multi -targeted kinase inhibitor, AT9283, currently in clinical trials for treating hematological malignancies, significantly inhibited Mm and Mtb burdens and lesions in THP-1 3D tuberculomas. The compound exhibited anti -A / A activity preferentially in the 3D tuberculoma milieu compared to traditional 2D monolayer cultures in a host-directed manner without apparent direct bacteria-targeting activity in axenic cultures. To determine whether AT9283 has efficacy against Mtb in vivo, we investigated anti-mycobacterial effects during early infection in the C3Heb / FeJ mouse model. We chose to examine early infection before the onset of peak adaptive immunity, as this is the time period best modeled by our simplistic THP-1 3D co-culture model. Activating the macrophage innate defenses in nascent granulomas will likely have the maximum impact in controlling the mycobacterial disease considering the delays in T-cell responses in the lungs. We infected mice with > Erdman by aerosol inoculating 10-15 CFUs in the lungs. Following the progression of infection for seven days, mice were treated with AT9283 at a dose of 35 or 17.5 mg / kg every day, five days per week, for three weeks. Four weeks after infection, mice were euthanized, and the left lungs were homogenized and plated for CFUs. The right middle lung lobes were fixed with 4% paraformaldehyde and investigated for histopathology. Mice treated with a dose of 35 mg / kg had significantly less bacterial burdens than DMSO-treated control mice (Fig. 30 A and B). Likewise, mice treated with 35 mg or 17.5 mg / kg exhibited no gross signs of toxicity and had fewer granuloma lesions. These results support the host-directed therapeutic potential of AT9283 for treating early Mtb infection. Additional work is required to demonstrate the efficacy of AT9283 during chronic infection and its potential as an adjunct therapy with standard anti-TB antibiotics for controlling TB infection and disease.Example 3Validation of ‘hit’ host-directed therapy compounds identified using 3D tuberculoma bioplatform in C3Heb / FeJ mice

[0504] Current approaches to experimental host-directed therapies (HDTs) against tuberculosis (TB) start with targeted preliminary screening of drugs in the 2D cell culture of Mycobacterium- infected human or animal macrophages or other cells in vitro for their capability to inhibit bacterial growth. However, this 2D cell-culture approach is limited by poor predictability of the drug efficacy and toxicity, the absence of accurate insights into the mode of action of selected compounds, and the failure of the system to mirror the microenvironment and drug penetration within the TB granuloma lesions. The model described herein mimics human 3D tuberculomas with high fidelity. The bioplatform has several advantages over existing in vitro 2D cell cultures, available granuloma models, and traditional methods. These include ease of generation and use, robustness, improved tractability and scalability, increased throughput and efficiency, reproducibility, cryo-shelf stability, and reduced animal usage and costs.

[0505] The 3D bioplatform also has clear advantages over macrophage 2D monolayer cell cultures used for primary screening. For example, several mycobacterial factors, microenvironments, and drug diffusion gradients are only induced in 3D granulomas but not in the infected macrophage 2D cultures. Therefore, additional therapeutic targets and true drug efficacy could be better revealed in the 3D tuberculous granulomas. The drug therapeutic efficacy assessment includes reducing the virulent Mtb burdens and resolving granuloma lesions and pathology in infected tissues. Mere identification of a ‘hit’ drug in the 2D in vitro cell cultures using mycobacterial burden reduction alone does not guarantee that the drug will be effective in vivo in the infected organs and suitable for the treatment of animal or human hosts by penetrating the tuberculous granuloma lesions and reducing bacterial burden and granuloma pathology.

[0506] It has been recognized that Mtb-infected 2D cell cultures cannot identify many drugs (e.g., pyrazinamide) effective in human tuberculous granulomas and TB patients. Therefore, to validate HDT compound screening results in the 3D in vitro tuberculoma bioplatform, several ‘hit’ compounds were investigated in vivo in the C3Heb / FeJ mouse model to inhibit virulent Mtb Erdman's burden and granuloma lesions. These HDT compounds have not been previously investigated in the Mtb-infected mammalian models, and their true efficacy in reducing virulentMtb burdens and tuberculous granuloma lesions in vivo in the lungs is unknown. Several mouse models have long been extensively used in the TB field to carry out preclinical drug evaluation studies and the development of improved treatment regimens. The C3HeB / FeJ mouse model was used because it, as opposed to BALB / c and C757BL / 6 mouse models, has been shown to develop necrotic and hypoxic tuberculous granulomas like those observed in humans.

[0507] To determine whether selected 10 potential HDT compounds have efficacy against Mtb in vivo, anti-mycobacterial effects in the C3Heb / FeJ mouse model were studied in terms of reduction of virulent Mtb Erdman burdens and granuloma lesions in the lungs. Female C3Heb / FeJ mice were infected (Fig. 31 A) with an aerosolized dose of 10-15 Mtb Erdman CFU using a Glas-Col inhalation exposure system (Fig. 3 IB). Seven days after aerosol infection, mice were treated with individual compounds at 35 mg / kg five days per week for three weeks(Fig. 31 A). Three weeks after initiation of the treatment, mice were euthanized, and the left lungs were homogenized and plated for CFU (Fig. 31C). The right inferior lung lobes were fixed with 4% paraformaldehyde and investigated for histopathology. In untreated Mtb-infected (control) mice, the bacterial burden was comparable between left and right lungs at the peak of infection at 4 weeks post-infection (Fig. 3 ID).

[0508] Of 10 selected HDT compounds, 9 significantly, albeit at varying degrees, reduced the bacterial burden in the 3D tuberculoma bioplatform of THP-1 monocytes and Mtb Erdman (Fig. 32A). One compound, dasatinib, was ineffective in the 3D tuberculoma bioplatform of Mtb Erdman. Since prior studies have demonstrated that nitazoxanide is highly plasma-protein-bound and may not leave the blood circulation for lung 3D tuberculoma penetration, tizoxanide, a more effective anti-parasitic metabolite of nitazoxanide, was used in mouse experiments. Significantly few bacteria (p < 0.05) were found in the lungs of mice treated individually with tizoxanide, quinacrine dihydrochloride, all-trans retinoic acid (ATRA) and AT9283 compared to DMSO- treated control mice (Fig. 32B). Although average bacterial burden in the lungs of vorinostat, lansoprazole, gefitinib, H89, and sitagliptin-treated mice was lower compared to DMSO-treated mice, the burden reducing effect was not statistically significant. Likewise, mice treated with AT9283, H89, sitagliptin, quinacrine dihydrochloride, lansoprazole, and tizoxanide had fewer and smaller size tuberculous granuloma lesions (Fig. 32C and Fig. 33). Dasatinib did not substantially reduce tuberculous granuloma numbers. Still, the size of the lesions was smaller. These results demonstrate for the first time the potential of AT9283, H89, sitagliptin, quinacrinedihy drochloride, lansoprazole, and tizoxanide to reduce tuberculomas in the lungs of the relevant mammalian model and support the HDT potential of these drugs in the 3D tuberculoma milieu in vivo.

[0509] Taken together, the results of the reduction of virulent Mtb Erdman burdens and tuberculous granuloma lesions in C3Heb / Fej mice collectively show the potential of tizoxanide, AT9283, quinacrine dihydrochloride, ATRA, H89, vorinostat, lansoprazole, and sitagliptin for inclusion in the drug regimens for treating or reducing pulmonary tuberculomas (Figs. 32 and 33).

Claims

CLAIMSWhat is claimed is:

1. A three-dimensional (3D) cell co-culture, comprising a plurality of human immune cells and a plurality of mycobacteria.

2. The 3D cell co-culture of claim 1, comprising mycobacteria-in-spheroid structures formed by the human immune cells and the plurality of mycobacteria.

3. The 3D cell co-culture of claim 1, wherein the human immune cells comprise cells selected from the group consisting of monocytes, macrophages, and peripheral blood mononuclear cells (PBMCs).

4. The 3D cell co-culture of claim 3, comprising monocytes, wherein the monocytes comprise THP-1 cells, U937 histiocytes, purified CD14+monocytes, or monocyte cell subsets in PBMCs.

5. The 3D cell co-culture of claim 1, wherein the mycobacteria comprise Mycobacterium marinum (Mm), Mycobacterium tuberculosis (Mtb), or a Mycobacterium tuberculosis complex species.

6. The 3D cell co-culture of claim 5, comprising Mm, wherein the Mm comprises strain 1218 or strain M.

7. The 3D co-culture of claim 5, comprising Mtb, wherein the Mtb comprises strain H37Rv, strain Erdman, strain CDC1551, strain Beijing F2, or a clinical isolate of a drug- susceptible or drug-resi slant Mtb.

8. The 3D co-culture of claim 1, comprising 500-5000 colony forming units (cfu) of mycobacteria per 105human immune cells, resulting in a multiplicity of infection (MOI) of O.OO5-O.O5.

9. The 3D co-culture of claim 8, comprising Mm strain 1218 or strain M, wherein the MOI is about 0.006-0.012.

10. The 3D co-culture of claim 8, comprising Mtb strain H37Rv, wherein the MOI is about 0.015-0.05.

11. The 3D co-culture of claim 10, wherein the MOI is about 0.025.

12. The 3D co-culture of claim 8, comprising Mtb Erdman or Beijing F2, wherein the MOI is about 012-0.05.

13. The 3D co-culture of claim 12, comprising Mtb Erdman, wherein the MOI is about 0.05.

14. The 3D cell co-culture of claim 5, wherein the mycobacteria express Green Fluorescent Protein (GFP), Red Fluorescent Protein (RFP), a fluorescent marker, or a luminescent marker.

15. The 3D cell co-culture of claim 14, wherein the RFP comprises tdTomato or a far- red fluorescent protein.

16. The 3D cell co-culture of claim 1, contained in a U-bottom shaped well comprising a hydrogel coating, wherein the hydrogel is hydrophilic, neutrally charged, and biologically inert.

17. The 3D cell co-culture of claim 16, wherein the hydrogel comprises a perfluorinated polymer, olefin, or a combination thereof.

18. The 3D co-culture of claim 17, wherein the hydrogel comprises a CORNING® PURAMATRIX™ Peptide Hydrogel.

19. The 3D co-culture of claim 16, wherein the hydrogel comprises one or more of collagen, fibrin, and alginate.

20. The 3D cell co-culture of claim 16, wherein the well is of a CORNING® UltraLow Attachment Spheroid Microplate or an S-BIO PRIMESURFACE® 3D Culture Spheroid Plate.

21. The 3D cell co-culture of claim 1, further comprising one or more additional cell type selected from the group consisting of A549 human lung epithelial cells, HUVEC-1 human fetal endothelial cells, HULEC human lung endothelial cells, and MRC-5 human lung fibroblasts or primary human epithelial, endothelial, and fibroblast cells.

22. The 3D cell co-culture of claim 1, made by a method comprising:(a) suspending the human immune cells in a Roswell Park Memorial Institute cell culture medium, further comprising one or more of L-glutamine, FBS, sodium pyruvate, and HEPES buffer;(b) mixing the mycobacteria with the human immune cells at a MOI of 0.005- 0.5; wherein the mixture is contained in a polystyrene U-bottom shaped ultra-low attachment well of a microplate, wherein the well comprises a hydrogel coating, and wherein the hydrogel is hydrophilic, neutrally charged, and biologically inert; and(c) incubating the mixture at 37°C for a time sufficient for the human immune cells and mycobacteria to form mycobacteria-in-spheroid structures and granuloma lesions.

23. A method of making a 3D cell co-culture comprising a plurality of human cells and a plurality of mycobacteria, the method comprising:(a) suspending the human immune cells in a cell culture medium, further comprising one or more of L-glutamine, FBS, sodium pyruvate, and HEPES buffer;(b) mixing the mycobacteria with the human immune cells at a MOI of 0.005- 0.5; wherein the mixture is contained in a well of a microplate; and(c) incubating the mixture at 37°C for a time sufficient for the human immune cells and mycobacteria to form mycobacteria-in-spheroid structures and granuloma lesions.

24. The method of claim 23, wherein the human immune cells comprise cells selected from the group consisting of monocytes, macrophages, and peripheral blood mononuclear cells (PBMC).

25. The method of claim 24, wherein the human immune cells comprise monocytes, and wherein the monocytes comprise THP-1 cells, U937 histiocytes, CD14+monocytes, or monocyte cell subsets in PBMCs.

26. The method of claim 23, wherein the mycobacteria comprise Mycobacterium marinum (Mm) or Mycobacterium tuberculosis (Mtb).

27. The method of claim 26, wherein the mycobacteria comprise Mm, wherein the Mm comprises strain 1218 or strain M.

28. The method of claim 27, wherein the Mm are mixed with the human immune cells at an MOI of about 0.006-0.012.

29. The method of claim 23, wherein the mycobacteria comprise Mtb, and wherein the Mtb comprise strain H37Rv, strain Erdman, strain CDC1551, strain Beijing F2, or a clinical isolate of a drug-susceptible or drug-resistant Mtb.

30. The method of claim 29, wherein the Mtb are mixed with the human immune cells at an MOI of about 0.012-0.05.

31. The method of claim 30, wherein the MOI is about 0.025.

32. The method of claim 29, wherein the Mtb comprise strain H37Rv and are mixed with the human immune cells at an MOI of about 0.015-0.05.

33. The method of claim 23, wherein the mycobacteria express GFP, RFP, tdTomato, a far-red fluorescent protein, or another fluorescent protein.

34. The method of claim 33, wherein the RFP comprises Tdtomato.

35. The method of claim 23, wherein the cell culture medium comprises Roswell Park Memorial Institute culture medium.

36. The method of claim 23, wherein the well is made of polystyrene.

37. The method of claim 23, wherein the well has a U-bottom shape.

38. The method of claim 23, wherein the well comprises a hydrogel coating.

39. The method of claim 38, wherein the hydrogel is hydrophilic, neutrally charged, and biologically inert.

40. The method of claim 39, wherein the hydrogel comprises a CORNING® PURAMATRIX™ Peptide Hydrogel.

41. The method of claim 39, wherein the well is of a CORNING® Ultra-Low Attachment Spheroid Microplate or an S-BIO PRIMESURFACE® 3D Culture Spheroid Plate.

42. The method of claim 23, further comprising contacting the mycobacteria-in- spheroid structures with an extracellular matrix (ECM).

43. The method of claim 42, wherein the ECM is added to the mycobacteria-in- spheroid structures 3 days after the human immune cells and the mycobacteria are mixed.

44. The method of claim 23, further comprising placing and storing the mixture of human immune cells and mycobacteria at a temperature of -80°C, and optionally at a temperature of < -160°C, within 30 minutes of mixing the human immune cells and the mycobacteria.

45. The method of claim 23, further comprising after mixing the human cells and the mycobacteria, incubating the mixture at 37°C for 16-72 hours; removing the cell culture medium; adding a cryopreservative to the co-culture of human cells and mycobacteria; and freezing the spheroid co-culture.

46. The method of claim 45, wherein the cryopreservative comprises 5% (v / v) dimethyl sulfoxide (DMSO) in RPML1640 or 3D cell culture medium.

47. The method of claim 46, further wherein the cryopreservative further comprises heat-inactivated fetal bovine serum.

48. A method of screening for a molecule capable of preventing, treating or reducing a mycobacterial infection, comprising:(a) contacting the 3D bioplatform of claim 2 with the molecule;(b) measuring one or more characteristics of the mycobacteria-in-spheroid structures; and(c) comparing the one of the one or more characteristics to a control, wherein a change in at least one of the one or more characteristics is indicative that the molecule is capable of treating or reducing the mycobacterial infection or one or more pathology features thereof.

49. The method of claim 48, wherein the one or more characteristics comprise one or more of the following:(a) the amount of fluorescence produced by the mycobacteria, wherein the mycobacteria express a fluorescent molecule;(b) integrity of granulomas;(c) mycobacterial growth;(d) granuloma formation;(e) granuloma growth;(f) granuloma number and size;(g) autophagolysosome formation;(h) inflammasomes and pyroptosis induction;(i) hypoxia induction or inhibition;(j) lysosomal acidification;(k) host cellular toxicity; and(l) an expression level of one or more genes or protein biomarkers expressed by the cells of the 3D bioplatform.

50. The method of claim 49, wherein the one or more characteristics comprise the amount of fluorescence produced by the mycobacteria, wherein the amount of fluorescence is indicative of mycobacterial burden, and wherein a reduction in mycobacterial burden for the molecule by a threshold value in comparison to the control is indicative that the molecule is capable of treating or reducing a mycobacterial infection or one or more pathology features thereof.

51. The method of claim 50, wherein the threshold value is a reduction of at least 25%.

52. The method of claim 50, wherein the threshold value is a z score less than -2.

53. The method of claim 52, wherein the threshold value is a z score less than -4.

54. The method of claim 50, wherein when the mycobacteria compriseMycobacterium marinum, the human immune cells comprise THP-1 cells, and wherein the MOI for the mycobacteria is 0.007-0.012.

55. The method of claim 50, wherein when the mycobacteria comprise Mycobacterium tuberculosis strain H37Rv, Erdman, or Beijing F2, the human immune cells comprise THP-1 cells, and wherein the MOI for the mycobacteria is 0.012-0.05.

56. The method of claim 55, wherein the mycobacteria comprise Mycobacterium tuberculosis strain H37Rv and the MOI is 0.025.

57. The method of claim 55, wherein the mycobacteria comprise Mycobacterium tuberculosis strain Erdman or Beijing F2 and the MOI is 0.05.

58. The method of claim 49, wherein the one or more characteristics comprise the expression level of one or more genes expressed by the cells of the 3D bioplatform, wherein a change in the expression level of the one or more genes, proteins or immune processes as compared to the control is indicative that the molecule is capable of treating or reducing a mycobacterial infection or one or more pathology features thereof.

59. The method of claim 58, wherein the control comprises uninfected human immune cells.

60. The method of claim 49, wherein the molecule is added about 0-6 days after the mycobacteria and the human immune cells were co-incubated.

61. The method of claim 49, wherein the one or more characteristics is observed one or more times over 1-14 days, optionally once on days 7, 9, and 12, after the 3D bioplatform is contacted with the molecule.

62. A method of treating or reducing tuberculoma lesions in a subject in need thereof, comprising administering to the subject an anti-mycobacterial agent selected from the group consisting of an anti-CDl la antibody, an anti-a407 integrin antibody, an anti-CD30 antibody, an anti-IGFIR antibody, an anti-IL-6R antibody, AT9283, Tizoxanide (active metabolite of nitazoxanide), Dasatinib, Quinacrine diHCL, All-trans Retinoic acid (ATRA), Vorinostat, Sitagliptin, and Lansoprazole.

63. The method of claim 62, wherein the subject has a mycobacterial infection.

64. A method of treating or reducing tuberculoma lesions in a subject in need thereof, comprising administering to the subject an anti-mycobacterial agent selected from the group consisting of an anti-CDl la antibody, an anti-a407 integrin antibody, an anti-CD30 antibody, an anti-IGFIR antibody, an anti-IL-6R antibody, AT9283, Tizoxanide (active metabolite of nitazoxanide), Dasatinib, Quinacrine diHCL, All-trans Retinoic acid (ATRA), Vorinostat, Sitagliptin, H89, and Lansoprazole.

65. The method of claim 64, wherein the subject has a mycobacterial infection.

66. A method of screening for a candidate agent that induces trained immunity in a human immune cell, comprising measuring the growth of the pathogenic mycobacteria in the 3D platform of claim 1 after co-culture of the human immune cells and the pathogenic mycobacteria, wherein the human immune cells have been contacted with the candidate agent, wherein a decrease in growth of the pathogenic mycobacteria or reduced granuloma lesion development by the co-culture is indicative that the candidate agent induces trained immunity in the human immune cell.

67. The method of claim 66, wherein the candidate agent is a vaccine, immunotherapeutic, biologic, or host-directed therapeutic.

Images