Antagonism of CXCR4 disrupts the interaction between the CXCR4 receptor on hematopoietic stem cells (HSCs) and the CXCL12 expressed by stromal cells in the bone marrow, which subsequently results in the shedding of HSCs to the periphery. Because of their profound immunomodulatory effects, HSCs have emerged as a promising therapeutic strategy for autoimmune disorders. We sought to investigate the immunomodulatory role of mobilized autologous HSCs, via target of the CXCR4-CXL12 axis, to promote engraftment of islet cell transplantation. Islets from BALB/c mice were transplanted beneath the kidney capsule of hyperglycemic C57BL/6 mice, and treatment of recipients with CXCR4 antagonist resulted in mobilization of HSCs and in prolongation of islet graft survival. Addition of rapamycin to anti-CXCR4 therapy further promoted HSC mobilization and islet allograft survival, inducing a robust and transferable host hyporesponsiveness, while administration of an ACK2 (anti-CD117) mAb halted CXCR4 antagonist-mediated HSC release and restored allograft rejection. Mobilized HSCs were shown to express high levels of the negative costimulatory molecule programmed death ligand 1 (PD-L1), and HSCs extracted from wild-type mice, but not from PD-L1 knockout mice, suppressed the in vitro alloimmune response. Moreover, HSC mobilization in PD-L1 knockout mice failed to prolong islet allograft survival. Targeting the CXCR4–CXCL12 axis thus mobilizes autologous HSCs and promotes long-term survival of islet allografts via a PD-L1–mediated mechanism.
Recent reports of clinical benefits following use of autologous hematopoietic stem cells (HSCs) in type 1 diabetes have sparked much interest in the utility and value of HSC therapy as a tool to induce tolerance (1). HSCs have also been used in other debilitating autoimmune diseases with the aim of resetting the patient’s immune system; indeed, HSC-based therapy has yielded success in the treatment of various autoimmune disorders, including multiple sclerosis (2), systemic sclerosis (3, 4), and Crohn’s disease (5). The primary objective in using HSCs has been to eliminate autoreactive immune cells, thus ensuring a properly functioning immune system. However, growing evidence suggests that autologous HSCs can induce central and peripheral immunological tolerance per se (6).
Preclinical studies have demonstrated that T cell-depleted bone marrow-resident CD34+ stem cells overcome MHC barriers in sublethally irradiated mice (7) and that murine HSCs may delete effector cells through Fas/Fas ligand interaction or via the TNF-α pathways, which are both present on HSCs (8, 9). Kared et al. (10) have recently demonstrated that murine HSCs may stimulate peripheral Foxp3+ regulatory T cell (Treg) expansion through both cell–cell contact activation of Notch signaling and through soluble factors, such as GM-CSF, which is produced at high levels by hematopoietic progenitors (10). With respect to human HSCs, Rachamim et al. (11) have shown that cells within the human CD34+ population are endowed with potent veto activity, referring to the ability of HSCs to neutralize precursors of cytotoxic T lymphocytes in an HLA-restricted and cell contact-dependent fashion (12, 13). HSCs have also been used to improve the outcome of solid organ transplantation through the induction of mixed hematopoietic chimerism (14). This strategy constitutes a unique approach to generate tolerance in solid organ transplantation without the need for long-term immunosuppressive therapy but also requires intense toxic conditioning strategies. To reduce the burden of these regimens, an attempt has been made to use megadoses of autologous stem cell transplants.
Recent efforts targeting the CXCR4–CXCL12 axis have been aimed at inducing shedding of HSCs to the periphery (15–18). HSCs express high levels of CXCR4, which keeps them anchored to the bone marrow where CXCL12 (or SDF-1α, the ligand for CXCR4) expression is high, particularly in stromal cells (19). We thus aimed to target the CXCR4–CXCL12 axis by blocking the CXCR4 receptor using a novel CXCR4 antagonist (20) to mobilize autologous HSCs in a murine model of islet transplantation. Our goal was to achieve HSC mobilization in our islet transplant recipients to improve allograft survival. This approach could have significant clinical applications, given that CXCR4 antagonists (i.e., Mozobil/plerixafor) are currently under investigation in a phase III trial to improve engraftment in bone marrow-transplanted patients.
Materials and Methods
C57BL/6, BALB/c, and DBA/1J mice were obtained from The Jackson Laboratory (Bar Harbor, ME) and cared for and used in accordance with institutional guidelines. Programmed death ligand 1 (PD-L1)−/− mice on a C57BL/6 background were provided by Dr. Arlene Sharp as previously published (21). Protocols were approved by the Harvard Medical School Institutional Animal Care and Use Committee.
Interventional animal studies
Mice were treated with a CXCR4 antagonist (NIBR1816, provided by Novartis, Basel, Switzerland) or vehicle at 30 mg/kg (1 mg/mouse) i.p. once per day for up to 14 d. Mobilization of murine HSCs, as demonstrated by expression of c-Kit+Lin−Sca-1+ (KLS), was then evaluated in the bone marrow and spleen by flow cytometry during CXCR4–CXCL12 targeting. KLS cells were detected by gating on lineage-negative cells (by staining for Gr-1, CD8, CD4, CD11b, and B220; Miltenyi Biotec, Auburn, CA) and then evaluating c-Kit/Sca-1 double-positive cells (both from BD Biosciences, San Jose, CA). For functional studies, mobilized HSCs were sorted using the above markers. Anti-CD117 (anti–c-Kit, ACK2 clone; Millipore, Billerica, MA) was used at a dose of 125 μg at days 0 and 5 to ablate HSCs. In some experiments, mice were pretreated with ACK2 (at days −5 and day 0). We also examined whether Treg depletion with anti-CD25 Ig at the time of transplantation (500 μg at days −6 and −1) was able to abrogate the effect of CXCR4 antagonist treatment.
Pancreatic islets were isolated by collagenase digestion, density gradient separation, and handpicking as described previously (22). Islets were transplanted under the renal capsule of mice rendered diabetic with streptozotocin (225 mg/kg, administered i.p.). Islets isolated from MHC-mismatched male BALB/c (H-2d) donors were transplanted under the kidney capsule of recipient C57BL/6 mice (H-2b) (850 islets/recipient). Rejection of islet allografts was defined as blood glucose levels of >250 mg/dl for at least 2 consecutive days.
CXCR4 ligand binding assay
Membranes prepared from the mouse pre-B cell line L1.2 were used as the source of mouse CXCR4. Membranes were prepared from cell homogenates in a buffer containing 20 mM HEPES, 1 mM EDTA (pH 7.4), and protease inhibitors followed by centrifugation at 28,000 × g. After a second round of homogenization and centrifugation, membranes were resuspended in 20 mM HEPES and 0.1 mM EDTA (pH 7.4) and stored in aliquots at −80°C. Murine CXCL12 was obtained from R&D Systems (Minneapolis, MN) and was 125I-labeled (GE Healthcare, Piscataway, NJ). For the assay the following ingredients were added to a 96-well OptiPlate: 10 μg membrane protein in 50 μl assay buffer (25 mM HEPES, 2 mM CaCl2, 5 mM MgCl2, 100 mM NaCl, 0.5% BSA, and protease inhibitors) followed by 0.5 mg wheat germ agglutinin-coated SPA beads (Amersham Biosciences) followed by 50 μl antagonist (or buffer) or 400 nM unlabeled murine CXCL12. Finally, 50 μl 125I-labeled murine CXCL12 in assay buffer was added (final concentration, 20–25 pM). The plates were incubated for 2 h at room temperature with continuous shaking. Following incubation, the plates were centrifuged for 10 min and measured in a TopCount NXT instrument (Packard Instruments, Boston, MA). Under these experimental conditions murine unlabeled CXCL12 displaced 125I-labeled human CXCL12 equally well (IC50 values, 110 ± 35 and 167 ± 52 pM, respectively).
Ca2+ mobilization assay
CXCL12-induced Ca2+ mobilization from intracellular stores was measured in L1.2 pre-B cells loaded with the Ca2+-sensing fluorochrome Fluo-4 (Molecular Probes/Invitrogen, Carlsbad, CA). Cells were incubated with fluorochrome at a concentration of 4 μM for 1 h at 37°C and then washed in assay buffer. Loaded cells were dispensed into microtiter wells, mixed with antagonist, incubated for 2 h, and subsequently dispensed into 384-well plates in quadruplicate and placed in a fluorescence image plate reader (FLIPR384; Molecular Devices, Sunnyvale, CA). Fluorescence of the cells was recorded for 20 s, after which CXCL12 was added (final concentration, 3 nM), and fluorescence was further recorded for 215 s. The minimal (baseline) and maximal fluorescence were used to calculate the inhibitory effect of CXCR4 antagonists. IC50 values were expressed as the inhibitor concentration that yielded 50% inhibition of the (Fmax − Fmin)/Fmin value measured in the absence of inhibitor.
Cell migration stimulated by CXCL12 was assessed in Transwell tissue culture inserts with porous polycarbonate membranes (5 μm pore size; Costar, Cambridge, MA). Target cells (500,000 L1.2 pre-B cells for mouse CXCR4 in 100 μl HBSS, 0.1% BSA) were placed in the upper chamber, and 600 μl buffer or chemoattractant (rCXCL12; final concentration, 10 nM) was placed in the lower chamber. After incubation for 4 h at 37°C, the migrated cells were counted in a FACSCalibur (BD Biosciences) by acquiring all events for 30 s. When inhibitor was tested, the compound was added to both the upper and the lower compartments to give the desired final concentration.
Actin polymerization assay in whole blood
Blood was collected in EDTA-coated tubes. CXCR4 inhibitors were diluted in PBS (Invitrogen) to the appropriate final concentration, and typically 3 μl CXCR4 antagonist solution was mixed with 100 μl whole blood in a 96-well plate (Costar). After incubation for 10 min at room temperature, 3 μl CXCL12 solution (to achieve a final concentration of 100 nM) was added to the sample and incubated for 25 s at room temperature. Ice-cold FACS lysing buffer (BD Biosciences) was added to the samples, followed by incubation for 10 min on ice. After several washes with PBS/0.5% BSA on ice, the blood cells were resuspended in 200 μl PBS containing 1% paraformaldehyde (Sigma-Aldrich, St. Louis, MO) and incubated for 5 min on ice. After centrifugation the cells were resuspended in 100 μl PBS/0.5% BSA, 1% Alexa 488 phalloidin (Invitrogen), and 1% l-α-palmitoyl-lysophosphatidyl-choline (Sigma-Aldrich) and incubated for 20 min on ice, protected from light. Cells were washed three times with PBS/0.5% BSA, and fluorescence was analyzed with a FACSCalibur by gating on granulocytes on forward light scatter versus side light scatter dot plots.
An ELISPOT assay was used to measure cytokine production by splenocytes extracted from treated or control mice (23). Briefly, Millipore immunospot plates (Millipore, Bedford, MA) were coated with capture Abs (BD Biosciences). Plates were blocked with 1% BSA to prevent nonspecific binding. Naive donor splenocytes were used as stimulators, while 5 × 105 C57BL/6 splenocytes isolated from CXCR4 antagonist-treated or untreated control islet-transplanted C57BL/6 mice were used as responders. After 4 d of culture in RPMI 1640 with 10% FCS penicillin/streptomycin at 37°C and 5% CO2, plates were washed, and biotinylated Abs specific for each cytokine were added to wells and incubated at 4°C for 12 h. Plates were then washed and incubated at room temperature with streptavidin-HRP for 2 h and developed with aminoethyl carbazole (Sigma-Aldrich) diluted in N,N-dimethylformamide. Spots were counted on an immunospot analyzer (Cellular Technology, Cleveland, OH). The supernatant of each culture was used for Luminex analysis.
MLR and anti-CD3/-CD28 stimulation
To measure the alloimmune response in the presence or absence of CXCR4 antagonist, cultured BALB/c dendritic cells (DCs) were used to stimulate C57BL/6 CD4+ T cells isolated from splenocytes by magnetic bead separation (Miltenyi Biotec) at a ratio of 1:1 DCs/splenocytes. Proliferation was measured at day 3 of incubation at 37°C and 5% CO2 following pulsing with [3H]TdR (PerkinElmer, Wellesley, MA) using a liquid scintillation counter. Lymphocytes were also stimulated with Con A as a positive control. An anti-CD3/-CD28 stimulation assay was performed as well; briefly, 5 × 105 splenocytes were plated with 1 μg/ml anti-CD3 and anti-CD28. CXCR4 antagonist was added to both assays at doses of 1, 10, and 100 μg/ml.
Rat anti-mouse CD3-PE, CD4-FITC, CD8-PE, CD25-PE, and CD62L-allophycocyanin were purchased from BD Biosciences. Foxp3-allophycocyanin was purchased from eBioscience (San Diego, CA). Cells recovered from spleens and peripheral lymphoid tissues were subjected to FACS analysis and run on a FACSCalibur. Foxp3 analysis was performed following overnight permeabilization of cells extracted from spleens and peripheral lymphoid tissue with commercially available Abs and gating on CD4+CD25+ cells. Data were analyzed with FloJo software version 6.3.2 (Tree Star, Ashland, OR).
Islet pathology and immunohistochemistry
Immunohistochemistry was performed with 5-μm-thick formalin-fixed, paraffin-embedded tissue sections. Briefly, slides were soaked in xylene, passed through graded alcohols, and immersed in distilled water. Slides were then pretreated with 10 mM citrate (pH 6.0) (Zymed Laboratories, South San Francisco, CA) or with 1 mM EDTA (pH 8.0) in a steam pressure cooker, followed by washing in distilled water for Ag retrieval. All further steps were performed at room temperature in a hydrated chamber. Slides were pretreated with peroxidase block (Dako, Carpinteria, CA) for 5 min to quench endogenous peroxidase activity. The following primary Abs were used (including company, clone/reference, dilution, and retrieval method, respectively): anti-CD3 (Cell Marque, Rocklin, CA; 1/1500, EDTA), anti-CD45/B220 (BD Biosciences, 1/200, citrate), anti-Foxp3 (eBioscience, 1/25, citrate), and anti-insulin (Dako, N1542, undiluted, EDTA). All primary Abs were applied to slides in Dako diluent for 1 h. Slides were then washed in 50 mM Tris-Cl (pH 7.4), and the appropriate HRP-conjugated secondary Ab (EnVision detection kits; Dako) was applied for 30 min. After further washing, immunoperoxidase staining was developed with a diaminobenzidine chromogen kit (Dako) per the manufacturer and counterstained with hematoxylin. Photomicrographs were taken on an Olympus BX41 microscope (Olympus, Center Valley, PA) at indicated magnifications with an Olympus Q-color5 digital camera and analyzed with Adobe Photoshop elements 2.0 (Adobe, San Jose, CA). All images were taken at ×400 original magnification. To detect GFP+HSCs after adoptive transfer, rabbit anti-GFP (Abcam, Cambridge, MA) was used at 1/3000 dilution with EDTA retrieval.
Adoptive and coadoptive cell transfer
To assess the robustness of tolerance toward alloantigen and the presence of infectious tolerance or hyporesponsiveness of the immune system, two different adoptive transfer experiments were performed. In the first experiment, the ability of splenocytes obtained from long-term tolerant-treated mice to reject islets in immunodeficient mice was evaluated. We adoptively transferred 10 × 106 splenocytes extracted from rapamycin plus CXCR4 antagonist-treated long-term tolerant C57BL/6 islet-transplanted mice into C57BL/6-RAG immunodeficient mice previously transplanted with BALB/c islets. Splenocytes extracted from C57BL/6 mice transplanted with BALB/c islets that had been rejected were used as controls for the adoptive transfer experiments, as previously reported (24). In a second experiment, we evaluated the ability of splenocytes obtained from long-term rapamycin plus CXCR4 antagonist-treated mice to control rejection mediated by splenocytes obtained from naive mice. We coadoptively transferred 10 × 106 splenocytes extracted from rapamycin plus CXCR4 antagonist-treated long-term tolerant C57BL/6 islet-transplanted mice and 10 × 106 splenocytes extracted from naive mice into C57BL/6-RAG immunodeficient mice previously transplanted with BALB/c islets. In another set of experiments, tolerant mice or controls (C57BL/6 [H-2b]) were transplanted with islet donor skin (BALB/c [H-2d]) or from a third-party donor (DBA/1J [H-2q]).
Data are expressed as means ± SEM. Kaplan–Meier analysis was used for survival analysis. To compare the two groups, we used a two-sided unpaired Student t test (for parametric data) or Mann–Whitney U test (for nonparametric data) according to distribution. A p value <0.05 (by two-tailed testing) was considered an indicator of statistical significance. Data were analyzed with an SPSS statistical package for Windows (SPSS, Chicago, Illinois).
NIBR1816 is a potent antagonist of murine CXCR4
We first evaluated the ability of the recently described small molecule CXCR4 antagonist NIBR1816 to block murine CXCR4 in vitro in various assays, as previously demonstrated for rat and human CXCR4 (20). The binding of 125I-labeled CXCL12 to CXCR4 (from murine L1.2 pre-B cell membranes) was inhibited by the presence of CXCR4 antagonist (IC50, 106.6 ± 9.1 nM). Ca2+ mobilization, from murine L1.2 pre-B cells loaded with the Ca2+-sensing fluorochrome Fluo-4, following the addition of CXCL12 was inhibited by CXCR4 antagonist (IC50, 15.7 ± 3.1 nM). CXCL12-stimulated migration of leukocytes in a chemotaxis assay was also inhibited by CXCR4 antagonist (IC50, 31.1 ± 5.5 nM). The antagonist was almost equally potent in inhibiting the CXCL12-induced actin polymerization in blood granulocytes (IC50, 11 ± 5 nM). We then evaluated the concentration and half-life in vivo in mice and found that the peak concentration of CXCR4 antagonist, after the i.p. injection of 1 mg, is reached at 6 h and that the half-life is ∼12 h.
CXCR4 antagonist mobilizes HSCs
HSC mobilization was evaluated in naive C57BL/6 mice after daily injections of 30 mg/kg/day for up to 14 d. Bone marrow and spleen were obtained from treated C57BL/6 mice at different time points (baseline, 6, 12, 18, 24 h, and 7, 14, 28 d postinjection), and HSC percentage was evaluated by FACS analysis (n = at least 3 mice/time point) by gating on KLS cells.
Six hours after a single injection of 30 mg/kg (∼1 mg), an increase in HSCs was evident in the bone marrow (Fig. 1A, 1C). HSCs increased from 2.28 ± 0.70% at baseline to 8.88 ± 2.65% 6 h after injection (p = 0.01); thereafter, the number of HSCs was generally decreased compared with the 6 h peak (Fig. 1A, 1C). At days 7 and 14 of treatment, a trend demonstrating a slight decrease in HSC percentage was noted in the bone marrow (p = NS), while a recovery of HSCs was found at day 28 (2 wk following discontinuation of treatment) (Fig. 1A, 1C). The frequency of HSCs was also examined in the spleen of treated mice. We noted an increase in the percentage of HSCs at days 7 and 14 posttreatment, but not at 6 h, as compared with baseline (baseline, 1.48 ± 0.08% versus day 7, 3.49 ± 0.93% [p = 0.004] versus day 14, 2.86 ± 0.97% [p = 0.03]) (Fig. 1B, 1D).
When CXCR4 antagonist was administered for >14 d (i.e., 21 d), no further HSC mobilization was noted in the spleen; this prolonged CXCR4 administration, however, induced a sustained reduction of HSC number in the bone marrow (data not shown).
CXCR4 antagonist diminishes the in vitro alloimmune response
To evaluate the effect of CXCR4 antagonist on the alloimmune response, we used increasing doses of CXCR4 antagonist in vitro in both an MLR and in an anti-CD3/-CD28 stimulation assay. Stimulator BALB/c DCs and responder C57BL/6 CD4+ T cells were incubated with or without CXCR4 antagonist at doses of 1, 10, and 100 μg/ml. When compared to control, treatment with CXCR4 antagonist reduced lymphocyte proliferation, as assessed by [3H]thymidine incorporation in an MLR, with a dose-dependent effect (n = 3, control versus all, p < 0.001) (Fig. 1E) and in an in vitro anti-CD3/-CD28 stimulation assay (n = 3, control versus 10 and 100 μg/ml, p < 0.001) (Fig. 1F).
CXCR4 antagonist prolongs islet allograft survival and mobilizes recipient HSCs
Fully allogeneic islets from BALB/c donors were transplanted under the kidney capsule of C57BL/6 mice chemically rendered diabetic using streptozotocin. CXCR4 antagonist treatment was initiated on the day of transplantation and was sustained for 14 d, with a single daily dose of 30 mg/kg (n = 11 mice). Islet allograft survival was prolonged compared with vehicle-treated control mice (n = 15 mice) (median survival time [MST] of 21 versus 12 d for CXCR4 antagonist-treated versus controls, p < 0.0001) (Fig. 2A).
We evaluated the levels of peripheral HSCs after islet transplantation in CXCR4 antagonist-treated and control C57BL/6 mice (n = at least 5 mice/group). Although islet transplantation per se did not increase HSCs in the spleen as evaluated at day 7 by flow cytometry (Fig. 2B), at day 7 following daily treatment with CXCR4 antagonist, an increase in HSC percentage was noted in the spleen of islet-transplanted C57BL/6 mice (KLS percentage at day 7 in CXCR4 antagonist-treated islet-transplanted mice, 2.17 ± 0.30 versus controls, 0.70 ± 0.05%, p < 0.001) (Fig. 2B).
Islets were recovered from CXCR4 antagonist-treated and control mice at 14 d after transplantation for histological studies. Control-rejecting mice displayed graft infiltration (Fig. 2D,1), absent insulin staining with disrupted islet morphology (Fig. 2D,2), and some CD3+ cells (Fig. 2D,3), which appeared to be Foxp3+, infiltrating the graft (Fig. 2D,4). Conversely, CXCR4 antagonist-treated mice appeared to have some extent of infiltration (Fig. 2E,1), but maintained well-preserved islet structure and insulin staining (Fig. 2E,2) with fewer CD3+ cells infiltrating the graft (Fig. 2E,3), which were found to be Foxp3+ cells (Fig. 2E 4).
To address the central issue of whether HSCs traffic to the graft in our model, we treated transgenic GFP-C57BL/6 mice (in which the β-actin promoter drives GFP expression in essentially all tissues) with CXCR4 antagonist. After 7 d of CXCR4 antagonist treatment, when HSC mobilization peaked in the spleen, we isolated mobilized GFP+HSCs from splenocytes by sorting for Lin−c-Kit+ cells, thus obtaining a population of highly enriched GFP+HSCs. We then injected 1 × 106 sorted GFP+HSCs i.v. into immunocompetent C57BL/6 mice, which had received an islet transplant 1 d prior to injection. Twenty-four hours after the injection of GFP+ HSCs, mice were sacrificed, and islet allografts and spleens were harvested for histological examination. Clusters of GFP+ HSCs were detected in the spleen and islet grafts (Fig. 3A, 3B), thus confirming that HSCs do indeed traffic to islet grafts. With the aim of finding a potential pathway to account for HSC recruitment to the graft, we evaluated the expression of CXCL12 (a chemoattractant for HSCs). CXCL12 expression in the graft was detected by immunohistochemistry at day 7 (Fig. 3C); this expression may confer the ability to recruit HSCs to islet grafts. We then evaluated HSC kinetics in CXCR4 antagonist-treated islet-transplanted and untransplanted mice in spleen, pancreatic lymph nodes, pancreata, and islet grafts. HSCs were mobilized in the spleen at days 7 and 14 during CXCR4 antagonist treatment (Fig. 3D) in transplanted and untransplanted mice (naive untreated versus all, p < 0.01). HSCs were detected in the lymph nodes and pancreata of treated mice and increased in both sites only after 14 d of treatment in islet-transplanted mice (day 14 transplanted mice versus naive untreated, p < 0.01) (Fig. 3E, 3F). Some HSCs were detected in the islet grafts of untreated mice, with a decrease in HSC percentage at day 7 and a return to values similar to baseline at day 14 (Fig. 3G).
Rapamycin synergizes with CXCR4 antagonist treatment
Establishing a safer immunosuppressive protocol is highly indicated in the current landscape of islet cell transplantation (25). In this study, we determined whether CXCR4 antagonist could synergize with a clinically applicable immunosuppressive drug, such as rapamycin. Islet allograft recipient mice were treated with rapamycin at a low clinical-grade dose (0.1 mg/kg) administered every other day (days 0, 2, 4, 6, 8, and 10), alone or in combination with CXCR4 antagonist as described above (30 mg/kg/day for 14 d). While C57BL/6 islet-transplanted mice treated with rapamycin experienced a prolongation of islet allograft survival compared with controls (MST of 26 versus >12 d, rapamycin [n = 7 mice] versus controls [n = 15 mice], p < 0.0001) (Fig. 4A), but ultimately rejected their islet grafts, mice that received a combination of rapamycin and CXCR4 antagonist experienced a considerable delay in allograft rejection (MST of 26 versus >150 d, rapamycin [n = 7 mice] versus rapamycin plus CXCR4 antagonist-treated [n = 5 mice], p = 0.01) (Fig. 4A). Indeed, 75% of C57BL/6 islet-transplanted mice treated with the combination of CXCR4 antagonist and rapamycin enjoyed indefinite allograft survival (>150 d) (Fig. 4A).
We then investigated whether adding rapamycin to our CXCR4 antagonist therapy would enhance HSC mobilization. Indeed, rapamycin synergized with CXCR4 antagonist in mobilizing HSCs (Fig. 4B). At day 7 after islet transplantation during rapamycin plus CXCR4 antagonist treatment, HSC percentage in the spleen was significantly increased compared with controls or in mice treated with CXCR4 antagonist alone (CXCR4 antagonist, 2.17 ± 0.30% versus CXCR4 antagonist plus rapamycin, 3.27 ± 0.19%, p = 0.03) (Fig. 4B).
Islets harvested from mice treated with rapamycin and CXCR4 antagonist (at >150 d) showed intact and barely infiltrated islets (Fig. 4C,1) with preserved insulin staining (Fig. 4C,2). Scarce CD3+ cells (Fig. 4C,3) were detected within the grafts, and the infiltrate consisted primarily of Foxp3+ cells (Fig. 4C 4).
We then evaluated the robustness of CXCR4 antagonist treatment. Spleens were harvested at day 150 in long-term normoglycemic combination-treated mice and in control rejecting mice. No difference was evident in the absolute number of Tregs or T effector cells between the two groups (data not shown), whereas a reduction in IFN-γ production was observed in our ELISPOT assay in which recipient splenocytes (from C57BL/6 mice) were stimulated with donor alloantigen (BALB/c naive splenocytes). Splenocytes harvested from long-term–tolerant CXCR4 antagonist-treated mice produced less IFN-γ upon donor alloantigen challenge compared with splenocytes obtained from control rejecting mice (92.3 ± 7.9 versus 173.2 ± 16.4, p = 0.0008) (Fig. 5A). Although the percentage of Tregs has not always consistently predicted the outcomes of tolerogenic mechanisms, transferable tolerance has been the determinant index of the presence of such mechanisms (26). We therefore recovered splenocytes from long-term–tolerant combination-treated mice 150 d after transplant and adoptively transferred 10 × 106 cells into immunodeficient C57BL/6 RAG mice that were transplanted 2 wk prior with islets from BALB/c donors. Splenocytes from untreated C57BL/6 naive mice, which served as controls, induced islet allograft rejection within 50 d when adoptively transferred into previously islet-transplanted C57BL/6 RAG mice (MST of 35 d, n = 5 mice) (Fig. 5B, group 3). However, >75% of C57BL/6 RAG islet-transplanted mice that received splenocytes from long-term–tolerant combination-treated mice in fact did not reject BALB/c islets (MST of >150 d, n = 5 mice, p = 0.02 versus control mice) (Fig. 5B, group 1). When 10 × 106 splenocytes from long-term–tolerant C57BL/6 mice were coadoptively transferred with 10 × 106 naive splenocytes into immunodeficient C57BL/6 RAG mice previously transplanted with islets from BALB/c donors, >50% of adoptively transferred mice also did not reject islet allografts (MST of >150 d, n = 5 mice, p = NS versus mice receiving adoptive transfer from naive splenocytes) (Fig. 5B, group 2). These results strongly suggest that in an in vivo setting, the combination of CXCR4 antagonist and rapamycin induces a robust and transferable tolerance. As further proof of the tolerance induced by our combination treatment, CXCR4 antagonist plus rapamycin-treated mice (85% tolerant in the long-term) or control rejecting mice were retransplanted with donor (BALB/c) or third-party (DBA/1J) skin grafts. Both control (MST: third party, 13 d; donor, 12 d), and CXCR4 antagonist plus rapamycin-treated mice (MST: third party, 12 d; donor, 14 d) promptly rejected donor and third-party skin grafts, thus confirming the presence of graft-specific tolerance.
CD25 depletion has no effect on CXCR4 antagonist-induced allograft survival
To confirm that the prolongation of islet allograft survival observed in CXCR4 antagonist-treated mice was Treg-dependent, we depleted Tregs using an anti-CD25 mAb. Treg ablation did not induce a loss of CXCR4 antagonist-mediated graft protection (MST of 18 d) (Fig. 2A).
Target of c-Kit and HSC ablation abrogates the effects of CXCR4 antagonist
To establish a link between mobilized HSCs and prolongation of islet allograft survival observed following treatment with CXCR4 antagonist, we performed islet transplantation using CXCR4 antagonist treatment combined with treatment with ACK2, an Ab that targets c-Kit (also known as CD117, the receptor for stem cell factor). ACK2 has been shown to halt bone marrow HSC mobilization in various models and to function by fully inhibiting c-Kit signaling and SCF-dependent stem cell proliferation, resulting in temporary HSC depletion (27, 28). When coinjected with CXCR4 antagonist, ACK2 abrogated the positive effect on islet allograft survival observed following target of the CXCR4–CXCL12 axis (MST of 21 versus 13 d, ACK2 plus CXCR4 antagonist-treated [n = 5] versus CXCR4 antagonist-treated [n = 11], p < 0.0001) (Fig. 2A). Of note, ACK2 treatment alone did not result in any change in survival compared with control mice (MST of 11.5 d, n = 5) (Fig. 2A). Injection of ACK2 also reduced CXCR4 antagonist-mediated HSC mobilization (HSCs at day 7 after islet transplantation: CXCR4 antagonist-treated mice, 2.17 ± 0.30% and CXCR4 antagonist plus ACK2-treated mice, 0.41 ± 0.07%, p < 0.001) (Fig. 2B), whereas ACK2 treatment alone had no effect on HSC mobilization (Fig. 2B). CXCR4 antagonist plus ACK2-treated mice displayed islets infiltrated by few CD3+Foxp3− cells (Fig. 2F), with a total lack of insulin staining (Fig. 2F,2) confirming that the appearance of Tregs in the graft was HSC-mediated. We then tested whether ACK2 treatment and HSC ablation prior to islet transplantation could affect islet allograft survival in untreated, CXCR4 antagonist-treated, and rapamycin-treated islet-transplanted mice. ACK2-pretreated mice displayed islet allograft survival (MST of 14 d, n = 4) comparable to untreated controls (p = NS). In CXCR4 antagonist-treated islet transplanted mice, ACK2 treatment (14 d) (but not ACK2 pretreatment [31 d]) abrogated the effect of CXCR4 antagonist by diminishing islet allograft survival (p = 0.005) (Fig. 2A). Rapamycin alone (0.1 mg/kg) administered every other day (at days 0, 2, 4, 6, 8, and 10) prolonged islet allograft survival (26 d) compared with controls (12 d) (p < 0.001) (Fig. 2C). ACK2 treatment (18.5 d) (but not ACK2 pretreatment [24 d]) abrogated the effect of rapamycin on islet allograft survival (p = 0.05) (Fig. 2C).
Furthermore, treatment with ACK2 abrogated the beneficial effect of the combination therapy of CXCR4 antagonist plus rapamycin on islet allograft survival (MST of 21 versus >150 d, CXCR4 antagonist plus rapamycin plus ACK2-treated [n = 5] versus CXCR4 antagonist plus rapamycin-treated [n = 5], p < 0.0001) (Fig. 4A). We also tested the effect of ACK2 administration on HSC mobilization in the CXCR4 antagonist plus rapamycin-treated group. When ACK2 was coinjected, a distinct abrogation of HSC mobilization was evident (KLS at day 7 after islet transplantation: CXCR4 antagonist plus rapamycin-treated mice, 3.27 ± 0.19% and CXCR4 antagonist plus rapamycin plus ACK2-treated mice, 0.80 ± 0.10%, p < 0.001) (Fig. 4B). Rapamycin alone induced a slight increase in the peripheral percentage of HSCs (p < 0.01 compared with controls) (Fig. 4B).
Role of PD-L1 in HSC-mediated effect
We evaluated the characteristics of CXCR4 antagonist-mobilized HSCs by FACS analysis. Lin− cells were isolated from splenocytes by microbead-negative selection using biotinylated lineage beads, followed by FACS sorting for c-Kit+ cells. Cells were isolated from islet-transplanted or naive CXCR4 antagonist-treated mice, after 7 and 14 d of CXCR4 antagonist treatment. As reported previously, these Lin−c-Kit+ cells are 50% positive for Sca-1 (Fig. 6A), indicative of murine HSCs (29). HSCs were then stained for specific positive and negative costimulatory molecules that have been shown to exert significant immunoregulatory roles in the alloimmune response. Interestingly, while most positive costimulatory molecules were found to be negative or scarcely expressed (CD40, CD80, CD86, PD-L2, ICOS, OX40, OX40L) (Fig. 6B), PD-L1 (a negative costimulatory molecule) was highly expressed by mobilized HSCs (58.0 ± 7.1%). Extracted HSCs also expressed CXCR4 (38.4 ± 4.2%) (Fig. 6B). The profiling of costimulatory molecules on HSCs obtained from CXCR4 antagonist-treated islet-transplanted and untransplanted treated mice at days 7 and 14 after transplantation confirmed that PD-L1 is constitutively expressed by HSCs (Fig. 6C), whereas most other receptors are absent or scarcely expressed (Fig. 6C). OX40L expression on HSCs increased transiently in untransplanted mice at day 7 (p < 0.05 versus all), and PD-1 expression on HSCs increased at day 7 in transplanted and untransplanted mice (p < 0.01 versus all) (Fig. 6C).
We then evaluated the in vivo percentage of c-Kit+PD-L1+ cells during CXCR4 antagonist treatment. c-Kit+PD-L1+ cells increased in bone marrow from C57BL/6 islet-transplanted mice 6 h after the initiation of CXCR4 antagonist treatment (30 mg/kg/day), but this increase was no longer evident at 12 and 24 h (baseline, 10.3 ± 4.4; 6 h, 31.5 ± 5.4; 12 h, 17.2 ± 5.1; and 24 h, 18.6 ± 5.8%; baseline versus 6 h, p = 0.04) (Fig. 6D).
To evaluate the immunoregulatory role of the PD-1 pathway in murine HSCs, we investigated the effect of mobilized HSCs from wild-type (WT) and PD-L1 knockout (KO) mice on the alloimmune response in vitro. A standard MLR assay was performed in which HSCs (from WT C57BL/6 or PD-L1 KO mice) were syngeneic to responder cells (CD4+ cells from C57BL/6 mice) but allogeneic to bone marrow-derived DCs (from BALB/c mice), as in our in vivo setting. While HSCs from WT C57BL/6 mice were capable of significantly abrogating the MLR response when added to culture (control, 27,000 ± 2,374 versus +50,000 HSCs, 16,329 ± 1,641, p = 0.003), HSCs from PD-L1 KO mice failed to do so (Fig. 7A, 7B).
To examine the in vivo functional role of PD-L1 on HSCs in the prolongation of islet allograft survival, we performed islet transplantation in PD-L1 KO mice with or without CXCR4 antagonist treatment. Untreated PD-L1 KO C57BL/6 mice rejected BALB/c islet allografts (MST of 22 d, n = 4 mice), albeit with a delay compared to WT C57BL/6 mice (MST of 12 d, n = 15 mice, p = 0.001) (Fig. 2A, 7C). In contrast to the prolongation observed in C57BL/6 WT mice, CXCR4 antagonist treatment failed to further delay islet allograft rejection in PD-L1 KO C57BL/6 mice (MST of 26 d, n = 4 mice, p = 0.001 versus WT C57BL/6 mice and p = NS versus PD-L1 KO untreated mice) (Fig. 7C).
Islet transplantation holds great promise as a potential cure for type 1 diabetes (25, 30–32), yet long-term survival of islet grafts remains problematic (20% at 5 y) (33). Thus, there exists an immense need for new therapeutic options to aid in prolonging islet allograft survival (22, 34, 35). In this study, we propose a novel approach to promote tolerance in an islet transplantation model by inducing the shedding of HSCs, using a CXCR4 antagonist (NIBR1816), with the goal of mobilizing the immunomodulatory ability of autogolous HSCs to dampen the alloimmune response and to prolong islet allograft survival. CXCL12–CXCR4 signaling effectively traps HSCs in the bone marrow (19). Through target of the CXCR4–CXCL12 axis with our novel small molecule NIBR1816, a CXCR4 antagonist, we were able to induce the shedding of HSCs to the periphery similar to others (20). We also show that mobilization of HSCs induced by our strategy prolongs islet allograft survival in a fully mismatched model of islet transplantation. Notably, the tolerogenic effect of CXCR4 antagonist can be abrogated in large part by blocking HSC mobilization induced by the CXCR4 antagonist through use of ACK2 (an anti-CD117 mAb), which targets c-Kit (the receptor of stem cell factor) (27, 28).
We then tested the synergistic effect of CXCR4 antagonist on a commonly used immunosuppressive drug, rapamycin. Our combined treatment significantly prolonged islet allograft survival; notably, 75% of islet-transplanted mice did not reject islet grafts until beyond 100 d post transplantation. Interestingly, rapamycin enhanced the mobilization of HSCs. The extent to which this mobilization of HSCs contributes to the prolongation of survival of allografts is important to address and requires future further investigation. Long-term tolerant mice (treated with a combination of rapamycin at clinical dose plus CXCR4 antagonist) developed robust hyporesponsiveness toward alloantigens, as demonstrated by our adoptive transfer experiments in C57BL/6-RAG immunodeficient mice, which did not reject BALB/c islet allografts when reconstituted with splenocytes from long-term–tolerant mice treated with rapamycin plus CXCR4 antagonist. Furthermore, among C57BL/6-RAG immunodeficient mice co-reconstituted with splenocytes from long-term–tolerant mice treated with rapamycin plus CXCR4 antagonist as well as with splenocytes from naive mice, >50% of mice did not reject islet grafts. This robust and transferable tolerance, taken together with the reduced production of IFN-γ in response to alloantigen in vitro, suggests the existence of an active hyporesponsiveness and a regulatory process. Unfortunately, tolerance was graft-specific; when islet-tolerant mice were transplanted with skin, they rejected both donor and third-party skin grafts. Our trafficking experiments using isolated GFP+ HSCs injected into islet-transplanted mice confirmed that autologous HSCs traffic to the lymphoid organs, to the pancreas, and to the graft, possibly exerting their immunomodulatory role locally. Although it is difficult to determine precisely which cell types are involved in the transfer of tolerance, HSCs and Tregs appear to be the primary candidates; however, myeloid and other progenitor cells may also be involved.
To better elucidate the mechanisms by which HSCs were able to induce this prolongation, we characterized mobilized HSCs for their expression of immunoregulatory molecules. Mobilized HSCs were shown to express the negative costimulatory molecule PD-L1 at substantial levels, while they appeared to be negative for most positive costimulatory molecules. Whereas HSCs from WT mice inhibited the alloimmune response in our MLR assay in vitro, HSCs extracted from PD-L1 KO mice failed to exert an inhibitory effect in our MLR. CXCR4 antagonist treatment also appeared ineffective in delaying graft rejection in PD-L1 KO mice. PD-L1 has been shown to be expressed on resting and activated B cells, T cells, myeloid cells, and DCs, to play a role in regulating the alloimmune response in vivo by inducing apoptosis of T effector cells, and to delay allograft rejection (36–41). We postulate that PD-L1 expression on HSCs plays a principal role in the tolerogenic effect of HSCs. Precise evaluation of trafficking of mobilized HSCs is problematic, as HSC may differentiate into myeloid or lymphoid lineages and thus alter their phenotype. Moreover, the half-life of HSCs has not yet been determined, and without a GFP+ HSC-specific mouse, accurate tracking of HSCs is difficult to achieve. Our data indicate that HSCs traffic to the spleen and appear to be present to a lower extent in the pancreatic lymph nodes, pancreas, and islet graft.
Using a CXCR4 antagonist, available for human use (i.e., Mozobil/plerixafor), to elicit shedding of HSCs has become an appealing strategy to mobilize HSCs and to facilitate HSC engraftment following bone marrow transplantation (16). Since HSCs have recently emerged as the key regulators of immune responses (42), our strategy is highly clinically relevant, as it would eliminate the need for bone marrow transplantation, which usually requires institution of chemotherapy and radiation therapy. In summary, we have demonstrated that targeting the CXCR4–CXCL12 axis with a CXCR4 antagonist can mobilize autologous HSCs, resulting in increased survival of islet allografts. This effect synergized with rapamycin, and HSCs likely exert their tolerogenic effect in part through the PD-L1 pathway.
Interestingly, we observed synergism between rapamycin and CXCR4 antagonist in mobilizing HSCs; as rapamycin is a well-known antiproliferative drug, this result was unexpected (43). Of note, recent works have shown that rapamycin is capable of halting and reverting age-related decline in HSC function (44). Investigators have demonstrated that mammalian target of rapamycin (mTOR) activity is increased in HSCs from old mice compared with those from young mice (44). In old mice, rapamycin increased life span and restored the self-renewal and hematopoiesis of HSCs (44), thus suggesting that mTOR signaling may be dangerous for HSCs and demonstrating the potential of mTOR inhibitors for restoring full competent hematopoiesis. Whether HSC mobilization has any utility in patients with oncogene/tumor suppressor mutations remains to be explored (45), and caution must be taken when considering mobilizing HSCs in patients with disorders of hematopoiesis.
It should be noted that CXCR4 blockade and HSC-mediated immunomodulation are not mutually exclusive. CXCR4 antagonist treatment may also affect several arms of effector and regulatory pathways, which are additionally important for transplant outcome. For instance, CXCR4 signaling could play an important role in the trafficking of Tregs and DCs, both of which are central to the pathogenesis of alloimmunity. Furthermore, c-Kit expressed on DCs has been shown to regulate the differentiation of Th cells (46, 47), so that anti–c-Kit treatment could also affect the outcome of transplantation. Although our data shed light on the effect of HSC release in organ transplantation and mechanisms by which HSC exert their function in prolonging graft survival, examining the relative contribution of these pathways is beyond the scope of this study but is nevertheless important to fully explore in future studies. It should be further noted that ACK2 treatment may also affect cells other than HSCs.
Our data could ultimately aid in the design of tolerogenic strategies in human islet cell transplantation, which is in distinct need of improvement.
We thank Novartis for providing NIBR1816.
Disclosures The authors have no financial conflicts of interest.
This work was supported by an American Society of Transplantation/Juvenile Diabetes Research Foundation Faculty grant, a Juvenile Diabetes Research Foundation Career Development award, and an American Society of Nephrology Career Development award (all to P.F.). A.V. is the recipient of an American Society of Transplantation/Juvenile Diabetes Research Foundation Fellowship grant. P.F. received support from a pilot and feasibility award from the Boston Area Diabetes Endocrinology Research Center (5P30DK57521). This work also was supported by Juvenile Diabetes Research Foundation Grant 4-2007-1065, a Juvenile Diabetes Research Foundation regular grant (to R.A.) and National Institutes of Health Grants P01 AI041521, R01 AI070820, R01 AI051559, and P01 AI056299 (to M.H.S.).
Abbreviations used in this paper:
hematopoietic stem cell
median survival time
mammalian target of rapamycin
programmed death ligand 1
regulatory T cell