Human MSCs were purchased from RoosterBio, Inc. (Frederick, MD), expanded in the recommended medium, and cryopreserved at the initial passage following receipt. Population doublings were limited to less than 15–20.

Whole blood was collected from healthy donors in citrate tubes according to an approved institutional standard operating procedure. PBMCs were purified using gradient centrifugation through Ficoll-Paque Premium (1.078 g/mL density, VWR, Lutterworth, Leicestershire, England) in Greiner Bio-One Leucosep tubes (Cat # 227290P, Greiner Bio-One North America Inc., Monroe, NC) as described by the manufacturer (45). Briefly, 50 mL of blood per donor was collected in citrated tubes by the USAISR Research Blood Bank; PBMCs were separated from plasma and red blood cells and washed twice with Dulbecco’s phosphate buffered saline (dPBS, Gibco #14190-144, without calcium or magnesium chloride). Both single donor (hSD-PBMC) and pooled donor (hPD-PBMC, 8–10 donors) PBMC preparations were prepared. PBMCs were resuspended at 1–2 × 10⁷ per mL in CryoStor10, frozen at 1°C/min cooling rate in a Nalgene™ Mr Frosty and stored in liquid nitrogen until use. Viability upon thaw was 94 ± 2.4% (n = 9) by 1:1 dilution with acridine orange propidium iodide (AOPI) stain and cell counting in the Cellaca MX High-Throughput Automated Cell Counter (Nexcelom Bioscience, Lawrence, MA).

Cryopreserved PBMCs were rapidly thawed; washed twice in PBMC media (RPMI without phenol red, Cat # R7509, Sigma Aldrich) containing 10% MSC qualified FBS (cat # 12662011, Thermo-Fisher), 1 mM Glutamax (Cat # 35050061, Thermo Fisher), and 10 mM HEPES, pH 7.4 (Cat # H0887, Sigma Aldrich); and resuspended at 1.5 × 10⁶ PBMCs per mL in PBMC media. In 96-well plate assays (Corning 96-well flat bottomed, Tissue culture treated, polystyrene plates, Sigma Aldrich Cat # CLS3595), 150,000 PBMCs were loaded per well with or without MSC co-culture in a volume of 200 µL. Proliferation was stimulated by the addition of 50 µL of stimulus as indicated in the text. Stimuli included phytohemagglutinin P (PHA-P, Sigma Aldrich Cat # L-1668) and phytohemagglutinin-L (PHA-L, Sigma Aldrich Cat # 431784). After 2, 4, 6, 24, 48, 72, or 96 h incubation at 37°C, 5% CO₂, cells in the media were resuspended and aliquots were taken for analysis. In designated experiments, a monocyte-depleted population of PBMCs was obtained by incubating the thawed, washed, resuspended PBMCs at 1.5 × 10⁶ in culture media in a T-75 tissue culture flask overnight. After incubation, monocyte-depleted PBMCs were removed, washed in culture media, and resuspended at 1.5 × 10⁶ cells per mL for use in experiments. After overnight incubation, monocyte-depleted PBMCs were generally 85% viable.

For routine experiments, MSCs were serially diluted starting at 50–60,000 cells/well and cultured in 96-well plates for 24 h (45). On Day 0, the attached MSCs were washed, re-fed with PBMC media, and incubated with or without 150,000 freshly thawed pooled PBMCs/well, in the presence or absence of stimulus. Actual PBMC : MSC ratios ranged from 2.5:1 to 202.5:1. In selected experiments, “licensing” of the MSCs was performed by incubating the cultured MSCs for the 72 h prior to harvest in 50 μg/mL interferon-gamma (IFN-γ).

The classic carboxyfluorescein succinimidyl ester (CFSE) assay was performed in 96-well plates as previously described (36, 45). MSC co-culture experiments were performed as above with PBMCs labeled with 5(6)-carboxyfluorescein diacetate N-succinimidyl ester (CFDA, Sigma #21888-25mg-F, St Louis, MO), converting CFDA to CFSE. After incubation with or without MSCs in the presence or absence of mitogen, at 37°C for 96 h, labeled PBMCs were assayed by flow cytometry. Cells were stained with anti-CD3 antibody, and to determine the lymphocyte population, uptake of the nucleic acid dye 7-aminoactinomycin D (7AAD; BioLegend Cat # 420404) discriminated live/dead staining. Specifically, cells were centrifuged at 400 × g for 6 min, resuspended in 50 µL of Hanks’ Buffered Saline Solution (HBSS) containing 2% bovine serum albumin (BSA) and then incubated with 20 µL of APC-CD3 antibody (BD Biosciences Cat# 555342, RRID : AB 398592) for 20 min at room temperature. After using the wash-only setting on the BD FACS Lyse Wash assistant, 5 µL of 7AAD was added and the cells were incubated for an additional 5 min before analysis on a BD FACS Canto II flow cytometer (BD Biosciences, Franklin Lakes, NJ) using an expanded FSC-A/SSC-A gate to capture the larger proliferating population (51). A total of 20,000 CD3+ events were captured when possible. Proliferation of PBMCs was shown by the dilution of CD3+/CFSE+ signal with diminished signal intensity (36, 45, 52, 53). The CFSE data were calculated as % proliferated of viable CD3+ population. Suppression was calculated as described previously (45) as 1 − ((Sample − UnstimControl)/(StimControl − UnstimControl)), where UnstimControl are unstimulated PBMCs without MSCs and StimControl is the value of mitogen-stimulated PBMCs without MSCs. In some calculations, MaxSignal, the highest value determined, was used instead of StimControl when the lowest cell numbers of MSCs in co-culture experiment synergized the PBMC immune response.

Proliferation of both MSCs and PBMCs was assessed by using ATP levels as a surrogate for cell number and quantitation of ATP levels using the bioluminescent reagent Cell Titer-Glo 2.0 (Promega Cat # G9242, Madison, WI) as described previously (45, 54). At 72 h, unless specified elsewhere, non-adherent PBMCs were resuspended with gentle trituration and 50 µL was were mixed with 50 µL of Cell Titer-Glo, 2.0 in a white Lumitrac plate (Greiner Bio-One Cat # 655075). Growth or attrition in the co-cultured MSCs was monitored by removal of suspension PBMCs, washing with 250 µL of dPBS, addition of 75 µL of dPBS to each well, followed by 75 µL of Cell Titer-Glo. After 30 min, 100 µL of the MSC samples was transferred to the white Lumitrac plate for assay. Luminescent signal was determined in an appropriate plate reader. The Perkin Elmer Victor3, 1420 multilabel counter, 1 s read, with Wallec software (Perkin Elmer, Waltham, MA), a Molecular Devices Spectramax M5, luminometry setting 5 repeats (Molecular Devices, San Jose, CA), and a GloMax Discover, quick read at 0.3 s integration (Promega, Madison, WI) were used interchangeably. A four-parameter logistic analysis of luminescence from contemporaneous ATP standards was used to determine ATP concentrations in the samples. Suppression was calculated as above.

Cryopreserved hPD-PBMCs were rapidly thawed, washed twice in PBMC media, and resuspended at 1.5 × 10⁶ PBMC per mL of PBMC media as above. In a 6-well plate, 4.5 × 10⁶ PBMCs in 2.5 mL were placed per well and treated as control (no addition), mitogen stimulated (15 µg/mL PHA-P with the addition of 7.5 µL of 5 mg/mL PHA-P) or apoptosis-induced (1 µM staurosporine with the addition of 2.5 µL of 1 mM staurosporine in DMSO; Sigma Aldrich Cat # S6942). At 24 and 48 h, PBMCs were resuspended and analyzed for viability, phosphatidylserine externalization, and caspase 3/7 activity.

After treatment as above, 0.6 × 10⁶ PBMC per tube were centrifuged at 400 × g for 6 min and the pellet was resuspended in 100 µL of Annexin V staining buffer (BioLegend Cat # 422201, San Diego CA) or 100 µL of HBSS (for lactadherin only samples) containing 5 µL of Human TruStain FcX™ (BioLegend Cat # 422302). After a 5-min room temperature incubation in the dark, the cells were then stained with 2.5 µL of Pacific Blue™ anti-human CD45 Antibody (BioLegend Cat#304029 RRID : AB_2174123, San Diego, CA) and 5 µL of Alexa Fluor® 647 Annexin V (BioLegend Cat # 640912, San Diego, CA) with and without 1 µL of FITC-bovine lactadherin (Prolytix, Essence Junction, VT) for 15 min at room temperature in the dark. The addition of 5 µL of 7AAD for a further 5 min in the dark was used to discriminate live from dead cells. Prior to analysis, the samples were diluted with 100 µL of HBSS (Lactadherin only) or 100 µL of Annexin V buffer (Annexin V ± lactadherin samples) and then analyzed on a BD FACS Canto II flow cytometer (BD Biosciences, Franklin Lakes, NJ). Alternatively, doubly labeled cells were analyzed on an AMNIS imaging flow cytometer (Luminex, Austin, TX). Data are expressed as percent of parent.

PBMCs with or without mitogen treatment and MSC co-culture were harvested and pooled from the 96-well plates at the times specified; 1 × 10⁶ cells were centrifuged at 600 × g for 6 min, resuspended in 100 µL of HBSS containing 5 µL of Human TruStain FcX™, and incubated for 5 min at room temperature. PBMCs were then stained with the following markers: 2.5 µL of APC/Cyanine7 anti-human CD3 Antibody (BioLegend cat # 300318, RRID : AB_314054, San Diego, CA) and 2 µL of bovine lactadherin (Prolytix Cat# BLAC-FITC, Essence Junction, VT) with or without 2.5 µL of Brilliant Violet 421™ anti-human CD4 antibody (BioLegend Cat #357424, RRID : AB_2721519, San Diego, CA). Cells were stained for 15 min at room temperature in the dark followed by a 5-min incubation using 5 µL of 7AAD (BioLegend, San Diego, CA). HBSS (100 µL) was added to the samples before analysis on a BD FACS Canto II flow cytometer (BD Biosciences, Franklin Lakes, NJ). Data are expressed as % positive cells of the live population. Suppression was calculated as above.

Caspase 3/7 activity was determined by automated cell counting. After incubation at 37°C for 2, 4, 6, 24, 48, or 72 h, control, mitogen-stimulated, and apoptosis-induced samples were resuspended, aliquoted into 0.6 mL of polypropylene micro-centrifuge tubes and mixed with 12.5 µL of a 1/50 dilution of NucView™ caspase 3/7 reagent in the ViaStain^TM−Cell Fitness Panel (Nexcelom Cat # CSK-V0024-1) for a further 30-min incubation at 37°C before transfer to a Cellaca MX plate and collection of brightfield and fluorescent images on a Cellaca MX automated cell counter using the caspase 3/7 apoptosis assay. Caspase active cells are expressed as percent of total cell number determined by brightfield analysis.

Activity of caspases 3/7, 8, and 9 were determined from singly cultured or co-cultured PBMCs and MSCs incubated in the presence or absence of mitogen as described above. At either 24 or 72 h, PBMCs in each well were triturated to suspend the PBMCs in media and 25-µL aliquots were diluted into 25 µL of dPBS in a white LUMITRAC assay plate; 50 µL of the appropriate caspase assay reagent was added (Promega Caspase-Glo 3/7, Cat # G8091; Caspase-Glo 8, Cat # G8201; and Caspase-Glo 9, Cat # G8211). After 30–60 min, luminescent readings were determined in a dedicated luminometer. Data are expressed as relative luminescent readings (RLU).

Tumor necrosis factor-alpha, interleukin 6, and interleukin 10 (TNFα, IL6, and IL10, respectively) were quantitated by multiplexed ELISA using the Ella Simple Plex platform from Bio-Techne (Minneapolis, MN) according to manufacturer’s recommendations. Conditioned media from stimulated or unstimulated PBMSC-MSC cocultures treated as described above in 96-well plates were frozen, thawed on the date of assay, and centrifuged at 20,000 × g for 5 min before analysis. Samples were diluted threefold in ELISA kit assay buffer; 60 µL from each sample was added to the dedicated ELISA plate and the plate loaded onto the automated ELISA. Cytokine levels were determined after 70 min. Alternatively, a rapid, no-wash ELISA (Lumit TNFα ELISA, Promega) was performed, which relies on proximity interactions of bound antibody to reconstitute a light-generating enzyme that can produce luminescence signals. Samples were either from frozen, thawed, centrifuged media from mitogen-stimulated assays or on samples taken directly from mitogen-stimulated samples without centrifugation and were assayed according to the manufacturer’s recommendations. For thawed frozen samples, 50 µL each of diluted standards or centrifuged media was added to a 96-well Lumitrac plate followed by the addition of 50 µL of a 2× antibody mix and then 60-min room temperature incubation, per the manufacturer’s recommendations. For freshly harvested samples, 80 µL of diluted standards or resuspended PBMCs was transferred to a 96-well Lumitrac plate with 20 µL of 5× antibody mix, per the manufacturer’s recommendations. After incubation, 25 µL of room temperature equilibrated, diluted detection substrate in buffer was added to each well. After brief mixing, and a 3- to 5-min incubation, the luminescence was read on a luminometer. The concentration of TNFα was determined from the linear standard curve. Suppression of the PBMC signal in the presence of MSCs was calculated as described above.

RT-PCR was performed as described previously (54). mRNA was isolated with the RNeasy Mini Kit (Qiagen, Germantown, MD) from washed MSCs after co-culture with PBMCs. cDNA was generated using the RT2 HT First strand kit (Qiagen). mRNA expression of target IDO and the housekeeping gene B2M was quantified by Real-Time PCR (RT-PCR) using validated mRNA primers (Quantitect Primer Assay, Qiagen), cDNA, and the i-Taq Universal SYBR Green Supermix (BioRad, Hercules, CA) on a PTC Tempo Thermal Cycler (BioRad) performed in 96-well clear multi-plates. Specific IDO mRNA expression was normalized to B2M gene expression.

ATP standards were fit to a four-parameter dose–response logistic curve using GraphPad Prism version 8.1.2 for Windows, GraphPad Software, Inc., San Diego, California USA (www.graphpad.com). TNFα standards were calculated internally (Simple Ella platform) or were determined by linear regression using GraphPad Prism, version 8.1.2 for Windows. Significant differences were determined with one- and two-way ANOVA using JMP version 15.2.0, from SAS Institute, Inc., Cary, NC, USA (www.jmp.com). Pairwise comparisons of means were made using Tukey HSD. Significance was set at p < 0.5. Unless otherwise noted, data are presented as average ±standard deviation.

When PBMCs undergo stimulation, the proportion of CD4+/CD3+ PBMCs decreases as CD8+/CD3+ increase (54–57). There are reports that MSCs induce apoptosis in co-culture (58). To determine if the CD3+ or CD3+/CD4+ population undergoing apoptosis might be a useful marker for immunomodulation by MSCs, we examined traditional markers of apoptosis especially for time frames earlier than 72–96 h (59, 60).

Caspase activation can be determined using NucView, a cell-permeant DNA binding dye with a linked DEVD peptide sequence (Asp-Glu-Val-Asp). Cells with caspase activity cleave the DEVD peptide, allowing the dye to bind DNA and display bright green fluorescence. In validating the assay, cells were monitored at 24, 48, and 72 h using an automated cell counter ( Supplementary Figure 1A ). Caspase 3/7 activity is evident 24 h after treatment with 15 µg/mL PHA-P, but not to the level of cells treated with staurosporine (Control, 5.5% ± 1.4%; PHA-P, 17.8% ± 3.6%; staurosporine, 26.5% ± 3.8%). Along with the observed increase in caspase activity, there is a concomitant decrease in viability ( Supplementary Figure 1B ) (Control, 86% ± 3.0%; PHA-P, 63% ± 7.2%; staurosporine, 34.4% ± 6.9%). By 48 h, caspase activity in PHA-stimulated PBMCs is stable (Control, 8.9% ± 1.14%; PHA-P, 21.2% ± 3.4%; staurosporine, 59.3% ± 7.4%) and viability has actually increased. In contrast, virtually all the cells are now dead after treatment with the known apoptosis inducer staurosporine (Control, 83.0% ± 1.6%; PHA-P, 72.3% ± 5.4%; staurosporine, 1.5% ± 2.14%). There is both a dose and time dependence for caspase 3/7 activity in PHA-stimulated PBMCs ( Supplementary Figure 1C ). Having established the validity of this assay in measuring caspase 3/7 activity on PBMCs, the effect of MSCs on PBMCs caspase 3/7 activity was determined. PBMCs were co-cultured with and without MSCs, with and without PHA stimulation. Caspase 3/7 activity by NucView labeling was determined at 2, 4, 6, 24, and 72 h after initiating co-culture. At 2 and 4 h, no significant difference with PHA stimulation is seen; however, by 6 h and for 24 h, PHA-stimulated PBMCs are significantly different from control PBMCs (p < 0.0001 and p = 0.001, respectively), although there was no significant difference with MSC co-culture at these times. There was significant immunosuppression of PBMC caspase activity by MSC co-culture only at the 72-h time points ( Figure 1 ).

Caspase 3/7 is only one family of caspases. Therefore, luminescent-based assays were used to detect activity by other caspases. These assays also rely on the cleavage of caspase specific tetrapeptides (DEVD, caspase 3/7; LEHD, caspase 9; LETD, caspase 8). In these assays, aminoluciferin becomes the substrate for a proprietary thermostable luciferase upon cleavage of the respective tetrapeptide, generating light that is detected with a luminometer. Using these luminescent assays for caspase activity, caspase activity of stimulated PBMCs (15 µg/mL PHA-P) was examined. Triplicate determinations for caspase 3/7, caspase 8, and caspase 9 were run on each sample. Data from three experiments are shown. These experiments were done in co-culture in the presence or absence of three different concentrations of one hBM-MSC preparation that had been allowed to attach for 24 h prior to co-culture with PBMCs to examine the ability of MSC co-culture to suppress caspase activation. Controls included PBMCs in the absence of stimulus and PHA-treated MSCs without PBMCs. Caspase activity was assayed on PBMC samples at 24 and 72 h ( Figure 2 ). In comparison to the previous method, caspase 3/7 showed little activation in stimulated PBMCs at 24 h ( Figure 2A ). Caspase 8 and caspase 9 do show slight increases in activity with stimulation (p = 0.0392 and p = 0.0758, respectively), but hBM-MSCs in co-culture show little suppression of PBMC caspase activity ( Figures 2B, C ). By 72 h after stimulation, PBMC caspase 3/7 levels are indistinguishable from those in unstimulated cells ( Figure 2D ). In this assay, caspase 3/7 activity is only evident when MSCs are present, and it is apparent that MSCs contribute the bulk of the signal, as seen in MSCs cultured alone ( Figure 2D ). However, for both caspase 8 and caspase 9 activity ( Figures 2E, F ), a significant, dose-dependent suppression of caspase activity is seen with MSC co-culture. MSCs co-cultured at the highest concentration significantly suppress caspase activity relative to that seen in PBMCs cultured alone (p < 0.001 for both caspase 8 and 9). While this suggests that monitoring caspase activity could be used in potency assays, the 72-h time frame does not markedly improve other existing potency assays following PBMC stimulation.

Phosphatidylserine (PS) externalization is commonly referred to as a marker of early apoptosis (60–63). However, PS externalization does not necessarily always lead to apoptosis and can instead be a hallmark of cellular shape change (64–66). Upon mitogenic stimulation, PBMCs undergo a “blast” change, accompanied by an increase in side scatter by flow cytometry ( Supplementary Figure 2 ) (55). Monitoring this blast size change has been suggested as a potency assay for MSCs (50). PS staining allows easy discrimination of this shape change. For this assay, we monitored PBMC PS externalization by lactadherin binding. While annexin V staining is the most commonly used stain for PS+, its calcium binding requirement can be problematic under some conditions. Lactadherin is a multifunctional secreted extracellular matrix glycoprotein with a stereo-specific binding affinity (67) for PS+ surfaces. It has a lower threshold affinity for PS+ binding than annexin V (68–71) and, as illustrated in Supplementary Figure 2A , shows more extensive binding to PHA-stimulated PBMCs compared to annexin V, although apoptosis-induced, staurosporine-treated cells show similar levels of annexin V and lactadherin binding. Lactadherin’s pattern of staining is similar to annexin V ( Figure 2B ). Live cells without PS binding show no annexin V binding; live cells staining with lactadherin and not annexin V show a diffuse PS+ staining pattern, while live cells staining with annexin V show similar staining with lactadherin.

The flow cytometry gating strategy for PS externalization is shown in Supplementary Figure 3 . For these studies, the FSA-A/SSC-A gate is widened to detect stimulated lymphocytes ( Supplementary Figure 3A ). Unstimulated ( Supplementary Figure 3A ) or stimulated ( Supplementary Figure 3C ) PBMCs at various time points are stained with fluorescent dyes and antibodies to distinguish live and dead cells, CD3+ cells, CD4+ cells, and PS+ cells with fluorescently labeled bovine lactadherin. With this gating strategy, a dose dependence for PS detection can be seen using two different forms of phytohemagglutinin—PHA-P, with two molecular forms containing both leucoagglutinin and erethroagglutinin activity, and PHA-L, which is the purified phytohemagglutinin-leukocyte reactive subunit. These two mitogens have different sensitivities, and lactadherin staining of PS+ cells shows this dose dependence. Supplementary Figure 4 also shows results of different gating strategies. Gating for both CD4+/CD3+ parent cells does not markedly increase sensitivity of PS+ staining over CD3+ parent cells alone. Supplementary Figures 4A–D demonstrates that this staining does not progress or change over time following treatment. The percentage of PS+ cells at 6 h and 24 h are not significantly different, whether analyzing the entire CD3+ population or the CD4+/CD3+ population subset. This is further illustrated in Supplementary Figure 4E with data from multiple different hPD PBMC preps. Supplementary Figure 4E shows average PS+/CD3+ live cells at 2, 4, 6, 24 and 72 h. At all time points, there is a significant increase in the PS+ population of CD3+ live cells with mitogen stimulation.

Figure 3A illustrates one representative MSC potency test by PS+ staining examined at 2, 4, 24, and 72 h. Levels of stimulated signal do not markedly increase from 2 to 24 h, with only a minor increase at 72 h. BM-MSC co-culture at the highest MSC concentrations relative to PBMC number diminishes PS+ staining, suggesting immunosuppression by the MSCs. Finally, when data from multiple experiments is expressed as PBMC suppression to allow comparison across different assay methods ( Figure 3B ), the BM-MSC suppression of PBMC PS+ signal at 24 h is remarkably similar to both a proliferation assay monitored by ATP at 72 h and by the dilution of CFSE signal in proliferating cells. By two-way ANOVA, there is a difference between these assays that does interact with the MSC concentration; however, these differences are only seen in the mid-range MSC concentrations. No differences exist at the highest MSC concentrations (PBMC : MSC ratios of 2.5:1). The PS+ staining assay most closely resembles the CFSE dilution assay.

PBMCs respond to mitogens with multiple cytokine responses (43, 54, 72–75). For example, co-culture of PBMCs with MSCs results in augmented IL6 cytokine release, increased IL10 release, and decreased TNFα release (54). The time frames commonly assayed range from 3 to 5 days. In a multiplex assay with TNFα, IL6, and IL10, both cytokines show a dose response in PBMCs alone at 24 h, with increasing stimuli resulting in increasing signal ( Figures 4A, C, E ). However, the response to MSC co-culture differs in that PBMCs respond to MSC co-culture with TNFα, a clear-cut decrease in signal ( Figure 4B ); PBMCs increase the secretion of IL10 in the presence of MSCs ( Figure 4D ); and IL6 secretion into the media is complicated by the strong synergy seen with PBMC-MSC co-culture even in the absence of mitogen ( Figure 4F ; Supplementary Figures 5, 6 ). For a potency assay involving IL6 and IL10, it is difficult to develop metrics when there does not appear to be a ceiling. TNFα response is the most straightforward cytokine response, and Chinnadurai et al. (43) found it to have the highest correlation to proliferation. Therefore, we focused on secretion of this cytokine into the media. Here, we assayed the level of TNFα secreted by PBMCs at 2, 4, 6, and 24 h following mitogen stimulation, used this in an MSC potency assay, and compared the results to a mitogen proliferation assay at 72 h. To determine TNFα concentrations, two separate rapid ELISA platforms were used; the Simple Ella automated ELISA platform from BioTechne and a luminescent, no wash ELISA from Promega.

In Supplementary Figure 7 , both time and dose dependence of TNFα in the media were determined by Simple Ella ELISA in the absence of MSC co-culture. Serial dilutions of either PHA-L or PHA-P were used as mitogens for hPD PBMC preparation. There is an expected dose response, and unlike the results for PS externalization, there is also a change in response over time.

Figure 5A shows TNFα concentrations from a representative PBMC : MSC co-culture experiment at 2, 4, 6, and 24 h. At early times (2–6 h after stimulation), low levels of TNFα are detected and co-culture with MSCs has little effect even at the highest relative level of MSCs (2.5:1: PBMC : MSC). A dose-responsive suppression of secreted TNFα by stimulated PBMCs is pronounced by 24 h. This signal is independent of hPD PBMC preparation. In Figure 5B , suppression data from multiple experiments performed in triplicate using five different hPD PBMCs show the same response for TNFα release at 24 h. This assay is robust and reproducible with an averaged intra-experimental error of 3.5%, an inter-experimental variation of 10.9%, and an average 27 fold increase in signal from control to stimulated values. The aggregated data for one BM-MSC preparation assayed both by TNFα release (at 24 h) and by PBMC proliferation (at 72 h) are graphed in Figure 6A . Both automated ELISA and a luminescent ELISA assay were compared with a 72-h proliferation assay utilizing ATP values of PBMCs. By two-way ANOVA, while there is the expected significant difference for PBMC : MSC ratio (p < 0.0001), no difference exists between assay type (p = 0.2369). A correlation of r ²= 0.74 is found ( Figure 6B ).

Ribeiro et al. (76) had developed an assay for LPS-driven TNFα production by monocytes. Therefore, we investigated whether the TNFα signal was derived from monocytes in the PBMCs. We performed a simple monocyte depletion of the PBMCs by incubating them overnight on tissue culture plastic, allowing the monocytes to attach. The next day, the remaining PBMCs were removed from the flask, washed, and compared in co-culture experiments with PBMCs freshly thawed and washed from cryopreservation. Figures 7A–D show the results of a representative experiment, performed in triplicate. Overnight incubation to monocyte depleted the PBMCs and lowered the initial background levels of TNFα secretion to near zero ( Figure 7A ). The time course from 2 to 24 h shows a similar pattern to non-monocyte-depleted PBMCs ( Figure 7B vs. Figure 5A ) and both 24-h suppression by TNFα secretion and 72-h suppression by ATP assay are similar ( Figures 7C, D , respectively). Thus, overnight incubation to deplete monocytes results in a PBMC population with an increase in TNFα secretion of ~260-fold above baseline vs. freshly thawed PBMCs with an increase in TNFα secretion above baseline of ~3.4 fold; the ultimate signal of TNFα secretion at 24 h for the monocyte-depleted cells is ~71% of the hPD PBMC preparations, and immunosuppression values show no significant difference between the two populations (average of two experiments performed in triplicate).