Flt3L treatment of bone marrow donors increases graft plasmacytoid dendritic cell content and improves allogeneic transplant outcomes

Abstract

Higher numbers of donor plasmacytoid dendritic cells (pDC) increased survival and reduced graft-versus-host disease (GvHD) in human recipients of unrelated donor bone marrow (BM), but not G-CSF peripheral blood grafts. In murine models, we have shown that donor BM pDC increase survival and decrease GvHD compared to G-CSF-mobilized pDC. To increase the content of pDC in BM grafts we studied the effect of FMS-like tyrosine kinase 3 ligand (Flt3L) treatment of murine BM donors on transplant outcomes. Flt3L treatment (300μg/kg/day) resulted in a schedule-dependent increase in the content of pDC in the marrow. Mice treated on days −4 and −1 had > 5-fold increase in pDC content without significant changes in numbers of HSC, T cells, B cells, and NK cells in the marrow graft. In an MHC mismatched murine transplant model, recipients of Flt3L-treated T-cell depleted BM (F-TCD BM) and cytokine-untreated T cells had increased survival and decreased GvHD scores with fewer Th1 and Th17 polarized T cells post-transplant compared with recipients of equivalent numbers of untreated donor TCD BM and T cells. Gene array analyses of pDC from Flt3L-treated human and murine donors showed upregulation of adaptive immune pathways and immunoregulatory checkpoints compared with pDC from untreated BM donors. Transplantation of F-TCD BM plus T cells resulted in similar tumor burden and prolonged survival compared with grafts from untreated donors in two murine GvL models. Thus, Flt3L treatment of marrow donors is a novel method to increase the content of pDC in allografts, increase survival, and decrease GvHD without diminishing the GvL effect.

INTRODUCTION

Hematopoietic stem cell transplantation (HSCT) is curative for patients with hematological malignances and bone marrow failure disorders.¹ HSC grafts are typically obtained from aspiration of bone marrow (BM) or from apheresis of granulocyte-colony stimulating factor (G-CSF) mobilized peripheral blood.^{2, 3} The major complications of allo-HSCT including graft rejection, disease relapse, and graft-versus-host disease (GvHD).^{4, 5} These adverse effects are initiated and regulated by both the content of donor immune cells in the graft and residual antigen presenting cells in the recipient.⁶ Results of BMTCTN 0201 showed that an increased content of donor plasmacytoid dendritic cells (pDC) in BM allografts increased survival and decreased GvHD, but not in G-CSF mobilized (G-mobilized) allografts.⁷ It has been reported that dendritic cell reconstitution post-transplant predicts outcomes including incidence of GvHD, relapse and death.⁸ We have observed that transplantation of purified BM pDC increased survival and decreased GvHD without affecting graft-versus-leukemia (GvL) in allogeneic murine transplant models compared to G-mobilized pDC (Hassan, 2018, submitted). Because pDC content of the marrow is variable among allogeneic donors, methods to increase the pDC content in marrow are attractive strategies to enhance survival and GvL while limiting GvHD.

Plasmacytoid dendritic cells can be identified as Lin⁻(CD3, CD14, CD16, CD19, CD20) HLADR⁺CD123⁺CD11c⁻CD33⁻ in humans and PDCA1⁺CD11c⁺B220⁺Lin⁻(CD3, CD11b, CD19, IgM, CD49b, Ter119) in mice.^{9, 10} (add see, peter, science, June 2017). Plasmacytoid dendritic cells play a significant role in both innate and adaptive immunity because they are the primary source of type 1 interferon in both humans and mice.¹¹ Recipient pDC are depleted following irradiation, allowing for examination of the effect of donor pDC on post-transplant GvHD and GvL.¹² Donor pDC have been shown to possess graft-facilitating functions including enhancing donor cell engraftment and survival post-transplant.¹³ The immunological status, inflammatory or immunosuppressive, of donor pDC that interact with donor T cells is paramount to their ability to limit GvHD.¹⁴ We have shown that pDC facilitate immunity through early post-transplant IL-12 secretion which enhances engraftment and GvL effect and late IFNγ response pathways that decrease GvHD via induction of IDO production and increased numbers of Treg.^15–17

FMS-like tyrosine kinase 3 ligand (Flt3L) is necessary for pDC differentiation and Flt3L treatment can be used to increase pDC content in vitro.^18–21 Although the use of CDX-301, a recombinant Flt3L, in mobilization of HSC has been studied, the effect of in vivo Flt3L administration alone on the content and immunological activity of pDC in bone marrow is yet to be determined.^{22, 23} We hypothesized that Flt3L treatment of bone marrow donors would increase the pDC content of the allograft and transplanting Flt3L-stimulated marrow grafts would enhance survival while limiting GvHD. We tested this hypothesis in an MHC mismatched C57BL/6➔B10.BR transplant model in which donors were treated with PBS or two injections of 300μg/kg of Flt3L. Herein, we report that treatment of donors with Flt3L increased the pDC content of the marrow graft, increased survival, and decreased GvHD in allo-transplant recipients compared to marrow grafts from PBS-treated donors. Additionally, we observed decreased Th1 and Th17 polarization in T cells recovered from FBM transplant recipients on day 3. Using FACS purified pDC, we show that Flt3L treatment of donors led to upregulation of adaptive immune pathways and immunoregulation checkpoints in donor pDC without a reduction in GvL effect. Therefore, Flt3L treatment is a novel method that increases pDC content in donor grafts, increases survival, and decreases GvHD in allogeneic transplantation.

METHODS

Mice

C57BL/6 (H-2K^b) and B10.BR (H-2K^k) mice were purchased from Jackson Laboratory (Bar Harbor, Maine). Male donor and recipient mice were 8–10 weeks and 10–12 weeks, respectively. National Institutes of Health animal care guidelines were used and approved by Emory University Institutional Animal Care and Use Committee.

Flt3L Treatment of Mice

C57BL/6 mice were treated with a variety of schedules of daily subcutaneous injections of PBS or 300μg/kg of recombinant human Flt3L (CDX-301) generously donated by CellDex Therapeutics (Hampton, NJ).

G-CSF Treatment

C57BL/6 mice were treated with five consecutive days of subcutaneous injections of PBS or 300μg/kg of recombinant G-CSF from Sandoz (Princeton, NJ).

Donor Cell Preparation

Donor mice were euthanized and femora and tibias of donor C57BL/6 mice were flushed with 2% FBS PBS. Biotinylated anti-mouse CD3 from BD Bioscience (San Jose, CA) was used for T cell depletion. T cell purification was performed by incubating with biotinylated B220, CD49b, Gr-1, Ter119 antibodies. T cell depletion and purification samples were then incubated with anti-biotin microbeads and negative selection with a LS Miltenyi MACS column was performed (Gladbach, Germany).

Flow Cytometry

Anti-mouse (CD3, CD11b, CD19, IgM, CD49b, Ter119) PE or CD11b PE-CY7, CD11c FITC or APC-CY7, B220 PERCP-CY5.5 and PDCA1 ef450 were used for pDC analysis and purchased from BD Bioscience (San Jose, CA), Biolegend (San Diego, CA), and eBioscience (San Diego, CA). Stimulation of pDC for cytokine profile analysis was done using 50μM of CpG (ODN 1585, Invivogen, San Diego, CA) of whole bone marrow in complete RPMI 1640 supplemented with 10% FBS, 100 U/mL of penicillin, 100μg/mL of streptomycin, and 50μM each of 2-mercaptoethonaol, nonessential amino acids, HEPES, and sodium pyruvate (complete media) in 10cm wells for 9 hours at 37°C. BD golgiplug was added at hour 3. Intracellular analysis of pDC was done using BD Bioscience cytofix/cytoperm kit and anti-mouse IDO PerCP-Cy5.5, IFNα FITC, IL-10 PECY7 and IL-12 APC antibodies.

Splenocytes were stained using anti-mouse CD3 FITC, CD4 PE-CF594, CD8 PERCP-CY5.5, and CD25 APC-CY7. T cells were stimulated with BD leukocyte activation cocktail and golgiplug for 6 hours. Intranuclear staining was done using the eBioscience fixation kit and Tbet PE-CY7, GATA3 PE, RORγT APC, and FoxP3 PE antibodies. Data were acquired with a FACS Aria (BD, San Jose, CA) and analyzed using FlowJo software (Tree Star, Ashland, Oregon).

In vitro T cell Activation

MACS purified splenic T cells from C57BL/6 mice were cultured with 2μL of anti-CD3/28 Dynabeads per 10⁶ T cells (ThermoFisher, Waltham, MA). A 1:2 ratio of PBS or Flt3L-treated pDC to T cells were incubated in 96 well plates in complete media for 72 hours at 37°C.

Transplantation

On day −1 recipients were irradiated twice at 5.5 Gy, with fractions separated by 3–4 hours, for a total of 11Gy.²⁴ Mice were transplanted with 5 × 10⁶ T cell depleted PBS or Flt3L-treated BM with or without 4 × 10⁶ T cells on day 0. GvHD monitoring used a 10-point scoring system including weight, attitude/activity, skin condition, hunching, and coat condition.²⁵

Tumor Cell Challenge and Bioluminescent Imaging

A generous gift by Dr. Bruce Blazar, an acute myeloid leukemia cell line, luciferase-transfected C1498, was used for graft-versus-leukemia experiments. On day 2, mice were lethally irradiated (11Gy), on day 1 injected with 50,000 C1498 cells, and on day 0 T cell depleted bone marrow or bone marrow from Flt3L treat bone marrow donors with or without untreated T cells were then transplanted. For bioluminescent imaging, 150μg/kg of D-Luciferin was injected and mice were imaged with IVIS spectrum (Perkin Elmer, Waltham, MA). Luminescence was measured in photons/sec/cm² and normalized to control mice (recipients of the same transplant without luciferase⁺ tumor).

Gene Array Analysis

Human subjects were treated with 75μg/kg of Flt3L (CDX-301) for 5 consecutive days and underwent large volume leukapheresis as part of an IRB-approved clinical trial (Clinical trial NCT022000380). Bone marrow samples from volunteer donors were collected as part of a separate IRB-approved clinical study of immune cells in the marrow (NCT02485639). Gene expression of human BM and Ftl3L mobilized peripheral blood FACS isolated pDC was assessed using Illumina HumanHT-12 v4 beadchip microarray. Data was preprocessed, quantile normalized, background-corrected and log2 transformed for downstream analysis.²⁶ RNASeq gene expression of murine BM and Flt3L-treated BM pDC was assessed. Preparation of cDNA was performed using the Takara SMART-Seq v4 low input RNA kit. The sequencing library was created using NEBNext Ultra II FS DNA kit. Samples were sequenced on next generation sequencing at 2 X 151 bp in the paired ends. Fastq reads were trimmed and filtered for quality and adapter contamination with trimmomatic. Post-filtered reads were mapped against Ensemble mouse GRCm38/mm10 reference genome and gencode Release M16 gene annotation using STARaligner. Sequencing results are available at the NCBI Sequence Read Archive, accession SRP155387. Expression quantification was obtained using HTSeq counts, DESeq normalized and log2 transformed for further analysis.²⁷ Differential expression analysis for both human and murine samples was performed using a moderated t-test.²⁶ Heatmaps were created using NOJAH (http://bbisr.shinyapps.winship.emory.edu/NOJAH/). Genes were determined to be significantly differentially expressed based on both a fold-change of 1.5 and an FDR cutoff of 0.05. Pathway analysis was performed using Cytoscape software v3.6.1 and ReactomeFI plugin.^{28, 29}

Statistical Analysis

Data were analyzed using Prism version 5 (Graphpad, San Diego, CA) for MAC and are displayed as mean+SD unless otherwise specified. Survival differences were calculated in a pairwise fashion using the log-rank test. Applicable data were compared by student-test and 1-way ANOVA. Significance was considered as a p-value of ≤ 0.05.

RESULTS

Flt3L administration expands pDC content in bone marrow in vivo

To examine whether Flt3L treatment could enhance pDC content in vivo, we tested the effect of Flt3L administration on murine bone marrow donors due to the known effect of Flt3L in stimulating differentiation and expansion of pDC.³⁰ We measured the content of HSC, pDC, T cell, B cell, and NK in bone marrow from mice treated with different dosing schedules of PBS or Flt3L (Table 1). The increased pDC content in marrow was proportionate to the number of Flt3L doses administered (Figure 1A). HSC, T cell, B cell, and NK cell content was not significantly affected in mice treated with less than 4 doses of Flt3L (Figure 1B–E). A schedule of 2 doses of Flt3L (schedule C) was chosen for all subsequent experiments because pDC content was increased 5-fold without significant differences in the content of other immune cells that might modulate GVHD, including HSC, T cells, B cells, and NK cells.

Table 1. Flt3L administration schedule.

Bone marrow controls were treated with PBS according to the schedule in the table. Experimental mice were treated with 300μg/kg of Flt3L following the schedule in the table.

Dosage Schedule (Days)
BM: Control
A: −7
B: −4
C: −4, −1
D: −7, −5, −3, −1
E: −7, −6, −5, −4, −3, −2, −1

Figure 1. Flt3L administration to bone marrow donors increased pDC content in graft.

Mice were treated with PBS or 300μg/kg of Flt3L according to the schedule in Table 1. (A) pDC, (B) HSC, (C) T cell, (D) B cell, and (E) NK cell content were measured by flow cytometry. n=6 per group, combined data from 2 independent experiments. *P < .05, **P < .01, ***P < .001.

Flt3L administration to bone marrow donors may affect homing and lineage, but not phenotype of pDC

Because homing of donor pDC to GvHD target organs is dependent on their chemokine receptor expression³¹ we measured expression of CCR4, CCR5, CCR7, CCR9, and CXCR4 on pDC from untreated BM and FBM.³² Mice were treated with PBS or 300μg/kg of Flt3L on days −4 and −1 and bone marrow was harvested. There was significantly less CCR9 and more CXCR4 expression in pDC from FBM compared with control pDC (Figure 2A), suggesting that ability of pDC to migrate to the gut and lymph nodes could be altered by treatment with Flt3L.^{33, 34}

Figure 2. Phenotype of pDC from BM versus FBM grafts are similar.

Mice were treated with PBS or 300μg/kg of Flt3L on days −4 and −1. (A) Surface marker expression of chemokine receptors was measured by flow cytometry. (B) Whole BM or FBM grafts were treated with 50μg of CpG for 9 hours at 37°C. Intracellular staining for cytokine and (C) IDO expression was measured by flow cytometry. n=3–6 per group, from 2 independent experiments. **P < .01.

To better understand the effect of pDC on T cell activation and polarization, we next examined the cytokine profiles and IDO production of pDC from BM and FBM grafts. Flt3L treatment did not result in substantive changes in the levels of IFNα, IL-12, IL-10 and IDO in FBM pDC compared with pDC from untreated BM (Figure 2B–C).

Recipients of FBM had increased survival and less GVHD compared to untreated BM recipients

We next performed a transplant using untreated BM or FBM with the addition of donor T cells form untreated mice to assess the effect of Flt3L-treatment of donor BM on post-transplant survival and GvHD. Lethally irradiated (11 Gy) B10.BR mice were transplanted with 5 million T cell depleted (TCD) BM from C57BL/6 donor mice treated with PBS or 300μg/kg of Flt3L on days −4 and −1 in combination with 4 million T cells from PBS-treated C57BL/6 donor mice. On average, TCD BM grafts consisting of 5 million nucleated cells from PBS-treated donors contained ~50,000 pDC and TCD FBM contained ~250,000 pDC. Mice that received allogeneic TCD FBM with the addition of donor T cells had significantly increased survival compared with recipients of TCD BM plus T cells (Figure 3A). Donor hematopoietic engraftment in all groups was near 100% (Figure 3B). Recipients of TCD FBM plus T cells had significantly lower GvHD than recipients of control TCD BM plus T cells (Figure 3C). Histological analysis of the liver and small intestine showed scant evidence for GvHD pathology in recipients of control TCD BM, without the addition of donor T cells. Using NIH criteria for histological diagnosis of GvHD^35–37, and euthanizing mice on day +30 post-transplant, GvHD pathology was not significantly different comparing scoring sections of small intestine or liver between recipients of TCD BM + T cells and TCD FBM + T cells (Figure 3D–E), although the small numbers of mice studies sacrified at this time point and possible differences in sampling of tissue limit the statistical power of this comparison.

Figure 3. Grafts from Flt3L treated donors increased survival and decreased GvHD.

C57BL/6 donor mice were treated with PBS or 300μg/kg of Flt3L on days −4 and −1. B10.BR recipient mice were transplanted with 5 million T cell depleted (TCD) BM or FBM cells with or without the addition of 4 million T cells. (A) Survival of murine transplants recipients. Recipient groups included TCD BM, TCD FBM, TCD BM + 4 million T cells (BMT), and TCD FBM + 4 million T cells (FBMT). *P < .05 represents significance between BMT and FBMT recipients using Kaplan-Meier survival analysis. (B) Chimerism of recipients 30 days post-transplant by groups as indicated. (C) Clinical GvHD scores of mice that received T cells. n=30 per group, from 2 independent experiments. *P < .05 represents significance between BMT and FBMT recipients using 2-way ANOVA. (D) Representative histopathological samples of small intestine and liver at day 15 from each treatment group photographed at 600X. Black squares denote apoptotic cells in the small intestine crypts. Histopathological scores of GvHD associated pathology. Score of 0: no pathology, 1: villi shortening, 2: apoptotic cells, 3: loss of crypts. Liver pathology included lymphocytic infiltration to the portal triad and rare apoptotic bile duct epithelial cells (black arrows). (E) Percentage of mice within each group that showed GvHD-associated pathology in the small intestine (F) or liver. n=4–6 per group, from 2 independent experiments.

To isolate the role of pDC from FBM in this transplant model, we transplanted B10.BR recipients with purified 5,000 HSC, 10⁶ T cells, and 50,000 FACS-isolated pDC from FBM or untreated C57BL/6 BM. Recipients of FBM pDC had 90% survival and lower GvHD scores compared to recipients of BM pDC, in which survival was 80% (p=NS; Supplemental Figure 1A–B).

Flt3L treated donors have different gene expression profiles than bone marrow donors

After observing a trend towards increased survival in the group that received an equal number of FACS isolated FBM pDC compared with untreated BM pDC in a model system using purified HSC, T cell, and pDC transplants we hypothesized that quantitative effects of Flt3L treatment on the increased number of pDC in donor marrow is not the sole determinant of improved transplant outcome following Flt3L treatment. Thus, to explore qualitative effects of Flt3L treatment on pDC we analyzed gene expression profiles of pDC from FBM versus BM from untreated donor mice. Plasmacytoid dendritic cells were isolated by FACS from the bone marrow of mice treated with PBS or 300μg/kg of Flt3L on days −4 and −1. RNA was obtained from these purified pDC and sequenced using next generation sequencing. Pairwise analyses were conducted on individual genes and immunological pathways using a p-value cutoff of < .05 (Supplemental Tables 1, 2). FBM pDC have distinct gene expression profiles from pDC from untreated BM grafts (Figure 4A–B). Immunological pathways that showed significant differences (p-value < .05) include the adaptive immune pathways and immune checkpoint pathways. Upregulation of genes in the adaptive immune pathway include APRIL, Hmox1, and Etv5, all of which may affect the ability of donor pDC to regulate T cell polarization and activation.³⁸ Immune checkpoint pathway genes, Bcl2, Cyclin D3, TIM-3, and ACK1, are upregulated in pDC from FBM grafts, suggesting an increased ability for immune cell regulation.³⁹ Finally, pDC from FBM grafts upregulated Tox1 and Prss16, genes that regulate T cell selection in the thymus (Table 2).^{40, 41}

Figure 4. FBM pDC have distinct gene expression profile from BM pDC.

Mice were treated with PBS and 300μg/kg of Flt3L on days −4 and −1. Plasmacytoid dendritic cells were isolated by FACS. RNA was sequenced by next generation sequencing. (A) Heatmap depicting the significantly differentially expressed genes in the murine samples using z-score scaling, 1-pearson correlation distance, and ward.D clustering. (B) Volcano plot of gene upregulation and downregulation (BM vs FBM). n=3 per group. (C) Healthy human donors were untreated or treated with 75μg/kg of CDX-301 (recombinant Flt3L) for 5 consecutive days. Untreated donors underwent bone marrow harvest and Flt3L-treated donors underwent leukapheresis on day 6. Plasmacytoid dendritic cells were isolated by FACS. RNA sequencing was done with Illumina HumanHT-12 v4 beadchip. Heatmap depicting the significantly differentially expressed genes in the human samples using z-score scaling, Euclidean distance, and complete clustering. (D) Volcano plot depicting upregulation and downregulation of genes (BM vs F-Apheresis). n=4–5 per group.

Table 2. Genes upregulated in FBM pDC.

Upregulated genes FBM vs BM in adaptive immune, immune checkpoint and thymic induced T cell selection pathways that are present in Figure 4B plotted Fold Change (logFC) by –log10(p-value).

Gene	Pathway	logFC	NegLog10pvalue
Bcl2	Immune Checkpoint	−1.66	2.27
Cyclin D3	Immune Checkpoint	−1.60	2.08
Etv5	Adaptive Immunity	−2.30	1.73
APRIL	Adaptive Immunity	−3.25	1.71
TIM-3	Immune Checkpoint	−1.79	1.56
Tox1	T cell selection	−4.24	1.53
Prss16	T cell selection	−3.12	1.48
Hmox1	Adaptive Immunity	−1.73	1.46
ACK1	Immune Checkpoint	−1.97	1.43

In order to better interpret the translational relevance of these murine studies, gene expression profiles were compared between FACS-isolated human pDC from untreated bone marrow donors and pDC isolated from apheresis products of Flt3L-treated sibling donors. Because we did not have access to the bone marrow from healthy human volunteers treated with Flt3L, we used grafts from sibling stem cell donors that underwent mobilization with 5 daily injections of 75 μg/kg of CDX-301 (recombinant Flt3L) as a single mobilization agent from an IRB-approved clinical study (NCT022000380) that assessed the efficacy of Flt3L in mobilizing stem cells, and compared gene expression to pDC isolated from untreated BM acquired on a separate clinical study (NCT02485639). RNA from pDC samples were sequenced by Illumina Chip (Supplemental Tables 3, 4). Similar to the murine gene array analysis, KLRF1 and SLAMF6 in the adaptive immune pathway and BCL2 and BIRC3 in the immune checkpoint pathway were upregulated in pDC from the Flt3L-treated donors compared with pDC from untreated bone marrow volunteers (Figure 4C–D) (Table 3). Additionally, toll like receptor (TLR) genes, including those in the TLR4 pathway, APP, MAP2K6, CD36, ACTG1, and IRF7, were downregulated in Flt3L-treated donor pDC, indicating decreased ability to stimulate innate immune pathways (Table 4).⁴²

Table 3. Genes upregulated in F-Apheresis pDC.

Upregulated genes F-apheresis vs. BM in adaptive immune and immune checkpoint pathways that are present in Figure 4D plotted Fold Change (logFC) by –log10(p-value).

Gene	Pathway	logFC	NegLog10pvalue
BCL2	Immune Checkpoint	−2.34	1.35
KLRF1	Adaptive Immunity	−3.70	2.53
SLAMF6	Adaptive Immunity	−2.74	2.16
BIRC3	Immune Checkpoint	−2.37	1.58

Table 4. Genes downregulated in F-Apheresis pDC.

Downregulated genes in toll-like receptor cascades including the TLR4 pathway that are present in Figure 4D plotted Fold Change (logFC) by –log10(p-value).

Downregulated in human F-Apheresis pDC	logFC	NegLog10pvalue
APP	3.67	2.02
MAPK6	1.85	1.47
CD36	3.14	2.93
ACTG1	1.91	1.32
IRF7	3.24	1.39

Flt3L bone marrow reduced T cell polarization post transplantation

Based upon the results of the gene array analyses of pDC from Flt3L treated mice and humans, we next studied the ability of pDC to interact with and influence donor T cell activation and immune polarization post-transplant in a non-paracrine cytokine fashion.^{43, 44} We examined the potential of pDC to stimulate or inhibit allo-activation of T cells by measuring surface expression of MHC II, CD86, and PDL1 on pDC. Plasmacytoid dendritic cells from Flt3L-treated donors expressed less MHC II and CD86 and expressed more PDL1 (Figure 5A). To determine whether changes in surface expression of co-stimulatory and co-inhibitory molecules affects T cell polarization, we performed a C57BL/6→B10.BR transplant with 5 million T cell depleted BM or FBM cells and 4 million untreated T cells. On day 3 post-transplant, donor T cells were analyzed for their content of T cell transcription factors by flow cytometry. Mice that received TCD FBM + 4 million T cells had donor T cells with significantly lower levels of Tbet and RoRγT expression on day 3 post-transplant, consistent with decreased Th1 and Th17 polarization (Figure 5B). To better determine what role pDC placed in the pattern of transcription factor expression in T cells post-transplant, we performed a C57BL/6→B10.BR using purified HSC, T cells, and purified populations of BM or FBM pDC, and again measured T cell transcription factor expression on day 3. The same trend of lower Tbet and RoRγT expression in T cells post-transplant was observed in this experiment as well (Supplemental Figure 2A). The effect of FBM pDC on T cells does not appear to limit T cell proliferation, as T cell proliferation was equivalent comparing the addition of FBM pDC versus untreated BM pDC to T cells stimulated with anti-CD3 and anti-CD28 beads (Supplemental Figure 3A–B).

Figure 5. Flt3L treatment of bone marrow donors decreased expression of T helper cell transcription factors in recipients.

Donor mice were treated with PBS or 300μg/kg of Flt3L on days −4 and −1. (A) CD86, MHC II, and PDL1 surface expression was measured by flow cytometry. (B) A C57BL/6→B10.BR transplant was performed where recipient mice were transplanted with 5 million T cell depleted BM or FBM cells with the addition of 4 million T cells. Intranuclear staining of transcription factors Tbet, GATA3, RoRγT, and FoxP3 of T cells was assessed by flow cytometry 3 days post-transplant. n=6 per group, combined data from 2 independent experiments. *P < .05.

GvL activity was not diminished in recipients of FBM

To determine whether Flt3L treatment of marrow donors affected the GvL activity of the allogeneic transplant, we compared the growth of leukemia cells in recipients of FBM compared with untreated BM using bio-luminescent imaging of luciferase+ leukemia cells and survival analyses of leukemia-bearing transplant recipients. First, C57BL/6 mice were irradiated (11Gy) on day −2, inoculated with 50,000 syngeneic luciferase⁺ C1498 tumor cells on day −1, and then transplanted with 5 million TCD BM or FBM cells plus 4 million T cells from MHC-mismatched B10.BR donors on day 0. Tumor burden was measured serially by bioluminescence. There was no significant difference in tumor burden comparing recipients of untreated TCD BM plus 4 million T cells versus TCD FBM plus 4 million T cells recipients (Figure 6A–C). Additionally, using the LBRM tumor line in C57BL/6→B10.BR transplant recipients, there was a significant prolongation of survival comparing recipients of TCD FBM plus 4 million T cells to recipients of TCD BM plus 4 million T cells (Supplemental Figure 4A–C).

Figure 6. No loss of GvL effect in recipients of FBM.

B10.BR→ C57BL/6 murine transplant recipients received 5 million T cell depleted bone marrow cells and 4 million T cells. Recipient mice received 50,000 luciferase⁺ C1498 cells. (A) Serial bioluminescent imaging of recipient mice. (B) Survival curve of murine transplant recipients. (C) Quantification of tumor burden. n=10–15 per group, from 2 independent experiments. *P < .05.

DISCUSSION

GvHD remains the most significant complication following HSCT.⁴⁵ Although recipients of bone marrow or G-CSF-mobilized grafts from unrelated donors have equal survival up to 7 years post-transplant, there is a higher incidence of chronic GvHD in recipients of G-CSF-mobilized grafts.^{46, 47} Furthermore, the BMTCTN 0201 study showed that recipients of marrow grafts containing higher numbers of pDC had increased survival and lower treatment related mortality due to less GvHD.⁷ Results of this study also showed that pDC content of the marrow graft varied greatly amongst volunteer bone marrow donors, raising the question of how to increase the content of immune-regulatory donor pDC in all bone marrow allografts. We report herein that administration of Flt3L to bone marrow donors increased pDC content of the graft and graft recipients of Flt3L-treated bone marrow or purified donor pDC from FBM have increased survival and less GvHD after allo-transplantation.

In our previous studies, we have shown that murine BM donor pDC limit GvHD without attenuating GvL.^{16, 17} Furthermore, we have unpublished data confirming that with transplantation of allo-grafts containing highly purified HSC, T cells, and pDC, recipients of BM pDC have increased survival and decreased GvHD incidence compared to recipients of G-CSF-mobilized pDC (Hassan 2018, submitted). Thus, data from the current study showed that Flt3L administration to bone marrow donors increased pDC content 5-fold spurred further characterization of the effects of Flt3L treatment on the quantity and quality of pDC in bone marrow and HSCT outcomes. Transplanting 5 million T cell depleted bone marrow cells from Flt3L-treated donors with the addition of 4 million T cells from untreated donors increased survival with decreased GvHD compared with transplantation of an equal number of untreated TCD BM cells and T cells.

Interestingly, although pDC from untreated BM and FBM have a similar phenotype, we show that FBM pDC have an enhanced cell-intrinsic ability to limit GvHD and a gene expression profile that supports greater immune-modulatory capacity compared with pDC from untreated BM in a C57BL/6B10.BR heterogenous transplant model with characteristics of both acute and chronic GvHD, a common aspect of HSCT recipients in a clinical setting (JCI 21). Of note, T cells from recipients of FBM pDC had less Th1 (Tbet) and Th17 (RoRγT) polarization than T cells from recipients of pDC from untreated BM (Figure 5B and Supplemental Figure 2A). Thus the upregulation of adaptive immune and immune checkpoint pathways in pDC from FBM may be responsible for their ability to regulate donor T cell immune polarization, decrease Th1 and Th17 polarization, and limit the incidence and severity of GvHD compared with pDC from untreated BM.^47–50 This coupled with increased expression of genes involved in positive and negative selection in the thymus may enable FBM pDC to induce post-transplant tolerance of donor stem cell-derived T cells. Additionally, pDC from Flt3L treated human donors have downregulated the expression of genes involved in TLR cascades and innate immune cell pathways compared with pDC from untreated BM, with the most significant decrease seen in expression of TLR4 pathway genes. Downregulation of the TLR4 pathway may also play a role in limiting GvHD in the gut, because gut GvHD can activate the release of LPS from bacteria, ultimately activating TLR4 on pDC, which can further aggravate injury in the gut and augment GVHD.^51–53 Thus, the downregulation of genes involved in TLR cascades and other innate immune pathways in FBM pDC may also play a role in the attenuation of GvHD following transplantation of pDC from FBM.

The present study has some limitations. Although the current studies focus on donor pDC and the effect of Flt3L administration to bone marrow donors on pDC, other donor cell types may contribute to the transplant outcomes following Flt3L treatment of BM donors. We have shown with one dose and one schedule of Flt3L administration (day −4 and day −1 with respect to marrow harvest), there is a significant increase in content and quality of the pDC. Although there was no significant change in content of HSC, NK cells, T cells, and B cells in Flt3L-treated donor bone marrow grafts, determining the effect of Flt3L administration on the cell-intrinsic qualities of these cells to interact with donor T cells and regulate immune responses may further clarify the mechanisms that result in increased survival and decreased GvHD in FBM recipients.⁴⁷ Thus, the clinical utility of Flt3L treatment of human bone marrow donors could be studied by characterizing the effect on immune cell content and quality within the bone marrow graft. Among volunteers treated with Flt3L, some previous phase 1 studies have shown safety of daily administration for a week or more.²² Additionally, we did not test the ability of peripheral blood mobilized Flt3L grafts to affect transplant outcomes, because we observed that HSC and pDC mobilization was not equivalent to G-CSF mobilization (Supplemental Figure 5A–B). Furthermore, Flt3L as a single agent for stem cell mobilization is not efficient (Clinical trial NCT022000380) and Flt3L would need to be combined with other agents such as CXCR4 antagonists for this approach to be feasible for clinical practice. Additionally, we compared gene expression in different sources of Flt3L-stimulated pDC in mice and humans, isolating FBM pDC from Ftl3L-treated mouse BM and F-apheresis pDC from human donor apheresis products due to the lack of available human FBM samples. Nevertheless, the similarities in gene expression between pDC from murine samples and human samples, and the striking findings of improved survival with less GVHD in the murine transplant models suggest that characterization of pDC from bone marrow of Flt3L-treated human donors is warranted in a planned clinical study to further validate the clinical translation potential of these findings.

To summarize, we report a novel method using Flt3L treatment of bone marrow donors to reduce the GvHD-promoting activity of marrow graft and marrow pDC and improve survival of allogeneic BM transplant recipients. Flt3L treatment increased the content and immune-regulatory capacity of pDC in bone marrow grafts, leading to decreased severity of GvHD in murine recipients. Notably, the reduction of GvHD activity following Flt3L treatment of BM donors was not associated with an attenuation of the GvL activity of donor T cells. Thus, the present pre-clinical data provide an impetus to test the clinical effect of Flt3L treatment of bone marrow donors as a novel method to improve transplant outcomes.

KEY POINTS.

• Flt3L treatment of murine bone marrow donors significantly increases pDC content

• Flt3L-treated bone marrow donor grafts increased survival and decreased GvHD

• Flt3L-treated bone marrow donor grafts do not diminish the graft-versus-tumor effect

• Flt3L-treated bone marrow donor pDC upregulate genes in adaptive immune pathways