Early detection of heterotopic ossification

Abstract

This disclosure relates to methods and systems for isolating, detecting, and / or analyzing circulating cells and particles, e.g., mesenchyme-derived cells or particles, present in the blood circulation, wherein the methods described herein can be used to determine whether the circulating cells or particles, e.g., mesenchyme-derived cells or particles, are associated with heterotopic ossification (HO). The methods can be used to provide an early diagnosis of HO, thus permitting early intervewntion, and to determine the efficacy of treatment.

Description

[0001] Early Detection of Heterotopic Ossification

[0002] CLAIM OF PRIORITY

[0003] This application claims the benefit of U.S. Provisional Patent Application Serial No. 63 / 596,720, filed on November 7, 2023. The entire contents of the foregoing are hereby incorporated by reference in their entireties.

[0004] TECHNICAL FIELD

[0005] This invention relates to the isolation of circulating cells and particles, e.g., mesenchyme-derived cells and particles, and their use to detect pathological bone formation, and more particularly to heterotopic ossification, e.g., after surgery, physical trauma, or burns.

[0006] BACKGROUND

[0007] Heterotopic ossification (HO), the abnormal growth of bone in soft tissues, is one of the most debilitating complications associated with severe burns, traumatic injuries, and joint replacement surgeries due to its insidious development. HO can also be caused by Fibrodysplasia Ossificans Progressiva (FOP), which is an extremely rare genetic connective tissue disorder characterized by the abnormal development of bone in areas of the body where bone is not normally present, such as the ligaments, tendons and skeletal muscles.

[0008] Although genetic forms of HO are rare, the incidence of trauma-induced HO is more common. In adults with traumatic tendon injuries, fractures, and sports injuries, the incidence range is -10-20%, whereas the incidence range is 60-70% for blast injuries, 20- 80% for joint, e.g., hip, knee, or elbow replacements, and more than 60% for major burns (1-6). HO severely diminishes the quality of life through pain, limited mobility, nerve damage, and poor wound healing. The gold-standard techniques for HO diagnosis include X-ray and computed tomography; however, these technologies only detect mature bone 6-8 weeks after injury when HO prevention or reversal is no longer possible. Currently, there are no technologies that can diagnose HO prior to irreversible symptoms such as joint contracture and pain or to guide early prophylactic strategies. Non-steroidal anti-inflammatory drugs, warfarin and pulse low-intensity electromagnetic field or radiation therapy are potential prophylactic measures. Bisphosphonates and excision are other potential treatments. While prophylactic strategies have been the focus of clinical studies, the only currently approved modalities are radiation therapy and non-steroidal anti-inflammatory medications. However, these treatments are non-specific and have off-target effects including fracture non-union, tissue fibrosis, malignancy, and wound healing complications (7, 8) making the need for precision medicine (only treating those at risk) a high priority. Additionally, the timing of initiation and termination of these prophylactic strategies remains unknown, and it is necessary to optimize treatment duration. Thus, a large percentage of patients (up to 80%) who will not go on to form HO end up receiving these potentially harmful prophylactic interventions. In addition to those treated unnecessarily, many patients are also treated at the wrong time and for the incorrect duration, given a lack of treatment guiding technologies. Thus, a point of care decision support tool to predict which patients will develop HO and assess the impact of treatment efficacy will provide a more “precision medicine” approach.

[0009] SUMMARY

[0010] The disclosure provides methods of isolating circulating cells and particles, such as mesenchyme-derived particles, and analyzing such cells and particles to diagnose and treat HO, along with microfluidic chip structures and systems that can be used with a group of select molecular markers, to enable the early detection of heterotopic bone formation in patients with surgeries, burns, or other physical trauma, or genetic mutations, before the bone is matured. Thus, the new methods and systems can be used to reduce, inhibit, and even prevent damage caused by HO. Thus, the new methods and systems provide a significant advantage over current diagnostic methods, which detect HO too late for prevention, leading to pain, surgeries, and loss of function for the patient.

[0011] The new methods provide point-of-care technology to detect mesenchyme-derived cells, and thus to diagnose and treat HO, e.g., at the bedside, based on circulating cells and particles, such as mesenchyme-derived cells, cell clusters, and particles, such as mesenchymal progenitor cells (MPCs), mesenchymal stem cells (MSCs), osteogenic or chondrogenic progenitor cells, extracellular vesicles, exosomes, and organelles. The new methods enable previously impossible observations (i.e., the observation of previously invisible MPCs), make use of novel biomarkers, and open new avenues with broad applications beyond the heterotopic ossification field.

[0012] In one aspect, the disclosure features methods for analyzing circulating cells and particles in a blood sample from a subject for association with heterotopic ossification (HO) by negative selection of circulating cells or particles. These methods include obtaining a blood sample from the subject, wherein circulating cells or particles in the blood sample comprise mesenchyme-derived cells or particles, fibrous tissue-derived cells or particles, cartilage-derived cells or particles, bone-derived cells or particles, or bone marrow-derived cells or particles; mixing the blood sample with magnetic beads and a binding agent that specifically binds to white blood cells (WBCs) for a time and under conditions sufficient for the binding agent to bind to the WBCs; flowing the blood sample through a size-based separation structure configured to direct red blood cells and platelets in the blood sample to a first waste outlet and flowing the remaining blood sample to an inertial focusing structure; flowing the remaining blood sample through the inertial focusing structure at a flow rate and for a distance sufficient to cause circulating cells and particles in the remaining blood sample to align in one or more streamlines within the remaining blood sample flowing through the inertial focusing structure; flowing the remaining blood sample with the circulating cells and particles aligned in one or more streamlines through a magnetophoresis structure for a time and distance sufficient to separate the WBCs bound to magnetic beads from the circulating cells and particles that are not bound to magnetic beads, flowing the bound WBCs into a second waste outlet, and flowing the remaining unbound circulating cells and unbound particles to a product outlet; and analyzing the unbound circulating cells or unbound particles from the product outlet for an HO-associated molecular signature comprising one or more of gene expression, protein expression, or post-translational modification, to determine whether the unbound circulating cells or unbound particles are HO-associated.

[0013] In another aspect, the disclosure features methods of analyzing circulating cells and particles in a blood sample from a subject for association with heterotopic ossification (HO) by positive selection of circulating cells or particles. These positive selection methods include obtaining a blood sample from the subject, wherein circulating cells or particles in the blood sample comprise mesenchyme-derived cells or particles, fibrous tissue-derived cells or particles, cartilage-derived cells or particles, bone-derived cells or particles, or bone marrow-derived cells or particles; mixing the blood sample with magnetic beads comprising a binding agent that specifically binds to circulating cells or particles and does not bind to white blood cells (WBCs) for a time and under conditions sufficient for the binding agent to bind to the circulating cells or particles; flowing the blood sample through a size-based separation structure configured to direct red blood cells and platelets in the blood sample to a first waste outlet and to direct the remaining blood sample to an inertial focusing structure; flowing the remaining blood sample through the inertial focusing structure at a flow rate and for a distance sufficient to cause circulating cells and particles in the remaining blood sample to align in one or more streamlines within the remaining blood sample flowing through the inertial focusing structure; flowing the remaining blood sample with the circulating cells and particles aligned in one or more streamlines through a magnetophoresis structure for a time and distance sufficient to separate the circulating cells and particles bound to magnetic beads from unbound WBCs and other unbound cells and unbound particles that are not bound to magnetic beads, flowing the unbound WBCs, other unbound cells, and unbound particles into a second waste outlet and flowing the bound circulating cells or bound particles to a product outlet; and analyzing the bound circulating cells or bound particles for an HO- associated molecular signature comprising one or more of gene expression, protein expression, or post-translational modification, to determine whether the bound circulating cells or bound particles are HO-associated.

[0014] In the negative selection methods, the binding agent that specifically binds to white blood cells can be an antibody that specifically bind to CD45, CD 16, or CD66b proteins found on the surface of white blood cells.

[0015] In the positive selection method, the binding agent that specifically binds to circulating cells or particles and not to WBCs or other cells or particles in the blood sample can be an antibody that specifically binds to plasma membrane proteins selected from the group consisting of PDGFRa, CD90, PGAP1, CDH11, GPX8, FBLN5, PRELP, LAMA4, or CADM3.

[0016] In all of the methods described herein, the HO-associated genes can be any one or more of the group consisting of Ackr3, Adm5, Asprvl, Cadm3, Cdhll, Collal, Colla2, Col3al, Col5al, Crtacl, Dact3, Dpysl3, Fmod, Gfpt2, Gpx8, Igfbp5, 1133, Lama4, Matn2, Mmp2, Msantd3, Tmeffl, Mxra5, Pcolce2, Pcyoxl, Penk, Pgapl, Plod2, Postn, Prelp, Rbms2, Trim2, Vnn3, Pcdhl9, or Dpp4. For example, one can select any one or more of Adm5, Asprvl, Colla2, Gjpt2, Gpx8, Matn2, Mxra5, Prelp, or Pcdhl9, or of Fmod, Col lai, Colla2, Col2al, Fgfr3, Dsp, Acvr2b, Inhba, 1133, Plod2, Ackr3, or Nt5c3b.

[0017] In some embodiments of these methods, the circulating cells or particles can be mesenchyme-derived cells or particles, e.g., one or more of mesenchymal progenitor cells (MPCs), mesenchymal stem cells (MSCs), osteogenic progenitor cells or chondrogenic progenitor cells. The circulating particles can be organelles or extracellular vesicles, e.g., one or more of microvesicles (MVs), exosomes, oncosomes, and apoptotic bodies.

[0018] In certain embodiments, the size-based separation structure can be or include an inertial exchanger configured to direct red blood cells, platelets, and particles in the blood sample to the first waste outlet and to direct the remaining blood sample to the inertial focusing structure.

[0019] In some embodiments, the size-based separation structure can be or include a deterministic lateral displacement array of microposts in a channel, wherein the array of microposts is configured to direct red blood cells, platelets, and particles in the blood sample to a first waste outlet and to direct the remaining blood sample to the inertial focusing structure.

[0020] In some embodiments, determining whether the circulating cells or particles are HO-associated circulating cells and particles includes analyzing the circulating cells or particles using droplet digital PCR, an immunoassay, or both. In certain embodiments, determining whether the circulating cells or particles are HO-associated circulating cells and particles includes analyzing the cells or particles using detection of antigens unique to HO-associated circulating cells or particles via fluorescently conjugated antibodies.

[0021] In certain embodiments, determining whether the circulating cells or particles are HO-associated circulating cells and particles includes analyzing the circulating cells or particles using HO-associated genes, transcripts, or proteins for differentiating HO- associated circulating cells or particles from non-HO-associated circulating cells or particles. In other embodiments, determining whether the circulating cells or particles are HO-associated circulating cells and particles includes analyzing the circulating cells or particles using single-cell RNA sequencing.

[0022] In certain embodiments, the new methods can further include detecting a quantity of the HO-associated circulating cells or particles, a quality of the HO-associated circulating cells or particles, or both, for determining a specific level of risk of heterotopic ossification. For example, in these methods, a quantity of HO-associated circulating cells in the blood sample from the subject that is higher than a reference quantity determined from HO-associated circulating cells in blood samples from a plurality of healthy individuals indicates the subject is at risk for developing HO.

[0023] In some embodiments, the methods further include treating a subject at risk for developing HO by administering an anti-HO treatment, e.g., the anti-HO treatment can be or include administering a non-steroidal anti-inflammatory drug, warfarin, an electromagnetic field, radiation therapy, or a bisphosphonate drug.

[0024] In certain embodiments, the methods can further include determining the efficacy of the anti-HO treatment, by taking a second blood sample at a first time, at or before starting the treatment, to determine a first level or number of HO-associated circulating cells or particles in the second blood sample of the subject; taking a third blood sample at a second time, later than the first time, to determine a second level or number of HO- associated circulating cells or particles in the third blood sample of the subject; and determining a level of efficacy of the treatment, wherein a decrease in a level or number of HO-associated circulating cells or particles at the second time compared to the level or number of HO-associated circulating cells or particles at the first time indicates that the treatment was effective.

[0025] In other embodiments, the efficacy of the anti-HO treatment can be determined by taking a second blood sample at a first time, at or before starting the treatment, to determine a first level or number of HO-associated circulating cells or particles in the second blood sample of the subject; taking a third blood sample at a second time, later than the first time, to determine a second level or number of HO-associated circulating cells or particles in the third blood sample of the subject; and determining a level of efficacy of the treatment, wherein a decrease in a level or number of HO-associated genes in circulating cells or particles at the second time compared to the level or number of HO- associated genes at the first time indicates that the treatment was effective.

[0026] In another aspect, this disclosure provides systems for isolating, analyzing, or both isolating and analyzing, circulating cells or particles from a blood sample from a subject, wherein the circulating cells or particles comprise mesenchyme-derived cells or particles, fibrous tissue-derived cells or particles, cartilage-derived cells or particles, bone-derived cells or particles, or bone marrow-derived cells or particles. These systems include a mixer for combining the blood sample with magnetic beads comprising a binding agent that specifically binds to either (i) circulating cells or particles and not to white blood cells (WBCs), or (ii) WBCs and not to other cells or particles, for a time and under conditions sufficient for the binding agent to bind; a size-based separation structure configured to direct red blood cells and platelets in the blood sample to a first waste outlet; an inertial focusing structure configured to align circulating cells and particles in the remaining blood sample into one or more streamlines within the remaining blood sample flowing through the inertial focusing structure; a magnetophoresis structure configured to separate cells or particles bound to magnetic beads from cells and particles not bound to magnetic beads, and directing the bound cells or bound particles into a second waste outlet and flowing unbound cells or unbound particles to a product outlet; and an analyzer configured to determine which of the circulating cells or particles are associated with a heterotopic ossification (HO)-associated molecular signature comprising one or more of gene expression, protein expression, or post-translational modification, to determine whether the circulating cells or unbound particles are HO- associated.

[0027] In certain embodiments of these systems, the size-based separation structure can be or include an inertial exchanger configured to direct red blood cells and platelets in the blood sample to a first waste outlet and to direct the remaining blood sample to the inertial focusing structure. In other embodiments, the size-based separation structure can be or include a deterministic lateral displacement array of microposts in a channel, wherein the array of microposts is configured to direct red blood cells and platelets in the blood sample to a first waste outlet and to direct the remaining blood sample to the inertial focusing structure.

[0028] In certain embodiments, the analyzer is or includes a system to encapsulate cells or particles in individual droplets and to perform ddPCR on each individual droplet to determine which circulating cells or particles are HO-associated. In other embodiments, the analyzer is or includes a system that performs RNA sequencing to determine whether the circulating cells or particles are HO-associated circulating cells and particles.

[0029] In another aspect, the present disclosure provides methods of detecting mesenchyme-derived cells in a blood sample. These methods include mixing a blood sample with magnetic beads and a binding agent that specifically binds to white blood cells (WBCs) for a time and under conditions sufficient for the binding agent to bind to the WBCs; flowing the blood sample through a size-based separation structure configured to direct red blood cells and platelets in the blood sample to a first waste outlet and flowing the remaining blood sample to an inertial focusing structure; flowing the remaining blood sample through the inertial focusing structure at a flow rate and for a distance sufficient to cause cells and particles in the remaining blood sample to align in one or more streamlines within the remaining blood sample flowing through the inertial focusing structure; flowing the remaining blood sample with the circulating cells and particles aligned in one or more streamlines through a magnetophoresis structure for a time and distance sufficient to separate the WBCs bound to magnetic beads from the cells and particles that are not bound to magnetic beads, flowing the bound WBCs into a second waste outlet, and flowing the remaining unbound cells and unbound particles to a product outlet; and analyzing the unbound cells or unbound particles from the product outlet for a mesenchyme-derived molecular signature comprising one or more of gene expression, protein expression, or post-translational modification, to determine whether the unbound cells or unbound particles are mesenchyme-derived. These methods can also be conducted using positive selection as described herein.

[0030] In some embodiments of these methods, the binding agent that specifically binds to white blood cells can be an antibody that specifically bind to CD45, CD 16, and CD66b proteins found on the surface of white blood cells.

[0031] In some embodiments, the genes associated with mesenchyme-derived cells include those selected from the group consisting of Ackr3, Adm5, Asprvl, Cadm3, Cdhll, Col lai, Colla2, ColSal, Col5al, Crtacl, Dact3, DpyslS, Fmod, Gfpt2, Gpx8, Igfbp5, 1133, Lama4, Matn2, Mmp2, Msantd3, Tmeffl, Mxra5, Pcolce2, Pcyoxl, Penk, Pgapl, Plod2, Postn, Prelp, Rbms2, Trim2, Vnn3, Pcdhl9, and Dpp4.

[0032] In certain embodiments, the mesenchyme-derived cells include one or more of mesenchymal progenitor cells (MPCs), mesenchymal stem cells (MSCs), osteogenic progenitor cells or chondrogenic progenitor cells, and / or the particles comprise organelles or extracellular vesicles, such as microvesicles (MVs), exosomes, oncosomes, and apoptotic bodies. In these methods, the circulating cells and particles can be as mesenchyme- derived cells or particles, fibrous tissue-derived cells or particles, cartilage-derived cells or particles, bone-derived cells or particles, or bone marrow-derived cells or particles. As used herein, fibrous tissue-derived cells arise from muscle, fat, tendon, or ligaments where HO can develop. The mesenchyme-derived cells or particles, can be MPCs, MSCs, osteogenic or chondrogenic progenitors, and mesenchyme-derived particles such as organelles or extracellular vesicles, e.g., microvesicles (MVs), exosomes, and apoptotic bodies, from a blood sample from a subject.

[0033] Analyzing cells, such as MPCs, as described herein, can be done, for example, using droplet digital polymerase chain reaction (ddPCR) and RNA sequencing can also be done with cells that are isolated using other known methods of isolation.

[0034] In some embodiments, the mesenchyme-specific and / or HO-associated genes include Ackr3, Adm5, Asprvl, Cadm3, Cdhll, Collal, Colla2, Col3al, Col5al, Crtacl, Dact3, Dpysl3, Fmod, Gjpt2, Gpx8, Igfbp5, 1133, Lama4, Matn2, Mmp2, Msantd3, Tmeffl, Mxra5, Pcolce2, Pcyoxl, Penk, Pgapl, Plod2, Postn, Prelp, Rbms2, Trim2, Vnn3, Pcdhl9, and Dpp4. In some embodiments, the mesenchyme-specific and HO-associated genes include Adm5, Asprvl, Colla2, Gjpt2, Gpx8, Matn2, Mxra5, Prelp, and Pcdhl9. In some embodiments, the mesenchyme-specific and HO-associated genes include Fmod, Collal, Colla2, Col2al, Fgfr3, Dsp, Acvr2b, Inhba, 1133, Plod2, Ackr3 (also known as Cxcr7), and Nt5c3b, and transcripts of these genes can be used to detect circulating mesenchyme-derived cells associated with HO comprising MPCs and MSCs.

[0035] In various implementations of these methods, the subject, such as a human or animal (e.g., cat, dog, mouse, rat, rabbit, monkey, ape, pig, cow, sheep, goat, or horse) subject, has a bum, surgery, and / or other traumatic injury or a heterotopic ossification- associated genetic mutation that leads to heterotopic ossification.

[0036] In some embodiments, the methods can further include detecting a quantity of the mesenchyme-derived cells or particles, e.g., MPCs, a quality of the mesenchyme-derived cells or particles, or both, for detecting a specific type of heterotopic ossification or other mesenchymal abnormality.

[0037] In some implementations, the magnetic beads specifically bind to WBCs and not to mesenchymal cells or the magnetic beads specifically bind to mesenchymal cells and not to WBCs. Any of these methods can further include separating the mesenchyme- derived cells, e.g., MPCs, or mesenchyme-derived particles, from other cells and / or particles in the blood sample to isolate the mesenchyme-derived cells or particles.

[0038] As used herein, “circulating cells or particles” are mesenchyme-derived cells or particles, fibrous tissue-derived cells or particles, cartilage-derived cells or particles, bone-derived cells or particles, or bone marrow-derived cells or particles.

[0039] As used herein, fibrous tissue-derived cells arise from muscle, fat, tendon, or ligaments” where HO can develop.

[0040] The present disclosure describes systems and methods that overcome the barriers to early diagnosis of HO and capture rare cMPCs by employing microfluidic technology with unprecedented sensitivity, cost-effectiveness, and bedside point-of-care potential. The new systems and methods can be used to capture ultra-rare cMPCs with high sensitivity in limited volume blood samples, differentiate cMPCs from other circulating mesenchymal cells that are not associated with heterotopic ossification, and can be used to elucidate the mechanism of formation of heterotopic ossification, as well as other mesenchymal disorders. Once the nature of a relationship is established as diagnostic for the specific connective tissue injury or disorder, different clinical options are provided that depend on the patient condition.

[0041] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs. Although methods and materials similar or equivalent to those described herein can be used in the practice or testing of the present invention, suitable methods and materials are described below. All publications, patent applications, patents, and other references mentioned herein are incorporated by reference in their entirety. In case of conflict, the present specification, including definitions, will control. In addition, the materials, methods, and examples are illustrative only and are not intended to be limiting.

[0042] The details of one or more embodiments of the disclosure are set forth in the accompanying drawings and the description below. Other features, objects, and advantages of the invention will be apparent from the description and drawings, and from the claims. DESCRIPTION OF DRAWINGS

[0043] FIG. 1 is a flow chart of the progression of HO from injury to heterotopic bone formation, and showing the window for best treatment chances as compared to the current window of diagnostic options, well after the best treatment window.

[0044] FIG. 2 is a schematic diagram that shows a blood vessel filled with circulating MPCs after a bum or traumatic injury, and the progression of the formation of ectopic bone, and the initial window for circulating MPF (cMPC) detection with a simple blood test using the methods disclosed herein.

[0045] FIG. 3 is a schematic representation of an example of a microfluidic chip as described herein and illustrating the key steps to capture circulating cell, e.g., mesenchyme-derived cells and particles.

[0046] FIG. 4 is a schematic flow chart of the progression of testing for a trauma, e.g., bum, or surgery patient according to the methods of liquid biopsy disclosed herein to provide an early diagnosis of whether the patient has a low or high risk of HO.

[0047] FIGs. 5A-5C are a series of representations of a microfluidic iChip™ equipped with non-equilibrium inertial separation array (NISA) that includes structures for inertial debulking, inertial focusing, and magnetic separation, e.g., magnetophoresis, which can be used in the methods and systems described herein.

[0048] FIGs. 5D-5G are a series of representations of another example of a disk-shaped microfluidic chip including structures for hydrodynamic size-based cell sorting, inertial focusing, and magnetophoresis for use with the methods and systems described herein.

[0049] FIG. 6 is a graph that shows the percentage of mouse bone marrow MPC recovery when isolated using the NISA-iChip of about 100%.

[0050] FIGs. 7A and 7B are a flow chart (7 A) and graph (7B) that show the format of a test in mice for a model of HO injury caused by a burn and trauma (which induces HO), no injury (no HO, as a control), and a burn that does not induce HO (another no HO control). The mouse blood samples are run through the NISA-iChip and then analyzed by flow cytometry. The graph in FIG. 7B shows the number of cells per ml of blood after the different treatments for the PDGFRa+ / CD90+ and CD45- cell subsets. The largest number of mesenchyme-derived cells were isolated from the induced HO mice, with far lower numbers of cells from the controls, as well as from blood samples from muscle fibrosis and traumatic brain injury subjects. FIGs. 8A-8C are a flow chart (8A), a pair of graphs (8B), and a series of microscope images (8C) showing another format of a test in mice for a model of HO injury caused by a burn and trauma (which induces HO), no injury (no HO, as a control), and a bum that does not induce HO (another no HO control). The mouse blood samples are run through the NISA-iChip and then analyzed by flow cytometry. The graphs in FIG. 8B show the number of cells per ml of blood after the induced HO and the right panel shows labeled cells with the PDGFRa+ / CD90+ cMPC gate. FIG. 8C shows a series of immunofluorescent histology microscope images that show an uninjured mouse hindlimb and 12 weeks post-injury hindlimb, with chevrons marking double-positive cells.

[0051] FIG. 9 is an image showing a uniform manifold approximation and projection (UMAP) representation of cells detected in microfluidics-enriched cells from a blood sample using scRNAseq as described herein. The MPCs are shown circled in the upper right hand quadrant.

[0052] FIGs. lOAis a schematic overview of an approach to identify cMPC-based markers for early detection of HO based on microfluidic sorting of cMPCs from whole blood samples taken from a mouse burn / tenotomy model.

[0053] FIG. 1 OB is a graphic representation of cMPC HO gene signatures for specific genes from blood samples taken from human patients who had undergone knee or hip arthroplasty, which can induce HO. Samples we taken on post-operative day 1 (POD1).

[0054] FIG. IOC is a graphic representation of cMPC HO gene signatures for specific genes from blood samples taken from a mouse B / T model of induced HO. Samples were taken on post-operative day 1 (DI), day 3 (D3), and day 7 (D7).

[0055] FIGs. 11A-11F are a series of graphs that show expression levels of select cMPC HO signature genes in human joint replacement surgery patients across different time points.

[0056] FIGs. 12A-12B are a series of graphs the show cMPC HO risk score performance. FIG. 12A is a pair of graphs including a bar graph that shows a threshold (dotted line) that achieved a 92% overall specificity and 91% overall sensitivity for this dataset from a mouse model, and another graph showing sensitivity vs specificity of the cMPC HO signature ROC. FIG. 12B shows a cMPC HO Score ROC curve one day post-surgery, showing that the methods described herein can predict HO weeks or months earlier than radiography methods.

[0057] FIG. 13 A is a representation of a microscope image of a mouse model of induced HO showing a control and BAPN-treated hind limb. P-aminopropionitrile (BAPN) is a pharmacological inhibitor of LOX (lysyl oxidase) and LOXLs (LOX-like proteins) and administration of BAPN targets MPCs and is used to treat HO.

[0058] FIG. 13B is a pair of bar graphs that show quantification of HO volume in mice treated with BAPN vs. vehicle control.

[0059] FIG. 13C is a bar graph that shows the cMPC HO score obtained as described herein decreases upon HO reducing treatment with BAPN (right-most bar). The B / S bar is a control that does not induce HO, and the B / T bar is the HO+ induced case.

[0060] FIG. 13D is a series of bar graphs that shows the results for a set of cMPC markers that accurately detect the reduction of HO due to treatment with BAPN at an early time point.

[0061] DETAILED DESCRIPTION

[0062] The present disclosure provides methods and systems that can include a microfluidic chip, along with binding agents that bind specifically to a group of selected molecular markers that enable detection of circulating cells and particles, i.e., mesenchyme-derived cells or particles, fibrous tissue-derived cells or particles, cartilage- derived cells or particles, bone-derived cells or particles, or bone marrow-derived cells or particles. As used herein, fibrous tissue-derived cells arise from muscle, fat, tendon, or ligaments” where HO can develop.

[0063] Such circulating cells and particles, e.g., mesenchyme-derived cells and particles, that are associated with heterotopic bone formation in subjects (heterotopic ossification, or HO) after surgeries, burns, or other physical trauma, or who have specific genetic mutations, can now be detected using the methods described herein before the heterotopic bone is matured. Thus, any damage that could otherwise be caused by HO can be reduced, inhibited, or even prevented. The new methods can detect the extremely rare mesenchyme-derived cells and particles in peripheral blood, even at locations remote from the site of bum or trauma.

[0064] Changes associated with HO occur much earlier than standard radiographic detection, however, these modalities require special imaging and are not regularly available (28-30). Currently, clinical detection of HO is performed with computed tomography (CT) (takes an average of 23 days after symptom development to detect HO)(31); bone scans with 99mTc-MDP (low specificity, which leads to difficulties in discriminating HO from other inflammatory, traumatic, or degenerative skeletal processes) and serum ALP measurement (sensitive, not specific, with alterations dependent on hepatic and renal function)(32). Thus, one of the major knowledge gaps in HO treatment is a lack of accurate early diagnostic modalities.

[0065] Precision medicine approaches to guide therapy for those patients who are developing HO as well as precise timing of treatment as described herein will allow for shorter treatment duration with minimization of adverse consequences, decreases in cost, and improvement in treatment adherence. Additionally, the in vivo blood-based biomarkers described herein will allow for targeted treatment on the basis of early identification of changes in MPC fate. Demonstration of therapeutic efficacy based on patient selection upon cMPC evidence of HO, as described herein, will dramatically enhance current treatment paradigms.

[0066] As shown in FIG. 1, current diagnostic methods detect heterotopic bone formation too late for prevention, leading to pain, surgeries, and loss of function for the patient. Although some mesenchymal cells have been detected in blood samples in prior HO models, we discovered molecular markers that enable detecting HO early. The use of microfluidic chips in these methods as described herein is also novel and enables the detection of mesenchyme-derived cells and particles from blood samples with high sensitivity and specificity. The disclosure further provides a description of a working protocol and the results as tested in mouse models of trauma-induced HO and of genetic HO (FOP).

[0067] When taken together, the data presented herein confirm the presence of cMPCs which are sensitive and specific to an HO-inducing injury. The data presented suggests that MPCs are mobilized from the HO site into the bloodstream after an HO-inducing injury and express HO-specific genes, long before HO can be detected radiographically. Beyond early diagnosis, our findings also validate cMPCs as a liquid biopsy to accurately assess treatment efficacy.

[0068] The present disclosure provides new point-of-care technology to detect HO, at the bedside and earlier than available radiographic diagnostics based on circulating cells and particles, such as mesenchyme-derived cells, cell clusters, and particles, e.g., mesenchymal progenitor cells (MPCs), mesenchymal stem cells (MSCs), osteogenic or chondrogenic progenitor cells, extracellular vesicles, exosomes, and organelles. The technology described herein enables previously impossible observations (i.e., the observation of previously invisible MPCs), reveal novel biomarkers, and open new avenues with broad applications beyond the heterotopic ossification field.

[0069] General Methodology

[0070] As shown in FIG. 2, circulating mesenchyme-derived cells, cell clusters, and particles, such as mesenchymal progenitor cells (MPCs), mesenchymal stem cells (MSCs), osteogenic or chondrogenic progenitor cells, extracellular vesicles, exosomes, and organelles, are stimulated by a bum or other traumatic injury, and are then stimulated to form clusters and eventually ectopic bone. By detecting these HO-stimulated mesenchyme-derived cells, EVs, exosomes, and organelles in peripheral blood of a patient at higher levels than in healthy subjects, one can make an early diagnosis of HO.

[0071] The new methods for isolating and / or analyzing mesenchyme-derived cells or particles, such as MPCs, MSCs, osteogenic or chondrogenic progenitors, and mesenchyme-derived particles such as organelles or extracellular vesicles, e.g., microvesicles (MVs), exosomes, and apoptotic bodies, from a blood sample from a subject can methods include obtaining a blood sample from the subject using any known techniques. In one embodiment, the blood sample is then mixed with magnetic beads including a binding agent, e.g., an antibody, such as a monoclonal antibody, that specifically binds to surface markers on white blood cells (WBCs), e.g., CD45, CD16, or CD66b, proteins found on the surface of white blood cells, and not specifically bind to the other cells or particles, for a time and under conditions sufficient for the binding agent to bind to the WBCs.

[0072] As shown in the schematic diagram of FIG. 3, in a system of negative depletion, the blood sample can be flowed into a size-based hydrodynamic cell sorting (e.g., deterministic lateral displacement) module comprising a microfluidic size-based separation structure configured to direct small cells and particles such as red blood cells (RBCs) and platelets in the blood sample to a first waste outlet (see upper right outlet in FIG. 3) using size-based, hydrodynamic cell sorting through an array of obstacles, and to direct the remaining blood sample to a second module comprising an inertial focusing structure (see sinusoidal channel along the left side). The remaining blood sample is then flowed through the second module at a flow rate and for a distance sufficient to cause cells and / or particles in the remaining blood sample to align in one or more streamlines within the remaining blood sample flowing through the inertial focusing channel.

[0073] For a description of such first and second modules, see, e.g., US Patent Nos. 8,784,012; 9,034,658; and 10,018,632. Such systems can interrogate as many as 30 million cells per second (see also references 1, 4, and 20).

[0074] Thereafter, the remaining blood sample containing with the cells and / or particles aligned in one or more streamlines is flowed through a third module comprising a magnetophoresis system for a time and distance sufficient to separate WBCs, which are bound to magnetic beads, from cells and particles, which are not bound to magnetic beads. For a description of such magnetophoresis systems, see, e.g., US Patent Nos. 9,278,353, 9,878,327; and 10,668,470. As further shown in FIG. 4, in such third modules, the WBCs (bound to magnetic beads) are automatically flowed towards a second waste outlet (WBCs, lower right outlet in FIG. 3) by the magnetic forces, e.g., provided by permanent magnets, or electromagnets, and other cells and particles (not bound to magnetic beads) flow naturally to a product outlet (cMPCs, lower left outlet in FIG. 3).

[0075] The microfluidic chip shown schematically in FIG. 3 includes elements or modules as described, for example, in US Patent Nos. 8,186,913; 8,784,012; 9,034,658; 9,610,582; 9,808,803; 9,895,694; and 10,018,632; and in US Published Patent Application No. US2016 / 0123858; which are all incorporated herein by reference in their entireties. The new systems can process high volumes of blood, e.g., 20 cc of blood in 30 minutes, to find extremely rare target cells, e.g., one target cell in 109blood cells (i.e., can find a needle in a haystack). Moreover, captured cells are healthy, and can be used for molecular analyses such as single cell RNA sequencing (scRNAseq).

[0076] In other embodiments, the systems described herein can also be run in a positive selection mode to isolate the target cells directly by using the same system, but run with binding agents that specifically bind to the circulating cells and particles, e.g., mesenchymal-derived cells and particles. For example, the bind agents can be antibodies that bind to plasma membrane proteins selected from the group consisting of PDGFRa, CD90, PGAP1, CDH11, GPX8, FBLN5, PRELP, LAMA4, and CADM3. In a further step, the cells or particles that exit from the product outlet are obtained and analyzed in a fourth analysis module, for mesenchyme-derived gene and / or protein expression, e.g., using any known nucleic acid or protein detection techniques, and then one determines whether the cells or particles express genes associated with HO ossification as described herein. In particular, mesenchyme-derived cells, such as MPCs, as described herein, can be analyzed, e.g., using droplet digital polymerase chain reaction (ddPCR) and / or RNA sequencing to detect the specific marker genes described herein.

[0077] The mesenchyme-specific and HO-associated genes include those selected from the group consisting of Ackr3, Adm5, Asprvl, Cadm3, Cdhll, Collal, Colla2, Col3al, Col5al, Crtacl, Dact3, Dpysl3, Fmod, Gfpt2, Gpx8, Igfbp5, 1133, Lama4, Matn2, Mmp2, Msantd3, Tmeffl, Mxra5, Pcdhl9, Pcolce2, Pcyoxl, Penk, Pgapl, Plod2, Postn, Prelp, Rbms2, Trim2, Vnn3, and Dpp4. In some embodiments, the presence of smaller subsets these genes, e.g., one, two, three, four, five, six, or more of these genes is analyzed. For example, one, two, or more of the following genes, which provide higher sensitivity and specificity for HO, can be analyzed, e.g., in parallel: Adm5, Asprvl, Colla2, Gjpt2, Gpx8, Matn2, Mxra5, Prelp, and Pcdhl9.

[0078] The genes described herein that are detected specifically in cMPCs have functional relevance for endochondral ossification and / or trauma-associated HO. For example, Fmod, encodes a collagen-binding small leucine-rich proteoglycan, and is expressed during fetal endochondral and intramembranous ossification36, articular cartilage formation37’38and its amount is correlated with the size of collagen fibrils in cartilage39. Ackr3, is a receptor of Cxcll2, is associated with mediating osteogenesis40, and is specifically expressed by MPCs with osteogenic and chondrogenic differentiation abilities derived from neural crest cells ex vivo41. Plod2 codes a collagen cross-linking enzyme that is highly upregulated in individuals with trauma-associated HO and in B / T mice23and its inhibition in MPCs reduces bone mineralization in vitro23. Colla2 encodes a characteristic extracellular matrix protein abundantly expressed in fibroblastic cells across many tissues42. 1133 is expressed by PDGFRa+ progenitors in muscle tissues, is enriched following acute injury to signal to immune cells for promotion of repair43, 44. Overall, these genes have functional relevance to tHO, and not detected in the bloodstream for healthy donors and mice without HO. Several mesenchymal collagen forms (Colla2, Collal, Col3al, and Col5aP) highly expressed in MPCs and other genes directly associated with extracellular matrix and ossification (Igfbp5, Matn2, Lama4, Mmp2) are also detected. Other genes that are included in the list are, cell adhesion associated genes such as Cdhll, Pcdhl9, Postn, Cadm3, and metabolic genes such as Gfpt2, Gpx8, Pcyoxl expressed in MPCs.

[0079] Therefore, screening these markers in circulation and analyzing them as described herein benefits subjects at risk for HO, future studies of tHO and other musculoskeletal pathologies, and for monitoring of prophylactic strategies. The liquid biopsy approach using cMPCs described herein can be used in HO prevention in subjects who are at risk, and can reduce the duration of treatment and the timing of treatment initiation.

[0080] As shown in FIG. 4, the new methods provide early detection of HO by conducting a simple blood test, a so-called liquid biopsy, for early detection of HO as described herein, to determine whether the patient is at high or low risk for developing HO. If there is a low risk, then the patient receives no prophylaxis against HO. If the patient is at high risk, then prophylaxis, e.g., in the form of NSAIDS, radiation, and / or immobilization, is provided to the patient. Thereafter, in some embodiments, the methods can include microfluidic and mass cytometry -based HO therapy monitoring, again, using the methods described herein, with follow-up diagnostics, such as CT scan and x-rays. The follow-up methods are ceased once the patient has been determined to be free of HO.

[0081] Treatment timing is important for several reasons, including: 1. Reduced treatment duration: by identifying the optimal treatment timing, we can minimize treatment length and monitor for recurrence. This targeted timing will minimize adverse consequences of therapeutics, improve patient adherence, and reduce costs. 2. Precision medicine-based patient selection: portable, point of care, non-invasive diagnostic modalities which can risk-stratify patients will allow for targeted treatments so that only patients at risk receive treatment. The potential impact of cMPC liquid biopsy is not limited to the patients at risk of HO. Aberrant cell fate and repair associated with PDGFRa+ mesenchymal cells are also observed in osteoarthritis45and fibrosis46. Early blood-based detection strategies, that can be routinely applied to patients at risk, is an active area of research for these diseases. Thus, the mesenchyme-derived cell and particle liquid biopsy methods described herein can provide direct and easy access to cMPCs for molecular, physiological, and pharmacological characterization in a vast array of acute or chronic diseases. The new systems are also portable and affordable point-of-care devices and will provide a significant clinical impact. For example, detecting the circulating mesenchyme- derived cells associated with HO after a fire or other serious accident, on the battlefield, or in emergency rooms will save millions of dollars every year in timely diagnosis and management, and help specific measures to be taken to reduce, inhibit, or even prevent HO’s debilitating effects. Thus, the presently disclosed microfluidic technology can transform the research, diagnosis, and management of HO by allowing large scale screening for and early detection of high-risk individuals, guiding treatment selections, and monitoring treatment efficacy.

[0082] Previous studies to isolate mesenchyme-derived cells such as MPCs have used flow cytometry with FACS, which has significant caveats: (a) MOCs are ultra-rare (few cells / ml whole blood), at or below the detection limit of flow cytometry; (b) FACS is labor intensive, expensive, time consuming, not available everywhere, and especially not at the bedside; and (c) cross-reactivity of established antibodies with platelets, lymphocytes, and other circulating mesenchymal progenitor cells can confound FACS.

[0083] The new microfluidic technology described herein offers distinct advantages over FACS. High-fidelity negative depletion of circulating blood cells eliminates the vast majority of platelets, erythrocytes (red blood cells (RBCs)), and white blood cells (WBC). In this way, the microfluidic chip achieves unprecedented enrichment of MPCs, greatly increasing its sensitivity to capture even a single cell in 109blood cells. Captured cells in the product are then interrogated (i.e., using ddPCR and immunolabeling) for a sensitive and specific molecular signature to distinguish the target cells. Moreover, because the blood is processed uniformly and physiologically, and the transit time is only 3 seconds, isolated cells are alive and healthy for accurate phenotypic characterization, in vitro (e.g., scRNAseq). In other implementations, positive selection can also be used.

[0084] Of course, cells such as MPCs can be analyzed as described herein, e.g., using ddPCR, even if the cells have been isolated using known methods and systems of isolation other than the microfluidic methods and systems described herein.

[0085] Microfluidic Systems - the iChip™

[0086] A useful system to capture circulating cells and particles, such as mesenchyme- derived cells and particles, should have high sensitivity and specificity, and be cheap, quick, and portable for the bedside (e.g., small clinics) and the field (e.g., ambulances, fire engines, or battlefields). We developed a microfluidic chip (so-called “iChip™”) for inertial focusing) to separate ultra-rare cells directly from whole blood by negative depletion (see, Ozkumur, Shah et al. 2013, Fachin, Spuhler et al. 2017). The device is independent of preselected surface markers on target cells (i.e., target antigenindependent), which is a major advance in rare cell isolation. However, in some implementations, the system can also be used for positive selection of the target cells when specific preselected surface markers are known for the target cells.

[0087] FIGs. 5A-5C show a microfluidic chip system that can be used in the methods described herein. This version of the microfluidic device utilizes non-equilibrium inertial separation array (NISA) to perform depletion of RBCs, platelets (PLT) and plasma. As shown in FIG. 5A, this system includes structures for size-based separation using inertial debulking, following by inertial focusing, and then magnetic separation. As shown in FIG. 5B, the inertial debulking structure (module) utilizes NISA optimized for removing RBCs, PLTs, plasma and keeping leukocytes and rare cells to pass onto the next structure - the inertial focusing channel. FIG. 5C shows a schematic that illustrates the physics of the NISA mechanism. The particle migrates away from the wall due to the wall lift force. After migration, a fraction of the flow is separated (top liquid that is siphoned off to waste), but the particle has migrated sufficiently lower to avoid siphoning. The migration trajectory is dictated by the particle diameter (a), flow velocity (U), island length (L) and siphon percentage (T). FIG. 5C also shows inlet and outlet images of the analytical device during separation of 5 and 10 pm fluorescent polystyrene particles to illustrate the mechanism.

[0088] FIGs. 5D-5G show another system (FIG. 5D) that includes the first three modules in the form of a circular disk-shaped substrate, wherein the first module provides hydrodynamic, size-based cells sorting through an array of specifically spaced and sized obstacles (see FIG. 5E), and the second module provides inertial focusing (FIG. 5F), and the third module provides magnetophoresis (FIG. 5G).

[0089] Key steps of the methods are illustrated in the schematic of FIG. 3 described above, and are described as follows:

[0090] (1) Magnetic tagging of target cells or particles with magnetic beads that include binding agents that specifically bind to the target cells or particles. For example, in a negative depletion (i.e., negative selection) mode, the target cells can be WBCs, and they can be bound to the magnetic beads (e.g., 1, 2, 3, 4, or 5 microns in diameter) using antibodies conjugated, e.g., biotin-conjugated, to the surface of the beads, such as anti- WBC antibodies, e.g., anti-CD45, anti-CD16, and anti-CD66b antibodies. In other embodiments, a positive selection mode is used in which the target cells are the cells of interest, e.g., mesenchyme-derived cells, which can be targeted using anti-PDGFRa, anti- CD90, anti-PGAPl, anti-CDHll, anti-GPX8, anti-FBLN5, anti-PRELP, anti-LAMA4, or anti-CADM3 antibodies coated, i.e., conjugated on the surface of, magnetic beads. Similarly, other binding agents, such as antibodies, that are known for specifically binding to mesenchyme-derived cells or particles can also be used.

[0091] (2) Size-based separation (in a first module or structure of the microfluidic chip), e.g., inertial exchanger or deterministic lateral displacement, for separation of small cells and particles (e.g., endothelial progenitor cells (EPCs), red blood cells (RBCs), and platelets) into Waste Outlet 1. This step is analogous to centrifugation or Ficoll gradient to prepare “buffy coat,” but with ultra-high precision.

[0092] (3) Inertial focusing (in a second module or structure of the microfluidic chip) of WBCs and target cells, e.g., endothelial cells, in a microfluidic inertial focusing channel to align the cells into one or more streamlines within the flowing blood sample (analogous to flow cytometry, but without the sheath flow), to facilitate high-fidelity deflection into waste or product channels with minimal magnetic moment.

[0093] (4) Magnetophoresis (in a third module or streucture) of i) WBCs into Waste Outlet 2 (negative depletion mode), leaving highly purified, untouched, and viable target cells and particles, such as mesenchyme-derived cells in the Product Outlet (similar to magnetic-activated cell sorting, but with extreme precision and sensitivity, without target cell injury), or ii) mesenchyme-derived cells into the Product Outlet, to sort them with high purity and sensitivity (positive selection mode).

[0094] The microfluidic chip (iChip™) can process large volumes of blood (40 ml / h) with unmatched throughput (20 million cells / sec), without losing target brain-derived cells and / or particles. The iChip, originally developed for circulating tumor cell detection, has been validated extensively to detect and separate even a single target cell in 1 ml whole blood. The present system is refined to capture certain circulating cells and / or particles, such as mesenchyme-derived cells and particles. The iChip is unique among other microfluidic methods for rare cell or particle isolation, because it yields the separated cells in a suspension amenable directly for subsequent imaging (hyperspectral fluorescent cell counting) or molecular analysis (ddPCR, scRNAseq). The viability and functionality of the separated cells have been tested extensively.

[0095] Other microfluidic methods rely on laminar flow of cells through antibody-coated microposts or microvortices generated by herringbone-shaped grooves to direct cells toward antibody-coated surfaces, where cells are immobilized and not readily available for imaging or single-cell molecular characterization. Other commercially available or experimental approaches to rare-cell separation, such as magnetically-activated-cell- sorting (MACS), Ficoll-Paque® or filtration techniques to separate WBCs from erythrocytes and other blood components, suffer from poor yield and purity and lack the rigorous and sensitive separation that is critical when the target cells are ultra-rare (e.g. 1- 100 / ml). The iChip overcomes these difficulties.

[0096] Other methods and systems of cell isolation can also be used, as long as the bulk of the RBCs, platelets, and WBCs are removed.

[0097] In another aspect, the disclosure provides methods of isolating and / or analyzing mesenchyme-derived cells or particles, such as MPCs, MSCs, and mesenchyme-derived particles, such as organelles or extracellular vesicles, e.g., microvesicles (MVs), exosomes, and apoptotic bodies, from a blood sample from a subject by using positive selection, rather than negative depletion. These methods include obtaining a blood sample from the subject; mixing the blood sample with magnetic beads comprising a binding agent that specifically binds to one or more specific types of cells or specific types of particles of interest, i.e., the mesenchyme-derived and HO-associated cells, and not to white blood cells (WBCs), for a time and under conditions sufficient for the binding agent to bind to the mesenchyme-derived cells or particles. As in the negative depletion methods described above, the blood sample is first through a first module comprising a microfluidic size-based separation system configured to direct small cells and particles such as red blood cells (RBCs) and platelets in the blood sample to a first waste outlet and to direct the remaining blood sample to a second module comprising an inertial focusing channel. The same systems described above can be used for the positive selection methods as well. Also as in the negative depletion methods, the remaining blood sample is flowed through the second module at a flow rate and for a distance sufficient to cause cells and / or particles in the remaining blood sample to align in one or more streamlines within the remaining blood sample flowing in the inertial focusing channel. Then the remaining blood sample with the cells and / or particles aligned in one or more streamlines is flowed through a third module comprising a magnetophoresis system for a time and distance sufficient to separate the specific types of cells or particles bound to magnetic beads from WBCs cells, other cells, and particles not bound to magnetic beads, and flowing the WBCs other cells, and particles not bound to magnetic beads into a second waste outlet and flowing the cells or particles bound to the magnetic beads to a product outlet. Again, the same systems described above for these steps can be used in the positive selection methods.

[0098] Next, the bound cells or particles are obtained from the product outlet and one determines which of the bound cells or particles originate in mesenchymal tissue. Then of those cells that are positively selected as originating from mesenchymal tissue are analyzed to determine which specific genes or proteins these positively selected cells express, using any known techniques. Then, in negative selection methods, one analyzes the unbound circulating cells or unbound particles, e.g., the mesenchyme-derived cells or particles, from the product outlet for an HO-associated molecular signature comprising one or more of gene expression, protein expression, or post-translational modification, to determine whether the unbound circulating cells or unbound particles are HO-associated. In positive selection methods, one analyzes the bound circulating cells or bound particles for an HO-associated molecular signature comprising one or more of gene expression, protein expression, or post-translational modification, to determine whether the bound circulating cells or bound particles are HO-associated. These analysis methods can be carried out using droplet digital PCR and RNA sequencing as described below.

[0099] Use of Droplet Digital PCR (ddPCR) to Quantify Mesenchyme-Derived Cells and Particles

[0100] Semi -quantitative RT-PCR analyses of ultra-rare circulating tumor cells have been inconsistent in part because of the relatively low sensitivity and specificity of RT-PCR when using whole blood. Indeed, ~1 target cell / million is below the detection limit of RT-PCR for a non-abundant transcript. Even a very low background transcription of highly tissue-specific transcripts by abundant blood cells becomes a confounder when target cells are present at such vanishingly low numbers. The susceptibility of qRT-PCR to the inhibitory effects of large amounts of non-specific template also adds to the large variability and inconsistencies in rare cell detection. To overcome these challenges, quantitative ddPCR is used after the initial enrichment of mesenchyme-derived cells and particles under RNA-preserving conditions as described in the methods above. ddPCR can be used to analyze mesenchyme-derived cells or particles, such as MPCs, isolated in the microfluidic systems described herein to determine which, if any, of the isolated cells or particles, are HO-associated. First, the cells, e.g., mesenchyme- derived cells, of which only a few may be HO-associated, are lysed and undergo WTA (Whole Transcriptome Amplification). Individual cells are encapsulated, e.g., using a system as described in, e.g., US Patent No. 9,068,181, which is incorporated herein by reference in its entirety.

[0101] Positive droplets are analyzed by ddPCR, which sequesters a small number of cDNA templates and PCR reaction reagents into aqueous droplets within an oil suspension, drastically increasing the effective concentration of the target transcript and allowing the differential expression of rare mesenchyme-derived and HO-associated genes to be leveraged. Partitioning the entire cDNA sample into these droplets followed by high-cycle PCR to amplify each template of interest maximally creates a digital readout of the number of positive droplets as a measure of the prevalence of each transcript of interest. By tabulating the total number of positive and negative droplets, and assuming the transcripts of interest follow a Poisson-distribution when partitioning into droplets, the absolute number of transcripts in the sample can be imputed. ddPCR can quantify multiple lineage-specific transcripts that are absent from background and hence denote the presence of mesenchyme-derived cells and particles that are also HO- associated. See, e.g., PCT WO 2016 / 154600, which is incorporated herein by reference in its entirety, for a description of ddPCR.

[0102] For example, a purified RNA sample and the reverse transcription reaction mix can be prepared and maintained on ice. The reaction mix can be composed of 5 pl Supermix® (ddPCR Supermix for Probes (No dUTP), Bio-Rad), 2 pl reverse transcriptase, 1 pl DTT, 1 pl each of the target primers / probes and 3 pl sample RNA. This mix then can be combined with the sample in the final step before being transferred to a ddPCR Cartridge (Bio-Rad), with 20 pl of the mix loaded into each sample well. Droplets are generated using, e.g., a Bio-Rad QC200 Droplet Generator. Following PCR reaction, data acquisition and analysis are conducted, e.g., using QX manager software version 3.0. The reads from probe assays measuring specific genes are summed, normalized to reflect the full RNA amount, and compared, e.g., using an unpaired t-test.

[0103] Methods of Detecting Subjects at Risk for HO and Guiding Treatment

[0104] The methods and systems described herein, including the use, in some embodiments, of a microfluidic iChip, can be used to isolate and analyze circulating mesenchymal progenitor cells (cMPCs) released into a subject’s peripheral blood. The methods can further include detecting a quantity of the HO-associated circulating cells or particles, a quality of the HO-associated circulating cells or particles, or both, for determining a specific level of risk of heterotopic ossification. For example, in these methods, a quantity of HO-associated circulating cells in the blood sample from the subject that is higher than a reference quantity determined from HO-associated circulating cells in blood samples from a plurality of healthy individuals indicates the subject is at risk for developing HO. Thus, a simple blood test can provide a sufficient amount of blood for analysis. RNA sequencing of cMPCs can be used to reveal the unique expression of HO-associated MPC genes observed as soon as 6 hours post HO-inducing injury, which is 41 days earlier than the only currently available diagnostic strategies, which involve x-ray exams.

[0105] A cMPC HO score can be used to evaluate the diagnostic ability of the liquid biopsy approach described herein for the early detection of HO. First, one selects the genes that will be the basis of cMPC HO score.

[0106] Gene Candidate Selection: Gene-level RNA quantification from Salmon files can be read into a count matrix, e.g., 60,609 genes and 21 samples, e.g., 10 of which are HO positive and 10 of which are HO negative. EdgeR and DESeq analysis pipelines can be applied to identify differentially expressed genes between HO positive and HO negative populations. For example, a database of 20,162 genes and their expression levels (TPM) in human PBMCs can be obtained from the Human Protein Atlas. 6,021 gene expressed at >5 TPM human PBMCs can be filtered from the differential expression analysis. For example, genes with log2(HO+ / HO-) > 2 can be considered to be differentially expressed.

[0107] A CPM of 13,262 genes from 1,049 single mouse MPCs are log-transformed, and mouse genes are mapped to their human orthologs using the R package orthogene. 11,285 genes (85%) where human orthologs existed are retained after mapping. An average CPM across all 1,049 cells is computed for each gene. MPC genes are then ranked in order of average single-cell expression.

[0108] HO signature gene candidates are then selected from the union of differential expression among clinical HO patients and controls and the highest ranked genes expressed in resident MPCs. These genes can be used to train a model using machine learning. cMPC HO Scoring Model Training: A matrix of log(CPM + 1) expression of the selected genes among a number of patients, e.g., 21 patient samples, is constructed, and a binary response column is added to indicate HO status as either a 1 or a 0. The data is partitioned into random training (70%) and validation (30%) sets. A linear Support Vector Machine (SVM) classifier can then be trained on these data using the svm function in the R package el 071 and validated using the predict function in the stats package. The receiver-operator characteristic (ROC) curve for this classifier was constructed and the area under the curve (AUC) computed.

[0109] In a mouse model described below in the Examples, this score yielded an overall sensitivity of 92% and an overall specificity of 91% for detecting HO. When applied to monitor HO-reducing prophylactic treatment with P-aminopropionitrile (BAPN), the new methods detect a significant reduction in the cMPC HO score, parallel to the eventual reduction in HO volume. Thus, the new methods and systems described herein can also be used for guiding and evaluating the efficacy of treatment for HO. These observations uncover a new paradigm for noninvasive early detection and screening of diseases that exhibit aberrant mesenchymal cell fate and repair that can enable high-throughput screening of high-risk patients as well as determine treatment efficacy.

[0110] In some embodiments, the methods further include treating a subject at risk for developing HO by administering an anti-HO treatment, e.g., the anti-HO treatment can be or include administering a non-steroidal anti-inflammatory drug, warfarin, an electromagnetic field, radiation therapy, or a bisphosphonate drug. While prophylactic strategies have been the focus of many clinical studies the only currently approved modalities are radiation therapy and non-steroidal antiinflammatory medications. However, these treatments are non-specific and have off- target effects including fracture non-union, tissue fibrosis, malignancy, and wound healing complications (7, 8) making the need for precision medicine (only treating those at risk) a high priority. Additionally, the timing of initiation and termination of these prophylactic strategies remains unknown, and it is necessary to optimize treatment duration. Thus, a large percentage of patients (up to 80%) who will not go on to form HO end up receiving these potentially harmful prophylactic interventions. In addition to those treated unnecessarily, many patients are also treated at the wrong time and for the incorrect duration, given a lack of treatment guiding technologies. Thus, a point of care decision support tool to predict which patients will develop HO and assess the impact of treatment efficacy will provide a more “precision medicine” approach.

[0111] In certain embodiments, when administering an anti-HO treatment, the methods can further include determining the efficacy of the anti-HO treatment, e.g., by taking a second blood sample at a first time, at or before starting the treatment, to determine a first level or number of HO-associated circulating cells or particles in the second blood sample of the subject; taking a third blood sample at a second time, later than the first time, to determine a second level or number of HO-associated circulating cells or particles in the third blood sample of the subject; and determining a level of efficacy of the treatment, wherein a decrease in a level or number of HO-associated circulating cells or particles at the second time compared to the level or number of HO-associated circulating cells or particles at the first time indicates that the treatment was effective.

[0112] In other embodiments, the efficacy of the anti-HO treatment can be determined by taking a second blood sample at a first time, at or before starting the treatment, to determine a first level or number of HO-associated genes in circulating cells or particles in the second blood sample of the subject; taking a third blood sample at a second time, later than the first time, to determine a second level or number of HO-associated genes in the third blood sample of the subject; and determining a level of efficacy of the treatment, wherein a decrease in a level or number of HO-associated genes in circulating cells or particles at the second time compared to the level or number of HO-associated genes at the first time indicates that the treatment was effective. Blood samples from people living with a rare genetic form of HO, fibrodysplasia ossificans progressiva (FOP) have also been shown to have increased circulatory osteogenic precursors during active episodes of HO formation when compared to stable disease (9). Indeed, previous studies detected HO-associated cMPCs following an HO- inducing burn / tenotomy (B / T) in our proven mouse model (10). Thus, the new methods and systems described herein can also be used to diagnose and monitor FOP patients.

[0113] EXAMPLES

[0114] The invention is further described in the following examples, which do not limit the scope of the invention described in the claims.

[0115] In experiments described below, transcripts of the following genes Ackr3, Adm5, Asprvl, Cadm3, Cdhll, Collal, Colla2, Col3al, Col5al, Crtacl, Dact3, Dpysl3, Fmod, (jfpt2, Gpx8, Igfbp5, 1133, Lama4, Matn2, Mmp2, MsantdS, Tmeffl, Mxra5, Pcolce2, Pcyoxl, Penk, Pgapl, Plod2, Postn, Prelp, Rbms2, Trim2, Vnn3, Pcdhl9, and Dpp4 were found to be highly specific for mesenchyme-derived cells that are HO-associated, and are generally absent in blood of healthy individuals who have not had a burn or other trauma that could lead to HO.

[0116] The following methods were used in the examples described below.

[0117] Mouse Injury Models

[0118] Gli lcreERTmice (JAX:007913), Adiponectin-cre mice (JAX:028020), Rosa-LSL- tdTomato mice (JAX:007909), and Collal(2.3kB)-GFP mice (JAX:013134) were purchased from Jackson laboratory. B / T injuries were performed as previously described10. Briefly, 8-10-week-old C57B6 mice (Jackson Laboratory) were provided buprenorphine sustained release for analgesia (dose) and isoflurane in oxygen (dose) for anesthesia. Mice were prepared for surgery by sharing the back and left hindlimb with clippers. Burn and tenotomy sites were cleaned by three alternating scrubs of ethanol and betadine. Anesthetic depth was assessed by hindlimb pinch prior to injury. The tenotomy was performed by making a small skin incision to expose the Achilles’ tenotomy, isolating the tendon using surgical scissors, and fully transecting the tendon at its midpoint. For B / S procedures, the tendon was isolated, but not transected. Skin was closed using absorbable 5-0 Vicryl® suture. Metal blocks were warmed to 60°C in a water bath and applied to the mouse back for 18 seconds. Mice were recovered from anesthesia in a cage warmed by a heating pad before returning to original cage. Mice were assessed for complications and pain daily for three days following injury. The surgeon was not blinded to surgical procedure.

[0119] BAPN Treatment

[0120] Following B / T or B / S, 8 mg / ml P-aminopropionitrile (BAPN) in 5% sucrose water, or 5% sucrose water control was provided to mice in bottles for ad libitum consumption. Bottles were shaken daily and replaced twice weekly. Mice were treated with BAPN to treat the induced HO until tissue harvest.

[0121] Microfluidic Enrichment

[0122] All experiments used iChips™ equipped with non-equilibrium inertial separation array (NISA)47. Blood samples were drawn in University of Texas Medical Center, into ACD tubes, transported overnight under normothermic conditions and processed within 24 hours of blood draw. Before enrichment, cells were incubated with antibodies against biotinylated CD45, CD16 and CD66b for a minimum of 30 minutes, and then with 1 pm streptavidin coated superparamagnetic beads for 20 minutes14, 48. Then, whole blood is passed through iChip and enriched product containing cMPCs is collected. We then performed a second step of enrichment by repeating reagent incubation and microfluidic enrichment steps.

[0123] Bulk RNA Sequencing

[0124] RNA was purified from lysed cells using Qiagen Micro RNeasy kit. We prepared libraries for bulk RNAseq® using the Plexwell® Rapid Library Preparation kit (Seqwell)49following manufacturer instructions. Sequencing was performed at MGH Nextgen Core Facility using Illumina NextSeq® 2000. Following debarcoding, the data was aligned and converted to reads files in Illumina Basespace® RNA-Seq Alignment App (v.2.0.2, Illumina, CA). Then, the data was normalized to counts per million (CPM), minimum reads threshold was set as 0.5 CPM, and differential expression analysis was performed by using edgeR50using DEApp51. To identify cMPC HO gene signature, genes increased in B / T with p<0.01 were filtered by removing genes expressed in PBMCs (>0.5nTPM in Human Protein Atlas52), is not detected in MPCs in HO anlagen (using dataset GEO: GSE126O6O10, and detected in healthy human donor microfluidics enriched blood samples. Statistical analyses were performed using Graphpad Prism 10 and JMP Version 16 (SAS Institute Inc.).

[0125] Single Cell RNA Sequencing iChip-enriched cells were captured and barcoded via 10X instrument at MGH Nextgen Core Facility (lOx Genomics Cell Ranger™ 7.1.0; via lOx Genomics Cloud Analysis) was used to perform sample de-multiplexing, barcode processing, and singlecell gene counting (alignment, barcoding, and unique molecular identifier [UMI] count)53. Reads were aligned to the mml0-2020-A reference genome as appropriate for the input dataset. Downstream analysis steps were performed using Seurat v554. Cells identified as from poor libraries or resulting from doublet cells were filtered by eliminating cells on both ends of the distribution, as well as cells with unusually high percent mitochondrial transcripts.

[0126] Normalization, scaling, and dimensional reductions using principal component analysis (PCA) and uniform manifold approximation and projection (UMAP) were performed using Seurat v5. Integrative analysis was performed using Harmony™ (55). Cell type labels were assigned based on characteristic relative marker gene expression levels between cell clusters using the FindAllMarkers function. Module scores were generated using the AddModuleScore function of Seurat to identify cell populations of interest. Module scores were calculated as the level of gene expression enrichment of a set gene list relative to a random control list, with higher module score values representing positive enrichment beyond background. Dimensional feature plots, violin plots, and dot plots were used to visualize gene expression within the population of interest.

[0127] Flow Cytometry iChip-isolated cells were labeled with a cocktail of antibodies in 0.5% BSA 0.1% F68-PBS at 4C. The antibody cocktail consisted of PDGFRa-SB600 (APA5, Invitrogen), CD90-eFluor450 (Thy-1.1, Invitrogen), CD45-BUV395 (30-F11, BD), Teri 19-BUV395 (Ter- 119, BD). Then, cells were washed and suspended in PBS 0.5% BSA before being analyzed on a FACS Aria® Fusion (BD Biosciences). For compensation, all antibodies were conjugated to OneComp eBeads® (Thermo Fisher Scientific). Unstained control cells and compensation beads were used for performing compensation. We used PDGFRa+CD90+ MPCs isolated from bone tissue using previously established protocols56, 57and PBMCs for determining the gating strategy. For transgenic mouse experiments, we also used G / z7Cre£7?r2-tdTomato+ control cells isolated from periosteum as compensation controls and for determining gates.

[0128] Microscopy

[0129] Freshly dissected tissues were fixed overnight in 4% paraformaldehyde in PBS at 4°C, then washed several times in PBS. Tissues were then decalcified in 10% EDTA (w / v) in PBS for 4 weeks, then paraffin embedded. 5 pm thick tissue sections were cut and adhered to glass slides. Sections were then deparaffinized and stained using the following procedure.

[0130] Sections were permeabilized for 10 minutes with 0.05% NP40 and 2.5% DMSO in hank’s balanced salt solution (HBSS), then washed in PBS and blocked in PBS with 5% normal donkey serum (Jackson Immunoresearch) for 1 hour. Slides were then stained overnight with combinations of the following antibodies: chicken anti-GFP (1-200, Aves Labs, GFP-1020), goat anti-dsRed (1-200, LifeSpan Biosciences, LS-C340696), rabbit anti-Aggrecan (1-35, EMD Millipore, AB1031). Secondary antibodies included donkey anti-chicken Alexa Fluor 488 (1-250, Jackson Immunoresearch), donkey anti-rabbit Alexa Fluor 647 (1-250, Jackson Immunoresearch), and donkey anti -goat CF-555 (1-500, Biotium). Sections were counterstained with DAPI to stain nuclei before mounting. Slides were mounted with Prolong Gold anti-fade reagent (Invitrogen). Images were acquired with a Zeiss LSM880 confocal microscope. Confocal images were processed and analyzed using Fiji (Image J) and Photoshop (Adobe Systems).

[0131] Example 1 - cMPCs for Use as Early Markers of Trauma-Induced HO

[0132] In this example, we set out to examine if cMPCs can be used as an early marker of trauma induced HO. First, we collected and analyzed blood samples from mice after trauma with and without HO to identify if there was a difference in cMPC number. Next, we set out to identify an HO-specific cMPC-based molecular signature which we used to create a scoring algorithm achieving a remarkable 100% specificity and 82% sensitivity in identifying early-stage H0+ samples. These findings provide compelling evidence supporting the feasibility of a liquid biopsy approach in early detection and surveillance of traumatic HO.

[0133] NISA-iChip Captures Mouse MPCs in Whole Blood

[0134] We tested how well the NISA-iChip captures mouse MPCs in whole blood. For this, we prepared cell suspensions from metaphysis / diaphysis regions of long bones of wild-type mice, rich in MPCs. We then FAC-sorted 1,000 CD45 PDGFRa+MPCs.

[0135] Next, we obtained whole blood (~1 ml, IO10cells) from a wild-type mouse, then spiked the blood and the same volume buffer with the sorted MPCs. We ran the spiked blood sample through NISA-iChip, to obtain the product.

[0136] Lastly, we performed a flow cytometry analysis of the fluorescent MPCs in the enriched product and compared to the spiked buffer. As shown in FIG. 6, we found that an average of 96±20% of all MPCs spiked into the whole blood were captured in the product (n=15). In addition, we counted recalcitrant cell populations in enriched blood from naive mice in comparison to the initial leukocyte numbers from blood counts to calculate the efficiency in depleting leukocytes. By manually counting the DAPI+ nucleated cell population, and dividing this number by the initial leukocyte count, we calculated a leukocyte depletion percentage of 99.97%, equivalent to 3.57+ / -0.06 log depletion (n=8). These data show that the current version of NISA-iChip captures mouse MPCs from whole blood samples with little loss of cMPCs.

[0137] Non-HO Inducing Musculoskeletal Injury Does Not Lead to cMPC Release

[0138] To test if cMPC release is specific to injury that leads to the heterotopic bone formation, we obtained blood samples from mice that went through other modes of injury: 1) a severe closed head concussive injury via weight drop that mimics human traumatic brain injury26(TBI) and 2) muscle fibrosis induced by intramuscular injection of cardiotoxin27. We have not observed cMPC shedding in these non-HO inducing models supporting the specificity of cMPCs for HO (See FIGs. 7A-7B).

[0139] In this part of the example, we tested whether MPC release in the blood is unique to HO. As shown in FIG. 7A, we utilized MPC-Chip equipped with non-equilibrium inertial separation array (NISA) to collect cMPC in the burn and tenotomy (B / T) mouse model. B / T injuries were performed on 8-10-week-old C57B6 mice (Jackson Laboratory) were provided buprenorphine sustained release for analgesia (dose) and isoflurane in oxygen (dose) for anesthesia. Mice were prepared for surgery by sharing the back and left hindlimb with clippers. Burn and tenotomy sites were cleaned by three alternating scrubs of ethanol and betadine. Anesthetic depth was assessed by hindlimb pinch prior to injury. The tenotomy was performed by making a small skin incision to expose the Achilles’ tenotomy, isolating the tendon using surgical scissors, and fully transecting the tendon at its midpoint.

[0140] For burn / sham injury (B) procedures, the tendon was isolated, but not transected. Skin was closed using absorbable 5-0 Vicryl suture. Metal blocks were warmed to 60 degrees C in a water bath and applied to the mouse back for 18 seconds. Mice were recovered from anesthesia in a cage warmed by a heating pad before returning to original cage. Mice were assessed for complications and pain daily for three days following injury. The surgeon was not blinded to surgical procedure. Blood samples were drawn into Acid-Citrate-Dextrose (ACD) tubes, transported overnight under normothermic conditions and processed within 24 hours of blood draw.

[0141] Before enrichment, cells were incubated with antibodies against biotinylated CD45, CD16 and CD66b for a minimum of 30 minutes, and then with 1 pm streptavidin coated superparamagnetic beads for 20 minutes. Then, whole blood is passed through MPC-Chip and enriched product containing cMPCs is collected, labeled with fluorochrome conjugated antibodies against PDGFRA, CD90, CD45, and analyzed using flow cytometry.

[0142] FIG. 7B shows flow cytometry quantification of cMPCs 24 hours post-B / T, when compared to naive, sham or non-HO inducing traumatic conditions, of microfluidics- enriched blood identified significantly increased CD45-CD1 lb-PDGFRa+CD90+ cMPCs. This data validates that PDGFRa / CD90 MPCs released into circulation specifically in B / T (HO forming injury), but not in healthy or other non-HO forming traumatic injuries (n=5 mice per condition, *p<0.05 by 1-way-ANOVA).

[0143] HO-Inducing Injury Leads to Early Release of PDGFRa+cMPCs

[0144] To investigate if peripheral blood contains cMPCs that can be used as biomarkers to enable early detection of HO, we performed either 1) a combination of 30% total body surface area (TBSA) burn injury followed by transection at the midpoint dorsal hind limb tendon (Achilles tenotomy) which results in HO or 2) a combination of 30% TBSA bum and a sham skin incision (B / S) which does not cause HO according to protocols previously published by our collaborative team, and collected blood samples at days 1, 3 or 7 after injury10. FIG. 8A illustrates the general approach for this mouse model method.

[0145] In particular, we performed the B / T HO model in transgenic GUI- and Adipoq- lineage tracing mice, isolated cMPCs as described herein, and quantified tdTomato cells in enriched blood samples, and performed flow cytometry as described above with respect to FIGs. 7A-7C, except here we also detected the tdTomato fluorophore.

[0146] The injuries were performed utilizing normothermic stabilization of whole blood for cMPC isolation using our established protocols20. We then isolated live cMPC populations using iChip and quantified CD45- cells expressing PDGFRa and CD90, a marker combination containing MPCs21, using flow cytometry.

[0147] In day 1 post-B / S control mice, we detected a mean of 4 PDGFRa+ / CD90+per 1 mL blood sample (data not shown, HO-associated MPCs; mean of n=5 mice), whereas day 1 post-B / T MPC numbers increased precipitously to 16 PDGFRa+ / CD90+(n=9) and 30 PDGFRa+ / CD90+at day 3-post B / T (n=9). This dramatic increase completely reversed by day 7 post-B / T, as the numbers of MPCs decreased to 3 PDGFRa+ / CD90+(n=5).

[0148] Next, we asked whether the cells identified in the circulation were derived from the HO site. To answer this question, we used two lineage tracing systems previously validated by our team(24). Specifically, we have shown that G / z'7-lineage cells mark bone progenitor cells in the periosteum whereas 4t / z / / -lineage cells mark bone progenitor cells in the bone marrow. To query if the cMPCs we isolated were from the periosteum or bone marrow, we activated the Cre system with tamoxifen using our previously validated models followed by a burn tenotomy. We then collected blood and isolated cMPCs as in the approach shown in FIG. 8A. Interestingly, we were able to detect GlilCreERT2cells that were also PDGFRa and CD90 positive, whereas we could not detect AdipoqCrecells indicating that the circulating cells most likely derived from Glil-CreER lineage cells (data not shown).

[0149] FIG. 8B shows two graphs in which the left graph shows the total number of tdTomato+ cells observed, and the right graph shows the percent of tdTomato labeled cells within PDGFRa / CD90 cMPC gate. We were able to detect G / z7CreERT2cells that were also PDGFRa and CD90 positive, whereas we could not detect AdipoqCKERcells, indicating that the circulating cells most likely derived from Glil-CreER lineage cells.

[0150] Next, we wanted to confirm that the GUI lineage cells marked the same cells at the HO site. Thus, we next employed our GlilCreERT2<X om Collal(2.3)GFP mice to lineage trace the GlilCreERT2cells to validate that they are the bone forming Collal cells. Indeed, prior to injury Gli lCreERTETdTom cells were located in the periosteum and peritenon, and co-stained for Collal (2.3)GFP. Next, we performed our burn / tenotomy model in these mice and analyzed the HO site 12 weeks after injury at which point mature heterotopic bone was noted to stain for both Gli lCreERT2;TdTom and Collal(2.3)GFP. This is consistent with previous studies that identified GlilCreERT2;TdTom cells to mark heterotopic bone25.

[0151] FIG. 8C shows the immunofluorescent histology of uninjured mouse hindlimb (first image) and 12 weeks post-injury hindlimb (three right side images) in GlilCreERT2:ROSA-LSL-TdTomato with chevrons marking double-positive cells (scale bars = 100 pm). When we quantified Collal(2.3)-GFP+ osteoblasts and Aggrecan+ chondrocytes, we found that 24.5%+ / -3.5 of chondrocytes and 3.4%+ / -12.6 of osteoblasts within the GlilCreERT2;TdTom+ population in HO anlagen (data not shown). Thus, for the first time, we have demonstrated that rare circulating MPCs are present in the blood after an HO-inducing injury and that these derive from the periosteum / peritenon.

[0152] Single Cell RNAseq Detects cMPCs Expressing HO Gene Expression Signature in B / T Blood Samples

[0153] MPC-Chip enriched cells were captured from blood samples isolated from B / T mice (HO+), and barcoded via 10X instrument at MGH Nextgen Core Facility. lOx Genomics Cell Ranger® 7.1.0 (via lOx Genomics Cloud Analysis) was used to perform sample de-multiplexing, barcode processing, and single-cell gene counting (alignment, barcoding, and unique molecular identifier [UMI] count).

[0154] Reads were aligned to the mml0-2020-A reference genome as appropriate for the input dataset. Downstream analysis steps were performed using Seurat v5. Cells identified as poor libraries or resulting from doublet cells were filtered by eliminating cells on both ends of the distribution, as well as cells with unusually high percent mitochondrial transcripts. Normalization, scaling, and dimensional reductions using principal component analysis (PC A) and uniform manifold approximation and projection (UMAP) were performed using Seurat v5.

[0155] Integrative analysis was performed using Harmony®. Cell type labels were assigned based on characteristic relative marker gene expression levels between cell clusters using the Find AllMarkers function. Module scores were generated using the AddModule Score function of Seurat to identify cell populations of interest. Module scores were calculated as the level of gene expression enrichment of a set gene list relative to a random control list, with higher module score values representing positive enrichment beyond background.

[0156] FIG. 9 shows the UMAP representation of cells detected in microfluidics-enriched sample, in which MPCs are circled in the upper right quadrant. The cell types and their total numbers of each type of cell are listed as well in the chart on the right side of the figure. This scRNAseq dataset confirms the finding of MPCs expressing the genes expected in microfluidics enriched blood samples on a single cell level and that these genes are not coming from other cell types. cMPCs Post-B / T Express Distinct HO-Associated Transcripts

[0157] To determine the molecular profile of cMPCs, we next performed RNAseq on iChip-enriched blood from mice post-B / T and B / S and cataloged differentially expressed cMPC genes using the approach shown schematically in FIG. 10 A, which is an overview of the approach to identify cMPC-based markers for early detection of HO. We performed microfluidic sorting of cMPCs from whole blood samples taken from model mice (bum / tenotomy B / T model or burn / sham incision as HO- control, blood volume ~1 ml analyzed) or patients undergoing joint replacement surgery (~5mL). Then, RNA was purified from lysed cells using Qiagen Micro RNeasy kit. We prepared libraries for bulk RNAseq using the Plexwell® Rapid Library Preparation kit (Seqwell) following manufacturer instructions. Sequencing was performed at MGH Nextgen Core Facility using Illumina NextSeq® 2000. Following debarcoding, the data was aligned and converted to reads files in Illumina Basespace® RNA-Seq Alignment App (v.2.0.2, Illumina, CA).

[0158] Then, the data was normalized to counts per million (CPM), minimum reads threshold was set as 0.5 CPM, and differential expression analysis was performed by using EdgeR. To identify cMPC HO gene signature, genes increased in B / T with p<0.01 were filtered by removing genes expressed in PBMCs (>0.5nTPM in Human Protein Atlas), and by removing genes not detected in mouse MPCs, and in MPCs in HO anlagen (using dataset GEO: GSE 126060), and detected in healthy human donor microfluidics enriched blood samples. The resulting genes were then used as the “cMPC HO signature” to test predicting HO formation early. This list of genes is shown in FIG. 10B.

[0159] Statistical analyses were performed using Graphpad® Prism 10 and JMP Version 16 (SAS Institute Inc.). We then removed genes expressed by leukocytes (TPM<5) and genes not expressed by mesenchymal progenitor cells (GEO: GSE126060).

[0160] We used the same approach to isolate and analyze cMPC HO gene signatures from blood samples taken from human knee or hip arthroplasty patients.

[0161] FIG. 10B shows expression levels of genes selected based on the filtering shown in FIG. 10A as a cMPC HO signature in joint replacement surgery patients that end up developing HO (HO+) when compared with patients that do not develop HO (HO-) based on follow up radiographic imaging analysis. The figure shows log(counts per million+1) normalized data collected from blood samples taken at post-operative day 1 (POD1). The results show sensitivity and specificity of genes across the list. These genes were combined to train a model to predict HO from blood samples.

[0162] FIG. 10C shows expression levels of cMPC HO signature genes in B / T tHO mouse model (n=5 for B / S, B / T DI and n=4 for B / S D3, B / T D7, n=7 for B / T D3) described above. The results validate the findings of these genes at POD1 in patients that end up developing HO. The expression levels of these genes change across different time points suggesting they may activate or be expressed in cMPCs only during a certain time post-surgery / trauma.

[0163] FIGs. 11 A-l IF show the expression levels of selected cMPC HO signature genes, Asprvl, Gjpt2, Gpx8, Matn2, Penk, and Prelp, in joint replacement surgery patients to highlight the detected expression levels across different time points. The graphs in these figures show log(counts per million+1) normalized data collected from blood samples taken before or during joint replacement surgery (pre-op, intra-op) or post-operative day 1 (POD1) or POD 14, using the methods described above as illustrated in FIG. 10 A.

[0164] The data shows that these genes are highly specific for HO when detected POD1. In addition, some patients that have a history of HO may show expression of these genes before operation (pre-op). We also detected these genes in blood samples taken during operation (intra-op), but this time point did not provide specificity for the majority of the markers. Also, we did not detect these genes from majority of blood samples taken at POD 14 time point. cMPC HO Risk Score Performance

[0165] Based on the expression levels in HO-negative mouse and samples (shown in FIGs. 10A-10C and 11 A-l IF described in the examples above), we determined a preliminary threshold for a positive score (patient at risk for HO) of about 1 log (CPM+1). FIG. 12A shows a bar graph of results for different experiments, B / S or B / T on days 1, 3, or 7, and the dotted line at about 1 log (CPM+1) is the determined threshold. As shown in the figure, all of the sham tests were below this line (except one data point on Day 3), and most of the induced HO results were above this line (except one data point on Day 1).

[0166] This score achieved 92% overall specificity and 91% overall sensitivity for this dataset (n=13 B / T mice, n=12 B / S mice).

[0167] The graph in FIG. 12B shows a ROC curve for cMPC HO score for B / S vs B / T conditions (Receiver Operating Characteristic (ROC) curve AUC = 0.9551+ / -0.0386, p=0.0001). The graph in FIG. 12C shows the results of the use of machine learning to train a cMPC HO predictive model for joint replacement patients. Showing average predictive performance, when model is trained and tested on a randomly selected training set (70% of all data) and tested on the validation dataset (30%) 100 times using 3 distinct algorithms (SVM: support vector machine, RF: random forest, LR: linear regression). SVM performed best with ROC AUC > 0.85.

[0168] These results show a HO scoring model with strong diagnostic power can be trained using a sensitive and specific cMPC gene list or molecular signature.

[0169] Example 2 - cMPCs for Use as Markers of Therapeutic Efficacy of Treatment for Trauma-Induced HO

[0170] To see if the technology described herein can be used to detect therapeutic efficacy of treatments for HO, we treated mice after trauma with a proven inhibitor of HO. Measuring the molecular phenotype and concentration of cMPCs - a “liquid biopsy” - can serve as a sensitive, minimally invasive diagnostic marker, or “footprint” of HO that can enable screening high risk patients as well as to evaluate therapeutic efficacy to more precisely guide early treatment. Just as liquid biopsies have emerged as critical tools to diagnose and monitor certain malignancies (11-19), we believe this technology holds great promise for musculoskeletal conditions that exhibit aberrant mesenchymal cell fate and repair such as HO. cMPCs HO-Associated Signature Decreases Significantly Due to HO-Reducing Prophylactic Treatment

[0171] We tested the feasibility of detecting a therapy response using cMPC-based liquid biopsy (FIGs. 13A-13D). This liquid biopsy approach can be used for monitoring patients during prophylactic treatments to assist in predicting early therapeutic response and guide the therapeutic window for each patient allowing for more precise treatment timing.

[0172] To do this, we applied our cMPC HO risk score to mice after B / T with or without lysyl oxidase inhibition via P-aminopropionitrile (BAPN) treatment, which we have shown to significantly decrease HO formation23. Mice were treated until tissue harvest which was on POD 3. FIG. 13A shows a micro-CT 3-dimensional reconstruction of mice after burn / tenotomy treated with BAPN or vehicle control. FIG. 13B is a pair of bar graphs that show the quantification of HO volume in mice treated with BAPN vs. vehicle control for total HO (left graph) and “floating” HO, meaning not connected to any other bone (right graph).

[0173] FIG. 13C shows a bar graph of cMPC HO scores obtained using the mean of expression levels of genes described above with respect to FIGs. 10B and IOC as well as FIGs. 11 A-l IF. The HO scores decrease upon HO reducing treatment with BAPN (n=5 BAPN treatment post-B / T, n=4 B / T (vehicle treatment), n=7 B / S (no HO)).

[0174] We devised a set of cMPC markers that accurately detects the reduction in HO due to treatment at an early time point using the data described in FIGs. 10A-10C. FIG. 13D shows the log(CPM+l) of these specific genes (Fmod, Colla2, Plod2, 1133, Cdh3 a.vA Ackr3) and their average for B / T animals with and without BAPN treatment (P0D3, *p<0.05). These data showed that mice treated with BAPN had a statistically significant expression in these key six transcripts in BAPN treated mice compared to control (FIG. 13D). The mean HO monitoring score shown in the right graph demonstrated a significant decrease in HO in the BAPN treated mice.

[0175] We further validated this finding by performing digital droplet PCR quantification of expression of cMPC HO signature genes. We directly used RNA extracted from enriched cMPCs to quantify the absolute number of Fmod, Colla2. Plod2. 1133 and Ackr3 transcripts by performing multiplexed ddPCR.

[0176] We found that the average number of HO-associated cMPC transcripts decreased 32.80 ± 28.60 following BAPN treatment (p=0.28, CI of 98.76 to -33.16) (data not shown). In summary, we have shown that not only can the cMPC risk score predict HO formation after injury compared to a non-HO injury, but this risk score can also accurately detect therapeutic efficacy.

[0177] REFERENCES

[0178] 1 / 48. Karabacak et al., Microfluidic, marker-free isolation of circulating tumor cells from blood samples. Nat Protoc, 2014. 9(3): p. 694-710.

[0179] 2. Blann et al., Circulating endothelial cells. Biomarker of vascular disease. Thromb Haemost, 2005. 93(2): p. 228-35.

[0180] 3. Yu et al., RNA sequencing of pancreatic circulating tumour cells implicates WNT signalling in metastasis. Nature, 2012. 487(7408): p. 510-3.

[0181] 4 / 19. Ozkumur et al., Inertial focusing for tumor antigen-dependent and - independent sorting of rare circulating tumor cells. Sci Transl Med, 2013. 5(179): p. 179ra47.

[0182] 5 / 12. Ting et al., Single-cell RNA sequencing identifies extracellular matrix gene expression by pancreatic circulating tumor cells. Cell Rep., 2014. 8(6): p. 1905-1918.

[0183] 6. Yu et al., Ex vivo culture of circulating breast tumor cells for individualized testing of drug susceptibility. Science, 2014. 345(6193): p. 216-220.

[0184] 7 / 13. Miyamoto et al., RNA-Seq of single prostate CTCs implicates noncanonical Wnt signaling in antiandrogen resistance. Science, 2015. 349(6254): p. 1351-6.

[0185] 8 / 15. Kalinich et al., An RNA-based signature enables high specificity detection of circulating tumor cells in hepatocellular carcinoma. Proc Natl Acad Sci U S A, 2017. 114(5): p. 1123-1128.

[0186] 9 / 16. Hong et al., Molecular signatures of circulating melanoma cells for monitoring early response to immune checkpoint therapy. Proc Natl Acad Sci U S A, 2018. 115(10): p. 2467-2472.

[0187] 10 / 18. Miyamoto et al., An RNA-Based Digital Circulating Tumor Cell Signature Is Predictive of Drug Response and Early Dissemination in Prostate Cancer. Cancer Discov, 2018. 8(3): p. 288-303.

[0188] 11. Mutlu et al., Oscillatory inertial focusing in infinite microchannels. Proc Natl Acad Sci U S A, 2018. 115(30): p. 7682-7687.

[0189] 12 / 47. Mishra et al., Ultrahigh-throughput magnetic sorting of large blood volumes for epitopeagnostic isolation of circulating tumor cells. Proc Natl Acad Sci U S A, 2020. 117(29): p. 16839-16847. 14. Stott et al., Isolation and characterization of circulating tumor cells from patients with localized and metastatic prostate cancer. Sci Transl Med, 2010. 2(25): p. 25ra23.

[0190] 15. Maheswaran et al., Detection of mutations in EGFR in circulating lung-cancer cells. N Engl J Med, 2008. 359(4): p. 366-77.

[0191] 16 / 11. Sullivan et al., Brain tumor cells in circulation are enriched for mesenchymal gene expression. Cancer Di scov, 2014. 4(11): p. 1299-309.

[0192] 17. Sari oglu et al., A microfluidic device for label-free, physical capture of circulating tumor cell clusters. Nat Methods, 2015. 12(7): p. 685-91.

[0193] 18. Stott et al., Isolation of circulating tumor cells using a microvortex-generating herringbone chip, Proc Natl Acad Sci U S A, 2010. 107(43): p. 18392-7.

[0194] 19. Nagrath et al., Isolation of rare circulating tumour cells in cancer patients by microchip technology. Nature, 2007. 450(7173): p. 1235-9.

[0195] 20 / 14. Fachin et al., Monolithic Chip for High-throughput Blood Cell Depletion to Sort Rare Circulating Tumor Cells. Sci Rep, 2017. 7(1): p. 10936.

[0196] 1. Zeckey et al., Risk of symptomatic heterotopic ossification following plate osteosynthesis in multiple trauma patients: an analysis in a level- 1 trauma centre. Scand J Trauma Resusc Emerg Med. 2009;17:55-. doi: 10.1186 / 1757-7241-17-55. PubMed PMID: 19825174.

[0197] 2. Mital et al., Ectopic bone formation in children and adolescents with head injuries: its management. Journal of pediatric orthopedics. 1987;7(l):83-90. Epub 1987 / 01 / 01. doi: 10.1097 / 01241398-198701000-00017. PubMed PMID: 3098781.

[0198] 3. Levi et al., Risk factors for the development of heterotopic ossification in seriously burned adults: A National Institute on Disability, Independent Living and Rehabilitation Research bum model system database analysis. J Trauma Acute Care Surg. 2015;79(5):870-6. doi: 10.1097 / TA.0000000000000838. PubMed PMID: 26496115.

[0199] 4. Nauth et al., Heterotopic ossification in orthopaedic trauma. J Orthop Trauma. 2012;26(12):684-8. doi: 10.1097 / BOT.0b013e3182724624. PubMed PMID: 23010648.

[0200] 5. Dey et al., The traumatic bone: trauma-induced heterotopic ossification. Transl Res. 2017;186:95-111. Epub 2017 / 06 / 15. doi: 10.1016 / j.trsl.2017.06.004. PubMed PMID: 28668522. 6. Kluger et al., Ossification in Childhood and Adolescence. Journal of Child Neurology. 2016;15(6):406-13.

[0201] 7. Merkely et al., Do Nonsteroidal Anti-Inflammatory Drugs Have a Deleterious Effect on Cartilage Repair? A Systematic Review. Cartilage. 2021;13(l_suppl):326s-41s. Epub 20190619. doi: 10.1177 / 1947603519855770. PubMed PMID: 31216865; PMCID: PMC8808836.

[0202] 8. Oh et al., Do Selective COX-2 Inhibitors Affect Pain Control and Healing After Arthroscopic Rotator Cuff Repair? A Preliminary Study. Am J Sports Med.

[0203] 2018;46(3):679-86. Epub 20171218. doi: 10.1177 / 0363546517744219. PubMed PMID: 29253346.

[0204] 9. Suda et al., Circulating osteogenic precursor cells in heterotopic bone formation. Stem Cells. 2009;27(9):2209-19. doi: 10.1002 / stem. l50. PubMed PMID: 19522009; PMCID: PMC3496263.

[0205] 10. Sorkin et al., Regulation of heterotopic ossification by monocytes in a mouse model of aberrant wound healing. Nat Commun. 2020;l 1(1):722. Epub 20200205. doi: 10.1038 / s41467-019-14172-4. PubMed PMID: 32024825; PMCID: PMC7002453.

[0206] 17. Kwan et al., A Digital RNA Signature of Circulating Tumor Cells Predicting Early Therapeutic Response in Localized and Metastatic Breast Cancer. Cancer Discov. 2018;8(10): 1286-99. Epub 2018 / 08 / 15. doi: 10.1158 / 2159-8290.Cd-18-0432. PubMed PMID: 30104333; PMCID: PMC6170694.

[0207] 20. Wong et al., Whole blood stabilization for the microfluidic isolation and molecular characterization of circulating tumor cells. Nat Commun. 2017;8(l): 1733.

[0208] Epub 20171123. doi: 10.1038 / s41467-017-01705-y. PubMed PMID: 29170510; PMCID: PMC5700979.

[0209] 21. Agarwal et al., Surgical Excision of Heterotopic Ossification Leads to Re- Emergence of Mesenchymal Stem Cell Populations Responsible for Recurrence. STEM CELLS Translational Medicine. 2017;6(3):799-806.

[0210] 22. Pagani et al., Novel Lineage-Tracing System to Identify Site-Specific Ectopic Bone Precursor Cells. Stem Cell Reports. 2021;16(3):626-40. Epub 20210218. doi: 10.1016 / j.stemcr.2021.01.011. PubMed PMID: 33606989; PMCID: PMC7940250. 23. Kang et al., The HIF-la / PL0D2 axis integrates extracellular matrix organization and cell metabolism leading to aberrant musculoskeletal repair. Bone Research.

[0211] 2024; 12(1): 17. doi: 10.1038 / s41413-024-00320-0.

[0212] 24. Jeffery et al., Bone marrow and periosteal skeletal stem / progenitor cells make distinct contributions to bone maintenance and repair. Cell Stem Cell. 2022;29(l 1): 1547- 61.e6. Epub 20221021. doi: 10.1016 / j.stem.2022.10.002. PubMed PMID: 36272401.

[0213] 25. Kan et al., Glil-labeled adult mesenchymal stem / progenitor cells and hedgehog signaling contribute to endochondral heterotopic ossification. Bone. 2018;109:71-9. Epub 20170621. doi: 10.1016 / j.bone.2017.06.014. PubMed PMID: 28645539; PMCID: PMC5801258.

[0214] 26. Khuman et al., Tumor necrosis factor alpha and Fas receptor contribute to cognitive deficits independent of cell death after concussive traumatic brain injury in mice. J Cereb Blood Flow Metab. 2011;3 l(2):778-89. Epub 20101013. doi: 10.1038 / jcbfm.2010.172. PubMed PMID: 20940727; PMCID: PMC3049532.

[0215] 27. Guardiola et al., Acute Skeletal Muscle Regeneration by Cardiotoxin Injection. J Vis Exp. 2017(119). Epub 20170101. doi: 10.3791 / 54515. PubMed PMID: 28117768; PMCID: PMC5407614.

[0216] 28. Perosky et al., Early detection of heterotopic ossification using near-infrared optical imaging reveals dynamic turnover and progression of mineralization following Achilles tenotomy and burn injury. J Orthop Res. 2014;32(l 1): 1416-23. Epub 20140802. doi: 10.1002 / jor.22697. PubMed PMID: 25087685; PMCID: PMC4408934.

[0217] 29. Peterson et al., Early detection of bum induced heterotopic ossification using transcutaneous Raman spectroscopy. Bone. 2013;54(l):28-34. Epub 2013 / 01 / 15. doi: 10.1016 / j.bone.2013.01.002. PubMed PMID: 23314070; PMCID: PMC3690323.

[0218] 30. Ranganathan et al., High-frequency spectral ultrasound imaging (SUSI) visualizes early post-traumatic heterotopic ossification (HO) in a mouse model. Bone. 2018. doi: 10.1016 / j.bone.2018.01.034. PubMed PMID: 29412179.

[0219] 31. Brownley et al., Characterization of Heterotopic Ossification Using Radiographic Imaging: Evidence for a Paradigm Shift. PloS one. 2015; 10(1 l):e0141432. doi:

[0220] 10.1371 / journal. pone.0141432. PubMed PMID: 26544555; PMCID: 4636348. 32. Citak et al., The roles of serum alkaline and bone alkaline phosphatase levels in predicting heterotopic ossification following spinal cord injury. Spinal Cord.

[0221] 2016;54(5):368-70. doi: 10.1038 / sc.2015.211. PubMed PMID: 26643987.

[0222] 33. Lin et al, Easy interpretation of heterotopic ossification demonstrated on bone SPECT / CT. Clin Nucl Med. 2014;39(l):62-3. doi: 10.1097 / RLU.0b013e318286824f. PubMed PMID: 24217541.

[0223] 34. Eekhoff et al., [18F]NaF PET / CT scan as an early marker of heterotopic ossification in fibrodysplasia ossificans progressiva. Bone. 2018;109: 143-6. Epub 20170818. doi: 10.1016 / j.bone.2017.08.012. PubMed PMID: 28826841.

[0224] 35. Ghanem et al., The added value of SPECT-CT in the detection of heterotopic ossification on bone scintigraphy. Skeletal Radiol. 2020;49(2):291-8. Epub 20190730. doi: 10.1007 / s00256-019-03286-x. PubMed PMID: 31363823.

[0225] 36. Gori et al., Fibromodulin is expressed by both chondrocytes and osteoblasts during fetal bone development. J Cell Biochem. 2001;82(l):46-57. doi: 10.1002 / jcb.1115. PubMed PMID: 11400162.

[0226] 37. Murphy et al., Distribution of cartilage molecules in the developing mouse joint. Matrix Biol. 1999;18(5):487-97. doi: 10.1016 / s0945-053x(99)00042-6. PubMed PMID: 10601736.

[0227] 38. Saamanen et al., Murine fibromodulin: cDNA and genomic structure, and age- related expression and distribution in the knee joint. Biochem J. 2001;355(Pt 3): 577-85. doi: 10.1042 / bj3550577. PubMed PMID: 11311118; PMCID: PMC1221771.

[0228] 39. Shu et al., Catabolism of Fibromodulin in Developmental Rudiment and Pathologic Articular Cartilage Demonstrates Novel Roles for MMP-13 and ADAMTS-4 in C-terminal Processing of SLRPs. Int J Mol Sci. 2019;20(3). Epub 20190129. doi: 10.3390 / ijms20030579. PubMed PMID: 30700002; PMCID: PMC6386837.

[0229] 40. Elmansi et al., DPP4-Truncated CXCL12 Alters CXCR4 / ACKR3 Signaling, Osteogenic Cell Differentiation, Migration, and Senescence. ACS Pharmacol Transl Sci. 2023;6(l):22-39. Epub 20221213. doi: 10.1021 / acsptsci.2c00040. PubMed PMID: 36659961; PMCID: PMC9844133.

[0230] 41. Jamal et al., Derivation and characterization of putative craniofacial mesenchymal progenitor cells from human induced pluripotent stem cells. Stem Cell Research.

[0231] 2018;33: 100-9. 42. Karlsson et al. A single-cell type transcriptomics map of human tissues. Sci Adv. 2021;7(31). Epub 20210728. doi: 10.1126 / sciadv.abh2169. PubMed PMID: 34321199; PMCID: PMC8318366.

[0232] 43. Kuswanto et al., Poor Repair of Skeletal Muscle in Aging Mice Reflects a Defect in Local, Interleukin-33 -Dependent Accumulation of Regulatory T Cells. Immunity.

[0233] 2016;44(2):355-67. Epub 20160209. doi: 10.1016 / j.immuni.2016.01.009. PubMed PMID: 26872699; PMCID: PMC4764071.

[0234] 44. Molofsky et al., Interleukin-33 in Tissue Homeostasis, Injury, and Inflammation. Immunity. 2015;42(6): 1005-19. doi: 10.1016 / j.immuni.2015.06.006. PubMed PMID: 26084021; PMCID: PMC4471869.

[0235] 45. Sono et al., Perivascular Fibro-Adipogenic Progenitor Tracing during Post- Traumatic Osteoarthritis. The American Journal of Pathology. 2020; 190(9): 1909-20. doi: 10.1016 / j.ajpath.2020.05.017.

[0236] 46. Harvey et al., A Tppp3+Pdgfra+ tendon stem cell population contributes to regeneration and reveals a shared role for PDGF signalling in regeneration and fibrosis. Nature Cell Biology. 2019;21(12): 1490-503. doi: 10.1038 / s41556-019-0417-z.

[0237] 49. Hahaut et al., Fast and highly sensitive full-length single-cell RNA sequencing using FLASH-seq. Nat Biotechnol. 2022;40(10): 1447-51. Epub 20220530. doi: 10.1038 / s41587-022-01312-3. PubMed PMID: 35637419; PMCID: PMC9546769.

[0238] 50. Robinson et al., edgeR: a Bioconductor package for differential expression analysis of digital gene expression data. Bioinformatics. 2010;26(l): 139-40. Epub 20091111. doi: 10.1093 / bioinformatics / btp616. PubMed PMID: 19910308; PMCID: PMC2796818.

[0239] 51. Li et al., DEApp: an interactive web interface for differential expression analysis of next generation sequence data. Source Code Biol Med. 2017;12:2. Epub 20170203. doi: 10.1186 / S13029-017-0063-4. PubMed PMID: 28174599; PMCID: PMC5291987.

[0240] 52. Uhlen et al., A genome-wide transcriptomic analysis of protein-coding genes in human blood cells. Science. 2019;366(6472). doi: 10.1126 / science.aax9198. PubMed PMID: 31857451.

[0241] 53. Zheng et al., Massively parallel digital transcriptional profiling of single cells. Nat Comms. 2017;8(l): 14049. doi: 10.1038 / ncomms 14049. 54. Hao et al., Dictionary learning for integrative, multimodal and scalable single-cell analysis. Nat Biotechnol. 2024;42(2):293-304. Epub 20230525. doi: 10.1038 / s41587-023- 01767-y. PubMed PMID: 37231261; PMCID: PMC10928517.

[0242] 55. Korsunsky et al., Fast, sensitive and accurate integration of single-cell data with Harmony. Nature Methods. 2019;16(12): 1289-96. doi: 10.1038 / s41592-019-0619-0. 56. Severe et al., Stress-Induced Changes in Bone Marrow Stromal Cell Populations

[0243] Revealed through Single-Cell Protein Expression Mapping. Cell Stem Cell.

[0244] 2019;25(4):570-83 e7. Epub 20190703. doi: 10.1016 / j .stem.2019.06.003. PubMed PMID: 31279774; PMCID: PMC6778015.

[0245] 57. Baryawno et al., Cellular Taxonomy of the Bone Marrow Stroma in Homeostasis and Leukemia. Cell. 2019;177(7): 1915-32 el6. Epub 20190523. doi:

[0246] 10.1016 / j.cell.2019.04.040. PubMed PMID: 31130381; PMCID: PMC6570562.

[0247] OTHER EMBODIMENTS

[0248] A number of embodiments of the invention have been described. Nevertheless, it will be understood that various modifications may be made without departing from the spirit and scope of the invention.

Claims

WHAT IS CLAIMED IS:

1. A method of analyzing circulating cells and particles in a blood sample from a subject for association with heterotopic ossification (HO) by negative selection of circulating cells or particles, the method comprising: obtaining a blood sample from the subject, wherein circulating cells or particles in the blood sample comprise mesenchyme-derived cells or particles, fibrous tissue-derived cells or particles, cartilage-derived cells or particles, bone-derived cells or particles, or bone marrow-derived cells or particles; mixing the blood sample with magnetic beads and a binding agent that specifically binds to white blood cells (WBCs) for a time and under conditions sufficient for the binding agent to bind to the WBCs; flowing the blood sample through a size-based separation structure configured to direct red blood cells and platelets in the blood sample to a first waste outlet and flowing the remaining blood sample to an inertial focusing structure; flowing the remaining blood sample through the inertial focusing structure at a flow rate and for a distance sufficient to cause circulating cells and particles in the remaining blood sample to align in one or more streamlines within the remaining blood sample flowing through the inertial focusing structure; flowing the remaining blood sample with the circulating cells and particles aligned in one or more streamlines through a magnetophoresis structure for a time and distance sufficient to separate the WBCs bound to magnetic beads from the circulating cells and particles that are not bound to magnetic beads, flowing the bound WBCs into a second waste outlet, and flowing the remaining unbound circulating cells and unbound particles to a product outlet; and analyzing the unbound circulating cells or unbound particles from the product outlet for an HO-associated molecular signature comprising one or more of gene expression, protein expression, or post-translational modification, to determine whether the unbound circulating cells or unbound particles are HO-associated.

2. A method of analyzing circulating cells and particles in a blood sample from a subject for association with heterotopic ossification (HO) by positive selection of circulating cells or particles, the method comprising:obtaining a blood sample from the subject, wherein circulating cells or particles in the blood sample comprise mesenchyme-derived cells or particles, fibrous tissue-derived cells or particles, cartilage-derived cells or particles, bone-derived cells or particles, or bone marrow-derived cells or particles; mixing the blood sample with magnetic beads comprising a binding agent that specifically binds to circulating cells or particles and does not bind to white blood cells (WBCs) for a time and under conditions sufficient for the binding agent to bind to the circulating cells or particles; flowing the blood sample through a size-based separation structure configured to direct red blood cells and platelets in the blood sample to a first waste outlet and to direct the remaining blood sample to an inertial focusing structure; flowing the remaining blood sample through the inertial focusing structure at a flow rate and for a distance sufficient to cause circulating cells and particles in the remaining blood sample to align in one or more streamlines within the remaining blood sample flowing through the inertial focusing structure; flowing the remaining blood sample with the circulating cells and particles aligned in one or more streamlines through a magnetophoresis structure for a time and distance sufficient to separate the circulating cells and particles bound to magnetic beads from unbound WBCs and other unbound cells and unbound particles that are not bound to magnetic beads, flowing the unbound WBCs, other unbound cells, and unbound particles into a second waste outlet and flowing the bound circulating cells or bound particles to a product outlet; and analyzing the bound circulating cells or bound particles for an HO-associated molecular signature comprising one or more of gene expression, protein expression, or post-translational modification, to determine whether the bound circulating cells or bound particles are HO-associated.

3. The method of claim 1, wherein the binding agent that specifically binds to white blood cells comprises an antibody that specifically bind to CD45, CD16, or CD66b proteins found on the surface of white blood cells.

4. The method of claim 2, wherein the binding agent that specifically binds to circulating cells or particles and not to WBCs or other cells or particles in the blood sample comprises an antibody that specifically binds to plasma membrane proteins selected from the group consisting of PDGFRa, CD90, PGAP1, CDH11, GPX8, FBLN5, PRELP, LAMA4, or CADM3.

5. The method of any one of claims 1 to 4, wherein the HO-associated genes are selected from the group consisting of Ackr3, Adm5, Asprvl, Cadm3, Cdhll, Collal, Colla2, Col3al, Col5al, Crtacl, Dact3, Dpysl3, Fmod, Gfpt2, Gpx8, Igfbp5, 1133, Lama4, Matn2, Mmp2, Msantd3, Tmeffl, Mxra5, Pcolce2, Pcyoxl, Penk, Pgapl, Plod2, Postn, Prelp, Rbms2, Trim2, Vnn3, Pcdhl9, or Dpp4.

6. The method of any one of claims 1 to 4, wherein the HO-associated genes are selected from the group consisting of Adm 5, Asprvl, Colla2, Gjpt2, Gpx8, Matn2, Mxra5, Prelp, or Pcdhl9.

7. The method of any one of claims 1 to 4, wherein the HO-associated genes are selected from the group consisting of Fmod, Collal, Colla2, Col2al, Fgfr3, Dsp, Acvr2b, Inhba, 1133, Plod2, Ackr3, or Nt5c3b.

8. The method of any one of claims 1 to 7, wherein the circulating cells or particles comprise mesenchyme-derived cells or particles.

9. The method of claim 8, wherein the mesenchyme-derived cells comprise one or more of mesenchymal progenitor cells (MPCs), mesenchymal stem cells (MSCs), osteogenic progenitor cells or chondrogenic progenitor cells.

10. The method of any one of claims 1 to 7, wherein the circulating particles comprise organelles or extracellular vesicles.

11. The method of claim 10, wherein the extracellular vesicles comprise one or more of microvesicles (MVs), exosomes, oncosomes, and apoptotic bodies.

12. The method of any one of claims 1 to 11, wherein the size-based separation structure comprises an inertial exchanger configured to direct red blood cells,platelets, and particles in the blood sample to the first waste outlet and to direct the remaining blood sample to the inertial focusing structure.

13. The method of any one of claims 1 to 11, wherein the size-based separation structure comprises a deterministic lateral displacement array of microposts in a channel, wherein the array of microposts is configured to direct red blood cells, platelets, and particles in the blood sample to a first waste outlet and to direct the remaining blood sample to the inertial focusing structure.

14. The method of any one of claims 1 to 13, wherein determining whether the circulating cells or particles are HO-associated circulating cells and particles comprises analyzing the circulating cells or particles using droplet digital PCR, an immunoassay, or both.

15. The method of any one of claims 1 to 13, wherein determining whether the circulating cells or particles are HO-associated circulating cells and particles comprises analyzing the cells or particles using detection of antigens unique to HO-associated circulating cells or particles via fluorescently conjugated antibodies.

16. The method of any one of claims 1 to 13, wherein determining whether the circulating cells or particles are HO-associated circulating cells and particles comprises analyzing the circulating cells or particles using HO-associated genes, transcripts, or proteins for differentiating HO-associated circulating cells or particles from non-HO- associated circulating cells or particles.

17. The method of any one of claims 1 to 13, wherein determining whether the circulating cells or particles are HO-associated circulating cells and particles comprises analyzing the circulating cells or particles using single-cell RNA sequencing.

18. The method of any one of claims 1 to 17, further comprising detecting a quantity of the HO-associated circulating cells or particles, a quality of the HO-associated circulating cells or particles, or both, for determining a specific level of risk of heterotopic ossification.

19. The method of claim 18, wherein a quantity of HO -associated circulating cells in the blood sample from the subject that is higher than a reference quantity determined from HO-associated circulating cells in blood samples from a plurality of healthy individuals indicates the subject is at risk for developing HO.

20. The method of any one of claims 1 to 19, further comprising treating a subject at risk for developing HO by administering an anti -HO treatment.

21. The method of claim 20, wherein the anti -HO treatment comprises administering a non-steroidal anti-inflammatory drug, warfarin, an electro-magnetic field, radiation therapy, or a bisphosphonate drug.

22. The method of claim 20 or claim 21, further comprising determining the efficacy of the anti-HO treatment, the method further comprising: taking a second blood sample at a first time, at or before starting the treatment, to determine a first level or number of HO-associated circulating cells or particles in the second blood sample of the subject; taking a third blood sample at a second time, later than the first time, to determine a second level or number of HO-associated circulating cells or particles in the third blood sample of the subject; and determining a level of efficacy of the treatment, wherein a decrease in a level or number of HO-associated circulating cells or particles at the second time compared to the level or number of HO-associated circulating cells or particles at the first time indicates that the treatment was effective.

23. The method of claim 20 or claim 21, further comprising determining the efficacy of the anti-HO treatment, the method further comprising: taking a second blood sample at a first time, at or before starting the treatment, to determine a first level or number of HO-associated genes in circulating cells or particles in the second blood sample of the subject; taking a third blood sample at a second time, later than the first time, to determine a second level or number of HO-associated genes in circulating cells or particles in the third blood sample of the subject; anddetermining a level of efficacy of the treatment, wherein a decrease in a level or number of HO-associated genes in circulating cells or particles at the second time compared to the level or number of HO-associated genes at the first time indicates that the treatment was effective.

24. A system for isolating, analyzing, or both isolating and analyzing, circulating cells or particles from a blood sample from a subject, wherein the circulating cells or particles comprise mesenchyme-derived cells or particles, fibrous tissue-derived cells or particles, cartilage-derived cells or particles, bone-derived cells or particles, or bone marrow-derived cells or particles, the system comprising: a mixer for combining the blood sample with magnetic beads comprising a binding agent that specifically binds to either(i) circulating cells or particles and not to white blood cells (WBCs), or(ii) WBCs and not to other cells or particles, for a time and under conditions sufficient for the binding agent to bind; a size-based separation structure configured to direct red blood cells and platelets in the blood sample to a first waste outlet; an inertial focusing structure configured to align circulating cells and particles in the remaining blood sample into one or more streamlines within the remaining blood sample flowing through the inertial focusing structure; a magnetophoresis structure configured to separate cells or particles bound to magnetic beads from cells and particles not bound to magnetic beads, and directing the bound cells or bound particles into a second waste outlet and flowing unbound cells or unbound particles to a product outlet; and an analyzer configured to determine which of the circulating cells or particles are associated with a heterotopic ossification (HO)-associated molecular signature comprising one or more of gene expression, protein expression, or post-translational modification, to determine whether the circulating cells or unbound particles are HO- associated.

25. The system of claim 24, wherein the size-based separation structure comprises an inertial exchanger configured to direct red blood cells and platelets in theblood sample to a first waste outlet and to direct the remaining blood sample to the inertial focusing structure.

26. The system of claim 24, wherein the size-based separation structure comprises a deterministic lateral displacement array of microposts in a channel, wherein the array of microposts is configured to direct red blood cells and platelets in the blood sample to a first waste outlet and to direct the remaining blood sample to the inertial focusing structure.

27. The system of any one of claims 24 to 26, wherein the analyzer comprises a system to encapsulate cells or particles in individual droplets and to perform ddPCR on each individual droplet to determine which circulating cells or particles are HO- associated.

28. The system of any one of claims 24 to 26, wherein the analyzer comprises a system that performs RNA sequencing to determine whether the circulating cells or particles are HO-associated circulating cells and particles.

29. A method of detecting mesenchyme-derived cells in a blood sample, the method comprising: mixing a blood sample with magnetic beads and a binding agent that specifically binds to white blood cells (WBCs) for a time and under conditions sufficient for the binding agent to bind to the WBCs; flowing the blood sample through a size-based separation structure configured to direct red blood cells and platelets in the blood sample to a first waste outlet and flowing the remaining blood sample to an inertial focusing structure; flowing the remaining blood sample through the inertial focusing structure at a flow rate and for a distance sufficient to cause cells and particles in the remaining blood sample to align in one or more streamlines within the remaining blood sample flowing through the inertial focusing structure; flowing the remaining blood sample with the circulating cells and particles aligned in one or more streamlines through a magnetophoresis structure for a time and distance sufficient to separate the WBCs bound to magnetic beads from the cells and particles thatare not bound to magnetic beads, flowing the bound WBCs into a second waste outlet, and flowing the remaining unbound cells and unbound particles to a product outlet; and analyzing the unbound cells or unbound particles from the product outlet for a mesenchyme-derived molecular signature comprising one or more of gene expression, protein expression, or post-translational modification, to determine whether the unbound cells or unbound particles are mesenchyme-derived.

30. The method of claim 29, wherein the binding agent that specifically binds to white blood cells comprises an antibody that specifically bind to CD45, CD 16, and CD66b proteins found on the surface of white blood cells.

31. The method of claim 29 or claim 30, wherein the genes associated with mesenchyme-derived cells include those selected from the group consisting of Ackr3, Adm5, Asprvl, Cadm3, Cdhll, Collal, Colla2, Col3al, Col5al, Crtacl, Dact3, Dpysl3, Fmod, Gjpt2, Gpx8, Igfbp5, 1133, Lama4, Matn2, Mmp2, Msantd3, Tmeffl, Mxra5, Pcolce2, Pcyoxl, Penk, Pgapl, Plod2, Postn, Prelp, Rbms2, Trim2, Vnn3, Pcdhl9, and Dpp4.

32. The method of any one of claims 29 to 31, wherein the mesenchyme- derived cells comprise one or more of mesenchymal progenitor cells (MPCs), mesenchymal stem cells (MSCs), osteogenic progenitor cells or chondrogenic progenitor cells.

33. The method of any one of claims 29 to 31, wherein the particles comprise organelles or extracellular vesicles.

34. The method of claim 33, wherein the extracellular vesicles comprise one or more of microvesicles (MVs), exosomes, oncosomes, and apoptotic bodies.

Images