Morphological features of IFN-γ–stimulatedmesenchymal stromal cells predict overallimmunosuppressive capacityMatthew W. Klinkera,1, Ross A. Markleina,1, Jessica L. Lo Surdoa, Cheng-Hong Weia, and Steven R. Bauera,2

aDivision of Cellular and Gene Therapies, Center for Biologics Evaluation and Research, US Food and Drug Administration, Silver Spring, MD 20993

Edited by Darwin J. Prockop, Texas A&M Health Science Center, College Station, TX, and approved February 6, 2017 (received for review October 28, 2016)

Human mesenchymal stromal cell (MSC) lines can vary significantlyin their functional characteristics, and the effectiveness of MSC-based therapeutics may be realized by finding predictive featuresassociated with MSC function. To identify features associated withimmunosuppressive capacity in MSCs, we developed a robust invitro assay that uses principal-component analysis to integratemultidimensional flow cytometry data into a single measurementof MSC-mediated inhibition of T-cell activation. We used this assayto correlate single-cell morphological data with overall immuno-suppressive capacity in a cohort of MSC lines derived from differentdonors and manufacturing conditions. MSC morphology after IFN-γstimulation significantly correlated with immunosuppressive capacityand accurately predicted the immunosuppressive capacity of MSClines in a validation cohort. IFN-γ enhanced the immunosuppres-sive capacity of all MSC lines, and morphology predicted the mag-nitude of IFN-γ–enhanced immunosuppressive activity. Together,these data identify MSC morphology as a predictive feature ofMSC immunosuppressive function.

immunosuppression | mesenchymal stromal cells | morphology |high-content imaging | interferon-gamma

Human mesenchymal stromal cells (MSCs) can potentlysuppress immune responses in vitro and in animal models of

human disease (1, 2), but to date MSC-based therapies haveproduced mixed results in clinical trials for treatment of in-flammatory and autoimmune diseases (3, 4). A major challengein the development of consistently effective MSC-based immu-nosuppressive therapies is that MSC lines derived from differentdonors and manufacturing processes (i.e., cell expansion) canpossess markedly dissimilar immunosuppressive function (3, 5, 6).Although methods exist to assess MSC immunosuppression in vitro,they are often based on only a few measured outcomes, assay cul-ture conditions, and donor MSC samples (5, 7–9). To improve uponthese methods, we developed an experimental and analyticalapproach to quantify MSC-mediated immune suppression usingprincipal-component analysis (PCA) to integrate multiple mea-surements of T-cell activation assessed at a range of MSC den-sities. This approach allowed us to determine a single value forimmunosuppressive capacity for MSC lines derived from twodifferent manufacturing processes and 13 independent donors.Another major challenge associated with MSC-based immune

therapies is the lack of well-defined predictive markers to iden-tify MSC lines with therapeutically relevant biological activitiesor manufacturing processes that produce more effective MSC-based products. Efforts have been made to identify MSC qualityattributes associated with immunosuppression (6, 7), but the ma-jority of clinical studies (10) rely upon the surface markers describedby Dominici et al. (11). Having previously shown that morphologycan predict MSC mineralization capacity (12), we hypothesized thatmorphological features associated with immunosuppression inMSCs could be identified and used to predict their performance inour quantitative immunosuppression assay.Using our quantitative method for assessing the overall im-

munosuppressive capacity of MSCs, we identified and quantified

differences in immunosuppression related to donor and passage.Additionally, we used this assay to demonstrate consistent en-hanced immunosuppression for MSC lines that were stimulatedwith IFN-γ, as has been reported previously (8, 13, 14). Usinghigh-content imaging of MSCs following IFN-γ stimulation,morphological features were identified that significantly correlatedwith immunosuppressive capacity and predicted the immunosup-pressive capacity of other MSC and non-MSC lines, as well as theeffects of IFN-γ pretreatment on immunosuppressive capacity.These findings highlight MSCmorphology as an attribute associatedwith immunosuppressive capacity and provide additional evidencethat morphological profiling can be used to predict MSC in vitrofunction (12, 15–17).

ResultsPCA Permits Quantitative Assessment of Overall MSC ImmunosuppressiveCapacity. To quantify the immunosuppressive capacity of MSC lines,105 peripheral blood mononuclear cells (PBMCs) from a heathyhuman donor (PBMC donor 1, “PBMC1”) were stimulated withT-cell–activating beads for 3 d in the presence of increasing con-centrations of MSCs. PBMCs were then collected, and activationwas assessed in CD4+ and CD8+ T cells by flow cytometry-basedmeasurements of proliferation [carboxyfluorescein succinimidylester (CFSE) dilution], secretion of the proinflammatory cytokines

Significance

Substantial evidence exists demonstrating the immunosup-pressive function of mesenchymal stromal cells (MSCs), butinconsistent clinical results suggest that better understandingof MSC-mediated immunosuppression and identification offeatures predictive of immunosuppressive capacity would ad-vance MSC-based therapeutics. In this work, we present a ro-bust analytical approach to quantify the immunosuppressivecapacity of MSCs by integrating high-dimensional flow cytometrydata from multiple experimental conditions into a single measureof immunosuppressive capacity. Additionally, we identified mor-phological features of MSCs that predicted immunosuppressivecapacity, as well as the magnitude of IFN-γ–mediated immuno-suppression enhancement. These improved methods of MSCcharacterization could be used to identify MSC preparations withdesired immunosuppressive capacity, as well as screen for pre-treatments that enhance their immunosuppressive function.

Author contributions: M.W.K., R.A.M., C.-H.W., and S.R.B. designed research;M.W.K., R.A.M.,and J.L.L. performed research; M.W.K., R.A.M., and J.L.L. contributed new reagents/analytictools; M.W.K., R.A.M., J.L.L., C.-H.W., and S.R.B. analyzed data; and M.W.K., R.A.M., andS.R.B. wrote the paper.

The authors declare no conflict of interest.

This article is a PNAS Direct Submission.

Freely available online through the PNAS open access option.1M.W.K. and R.A.M. contributed equally to this work.2To whom correspondence should be addressed. Email: steven.bauer@fda.hhs.gov.

This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.1073/pnas.1617933114/-/DCSupplemental.

E2598–E2607 | PNAS | Published online March 10, 2017 www.pnas.org/cgi/doi/10.1073/pnas.1617933114

IFN-γ and tumor necrosis factor-α (TNF-α), and surface expres-sion of the activation marker CD25 (Fig. 1). Our initial studieswere carried out using early-passage cells (passage 3) from sixindependent MSC lines (three each from two independent com-mercial sources; see Table S1 for descriptions of all cell lines used).MSCs from all cell lines suppressed activation of CD4+ and CD8+

T cells in a dose-dependent manner, although the extent of sup-pression varied among MSC lines (Fig. 1).We observed that the relative magnitude of immunosuppres-

sion mediated by MSC lines differed between T-cell activationmeasurements (Fig. 1). Whereas proliferation was maximallyreduced by 80–90% in CD4+ and CD8+ cells by the highestconcentrations of MSCs (Fig. 1, % CFSE Diluted), the frequencyof CD25+ cells was only reduced by 20–40% for the same cellconcentrations (Fig. 1, % CD25+). Additionally, the relativeranking of MSC lines in our cohort based upon magnitude ofimmunosuppression differed between T-cell activation measure-ments as well as between MSC concentrations. For example, MSCline PCBM1632 had the greatest observed suppression of T-cellproliferation for both CD4+ and CD8+ cells at the majority of celldensities (Fig. 1, % CFSE Diluted), but this was not the case forthe percentage of IFN-γ–producing cells at multiple cell densities(Fig. 1, % IFN-γ+). Furthermore, cell density had a dramatic effecton the observed ranking of MSC immunosuppressive capacitybased on percentage of TNF-α–producing CD4+ T cells wherePCBM1632 had the most significant reduction at the lowest celldensity, but at the highest cell density had one of the least sig-nificant reductions in TNF-α production (Fig. 1, % TNF-α+). Fi-nally, we observed that low cell densities could enhance ratherthan inhibit some activation parameters above that of positivecontrols (>100% activation), especially IFN-γ secretion amongCD4+ T cells (Fig. 1, % IFN-γ+). Therefore, using data from a

single T-cell activation measure or from a single MSC concen-tration to quantify immunosuppressive capacity could lead tospurious conclusions.To address these issues, we developed an analytical approach

that incorporates various measures of T-cell activation and in-tegrates data obtained from multiple MSC concentrations into asingle value representing a given MSC line’s overall immuno-suppressive capacity (Fig. 2A). The initial step in this approachuses unsupervised PCA to effectively reduce the dimensionalityof the T-cell activation response and determine a compositevariable for overall T-cell activation. A total of 16 activationvariables [% CFSE diluted, % CD25+, % IFN-γ+, % TNF-α+,and associated median fluorescence intensity (MFI) values foreach marker for both CD4+ and CD8+ T cells] were used toconstruct principal components (Fig. 2A). All replicates of bothcontrol and experimental conditions were used for PCA. Whenthe overall experimental results were plotted in PC space (Fig. 2B),we observed that positive and negative T-cell activation controlsclearly separated along principal component 1 (PC1), and exper-imental conditions fell between the positive and negative controls.The extent of T-cell activation observed in samples containing thelowest concentration of MSCs is nearly indistinguishable frompositive control samples, but as the concentration of MSCs in-creases, samples cluster farther away from the positive controlsand toward negative controls (Fig. 2B). PC1 accounted for 60–80% of the observed variance in all experiments using PBMCsfrom a single donor, suggesting that PC1 is a suitable measurementof overall T-cell activation (Fig. 2B).

Cell Line- and Passage-Dependent Differences in MSC Immunosuppression.Using PC1 as a composite variable for T-cell activation, we de-veloped an area-under-curve (AUC) analysis to incorporate T-cell

Fig. 1. The relative immunosuppressive capacity of MSC lines varies according to T-cell activation measurement and MSC concentration. A constant numberof human PBMCs from a single donor (105 per well) were stimulated with T-cell–activating beads in the presence of increasing concentrations of cells from sixindependent MSC lines. After 3 d, activation of CD4+ and CD8+ T cells was measured by proliferation (% CFSE diluted), CD25 expression (% CD25+), andcytokine secretion (% IFN-γ+ and % TNF-α+). Percent activation was calculated for each value relative to positive and negative control samples. The range ofthe y axes differs between graphs because low concentrations of some MSC lines enhanced activation above that observed in positive controls for someactivation measurements. Dotted lines represent 100% activation. Data shown are the mean and SEM values of biological experiments performed with fivereplicates per condition.

Klinker et al. PNAS | Published online March 10, 2017 | E2599

APP

LIED

BIOLO

GICAL

SCIENCE

SPN

ASPL

US

activation data from multiple MSC concentrations into a singlevalue representing a given MSC line’s overall immunosuppressivecapacity. PCA using all available data were used to determine thecomposite T-cell activation value PC1, which was then trans-formed relative to control samples and plotted vs. MSCs per well(Fig. 2C). The AUC for each MSC line/passage group was cal-culated using the average transformed PC1 value for each con-centration of MSCs (SI Materials and Methods). For ease ofpresentation and to facilitate interexperimental comparisons, AUCvalues were transformed such that a hypothetical MSC line thatshowed complete immunosuppression (comparable to negativecontrol) at all cell concentrations would have an AUC of 0,whereas a hypothetical MSC line with no observed immunosup-pression (comparable to positive control) would have an AUCvalue of 100. By convention, AUC values are representative of theobserved level of overall T-cell activation, and therefore lowerAUC values represent stronger immunosuppression and higher

AUC values represent weaker immunosuppression. We used thiscombined PCA/AUC analysis to quantify the overall immunosup-pressive capacity of early- and late-passage cells from each MSC linein our initial cohort (Fig. 2 C and D). We tested both early- and late-passage cells (p3 and p7 for all lines except 127756 where late pas-sage represents p5) from all six of our initial cohort of MSC lines toincrease the observed dynamic range of immunosuppression, as im-munosuppressive capacity is reduced at higher passages (Fig. 2D, P <0.0001, two-way ANOVA).

MSC Morphology Is Altered by Exposure to IFN-γ. The immuno-suppressive capacity of MSCs is enhanced by exposure to theproinflammatory cytokine IFN-γ (18, 19), and we hypothesizedthat MSCs would undergo characteristic morphological changesfollowing exposure to IFN-γ that could be informative of theiroverall immunosuppressive capacity. To test this hypothesis,we used an experimental design in which morphological and

Fig. 2. Immunosuppressive capacity of MSC lines is quantified by integrating data from multiple MSC concentrations using a PCA-generated composite T-cellactivation measurement. (A) Activation data are collected from both CD4+ and CD8+ T cells by flow cytometry after 3 d of stimulation in the presence orabsence of MSCs. Histograms depict representative data from unstimulated T cells (red histograms) and activated T cells (blue histograms). Data from allexperimental conditions were then analyzed by PCA. (B) Five concentrations of early-passage cells from six independent MSC lines (500, 1,000, 2,000, 5,000,and 10,000 MSCs per well) were cocultured with a constant number of PBMCs from a single donor (105 per well) and T-cell–activating beads. PBMCs wereactivated for 3 d, and T-cell activation measurements (CFSE dilution, cytokine secretion, and CD25 expression) in CD4+ and CD8+ T cells were used to constructprincipal components. Each dot represents a single replicate (five replicates per cell concentration), and the size of each dot corresponds to the number ofMSCs per well. The first principal component (PC1) was used as an unbiased composite measure for overall T-cell activation. (C) Activation data from T cellsstimulated in the presence of a titration of early- and late-passage cells from six independent MSC lines were analyzed as in B. The first principal component(PC1) constructed from these data was then transformed relative to positive and negative control samples and plotted against MSC concentration. The shadedarea beneath the line represents the area under curve (AUC). (D) AUC for early- and late-passage cells for each of the six MSC lines was calculated from datashown in C. AUC values were transformed relative to positive and negative controls so that an AUC value of 100 represents a lack of observed MSC-mediatedsuppression, and a value of 0 corresponds to complete MSC-mediated suppression at all cell concentrations. The AUC of early-passage MSC lines was sig-nificantly lower than that of late-passage MSC lines (P < 0.0001, two-way ANOVA). Data shown are the mean and SEM values of biological experimentsperformed with five replicates per condition.

E2600 | www.pnas.org/cgi/doi/10.1073/pnas.1617933114 Klinker et al.

immunosuppressive assays were performed in parallel with early-and late-passage MSCs (Fig. 3) to identify morphological fea-tures associated with immunosuppression.The overall morphological signatures of six MSC lines under

control and IFN-γ–stimulated conditions (10 and 50 ng/mL)were determined using both unsupervised (A) and supervised (B)PCA (Fig. 4 A and B). Unsupervised PCA used all 558 morpho-logical features, whereas supervised PCA used a subset of 21features (Fig. 4, top right table) found to be significantly different(P < 0.0001) between unstimulated (0 ng/mL IFN-γ) and stimu-lated groups for all cell lines. The mean of PC1 was significantlydifferent (P < 0.0001) between both concentrations of IFN-γ andthe unstimulated group for both unsupervised (Fig. 4C) and su-pervised (Fig. 4D) PCA. Moreover, there was no significant dif-ference between the overall morphological profiles of MSCscultured in 10 and 50 ng/mL IFN-γ. Representative cell imagesfrom a single MSC line for each condition illustrate the charac-teristic morphological changes observed upon IFN-γ stimulation,for example, increase in cell and nuclear aspect ratio/eccentricity,decrease in cell and nuclear form factor (Fig. 4E). Fig. 4F furtherhighlights the distinct separation in the overall single-cell mor-phological profiles of unstimulated and stimulated MSCs using

PC1, but also reveals the existence of single-cell heterogeneitywithin each population.

Morphological Features of MSCs After IFN-γ Stimulation Correlatewith Immunosuppressive Capacity. Having confirmed that MSCsalter their overall morphology upon IFN-γ stimulation, we soughtto identify morphological features that correlated with our im-munosuppression results and could be used to predict a givenMSC line’s overall immunosuppressive capacity. In the absence ofIFN-γ stimulation, no morphological features of MSCs correlatedwell with immunosuppressive capacity. In contrast, we found thatmultiple morphological features of IFN-γ–stimulated MSCs weresignificantly correlated with immunosuppressive capacity (Fig. 5).For both concentrations of IFN-γ, similar correlation trends wereobserved for five cellular features (perimeter, form factor, maxi-mum feret diameter, nucleus/cytoplasm ratio, and Δminor axis

Fig. 3. Multiple MSC lines from independent donors are culture expandedand seeded for simultaneous morphological assessment and immunosup-pressive capacity using a coculture assay with human PBMCs. For morpho-logical assessment, MSCs are precultured for 24 h in control growth mediumand then exposed to 0, 10, and 50 ng/mL IFN-γ medium for 24 h. After 24 h,cells are fixed and analyzed for high-dimensional cellular (FITC-maleimide,green) and nuclear (Hoechst, blue) morphological features, and a charac-teristic morphological signature is obtained for each MSC line sample. Linesdemarcate cell and nuclear boundaries identified in image analysis. Thesemorphological signatures are then correlated with AUC values to identifymorphological features that can predict MSC immunosuppressive capacity.

Fig. 4. MSCs exhibit distinct morphological response upon stimulation withIFN-γ. Unsupervised (A) and supervised (B) principal-component analysis(PCA) performed on six MSC lines based on their morphological responses tounstimulated 0 ng/mL (black), 10 ng/mL (orange), and 50 ng/mL (red) IFN-γ .The table summarizes differences in cell- and nucleus-associated morpho-logical features (used in supervised PCA) upon IFN-γ stimulation. C and Dpresent the mean ± SD principal component 1 position (PC1 from A and B)for each condition using unsupervised and supervised morphological fea-tures, respectively (*P < 0.0001, significantly different from unstimulated).PC1 accounted for 47.1% (unsupervised) and 53.7% (supervised) of the datavariance. Data in A–D represent six cell lines (biological replicates) from fiveindependent experiments (n = 30 total points). (E) Separate PCA performedusing single-cell morphological data (>1,000 cells presented in density plotsof PC space) from one MSC line (PCBM1632) under unstimulated (black) andstimulated (10 ng/mL IFN-γ, orange) conditions. Representative cells werechosen randomly from the centroids and are shown for each condition.(F) Histogram representation of PC1 from single-cell data shown in E.

Klinker et al. PNAS | Published online March 10, 2017 | E2601

APP

LIED

BIOLO

GICAL

SCIENCE

SPN

ASPL

US

length). Of particular note, the differential feature Δminor axislengthcell was positively correlated with AUC, as a larger Δminoraxis lengthcell for a given MSC sample was associated with higherAUC values and thus lower immunosuppressive capacity. Pre-viously, we observed that this differential feature correlatedsimilarly with osteogenic potential as larger Δminor axis lengthcellwas associated with the absence of mineralization (12).The three morphological features most significantly correlated

with immunosuppressive capacity are presented in Fig. 5B in theform of 4D graphs. MSC lines with high immunosuppressivecapacity (low AUC values) clustered in the region that corre-sponds to a morphological profile of low cell perimeter, low cellmaximum feret diameter, and high nucleus/cytoplasm ratio afterIFN-γ stimulation (Fig. 5B). A linear regression model con-structed using these three features was significantly correlatedwith immunosuppression (Fig. S1, P < 0.0001, R = 0.78), whereasa model made using the same three features in unstimulatedcontrols did not correlate with immunosuppressive response(Fig. S1, P > 0.02, R = 0.55). Representative cells from both highand low immunosuppressive MSC lines after stimulation with10 ng/mL IFN-γ are shown in Fig. 5 C and D, respectively.

MSC Morphology After IFN-γ Stimulation Predicts Overall Immuno-suppressive Capacity. We sought to corroborate our initial find-ings using new MSC lines and with PBMCs isolated from adifferent donor (Table S2). Because the susceptibility of T cells toMSC-mediated suppression can differ between individuals (8, 20),

we tested several MSC lines with PBMCs from both donors andused these results to normalize AUC values obtained from ex-periments using PBMCs from our second donor (PBMC2). Tcells from PBMC2 were more susceptible to MSC-mediatedimmune suppression, as AUC values for MSC lines tested withPBMC2 cells were consistently lower than AUC values obtainedwhen the same MSC lines were tested with T cells from PBMC1(Fig. S2).Following normalization of the AUC values between experi-

mental datasets (Fig. S2), AUC values for the validation datasetalso demonstrated high correlation (Fig. 6A) with the top threemorphological features identified from the test dataset (cell pe-rimeter, nucleus/cytoplasm ratio, max feret diameter). PredictedAUC values were calculated for the validation dataset using themodel constructed from the test dataset (Fig. S1) and comparedwith the actual experimental AUC values. We found that AUCvalues predicted by the model were highly correlated with AUCvalues obtained experimentally (Fig. 6B, P < 0.001, R = 0.72 forAUCpred vs. AUCexp), and closely followed the line of unity. Thiscorrelation of predicted vs. experimental AUC values was evi-dent across the full dynamic range of immunosuppressive ca-pacity (Fig. 6C).Although these morphological features were useful predictors

for immunosuppression with MSC lines expanded using ourstandard conditions [α-MEM with 16.6% (vol/vol) lot-selectedFBS], they did not effectively predict the immunosuppressivecapacity of MSCs expanded by a different process [RoosterBio

Fig. 5. Morphological response of MSCs to IFN-γ correlates with immunosuppression. (A) Correlation of individual morphological features from MSCsstimulated by 10 ng/mL IFN-γ (Top) and 50 ng/mL IFN-γ (Bottom) with quantitative AUC immunosuppression results. Each point represents a single MSC line/passage with its associated median morphological feature and AUC values. (B) Four-dimensional graphs showing correlation of the top three morphologicalfeatures (perimeter, nucleus/cytoplasm ratio, maximum feret diameter) with AUC (color map shown below graphs) using morphological data from both 10and 50 ng/mL IFN-γ conditions. Representative cells stimulated with 10 ng/mL IFN-γ from a high immunosuppressive (low AUC) MSC line (C) and a low im-munosuppressive (high AUC) cell line (D). Data in A and B represent morphological data from five independent experiments using six MSC lines (n = 30 totalpoints).

E2602 | www.pnas.org/cgi/doi/10.1073/pnas.1617933114 Klinker et al.

Fig. 6. Morphological response of MSCs to IFN-γ predicts degree of immunosuppression. (A) Four-dimensional graph showing correlation of the top threemorphological features (perimeter, nucleus/cytoplasm ratio, maximum feret diameter) with AUC (color map shown below) using morphological data from10 ng/mL IFN-γ conditions for additional MSC lines (n = 16 biological replicates from one experiment) with a different PBMC donor (see Table S2 for descriptionsof cell lines). (B) Predictive model for AUC developed using linear regression on the top three highly correlated morphological features and AUC values fromtest set (perimeter, nucleus/cytoplasm ratio, and maximum feret diameter) and used to predict the AUC of validation set cell lines based on their earlymorphological signatures. Predicted AUC values (AUCpred) vs. experimental AUC values (AUCexp) are shown overall for the validation dataset (16 different cellline samples from one experiment) with linear fit (solid line) and reference AUCpred = AUCexp line for comparison. (C) Predicted (black) vs. experimental (white)AUC values shown in bar graph for each individual sample (ordered Left to Right from high to low immunosuppression). (D) The enhancing effect of IFN-γstimulation on MSC immunosuppression was demonstrated using eight MSC lines, which consisted of the six cell lines in Table S1 and two RB MSC lines.Overall immunosuppression response (transformed relative to inactivated and activated PBMCs) presented as PC1 for each MSC line at a range of cell densities(x axis). Five replicates were used for each cell concentration. Shaded region for unstimulated (black) and 10 ng/mL IFN-γ stimulated (orange) used for AUCanalysis of overall immunosuppressive capacity. (E) AUC immunosuppression scores for each MSC line for unstimulated (−IFN-γ, black) and stimulated (+IFN-γ,orange). Data presented as mean ± SEM (n = 5 per condition). Paired t test showed significant difference in AUC values for unstimulated and stimulatedconditions for each cell line (P < 0.0001). (F) Quantification of the change in immunosuppression [−ΔAUC = −(AUCstimulated – AUCunstimulated], where a larger–ΔAUC indicates a greater improvement in immunosuppression due to stimulation by IFN-γ. Correlation of –ΔAUC values with early differential morpho-logical features ΔMajor Axis LengthNucleus,10–0 (G) and ΔPerimeterNucleus,10–0 (H), which were found to be predictive of the change in immunosuppression uponIFN-γ stimulation.

Klinker et al. PNAS | Published online March 10, 2017 | E2603

APP

LIED

BIOLO

GICAL

SCIENCE

SPN

ASPL

US

(RB) MSC lines, Table S3], and none of the predictive mor-phological features previously identified (cell perimeter, nucleus/cytoplasm ratio, and maximum feret diameter) correlated withAUC values in RB MSC lines (Fig. S3A). Instead, unique mor-phological features were found to correlate with immunosup-pression in RB MSC lines: differential nuclear morphologicalfeatures extent, form factor, and minimum feret diameter (Fig.S3B). Although the same characteristic morphological changeupon IFN-γ stimulation was observed as with the other cell lines(increase in aspect ratio and decrease in nucleus/cytoplasm ratio),this finding suggests that predictive morphological features maydiffer depending on the manufacturing process.

Increased Immunosuppression Following IFN-γ Pretreatment Can BePredicted by MSC Morphology. We next examined the effects ofIFN-γ pretreatment on MSC immunosuppressive capacity bystimulating eight MSC lines with 10 ng/mL IFN-γ for 24 h beforePBMC coculture. As expected, IFN-γ pretreatment significantlyenhanced the immunosuppressive capacity of all MSC lines (Fig.6 D and E, P < 0.0001, paired t test). The difference in AUCvalues obtained under control and IFN-γ–stimulated conditionswas used to quantify the magnitude of this enhancement for eachMSC line as –ΔAUC, where larger values indicate an increasedeffect of IFN-γ stimulation on immunosuppressive capacity (Fig.6F). Although this enhancement in immunosuppression could bepredicted by noting the difference in the panel of morphologicalfeatures in Fig. 4, we sought to correlate the magnitude of thisenhanced immunosuppression due to IFN-γ stimulation withMSCmorphology. We found thatΔmajor axis lengthnucleus,10–0 andΔperimeternucleus,10–0 were highly correlated with IFN-γ–inducedenhancement of immunosuppression and could potentially beused to predict the effects of other soluble factor pretreatments(Fig. 6 G and H). Of note, both features exhibited consistentcorrelative trends to other differential features in that a greaterresponse to IFN-γ (i.e., increased Δmajor axis lengthnucleus andΔperimeternucleus) correlated with a smaller increase in immu-nosuppression. These results suggest that morphological profilingcan predict not only a given MSC line’s inherent immunosup-pressive capacity, but also the magnitude of IFN-γ–mediatedenhanced immunosuppression.

DiscussionMSC-based cellular therapies continue to be the subject of in-tense research despite a lack of clearly demonstrated clinicaleffectiveness in treating inflammatory disorders (2, 4). A majorcontributor to these inconsistent clinical outcomes may be theinherent variability in the immunosuppressive capacity of MSClines due to differing tissues of origin, manufacturing processes,and donor-specific differences (10, 21–23). Although this func-tional heterogeneity is well known, no reliable predictors of MSCimmunosuppressive capacity have been identified (3, 24). Anothersignificant challenge facing development of MSC-based therapiesis the lack of well-defined, robust assays for assessing their im-munosuppressive function. Current assays for assessing MSC im-munosuppressive capacity exist; however, they often rely onquantification of only a few markers in a single culture condition(8, 9, 20). Based on these limitations, we developed a robust,quantitative method of assessing the immunosuppressive capacityof human bone marrow-derived MSCs and further demonstratedthat morphological features of IFN-γ–stimulated MSCs can beused to predict their overall immunosuppressive capacity.Criteria for a standardized immunosuppressive capacity assay

for MSC-based products have been suggested (5, 6), but a spe-cific protocol for quantification of immunosuppressive capacityhas not yet gained wide acceptance. Measuring the activation ofT cells stimulated in the presence or absence of MSCs in vitro isa commonly used assay for assessing the effects of stimulatoryand inhibitory factors on MSC-mediated immune suppression;

however, this approach is rarely used to compare the inherentimmunosuppressive capacity of MSC lines, and interexper-imental variability can make it difficult to reliably distinguishbetween MSC lines with similar immunosuppressive capacities(5). Although T-cell proliferation is often used to measure ac-tivation in MSC immunosuppression studies, other measures ofactivation such as cytokine secretion and up-regulation of growthfactor receptors like CD25 more directly assess activated T-celleffector functions (25, 26). In this study, we assessed cytokinesecretion (IFN-γ and TNF-α) and CD25 expression in additionto proliferation in both CD4+ and CD8+ T cells activated in thepresence of MSCs (Fig. 1). We found that considering only asingle activation measurement to compare immunosuppressivecapacity could lead to erroneous conclusions, as the magnitudeof suppression differs between activation measurements, andsuppression of some activation measurements could be observedat lower MSC concentrations than others (Fig. 1).To avoid some of these limitations, we developed a method for

quantifying the overall immunosuppressive capacity of MSClines that is robust to interexperimental variability. For eachexperiment, we generated a composite T-cell activation variablefrom multiple activation measurements, allowing informationregarding effector functions and proliferation to be incorporatedinto a single measure of global T-cell activation (Fig. 2 A and B).Similar to other studies, we observed that the magnitude of T-cell responses to stimulation can differ substantially betweenindependent PBMC donors (27, 28), and when multiple mea-surements of T-cell activation are considered, each independentPBMC donor shows a unique profile of T-cell activation. Byusing PCA to produce a composite activation variable specific tothe PBMC donor in a given experiment, our method accounts forthese unique profiles by considering each activation measure-ment in an unbiased manner and incorporating those with themost variation into the first principal component (PC1; Fig. 2A).The presence of these unique activation profiles can be observedwhen data collected from two individual PBMC donors are usedtogether to construct principal components, as PC1 explainsmuch of the variation due to MSC-mediated immune suppres-sion, whereas the second principal component separates the re-sults by PBMC donor (Fig. S4). Using an unbiased compositemeasurement of T-cell activation rather than a single measure-ment reduces the impact of PBMC donor-specific variability thatcan impede interexperimental comparisons.Our method for quantifying the immunosuppressive capacity

of MSC lines also integrates data from multiple concentrationsof MSCs using an AUC analysis. Incorporating data from a dy-namic range of immunosuppression for each MSC line providesa robust and repeatable measurement of immunosuppressivecapacity for a given MSC line and avoids any dubious conclu-sions about the relative immunosuppressive capacity of MSClines that could be made when only one or a few MSC concen-trations are considered. For some MSC lines, we observed anenhancement of T-cell activation at low MSC concentrationseven though higher concentrations of MSCs from the same lineshowed clear immunosuppression (Fig. 2C). This enhancementhas been observed by others as well, and its incorporation intothe AUC gives a broader measure of the immunosuppressivecapacity of a given MSC line that may have biological relevancein some in vivo scenarios where MSCs present in small numberscould potentially increase rather than decrease inflammatorysymptoms (27, 29). The AUC for a given MSC line differedbetween experiments in which independent PBMC donors wereused due to donor-specific differences in T-cell susceptibility toMSC-mediated immune suppression (Fig. S2). However, al-though the absolute AUC value for a given MSC preparationdiffered when independent PBMC donors were used, the relativevalue of MSC lines in our cohort was generally consistent re-gardless of the PBMC source (Fig. S2). These results suggest that

E2604 | www.pnas.org/cgi/doi/10.1073/pnas.1617933114 Klinker et al.

the presence of a “reference” immunosuppressive cell line ineach experiment would be beneficial to make direct comparisonsof MSC lines tested in different experiments (8). A similar strategycould also facilitate comparing experimental results from differentlaboratories by using results obtained for shared “cell referencematerials” to normalize results between laboratories (5, 30). Al-though we limited our experiments to MSC-mediated suppressionof T cells in this study, this multivariate PCA/AUC approach couldbe adapted for the study of MSC-mediated suppression of othertypes of immune cells, such as B cells or monocytes.MSC-mediated immunosuppression is commonly measured

using only proliferation in CD4+ T cells, with many studiesmeasuring suppression at only a few MSC:T-cell ratios. We there-fore chose to compare our comprehensive AUC approach to themore commonly used approach assessing only proliferation inCD4+ T cells (Fig. S5). Using data combined from multiple ex-periments, we correlated AUC values with percentage suppressioncalculated using only CFSE dilution among CD4+ T cells fromthese same experiments. Although CFSE dilution from some celldensities correlated moderately with AUC (R2 = 0.58 for 2,000MSCs per well, R2 = 0.70 for 5,000 MSCs per well; R2 = 0.61 for10,000 MSCs per well), CFSE dilution alone was unable to accountfor all of the variation in immunosuppressive capacity observed bythe AUC metric (Fig. S5). Therefore, although less comprehensivemeasures of MSC-mediated immunosuppression are appropriatefor detecting qualitative differences in immunosuppressive capacityamong MSC lines, studies using such methods will have reducedpower to detect quantitative differences or differentiate betweenMSC lines with comparable immunosuppressive capacities.Our AUC results demonstrate that immunosuppressive ca-

pacity decreases with passage for each MSC line and provides acomprehensive and quantitative description of that difference(Fig. 2D). We observed that even the least-immunosuppressiveearly-passage MSC line showed better suppression than themost-suppressive late-passage MSC line, and the combined co-hort of early- and late-passage lines provided a large dynamicrange of immunosuppressive capacity. The immunosuppressionresults collected in this study encompass data on MSC lines pro-duced from three different commercial sources and includesdonor-matched suppression results from both early- and late-pas-sage cells as well as unstimulated and IFN-γ–treated conditions.This study contains a substantial amount of information fromcharacterization of in vitro MSC-mediated immunosuppression,which may be important for future correlational studies looking forother predictive factors associated with immunosuppressive ca-pacity, including gene expression and secretome analyses.Concomitant with the challenges of effectively characterizing

the overall immunosuppressive capacity of MSCs is the inabilityto identify cell characteristics that can effectively predict thisbehavior in vitro and in vivo. We previously demonstrated thatearly morphological features can serve as predictors of MSCmineralization capacity (12) and sought to extend this predictivecharacterization to immunosuppressive capacity by assessingMSC morphology in culture conditions relevant to MSC-mediatedimmune suppression. We chose to investigate the morphologicalresponse of MSCs to the cytokine IFN-γ as it is abundant at sitesof inflammation, enhances the immunosuppressive function ofMSCs, and is required for optimal MSC-mediated immunosup-pression in some animal disease models (13, 31, 32). We foundthat, upon IFN-γ stimulation, all MSC lines adopted morpholog-ical phenotypes distinct from unstimulated MSCs, with IFN-γ–stimulated MSCs displaying increased elongation (increasedaspect ratio/eccentricity), increased complexity (decrease in formfactor and solidity), and lower nucleus/cytoplasm ratios, all within24-h stimulation.IFN-γ receptors signal in part through the JAK–STAT pathway,

which mediates transcriptional changes through translocation ofSTAT proteins into the nucleus (33). Evidence of IFN-γ–stimulated

morphological changes has been found in the case of monocyteswhere enhanced actin polymerization was observed in the pres-ence of IFN-γ (34). In other cell types (HepG2 and HeLa),differences in cell adhesion signaling (drug and matrix-induced)that led to adoption of distinct morphologies were discovered tosignificantly impact IFN-γ signaling as actin cytoskeletal arrange-ment was directly linked to STAT1 dephosphorylation (35). Thissuggests that morphology not only serves as a phenotypic readoutfor IFN-γ treatment, but could also play a direct role in the mag-nitude of the MSC immunosuppression response as IFN-γ signalingpromotes production of immunomodulatory factors such as indo-leamine 2,3-dioxygenase, prostaglandin E2, and IL-6 (20, 36, 37).MSCs in coculture with PBMCs are exposed to T-cell–derived

IFN-γ (as flow cytometry confirmed), and their ability to sup-press T-cell activation was highly correlated with their mor-phology after 24 h of IFN-γ treatment. This correlation wasdemonstrated with six independent MSC lines at both early andlate passages, and prediction models constructed with morpho-logical data from these experiments were able to corroborate andaccurately predict the immunosuppressive capacity of additionaluntested MSC lines from multiple donors (and a control fibro-blast cell line). Although increase in cell size with extended culturehas been widely observed for MSCs (12, 38–40), overall cell areaalone was a poor predictor of immunosuppression (P > 0.11) andassessment of more subtle cell morphological features upon IFN-γstimulation (perimeter, maximum feret diameter, and form factor)better predicted overall immunosuppression (Fig. 5). This furtherunderscores the importance of designing functionally relevantmorphological profiling tools based on the context of the assay,that is, extracting information about MSC immunosuppression byobserving morphology in response to a stimulus (IFN-γ) that en-hances the immunosuppressive function of MSCs.Using data compiled from publications using the same MSC

lines used in this study (12, 38, 41), we directly compared thepotential predictive power of several commonly used MSCmarkers to that of the morphological features identified in thisstudy. We found that only the morphological features (upon 24-hIFN-γ stimulation) significantly correlated with immunosuppres-sive capacity, whereas the expression of traditional MSC surfacemarkers and trilineage differentiation capacity were not signifi-cantly correlated with our AUC results (Table S4). These resultsemphasize the importance of using functionally relevant assays tomore effectively predict the immunosuppressive capacity of anMSC population. Other researchers could potentially use thismorphological profiling approach to better characterize their celllines and predict the immunosuppressive capacity of different celllines as well as the effect of changes in manufacturing conditions(i.e., culturing conditions) on immunosuppressive capacity byquantifying morphological responses of cells to IFN-γ.MSCs expanded with different manufacturing processes did

not demonstrate the same correlation as our original dataset(Fig. S3). However, unique morphological features that corre-lated with immunosuppression were identified for these MSClines. Although we tested fewer MSC lines expanded by thissecond manufacturing process, this result may indicate thatmorphological characterization must be performed with eachmanufacturing process to establish useful criteria for identifyingMSC preparations with desired immunosuppressive potential.Similarly, MSCs from other tissue sources (adipose, umbilicalcord, dental pulp) may possess tissue-specific morphologicalphenotypes and responses that will need to be characterized toidentify potential morphological quality attributes unique tothese cell types (42). As this study tested only three MSC linesproduced by this second manufacturing process, further studywill be required to confirm that unique morphological profiles arecorrelated with immunosuppressive capacity in bone marrow-derived MSC lines produced by varying sources or manufactur-ing processes.

Klinker et al. PNAS | Published online March 10, 2017 | E2605

APP

LIED

BIOLO

GICAL

SCIENCE

SPN

ASPL

US

This work also demonstrated that changes in morphology werenot only associated with inherent immunosuppressive capacity,but could also be used to determine the effects of IFN-γ pre-treatment on MSC immunosuppression (Fig. 6). Using our AUCanalysis, we observed consistent enhancement of immunosup-pression for eight MSC lines when samples were stimulated withIFN-γ for 24 h before coculture with PBMCs. The presence ofthis conserved trend of enhanced immunosuppression could bepredicted based on the overall difference in morphological signa-tures (Fig. 4, control vs. stimulated). Furthermore, the magnitudeof this enhancement highly correlated with the differential fea-tures perimeternucleus and major axis lengthnucleus (Fig. 6 G andH). Nuclear morphology relates to stemness and differentiation(43, 44), and there is direct coupling of the nucleus to the cy-toskeleton through proteins such as Lamin A/C (44, 45). Evi-dence suggests that nuclear morphological changes associatedwith a biological function may be mediated by epigenetic mod-ifications (46–48). The effect of IFN-γ stimulation on permissivechromatin (at the IDO promoter) associated with MSC immu-nosuppressive function was demonstrated recently with severalof the MSC lines used in this study and may be related to thechange in nuclear morphology associated with IFN-γ treatment(47). This further supports the notion that nuclear morphologicaldata can be informative of MSC immunosuppressive capacityand treatment conditions that enhance this capacity. Again, aswas the case with our osteogenic morphological analysis (12),larger differential values (i.e., a greater morphological responseto IFN-γ) correlated with a reduced functional response to stim-ulation. Using this technique, a high-throughput assessment ofstimulatory regimens could be performed by analyzing the mor-phological response of MSCs, which could facilitate discovery ofoptimal conditions for preconditioning MSCs for immunother-apies. This approach could potentially identify “personalized”stimulatory regimens for a given patient or disease indication andresult in improved clinical outcomes using autologous or allo-geneic MSC-based products.Although derived from high-content, single-cell measure-

ments, our morphological correlations with immunosuppressionstill represent population-based techniques, which require fur-ther investigation and refinement to better address inherentcharacterization difficulties associated with MSC heterogeneity.Preliminary evidence exists that demonstrates morphologicalsubpopulations with distinct functional capacities (proliferationand differentiation), and continued investigations into the emer-gence of heterogeneity will facilitate improved characterizationand understanding of cell therapy products (49–51). Improvementson single-cell morphological analysis using high-dimensional single-cell classification techniques such as SPADE or viSNE couldallow for identification of morphologically distinct subpopulationsof MSCs that could be enriched for immunotherapies (52). Theseapproaches could also provide insight into the progression andestablishment of MSC heterogeneity as it relates to differencesin donor (or tissue) source and in manufacturing (40, 53). Ad-vancements in live-cell tracking and monitoring in vitro wouldfurther allow for direct correlation of early morphological phe-notypes with assay outcomes (15, 54).Improved understanding of morphological signatures in the

context of stem cell heterogeneity and specific biological func-tions will supplement current MSC characterization methodsused in clinical applications involving immunomodulation. Mor-phology has been identified in many studies as a predictive markerof osteogenic capacity (15, 16); however, limited research existsthat demonstrates a correlation of morphology with other MSCfunctions such as immunosuppression. Further improvement onthe methods for morphological assessment and applicability acrossmultiple donors is necessary to fully demonstrate its possible utilityas a predictor of not only in vitro but also in vivo MSC immuno-suppressive capacity. This work also provides further insight into

the establishment of analytical approaches making use of anMSC internal “reference cell line” (8), as we were able to de-termine the relative immunosuppressive capacity of validationMSC lines based on normalization of data acquired fromMSCs and PBMCs derived from different donors. Our de-velopment of predictive models of in vitro immunosuppressionbased on morphological data followed by successful applica-tion to new MSC lines lays the foundation for further studyinto identifying morphological signatures that can predictMSC potency.

Materials and MethodsCell Lines and MSC Manufacturing/Expansion. Human bone marrow-derivedstem cells were obtained from six different donors purchased from Lonza(127756, 167696, 8F3560) and AllCells (PCBM1632, PCBM1641, PCBM1662) atpassage 2 (see Table S1 for donor specifications). Additional cell lines usedfor validation are described in Tables S2 and S3. MSC culture conditions werechosen based on well-established protocols (55). Most of the MSC lines usedin this study have been extensively characterized and further informationabout the donors’ surface marker expression, genomic, epigenetic, andproteomic profiles, as well as performance in multiple bioassays for MSCfunction has been published previously (12, 38, 47, 48, 56–58). Detailed MSCculturing and expansion methods are described in SI Materials and Methods.All cell lines used in this work possessed viability >95% (assessed usingTrypan Blue) before plating for immunosuppression coculture and mor-phological assessment.

Quantitative Assessment of MSC Immunosuppressive Capacity. PBMCs fromhealthy human donors (105 per well) were stimulated with an equal numberof T-cell–activating beads (Life Technologies; 11131D) in the presence of fiveconcentrations of each MSC line (500, 1,000, 2,000, 5,000, and 10,000 cellsper well) in 96-well plates. After 3 d, PBMCs were harvested and activationwas assessed in both CD4+ and CD8+ T cells by proliferation (CFSE dilution),CD25 expression, and production of IFN-γ and TNF-α. PCA was used togenerate a composite T-cell activation metric (PC1) using all available flowcytometry data for a given experiment. Transformed PC1 values were thenplotted over MSC concentration for each MSC line, and the AUC calculatedfrom these plots was used to quantify the immunosuppressive capacity ofeach MSC line. AUC values are inversely proportional to immunosuppressivecapacity, and all AUC values were transformed for ease of presentation suchthat a hypothetical MSC line that showed complete immune suppression atall concentrations would have an AUC of 0, and an MSC line with no im-munosuppressive capacity would have an AUC value of 100. More detaileddescriptions of this experimental and analytical approach are found in SIMaterials and Methods and Tables S5–S7. All MSC and PBMC samples werepurchased from commercial sources with appropriate informed consent asindicated on the certificates of analysis. Research on commercially obtainedhuman cell lines is exempt from FDA IRB (Research Involving Human SubjectsCommittee) oversight.

Determination of an Overall MSC Morphological Signature. For MSC mor-phological analysis, cells were seeded at a density of 1,000 cells per well (fourtotal wells per experimental group) in 12-well plates (Corning) and culturedfor 24 h in growth medium. Growth medium was replaced with growthmedium containing 0, 10, or 50 ng/mL IFN-γ (Life Technologies) and culturedfor an additional 24 h and then fixed with 4% (wt/vol) paraformaldehyde for15 min. Cell and nuclear morphology were assessed using fluorescein (FITC)-maleimide (Life Technologies) and Hoechst (Sigma-Aldrich), respectively.Briefly, fixed samples were incubated with 20 μM FITC-maleimide for 30 min,washed with PBS, incubated with 1 mg/mL Hoechst for 5 min, and washedwith PBS before imaging. Samples for morphological analysis were imagedusing an inverted Nikon Ti-U microscope with automated stage (Prior) andfilters (Chroma Technology). At least 1,000 cells were assessed for each ex-perimental group, with approximately equal numbers of cells analyzed fromeach of four replicate wells. Automated quantification of cellular and nu-clear shape features was performed using CellProfiler (59) to obtain quan-titative morphological data for each cell consisting of 46 cellular shapefeatures and 46 nuclear shape features (Tables S8 and S9). Overall mor-phological signatures were constructed for each group by taking the medianvalue of the 93 total cellular and nuclear features in each IFN-γ condition fora total of 279 morphological features. More detailed descriptions of thisexperimental and analytical approach are found in SI Materials and Methodsand Table S6.

E2606 | www.pnas.org/cgi/doi/10.1073/pnas.1617933114 Klinker et al.

ACKNOWLEDGMENTS. We thank Drs. Raj Puri, Ian Bellayr, and NirjalBhattarai for review of this manuscript and Drs. Johnny Lam, Ian Bellayr,and Eva Rudikoff for technical assistance in producing most cell lines used inthis study. M.W.K. was supported in part by appointment to the ResearchParticipation Program at Center for Biologics Evaluation and Researchadministered by the Oak Ridge Institute for Science and Education through

US Department of Education and US Food and Drug Administration. Thiswork was also supported in part by the Food and Drug AdministrationModernizing Science grant program, a Biomedical Advanced Researchand Development Authority grant, a grant from the Medical Countermea-sures Initiative, and research funds from the Division of Cell and GeneTherapies.

1. Klinker MW, Wei C-H (2015) Mesenchymal stem cells in the treatment of in-flammatory and autoimmune diseases in experimental animal models. World J StemCells 7(3):556–567.

2. English K (2013) Mechanisms of mesenchymal stromal cell immunomodulation.Immunol Cell Biol 91(1):19–26.

3. Sharma RR, Pollock K, Hubel A, McKenna D (2014) Mesenchymal stem or stromal cells:A review of clinical applications and manufacturing practices. Transfusion 54(5):1418–1437.

4. Griffin MD, et al. (2013) Concise review: Adult mesenchymal stromal cell therapy forinflammatory diseases: How well are we joining the dots? Stem Cells 31(10):2033–2041.

5. Galipeau J, et al. (2016) International Society for Cellular Therapy perspective onimmune functional assays for mesenchymal stromal cells as potency release criterionfor advanced phase clinical trials. Cytotherapy 18(2):151–159.

6. Krampera M, Galipeau J, Shi Y, Tarte K, Sensebe L; MSC Committee of the In-ternational Society for Cellular Therapy (ISCT) (2013) Immunological characterizationof multipotent mesenchymal stromal cells—The International Society for CellularTherapy (ISCT) working proposal. Cytotherapy 15(9):1054–1061.

7. Rizzo R, et al. (2011) A simple method for identifying bone marrow mesenchymalstromal cells with a high immunosuppressive potential. Cytotherapy 13(5):523–527.

8. Salem B, et al. (2015) Quantitative activation suppression assay to evaluate humanbone marrow-derived mesenchymal stromal cell potency. Cytotherapy 17(12):1675–1686.

9. Copland I, Qayed M, Garcia M, Galipeau J, Waller E (2015) Bone marrow mesenchymalstromal cells from patients with acute and chronic graft-versus-host disease deploynormal phenotype, differentiation plasticity, and immune-suppressive activity. BiolBlood Marrow Transplant 21(5):934–940.

10. Mendicino M, Bailey AM, Wonnacott K, Puri RK, Bauer SR (2014) MSC-based productcharacterization for clinical trials: An FDA perspective. Cell Stem Cell 14(2):141–145.

11. Dominici M, et al. (2006) Minimal criteria for defining multipotent mesenchymalstromal cells. The International Society for Cellular Therapy position statement.Cytotherapy 8(4):315–317.

12. Marklein RA, et al. (2016) High content imaging of early morphological signaturespredicts long term mineralization capacity of human mesenchymal stem cells uponosteogenic induction. Stem Cells 34(4):935–947.

13. Dazzi F, Krampera M (2011) Mesenchymal stem cells and autoimmune diseases. BestPract Res Clin Haematol 24(1):49–57.

14. Valencic E, Piscianz E, Andolina M, Ventura A, Tommasini A (2010) The immuno-suppressive effect of Wharton’s jelly stromal cells depends on the timing of their li-censing and on lymphocyte activation. Cytotherapy 12(2):154–160.

15. Matsuoka F, et al. (2013) Morphology-based prediction of osteogenic differentiationpotential of human mesenchymal stem cells. PLoS One 8(2):e55082.

16. Treiser MD, et al. (2010) Cytoskeleton-based forecasting of stem cell lineage fates.Proc Natl Acad Sci USA 107(2):610–615.

17. Chen D, et al. (2016) Machine learning based methodology to identify cell shapephenotypes associated with microenvironmental cues. Biomaterials 104:104–118.

18. Krampera M, et al. (2006) Role for interferon-gamma in the immunomodulatoryactivity of human bone marrow mesenchymal stem cells. Stem Cells 24(2):386–398.

19. Meisel R, et al. (2004) Human bone marrow stromal cells inhibit allogeneic T-cell re-sponses by indoleamine 2,3-dioxygenase-mediated tryptophan degradation. Blood103(12):4619–4621.

20. Solchaga LA, Zale EA (2012) Prostaglandin E2: A putative potency indicator of theimmunosuppressive activity of human mesenchymal stem cells. Am J Stem Cells 1(2):138–145.

21. Phinney DG (2007) Biochemical heterogeneity of mesenchymal stem cell populations:Clues to their therapeutic efficacy. Cell Cycle 6(23):2884–2889.

22. Ho AD, Wagner W, Franke W (2008) Heterogeneity of mesenchymal stromal cellpreparations. Cytotherapy 10(4):320–330.

23. Phinney DG (2012) Functional heterogeneity of mesenchymal stem cells: Implicationsfor cell therapy. J Cell Biochem 113(9):2806–2812.

24. Galipeau J, Krampera M (2015) The challenge of defining mesenchymal stromal cellpotency assays and their potential use as release criteria. Cytotherapy 17(2):125–127.

25. Kim HP, Imbert J, Leonard WJ (2006) Both integrated and differential regulation ofcomponents of the IL-2/IL-2 receptor system. Cytokine Growth Factor Rev 17(5):349–366.

26. Liao W, Lin JX, Leonard WJ (2011) IL-2 family cytokines: New insights into the complexroles of IL-2 as a broad regulator of T helper cell differentiation. Curr Opin Immunol23(5):598–604.

27. Bloom DD, et al. (2015) A reproducible immunopotency assay to measure mesen-chymal stromal cell-mediated T-cell suppression. Cytotherapy 17(2):140–151.

28. Ketterl N, et al. (2015) A robust potency assay highlights significant donor variation ofhuman mesenchymal stem/progenitor cell immune modulatory capacity and ex-tended radio-resistance. Stem Cell Res Ther 6:236.

29. Le Blanc K, Tammik L, Sundberg B, Haynesworth SE, Ringdén O (2003) Mesenchy-mal stem cells inhibit and stimulate mixed lymphocyte cultures and mitogenic re-sponses independently of the major histocompatibility complex. Scand J Immunol57(1):11–20.

30. Viswanathan S, et al. (2014) Soliciting strategies for developing cell-based referencematerials to advance mesenchymal stromal cell research and clinical translation. StemCells Dev 23(11):1157–1167.

31. Polchert D, et al. (2008) IFN-gamma activation of mesenchymal stem cells for treat-ment and prevention of graft versus host disease. Eur J Immunol 38(6):1745–1755.

32. Schurgers E, Kelchtermans H, Mitera T, Geboes L, Matthys P (2010) Discrepancy be-tween the in vitro and in vivo effects of murine mesenchymal stem cells on T-cellproliferation and collagen-induced arthritis. Arthritis Res Ther 12(1):R31.

33. Bach EA, Aguet M, Schreiber RD (1997) The IFN gamma receptor: A paradigm forcytokine receptor signaling. Annu Rev Immunol 15:563–591.

34. Frausto-Del-Río D, et al. (2012) Interferon gamma induces actin polymerization, Rac1activation and down regulates phagocytosis in human monocytic cells. Cytokine 57(1):158–168.

35. Chen Z, et al. (2011) Negative regulation of interferon-γ/STAT1 signaling through cell ad-hesion and cell density-dependent STAT1 dephosphorylation. Cell Signal 23(8):1404–1412.

36. Chinnadurai R, Copland IB, Patel SR, Galipeau J (2014) IDO-independent suppressionof T cell effector function by IFN-γ-licensed human mesenchymal stromal cells.J Immunol 192(4):1491–1501.

37. Zimmermann JA, McDevitt TC (2014) Pre-conditioning mesenchymal stromal cell spher-oids for immunomodulatory paracrine factor secretion. Cytotherapy 16(3):331–345.

38. Lo Surdo JL, Millis BA, Bauer SR (2013) Automated microscopy as a quantitativemethod to measure differences in adipogenic differentiation in preparations of hu-man mesenchymal stromal cells. Cytotherapy 15(12):1527–1540.

39. Lee WC, et al. (2014) Multivariate biophysical markers predictive of mesenchymalstromal cell multipotency. Proc Natl Acad Sci USA 111(42):E4409–E4418.

40. Whitfield MJ, Lee WC, Van Vliet KJ (2013) Onset of heterogeneity in culture-expandedbone marrow stromal cells. Stem Cell Res (Amst) 11(3):1365–1377.

41. Lo Surdo JL (2013) Toward improved characterization of human mesenchymal stemcells for use in cellular therapies: A regulatory science perspective. PhD dissertation(University of Pittsburgh, Pittsburgh).

42. Lv F-J, Tuan RS, Cheung KM, Leung VY (2014) Concise review: The surface markers andidentity of human mesenchymal stem cells. Stem Cells 32(6):1408–1419.

43. Pajerowski JD, Dahl KN, Zhong FL, Sammak PJ, Discher DE (2007) Physical plasticity ofthe nucleus in stem cell differentiation. Proc Natl Acad Sci USA 104(40):15619–15624.

44. Talwar S, Kumar A, Rao M, Menon GI, Shivashankar GV (2013) Correlated spatio-temporal fluctuations in chromatin compaction states characterize stem cells. BiophysJ 104(3):553–564.

45. Shin J-W, et al. (2013) Mechanobiology of bone marrow stem cells: From myosin-IIforces to compliance of matrix and nucleus in cell forms and fates. Differentiation86(3):77–86.

46. Chambliss AB, Wu P-H, Chen W-C, Sun SX, Wirtz D (2013) Simultaneously defining cellphenotypes, cell cycle, and chromatin modifications at single-cell resolution. FASEB J27(7):2667–2676.

47. Rovira Gonzalez YI, Lynch PJ, Thompson EE, Stultz BG, Hursh DA (2016) In vitro cy-tokine licensing induces persistent permissive chromatin at the indoleamine 2,3-dioxygenase promoter. Cytotherapy 18(9):1114–1128.

48. Stultz BG, et al. (2016) Chromosomal stability of mesenchymal stromal cells during invitro culture. Cytotherapy 18(3):336–343.

49. Colter DC, Sekiya I, Prockop DJ (2001) Identification of a subpopulation of rapidly self-renewing and multipotential adult stem cells in colonies of human marrow stromalcells. Proc Natl Acad Sci USA 98(14):7841–7845.

50. Prockop DJ, Sekiya I, Colter DC (2001) Isolation and characterization of rapidly self-renewing stem cells from cultures of human marrow stromal cells. Cytotherapy 3(5):393–396.

51. Sekiya I, et al. (2002) Expansion of human adult stem cells from bone marrow stroma:Conditions that maximize the yields of early progenitors and evaluate their quality.Stem Cells 20(6):530–541.

52. Fienberg HG, Nolan GP, eds (2014) High-Dimensional Single Cell Analysis (Springer,Berlin), Vol 377.

53. Rennerfeldt DA, Van Vliet KJ (2016) When colonies are not clones: Evidence andimplications of intracolony heterogeneity in mesenchymal stem cells. Stem Cells 34(5):1135–1141.

54. Phadnis SM, et al. (2015) Dynamic and social behaviors of human pluripotent stemcells. Sci Rep 5:14209.

55. Prockop DJ, Phinney D, Bunnell B (2008) Mesenchymal Stem Cells—Methods andProtocols. Methods in Molecular Biology (Humana, Totowa, NJ), Vol 449.

56. Bellayr IH, et al. (2014) Gene markers of cellular aging in human multipotent stromalcells in culture. Stem Cell Res Ther 5(2):59.

57. Bellayr IH, Marklein RA, Lo Surdo JL, Bauer SR, Puri RK (2016) Identification of pre-dictive gene markers for multipotent stromal cell proliferation. Stem Cells Dev 25(11):861–873.

58. Mindaye ST, Ra M, Lo Surdo JL, Bauer SR, Alterman MA (2013) Global proteomicsignature of undifferentiated human bone marrow stromal cells: Evidence for donor-to-donor proteome heterogeneity. Stem Cell Res (Amst) 11(2):793–805.

59. Carpenter A, et al. (2006) CellProfiler: Image analysis software for identifying andquantifying cell phenotypes. Genome Biol 7(Software):R100.

Klinker et al. PNAS | Published online March 10, 2017 | E2607

APP

LIED

BIOLO

GICAL

SCIENCE

SPN

ASPL

US