Systems and methods for label-free tracking of human somatic cell reprogramming

Abstract

Systems and methods are provided for identifying the current reprogramming status and predicting future reprogramming status for reprogramming intermediate cells (i.e., somatic cells undergoing reprogramming). Label-free autofluorescence measurements are combined with machine learning techniques to provide highly accurate identification of the current reprogramming status and prediction of future reprogramming status. Identification of the current reprogramming status utilizes metabolic endpoints from an autofluorescence data set. Prediction of future reprogramming status utilizes a pseudo-time axis constructed from autofluorescence data of reprogramming intermediate cells with known reprogramming status.

Description

[Technical field]

[0001] (CROSS REFERENCE TO RELATED APPLICATIONS) This application claims priority to U.S. Provisional Patent Application No. 63 / 257,034, filed October 18, 2021, which is incorporated by reference herein for all purposes. (Statement Regarding Federally Funded Research) Not applicable [Background technology]

[0002] The derivation of patient-specific induced pluripotent stem cells (iPSCs) from their somatic cells via reprogramming generates a unique self-renewing cell source for disease modeling, drug discovery, toxicology, and personalized cell therapy. These cells retain the patient's genome to facilitate elucidation of the genetic causes of disease and are immunologically matched to the patient to facilitate engraftment of any cellular therapy generated from these cells. There is a need for new platforms for biomanufacturing iPSCs that are integration-free, rapid, scalable, and easily transferable to GMP-compatible conditions. Summary of the Invention

[0003] In one aspect, the present disclosure provides a somatic cell reprogramming tracking device. The device includes a cell analysis observation zone, an autofluorescence spectrometer, a processor, and a non-transitory computer readable medium. The cell analysis observation zone is adapted to receive the reprogramming intermediate cells and present the reprogramming intermediate cells for individual autofluorescence matching. The autofluorescence spectrometer is configured to acquire an autofluorescence data set regarding the reprogramming intermediate cells positioned in the cell analysis observation zone. The autofluorescence spectrometer includes a light source, a photon counting detector, and photon counting electronics. The processor is in electronic communication with the autofluorescence spectrometer. The non-transitory computer readable medium is accessible to the processor and has instructions stored thereon. The instructions, when executed by a processor, cause the processor to: a) receive an autofluorescence data set; and b) identify a current reprogramming status of the reprogramming intermediate cell based on a current reprogramming prediction, the current reprogramming prediction calculated using at least a portion of the autofluorescence data set, the current reprogramming prediction calculated using at least one metabolic endpoint and optionally at least one nuclear parameter of the autofluorescence data set as input, the at least one metabolic endpoint being determined by a flavin adenine dinucleotide (FAD) mean fluorescence lifetime (τ m ), the shortest lifetime amplitude component of FAD (α1), the shortest fluorescence lifetime component of FAD (τ1), the longest fluorescence lifetime component of FAD (τ2), or a combination thereof.

[0004] In another aspect, the present disclosure provides a method for characterizing the progression of somatic cell reprogramming, the method comprising the steps of: a) optionally receiving a population of reprogramming intermediate cells having an unknown reprogramming status; b) acquiring an autofluorescence data set from the reprogramming intermediate cells of the population of reprogramming intermediate cells; and c) identifying a current reprogramming status of the reprogramming intermediate cells based on a current reprogramming prediction, the current reprogramming prediction being calculated using at least a portion of the autofluorescence data set, the current reprogramming prediction being calculated using at least one metabolic endpoint and optionally at least one nuclear parameter of the autofluorescence data set as inputs, the at least one metabolic endpoint being a flavin adenine dinucleotide (FAD) mean fluorescence lifetime (τ m ), the shortest lifetime amplitude component of FAD (α1), the shortest fluorescence lifetime component of FAD (τ1), the longest fluorescence lifetime component of FAD (τ2), or a combination thereof.

[0005] In another aspect, the disclosure provides a method for characterizing the progression of somatic cell reprogramming. The method includes any of the following steps: a) receiving a population of reprogramming intermediate cells having an unknown reprogramming status; b) acquiring an autofluorescence data set for each reprogramming intermediate cell of the population of reprogramming intermediate cells, each autofluorescence data set including autofluorescence lifetime information; and c1) physically isolating a first portion of the population of reprogramming intermediate cells from a second portion of the population of reprogramming intermediate cells based on a current reprogramming prediction, wherein each reprogramming intermediate cell of the population of reprogramming intermediate cells is placed in the first portion when the current reprogramming prediction exceeds a predetermined threshold and is placed in the second portion when the current reprogramming prediction is less than or equal to the predetermined threshold; or c2) generating a report including the current reprogramming prediction, the report optionally identifying a proportion of the population of reprogramming intermediate cells having a current reprogramming prediction that exceeds a predetermined threshold, wherein the current reprogramming prediction is calculated using at least one metabolic endpoint and optionally at least one nuclear parameter of the autofluorescence data set as input, wherein the at least one metabolic endpoint is determined based on a flavin adenine dinucleotide (FAD) mean fluorescence lifetime (τ m ), the shortest lifetime amplitude component of FAD (α1), the shortest fluorescence lifetime component of FAD (τ1), the longest fluorescence lifetime component of FAD (τ2), or a combination thereof.

[0006] In another aspect, the present disclosure provides a method for generating a pseudo-temporal reprogramming pathway map, the method comprising the steps of: a) receiving an autofluorescence data set and optionally a nuclear data set for a plurality of reprogramming intermediate cells, the autofluorescence data set corresponding to a pseudo time point along a pseudo time axis of reprogramming; b) constructing a pseudo-temporal single-cell trajectory associated with each of the plurality of reprogramming intermediate cells based on the received autofluorescence data set associated with each of the plurality of reprogramming intermediate cells, each of the pseudo-temporal single-cell trajectories including a current reprogramming prediction associated with each of the predetermined pseudo time points, the current reprogramming prediction being calculated using at least a portion of the autofluorescence data set and optionally using at least a portion of the nuclear data set, the current reprogramming prediction being calculated using as input at least one metabolic endpoint of at least one of the autofluorescence data sets and optionally at least one nuclear parameter of at least one of the nuclear data sets, the at least one metabolic endpoint being calculated using as input a flavin adenine dinucleotide (FAD) mean fluorescence lifetime (τ m ), FAD shortest lifetime amplitude component (α1), FAD shortest fluorescence lifetime component (τ1), nicotinamide adenine dinucleotide and / or reduced nicotinamide dinucleotide phosphate adenine dinucleotide (NAD(P)H) shortest lifetime amplitude component (α1), NAD(P)H shortest fluorescence lifetime component (τ1), NAD(P)H longest fluorescence lifetime component (τ2), or combinations thereof; c) compiling the constructed pseudo-time single-cell trajectories into a single compiled data set; d) identifying clusters, branching events, and / or disconnected branches within the single compiled data set; and e) identifying correlations between clusters, branching events, and / or disconnected branches, and current or future reprogramming status within the single compiled data set, thereby generating a pseudo-time reprogramming pathway map that is used to predict future reprogramming based on the correlations. [Brief description of the drawings]

[0007] [Figure 1] 1 is a flow chart illustrating a method according to an aspect of the present disclosure. [Diagram 2] 1 is a flow chart illustrating a method according to an aspect of the present disclosure. [Diagram 3] 1 is a flow chart illustrating a method according to an aspect of the present disclosure. [Figure 4] 1 is a flow chart illustrating a method according to an aspect of the present disclosure. [Diagram 5] FIG. 2 is a block diagram of a device according to an aspect of the present disclosure. [Figure 6] Plot of the trajectory of a reprogrammed EPC constructed from metabolic and nuclear parameters based on UMAP dimensionality reduction using Monocle, showing the four branch points colored by cell type. [Figure 7] FIG. 13 is a plot of the trajectory of a reprogrammed EPC constructed from metabolic and nuclear parameters based on UMAP dimensionality reduction using Monocle, showing the four branch points colored by pseudotime. [Figure 8] A collection of Monocle UMAP plots showing the clustering of reprogrammed EPCs. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0008] Before describing the present invention in further detail, it is to be understood that the present invention is not limited to the particular embodiments described. It is also to be understood that the terminology used herein is for the purpose of describing particular embodiments only, and is not intended to be limiting. The scope of the present invention will be limited only by the claims. As used herein, the singular forms "a," "an," and "the" include plural embodiments unless the context clearly indicates otherwise. Specific structures, devices, and methods for modification of biomolecules are disclosed. It should be apparent to one of ordinary skill in the art that many additional modifications beyond those already described are possible without departing from the concept of the invention. In interpreting this disclosure, all terms should be interpreted in the broadest possible manner consistent with the context. Variations of the term "comprise" should be interpreted to refer to elements, components, or steps in a non-exclusive manner, such that a referenced element, component, or step may be combined with other elements, components, or steps not expressly referenced. An embodiment that is said to "comprise" certain elements also contemplates "consisting essentially of" and "consisting of" those elements. When two or more ranges for a particular value are recited, the disclosure contemplates all combinations of the upper and lower limits of those ranges not expressly recited. For example, recitation of values ​​between 1 and 10 or between 2 and 9 also contemplates values ​​between 1 and 9 or between 2 and 10.

[0009] As used herein, the term "FAD" refers to flavin adenine dinucleotide. The term "memory" as used herein includes non-volatile media, such as magnetic media or hard disks, optical storage devices, or flash memory; volatile media, such as system memory, such as random access memory (RAM), such as DRAM, SRAM, EDO RAM, RAMBUS RAM, DR RAM, etc.; or installation media, such as software media, such as CD-ROMs, or floppy disks, on which programs may be stored and / or data communications may be buffered. The term "memory" may also include other types of memory or combinations thereof. As used herein, the term "NAD(P)H" refers to reduced nicotinamide adenine dinucleotide and / or reduced nicotinamide dinucleotide phosphate.

[0010] As used herein, "nuclear parameters" refer to measured genomic or clustering properties of the nucleus of a cell of a subject, determined by analyzing acquired images of the cell of a subject. As used herein, the term "processor" may include one or more processors and memory and / or one or more programmable hardware elements. As used herein, the term "processor" is intended to include any processor, CPU, GPU, microcontroller, digital signal processor, or other type of device capable of executing software instructions. The term "pseudotime" / "pseudotimeline" as used herein refers to the dimension of progression along a cell reprogramming pathway. Pseudotime is related to time in that progression is directional in the same way as time, but differs from time in that it is not related to a fixed time course. Two different cells may take different lengths of actual time to go through the same length of pseudotime.

[0011] The term "redox ratio" or "optical redox ratio" as used herein refers to the ratio of NAD(P)H fluorescence intensity to FAD fluorescence intensity; the ratio of FAD fluorescence intensity to NAD(P)H fluorescence intensity; the ratio of NAD(P)H fluorescence intensity to any arithmetic combination that includes FAD fluorescence intensity; or the ratio of FAD fluorescence intensity to any arithmetic combination that includes NAD(P)H fluorescence intensity. In certain instances, the redox ratio or optical redox ratio refers to the ratio of NAD(P)H fluorescence intensity to the sum of NAD(P)H and FAD fluorescence intensities. As used herein, "somatic cell" refers to any non-germ cell. In this disclosure, the term somatic cell can refer to a single cell that is subjected to reprogramming. A somatic cell undergoing reprogramming can be identified by a single name, "reprogramming intermediate (IM) cell," with the understanding that the cell's cell type can be changed.

[0012] Autofluorescence endpoints include photon counts / intensity and fluorescence lifetime. The fluorescence lifetime of a cell can be a single value, the average fluorescence lifetime, or can be influenced by the lifetime values ​​of multiple subspecies with different lifetimes. In this case, multiple lifetime and lifetime component amplitude values ​​are extracted. Both NAD(P)H and FAD can exist in quenched (short lifetime) and unquenched (long lifetime) configurations; therefore, the fluorescence decays of NAD(P)H and FAD fit into two components. In general, NADH and FAD fluorescence lifetime decays follow a two-component exponential decay, I(t)=α1e -t / τ1 +α2e -t / τ2 +C, where I(t) is the fluorescence intensity as a function of time t after the laser pulse, α1 and α2 are the fractional contributions of the short and long lifetime components, respectively (i.e., α1 + α2 = 1), τ1 and τ2 are the short and long lifetime components, respectively, and C accounts for background light. However, the lifetime decay can be fitted to more components (theoretically, any number of components, but in practice up to about 5-6), which may allow quantification of additional lifetimes and component amplification. By convention, lifetimes and amplitudes are numbered from short to long values, but this may be reversed. The mean lifetime can be calculated from the lifetime components (τ m =α1τ1+α2τ2...). Fluorescence lifetimes and lifetime component amplitudes can also be approximated from frequency domain data and a gated camera / detector. For gated detection, α1 may be approximated by dividing the intensity detected in the early bin by the late bin. Alternatively, fluorescence anisotropy can be measured by polarization-sensitive detection of autofluorescence, thus identifying free NAD(P)H as having short rotational diffusion times in the range of 100-700 picoseconds.

[0013] FADα1 refers to the combined FAD contribution and is the shortest lifetime that is not dominated (i.e., greater than 50%) by instrument response and / or scattering. FADα1 is the contribution associated with FAD lifetime values ​​of 50-1500 ps, ​​50-1000 ps, ​​or 50-600 ps. For clarity, claims herein that include a characterization of a "shortest" lifetime cannot be avoided by defining the lifetime value to include a sacrificial shortest lifetime that is dominated by instrument response and / or scattering. FADτ1 refers to the bond FAD lifetime, which is the shortest lifetime that is not dominated (i.e., greater than 50%) by instrument response and / or scattering. FADτ1 is an FAD lifetime value between 50 and 1500 picoseconds, between 50 and 1000 picoseconds, or between 50 and 600 picoseconds. For clarity, claims herein that include a characterization of a "shortest" lifetime cannot be circumvented by defining the lifetime value to include a sacrificial shortest lifetime that is dominated by instrument response and / or scattering. FADτ2 refers to the free FAD lifetime and is the longest lifetime that is not dominated (i.e., greater than 50%) by instrument response and / or scattering. FADτ2 is an FAD lifetime value between 1000 and 4000 picoseconds, between 1000 and 3000 picoseconds, or between 1500 and 3000 picoseconds. For clarity, claims herein that include a characterization of a "longest" lifetime cannot be circumvented by defining the lifetime value to include a sacrificial shortest lifetime that is dominated by instrument response and / or scattering. FADτ m =α1·τ1+(1-α1)·τ2 Various aspects may be described herein in terms of various functional components and process steps. It should be understood that such components and steps may be realized by any number of hardware components configured to perform the specified functions.

[0014] (method) The present disclosure provides various methods. It should be understood that the various methods are suitable for use in other methods. Likewise, it should be understood that the various methods are suitable for use in systems described elsewhere herein. When a feature of the present disclosure is described with respect to a given method, it is expressly contemplated that the feature is also useful in the other methods and systems described herein, unless the context clearly indicates otherwise.

[0015] The method described herein includes two different types of predictions regarding the reprogramming status of a given reprogramming intermediate cell. First, there is a current reprogramming prediction, which provides a computer-generated prediction regarding the current status of reprogramming in a given reprogramming intermediate cell of a subject. For example, the current reprogramming prediction may indicate that the given reprogramming intermediate cell has been reprogrammed into an induced pluripotent stem cell. Second, there is a future reprogramming prediction, which provides a computer-generated prediction regarding the future status of reprogramming in a given reprogramming intermediate cell of a subject based on a pseudo-temporal reprogramming pathway map (discussed in more detail below). The prediction regarding the future status may be based on the current reprogramming prediction. In other words, the future reprogramming prediction may indicate that the given reprogramming intermediate cell is likely or unlikely to eventually be reprogrammed into an iPSC. This future prediction is based on machine learning analysis of a large population of cells undergoing reprogramming, which identified that cells encompassed in a given cluster on the pseudo-temporal reprogramming pathway map have a given probability of taking various possible reprogramming pathways available.

[0016] With reference to FIG. 1 , the present disclosure provides a method 100 for characterizing the progression of somatic cell reprogramming. At process block 102, the method 100 may include receiving a population of reprogramming intermediate cells having an unknown reprogramming status. The population of reprogramming intermediate cells may itself be included in a broader population of cells that includes some cells that are not reprogramming intermediate cells. At process block 104, the method 100 includes acquiring an autofluorescence data set for each reprogramming intermediate cell of the population of reprogramming intermediate cells. At process block 106, the method 100 includes identifying a current reprogramming status of each of the reprogramming intermediate cells based on a current reprogramming prediction. The current reprogramming prediction is calculated using at least a portion of the autofluorescence data set. The current reprogramming prediction is calculated using at least one metabolic endpoint of the autofluorescence data set and, optionally, using at least one nuclear parameter as input. The at least one metabolic endpoint is calculated using FADτ. m, FADτ1, FADτ2, FADα1, or combinations thereof. After process block 106, method 100 can proceed to process block 108 or 110 depending on the desired outcome. In some cases, method 100 proceeds to process block 108 and process block 110 in either order. Although process blocks 108 and 110 are optionally shown and described together, method 100 includes either process block 108 or process block 110. At optional process block 108, method 100 may include physically isolating a first portion of the population of reprogramming intermediate cells from a second portion of the population of reprogramming intermediate cells based on a current reprogramming prediction, where each reprogramming intermediate cell of the population of reprogramming intermediate cells is placed in the first portion when the current reprogramming prediction exceeds a predetermined threshold and is placed in the second portion when the current reprogramming prediction is less than or equal to the predetermined threshold. At optional process block 110, method 100 may include generating a report including the current reprogramming prediction. The report may include identifying a percentage of the population of reprogramming intermediate cells that have a current reprogramming prediction above a predetermined threshold.

[0017] With reference to FIG. 2 , the present disclosure provides a method 200 for characterizing a reprogramming status of reprogramming intermediate cells. At optional process block 202, the method 200 may include receiving a population of reprogramming intermediate cells having an unknown reprogramming status. At process block 204, the method 200 includes acquiring an autofluorescence data set from the reprogramming intermediate cells of the population of reprogramming intermediate cells. At process block 206, the method 200 includes calculating a current reprogramming prediction using at least a portion of the autofluorescence data set. The current reprogramming prediction is calculated using at least one metabolic endpoint and, optionally, using at least one nuclear parameter. The at least one metabolic endpoint may be FADτ. m, FADα1, FADτ1, FADτ2, or a combination thereof. At process block 208, the method 200 includes identifying a current reprogramming status of the reprogramming intermediate cell based on the current reprogramming prediction.

[0018] Methods 100 and 200 are interrelated and may be utilized together, for example, method 200 may be utilized within method 100. Aspects described with respect to method 100 may be utilized with method 200, and vice versa, unless the context clearly indicates otherwise. Methods 100 and 200 may further include identifying a future reprogramming status of one or more reprogramming intermediate cells, either alone or within a population of reprogramming intermediate cells. The identification of the future reprogramming status is based on a current reprogramming prediction and a pseudotemporal reprogramming pathway map based on machine learning analysis of an autofluorescence data set acquired for reprogramming intermediate cells with known reprogramming status over the course of a pseudotemporal trajectory. In some cases, the pseudotemporal reprogramming pathway map is generated by method 400, as discussed below.

[0019] The autofluorescence data set acquired at process block 104 or 204 can be acquired in a variety of ways, as would be understood by one skilled in the art of spectroscopy with the knowledge of this disclosure and their own knowledge gained from the field. For example, the autofluorescence data can be acquired from fluorescence decay data. In another example, the autofluorescence data can be acquired by gating a detector (e.g., a camera) to acquire data at specific times throughout the decay to approximate the autofluorescence end point described herein. As yet another example, a frequency domain approach can be used to measure lifetime. Alternatively, fluorescence anisotropy can be measured by polarization sensitive detection of the autofluorescence, thus identifying free NAD(P)H as having short rotational diffusion times in the range of 100-700 picoseconds. The particular way in which the autofluorescence data is acquired is not intended to limit the scope of the invention, so long as the lifetime information necessary to determine the autofluorescence end point required for the methods described herein can be appropriately measured, estimated, or determined in any manner. One example of a suitable autofluorescence data set acquisition is described in the Examples section below. The physical isolation operation of optional process block 108 is responsive to a current reprogramming prediction determined from the acquired autofluorescence data set. If the current reprogramming prediction exceeds a predetermined threshold for a given reprogramming intermediate cell, the reprogramming intermediate cell is placed in a first portion. If the current reprogramming prediction is less than or equal to a predetermined threshold for a given reprogramming intermediate cell, the reprogramming intermediate cell is placed in a second portion. The result of this physical isolation is that the first portion of the population of reprogramming intermediate cells is significantly enriched in reprogramming intermediate cells having a given reprogramming status (e.g., successfully reprogrammed as iPSCs), whereas the second portion of the population of reprogramming intermediate cells is significantly devoid of reprogramming intermediate cells having the given reprogramming status.

[0020] In some cases, the physical isolation operation of optional process block 108 may include isolating the cells into 3, 4, 5, 6, or more portions. In these cases, the various portions will be separated by a number of predefined thresholds, one less than the number of portions (i.e., 3 portions = 2 predefined thresholds). The portion whose current reprogramming prediction exceeds all of the predefined thresholds (i.e., exceeds the highest threshold) contains the greatest concentration of reprogramming intermediate cells with a given reprogramming status. The portion whose current reprogramming prediction cannot exceed any of the predefined thresholds (i.e., cannot exceed the lowest threshold) contains the lowest concentration of reprogramming intermediate cells with a given reprogramming status. The use of multiple predefined thresholds can provide for the preparation of portions of a population of reprogramming intermediate cells having very high or very low concentrations of reprogramming intermediate cells with a given reprogramming status. In some cases, the physical isolation operation of optional process block 108 (or the entire separation aspect of method 100, as would be understood by one of ordinary skill in the art of cell isolation) may include isolating other types of cells, such as red blood cells or the like, or various types of necrotic debris, which are therefore not included in the portion containing the reprogramming intermediate cells.

[0021] The current reprogramming prediction is calculated using at least one metabolic endpoint of the autofluorescence data set and may include at least one nuclear parameter for each reprogramming intermediate cell of the population of reprogramming intermediate cells as input. The current reprogramming prediction is calculated using an equation created by a machine learning process on data of a population of reprogramming intermediate cells with known reprogramming status using at least one metabolic endpoint and optionally at least one nuclear parameter as variables. In some cases, the reprogramming prediction can have different predictive potential for different reprogramming statuses (i.e., can be more predictive for iPSC status and other statuses). At least one metabolic endpoint was determined by measuring the FAD mean fluorescence lifetime (τ m), FAD shortest lifetime amplitude component (α1), FAD shortest fluorescence lifetime component (τ1), FAD longest fluorescence lifetime component (τ2), or a combination thereof. At least one metabolic endpoint may be: NAD(P)H fluorescence intensity; FAD fluorescence intensity; optical redox ratio (i.e., NAD(P)H / [NAD(P)H+FAD], see definition above); NAD(P)H shortest lifetime amplitude component or NAD(P)Hα1; NAD(P)H average fluorescence lifetime or NAD(P)Hτ m ; the FAD(P)H shortest fluorescence lifetime or NAD(P)Hτ1; the second shortest fluorescence lifetime of NAD(P)H or HAD(P)Hτ2.

[0022] The at least one nuclear parameter can include the area of ​​the nucleus of the reprogramming intermediate cell, the perimeter of the nucleus of the reprogramming intermediate cell, a nuclear shape index, the average distance from any pixel within a nucleus in a reprogramming intermediate cell to the nearest pixel outside the nucleus, the percentage of pixels located in a convex hull (i.e., the smallest convex polygon that fits around the nucleus) that are also located within the nucleus of the reprogramming intermediate cell, the percentage of pixels within a bounding box of the nucleus (i.e., the smallest rectangle that encloses the nucleus) that are also located within the nucleus of the reprogramming intermediate cell, the total number of nuclei adjacent to the nucleus of the reprogramming intermediate cell, the distance from the reprogramming intermediate cell nucleus to the nearest neighboring nucleus, or a combination thereof. In some cases, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18 or more inputs are used.

[0023] Method 100 or method 200 can sort the reprogramming intermediate cells into EPC, IM, and iPSC categories based on their current reprogramming status. Method 100 or method 200 can provide surprising accuracy in classifying the current reprogramming status of somatic cells. The accuracy can be at least 85%, at least 87.5%, at least 90%, at least 92.5%, at least 95%, at least 96%, at least 97%, or at least 98%. One non-limiting example of measuring accuracy includes performing method 100 or method 200 on a given cell with an unknown current reprogramming status, and then using one of the traditional methods to determine the reprogramming status (typically resulting in a destructive method) for a statistically significant number of cells. Method 100 or method 200 can be performed without using fluorescent labels to bind the reprogramming intermediate cells. Method 100 or method 200 can be performed without immobilizing the reprogramming intermediate cells.

[0024] With reference to FIG. 3, the present disclosure provides a method 300 for creating a pseudo-temporal reprogramming pathway map. At process block 302, the method 300 includes receiving autofluorescence data sets for a plurality of reprogramming intermediate cells undergoing reprogramming, the autofluorescence data sets corresponding to pseudo time points along a pseudo time axis of reprogramming. At process block 304, the method 300 includes constructing a pseudo-temporal single-cell trajectory associated with each of the plurality of reprogramming intermediate cells based on the received autofluorescence data sets associated with each of the plurality of reprogramming intermediate cells. Each of the pseudo-temporal single-cell trajectories includes a current reprogramming prediction associated with each of the predetermined pseudo time points. The current reprogramming prediction is calculated as described herein. The current reprogramming prediction used in this step of the method 300 is calculated using at least one metabolic endpoint and optionally at least one nuclear parameter as input. A metabolic endpoint as used herein is defined as FADτ. m, FADα1, FADτ1, NAD(P)Hα1, NAD(P)Hτ1, NAD(P)Hτ2, or combinations thereof. At process block 306, the method 300 includes compiling the constructed pseudo-time single-cell trajectories into a single compiled data set. At process block 308, the method 300 includes identifying clusters, bifurcation events, and / or disconnected branches within the single compiled data set. At process block 310, the method 300 includes identifying correlations between the clusters, bifurcation events, and / or disconnected branches and current or future reprogramming situations within the single compiled data set, thereby generating a pseudo-time reprogramming pathway map for use in predicting future reprogramming based on the correlations.

[0025] For this purpose, differentiated cells, called erythroid progenitor cells or EPCs, are isolated from human blood and reprogrammed into induced pluripotent stem cells or iPSCs. The cells undergoing reprogramming are imaged for various cellular metabolic and nuclear parameters using autofluorescence microscopy. These parameters are then used to train a supervised learning algorithm. This algorithm can then identify the current status of the reprogrammed cells, i.e., starting EPCs, partially reprogrammed intermediate cells (IMs) or fully reprogrammed iPSCs, based solely on the metabolic and nuclear parameters. These parameters can further be used to predict the future reprogramming status of the cells, which is achieved through the construction of single-cell reprogramming trajectories using machine learning. Briefly, at the start of reprogramming, differentiated cells from blood (called erythroid progenitor cells or EPCs) had variable metabolic and nuclear characteristics. These cells could be separated into three groups, called clusters 1, 2, and 3, according to these characteristics. Single cells in cluster 2 are reprogrammed and progress through epigenetic reprogramming. However, cells in clusters 1 and 3 behave differently and do not progress through reprogramming even when exposed to reprogramming factors.

[0026] For cells from cluster 2 that complete reprogramming, these pluripotent stem cells, called iPSCs, have two different types of metabolic and nuclear characteristics. These two types of iPSCs can be grouped into two clusters: clusters 7 and 10. Cells in culture that have undergone reprogramming without being fully reprogrammed into iPSCs can also be distinguished with respect to their metabolic and nuclear characteristics. These cells are called intermediate cells, IM, and are metabolic and nuclear characteristics. Among these various cell types, trajectories can be created to infer the sequence of changes in these characteristics as the cells are reprogrammed. By following various trajectories, we find that cells in clusters 6 and 8 undergo significant changes in metabolic and nuclear characteristics that do not lead to successful reprogramming into iPSCs. By following several other trajectories, we identify various common branches that summarize the changes that single cells undergo in culture. The points where these branches of trajectories separate are called branching points. There are three distinct branching points that can be identified during reprogramming culture. Single cells that go to the right at branch points 1, 2, and 3 are fully reprogrammed into iPSCs within 25 days of reprogramming, whereas cells that go to the left at branch points 1 and 3 remain at an intermediate stage. Thus, metabolic and nuclear properties of single cells and changes during culture can be used to identify cells that will be successfully reprogrammed from those that will not.

[0027] With reference to FIG. 4, the present disclosure provides a method 400 of administering reprogrammed reprogramming intermediate cells to a subject in need thereof. At process block 402, method 400 includes method 100 or method 200 described above, providing a first portion of a population of reprogramming intermediate cells enriched in a current or future reprogramming state (when optional process block 108 is utilized), or providing a report identifying a percentage of reprogramming intermediate cells having a given current or future reprogramming state (when optional process block 110 is utilized). At optional process block 404, method 400 may include modifying the first portion of the population of reprogramming intermediate cells or the population of reprogramming intermediate cells. The modification may include gene editing. At process block 406, method 400 includes administering the first portion of the population of reprogramming intermediate cells to the subject if the cells have been sorted, or administering the population of reprogramming intermediate cells to the subject if the cells have not been sorted.

[0028] The somatic cells from which the reprogramming intermediate cells are generated can be blood cells, including peripheral blood cells, umbilical cord blood cells, bone marrow cells, and the like, skin biopsy cells such as keratinocytes, epithelial cells such as those excreted in the urine, and other cells that are understood to be useful in reprogramming. The reprogramming intermediate cells can be collected from the subject to which they are administered before being selected. The reprogramming intermediate cells can be directly introduced into the subject or can be subjected to additional processing before being introduced into the subject. In some cases, the reprogramming intermediate cells can be modified to contain a chimeric antigen receptor (CAR). The somatic cells from which the reprogramming intermediate cells arise can be harvested from a donor. Somatic cells can be reprogrammed according to methods that are understood by those skilled in the art of reprogramming and / or iPSC. Non-limiting examples of reprogramming methods include viral factors, CRISPR, and the like. It is also contemplated that the delivery method for reprogramming is not a limitation of the present disclosure. Delivery forms can be via episomes, lentivirus, mRNA delivery, or other forms that are understood by those skilled in the art. The method described herein provided the inventors with unexpected results. First, it was unclear whether the acquired fluorescence data could completely classify somatic cell reprogramming. Second, it was unclear whether insight into future reprogramming situations could be gained by classifying current reprogramming situations using the pseudo-temporal reprogramming pathway map described herein. Third, it was unexpected that FAD life span could classify somatic cell reprogramming, since these parameters do not provide discriminatory ability in other contexts.

[0029] (system) The present disclosure also provides systems, which may be suitable for use with the methods described herein. When features of the present disclosure are described with respect to a given system, it is expressly contemplated that the features can be combined with other systems and methods described herein, unless the context indicates otherwise. Referring to FIG. 5, the present disclosure provides a somatic cell sorting device 500. The device 500 includes an observation zone 506. The observation zone 506 is adapted to receive a cell analysis pathway 502, a cell culture (not shown), or other device or system capable of presenting reprogramming intermediate cells for optical interrogation. The device 500 includes a processor 512 and a non-transitory computer-readable medium 514, such as a memory. In some configurations, the processor 512 can be or otherwise include a field programmable gate array (FPGA). In configurations where the processor 512 is an FPGA, an additional processor (not shown) may be included to acquire images.

[0030] The device 500 may include a cell analysis pathway 502. The cell analysis pathway 502 includes an inlet 504, an observation zone 506, and an outlet 505. The device 500 may include a cell sorter 508. The observation zone 506 is coupled to the inlet 504 downstream of the inlet 504 and is coupled to the outlet 505 upstream of the outlet 505. The device 500 also includes a single cell autofluorescence spectrometer 510. The device 500 may further include an optional cell picking apparatus (not shown). Inlet 504 can be any nanofluidic, microfluidic, or other cell sorting inlet. Those skilled in the art of fluidics will be knowledgeable of suitable inlets 504, and this disclosure is not intended to be bound by one particular implementation of inlet 504. The outlet can be any nanofluidic, microfluidic, or other cell sorting outlet. Those skilled in the art of fluidics will be knowledgeable of suitable outlets 505, and this disclosure is not intended to be bound by one particular implementation of the outlet 505.

[0031] Observation zone 506 is configured to present the reprogramming intermediate cells for individual autofluorescence decay interrogation. Those skilled in the art will be knowledgeable of suitable observation zones 506, and the present disclosure is not intended to be bound by one particular implementation of observation zone 506. The optional cell sorter 508 has a sorter inlet 516 and at least two sorter outlets 518. The cell sorter is coupled to the observation zone 506 via the sorter inlet 516 downstream of the observation zone 506. The cell sorter 508 is configured to selectively direct cells from the sorter inlet 516 to one of the at least two sorter outlets 518 based on a sorting signal. The inlet 504, observation zone 506, outlet 505, and optional cell sorter 508 can be components known to those skilled in the art to be useful in high throughput cell screening devices or flow sorters, including commercial flow sorters. The cell analysis pathway 502 can further optionally include a flow conditioner, as would be understood by those skilled in the art. The flow conditioner can be configured to provide a flow of cells through the observation zone at a rate that allows an autofluorescence data set to be acquired by the autofluorescence spectrometer 510. A useful discussion of fluid sorting that can be used in conjunction with the present disclosure is Shields et al., "Microfluidic cell sorting: a review of the advances in the separation of cells from debulking to rare cell isolation," Lab Chip, 2015 Mar 7; 15(5): 1230-49, which is incorporated herein by reference in its entirety.

[0032] Any cell picking device can perform a similar function to any cell sorter 508, i.e., isolating cells based on a sorting signal. The cell picking device can be automated. One example of a suitable cell picking device includes the ALS CellCelector™, available from ALS Automated Lab Solutions GmbH, Jena, Germany. The autofluorescence spectrometer 510 includes a light source 524 , a photon-counting detector 526 , and photon-counting electronics 528 . Autofluorescence spectrometer 510 can be any spectrometer suitable for acquiring an autofluorescence data set, as will be appreciated by those skilled in the art of optics. Suitable light sources 524 include, but are not limited to, lasers, LEDs, lamps, filtered light, fiber lasers, and the like. The light source 524 can be pulsed, including naturally pulsed light sources, and continuous light sources that are optically modulated in short or other ways by external components.

[0033] The light source 524 can provide pulses of light having a full width at half maximum (FWHM) pulse width of a duration appropriate to achieve the spectroscopic goals described herein, as can be understood by one of ordinary skill in the art of spectroscopy, in some cases the FWHM pulse width is at least 1 femtosecond, at least 5 femtoseconds, at least 10 femtoseconds, at least 25 femtoseconds, at least 50 femtoseconds, at least 100 femtoseconds, at least 200 femtoseconds, at least 350 femtoseconds, at least 500 femtoseconds, at least 750 femtoseconds, at least 1 picosecond, at least 3 picoseconds, at least 5 picoseconds, at least 10 picoseconds, at least 20 picoseconds, at least 50 picoseconds, or at least 100 picoseconds. In some cases, the FWHM pulse width is up to 10 ns, up to 1 ns, up to 900 ps, ​​up to 750 ps, ​​up to 600 ps, ​​up to 500 ps, ​​up to 400 ps, ​​up to 250 ps, ​​up to 175 ps, up to 100 ps, ​​up to 75 ps, up to 60 ps, ​​up to 50 ps, ​​up to 35 ps, up to 25 ps, up to 20 ps, ​​up to 15 ps, up to 10 ps, ​​or up to 1 ps.

[0034] The light source 524 can emit wavelengths that are tuned for the absorption of NAD(P)H and / or FAD. In some cases, the wavelengths are at least 340 nm, at least 345 nm, at least 350 nm, at least 355 nm, at least 360 nm, at least 365 nm, or at least 370 nm. In some cases, the wavelengths are up to 415 nm, up to 410 nm, up to 405 nm, up to 400 nm, up to 395 nm, up to 390 nm, up to 385 nm, or up to 380 nm. In some cases, the wavelengths are between 360 nm and 415 nm, between 350 nm and 410 nm, or between 370 nm and 380 nm. In some cases, the wavelengths are 375 nm. In some cases, the wavelengths are two or three times these wavelength values ​​(i.e., half or one-third the frequency). It should be understood that pulsed light sources inherently have some bandwidth and are therefore never strictly monochromatic. Thus, when "wavelength" is used herein, it refers to either the wavelength of the peak intensity or the weighted average wavelength. In some cases, the pulsed light source 524 is a UV pulsed diode laser. In some cases, the pulsed light source has a wavelength that is twice the peak absorption wavelength of NAD(P)H and / or FAD with an ultrashort pulse duration, such that fluorescence excitation is achieved through a two-photon excitation event, as will be understood by those skilled in the art of optics. The photon-counting detector 526 may be any detector capable of suitably detecting single photons and providing an analog or digital output representative of the detected photons. Examples of photon-counting detectors 526 include, but are not limited to, photomultiplier tubes, photodiodes, avalanche photodiodes, single-photon avalanche diodes (SPADs), charge-coupled devices, combinations thereof, and the like.

[0035] The photon counting electronics 528 can include electronics that one skilled in the art would understand to be suitable for use with the single photon detector 526 to generate the data sets described herein. Examples of suitable photon counting electronics 528 include, but are not limited to, field programmable gate arrays (FPGAs), dedicated digital signal processors (DSPs) with digitizers and time-to-digital converters, time-correlated single photon counting (TCSPC) electronics boards with time-to-amplitude and analog-to-digital converter electronics (such as those implemented by Becker&Hickl, Berlin, Germany), combinations of these, and the like. The autofluorescence spectrometer 510 is controlled by the processor 512 either directly (i.e., the processor 512 communicates with and receives signals directly from the spectrometer 510) or indirectly (i.e., the processor 512 communicates with a sub-controller specific to the spectrometer 510, and the signal from the spectrometer 510 may or may not be modified before being sent to the processor 512). The autofluorescence data set may be acquired by known spectroscopic methods. Fluorescence lifetime images may also be acquired by known imaging methods, and those acquired images may be used by the systems and methods described herein, as may be understood by those skilled in the art of spectroscopy. The device 500 may include various optical filters tuned to isolate the autofluorescence signal of interest. The optical filters may be tuned to the autofluorescence wavelengths of NAD(P)H and / or FAD.

[0036] The autofluorescence spectrometer 510 can be configured to acquire an autofluorescence data set from the electrical output of the detector 526 at a repetition rate that will be understood by one of ordinary skill in the art of spectroscopy to be appropriate to provide adequate sampling for observing the dynamics disclosed herein. In some cases, the repetition rate can be at least 1 kHz, at least 5 kHz, at least 10 kHz, at least 30 kHz, at least 50 kHz, at least 100 kHz, at least 500 kHz, at least 750 kHz, at least 1 MHz, at least 4 MHz, at least 7 MHz, at least 10 MHz, at least 15 MHz, at least 20 MHz, at least 50 MHz, at least 100 MHz, at least 500 MHz, or at least 1 GHz. In some cases, the repetition rate can be up to 1 THz, up to 800 GHz, up to 500 GHz, up to 250 GHz, up to 150 GHz, up to 100 GHz, up to 70 GHz, up to 50 GHz, up to 25 GHz, up to 15 GHz, up to 10 GHz, up to 6 GHz, up to 2 GHz, up to 1 GHz, up to 750 MHz, up to 500 MHz, up to 400 MHz, up to 250 MHz, up to 175 MHz, or up to 100 MHz. Although there may be drawbacks associated with oversampling, in principle the present disclosure can function with sampling rates as high as can be realized with existing technology. The repetition rates specified herein are based on the current state of the art at the time this disclosure was prepared and filed, and are not intended to be limiting in case future developments facilitate faster repetition rates.

[0037] The pulsed light source 524 can be configured to operate at a pulse repetition rate adapted to acquire the required fluorescence lifetime information. The maximum pulse repetition rate is limited by the fluorescence lifetime of the fluorochrome of interest. The fluorescence decay must subside completely by the time the next pulse of light is introduced to the sample to avoid ambiguity regarding the source of the data set (i.e., was this particular fluorescence photon initiated by the most recent excitation pulse of light or the one that preceded it?). The pulsed light source 524 can have a pulse repetition rate of up to 100 MHz, up to 80 MHz, up to 60 MHz, or up to 40 MHz. The lower limit of the pulse repetition rate is more practical in terms of reducing the overall sampling time, but in theory the data could be acquired very slowly for several reasons to do so. The device 500 may optionally include an optical microscope 520 for acquiring visual images of cells positioned in the observation zone 506 or elsewhere along the cell analysis path 502 .

[0038] The device 500 can optionally include a cell size measuring tool 522. The cell size measuring tool 522 can be any device capable of measuring the size of a cell, including but not limited to an optical microscope, such as the optical microscope 520. In some cases, the optical microscope and the cell size measuring tool 522 are the same subsystem. In some cases, the autofluorescence spectrometer 510 and the optical microscope 520 can be integrated into a single optical subsystem. In some cases, the autofluorescence spectrometer 510 and the cell size measurement tool 522 can be integrated into a single optical subsystem. Some aspects of the methods described herein can operate without utilizing cell size as an input to the convolutional neural network, but may be useful for measuring cell size for other purposes.

[0039] The processor 512 is in electronic communication with the spectrometer 510. The processor 512 is also in electronic communication with the optional cell sorter 508, the optional light microscope 520, and the optional cell size measurement tool 522, when present. The non-transitory computer readable medium 514 has instructions stored thereon that, when executed by the processor, cause the processor to perform at least a portion of the methods described herein. The equations into which the first and second phasor coordinates are input may also be stored on the non-transitory computer readable medium 514. The non-transitory computer readable medium 514 may be present in the device 500 or may be separate from the device, as long as it is accessible by the processor 512. The device 500 is substantially free of fluorescent labels (i.e., the cell analysis pathway 502 does not include an area for mixing the cells with the fluorescent labels). The device 500 is substantially free of fixation agents for binding and immobilizing the reprogramming intermediate cells. EXAMPLES

[0040] Example 1 Before describing this example in detail, it should be understood that many of the observations described herein are supported by data that can be provided to the examiner upon request. For the sake of brevity, much of the raw data and images have been omitted from this description, as these data and images do not provide a better understanding of the invention and merely support some of the statements made and conclusions drawn. Materials and Methods EPC isolation and cell culture EPCs were isolated from fresh peripheral human blood obtained from healthy donors (Interstate Blood Bank, Memphis, Tenn.). Blood was processed within 24 hours of collection, and hematopoietic progenitor cells were extracted from whole blood using negative selection (RosetteSep; STEMCELL Technologies) and cultured in polystyrene tissue culture plates in Erythroid Growth Medium (STEMCELL Technologies) for 10 days to expand EPCs. On day 10, expanded EPCs were examined by staining with APC anti-human CD71 antibody (334107; Biolegend; 1:100) and incubating for 1 hour at room temperature. Data were collected on an Attune Nxt flow cytometer and analyzed with FlowJo.

[0041] Micropattern design and PDMS stamp generation First, a template with the feature design was created in AutoCAD (Autodesk). The template was then sent to Advance Reproductions Corporation, MA to fabricate a photomask and a 6-inch patterned Si wafer was fabricated by the Microtechnology Core, University of Wisconsin-Madison, WI. Using soft photolithography techniques, the Si wafer was spin-coated with SU-8 negative photoresist (MICRO CHEM) and exposed to UV light. The Si mold was then developed in SU-8 developer (Sigma) for 45 minutes to obtain 150 μm tall features. The Si mold was then washed with acetone and isopropyl alcohol. The elastomeric stamp used in microcontact printing was made by standard soft lithography techniques. The silicon mold was made inert by exposing it to (tridecafluoro-1,1,2,2-tetrahydrooctyl)trichlorosilane vapor overnight. Polydimethylsiloxane (Sylgard 184 Silicone Elastomer Base, 3097366-1004, Dow Corning; PDMS) was prepared with a 1:10 curing agent ratio (Sylgard 184 Silicone Elastomer Curing Agent, 3097358-1004, Dow Corning) and degassed in vacuum for 30 minutes. PDMS was then poured onto the SU-8 silicon mold on a hot plate and baked overnight at 60°C to create the PDMS stamp.

[0042] Construction of μCP well plates A microcontact patterning (μCP) substrate was constructed based on previous work. Briefly, a polydimethylsiloxane (PDMS) stamp with 300 μm radius circular features was coated with Matrigel (WiCell Research Institute) for 24 h. After 24 h, the Matrigel-coated PDMS stamp was dried with N2 and placed on a 35 mm cell culture-treated ibiTreat dish (81156; Ibidi). A 50 g weight was added to the top of the PDMS stamp to ensure even pattern transfer from the Matrigel-coated PDMS stamp to the ibiTreat dish. The device was incubated at 37 °C for 2 h. A 35 mm ibiTreat dish was then backfilled with PLL(20 kDa)-g-(3.5)-PEG(2 kDa) (Susos), a grafted polymer solution with a 20 kDa PLL backbone and 2 kDa PEG side chains, with a grafting ratio of 3.5 (average PLL monomer units per PEG side chain), at 0.1 mg / mL in 10 mM HEPES buffer for 30 min at RT. The ibiTreat dish was then washed with PBS and sterilized by exposure to UV light for 15 min to obtain a micropatterned substrate.

[0043] Reprogramming On day 10, EPCs were electroporated with four episomal reprogramming plasmids encoding Oct4, shRNA knockdown of p53 (#27077; Addgene); Sox2, Klf4 (#27078; Addgene); L-Myc, Lin28 (#27080; Addgene); and miR302-367 cluster (#98748; Addgene) using the P3 Primary Cell 4D-Nucleofector kit (Lonza) and EO-100 program. Electroporated EPCs were plated on micropatterned substrates with erythroid growth medium (STEMCELL Technologies) at a seeding density of 2000k cells / dish. Cells were supplemented with ReproTeSR (STEMCELL Technologies) every other day starting on day 3 without removing all medium from the wells. On day 9, the medium was completely switched to ReproTeSR, and ReproTeSR medium was changed daily starting from day 10.

[0044] Isolation of iPSCs To isolate high-quality iPSC lines, candidate colonies were picked from the micropatterns using a 200 μL micropipette tip and transferred to Matrigel-coated polystyrene tissue culture plates in mTeSR1 medium (WiCell Research Institute). If additional purification was required, one additional manual picking step with a 200 μL micropipette tip was performed. During picking and subsequent passaging, the medium was often supplemented with the Rho kinase inhibitor Y-27632 (Sigma-Aldrich) at a 10 μM concentration to stimulate cell survival and establish clonal lineages. iPSCs derived from EPCs were maintained in mTeSR1 medium on Matrigel-coated polystyrene tissue culture plates and passaged every 3–5 days in ReLeSR (STEMCELL Technologies). All cells were maintained at 37 °C and 5% CO2.

[0045] Antibodies and staining All cells were fixed with 4% paraformaldehyde in PBS (Sigma-Aldrich) for 15 min, permeabilized with 0.5% Triton-X (Sigma-Aldrich) for >4 h at room temperature, and then stained. Nuclei were stained using Hoechst (H1399; Thermo Fisher Scientific, Waltham, MA) at 5 μg / mL with a 15 min incubation at room temperature. Primary antibodies were attached overnight at 4 °C in blocking buffer of 5% donkey serum (Sigma-Aldrich) at the following concentrations: anti-laminin (L9393; Sigma-Alrich) 1:500; TRA-1-60 (MAB4360; EMD Millipore, Burlington, MA) 1:100; Nanog (AF1997; R&D Systems) 1:200; CD71 (334107; Biolegend) 1:100. Secondary antibodies were obtained from Thermo Fisher Scientific and added at concentrations of 1:400–1:800 in blocking buffer with 5% donkey serum for 1 h at room temperature. A Nikon Eclipse Ti epifluorescence microscope was used to acquire single 10× images of each micropattern, and a Nikon AR1 confocal microscope was used to acquire 60× stitched images of each micropattern using a z-plane proximal to the micropatterned substrate for reprogramming studies.

[0046] Autofluorescence imaging of NAD(P)H and FAD Fluorescence lifetime imaging (FLIM) was performed at different time points during reprogramming with an Ultima two-photon microscope (Bruker) consisting of an ultrafast tunable excitation laser source (Insight DS+, Spectra-Physics) coupled to a Nikon Ti-E inverted microscope equipped with time-correlated single-photon counting electronics (SPC-150, Becker&Hickl). The laser source allows sequential excitation of NAD(P)H at 750 nm and FAD at 890 nm. NAD(P)H and FAD images were acquired through 440 / 80 nm and 550 / 100 nm bandpass filters (Chroma), respectively, using a gallium arsenide phosphide (GaAsP) photomultiplier tube (PMT; H7422, Hamamatsu). The laser power at the sample was approximately 3.5 mW for NAD(P)H and 6 mW for FAD. Lifetime imaging using time-correlated single-photon electronics (SPC-150, Becker & Hickl) was performed with Prairie View Atlas Mosaic Imaging (Bruker Fluorescence Microscopy) to acquire global μ-features. Fluorescence lifetime decays in 512 bins were acquired across a 512 × 512 pixel image with a pixel dwell time of 4.8 μs and a match period of 60 s. Photon counting rates were approximately 1–5 × 10 5 and was monitored during image acquisition to ensure that photobleaching did not occur. All samples were placed in a stage-op incubator and illuminated through a 40× / 1.15NA objective (Nikon). The short lifetime of red blood cell fluorescence at 890 nm was used as the instrument response function, with a full width at half maximum of 240 picoseconds. YG fluorescent beads (τ = 2.13 ± 0.03 ns, n = 6) were imaged daily as a fluorescence lifetime standard.

[0047] Image analysis The fluorescence lifetime decay was analyzed via SPCImage software (Becker & Hickl) to extract the fluorescence lifetime components. A threshold was used to eliminate pixels with low fluorescence signal (i.e., background). The fluorescence lifetime decay was deconvolved from the instrument response function and fitted to a biexponential decay model, I(t) = α1e-t / τ 1+α2e -t / τ The fluorescence decays of NAD(P)H and FAD were fitted to the equation: 2+C, where I(t) is the fluorescence intensity as a function of time t after the laser pulse, α1 and α2 are the fractional contributions of the short- and long-lived components, respectively (i.e., α1+α2=1), τ1 and τ2 are the short- and long-lived components, respectively, and C accounts for background light. Both NAD(P)H and FAD can exist in quenched (short-lived) and unquenched (long-lived) configurations; therefore, the fluorescence decays of NAD(P)H and FAD are fit to two components. Fluorescence intensity images were generated by integrating photon counts into the fluorescence decays at every pixel.

[0048] Images were analyzed at the single cell level and assessed for cellular heterogeneity. A pixel trainer was trained on 15 images using ilastik software (Berg, S. et al. ilastik: interactive machine learning for (bio)image analysis. Nat. Methods 16, 1226-1232 (2019)) to identify pixels within nuclei in the NAD(P)H images. An object classifier was then used to identify nuclei in the NAD(P)H images using a pixel classifier with the following parameters: Method=Simple, Threshold=0.3, Smooth=1, Size Filter Min=15 pixels, Size Filter Max=500 pixels. A customized CellProfiler (Carpenter, AE et al. CellProfiler: image analysis software for identifying and quantifying cell phenotypes. Genome Biol. 7, R100 (2006)) pipeline was then used to obtain metabolic and nuclear parameters. The CellProfiler pipeline applied the following steps: Primary objects (nuclei) were input from ilastik. Secondary objects (cells) were then identified in the NAD(P)H intensity images by outward propagation of the primary objects. The cytoplasmic mask was determined by subtracting the nuclear mask from the cell mask. The cytoplasmic mask was applied to all images to determine single-cell redox ratios and NAD(P)H and FAD lifetime parameters. A total of 11 metabolic parameters were analyzed for the cytoplasm of each cell: optical redox ratio, [NAD(P)H], NAD(P)Hα1, NAD(P)Hτ1, NAD(P)Hτ2, NAD(P)Hτ m , [FAD], FADα1, FADτ1, FADτ2, FADτ m A total of eight kernel parameters were analyzed for each kernel: Area, Perimeter, MeanRad, NSI, Robustness, Range, #Neigh, 1st Neigh.

[0049] The optical redox ratio (the fluorescence intensity of NAD(P)H divided by the combined intensity of NAD(P)H and FAD) and the mean fluorescence lifetimes (τm Representative images of α1τ1+α2τ2) were calculated using Fiji software. UMAP Clustering Clusters of cells across EPCs, IMs, and iPSCs were represented using Uniform Manifold Approximation and Projection (UMAP). UMAP dimensionality reduction (McInnes, L., Healy, J., Saul, N. & Grossberger, L. UMAP: Uniform Manifold Approximation and Projection. J. Open Source Softw. 3, 861 (2018)) was applied to the 11 OMI parameters (optical redox ratios, NAD(P)H τ m , τ1, τ2, α1, α2;FAD τ m Projections in 2D space were realized using R for all 8 kernel parameters (area, perimeter, MeanRad, NSI, robustness, extent, #Neigh, 1stNeigh). The following parameters were used for UMAP visualization: "n_neighbors": 20; "min_dist": 0.3, "metric": Jaccard, "n_components": 2.

[0050] z-score hierarchical clustering The z-score of each metabolic and nuclear parameter for each cell was calculated. z-score = (μ observed -μ row ) / σ row Wherein, μ observed is the average value of each parameter for each cell; μ row is the average value of each parameter across all cells, and σ row is the standard deviation of each parameter across all cells. Heatmaps of z-scores for all OMI variables were generated to visualize the differences in each parameter between different cells. Dendrograms show clustering based on the similarity of the average Euclidean distance across all variable z-scores. Heatmaps and associated dendrograms were generated in Python.

[0051] Classification method Random forest, simple logistic, k-nearest neighbor (IBk), and naive Bayes classifiers were trained to classify reprogrammed cells into EPCs, IMs, and iPSCs using Weka software (Holmes, G., Donkin, A. & Witten, IH WEKA: a machine learning workbench. in Proceedings of ANZIIS '94 - Australian New Zealand Intelligent Information Systems Conference 357-361 (IEEE, 1994). doi:10.1109 / ANZIIS.1994.396988). All data were randomly partitioned into training and testing data sets using 15-fold cross-validation with training and testing proportions of 93.3% (1994 cells) and 6.7% (143 cells), respectively. Each model was replicated 100 times; new training and testing data were generated before each iteration. Parameter weights for metabolic and nuclear parameters were extracted using the GainRatioAttributeEval function in Weka to determine the contribution of each variable to the trained classification model. One-to-many receiver operating characteristic (ROS) curves were generated to evaluate the classification model performance for classification of the test set data, which is the average of 100 iterations of randomly selected data from the training and test sets. All of the ROC curves shown were constructed from the test data set using models created from the training data set.

[0052] Karyotype analysis Cells cultured for at least five passages were grown to 60-80% confluency and shipped to WiCell Research Institute, Madison, WI for karyotyping. G-banded karyotyping was performed using standard cytogenetic protocols. Metaphase preparations were digitally acquired with Applied Spectral Imaging software and hardware. For each cell line, 20 GTL-banded metaphases were counted and a minimum of five were analyzed and karyotyped. Results were reported according to guidelines established by the International System for Cytogenetic Nomenclature 2016.

[0053] statistics p-values ​​were calculated using the non-parametric Kruskal-Wallis test for multiple unmatched comparisons in GraphPad Prism software. Statistical tests were considered significant at α≦0.05. Technical repeats are defined as distinct μ features in an experiment. Biological replicates are experiments performed on different donors. No a priori power calculations were performed.

[0054] result Establishing reprogramming on microcontact printed substrates. We first designed a microcontact printing (μCP) substrate to spatially control the adhesion of EPCs undergoing reprogramming. The μCP substrate is formed by coating 300 μm radius circular areas, called μ-features, on a 35 mm ibiTreat dish that allows cell adhesion. The remaining areas of the dish are then backfilled with a polycationic graft copolymer, PLL-g-PEG, which resists protein adsorption and prevents cell adhesion in these areas. The ibiTreat dish is made of a gas-permeable material, allowing for the maintenance of carbon dioxide or oxygen exchange during cell culture, and has high optical quality. These properties make the dish suitable for two-photon microscopy during reprogramming. To verify the stability of the circular areas coated with Matrigel, we immunostained for laminin, the main component of Matrigel. Fluorescence imaging showed that laminin was consistently present within the circular μ-features, indicating uniform patterning of Matrigel. We next assessed the ability of the μCP substrate to allow cell attachment by seeding two different cell types: human dermal fibroblasts (HDF) and H9 human embryonic stem cells (H9 ESC). We observed that both HDF and ESC remained viable, attached and were confined to the circular μ-features, indicating that the μCP substrate can spatially control cell attachment.

[0055] We then isolated peripheral blood mononuclear cells (PBMCs) from peripheral blood of healthy human donors and further enriched for EPCs. We examined enrichment of EPCs by flow cytometry with the erythroid cell surface marker CD71. Flow cytometry confirmed the presence of enriched EPCs by showing that >98% of cells expressed CD71 at day 10 of culture. To initiate reprogramming, we electroporated EPCs with four episomal reprogramming plasmids encoding Oct4, shRNA knockdown of P53, Sox2, Klf4, L-Myc, Lin28, and miR302-367 cluster; and plated them on μCP substrates. We assessed the ability of μCP substrates to sustain long-term reprogramming studies by performing high-content imaging to track individual μ features (>30 μ features per 35 mm dish) longitudinally at multiple time points during the reprogramming process over approximately 3 weeks. Day 22 was taken as the reprogramming end point, as at this time there were several iPSC colonies without significant overgrowth within the μ features, allowing analysis at single-cell resolution. While starting EPCs are non-adherent, reprogramming intermediate cells (IM) and end point iPSCs adhere to the circular μ features within the μCP substrate, indicating that the μCP substrate can support EPC reprogramming. Overall, the μCP platform provides unique spatial control over reprogramming cells and enables high-content quantitative imaging of reprogramming.

[0056] OMI reveals distinct metabolic changes during reprogramming. Metabolic state plays an important role in regulating iPSC reprogramming and pluripotency and can be monitored non-invasively via OMI. NADH is an electron donor and FAD is an electron acceptor, both of which are present in all cells as coenzymes and provide energy for metabolic reactions. For example, glycolysis in the cytoplasm generates NADH and pyruvate, while OXPHOS consumes NADH and produces FAD. Therefore, autofluorescence imaging of NADH and FAD can be dynamically responsive to the redox state of cells and is affected by many reactions. We followed the autofluorescence dynamics of NAD(P)H and FAD by performing OMI on μCP substrates at various time points during EPC reprogramming. In these images, nuclei remain dark as NAD(P)H is primarily located in the cytosol and mitochondria, and FAD is primarily located in mitochondria. NAD(P)H images were used as input to ilastik software (cited above) to identify nuclei. Identified nuclei were then used as input to a high-content CellProfiler software (cited above) pipeline to segment the cytoplasm and measure various metabolic and nuclear parameters. Overall, a total of 11 metabolic parameters (optical redox ratio, [NAD(P)H], NAD(P)Hα1, NAD(P)Hτ1, NAD(P)Hτ2, NAD(P)Hτ m , [FAD], FADα1, FADτ1, FADτ2, FADτ m ), and eight nuclear parameters (Molugu, K. et al. Tracking and Predicting Human Somatic Cell Reprogramming Using Nuclear Characteristics. Biophys. J. 118, 2086-2102 (2020)) (area, perimeter, MeanRad, NSI, robustness, extent, #Neigh, firstNeigh) were measured by the analysis pipeline. Immunofluorescence labeling further validated the cell type at these different time points, namely EPC (CD71 + , Nanog - ), IM(CD71 - , Nanog - ), and iPSCs (CD71 - , Nanog + ) was examined. NAD(P)H and FAD autofluorescence imaging revealed metabolic differences between initiating EPCs, intermediates (IMs), and iPSCs.

[0057] We observed a significant increase in the optical redox ratio (iPSC>IM>EPC) during the reprogramming process, indicating that EPCs are more oxidized than IMs and iPSCs. Furthermore, we noticed that patterned IMs and iPSCs had significantly higher optical redox ratios compared to their non-patterned counterparts. This observation is consistent with previous studies showing that mechanical cues can regulate their relative use of glycolysis and may require further investigation. We then observed that the NAD(P)H and FAD lifetime components undergo biphasic changes during the process of reprogramming. The FAD lifetime component undergoes more significant changes relative to the NAD(P)H component. The proportion of protein-bound FAD (FADα1) undergoes a decrease from EPCs to IMs and then increases from IMs to iPSCs, which may reflect an OXPHOS burst. FADτ m is inversely proportional to FADα1, and therefore FADτ m Similar biphasic changes occur in nuclear parameters during reprogramming, consistent with our previous studies.

[0058] We noticed that H9 ESCs had similar metabolic and nuclear parameters to iPSCs, as expected. Fibroblasts had significantly different metabolic parameters than EPCs (Fig. 1, Supplementary Figure 2). This could be because 1) fibroblasts adhere but initial EPCs do not, and 2) fibroblasts and EPCs have different proliferation rates and energy requirements. Taken together, autofluorescence imaging of NAD(P)H and FAD showed striking changes during reprogramming. OMI enables sorting of reprogrammed cells with high accuracy.

[0059] Uniform Manifold Approximation and Projection (UMAP) (cited above), a dimensionality reduction technique, was used to visualize how cells cluster in 2D space from metabolic and nuclear parameters. Neighborhoods were defined through the Jaccard similarity coefficient calculated across metabolic and nuclear parameters. UMAP was chosen over t-distributed stochastic neighbor embedding (t-SNE) because UMAP separates EPCs, IMs, and iPSCs better than t-SNE. Furthermore, UMAP has faster speed and the ability to include non-metric distance functions, preserving the global structure of the data. We used UMAP to visualize how cells cluster exclusively from 11 metabolic parameters and exclusively from 8 nuclear parameters. Although these UMAP displays revealed separation of EPCs, IMs, and iPSCs; UMAPs created using both metabolic and nuclear parameters provided a cleaner separation between EPCs, IMs, and iPSCs. We also plotted heat map displays of the z-scores of metabolic and nuclear parameters at the donor level to examine donor heterogeneity. Taken together, the clustering of 11 metabolic and 8 nuclear parameters by using UMAP, as well as the z-score heat map clustering, showed that EPCs, IMs, and iPSCs could be differentiated based on these parameters.

[0060] Classification models were then developed based on 11 metabolic and 8 nuclear parameters to predict the reprogramming status of cells, i.e., EPCs, IMs, or iPECs. Supervised machine learning classification (Naive Bayes, K-nearest neighbors) and regression algorithms (Logistic regression, and Random Forest - see Amancio, DR et al. A Systematic Comparison of Supervised Classifiers. PLoS ONE 9, e94137 (2014)) were implemented to test the prediction accuracy for iPSCs when all metabolic and nuclear parameters are used. To protect against overfitting, the classification models were trained using 15-fold cross-validation on single-cell data from three different donors with assigned reprogramming status with morphological characteristics, and tested on the same cell CD71 and Nanog staining validation data from three donors (fully independent and non-overlapping observations). One-to-many receiver operating characteristic (ROC) curves of the test data revealed the highest classification accuracy for predicting iPSCs (area under the curve AUC=0.993), IMs (AUC=0.993), and EPCs (AUC=0.999) when the random forest classification model was used. Therefore, we used the random forest classification model for further analysis in this study.

[0061] The gain ratio analysis was performed for the FAD lifetime components FADα1, FADτ1, and FADτ m We found that τ is the most important parameter for classifying the reprogramming status of cells. This is consistent with the observation that the FAD lifespan components are significantly different among EPCs, IMs, and iPSCs. We then plotted the precision score (selected based on the value of the gain ratio for the random forest classifier) ​​as a function of the number of parameters used for classification. This plot revealed that the precision score increases with the number of parameters up to 8 parameters, after which it plateaus. We further found that τ using only the FAD lifespan variables (FAD τ mWe noticed that high classification accuracy could be achieved in predicting iPSCs (area under the curve AUC=0.944), IMs (AUC=0.968), and EPCs (AUC=0.998; collected only in the FAD channel). Using only the FAD lifetime parameters, we ensured a minimum imaging time of 2.5 mm per μ feature, with no additional dependency on intensity parameters associated with even higher variability due to confounding factors of intensity levels (throughput, detector gain, and inner filter effects due to laser power). Thus, the FAD lifetime parameters alone are sufficient to predict the reprogramming status of cells.

[0062] Pseudotemporal ordering of single cells reveals heterogeneous cell populations. To study the heterogeneity of reprogramming μ features, we used 11 metabolic and 8 nuclear parameters to construct pseudotemporal single-cell trajectories of cell reprogramming using the Monocle3 program (see Trapnell, C. et al. The dynamics and regulators of cell fate decisions are revealed by pseudotemporal ordering of single cells. Nat. Biotechnol. advance online publication, (2014) and Qiu, X. et al. Reversed graph embedding resolves complex single-cell trajectories. Nat. Methods 14, 979-982 (2017)), a trajectory estimation method that learns the combinatorial changes that each cell must go through as part of the reprogramming process and subsequently places each cell in its correct place in the trajectory. The trajectories constructed using this method were distributed across 10 clusters and consisted of EPCs, IMs, and iPSCs, with four branching events, and disconnected branches (Figures 6-8). Notably, the pseudotime trajectory, colored by the actual reprogramming time points, showed that the pseudotime progressed along the timeline of the actual reprogramming (Figure 7). The initiating EPCs were heterogeneous and occupied three clusters (clusters: 1, 2, 3). Cluster 2 consisted of the initiating EPCs undergoing reprogramming, whereas clusters 1 and 3, which constituted branches separate from EPCs, could not readily permit reprogramming. iPSCs, regardless of the reprogramming time point, mainly occupied two clusters (clusters: 7, 10), whereas IMs belonged to several clusters (clusters: 4, 5, 6, 8, 9), of which clusters 6 and 8 were concentrated in unsuccessful reprogramming branches (Figure 3C). Overall, the single-cell trajectory maps showed the heterogeneity of reprogramming, i.e., cells going right at branch points 1, 2, and 3 (Figures 7, 8) were fully reprogrammed into iPSCs within 25 days of reprogramming initiation, whereas cells going left at branch points 1 and 3 (Figures 7, 8) remained in an intermediate stage.

[0063] Subsequent heatmap analysis of the clusters in the single-cell reprogramming trajectory map reveals that the clusters show correlation patterns based on their reprogramming status, i.e., EPCs (cluster: 2) are highly correlated with early IMs (cluster: 4), while late IMs (cluster: 5, 6, 8, 9) demonstrate high correlation with iPSCs (cluster: 7). When we compared IMs undergoing reprogramming (cluster: 9) and IMs not reprogrammed to iPSCs (cluster: 6), we noticed differences in their NAD(P)H lifetime components, indicating that these parameters may play a role in determining the fate of reprogrammed cells. To further examine the parameters that differentiated the cell clusters, we performed spatial correlation analysis using the statistical Moran's I, which tells us whether cells located close on the trajectory will have similar (or dissimilar) expression levels for the parameters tested. When the parameters were ranked by Moran's I, the FAD lifetime parameters (FAD τ1, τ2, τ m) were found to be the most important in distinguishing the clusters, followed by the NAD(P)H lifetime parameters (NAD(P)H τ2, α1, τ1). This result is consistent with the high gain ratio values ​​for the FAD lifetime parameters, and the observed FAD lifetime parameters differ significantly among EPCs, IMs, and iPSCs.

[0064] Overall, FAD parameters are important to distinguish EPCs, IMS, and iPSCs; whereas NAD(P)H parameters are key to determine the final reprogramming fate of cells. When we plotted the identified key metabolic parameters as a function of pseudotime, we observed that they undergo biphasic changes during reprogramming, which may represent OXPHOS bursts. These pseudotime trajectory plots provided increased temporal resolution compared to our previous plots. Taken together, our reprogramming trajectory analysis provided insights into the heterogeneity of reprogramming at high temporal and single-cell resolution.

[0065] Isolation of high-quality iPSCs. The ultimate goal of any reprogramming platform is to successfully isolate iPSCs that can be used in downstream applications. Using a combination of OMI, the μCP platform, and the machine learning model developed in this study, we were able to successfully isolate high-quality iPSCs. First, we used OMI to track metabolic and nuclear parameters of μ features throughout the reprogramming process. Second, we used our random forest classification model to predict the reprogramming status of tracked μ features. Third, we estimated pseudotime during the reprogramming process to monitor the progress of μ features along the reprogramming trajectory. Finally, we performed immunostaining on μ features, and showed that the predictions of reprogramming status made by the machine learning model correlated well with actual staining. We then isolated iPSCs from the μCP platform based on the predictions made by the random forest classification model. The physical separation of the micropatterns relative to each other resulted in easy picking and isolation of fully reprogrammed iPSCs, combined with a high percentage of predicted iPSC cells, even up to 100%, across the entire μ feature. We further confirmed that the isolated iPSCs expressed pluripotency markers and did not show genomic abnormalities, indicating that our reprogramming platform can be used to generate genetically stable iPSC lines.

[0066] Consideration Here we report a noninvasive, high-throughput, quantitative and label-free imaging platform for predicting the reprogramming outcome of EPCs by combining micropatterning, live-cell autofluorescence imaging, and automated machine learning. Using a random forest classification model with 11 metabolic and 8 nuclear parameters, we are able to predict the reprogramming status of EPCs at any time point during reprogramming with a prediction accuracy of about 95% and a model performance of about 0.99 (AUC of ROC). Furthermore, we provide a single-cell roadmap of EPC reprogramming, which reveals the diverse cell fate trajectories of individual reprogrammed cells (Figures 6-8).

[0067] Recent evidence indicates that metabolic changes during reprogramming include a decrease in OXPHOS and an increase in glycolysis, together with a transient hyperactive metabolic state called the OXPHOS burst. This OXPHOS burst occurs early in reprogramming, exhibits characteristics of both high OXPHOS and high glycolysis, and may be a regulatory trigger for a global shift in reprogramming. These changes are accompanied by changes in the amounts of corresponding metabolites and have been confirmed by genome-wide analysis of gene expression, protein levels, and metabolomic profiling. Shifts in cellular metabolism affect enzymes that control epigenetic organization, which may affect chromatin reorganization and provide the basis for changes in nuclear morphology as well as gene expression during reprogramming. Consistent with these studies, the redox ratio increases during reprogramming, which may indicate an increase in glycolysis during reprogramming.

[0068] The changes in NAD(P)H and FAD lifetime parameters that occur during reprogramming may reflect changes in quencher concentrations such as oxygen, tyrosine, or tryptophan, or changes in local temperature and pH. In particular, biphasic changes in metabolic and nuclear parameters may be temporarily increased due to increased generation of ROS by mitochondria during OXPHOS bursts. The generated ROS further acts as a signal to activate nuclear factor (erythroid-derived 2)-like 2 (NRF-2), which then induces hypoxia-inducible factor (HIF) to promote glycolysis during reprogramming by increasing the expression levels of glycolysis-related genes. Furthermore, the importance of FAD parameters for distinguishing various reprogramming cell types can point to significant changes in the mitochondrial environment during reprogramming. Differences in NAD(P)H lifetime parameters between IMS that have successfully undergone reprogramming and those that have not may suggest a role for NAD(P)H in influencing reprogramming barriers, requiring further investigation.

[0069] Classification analysis revealed that models trained on all 11 metabolic and 8 nuclear parameters yielded the highest accuracy in classifying the reprogramming status of cells. Random forest classification using only FAD lifetime parameters yielded relatively high ROC AUC values. Furthermore, FAD lifetime parameters were more accurate in predicting reprogramming status than using only nuclear parameters, which can be obtained using wide-field or confocal fluorescence microscopy. Imaging only FAD lifetime parameters instead of imaging all parameters significantly reduced imaging time from 7 min to 2.5 min per μfeature. This is especially useful when a large number of μfeatures are being attempted to assess iPSC quality at a manufacturing scale, eliminating the variability associated with intensity measurements. Our single-cell reprogramming trajectory maps (Figures 6-8), constructed based on metabolic and nuclear parameters, could show that the reprogramming process proceeds by a combination of selective and stochastic models. There seems to be a certain percentage of initiating EPCs that are robust to reprogramming, supporting the selective model of reprogramming, but there is also a certain percentage of intermediate cells at various stages of reprogramming that are not fully reprogrammed into iPSCs, supporting the stochastic model of reprogramming.

[0070] Many of the current studies to understand the heterogeneity during reprogramming rely on bulk or single cell analysis techniques. Bulk samples obscure the variability in both the starting cell population and during fate conversion due to variable kinetics and low efficiency of reprogramming; however, single cell techniques destroy the cell's microenvironment, resulting in significant changes in the biophysical properties of cells undergoing reprogramming. Our reprogramming platform overcomes these challenges by using a combination of the μCP platform, OMI, and Monocle algorithm. First, the μCP platform ensures an intact microenvironment for reprogramming cells, which also allows analysis at the single cell level. Second, OMI has several spatial and temporal resolution advantages compared to traditional assays that allow further insight into the heterogeneity of reprogramming. OMI can be performed with high resolution to allow measurements at the single cell level, is non-destructive, allows measurements of the spatial integrity of adjacent cells, and also has high temporal resolution that allows for time-course studies of reprogramming. Finally, pseudo-time trajectories using the Monocle algorithm overcome the problem of reprogramming trajectories constructed based on absolute time points, which ignore the asynchrony of the reprogramming process. Overall, these methods can help identify somatic cells or early reprogramming cells that are robust to reprogramming, and thus can be used to increase the success rate of iPSC generation from patient-derived primary cells or cell lines.

[0071] Our reprogramming platform can contribute to the long-term commercial success of iPSC-derived therapies by moving the iPSC manufacturing process from labor-intensive, time-consuming, error-prone, high-risk laboratory bench protocols to industrial-scale GMP-compliant manufacturing systems. First, the μCP platform involves direct ECM printing onto optically transparent substrates, without any gold coating as in traditional microcontact printing methods, making the process cost-effective and even simpler by eliminating the need for cleanroom use. Using this μCP platform, we can also isolate fully pluripotent iPSCs without any genomic abnormalities, showing that the platform is adaptable for biomanufacturing GMP-grade iPSCs. Second, we used erythroid progenitors isolated from peripheral blood as the starting cell type for reprogramming due to their lack of genomic rearrangements and demonstrated reprogramming ability. Furthermore, blood collection is a minimally invasive procedure, and the collected cells are naturally replaced as the tissue self-renews, making it suitable for iPSC generation. Third, we used a combination of the aforementioned oriP / EBNA-based virus-free and non-integrated episomal reprogramming plasmids, which avoids the safety concerns surrounding the use of integrated viral vectors. Furthermore, these episomes are lost at about 5% per cell generation due to defects in plasmid synthesis and partitioning, and thus plasmid-free iPSCs can be easily isolated for clinical application. We also used xeno-free components for feeder-free reprogramming and maintenance of iPSCs to eliminate inconsistencies arising from the undefined nature of xeno-components and ensure GMP compliance. Finally, autofluorescence imaging techniques are label-free, unlike other methods of studying metabolism such as electron microscopy, immunocytochemistry, and colorimetric metabolic assays. They also allow non-destructive real-time monitoring of live cells with lower sample phototoxicity compared to single-photon excitation.Taken together, the processes of μCP platform fabrication, reprogramming, autofluorescence imaging, iPSC identification based on machine learning models, and iPSC isolation can all be automated and expanded to different reprogramming methods such as mRNA, Sendai virus; to other starting cell types such as fibroblasts, keratinocytes; to other parameters (cell morphology and mitochondrial structure), and to other processes such as differentiation; making it an attractive platform for the biomanufacturing of iPSCs and iPSC-derived cells on an industrial scale.

[0072] Overall, we have developed a high-throughput, non-invasive, rapid and quantitative method to predict cellular reprogramming status and study reprogramming heterogeneity. Our work shows that OMI can predict cellular reprogramming status and enable real-time monitoring during iPSC manufacturing, thereby helping to identify high-quality iPSCs in a timely and cost-effective manner. Similar techniques may impact other areas of cell manufacturing, such as direct reprogramming, differentiation, and cell lineage development, thus contributing to the rapid development of regenerative medicine and precision medicine applications from bench to bedside.

Claims

1. a cell analysis observation zone adapted to receive reprogramming intermediate cells and present said reprogramming intermediate cells for individual autofluorescence interrogation; an autofluorescence spectrometer configured to acquire an autofluorescence data set regarding the reprogramming intermediate cells positioned in the cell analysis observation zone, the autofluorescence spectrometer including a light source, a photon-counting detector, and photon-counting electronics; a processor in electronic communication with the autofluorescence spectrometer; a non-transitory computer-readable medium accessible to said processor and having instructions stored thereon; wherein the instructions, when executed by the processor, cause the processor to: a) receiving the autofluorescence data set; and b) determining a current reprogramming status of the reprogramming intermediate cell based on a current reprogramming prediction, wherein the current reprogramming prediction is calculated using at least a portion of the autofluorescence dataset, wherein the current reprogramming prediction is calculated using at least one metabolic endpoint and optionally at least one nuclear parameter of the autofluorescence dataset as input, wherein the at least one metabolic endpoint is a flavin adenine dinucleotide (FAD) mean fluorescence lifetime (τ m ), FAD shortest life amplitude component (α 1 ), FAD shortest fluorescence lifetime component (τ 1 ), the FAD longest fluorescence lifetime component (τ 2 ), or a combination thereof, Somatic cell reprogramming tracking device.

2. The somatic cell reprogramming tracking device of claim 1, wherein the instructions, when executed by the processor, cause the processor to: c) identify a future reprogramming status of the reprogramming intermediate cell based on the pseudotime trajectory of cell reprogramming based on machine learning analysis of an autofluorescence dataset acquired for a reprogramming intermediate cell having a known reprogramming status over the course of the pseudotime trajectory.

3. The somatic cell reprogramming tracking device of claim 2, wherein the instructions, when executed by the processor, cause the processor to: c) identify the future reprogramming status of the reprogramming intermediate cell based on the current reprogramming prediction and the pseudotime trajectory of cell reprogramming.

4. The somatic cell reprogramming tracking device of any one of claims 1 to 3, wherein the device further comprises a cell analysis platform adapted to receive a cell culture containing the reprogramming intermediate cells of the subject, and the cell analysis platform is adapted to position the cell culture containing the reprogramming intermediate cells of the subject in the observation zone.

5. 4. The somatic cell reprogramming tracking device of claim 1, wherein the device further comprises a cell analysis pathway further comprising: (i) an inlet; (ii) an observation zone connected to the inlet downstream of the inlet and configured to present reprogramming intermediate cells for individual autofluorescence interrogation; and (iii) an outlet connected to the observation zone downstream of the observation zone.

6. The somatic cell reprogramming tracking device of claim 5 , wherein the cell analysis pathway comprises a microfluidic pathway or a nanofluidic pathway.

7. The somatic cell reprogramming tracking device of claim 5 , wherein the somatic cell reprogramming tracking device further comprises a flow regulator coupled to the inlet.

8. 6. The somatic cell reprogramming tracking device of claim 5, wherein the flow regulator, when positioned in the observation zone, is configured to provide a flow of cells through the observation zone at a rate that enables the autofluorescence data set regarding the reprogramming intermediate cells to be acquired by the autofluorescence spectrometer.

9. The somatic cell reprogramming tracking device of claim 5 , wherein the cell analysis pathway does not include a fluorescent label for binding to the reprogramming intermediate cells.

10. The somatic cell reprogramming tracking device of claim 5 , wherein the cell analysis pathway does not include an immobilization agent for binding and immobilizing the reprogramming intermediate cells.

11. 6. The somatic cell reprogramming tracking device of claim 5, further comprising a cell sorter having a sorter inlet and at least two sorter outlets, the cell sorter coupled to the cell analysis pathway via the outlet downstream of the observation zone, the cell sorter configured to selectively direct cells from the sorter inlet to one of the at least two sorter outlets based on a sorting signal, the processor in electronic communication with the cell sorter, and the instructions, when executed by the processor, further cause the processor to provide the sorting signal to the cell sorter based on the current reprogramming prediction.

12. The somatic cell reprogramming tracking device of claim 1 , wherein the somatic cell reprogramming tracking device further comprises a cell picking device.

13. The somatic cell reprogramming tracking device of claim 12 , wherein the cell picking apparatus automatically picks cells based on the current reprogramming prediction.

14. 14. The somatic cell reprogramming tracking device of claim 13, wherein the processor and the physical dimensions and flow rate of the cell analysis pathway are adapted to provide the sorting signal to the cell sorter before the reprogramming intermediate cells reach the cell sorter.

15. 2. The somatic cell reprogramming tracking device of claim 1, wherein the light source emits light having a wavelength adjusted to excite fluorescence from reduced nicotinamide adenine dinucleotide and / or reduced nicotinamide dinucleotide phosphate (NAD(P)H) and / or FAD.

16. 16. The somatic cell reprogramming tracking device of claim 15, wherein the wavelength is at least 340 nm, at least 345 nm, at least 350 nm, at least 355 nm, at least 360 nm, at least 365 nm, or at least 370 nm, at most 415 nm, at most 410 nm, at most 405 nm, at most 400 nm, at most 395 nm, at most 390 nm, at most 385 nm, or at most 380 nm, or the wavelength is between 340 nm and 415 nm, between 350 nm and 410 nm, or between 370 nm and 380 nm, or the wavelength is 375 nm.

17. the light source is a pulsed light source, and the pulsed light source has a pulse duration of at least 1 femtosecond, at least 5 femtoseconds, at least 10 femtoseconds, at least 25 femtoseconds, at least 50 femtoseconds, at least 100 femtoseconds, at least 200 femtoseconds, at least 350 femtoseconds, at least 500 femtoseconds, at least 750 femtoseconds, at least 1 picosecond, at least 3 picoseconds, at least 5 picoseconds, at least 10 picoseconds, at least 20 picoseconds, at least 50 picoseconds, or at least 10. The somatic cell reprogramming tracking device of claim 1, wherein the somatic cell reprogramming tracking device emits light having a full width at half maximum pulse width of at most 100 picoseconds, and up to 1 nanosecond, up to 900 picoseconds, up to 750 picoseconds, up to 600 picoseconds, up to 500 picoseconds, up to 400 picoseconds, up to 250 picoseconds, up to 175 picoseconds, up to 100 picoseconds, up to 75 picoseconds, up to 60 picoseconds, up to 50 picoseconds, up to 35 picoseconds, up to 25 picoseconds, up to 20 picoseconds, up to 15 picoseconds, up to 10 picoseconds, or up to 1 picosecond.

18. 10. The somatic cell reprogramming tracking device of claim 1, wherein the light source is a diode laser.

19. The autofluorescence spectrometer may be configured to operate at frequencies of at least 1 kHz, at least 5 kHz, at least 10 kHz, at least 30 kHz, at least 50 kHz, at least 100 kHz, at least 500 kHz, at least 750 kHz, at least 1 MHz, at least 4 MHz, at least 7 MHz, at least 10 MHz, at least 15 MHz, at least 20 MHz, at least 50 MHz, at least 100 MHz, at least 500 MHz, or at least 1 GHz, and up to 10 THz, up to 1 THz, up to 8 THz, 2. The somatic cell reprogramming tracking device of claim 1, wherein the device is configured to acquire the autofluorescence dataset at a repetition rate of up to 700 MHz, up to 500 GHz, up to 250 GHz, up to 150 GHz, up to 100 GHz, up to 70 GHz, up to 50 GHz, up to 25 GHz, up to 15 GHz, up to 10 GHz, up to 6 GHz, up to 2 GHz, up to 1 GHz, up to 750 MHz, up to 500 MHz, up to 400 MHz, up to 250 MHz, up to 175 MHz, or up to 100 MHz.

20. 10. The somatic cell reprogramming tracking device of claim 1, wherein the photon counting detector is a photomultiplier tube, a photodiode, an avalanche photodiode, a single-photon avalanche diode, a charge-coupled device, or a combination thereof.

21. 2. The somatic cell reprogramming tracking device of claim 1, wherein the photon counting electronics comprises a field programmable gate array, a dedicated digital signal processor with digitizer and time-to-digital converter, a time-correlated single photon counting electronics board with time-to-amplitude and analog-to-digital converter electronics, or a combination thereof.

22. 10. The somatic cell reprogramming tracking device of claim 1, wherein the autofluorescence spectrometer comprises a detector side filter configured to transmit the fluorescent signal of the subject.

23. 23. The somatic cell reprogramming tracking device of claim 22, wherein the detector side filter is configured to transmit NAD(P)H fluorescence and / or FAD fluorescence.

24. 10. The somatic cell reprogramming tracking device of claim 1, wherein the somatic cell reprogramming tracking device further comprises a cell size measurement tool configured to measure cell size and to communicate the cell size to the processor.

25. 25. The somatic cell reprogramming tracking device of claim 24, wherein said cell size measuring tool is a microscope.

26. 2. The somatic cell reprogramming tracking device of claim 1, wherein the somatic cell reprogramming tracking device further comprises a cell imager configured to acquire images of cells positioned within the observation zone and to communicate the images to the processor.

27. 10. The somatic cell reprogramming tracking device of claim 1, wherein the autofluorescence spectrometer is adapted to measure the autofluorescence of reprogramming intermediate cells without requiring the use of fluorescent labels.

28. The somatic cell reprogramming tracking device of claim 1 , wherein the instructions, when executed by the processor, further cause the processor to generate a report including the current reprogramming prediction for the reprogramming intermediate cells analyzed by the device.

29. 1. A method for characterizing the progression of somatic cell reprogramming, comprising: a) optionally receiving a population of reprogramming intermediate cells with unknown reprogramming status; b) acquiring an autofluorescence data set from a reprogramming intermediate cell of said population of reprogramming intermediate cells; and c) determining the current reprogramming status of said reprogramming intermediate cells based on the current reprogramming prediction; wherein the current reprogramming prediction is calculated using at least a portion of the autofluorescence dataset, and the current reprogramming prediction is calculated using at least one metabolic endpoint and optionally at least one nuclear parameter of the autofluorescence dataset as input, wherein the at least one metabolic endpoint is a flavin adenine dinucleotide (FAD) mean fluorescence lifetime (τ m ), FAD shortest life amplitude component (α 1 ), FAD shortest fluorescence lifetime component (τ 1 ), the FAD longest fluorescence lifetime component (τ 2 ), or a combination thereof, method.

30. The method comprises: d) identifying future reprogramming states of the reprogramming intermediate cells based on a pseudotemporal reprogramming pathway map based on machine learning analysis of autofluorescence datasets acquired for reprogramming intermediate cells with known reprogramming states over the course of the pseudotemporal trajectory.

30. The method of claim 29, further comprising:

31. 31. The method of claim 30, wherein said identifying said future reprogramming situation is further based on said current reprogramming prediction.

32. 1. A method for classifying somatic cell reprogramming, comprising: a) receiving a population of reprogramming intermediate cells having an unknown reprogramming status; b) acquiring an autofluorescence data set for each reprogramming intermediate cell of the population of reprogramming intermediate cells, each autofluorescence data set comprising autofluorescence lifetime information; and c1) physically isolating a first portion of the population of reprogramming intermediate cells from a second portion of the population of reprogramming intermediate cells based on a current reprogramming prediction, wherein each reprogramming intermediate cell of the population of reprogramming intermediate cells is placed in the first portion when the current reprogramming prediction exceeds a predetermined threshold, and is placed in the second portion when the current reprogramming prediction is less than or equal to the predetermined threshold; or c2) generating a report including said current reprogramming prediction, said report optionally identifying the proportion of said population of reprogramming intermediate cells that have said current reprogramming prediction above said predetermined threshold. including any of the following: The current reprogramming prediction is calculated using at least one metabolic endpoint and optionally at least one nuclear parameter of the autofluorescence dataset as input, wherein the at least one metabolic endpoint comprises a flavin adenine dinucleotide (FAD) mean fluorescence lifetime (τ), a FAD shortest lifetime amplitude component (α), a FAD shortest fluorescence lifetime component (τ), a FAD longest fluorescence lifetime component (τ), or a combination thereof. method.

33. d) identifying future reprogramming states of said reprogramming intermediate cells based on a pseudotemporal reprogramming pathway map based on machine learning analysis of autofluorescence datasets acquired for reprogramming intermediate cells with known reprogramming states over the course of their pseudotemporal trajectories; 33. The method of claim 32, further comprising:

34. 34. The method of claim 33, wherein said identifying said future reprogramming situation is further based on said current reprogramming prediction.

35. 30. The method of claim 29, wherein said method does not include the use of a fluorescent label to bind to said reprogramming intermediate cells.

36. 30. The method of claim 29, wherein the method does not include immobilizing the reprogramming intermediate cells.

37. Method according to claim 29, wherein said method is adapted to use the device and / or one or more features of said device of claims 1 to 3.

38. 1. A method of administering reprogrammed somatic cells to a subject in need thereof, comprising: e) the method of claim 32, comprising step c1); and f) introducing said first portion of said population of reprogramming intermediate cells into said subject. A method comprising:

39. 39. The method of claim 38, wherein the first portion of the population of reprogramming intermediate cells is modified prior to step f).

40. 1. A method of administering reprogrammed somatic cells to a subject in need thereof, comprising: e) the method of claim 32, comprising step c2); and f) introducing the population of reprogramming intermediate cells into the subject in response to a rate exceeding a second predetermined threshold. A method comprising:

41. 41. The method of claim 40, wherein the population of reprogramming intermediate cells is modified prior to step f).

42. 1. A method for generating a pseudotemporal reprogramming pathway map, comprising: a) receiving autofluorescence data sets and optionally nuclear data sets for a plurality of reprogramming intermediate cells, the autofluorescence data sets corresponding to pseudo-time points along a pseudo-timeline of reprogramming; b) constructing a pseudo-time single-cell trajectory associated with each of the plurality of reprogramming intermediate cells based on the received autofluorescence data sets associated with each of the plurality of reprogramming intermediate cells, wherein each of the pseudo-time single-cell trajectories includes a current reprogramming prediction associated with each of the plurality of reprogramming intermediate cells at a respective predetermined pseudo-time point, wherein the current reprogramming prediction is calculated using at least a portion of the autofluorescence data sets and optionally at least a portion of the nuclear data sets, and wherein the current reprogramming prediction is calculated using as input at least one metabolic endpoint of at least one of the autofluorescence data sets and optionally at least one nuclear parameter of at least one of the nuclear data sets, and wherein the at least one metabolic endpoint is determined based on a flavin adenine dinucleotide (FAD) mean fluorescence lifetime (τ m ), FAD shortest life amplitude component (α 1 ), FAD shortest fluorescence lifetime component (τ 1 ), nicotinamide adenine dinucleotide and / or reduced nicotinamide dinucleotide phosphate adenine dinucleotide (NAD(P)H) shortest lifetime amplitude component (α 1 ), the NAD(P)H shortest fluorescence lifetime component (τ 1 ), NAD(P)H longest fluorescence lifetime component (τ 2 ), or a combination thereof; c) compiling the constructed pseudo-time single-cell trajectories into a single compiled data set; d) identifying clusters, branching events, and / or disconnected branches within the single compiled data set; and e) identifying correlations between the clusters, branching events, and / or disconnected branches, and current or future reprogramming situations within the single compiled data set, thereby generating the pseudo-temporal reprogramming pathway map used to predict future reprogramming based on the correlations. A method comprising:

43. 44. The method of claim 30 or 33, wherein the pseudotemporal reprogramming pathway map is provided by the method of claim 42.

44. A somatic cell reprogramming tracking device described in any one of claims 1 to 3, wherein the autofluorescence data set is adjusted to wavelengths corresponding to NAD(P)H fluorescence and / or FAD fluorescence.

45. A somatic cell reprogramming tracking device as described in any one of claims 1 to 3, wherein the at least one metabolic endpoint comprises NAD(P)H autofluorescence intensity, FAD autofluorescence intensity, redox ratio, NAD(P)H τ 1 , NAD(P)H τ 2 , NAD(P)H α 1 , FAD τ 1 , FAD τ 2 , FAD α 1 , NAD(P)H τ m , or FAD τ m .

46. ​​The method of claim 46, wherein the current reprogramming prediction is calculated using the at least one metabolic endpoint and the at least one nuclear parameter as the input; the at least one nuclear parameter comprises at least one of: an area of ​​the nucleus of the reprogramming intermediate cell; a perimeter of the nucleus of the reprogramming intermediate cell; a nuclear shape index; an average distance of any pixel within the nucleus of the reprogramming intermediate cell to the nearest pixel outside the nucleus; a percentage of pixels located in a convex hull that are also located within the nucleus of the reprogramming intermediate cell; a percentage of image pixels located in a bounding box that are also located within the nucleus of the reprogramming intermediate cell; a total number of objects or other nuclei adjacent to the reprogramming intermediate cell nucleus; or a distance from the reprogramming intermediate cell nucleus to the nearest object and / or nearest other nucleus. The somatic cell reprogramming tracking device according to any one of claims 1 to 3.

47. A method according to any one of claims 29 to 36 or claim 42, wherein the autofluorescence data set is adjusted to wavelengths corresponding to NAD(P)H fluorescence and / or FAD fluorescence.

48. A method described in any one of claims 29 to 36 or claim 42, wherein the at least one metabolic endpoint comprises NAD(P)H autofluorescence intensity, FAD autofluorescence intensity, redox ratio, NAD(P)H τ 1 , NAD(P)H τ 2 , NAD(P)H α 1 , FAD τ 1 , FAD τ 2 , FAD α 1 , NAD(P)H τ m , or FAD τ m .

49. The method of claim 49, wherein the current reprogramming prediction is calculated using the at least one metabolic endpoint and the at least one nuclear parameter as the input; the at least one nuclear parameter comprises at least one of: an area of ​​the nucleus of the reprogramming intermediate cell; a perimeter of the nucleus of the reprogramming intermediate cell; a nuclear shape index; an average distance of any pixel within the nucleus of the reprogramming intermediate cell to the nearest pixel outside the nucleus; a percentage of pixels located in a convex hull that are also located within the nucleus of the reprogramming intermediate cell; a percentage of image pixels located in a bounding box that are also located within the nucleus of the reprogramming intermediate cell; a total number of objects or other nuclei adjacent to the reprogramming intermediate cell nucleus; or a distance from the reprogramming intermediate cell nucleus to the nearest object and / or nearest other nucleus.

43. The method of any one of claims 29 to 36 or claim 42.

Images