Heparan sulfate, an endogenous TLR4 agonist, promotes acute GVHD after allogeneic stem cell transplantation

Abstract

Graft-versus-host disease (GVHD) remains the most common cause of nonrelapse-related morbidity and mortality after allogeneic hematopoietic stem cell transplantation (allo-HSCT). Although T-cell depletion and intensive immunosuppression are effective in the control of GVHD, they are often associated with higher rates of infection and tumor recurrence. In this study, we showed that heparan sulfate (HS), an extracellular matrix component, can activate Toll-like receptor 4 on dendritic cells in vitro, leading to the enhancement of dendritic cell maturation and alloreactive T-cell responses. We further demonstrated in vivo that serum HS levels were acutely elevated at the onset of clinical GVHD in mice after allo-HSCT. Treatment with the serine protease inhibitor α1-antitrypsin decreased serum levels of HS, leading to a reduction in alloreactive T-cell responses and GVHD severity. Conversely, an HS mimetic that increased serum HS levels accelerated GVHD. In addition, in patients undergoing allo-HSCT for hematologic malignancies, serum HS levels were elevated and correlated with the severity of GVHD. These results identify a critical role for HS in promoting acute GVHD after allo-HSCT, and they suggest that modulation of HS release may have therapeutic potential for the control of clinical GVHD.

Figures

Figure 1

HS is a potent stimulator of alloreactive T-cell responses through the TLR4- and MyD88-dependent activation of DCs. (A) TLR and NLR agonists were assayed in allogeneic T-cell proliferation assay between purified T cells (2 × 105/well) from C57BL/6 mice and bone marrow–derived BALB/c DCs (2.5 × 104/well). Cells were cocultured either alone (media) or in the presence of LPS (100 ng/mL), Pam3CSK4 (2 μg/mL), hyaluronan (HA; 100 μg/mL), sonicated-HA (sHA; 100 μg/mL), fibronectin (FN; 100 μg/mL), fibrinogen (Fbn; 100 μg/mL), HS (100 μg/mL), Hsp70 (5 μg/mL), HMGB1 (1 μg/mL), C12-iE-DAP (1 μg/mL), or L18-MDP (1 μg/mL) for 72 hours and then pulsed [3H]thymidine for 16 hours. Proliferation was determined by 3H incorporation and results are expressed as cpm ± SEM. Baseline alloreactivity is indicated by the dotted line (*P < .05 compared with media alone). (B) Proliferation performed as described in panel A ± the addition of the LPS inhibitor polymyxin B (PMB; 10 μg/mL; *P < .05). (C) Proliferation assay performed as described in panel A with purified responder T cells (R) from either WT (+) or MyD88−/− (−) C57BL/6 mice were cocultured with DC stimulators (S) from either WT (+) or MyD88−/− (−) BALB/c mice (*P < .05 compared with media alone in S+/R+ group. (D-E) Analysis of proliferation and IFN-γ production in proliferation assays performed as described in panel A using WT, TLR4−/−, and MyD88−/− BALB/c DCs as stimulators and purified C57BL/6 T cells as responders (*P < .05). Results are representative of 3 independent experiments.

Figure 2

HS promotes DC maturation and production of proinflammatory cytokines via the TLR4-MyD88 pathway. (A) WT, TLR4−/−, or MyD88−/− BALB/c DCs (2 × 105/well) were stimulated with LPS (100 ng/mL), HS (100 μg/mL), or Pam3CSK4 (2 μg/mL), or left unstimulated (media) for 24 hours and then measured for surface expression of costimulatory molecules CD40 and CD80 by FACS analysis. (B-C) WT, MyD88−/−, and TLR4−/− BALB/c cultured DCs were cocultured with media alone, LPS, HS, or Pam3CSK4 as described in panel A, and culture supernatants were tested for IL-6 (B) and IL-12 (C) by ELISA. Data are representative of 3 independent experiments (*P < .05 compared with media alone). (D-E) Assay of DC production of IL-6 and IL-12 ± the addition of the LPS inhibitor PMB (10 μg/mL) after stimulation with media, HS, or LPS (*P < .05).

Figure 3

Serum HS is highly elevated at the onset of GVHD. Lethally irradiated BALB/c recipients received either 1 × 107 B10.D2 TCD-BM only (allo-BM), 1 × 107 B10.D2 TCD-BM and 5 × 106 B10.D2 LCs (allo-BM + LC), or 1 × 107 BALB/c TCD-BM and 5 × 106 BALB/c LCs (Syn-BM + LC). (A) After transplantation, serum HS concentrations were determined by ELISA at the indicated time points; n = 2-5 samples per time point (*P < .05 comparing allo-BM + LC and allo-BM at the indicated time point. (B) To determine the half-maximal effective concentration of HS on DC stimulation, BALB/c DCs (2 × 105/well) were cultured for 24 hours with differing concentrations of HS in triplicate, and IL-6 production was tested by ELISA. Results are representative of 3 independent experiments.

Figure 4

A1AT decreases serum HS levels and improves the outcome of GVHD after allo-HSCT. (A) Serum HS concentrations were determined by ELISA at the indicated time points after allo-HSCT (B10.D2→BALB/c; 1 × 107 B10.D2 TCD-BM and 5 × 106 B10.D2 LC) treated with A1AT (2 mg) or PBS every 3 days by intraperitoneal injection, starting 1 day before transplantation; n = 3 per data point (*P < .05). Survival (B) and GVHD clinical score (C) of allo-BM only (n = 5) or allo-BM + LC treated with A1AT (n = 8) or PBS (n = 5). Data are from 1 of 2 independent experiments with identical results. GVHD pathology score (D) and representative H&E histology (E) of BALB/c recipients of B10.D2 (allo) TCD-BM + LC treated with PBS or with A1AT (n = 6 per group; bar denotes 100 μm (*P = .05). The micrographs were taken from H&E sections (200× magnification) using the 20× PlanApochromatic objective with an Olympus Vanox-AHBS-3 microscope. The camera used is Olympus DP-70 with its own acquisition software. (F) Survival of lethally irradiated C57BL/6 recipients of 1 × 107 C3H.SW TCD-BM (allo-BM) and 5 × 106 C3H.SW LC administered A1AT (n = 10) or PBS (n = 9). (G) Improvement in GVHD survival by A1AT is dependent on host TLR4 expression. Survival of BALB/c recipients of allo-BM + LC from B10.D2 donors (n = 11) and TLR4−/− BALB/c recipients of allo-BM + LC from B10.D2 donors administered A1AT (n = 8) or PBS (n = 14) every 3 days by intraperitoneal injection, starting 1 day before transplantation.

Figure 5

HS mimetic increases serum HS levels and increases CD8 T-cell proliferation in allo-HSCT recipients. BALB/c recipient of B10.D2 allo-BM + LC were treated with subcutaneous injections of the HS mimetic PG545 (20 mg/kg in PBS), or PBS control once weekly, beginning 1 day before allo-HSCT. (A) Serum HS levels were determined by ELISA on the indicated days after transplant; n = 2-4 samples each (*P < .05). (B) BrdU uptake by CD8 T cells 6 days after allo-HSCT. Average and SEM are plotted; n = 3 per group (*P < .05). (C) Survival analysis of allo-BM + LC+PG545 (n = 10) compared with allo-BM + LC + PBS (n = 10).

Figure 6

Persistence of recipient MHC class II–expressing cells after allogeneic HSCT. Lethally irradiated C57BL/6 recipients received 107 TCD-BM + 106 LC from either allogeneic BALB/c donors or syngeneic C57BL/6 donors. Fourteen days after transplant, recipient mice were injected with 2 × 106 CFSE-labeled lymphocytes from 4C TCR-tg mice (direct allospecificity toward the BALB/c MHC class II molecule I-Ad) that were on the Ly5.1 congenic background. Recipient intrahepatic lymphocytes were harvested 3 days later and FACS-analyzed. (A) Schematic of experiment. (B) FACS gates for detection of 4C TCR-tg T cells and CFSE analysis. Results shown are representative of 4 mice in each group.

Figure 7

HS is elevated in serum samples of human allo-HSCT recipients with GVHD. (A) Serum samples from allo-HSCT recipients were tested for HS by ELISA. Patients were divided into 3 groups: no GVHD (grade 0; n = 8), mild GVHD (grade 1-2; n = 17), and moderate to severe GVHD (grade 3-4; n = 11; *P = .003, **P = .0009, ***P = .01). (B) Serum HS levels relative to time of diagnosis of GVHD in patients with grades 1 to 2 and grade 3 to 4 GVHD. Average ± SEM plotted (*P = .01).