Human pluripotent stem cells, OP9 cells, MS-5 cells and UCB CD34⁺ cells

hiPSC L71.019 [15], H9 hESC and H1 hESC [16] were cultured on Matrigel coated plates, in mTeSR1 media (Catalog 85850, STEMCELL technologies, Vancouver, Canada). Cells were weekly passed as clumps by a non-enzymatic method (Catalog 130-104-688, XF passaging solution, Miltenyi GmbH, Bergisch-Gladbach, Germany). Cells were regularly tested for mycoplasma. hESC were used under agreement from Agence de la Biomédecine RE17-007. hiPSC L71.19 were derived from fibroblasts obtained from Lonza (Catalog CC-2511, Lonza, Basel, Switzerland), originally prepared through anonymised skin samples.

OP9-GFP stromal cells were kindly provided by Juan-Carlos Zuniga-Pflucker, University of Toronto. OP9 cells were cultured in OP9 media consisted in αMEM (Catalog 22561–021, Gibco; Thermo Fisher Scientific, Inc., Waltham, MA, USA) containing 20% foetal calf serum (FCS, Catalog 10270–106, Gibco; Thermo Fisher Scientific, Inc.), 1% penicillin/streptomycin (Catalog 15140–122, Gibco; Thermo Fisher Scientific, Inc.) on gelatin coated plates.

MS-5 stromal cells were obtained from DSMZ (ACC 441, DSMZ, Brunswick, Germany). MS-5 cells were cultured in αMEM containing 10% FCS, 1% penicillin/streptomycin on gelatin coated plates. All cells were cultured at 37°C under 5% CO2 and atmospheric O2 conditions.

Human umbilical cord blood (UCB) was obtained from Nantes University Hospital, after an informed and written consent. The human biological samples were collected during the patients care. These samples were integrated into the collections of human biological samples attached to the "Immunology" research program declared on 21/12/2012 under the n° DC-2012-1555 and in the following amending declarations (DC-2013-1832; DC2014-2206 and DC-2017-2987 currently pending) at the Ministry of Research and having obtained a favorable decision from the CPP Ouest IV on 06/03/2013. UCB CD34⁺ cells were isolated by positive selection on magnetic columns according to the manufacturer’s instructions (Catalog 130-046-702, CD34 MicroBead Kit, Miltenyi Biotec).

B cell differentiation

The first step was to generate CD34⁺ progenitors. Confluent hPSCs were harvested after dispase treatment (Catalog 07913, STEMCELL technologies) 5min at 37°C. Then, 1–1.5x10⁶ cells were seeded as clumps in 10cm dishes on a monolayer of over-confluent OP9 stroma. Cells were cultured as previously described [14, 17] in αMEM containing 20% FCS, 1% penicillin/streptomycin and 100μM monothioglycerol (MTG) (Catalog M1753, Sigma-Aldrich, Saint-Louis, MO, USA). Total culture media was changed the first day to remove unattached cells and 50% of the media was systematically renewed at Day 4, 6 and 8. According to the different experiments, media was supplemented with BMP4 (Catalog 120-05ET, Peprotech, Rocky Hill, CT, USA), VEGF (Catalog 130-094-030), SCF (Catalog 130-096-695), Flt3L (Catalog 130-096-479) and IL-3 (Catalog 130-095-070) (all four from Miltenyi Biotec) respectively. At Day 10, cells were harvested by treatment with Trypsine-EDTA 0.05% during 15-20min at 37°C. After filtration through 40μm pores, CD34⁺ cells were isolated by positive selection on magnetic columns according to the manufacturer’s instructions.

CD34⁺ cells were then seeded onto a monolayer of MS-5 stromal cells. MS-5 cells were previously irradiated at 50Gy one day before and seeded at a density of 5x10⁵ cells on gelatinised MW6 plates well. 5x10⁴ CD34⁺ cells were seeded per well in 4ml of culture medium and cultured for 21 days at 37°C. Culture media consisted in αMEM containing 10% FCS, 1% penicillin/streptomycin, 100μM MTG and SCF (50ng/ml), Flt3L (50ng/ml), IL-7 (20ng/ml, Catalog 130-095-363) and IL-3 (10ng/ml) (all from Miltenyi Biotec). Culture media was changed every 5 days with the previous cytokine except IL-3.

Flow cytometry analysis

For CD34⁺ cell characterization, differentiated hPSCs at Day 10 and UCB CD34⁺ cells were stained with Fixable Yellow Dead Cell (Catalog L34967, Invitrogen; Thermo Fisher Scientific, Inc.) and the following antibodies CD34-PECy7 (Catalog 560710, clone 581; BD Biosciences, San Jose, CA, USA), CD43-APC (Catalog 560198, clone 1G10; BD Biosciences), CD31-PE (Catalog 555446, clone WM59; BD Biosciences) and CD45-APC-Cy7 (Catalog 557833, clone 2D1; BD Biosciences). For B cell characterization, cells were stained with Fixable Yellow Dead Cell (Invitrogen; Thermo Fisher Scientific, Inc.) and the following antibodies CD19-BUV395 (Catalog 563549, clone SJ25C1; BD Biosciences), CD10-BB700 (Catalog 745745, clone W8E7; BD Biosciences), CD45-PECy7 (Catalog 557748, clone HI30; BD Biosciences), CD34-BV711 (Catalog 745543, clone 8G12; BD Biosciences), IgM-FITC (Catalog 555782, clone G20-127; BD Biosciences) and CD79a-BV421 (Catalog 562852, clone HM47; BD Biosciences). Extracellular staining was performed during the viability staining in order to limit the loss of cells with 5.10⁴ cells in PBS during 20min at 4°C. Intracellular staining of CD79a was performed in Perm Wash Buffer (Catalog 88-8824-00, eBioscience, Vienna, Austria) (30 min, 4°C) after fixation/permeabilization (eBioscience) (30 min, 4°C). Supervised analysis was performed with FlowJo v.10 software on CD43, CD45 and CD31 on CD34⁺ gated cells after doublet and dead cells exclusion. Fluorescence minus one (FMO) was used to gate on marked cells.

Gene expression analysis

At the different steps of the differentiation protocols, RNA was extracted from total cells with a minimum of 3.10⁵ cells using the RNeasy Kit (Catalog 74106, Quiagen, Hilden, Germany) according to the manufacturer’s instructions. Reverse transcription was performed with 0.5μg of RNA according to the manufacturer’s instructions (Catalog 4368814, High Capacity cDNA Reverse Transcription Kit, Applied Biosystems; Thermo Fisher Scientific, Inc). Expression analysis of PAX5, VPREB1, IGLL1, IL-7RA, EBF1, RAG1, DNTT and POU5F1 (OCT4) was performed using TaqMan™ Gene Expression Assays (Applied Biosystems, Thermo Fisher Scientific, Inc). References of TaqMan probes are described in Table 1. PCR reactions were carried out with cDNA diluted to 1/5 in a final volume of 10μl to avoid PCR inhibition. Gene expression levels were calculated using the 2^-ΔΔCT method [18]. Gene expressions were reported to the expression of both the housekeeping genes GAPDH and HPRT and to a calibrator with undetermined CT or CT with a value higher than 35, set to 40. Gene expression analysis were also performed on tonsil B cells as control of B cell gene expression. Human tonsils were obtained from the Plateau technique médico-chirurgical (PTMC) of the Nantes University Hospital after an informed and written consent was obtained. Tonsil cells were mechanically isolated freshly after their reception. Total B cells were then purified by negative selection on magnetic columns according to the manufacturer’s instructions (B Cell Isolation Kit II; Miltenyi Biotec).

Table 1

Target genes	References
PAX5	Hs00277134_m1
VPREB1	Hs00356766_g1
IGLL1	Hs00252263_m1
IL7R	Hs00902334_m1
EBF1	Hs01092694_m1
RAG1	Hs00172121_m1
DNTT	Hs00172743_m1
POU5F1 (OCT4)	Hs00999632_g1

Statistical analyses

Statistical analyses were performed using GraphPad Prism (GraphPad Software, San Diego, CA). Significance of difference between nonparametric variables were assessed with a Mann-Whitney test and the differences were considered statistically significant at p < 0.05. All results are expressed as means ± SEMs.

Assessing B cell differentiation

Our goal was to generate B cells from hPSCs. To do so, we started by implementing a published protocol, outlined in Fig 1A [14]. Briefly, hPSCs (H9 hESC and hIPSC L71.019) were differentiated toward CD34⁺ hematopoietic progenitors during 10 days onto a monolayer of OP9 stromal cells. Then, CD34⁺ cells were cultured onto a monolayer of MS-5 stromal cells for B cell differentiation for 21 days, in the presence of IL-7, IL-3, SCF and Flt3L. B cell differentiation was measured by the frequency of CD10⁺CD19⁺ B cells obtained after differentiation. CD34⁺ cells isolated from umbilical cord blood (UCB) were cultured in parallel in the same conditions, as positive control. Starting from H9 hESC (Fig 1B) and from H1 hESC (S1A Fig), we obtained 5% of CD34⁺ and they did not express CD43, a crucial intermediate phenotype for B cell differentiation [14]. This resulted in an impairment of B cell differentiation considering that less than 0.2% of CD10⁺CD19⁺ B cells were generated after 21 days of differentiation from these CD34⁺ (Figs 1C and S1B). In contrast and using the same protocol and conditions, UCB CD34⁺ cells efficiently engaged into CD10⁺CD19⁺ B cells (7.1%) at Day 21, without IgM expression (Fig 1C).

Fig 1

CD34⁺ generation from hPSC without cytokines and their low potential in B cell differentiation.

(A) Protocol of hPSC differentiation into B cells as previously described [14]. (B) CD34⁺ differentiation from H9 hESC onto OP9 stroma at Day 10. CD34/CD43 expression analyzed by flow cytometry on viable differentiated hESC is shown in the left as a dot plot from one representative experiment and the proportion of CD34⁺ cells (in black) and CD34⁺CD43⁺ (in red) are shown on the right panel, n = 9 experiments. (C) B cells differentiation at Day 21 from UCB CD34⁺ cells (Upper panel) and CD34⁺-derived hESC (middle panel). CD10⁺CD19⁺ B cells proportion and their IgM expression was analyzed on viable CD45⁺ cells, dot plot from one representative experiment is shown. Histogram of CD10⁺CD19⁺ B cells proportion is presented in the lower panel (n = 3 experiments from both UCB CD34⁺ cells and hESC-derived CD34⁺).

Cytokines trigger the generation of CD34⁺CD43⁺ hematopoietic progenitors with low ability to differentiate into CD19⁺ B cells

Considering these first data, we set out to improve the generation of CD34⁺CD43⁺ on OP9 monolayers during CD34⁺ differentiation. Our strategy consisted in adding cytokines to the OP9 media, to trigger optimal hematopoietic differentiation from hESC. VEGF, SCF and Flt3L were kept in the medium from Day 1 to Day 10 at 10ng/ml with or without BMP4 (5ng/ml) and IL-3 (10ng/ml) as detailed in Fig 2A. We found that a significant increase in CD34⁺CD43⁺ proportion using VEGF, SCF, Flt3L and BMP-4 combined with IL-3 from Day 4 to Day 10 (Fig 2B, p = 0.0364) from 0.05% to 0.8%. Interestingly, the use of VEGF, SCF and Flt3L combined with IL-3 from Day 4 to Day 10 and without BMP4 permits a higher improvement of CD34⁺ differentiation, increasing the yield of CD34⁺CD43⁺ up to 6%, not significantly at 10ng/ml because of the small sample size (Fig 2B) but significantly at 25ng/ml (S1C Fig, p = 0.0091). However, B-cell differentiation conducted with these progenitors until Day 21 resulted in 0.2% of CD10⁺CD19⁺ B cells (Fig 2C). These results were also obtained when all cytokines were set to 25ng/ml during the CD34⁺ differentiation (S1C and S1D Fig) suggesting that hESC-derived CD34⁺CD43⁺ cells were not in a state permissive to B cell differentiation.

Fig 2

Cytokines allow CD34⁺CD43⁺ differentiation from stem cells.

(A) Table of the cytokine conditions tested according their presence from Day 1 to Day 3 and Day 4 to Day 10 for the CD34⁺ differentiation (cytokines were used at 10ng/ml except for BMP4 at 5ng/ml). (B) Percentage of CD34⁺CD43⁺ cells at the end of the CD34⁺ differentiation according to the cytokine conditions used (n = 2 or 3 independent experiments) or without cytokines as control (Ctrl, n = 9 independent experiments). The significance of differences between the condition C and the Ctrl was determined using a Mann-Whitney Test; *p < 0.05. (C) Percentage of CD10⁺CD19⁺ B cells at Day 21 of B cell differentiation from hESC-derived CD34⁺ with (Condition D; n = 2 independent experiments) or without cytokine (Ctrl, n = 3 independent experiments).

To further improve our protocol of differentiation, we kept the same cytokine cocktail than the one allowing the best CD34⁺CD43⁺ differentiation (i.e. VEGF, SCF, Flt3L constant and combined with IL-3, from Day 4 to 10), and we tested different concentrations of the cytokines. Starting from H9 hESC, we found that the concentration of 25ng/ml significantly increase the proportion of CD34⁺ (p = 0.0364) and CD34⁺CD43⁺ progenitors (p = 0.0091) (Fig 3A). However, while not significant because of the small sample size (n = 2), the concentration of 10ng/ml for our cytokine cocktail triggered the highest proportion of CD34⁺ (22.8%) and CD34⁺CD43⁺ progenitors (4.7%) (Fig 3A), without increasing the final CD10⁺CD19⁺ B cell proportion (< 0.2%) (Fig 3B). Starting from hIPSC, there was no difference in CD34⁺CD43⁺ proportion according to the different concentrations of the cytokine cocktail used (Fig 3C), and as previously this resulted in only a low level of final CD10⁺CD19⁺ B cell proportion (0.7% at 10ng/ml, Fig 3D).

Fig 3

Cytokines at 10ng/ml permit the best B cell differentiation.

hESC (Upper panels) and hIPSC (lower panels) were used for CD34⁺ differentiation with our protocol of cytokines addition (the cytokine condition D, as detailed in Fig 2A) at different concentration. (A, C) Histograms show the proportion of CD34⁺ (left panel) and CD34⁺CD43⁺ (right panel) after CD34⁺ differentiation from (A) hESC with cytokines (n = 1 to 3 independent experiments) or without cytokines as control (Ctrl, n = 9 independent experiments) and from (C) hIPSC (n = 1 to 4 independent experiments) The significance of differences between the condition at 25ng/mL and the Ctrl was determined using a Mann-Whitney Test; *p < 0.05; **p < 0.01. (B, D) B cells proportion at Day 21 after B cell differentiation from the corresponding CD34⁺ cells is shown from (B) hESC-derived CD34⁺ cells with cytokines (n = 1 or 2 independent experiments) or without cytokines as control (Ctrl, n = 3 independent experiments) during CD34⁺ differentiation and from (D) hIPSC-derived CD34⁺ (n = 1 to 3 independent experiments).

Altogether our results indicate that modifying cytokine concentrations efficiently increased the CD34⁺CD43⁺ proportion in H9 hESC but did not significantly improve the subsequent CD10⁺CD19⁺ B-cell differentiation suggesting that CD43 expression by CD34⁺ progenitors is not sufficient to engage them toward B cell differentiation.

hPSC-derived CD34⁺ generated Pro-B and Pre-B cells on MS-5 stroma

During B cell differentiation, CD45⁺CD10⁺ common lymphoid progenitors (CLP) differentiate into CD79a⁺CD19^- Pre-Pro B cells and then into Pro-B and Pre-B cells included in the CD79a⁺CD19⁺ fraction. While using our protocol to generate CD34⁺ with the same cytokine condition at 10ng/ml, we wanted to better follow and appreciate their B cell differentiation. We thus analyzed during B cell differentiation, CD10⁺CD19⁺ B cell proportion among CD45⁺ cells and we looked at CD79a expression among the CD45⁺CD10⁺ lymphoid progenitors allowing us to distinguish CLP which are CD79a^-CD19^-, CD79a⁺CD19^- Pre-Pro B cells and the CD79a⁺CD19⁺ fraction composed of the Pro and Pre-B cells as detailed in Fig 4A. Starting B cell differentiation from UCB CD34⁺ cells, we obtained 30% CLP, 50.74% Pre-pro B cells and 15% Pro-B and Pre-B cells among the CD45⁺CD10⁺ cell population (Fig 4B). During B cell differentiation of hESC-derived CD34⁺, we observed that all CD45⁺CD10⁺ were CLP at Day 7 and their frequency decreased to 82% at Day 21. Pre-pro B cells appeared between Day 7 and Day 14 and increased from 6.47% to 10.9% of CD45⁺CD10⁺ cells, from Day 14 to Day 21. Finally, the fraction consisting of Pro-B cells and Pre-B cells represented 1.25% of CD45⁺CD10⁺ at Day 14 and 10.7% of CD45⁺CD10⁺ at Day 21 (Fig 4C). These data show that CD34⁺ hematopoietic progenitors were engaged into B lymphopoiesis reaching the stage of Pro-B and Pre-B cells, albeit in limited amounts (S3 Fig).

Fig 4

Gating strategy of B cell progenitors.

(A) B cell characterization during B cell differentiation of UCB CD34⁺ cells using CD10⁺CD19⁺ total B cells gated on CD45⁺ (upper panel) and using CD79a^-CD10^- CLP, CD79a⁺CD10^- Pre-pro-B cells and CD79a⁺CD19⁺ both Pro-B and Pre-B cells gated on CD45⁺CD10⁺ (lower panel). (B) B cell progenitors proportion in 2 independent experiments. (C) B cell progenitors proportion at Day 7 (D7), 14 (D14) and 21 (D21) of B cell differentiation from hESC-derived CD34⁺ (n = 1 or 2 independent experiments).

hESC-derived cells express genes implicated in B cell commitment during B cell differentiation

We analyzed gene expression known to be important for B cell commitment in the course of B cell differentiation from CD34⁺ progenitors after 7, 14, and 21 days of culture on MS-5 stroma. We thus extracted RNA of total cells at these different time points and analyzed the transcripts expression of the following genes PAX5, EBF1, VPREB1, IGLL1, IL-7RA and RAG1 (Fig 5). While OCT4 (POU5F1) was used as pluripotency control, undifferentiated H9 hESC and tonsil B cells served as control of gene expression. At Day 21, B cell differentiation from UCB CD34⁺ cells was characterized by a low gene expression for PAX5, EBF1, IL-7RA and OCT4 and a high gene expression for VPREB1, IGLL1 and RAG1 (Fig 5A). Regarding differentiated hESC from Day 7 to Day 21, they harbor a low PAX5 gene expression, a decreasing gene expression for EBF1 and a slight increase for IL-7RA with their overall expression level similar to those obtained from UCB CD34⁺ cells (Fig 5B). We observed an expression of VPREB1 gene starting at Day 14 while IGLL1 gene expression is present at Day 7 and decreased during the differentiation. This is associated with an increase in RAG1 expression from Day 7 to Day 21, known to be important for immunoglobulin (Ig) gene rearrangement leading to the formation of the Pre-BCR. IL-7RA, IGLL1 and VPREB1 expression levels were lower than during B cell differentiation from UCB CD34⁺ cells, this could be explained by the lower number of B cells generated from hESC. Finally, during B cell differentiation from hESC-derived CD34⁺ cells, OCT4 was not expressed compared to the undifferentiated hESC consistent with their exit from pluripotency. Thus, while we obtained few B cells from H9 hESC, these results are concordant with a normal process of early B cell differentiation (S3 Fig).

Fig 5

Gene expression at different time point of B cell differentiation.

Gene expression analysis by Taqman expression assay of PAX5, EBF1, VPREB1, IL-7RA, IGLL1, RAG1 and OCT4 were performed on total cells at different time point during B cell differentiation. Histograms of 2^-DDCT value display gene expression analysis of cells (A) at Day 21 of B cell differentiation from UCB CD34⁺ cells (n = 2 to 3 independent experiments) and (B) at Day 0 (D0), 7, 14 and 21 of B cell differentiation from hESC-derived CD34⁺ with undifferentiated hESC and tonsil B cells (LB) as controls (n = 1 to 2 independent experiments).

UCB CD34⁺ cells exhibit a CD31^intCD45^int phenotype and differentiate into CD19⁺ B cells

The CD34⁺ differentiation is a crucial step to obtain hematopoietic progenitors in sufficient amount and abilities to generate B cells on MS-5 stroma. While UCB CD34⁺ cells succeeded to allow great B cell differentiation on MS-5 stroma (7%), CD34⁺ generated from both hESC and hIPSC did not reach this efficiency. To better characterize the progenitors that could differentiate into B cells, we extended our profiling of the CD34⁺ cells and we measured the expression of CD31, an endothelial marker and CD45, a marker of hematopoietic cells expressed after CD43 during hematopoiesis on UCB CD34⁺ cells. We observed that UCB CD34⁺ cells expressed CD43 and were CD31^intCD45^int (Fig 6A). In contrast, CD34⁺ differentiated from hESC and hIPSC using our same protocol as before with cytokines at 10ng/ml, expressed few CD45 and were mainly CD31^hi (Fig 6B). In addition, variation in cytokine concentrations did not increase CD31^intCD45^int proportion in CD34⁺ differentiated neither from hESC nor hIPSC (Fig 6C). These results on UCB CD34⁺ cells suggest that hematopoietic progenitors possess a CD31^intCD45^int phenotype that should be reached by differentiated hPSC before engaging B cell differentiation.

Fig 6

CD31^intCD45^int are important progenitors for B cells differentiation.

Characterization of CD34⁺ progenitors by their expression analysis of the following markers CD45, CD43 and CD31. FACS analysis was performed from (A) UCB CD34⁺ cells (Upper panel) with unstained cells as control (lower panel), dot plot representative of 2 experiments and from (B) hESC-derived CD34⁺ (Upper panel) and hIPSC-derived CD34⁺ (lower panel) using our CD34⁺ differentiation protocol with cytokines at 10ng/ml, dot plot representative of 2 experiments for hESC and 4 experiments for hIPSC. (C) Histograms show CD31^intCD45^int proportion in hESC (upper panel) and hIPSC (lower panel) at Day 10 of CD34⁺ using our CD34⁺ differentiation protocol with cytokines at different concentrations (n = 1 to 4 independent experiments).

On this basis, we hypothesized that CD34⁺ progenitors differentiated from hPSC were not at an optimal differentiation stage before engaging the B cell differentiation. Therefore, we stopped the CD34⁺ differentiation from hIPSC at 8, 10 and 12 days and compared the CD34⁺ progenitor’s ability to differentiate into B cells. As expected, we observed an increase in the proportion of CD43⁺CD45^int, CD43⁺CD45^hi, CD31^intCD45^int and a decrease in the proportion of CD43⁺CD45^- from Day 8 to Day 12 of CD34⁺ differentiation (Fig 7A). Interestingly, the proportion of CD31^hiCD45^int increased faster than the proportion of CD31^intCD45^int and reached a similar proportion at Day 12. However, despite a higher proportion of CD31^intCD45^int at Day 12, CD10⁺CD19⁺ B cells proportion was not increased upon B cell differentiation from these hIPSC-derived CD34⁺ cells (Fig 7B).

Fig 7

Day 10 of CD34⁺ differentiation permits the better B cell differentiation.

(A) CD34⁺ differentiation from hIPSC, with our protocol of cytokine addition (cytokine condition D, as detailed in Fig 2A) at 10ng/ml, was stopped at Day 8, Day 10 and Day 12 to analyze the expression of CD34⁺ cells for the following markers CD45, CD43 and CD31. Dot plots from one experiment for Day 8 and Day 12 and 4 independent experiments for Day 10 are shown in the left panel. Histogram of the CD31^intCD45^int proportion among the CD34⁺ at different time of the CD34⁺ differentiation is shown. (B) Histogram represents the B cell proportion at the end of the B cell differentiation using the corresponding hIPSC-derived CD34⁺ (n = 1 to 3 independent experiments).

Altogether, our results point out the difficulties to obtain hematopoietic progenitors with strong lymphoid abilities from hPSC. We obtained only low B cell proportion from hESC and hIPSC. We show that while CD43 expression is needed in hPSC-derived CD34⁺ cells, CD43 is not sufficient for differentiation into B cells. UCB CD34⁺ cells easily differentiate into B cells and harbor a CD31^intCD45^int phenotype.