In vivo immune modulatory activity of hepatic stellate cells in mice

Authors

  • Cheng-Hsu Chen,

    1. Thomas E. Starzl Transplantation Institute, Department of Surgery, University of Pittsburgh, Pittsburgh, PA
    Search for more papers by this author
    • Cheng-Hsu Chen and Liang-Mou Kuo contributed equally to this study.

  • Liang-Mou Kuo,

    1. Thomas E. Starzl Transplantation Institute, Department of Surgery, University of Pittsburgh, Pittsburgh, PA
    Search for more papers by this author
    • Cheng-Hsu Chen and Liang-Mou Kuo contributed equally to this study.

  • Yigang Chang,

    1. Division of Immunogenetics, Children's Hospital of Pittsburgh, Pittsburgh, PA
    Search for more papers by this author
  • Wenhan Wu,

    1. Thomas E. Starzl Transplantation Institute, Department of Surgery, University of Pittsburgh, Pittsburgh, PA
    Search for more papers by this author
  • Christina Goldbach,

    1. Department of Cell Biology and Physiology, University of Pittsburgh, Pittsburgh, PA
    Search for more papers by this author
  • Mark A. Ross,

    1. Department of Cell Biology and Physiology, University of Pittsburgh, Pittsburgh, PA
    Search for more papers by this author
  • Donna B. Stolz,

    1. Department of Cell Biology and Physiology, University of Pittsburgh, Pittsburgh, PA
    Search for more papers by this author
  • Liepin Chen,

    1. Department of Dermatology and Oncology, Johns Hopkins University School of Medicine, Baltimore, MD
    Search for more papers by this author
  • John J. Fung,

    1. Thomas E. Starzl Transplantation Institute, Department of Surgery, University of Pittsburgh, Pittsburgh, PA
    2. Department of General Surgery, Cleveland Clinic Foundation, Cleveland, OH
    Search for more papers by this author
  • Lin Lu,

    Corresponding author
    1. Thomas E. Starzl Transplantation Institute, Department of Surgery, University of Pittsburgh, Pittsburgh, PA
    2. Department of General Surgery, Cleveland Clinic Foundation, Cleveland, OH
    3. Department of Immunology, Cleveland Clinic Foundation, Cleveland, OH
    • Department of Immunology, Lerner Research Institute, Cleveland Clinic Foundation, 9500 Euclid Avenue, NE40, Cleveland, OH 44195
    Search for more papers by this author
    • fax: 216-444-8372

  • Shiguang Qian

    Corresponding author
    1. Thomas E. Starzl Transplantation Institute, Department of Surgery, University of Pittsburgh, Pittsburgh, PA
    2. Department of General Surgery, Cleveland Clinic Foundation, Cleveland, OH
    3. Department of Immunology, Cleveland Clinic Foundation, Cleveland, OH
    • Department of Immunology, Lerner Research Institute, Cleveland Clinic Foundation, 9500 Euclid Avenue, NE40, Cleveland, OH 44195
    Search for more papers by this author
    • fax: 216-444-8372


  • Potential conflict of interest: Nothing to report.

Abstract

Accumulating data suggest that hepatic tolerance, initially demonstrated by spontaneous acceptance of liver allografts in many species, results from an immune regulatory activity occurring in the liver. However, the responsible cellular and molecular components have not been completely understood. We have recently described profound T cell inhibitory activity of hepatic stellate cells (HSCs) in vitro. In this study, we demonstrate in vivo evidence of immune modulatory activity of HSCs in mice using an islet transplantation model. Co-transplanted HSCs effectively protected islet allografts from rejection, forming a multi-layered capsule, which reduced allograft immunocyte infiltrates by enhancement of apoptotic death. The immune modulation by HSCs appeared to be a local effect, and regulated by inducible expression of B7-H1, an inhibitory molecule of B7 family. This may reflect an intrinsic mechanism of immune inhibition mediated by liver-derived tissue cells. In conclusion, these results may lead to better understanding of liver immunobiology and development of new strategies for treatment of liver diseases. (HEPATOLOGY 2006;44:1171–1181.)

Hepatic tolerance was initially recognized by spontaneous acceptance of liver allografts in a number of species.1–4 Our laboratory focused on the mechanistic insights into liver transplant tolerance in mice and demonstrated that it is not attributable to a deletion of T cell clones specific to donor antigens, because lymphocytes isolated from tolerant mice respond well to donor antigen stimulation in vitro.5 Whereas histological findings are notable for an abundance of CD4 and CD8 T cell infiltrates during the first week after transplantation, they are rapidly diminished via apoptotic death.6–8 Similarly, in a nontransplant model, after injection of a specific antigenic peptide into TCR transgenic mice, deletion of specific CD8+ T cells from the periphery was observed, as they accumulated in the liver and underwent subsequent apoptosis.9 Combined, these results suggest a regulatory mechanism residing in the liver that may be responsible for its immune modulation.10 However, the cellular components of the liver responsible for this are not completely understood.

We have recently demonstrated that activated hepatic stellate cells (HSCs) effectively inhibit T cell responses in vitro, which is mediated by induction of T cell apoptosis.11 Thus, subsequently studying whether HSCs can exert immune regulatory activities in vivo is logical. However, it is difficult to obtain direct evidence in the liver because approaches to specifically inhibit the effect of HSCs have not been defined. In this study, we used an islet allograft transplant model and demonstrated that co-transplantation with HSCs effectively protected islet allografts from rejection by forming a multi-layered immune barrier around the islets and reducing infiltrating T cells. Interestingly, HSCs isolated from B7-H1 knockout (KO) mice lost the protective effect on co-transplanted islet allografts, indicating a critical role of B7-H1 in immune regulatory activity of HSCs.

Abbreviations

HSC, hepatic stellate cell; KO, knockout; NPC, nonparenchymal cells; mAb, monoclonal antibody; IL, interleukin; MLR, mixed lymphocyte reaction; DC, dendritic cell; CTL, cytotoxic T lymphocyte; CFSE, carboxyfluoroscein succinimidyl ester; SMA, smooth muscle actin; TUNEL, terminal deoxynucleotidyl transferase-mediated nick-end labeling; MST, median survival time; IFN-γ, interferon-gamma; TGF-β, transforming growth factor beta.

Materials and Methods

Animals.

Male C57BL/6 (B6; H2b), C3H (H2k), BALB/c (H2d), B10.BR (H2k) and enhanced green fluorescent protein (EGFP) transgenic mice [C57BL/6-TgN(ACTbEGFP)10sb, H2b] were purchased from the Jackson Laboratory (Bar Harbor, ME). B6. B7-H-1 KO mice were provided by Dr. Lieping Chen at Johns Hopkins University Medical School, Baltimore, MD. All animals were maintained in the specific pathogen-free facility of the University of Pittsburgh Medical Center and provided with Purina rodent chow and tap water ad libitum and used at 8 to 10 weeks of age following NIH guidelines.

Preparation of HSCs.

HSCs were isolated from the mouse liver nonparenchymal cells (NPC) as previously described12 with some modifications.11 The isolated HSCs were cultured (105/mL) in a Nunclon cell culture flask (25 cm2 surface area) (NUNC A/S, Roskilde, Denmark) with RPMI-1640 (Mediatech Inc., Herndon, VA) supplemented with 10% fetal bovine serum and 10% horse serum in 5% CO2 in air at 37°C for 2 or 7 days. Cell viability was greater than 90% as determined by trypan blue exclusion. The purity of HSCs was determined by desmin immunostaining and the typical light microscopic appearance of the lipid droplets.13

Flow Cytometry.

Expression of cell surface molecules was analyzed using an EPICS ELITE flow cytometer (Coulter Corporation, Hialeah, FL) and stained with anti-CD45 (rat IgG2b), -CD54 (hamster IgG), or -H-2Kb, -I-Ab(both mouse IgG2a), respectively (all from BD PharMingen, San Diego, CA). Expression of B7-H1 was identified by anti-B7-H1 monoclonal antibody (mAb) (rat IgG2a, eBioscience, San Diego, CA). The appropriate isotype control antibodies were used.

Dendritic Cell Culture.

As previously described,14 bone marrow cells were isolated from mouse femurs and tibias, after lysis of red blood cells, cultured in RPMI-1640 complete medium in the presence of mouse recombinant (r) granulocyte-macrophage colony-stimulating factor (4 ng/mL) and interleukin-4 (IL-4) (1,000 unit/mL) (both from Schering-Plough, Kenilworth, NJ). Nonadherent cells were released spontaneously from the proliferating cell clusters. These cells were then harvested, washed, and resuspended in medium.

Mixed Lymphocyte Reaction.

Nylon wool–eluted B6 spleen T cells (2 × 105/well in 100 μL) were cultured in triplicate in 96-well round-bottom microculture plates (Corning Inc., Corning, NY) with graded doses of γ-irradiated (20 Gy; X-ray source) DCs from BALB/c mice in RPMI-1640 complete medium, for 3 to 4 days in 5% CO2 in air. [3H]TdR (1 μCi/well) was added for the final 18 hours, and incorporation of [3H]TdR into DNA was assessed by liquid scintillation counting. Results were expressed as mean counts per minute ± 1 SD. To examine the effect of HSCs on T cell proliferation, γ-irradiated (50 Gy) HSCs were added into the cultures at the beginning of the culture.

In Vitro Cytotoxic T Lymphocyte Assay.

B6 spleen T cells cultured with γ-irradiated (20 Gy) BALB/c dendritic cells (DCs) at a ratio of 10:1 for 5 to 6 days were used as effectors. P815 (H2d), EL4 (H2b), or R1.1 (H2k) lymphoma cell lines (all from American Type Culture Collection, Manassas, VA) labeled with 100 μCi Na251CrO4 (NEN, Boston, MA) were used as donor-specific, syngeneic or third-party targets, respectively. They were plated at 5 × 103 cells/well in 96-well round-bottom culture plates (Corning Inc.). Serial 2-fold dilutions of effector cells were added (total volume of 200 μL/well). The percentage of 51Cr release was determined after incubating the plates for 4 hours at 37°C in RPMI-1640 complete medium in 5% CO2 in air. Maximum 51Cr release was determined by lysis of the target cells.

In Vivo Cytotoxic T Lymphocyte Assay.

Cytotoxic T lymphocyte (CTL) activity against islet donor-specific targets was assessed as described15 with some modification. Briefly, BALB/c mouse splenocytes (107/mL) were labeled with a low concentration (0.5 μmol/L) of carboxyfluorescein succinimidyl ester (CFSE) (Molecular Probes, Eugene, OR) and used as donor targets (CFSElow). B6 splenocytes labeled with a high concentration (5 μmol/L) of CFSE were used as a syngenic control (CFSEhigh). Ten million cells of each population were mixed, and intravenously injected into naïve B6 mice (controls) or B6 recipients of islets (BALB/c) alone or plus HCSs (B6). Three hours later, CFSElow and CFSEhigh cells in spleen were analyzed by flow cytometry (100,000 events) using ModFit LT cell cycle analysis software (Verity Software House, Topsham, ME). Percent specific lysis of target cells were calculated as: 100 − [(percentage CFSEhigh/percentage CFSElow) × 100].

Islet Transplantation.

Diabetes was induced in recipients with a single intraperitoneal injection of streptozotocin (220 mg/kg body weight, Sigma-Aldrich, St. Louis, MO). Only mice with nonfasting blood glucose levels exceeding 350 mg/dL were used as recipients. Islets were isolated from donor pancreas by collagenase V (Sigma-Aldrich) digestion as previously described16 with slight modifications.17 After separation on a Ficoll gradient (Type 400, Sigma-Aldrich), the islets were purified by hand picking. Three hundred freshly isolated islets alone or mixed with 1 or 3 × 105 HSCs were aspirated into polyethylene tubing (PE-50, Becton Dickinson, Parsippany, NJ), and pelleted by centrifugation for 2 minutes, and then gently placed under the subcapsular space of recipient kidney. Transplantation was considered successful if the nonfasting blood glucose returned to and remained normal (<150 mg/dL) for the first 4 days after transplantation. Tail vein nonfasting blood glucose was monitored after transplantation, and the first day of 2 consecutive readings of blood glucose greater than 350 mg/dL was defined as the date of diabetes onset. No immunosuppressive reagents were administered throughout the experiments.

Immunohistochemistry.

CD4, CD8 T cells, and HSCs were identified in cryostat sections using biotinylated rat anti-mouse CD4, CD8 (BD PharMingen) and mouse anti-mouse α smooth muscle actin (SMA) mAb (IgG2a, Sigma-Aldrich), respectively. The isotype and species-matched irrelevant mAbs were used as controls. Apoptotic cells were identified by terminal deoxynucleotidyl transferase-mediated dUTP nick-end labeling (TUNEL). Each experiment was performed with a negative control (without dUTP) and a positive control (10 minute pretreatment of slides with 1 mg/mL DNase in reaction buffer). The color was developed by an enzyme reaction using avidin-biotin-alkaline phosphatase complex as the substrate. The slides were counterstained with Harris's hematoxylin and mounted with Crystal mount (Biomeda Corp., Foster City, CA). For quantification, the positive cells were counted under a microscope, and a total of 30 high-power fields (HPFs) were randomly selected in each group. The immunofluorescence protocols were performed to determine expression of CD3, CD31 (both from BD Pharmingen), insulin (Santa Cruz Biotechnology, Santa Cruz, CA), α-SMA (Sigma-Aldrich), and enhanced green fluorescent protein (EGFP) (Abcam, Cambridge, MA). Slides were analyzed using an Olympus BX 51 fluorescence microscope (Olympus America, Center Valley, PA). Quantification of CD31, insulin, and CD3 was performed using MetaMorph software.

Statistical Analysis.

The parametric data were analyzed by Student t test (2-tailed). Graft survival between groups of transplanted animals was compared using the log-rank test. Values of P < .05 were considered statistically significant.

Results

Isolation and Activation of HSCs.

HSCs isolated from B6 mouse liver NPC with purity greater than 95% determined by desmin immunostaining expressed MHC class I, but not CD45 (Fig. 1A). HSCs were activated by culture for 7 days and became positive expression of αSMA, B7-H1, and ICAM-1 as previous described.11 The immune inhibitory function of activated HSCs was examined by addition of HSCs in a mixed leukocyte reaction (MLR) assay. Figure 1B showed that activated HSCs markedly inhibited T cell thymidine uptake, but not the quiescent HSCs as previously demonstrated.11

Figure 1.

Hepatic stellate cell (HSC) isolation and activation. (A) HSCs were isolated from B6 mouse liver nonparenchymal cells by gradient centrifugation as described in Methods. Flow cytometric analyses (histograms) showed that freshly isolated HSCs (filled) were distinct from hepatic lymphocytes (opened) in not expressing CD45 (lymphocyte marker). Open dotted areas are isotype controls. (B) HSCs were cultured in an uncoated plastic flask for 2 days [quiescent (qHSCs)] or 7 days [activated (aHSCs)]. Irradiated HSCs (50 Gy) were added at the beginning into an MLR culture in which splenic T cells (2 × 105) from B6 (H2b) mice and irradiated (20 Gy) BALB/c (H2d) DCs (B6 DCs and HSCs only were used as controls) were cultured at a T:DC:HSC ratio of 40:2:1 for 3 days. Compared with the MLR culture without HSCs, T cell thymidine uptake stimulated by allogeneic DCs was almost completely inhibited by activated HSCs. (P < .05), but not by quiescent HSCs (P > .05). cpm, counts per minute.

Co-transplantation with HSCs Protects Islet Allografts from Rejection.

HSCs from B6 (H2b) mice were mixed with 300 islets isolated from BALB/c (H2d) pancreata, and transplanted under the capsule of the left kidney of diabetic B6 recipients. Co-transplantation with activated (cultured for 7 days), but not quiescent (cultured for 2 days) HSCs (3 × 105) markedly prolonged islet allograft survival (median survival time [MST] >60 days, compared with islet grafts alone (MST = 13 days, P < .01). 55% (6 of 11) of recipients that received islets plus activated HSCs remained euglycemic for more than 60 days, whereas none (0 of 9) of recipients with islet-only grafts remained euglycemic by 17 days after transplantation (Table 1). Removal of the kidney bearing the islet grafts was performed in all animals maintaining euglycemia for more than 60 days at various times (from day 65 to 170 after transplantation), resulting in a prompt return to hyperglycemia (Fig. 2A), indicating the presence of functional transplanted islets. This was further supported by immunohistochemical staining showing that the removed kidney contained insulin-positive islets (Fig. 2A). The required number of activated HSCs to obtain this effect was 3 × 105 because the MST of islets plus 1 × 105 HSCs was only 18 days (Table 1, P < .05 compared with islet alone group). In addition, HSCs needed to be syngeneic to the recipient because co-transplantation with BALB/c (allogeneic) or C3H (H2k, third party) HSCs slightly, but not statistically significantly, prolonged survival of islet allografts (Table 1, both P > .05 compared with islet alone group). To rule out the possibility that this protective activity of HSCs was specific to the B6 strain, similar experiments were performed in other strains of mice including BALB/c and B10.BR (H2k). HSCs showed a similar protective effect on the co-transplanted islet allografts (data not shown).

Table 1. Effect of Co-transplantation With HSCs on Survival of BALB/c Islet Allografts in B6 Recipients
GroupHSCNumber (×105)Islet Graft Survival (days)MSTa(days)
  • a

    MST, median survival time.

  • b

    Islets (BALB/c) allografts alone were transplanted under the renal capsule of left kidney of B6 recipients.

  • c

    HSCs were cultured for 7 days, then mixed with islet allografts, and co-transplanted under the renal capsule of left kidney of B6 recipients.

  • d

    HSCs were cultured for 2 days, then mixed with islet allografts, and co-transplanted under the renal capsule of left kidney of B6 recipients.

  • e

    Euglycemia maintained for more than 60 days posttransplantation, and the animals were killed on the indicated day for further studies.

  • f

    HSCs were implanted under the renal capsule of the opposite side (right) kidney.

1Noneb9 (x2), 10, 12, 13, 15, 16 (x2), 1713
2B6c3.010, 11 (x3), 13, 15, 16 (x2), 18, 2014
3B6d1.010, 12, 16 (x2), 18 (x2), 21, 24, 2818
4B6d3.030, 35, 48 (x2), 50, 65e, 77 (x2)e, 100e, 122e, 170e>60
5BALB/cd3.012, 14, 21 (x3), 23, 24, 26, 30, 3622
6C3Hd3.013, 15, 20, 26 (x2)20
7B6df3.010, 11, 12, 13, 1812
8B6, B7-H1 KOd3.016, 17, 26, 26, 3726
Figure 2.

Co-transplantation with HSCs provides local protection of islet allografts. (A) Blood glucose levels in chemically induced diabetic B6 mice were shown before (day 0) and after transplantation of 300 BALA/c islets alone under the left renal capsule (black triangles). The islet graft destruction was indicated by recurrence of hyperglycemia on day 15 after transplantation, and very few insulin-positive islets with massive infiltrates were seen in histochemical examination (upper right). In contrast, after transplantation of 300 BALB/c islets plus 3 × 105 B6 HSCs under the left renal capsule (black squares), euglycemia was maintained until the left kidney was removed on day 65 after transplantation. Immunochemistry showed the presence of numerous insulin-positive islets under the renal capsule (lower right). (B) Co-transplantation with HSCs did not affect in vivo–specific CTL activity. BALB/c mouse splenocytes (106) labeled with low CFSE amounts (0.5 μmol/L, used as donor-specific targets) were mixed with the same number of B6 splenocytes labeled with high CFSE amounts (5 μmol/L, used as internal controls). Both CFSElow and CFSEhigh cell populationss were clearly identified by flow cytometry. The mixed cells were then injected intravenously into normal B6 mice (none) or B6 recipients that had been transplanted with islets (BALB/c) alone or islet (BALB/c) plus HCSs (B6) under the left renal capsule for 12 days. The incidence of CFSElow and CFSEhigh cells in spleen was determined by flow cytometry at 3 hours after injection. Target lysis was calculated based on the incidence of CFSElow and CFSEhigh cells as described in Methods. Cytotoxic T lymphocytes (CTL) against BALB/c targets in normal B6 mice was 62.5%. Transplantation with islet allografts alone enhanced the specific CTL activity to 82%. Co-transplantation with HSCs had no effect on this CTL activity (80.6%). (C) To determine whether co-transplanted HSCs exerted a remote effect on islet allografts transplanted in the opposite side kidney, diabetic B6 recipients received 300 BALB/c islets plus 3 × 105 B6 HSCs under the left renal capsule, and 300 BALB/s islets alone under the right renal capsule. Animals were killed on day 18 after transplantation. The graft sections were immunostained by anti-insulin monoclonal antibody (mAb) (red). All data are representative of 3 separate experiments.

Co-transplanted HSCs Do Not Confer a Systemic Immune Suppressive Effect.

To distinguish local or systemic effect, we examined the impact of co-transplanted HSCs on in vivo specific CTL activity against donor (BALB/c) targets. BALB/c CFSElow splenocytes (donor targets) were mixed with same number of B6 CFSEhigh splenocytes (control targets), and then injected into normal B6 (control) and B6 recipients that had transplanted with BALB/c islet alone or plus HCSs (B6) for 12 days. The specific lysis of donor targets was analyzed by flow cytometry 3 hours after injection. CTL against BALB/c targets in normal B6 mice was 62.5% (Fig. 2B), reflecting a level of allogeneic CTL activity in normal recipients. Transplantation with islet allografts alone increased the specific CTL to 82%, whereas co-transplantation with HSCs did not alter allogeneic CTL activity (80.6%) (Fig. 2B), suggesting lack of systemic immunosuppressive effect. This was supported by the fact that HSCs transplanted into the right kidney failed to protect islet allografts in left kidney (Table 1, group 6, P > .05 compared with group 1). To exclude the possibility that transplanted syngeneic HSCs lacking persistent stimulation by infiltrating lymphocytes are inactivated, we performed the following experiments (n = 3). HSCs were mixed with allogeneic (BALB/c) islets and transplanted into left kidney, and an additional 300 islets without HSCs were transplanted into right kidney. Removal of the right kidney containing islet grafts alone on day 18 did not change euglycemic status, but when the left kidney bearing HSC and islets was removed, hyperglycemia was promptly restored, suggesting that HSCs cannot remotely protect islet allografts in the other side kidney. Immunohistochemical analysis confirmed insulin-positive islets surrounded by HSC under the renal capsule of the left kidney, whereas very few functional islets were seen in the right kidney (Fig. 2C). To examine the status of donor-specific tolerance, skin grafts from BALB/c or C3H (third-party strain) were transplanted onto three B6 recipients whose islet allografts (BALB/c) had survived for more than 100 days. Survival of BALB/c skin grafts on B6 recipients accepting BALB/c islet grafts was not significantly prolonged compared with controls (data not shown), indicating an absence of specific systemic tolerance.

Co-transplanted HSCs Formed a Multiple-Layered Capsule Surrounding the Islets.

To determine how HSCs provided a local protective effect, the long-term surviving islet grafts were harvested for immunohistochemical examinations. Interestingly, the islet allografts were surrounded by a multilayered capsule, which stained positive for α-SMA (Fig. 3A). The anatomic relationship of the α-SMA–positive capsule and islet allografts were shown by multiple-color staining using anti–α -SMA, anti-insulin, and Hoescht (blue for nuclei). In long-term euglycemic animals, functional islet allografts (identified by positive insulin staining) were encapsuled by multiple layers of α-SMA–positive cells (Fig. 3B). To ascertain the origin of these α-SMA+ cells whether they were derived from transplanted HSCs or recruited from the host cells, BALB/c islet allografts were co-transplanted with fluorescent HSCs from EGFP mice (B6 background) (Fig. 3C) into B6 recipients. Euglycemic animals were killed on either day 21 or day 100 after transplantation. The removed kidneys with islets were examined. These showed that the fluorescent cells formed a multilayered capsule in day 21 (Fig. 3D) and day 100 animals in which healthy-appearing islets with positive insulin staining were surrounded by EGFP-positive cells (Fig. 3E), indicating that the formed capsule originated mainly from the co-transplanted HSCs.

Figure 3.

Formation of an alpha smooth muscle actin (α-SMA)–positive capsular rim surrounding islets after co-transplantation with islet plus hepatic stellate cells (HSCs). Chemically induced diabetic B6 recipients were co-transplanted with 300 BALB/c islets plus 3 × 105 B6 HSCs under left renal capsule and maintained normal blood glucose levels. (A) The animal was killed on day 122 after transplantation. The cryostat sections of the left kidney were stained with anti-SMA monoclonal antibody (mAb). A complete multiple-layered SMA-positive capsule was formed around the islets which were morphologically healthy. (B) The animal was killed on day 100 after transplantation. The cryosections of the left kidney were stained with fluorescence-labeled anti–α- SMA (green) and insulin (red). The nuclei were stained blue. (C-E) To examine the role of co-transplanted HSCs in the formation of the capsule surrounding the islets, chemically induced diabetic B6 recipients received 300 BALB/c islets plus 3 × 105 B6 HSCs from EGFP transgenic mice (B6 background) under left renal capsule. Animals were killed on days 21 and 100. (C) HSCs from EGFP transgenic mice were examined under fluorescent microscopy showing green fluorescence. (D) The cryostat sections of the grafts from an animal killed on day 21 after transplantation were examined under fluorescent microscopy showing a ringed structure of fluorescent (green) cells, suggesting formation of capsular rim by co-transplanted HSCs. (E) The cryostat sections of the grafts from an animal killed on day 100 after transplantation were stained with anti–green fluorescent protein (GFP) (green), anti-insulin (red), and Hoescht stain for nuclei (blue), and examined under fluorescent microscopy. The picture indicated the close approximation of co-transplanted HSCs with the islet allografts. The data are representative of 3 separate experiments.

Co-transplantation with HSCs Attenuates T Cell Infiltration in Islet Grafts.

The grafts from the recipients of islets allografts alone or islets allografts plus HSCs 12 days after transplantation were staining with anti-CD4 and anti-CD8 mAbs. Both CD4+ and CD8+ cells were significantly reduced in islet allografts co-transplanted with HSCs (Fig. 4A), indicating a role of HSCs in attenuating T cell infiltration. To illustrate the apoptotic activity, nephrectomy specimens bearing the HSC/islet grafts were examined by TUNEL staining. The apoptotic mononuclear cells were significantly increased in HSCs co-transplanted grafts (Fig. 4B), suggesting that induction of apoptotic death of infiltrating cells may be one of the underlying mechanisms maintaining local tolerance. The anatomic relationship of infiltrated T cells and co-transplanted HSCs was demonstrated by the sequential sections and double staining using anti-insulin (green) and anti–α -SMA or anti-CD3 (red). α-SMA–positive cells formed a capsule around the islets by day 12 and were associated with lighter graft infiltration of CD3+ cells. Most of the CD3+ cells appeared to be trapped within these capsules (Fig. 4C).

Figure 4.

Acceptance of islet allografts co-transplanted with HSCs is associated with decreased T cell infiltration and enhanced apoptotic activity. Chemically induced diabetic B6 recipients were transplanted with 300 BALB/c islets alone or co-transplanted with same number BALB/c islets plus 3 × 105 B6 HSCs under the left renal capsule (n = 3 in each group). Animals were killed on day 12 after transplantation. (A and B) The cryostat sections of the grafts were stained with biotinylated rat anti-CD4 or anti-CD8 monoclonal antibody (mAb). The apoptotic cells were identified by terminal deoxynucleotidyl transferase-mediated nick-end labeling (TUNEL) staining. The positive-staining cells were counted by microscopy. A total of 30 high-power fields (HPFs) were randomly selected in each group. Data are expressed as mean positive cells per HPF ± SD. (C) Graft sections were stained with fluorescent anti-insulin (green) and anti–alpha smooth muscle actin (αSMA) (red, left panels) or anti-CD3 (red, right panels). The nuclei were stained with Hoescht stain (blue). In contrast to islet graft alone, which showed heavy infiltration of CD3+ T cells in the grafts, co-transplantation with HSCs demonstrated formation of multiple layers of αSMA cells surrounding the islets with much less infiltration of CD3+ cells. These CD3+ cells were trapped in the capsular layers composed of αSMA-positive cells.

Effect of Co-Transplantation with HSCs on Islet Revascularization.

Other than overcoming of immune rejection, rapid and adequate islet revascularization is another crucial factor for the survival and function of islet grafts. We next evaluated the effect of co-transplanted HSC on islet revascularization. Under double staining with anti-CD31 (red) and anti-insulin (green), the islet grafts that were co-transplanted with HSCs exhibited more CD31+ cells compared with the islets alone group either at day 5 or day 14 after transplantation (Fig. 5A). CD31+ cells were largely located at the edge of islet grafts in the islets alone group, whereas, with co-transplantation of HSCs, many CD31+ cells were seen within the islet grafts (Fig. 5A). Quantitative data showed that the CD31 staining intensity in islet alone group was significantly lower than in the co-transplant group (Fig. 5B, P < .05 on day 5, P < .01 on day 14). As expected, the islets transplanted with HSCs displayed significantly higher insulin content than that islets alone (Fig. 5C, P < .05). To determine whether the enhancement of CD31 expression was simply attributable to survival of islets or the presence of co-transplanted HSCs, syngeneic islets without HSCs were transplanted. The data shown in Fig. 5B showed that the intensity of CD31 in syngeneic islets was comparable to allo-islet plus HSCs (P > .05 on both day 5 and day 14), suggesting that the revascularization is related to survival of islet grafts, and the presence of HSCs per se does not further enhance it.

Figure 5.

Co-transplantation with HSCs enhances survival of islet allograft, but not vascularization. The islet grafts were retrieved from B6 recipient diabetic mice 5 or 14 days after transplantation of BALB/c (allogeneic) 300 islets alone or mixed with 3 × 105 B6 HSCs. Transplants of B6 islets alone were used as syngeneic controls (n = 3 in each group). The sections were stained with anti-CD31 (red) and anti-insulin (green). (A) Histochemistry of islet grafts (14 days after transplantation). (B) Graft anti-CD31 immunostaining intensity is shown. The relative intensity of immunostaining with anti-CD31 monoclonal antibodies (mAb) within islet grafts was determined by computerized morphometric analysis. Anti-CD31 staining intensity in the allo-islet plus HSCs group was significantly higher than in the group with allografts alone on either day 5 or day 14 (both P < .05), but was similar to that in the group with syngeneic islets alone (P > .05). (C) Graft insulin content. The intensity of immunostaining with anti-insulin mAb within islet grafts was compared. The kidney transplanted with allo-islets plus HSCs displayed significantly higher insulin content than that transplanted with allo-islets alone (P < .05).

Involvement of B7-H1 in Immune Regulation Activity of HSCs.

We have previously shown that quiescent HSCs (cultured for 2 days) expressed low levels of B7-H1, whereas expression of B7-H1 was markedly up-regulated after in vitro activation by activated T cells or interferon gamma (IFN-γ).11 To define the role of B7-H1, we used B6. B7-H1 KO mice this study. Deficiency of B7-H1 expression did not alter HSC proliferation (cell yield number) or expression of α-SMA in culture compared with HSCs isolated from normal B6 mice (data not shown). In a 4-hour 51Cr release assay, the presence of wild-type HSCs markedly suppressed generation of specific CTL activity against allogeneic targets, which was almost totally reversed by using B7-H1−/− HSCs, or when function of B7-H1 was blocked by anti–B7-H1 mAb (Fig. 6). This was clearly demonstrated in vivo when co-transplantation of HSCs from B7-H1−/− mice largely lost their capacity to protect islet allografts from rejection (Table 1, P < .05, group 8 versus 3; P > .05, group 8 versus 1). Immunochemical results showed that islet allografts were rejected, but the co-transplanted B7-H1−/− HSCs (syngeneic to recipients) survived when the animal became diabetic (Fig. 6B).

Figure 6.

The role of B7-H1 expression in the immune inhibitory effect of HSCs. (A) Effect on inhibition of CTL activity. B6 (H2b) spleen T cells (2 × 105) cultured with γ-irradiated (20 Gy) BALB/c (H2d) allogeneic dendritic cells (DCs) at a ratio of 10:1 for 5 to 6 days were used as effectors. To determine the effect of HSCs on T cell responses, γ-irradiated (50 Gy) HSCs from wild-type B6 (white squares), wild-type B6 with anti–B7-H1 blocking monoclonal antibody (mAb) (15 μg/mL, rat IgG2a, eBioscience, San Diego, CA) (black ×) or isotype Ab (white circles), or B6.B7-H1 knockouts (white triangles) were added at HSC:T ratio of 1:40 at the beginning of the culture. P815 (H2d), EL4 (H2b), R1.1 (H2k), or lymphoma cells labeled with 100 μCi Na251CrO4 (NEN, Boston, MA) were used as donor-specific, syngeneic, or third-party targets, respectively. They were plated in 96-well round-bottom plates and cultured with effectors at E:T ratios of 40:1 and 20:1 in a total volume of 200 μL/well. CTL activity was examined in a 4-hour 51Cr release assay. No cytotoxicity was generated against syngeneic EL4 (H2b) or third-party R1.1 (H2k) targets (data not shown). (B) Immunochemistry. Sections from islet allografts (B6) co-transplanted with B7H1−/− (became diabetic at day 26) or wild-type HSCs (nondiabetic at day 65) were stained with anti-insulin (green) and anti–α-SMA (red). The nuclei were counterstained with Hoechst stain (blue).

Discussion

We demonstrated that activated HSCs suppress T cell thymidin uptake stimulated by allogeneic DC in vitro. The data in this study provides evidence of immune regulatory functions of HSCs in vivo. More than half (55%)of diabetic mice achieved long-term (>60 days) normoglycemia with transplantation of composite islet and activated HSC cellular grafts, whereas none of islet-only recipients remained normoglycemia more than 17 days. Speculating that HSCs may play an important role in modulating immune responses in the liver, an immune-privileged organ that was initially recognized by spontaneous acceptance of liver allografts.4. The mechanism of HSC immune function within the liver may differ from that seen in the testis, another site considered to provide immune privilege. Islet allografts survive without immunosuppression when transplanted into allogeneic testis.18, 19 Co-transplantation of testicular Sertoli cells also protected islet grafts from rejection in a similar manner to HSCs19; however, the protective effect of HSCs appears to be more potent than that of Sertoli cells because the required cell number of HSCs (3 × 105) was approximately one-tenth of Sertoli cells (4 × 106 to 11 × 106).20, 21 Differences are also manifest by local versus systemic effects. HSCs implanted distantly from islet allografts did not show protection. However, Sertoli cells protected islets transplanted into the contralateral kidney.22 Moreover, in our model, co-transplanted HSCs needed to be syngeneic (to the host) (Table 1), whereas allogeneic Sertoli cells prolonged survival of islet allografts,23 suggesting involvement of different mechanisms. Sertoli cells may suppress the immune response through Fas or transforming growth factor beta (TGF-β) pathways,22 by which the allogeneic Sertoli cells escape from host immune attack, whereas allogeneic HSCs may not be able to escape host immune attack. The results shown in Table 1 also indicate that, in contrast to activated HSCs (cultured for 7 days), co-transplantation with quiescent HSCs (cultured for 2 days) showed no protective effect on survival of islet allografts. This is not surprising because quiescent HSCs express very few immune regulatory molecules, including B7-H1, TGF-β, and IL-10, as reported in our previous study.11

Although the exact mechanisms are not completely understood, our data showed that formation of capsules surrounding the islet allografts were composed of multiple layers of α-SMA–positive cells, and were associated with a significant reduction of graft T cell infiltration. These α-SMA–positive cells showed the origin of co-transplanted HSCs by using of EGFP mice. The reduction of infiltrating T cell number was not likely to have resulted from a lack of chemotactic factors because activated HSCs produce numerous chemokines and attract leukocytes24, 25; rather, it was a result of enhanced apoptosis activity in infiltrated mononuclear cells (Fig. 4B). These findings are consistent with our previous observations that HSCs induced apoptotic death of activated T cells in vitro.11 Formation of the protective capsules by co-transplanted HSCs may also explain why HSCs more effectively protect islet allografts than Sertoli cells, which do not form capsules. HSCs are well known to participate in the process of fibrogenesis occurring in the injured livers that distorts the normal hepatic architecture.26, 27 Treatment of liver fibrosis has aimed to inhibit the accumulation of activated HSCs.26 However, the data of this study suggest that liver fibrogenesis participated in by HSCs may reflect a protective mechanism to prevent hepatocytes from overwhelming immunological injury. Therefore, the strategy of anti-fibrotic therapy needs to be cautiously re-evaluated.

Our data also suggest an important role of induction of B7-H1 expression in allowing HSCs to mediate an immune regulatory function. Quiescent HSCs do not express B7-H1, which is markedly up-regulated after activation by various stimulators including IFN-γ and activated T cells.11 This is consistent with other reports; B7-H1 is inducible in a variety of organs, including nonlymphoid tissues.28 HSCs deficient in B7-H1 lost the suppressive activity in inhibiting generation of specific CTL activity in vitro (Fig. 6) and failed to prolong survival of co-transplanted islet allografts (Table 1). This may reflect the different effect of HSCs on naïve and activated T cells. In contrast to their inhibitory effect on activated T cells, HSCs demonstrate modest stimulatory activity when cultured with naïve allogeneic T cells (Qian et al., unpublished data). This has also been shown in human HSCs.29 B7-H1 ligation may contribute to inducing hyporesponsiveness of activated T cells through promoting their apoptotic death because the B7-H1 receptor PD-1 is inducibly expressed on activated T cells.30 B7-H1 has been shown to negatively regulate the immune system via downregulation at the effector phases. Ligation of PD-1, the receptor of B7-H1, and TCR leads to rapid phosphorylation of SHP-2, a phosphatase that attenuates TCR signaling, which negatively regulates cytokine synthesis.28, 31–33 A recent study demonstrating that the deficiency of B7-H1 results in accumulation of CD8+ T cells in the liver suggested a role of B7-H1 in regulating T cell homeostasis.34 However, B7-H1 may not be the only molecule involved in the immunosppressive activity of HSCs, because blockage of B7-H1 ligation by addition of anti–B7-H1 mAb can only partially reverse the inhibition of T cell proliferation mediated by HSC in an MLR assay. In keeping with this hypothesis, anti–B7-H1 mAb only partially reverses T cell apoptosis triggered by HSC.11 Therefore, other pathways are expected to be involved. Activated HSC also produce IL-10 and TGF-β, cytokines with inhibitory and regulatory property in immune responses, which may provide synergizing signals with B7-H1 in inhibition of immune responses.

Whereas organ transplantation has been successful for decades, the outcome of cell transplantation remains largely disappointing. This is the case in animal models. Thus, liver allografts are spontaneously accepted in many species,1–4 whereas hepatocyte allografts are promptly destroyed, succumbing to immune-mediated destruction because syngeneic hepatocytes survive long-term.35, 36 These observations suggest that the presence of organ NPC may attenuate immune attacks. Lack of such mesenchymal derived organ-based cell populations may contribute to poor outcomes of cell transplants. This was hinted at in previous clinical islet transplant studies. A significant positive correlation of islet allograft survival was observed when the number of ductal–epithelial cells transplanted was greater in the purified islet preparation.37 This was not attributable to the ductal cells containing islet progenitor cells, because β-cell renewal in adults does not originate from islet progenitors.38

To avoid immune attacks to islet transplants, some investigators advocate using an “immunoisolation” approach by encapsulation using synthetic semipermeable membranes.39 They are however, not completely biocompatible39 and also result in insufficient oxygen and nutrient transport to the islets.40, 41 Our results may provide a method to develop biological encapsulation, using cells that are syngeneic and cause no biocompatibility problems. A better understanding of the paradigm of local immunomodulation provided by HSCs may lead to development of novel strategies for cell transplantation and treatment of liver diseases.

Ancillary