Exopolysaccharide-treated Dendritic Cells Effectively Ameliorate Acute Graft vs Host Disease

Abstract

Graft versus host disease (GvHD) is one of the primary and often lethal complications of allogenic hematopoietic stem cell transplantation (HSCT). Prophylactic regimens for GvHD are given as standard pre-transplantation therapy; however, up to 50% of these patients develop acute GvHD (aGvHD) and require additional immunosuppressive intervention. Using a mouse GvHD model, we previously showed that injecting mice with exopolysaccharide (EPS) from Bacillus subtilis prior to GvHD induction significantly increased 80-day survival after transplantation of complete allo-MHC mismatched cells. To ask whether EPS might also ameliorate GvHD in humans, we used humanized NSG-HLA-A2 mice and induced GvHD by i.v. injecting A2^neg human PBMC. Because we could not inject human donors with EPS, we transferred EPS-pretreated dendritic cells (DC) to inhibit aGvHD. We derived DC from CD34⁺ human cord blood cells, treated them with EPS, and then injected them together with PBMC into the NSG-HLA-A2 mice. We found that all mice that received untreated DC were dead by day 35, whereas 25% of mice receiving EPS-treated DC survived. This DC cell therapy could be readily translatable to humans because we can generate large numbers of human EPS-DC and use them as an “off the shelf” treatment for patients undergoing HSCT.

Introduction

Hematologic malignancies are often treated with allogeneic hematopoietic stem cell transplantation (HSCT) with curative intent. However, HSCT is complicated by acute graft-vs-host disease (aGvHD), a severe, potentially lethal complication initiated when donor alloreactive T cells attack host cells and organs^1,2. Furthermore, aGvHD is a risk factor for developing chronic (cGvHD)³; thus, preventing aGvHD dramatically decreases the likelihood of developing cGvHD. Although prophylactic regimens for aGvHD are standard pre-transplantation therapy, up to 50% of these patients treated with calcineurin-based regimens develop grade II or higher aGVHD (grade III-IV)¹, and require additional immunosuppressive intervention⁴. GvHD is caused by activation of donor allogeneic T cells in response to molecules on host cells. This T cell activation results in the recognition and the destruction of the recipient tissues and organs by the donor immune effector cells primarily affecting the skin, liver, and gastrointestinal tract^5,6. T cell activation also causes inflammation and a cytokine storm that further exacerbates disease⁷. Two standard regimens used clinically to prevent this complication include: a calcineurin inhibitor with short course methotrexate, or post-transplant cyclophosphamide (PTCy) with a calcineurin inhibitor and mycophenolate^1,6,8. While prophylaxis with calcineurin inhibitors combined with methotrexate has been the customary way to prophylaxis patients for decades^1,6, post-transplant high dose cyclophosphamide given on days +3 and +4 after transplant has recently been shown to decrease the rates of both acute and chronic GVHD in a Phase 3 comparison trial to tacrolimus and methotrexate⁸. However, in this pivotal trial, there was no difference in disease-free or overall survival for the PTCy group, as higher rates of organ failure, primary graft failure and hemorrhage likely directly related to cyclophosphamide toxicity, offset the benefits from a lower rate of deaths from acute and chronic GVHD. Thus, a new method of aGVHD amelioration without chemotherapy toxicity such as the novel immunosuppressive method described here could improve not only aGVHD rates but overall survival.

Cellular therapy is an attractive treatment for aGvHD⁴. However, cell therapy in animal models using CD34⁺ monocytes⁹, regulatory T cells¹⁰, mesenchymal stem cells¹¹, DC^12–14, or myeloid-derived suppressor cells¹⁵ has had limited success thus far¹⁶. DC treated with vitamin D can delay GvHD in humanized NSG mice, indicating that DC have the potential to protect from GvHD¹².

We recently discovered a probiotic exopolysaccharide (EPS) from Bacillus subtilis, a harmless soil bacterium, that has potent cellular immune suppressive activities in a variety of disorders, including allergy¹⁷, colitis^18–20, sepsis^21,22 and GvHD²³. EPS induces tolerogenic DC and anti-inflammatory macrophages, which inhibit proliferation of activated T cells^17–20,23. In mice, transfer of in vitro EPS-treated bone marrow-derived DC (BMDC) ameliorates allergic eosinophilia¹⁷, and peritoneal macrophags isolated from EPS-treated mice and transferred to recipient mice protect from disease caused by an enteric pathogen^17,20. In a mouse GvHD model, injection of EPS prior to GvHD induction increased 80-day survival from 10% (in control, PBS-treated mice) to 70% (in EPS-treated mice) after transplantation of complete allo-MHC mismatched cells²³. Here we investigated how EPS-treated human DC affect human allo- and xeno-reactive T cells and whether they can ameliorate GvHD in humanized mice.

Materials and Methods

Mice and reagents

We housed all mice in pathogen-free conditions in the animal facility at Loyola University Chicago (Maywood, IL). We purchased NSG-HLA-A2 (strain # 009617) mice from the Jackson Laboratory (Bar Harbor, ME) or bred them in-house. We utilized age- and sex-matched 8- to 12-wk-old mice for all experiments, and all experiments were performed according to protocols approved by the Institutional Animal Care and Use Committee at Loyola University Chicago. Base media and supplements were purchased from Life Technologies (Grand Island, NY); StemSpan SFEM II media, human AB serum and Stemregenin1(SR1) from STEMCELL Technologies (Vancouver, Canada); bacterial media from Becton Dickinson (Franklin Lakes, NJ), all antibodies (Ab) from BioLegend (San Diego, CA) or eBioscience (San Diego, CA); and all cytokines from PeproTech (Thermo Fisher Scientific, Waltham, MA).

Preparation of B. subtilis–derived EPS

We prepared EPS from B. subtilis DS991 (sinR:tasA), which overproduces and secretes EPS. ΔEPS is carbohydrate prepared from B. subtilis DS5187 (sinRtasAepsH mutant), which does not produce EPS^19,24. Daniel Kearns and Andrew Burrage (Indiana University, Bloomington, IN) generously provided both strains. We plated bacteria at OD₆₀₀ 0.6–0.7, on Luria–Bertani agar plates (150 mm) for 4 h, harvested bacterial lawns, and treated supernatants with DNase (67 μg/ml), RNase (330 μg/ml), and protease K (40 μg/ml). We precipitated carbohydrates with 75% EtOH and, after resuspension, EPS was purified by gel filtration on Sephacryl S-500. We pooled carbohydrate-positive fractions identified by a modified phenol sulfuric acid assay^25,26, dialyzed them extensively, and air-dried them. We assessed EPS purity by the lack of detectable protein or nucleic acid (O.D._260/280), and the absence of cytotoxicity in vitro. We sterile-filtered all EPS preparations and confirmed by ELISA that they had undetectable levels of endotoxin (LPS). Each preparation was assessed for its capacity to induce peritoneal M2 macrophages, as previously described²⁰, or to inhibit MLR, as described below.

CD34 cell expansion and derivation of dendritic cells

We expanded umbilical cord blood CD34+ cells (Lonza Group Ltd., Switzerland) for 7 days in expansion medium (StemSpan SFEM II, 0.75μM Stemregenerin, 1x penicillin-streptomycin, 50 ng/ml of Flt3L, SCF, TPO and IL-6). After one week, we froze cells (10⁶/ml) in Cryostor media (StemCell Technology, Seattle, WA). We thawed one-week-expanded CD34+ cells and expanded them for another week in expansion medium. The two-week-expanded CD34+ cells (100,000/ml) were plated in DC-derivation medium consisting of 1:1:1 mixture of aMEM:IMDM:RPMI with 10% FBS, 2% human AB serum with 100 ng/ml hGM-CSF, and 50 ng/ml hIL-4. One-half of the medium was exchanged every 2 days. On day 12 we pre-treated cells with EPS or PBS overnight before using them in aGvHD experiments or in mixed lymphocyte reaction (MLR) testing. Monocyte-derived DCs were generated by isolating CD14⁺ cells from human PBMCs using MojoSort Human CD14⁺ Monocytes isolation kit (Biolegend, San Diego, CA) according to the manufacturer’s instructions, then culturing them in RPMI with 10% FBS with 50 ng/ml of hGM-CSF and 25 ng/ml hIL-4 for 10 days. We exchanged one-half of the medium every other day.

GvHD and MLR

We purchased peripheral blood mononuclear cells (PBMC) isolated from heathy HLA-A2 negative donors (STEMCELL Technologies, Vancouver, Canada). To induce GvHD in 8- to 12-wk-old NSG-HLA-A2 mice, we conditioned them with 2.5 Gy from an x-ray source (RS 2000 Biological Research Irradiator) and 4 hrs later i.v. injected them with 4×10⁶ PBMCs alone, or as a mixture of 4×10⁶ PBMCs with 10⁶ CD34-DCs from a different donor; the CD34-DCs were pretreated with EPS (60 μg/ml) or PBS for 16 hrs. For MLR, we labeled PBMCs with 5 μM CellTrace Violet (CTV) (Thermo Fisher Scientific, Waltham, MA) according to the manufacturer’s instructions and cultured them either alone (2.5 × 10⁵ cells, negative control) or with CD34-derived DCs (7.5 × 10⁴ in 96-well U-bottom plates) with EPS (60 μg/ml) or ΔEPS (60 μg/ml). In some experiments, we added 3^rd party irradiated PBMCs to the cultures as initiators. Four days later, T cell proliferation and activation were measured by flow cytometry. We evaluated two different CD34⁺ cell donors along with multiple (> 8) hPBMC donors.

Flow cytometry

For flow cytometry, we treated cells with anti–CD16/32 Fc Block, then stained with Ab specific for CD4, CD8, CD25 and CD44. Cells were analyzed using a FACSCanto II or LSRFortessa flow cytometer (BD Biosciences). We measured proliferation based on the dilution of the CTV dye, as determined by flow cytometry, and analyzed the data using FlowJo software (BD Biosciences).

Statistical analysis

We used an unpaired, two-tailed Student t-test, unless otherwise noted, with p<0.05 considered statistically significant. We modeled survival curves using Kaplan-Meier methods, and applied a log-rank test to make comparisons between groups. We utilized Graphpad Prism 5.0 (Graphpad Software, San Diego, CA, USA) to make our statistical analyses.

Results

EPS effect on human dendritic cells

EPS inhibits T cell proliferation in murine MLR cultures by converting BMDC into inhibitory DC²³. To evaluate if EPS also affects human cells, we investigated its effects in MLR on DC derived from cord blood CD34⁺ hematopoietic stem cells (CD34-DC; Fig 1A) or from CD14⁺ peripheral blood monocytes (moDC; Fig 1B). We used these DC as initiators in MLR cultures, along with CTV-labeled PBMC from allogeneic donors. To determine if EPS converts DC into inhibitory cells, 60 μg/ml of EPS, or the negative control, ΔEPS, was added to the cultures and T cell activation and proliferation were measured 4 days later by flow cytometry. We found that CD34-DC inhibited alloreactive T cell proliferation by more than 50%, compared to the same cultures without EPS added, or with ΔEPS added as a negative control (Fig 1C and D). Cultures with CD14⁺-derived moDC and EPS did not inhibit T cell proliferation (Fig 1E). To ask whether the observed inhibitory effects of CD34-DC were due to DC only and if EPS-treated DC retain their inhibitory properties after EPS is removed from cultures, we pre-treated CD34-DC with EPS overnight and added them to MLR cultures after washing away EPS. We found that these DC still inhibited proliferation of alloreactive T cells in MLR cultures (Fig 1F) to a similar extent as that observed when EPS was present throughout the culture period (Fig 1D). These results indicate that EPS converts human CD34-derived DC into inhibitory DC and that these cells significantly inhibit proliferation of alloreactive T cells in MLR cultures. Importantly, we also showed that human CD34-DC retain their inhibitory properties after EPS is removed, leading us to hypothesize that EPS-treated CD34-DC cells (EPS-DC) might be used as a cell-based therapy to ameliorate human GvHD.

Fig 1. EPS-DC derived from CD34⁺ cord blood cells inhibit proliferation of alloreactive CD4 T cells in MLR.

Diagrams showing dendritic cell derivation (A) from CD34⁺ cord blood cells, (B) from human monocytes. (C) Representative flow cytometry of CD4 T cell proliferation in MLR cultures with CD34-DC: not-treated (NT, left panel), ΔEPS-treated (negative control, middle panel) and EPS-treated (right panel, EPS present throughout the culture period). (D–F) Quantification of relative proliferation of alloreactive CD4 T cells in MLR cultures with CD34-DC and EPS present throughout the culture period (D), with moDCs and EPS present throughout the culture period (E), and with CD34-DC pre-treated with EPS for 16 hours and washed out prior to adding to MLR cultures (F). N > 3 independent experiments. **p < 0.01, ***p < 0.001, ****p < 0.0001, unpaired t test.

EPS-treated CD34-DC prolong survival of NSG-HLA-A2 mice experiencing GvHD

To investigate whether EPS-DC can ameliorate GvHD, we used a humanized mouse model of induced GvHD in which human PBMC (hPBMC) are transferred to lightly irradiated NSG-HLA-A2 mice. To determine if EPS-DC can reduce GvHD-related mortality, we induced GvHD with 4×10⁶ hPBMC and, in separate cohorts of mice, simultaneously injected CD34-DC pre-treated 16 hr with EPS, or PBS as a negative control, at a 4:1 ratio of hPBMC to DC (Fig 2A). All control mice that received PBS-treated CD34-DC died between 10 and 36 days after disease induction (Fig 2B). Mice that received EPS-DC had significantly improved survival. In the control cohort that received only hPBMCs (4×10⁶), the median survival was 23 days, which was not significantly different from the control group that received PBS-treated CD34-DC and hPBMC (21 days). The similar survival dynamic (Fig 2B), as well as similar median days of lethal induction, 21 and 23 days, in mice that received hPBMC with untreated DC (NT-DC) and hPBMC alone respectively, indicate that adding HLA mismatched DC, that were derived from cord blood CD34⁺ cells, do not exacerbate GvHD in humanized mice.

Fig 2. Injection of EPS-DC derived from CD34⁺ cord blood cells extends survival of humanized GvHD mice.

(A) Schematic of the GvHD induction and treatment. (B) Kaplan–Meier survival analysis of GVHD mice that received PBMCs alone (PBMC, gray line), PBMCs plus PBS-treated CD34-DCs (NT DC, dash line), or PBMCs plus EPS-treated DCs (EPS DC, black line). The log–rank test was performed between EPS-DC and NT-DC (EPS-DC and PBMC) (**p>0.005) and NT-DC and PBMC (ns). Data presented from >3 independent experiments. (C) Body weight changes after the GvHD induction in mice that received NT-DC (grey line) and EPS-DC (black line). (D) GVHD clinical scores. *p < 0.05, **p < 0.01, ***p < 0.001 as determined by unpaired t test. Data in C and D are from a single experiment with male and female mice (total N = 15) in each group. In these experiments we used CD34⁺ cells from 3 donors, all of them were HLA-A2-negative.

In a separate GvHD experiment, we transferred 2×10⁶ hPBMC and 0.5×10⁶ untreated or EPS-treated CD34-DC, and mice were analyzed for weight loss and a clinical score. We found that mice that received EPS-DC lost significantly less weight than those that received untreated DC (Fig 2C). Similarly, mice with EPS-DC showed essentially no evidence of disease in the first 30 days as measured by the clinical score, whereas mice receiving untreated DC had evidence of disease based on the clinical score (Fig 2D). After 30 days, however, the clinical score of EPS-DC mice increased, suggesting that in this model, another dose of EPS-DC could further protect the mice from disease. The fact that most mice in this experiment survived past 50 days compared to mice shown in Fig 2B, is likely due to the decreased number of PBMC administered. We conclude that EPS-DC significantly increased survival, and decreased GvHD symptoms.

Generating EPS-DC from human cord blood CD34⁺ stem cells

To meet clinical applications for using EPS-DC to prevent GvHD, we would need to generate large numbers of DC. It is well established that CD34⁺ HSC can give rise to DC in vitro, however, the number of CD34⁺ HSC in circulation is low (<0.1% of PBMC). Although this number is higher in umbilical cord blood, the volumes of clinically obtained cord blood are small. Human cord blood CD34⁺ cells are commercially available, and we used a simple protocol to expand cord blood CD34⁺ cells in serum free medium using a carefully titrated cocktail of SCF, Flt3L, TPO, IL6 and StemRegenin 1 (SR1), an aryl hydrocarbon receptor antagonist. Using CD34⁺ cells from four different donors, we found that after 3 weeks in culture with the cytokines, the number of CD34⁺ cells expanded by more than 2 ×10⁴-fold (Table 1). These data show that we can produce >2×10⁹ CD34⁺ cells from 10⁵ CD34⁺ cells.

Table 1.

Expansion of CD34+ cells

Donor #	Fold expansion in 3 weeks	Total yield/105 cells
	Total
1	15,000	1.5 × 109
2	2,000	2.0 × 108
3	27,903	2.8 × 109
4	61,200	6.1 × 109
Average	26,525	2.7 × 109

Characterization of CD34-derived dendritic cells and their EPS uptake

By flow cytometry, essentially none of the CD34⁺ cells had HLA-DR or CD11c DC markers after the 3 wk expansion, but after differentiating these cells to DC, we found that by day 5 and 12, 50% and 72%, respectively, were HLA-DR⁺ CD11c⁺ (Fig 3A). The 12-day DC were essentially all conventional DC (cDC), with ~20% having cDC1 markers and >70% having cDC2 markers (Fig 2A). Using fluorescently labeled EPS we tested which cells take-up EPS and found that essentially all DC take up EPS, but cDC2 take up considerably more than do cDC1 (Fig 3B). The identity of the contaminating ~20% HLA-DR⁻ CD11c⁻ cells which were also CD34⁻, CD14⁻ and CD3⁻, is not known, but they did not take up EPS, and likely did not contribute to protection by EPS. We also analyzed EPS-DC for upregulated expression of the inhibitory molecules (PD-L1 and PD-L2), that are upregulated on mouse DC after treatment with EPS. The expression of both PD-L1 and PL-L2 was increased in EPS-DC (Fig 2C), as expected.

Fig 3. DC composition, EPS uptake and phenotypic changes induced by EPS.

(A) Percentage of HLA-DR⁺CD11c⁺ DC in the initial culture of expanded CD34⁺ cells (left panel); after 5 and 12 days of culture with hGM-CSF and hIL-4 (two middle panels); cDC1 (CD141+) and cDC2 (CD1c) in the 12 day DC cultures. (B) EPS uptake by DC, non-DC, cDC1, cDC2 and double negative (non-cDC1/cDC2) cells. (C) Expression of inhibitory PD-L1 and PL-L2 by untreated (NT) and EPS-treated DC; data compiled from 3 independent experiments (*p < 0.05, ***p < 0.001 as determined by unpaired t test. All experiments were repeated ≥ 3 times.

Shown representative flow plots from > than 3 experiments.

Discussion

HSCT is often the only curative option to treat life-threatening hematologic diseases, including leukemias, hemoglobinopathies, and some forms of anemia, including sickle cell anemia. However, GvHD is a main cause of morbidity and mortality after HSCT, severely limiting its curative potential. The high incidence of GvHD, with greater than 10% of patients dying from this complication²⁸ despite current prophylactic treatments, underscores the unmet need for new therapies to prevent and/or treat GvHD. We discovered that EPS from a harmless probiotic soil bacterium, Bacillus subtilis, induces tolerogenic DC (EPS-DC) that prevent activated T cells from proliferating in MLR cultures. We used a humanized NSG-HLA-A2 transgenic mouse model and found that injecting EPS-DC when transferring hPBMCs significantly increased survival from a median of 22 days with NT-DC to 31 days with EPS-DC. Two of the 32 EPS-DC-treated mice survived at least 100 days. In addition, EPS-DC-treated mice showed significantly less weight loss and clinical symptoms than did NT-DC-treated mice. The decrease in clinical score was most apparent in the first month after transplantation, suggesting that more than one dose of EPS-DC may be needed to optimize treatment.

Cell therapy is an attractive approach to treat GvHD^{9,10,12–16,29,30}. Administering mesenchymal stem cells has shown variable success in treating patients with steroid-refractory GvHD^29,30. However, in a recent Phase III trial, significant improvement was observed only in pediatric and not in adult patients³⁰. Treg cells have also shown promise in treating GvHD in humans, but no large-scale clinical trial results are available. Studies in humans are performed primarily in patients with established GvHD, and it seems likely that more success will come by preventing GvHD instead of treating it³¹. DC, specifically tolerogenic DC, have the potential to prevent GvHD by inhibiting alloreactive T cell activation and proliferation. In a pre-clinical model, vitamin D-treated DC delayed GvHD in humanized NSG mice, but they did not enhance survival¹². In contrast, the EPS-DC described here appear to be tolerogenic DC that inhibit alloreactive T cells in vitro (Fig. 1C, D and F) and prolong survival in a humanized GvHD model (Fig. 2B). Unique subsets of DC have been described, including conventional DC (cDC), moDC, and plasmacytoid DC, each with different phenotypes and functions. Tolerogenic EPS-DC are derived from human cord blood CD34⁺ stem cells and are mostly cDC, especially cDC2³², whereas VitD-DC and other tolerogenic DC used in the clinic are moDC, derived from peripheral blood monocytes^33–36. Of note, VitD-DC were administered in a different ratio and by a different route than EPS-DC which also could explain differences in the survival benefits¹². EPS does not induce tolerogenic functions in moDC (Fig 1C), underscoring the novel mechanisms that EPS employs to tolerize a specific subset of DC.

We do not know how EPS-DC increased survival of humanized GvHD mice. Activated donor allogeneic T cells cause GvHD as they attack recipients’ organs. In a mouse GvHD model using biosensor mice, we showed that injecting EPS inhibited alloreactive T cell proliferation in vivo²³. In the clinic, the current treatment of choice to prevent GvHD is post-transplant administration of PTCy. Its use was shown to reduce the incidence of severe aGvHD to <10% and appears to reduce the incidence of cGvHD as well^37–40. Wacshmuth et al.,⁴¹ demonstrated that the protective effect of PTCy in an MHC-haploidentical HSCT model resulted from dampening alloreactive donor T cell proliferation, primarily through regulatory T cell enhancement. Here, we showed that human EPS-treated DC which inhibit alloreactive T cell proliferation in vitro in MLR cultures have increased expression of inhibitory molecules PD-L1 and PD-L2. These molecules may contribute to the mechanism by which EPS-DC ameliorate GvHD in humanized mice. In this case, EPS-DC could be directly interacting with donor T cells and inhibiting their activation and proliferation as occurs in MLR cultures. Moreover, EPS-DC could decrease systemic inflammation in GvHD mice leading to decreased donor T cell activation. Recent studies demonstrated that extracellular factors including tissue microenvironment, which can be altered by host microbiota or microbiota-derived metabolites such as short-chain-fatty acids, affect allogeneic immune responses^42,43. In our study, EPS-DC may function by altering the tissue microenvironment and/or by making T cells tolerant to pro-inflammatory microenvironments. Thus, EPS-DC may exert a similar effect as PTCy on alloreactive T cells, as they directly inhibit proliferation of activated T cells, including alloreactive T cells²³. In strong contrast to PTCy, an alkylating agent that alters nucleic acid and can be toxic at higher concentrations, we have no evidence that EPS-DC are toxic. We further expect that EPS-DC treatments would be short-term, and we hypothesize that they can serve as a non-toxic alternative to PTCy to prevent GvHD.

Using EPS-DC to prevent GvHD in humans will require a standardized protocol for generating DC from cord blood CD34⁺ cells and treating them with EPS, after which they can be cryopreserved until they are thawed and infused. Large numbers of EPS-DC would be needed to meet the clinical application for using them to prevent GvHD. CD34⁺ HSC can give rise to DC after culture with GM-CSF and IL-4, but only small numbers of CD34⁺ HSC can be obtained from peripheral blood or umbilical cord blood. We developed a simple protocol to expand CD34⁺ cells more than 2×10³-fold in 2 weeks. We estimate that a therapeutic dose of 10⁷–10⁸ EPS-DC would be necessary for humans undergoing HSCT and we demonstrated we could produce this number of DC from fewer than 10⁵ CD34⁺ cells. After overnight treatment with EPS, the EPS-DC can be cryopreserved and, upon thawing, transferred to pre-conditioned human patients simultaneously with donor HSC. Although in our current study only 30% of mice survived, we should note that these mice received only a single injection of EPS-DC. We hypothesize that survival could significantly increase with additional doses of EPS-DC following HSCT.

Here we described how we generated EPS-treated DC and demonstrated that they ameliorate GvHD in a humanized model of the disease. Using EPS-DC as a cell-based therapy for GvHD prophylaxis has a distinct advantage and is likely safe because the cells are “normal” and not genetically manipulated. An advantage of using cord blood CD34⁺ HSC, is that there are numerous samples currently available in cord blood banks, world-wide, due to the increase in alternative donors available for those without a matched sibling or unrelated donor. These cord blood cells can be expanded >10³-fold, and large stocks of EPS-DC can be generated and cryopreserved. This approach allows use of an off-the-shelf product, making this a potentially useful product to prevent lethal aGvHD in combination with standard and less toxic, but less effective, GVHD prophylactic regimens other than PTCy or together with PTCy, potentially allowing for PTCy to be administered at lower and less toxic doses.

There are two possible strategies where these EPS-DC could be used clinically. First, as suggested by the strategy used to treat the mice here, one design would be to develop an off the shelf EPS-DC product made from umbilical cord blood stem cells and infused timewise as PTCy is, i.e. shortly after the infusion of donor stem cells, to prevent aGVHD. Ex vivo expansion of UCB CD34⁺ stem cells using cytokines in the presence of nicotinamide has recently been approved by the US FDA for clinical transplantation based on a Phase 3 trial showing a 130 fold expansion of CD34⁺ cells (median 6.6 × 10⁸ CD34⁺ cells) with resulting superior engraftment as compared to unexpanded UCB grafts⁴⁴, thus sufficient CD34⁺ cells could be obtained from these expanded stem cell to manufacture EPS-DC for multiple patients, with wide spread availability to all transplant programs. There are, however, potential risks of using third party DC for this purpose as they express low levels of MHC, although such third-party tolerogenic DC have been tested in a murine model of GVHD and appear to be effective in prolonging survival^5,6. This benefit however may potentially be dose-dependent thus careful clinical studies would need to be performed.

The second strategy would be to develop specific one-of-a kind donor derived EPS-DCs from CD34⁺ peripheral blood G-CSF-mobilized stem cell grafts. This technique could be used for patients undergoing allografts from matched related and unrelated donors, but perhaps not for mismatched or haplo-identical transplants for the reasons mentioned above. Again, as the goal for routine CD34⁺ cell dose mobilized for transplantation is 5 × 10⁶/kg, sufficient stem cells should be available for the production of these cells as indicated in our results. These patient specific EPC-DC cells would however be required to be produced in a cGMP facility likely locally, thus limiting their availability and increasing the costs of transplant, at centers capable of making these cells.

Highlights.

• B. subtilis exopolysaccharide-treated DC (EPS-DC) upregulate inhibitory molecules

• EPS-DC inhibit proliferation of alloreactive T cells in mix lymphocyte reaction cultures

• EPS-DC ameliorate graft versus host disease in humanized mice

Abbreviations:

GvHD

graft-versus-host disease

CTV

CellTrace Violet

DC

dendritic cell

cDC

conventional DC

EPS

exopolysaccharide

EPS-DC

DC treated with EPS

moDC

monocyte-derived DC

HSCT

hematopoietic stem cell transplantation

MLR

mixed lymphocyte reaction

aGvHD

acute GvHD

cGvHD

chronic GvHD

HSC

hematopoietic stem cell

BMDC

bone marrow-derived DC

Data availability statement

Data are available from the corresponding author upon reasonable request.