Activated Protein C Preserves Functional Islet Mass After Intraportal Transplantation: A Novel Link Between Endothelial Cell Activation, Thrombosis, Inflammation, and Islet Cell Death

Male C57BL/6 mice (The Jackson Laboratories, Bar Harbor, ME) were used as islet donors and recipients. Surgical and nonsurgical procedures were carried out in accordance with the National Institutes of Health Guide for the Care and Use of Animals, under supervision of the Institutional Animal Care and Use Committee. Four days before islet transplantation, recipients were rendered diabetic by a single intraperitoneal injection of freshly reconstituted streptozotocin (Sigma, St. Louis, MO) at 200 mg/kg i.p. Blood glucose levels were obtained after tail snipping using a portable glucose meter (Roche AccuCheck III; Roche Diagnostics, Indianapolis, IN). Diabetes was defined as nonfasting glucose levels >250 mg/dl for ≥3 consecutive days.

Pancreatic islets were isolated and purified following standard techniques by intraductal injection of rodent Liberase (Roche Diagnostics) (30). The number of islets within each size class was converted to the standard number of islets of 150-μm diameter equal in volume to the sample (islet equivalents [IEQs]). Purity was assessed by comparing the relative quantity of dithizone-stained endocrine tissue with unstained exocrine tissue. Islet viability was assessed following staining with acridine orange (100 μg/ml) and ethidium bromide (100 μg/ml) and subjected to two-color fluorescence microscopy. Only islet isolations with >90% viability and >90% purity were used. After isolation, islets were cultured for 24 h in RPMI 1640 (Sigma) supplemented with l-glutamine (Sigma), 100 units/ml of penicillin, 100 mg/ml of streptomycin (Life Technologies, Grand Island, NY), and 10% FCS (Sigma). Then, islets were recounted and injected at different doses into the portal vein as described (30).

Diabetic animals were randomly divided into two groups (experimental and control). Each animal in the experimental group received purified recombinant mouse APC (mAPC) at 0.2 mg/kg i.v. (15) 1 h before islet transplantation. This dose was similar to that in previous studies (22,29) that had revealed beneficial effects of mAPC on stroke and ischemia/reperfusion injury in mice. The control group received vehicle only. Within each group, some animals received an optimal islet dose (500 IEQ/mouse) or suboptimal doses (250 or 150 IEQ). As indicated, additional animals received media instead of islets and were also used as controls.

Blood samples were obtained from the tail vein at indicated experimental time points after transplantation into glass tubes containing heparin, placed on ice, and centrifuged within 10 min. Plasma was separated, immediately frozen, and stored at −70°C until analysis. Proinsulin levels were determined by enzyme-linked immunosorbent assay (ELISA) using a monoclonal antibody against mouse proinsulin (Clone D3E7; Research Diagnostics, Flanders NJ) as described (12). Proinsulin levels were used as an established marker of islet injury (12). Mouse soluble ICAM (sICAM)-1, soluble E-selectin, and soluble P-selectin were assessed in plasma samples collected 12 h after transplantation by ELISA (Quantikine; R & D Systems, Minneapolis, MN) following the manufacturer’s instructions. This time point was chosen because it corresponds to the peak of sICAM-1, soluble E-selectin, and soluble P-selectin in our model of PIT (data not shown).

Liver samples were obtained 6 h after transplantation, snap frozen in liquid nitrogen, and stored at −70°C until use. Total mRNA was prepared using the RNAeasy 96 Total RNA Isolation Kit (Qiagen, Valencia, CA) according to the manufacturer’s protocol. RNA concentration was estimated from absorbance at 260 nm. The expression level of specific mRNAs was quantitated (Quantikine Mouse mRNA; R&D Systems) following the manufacturer’s instructions. One attomole of mouse IL-1β mRNA is ∼430 fmg, and 1 amol of TNF-α mRNA is ∼230 fmg.

Fibrin deposition was quantified in the whole liver by Western blotting with antibodies against fibrin II (NYBT2G1, dilution 1:500; Accurate Chemical & Scientific, Westbury, NY) as previously described (22,29). Briefly, equal amounts of protein (10 μg/lane), measured by Bradford assay (Bio-Rad Laboratories, Hercules, CA), were electrophoresed on 10% SDS-polyacrylamide gels, transferred onto nitrocellulose membranes, and probed with the described antibody. The relative abundance of each protein was determined by scanning densitometry using β-actin as an internal control. Fibrin abundance was expressed relative to intrahepatic fibrin concentration assessed in control diabetic mice, whose value was arbitrarily assigned as 1. Experiments demonstrated that 6 h was required for the peak of intrahepatic fibrin deposition after syngeneic PIT (data not shown); therefore, this time-point was selected for the experiments.

Portal vein endothelial cells (PVECs) were isolated and characterized as described previously (19). Briefly, rings of the portal vein (C57BL/6 mice) were placed endothelial side down on Petri dishes and incubated overnight in a humidified incubator at 37°C in a 95% air/5% CO₂ atmosphere. Dulbecco’s modified Eagle’s medium (Life Technologies) supplemented with l-glutamine, 25 mmol/l HEPES, 10% fetal bovine serum, 100 units/ml of penicillin, and 100 mg/ml of streptomycin was used to keep the tissue moist. One day later, 1 ml of media was added per 10 cm² of tissue culture dish. Culture continued for 5–7 days until PVECs migrated onto the culture dish. Then, the rings were removed. Immunofluorescence staining of factor VIII antigen was used to identify PVECs (19). Following these procedures, it was possible to obtain PVECs with purity >85% after 6–10 days in culture (data not shown).

Hand-picked islets (C57BL/6 mice) were cocultured with PVECs at 1 IEQ/100 PVECs (19) in the presence of mAPC (50 nmol/l) or vehicle, as described (22). This ratio is comparable with the dose of islet cells normally infused into the portal vein during PIT. After 6 h in culture, islets were removed, and nuclear extracts were immediately prepared from trypsinized PVECs using a nuclear extraction kit (Transfactor Extraction Kit; BD Biosciences Clontech Laboratories, Palo Alto, CA) following the manufacturer’s instructions. DNA binding of NF-κB p50 was assessed in nuclear extracts using an ELISA-based format (BD Mercury NF-κB p50, Transfactor Kit; BD Biosciences Clontech Laboratories) following the manufacturer’s instructions. This time point was selected based on preliminary experiments that had shown maximal activity (data not shown). In addition, TNF-α was determined from tissue culture supernatants 24 h after the initial exposure of PVECs to isolated islets by ELISA (Quantikine Mouse TNF-α; R&D Systems) following the manufacturer’s instructions.

Livers from islet transplant recipients were processed 6 h after the transplant for histological evaluation of fibrin and leukocyte deposition surrounding the islets as described (30). In addition, liver samples were assessed for the presence of insulin 10 days after the transplant. Briefly, tissues were fixed in paraformaldehyde or snap frozen in liquid nitrogen. Five-micron-thick sections were mounted on glass slides. Sections were stained with horseradish peroxidase–conjugated anti-fibrin II antibody (NYB-T2G1), anti-CD11b (Dako, Glostrup, Denmark), and anti-insulin (Anti-insulin antibody diluted in Tris-buffered saline containing 0.1% BSA; ICN Biomedicals, Aurora OH) as described (30). The slides were counterstained with hematoxylin and eosin.

Serum hyaluronic acid clearance via endothelial cell receptor CD44 is a sensitive assay to evaluate the function of endothelial cells of the hepatic sinusoidal space after PIT (14,15). Hyaluronic acid (300 μg/kg i.v.; Sigma) was given 6 h after PIT. Blood samples were collected in heparinized tubes from the tail vein before and at 5, 30, 180, 540, 1,440, and 2,880 min after hyaluronic acid administration. Plasma was separated, immediately frozen, and stored at −70°C until analysis. Serum hyaluronic acid levels were determined using a commercially available ELISA (Corgenix, Westminster, CO).

Nonfasting glucose levels were assessed as described above. In addition, recipients were subjected to metabolic testing 10 days after PIT. Glucose disposal rates (K_g) were calculated from intraperitoneal glucose tolerance testing (16-h fast recipients, 1 g glucose/kg body wt i.p.) as previously reported (31). Arginine-stimulated acute insulin release (AIR_arginine) was performed in overnight-fast recipients after intravenous injection of arginine (0.3 g/kg body wt) infused in 30 s. Thirty microliters of whole blood were collected in 3 μl heparin (5,000 units/ml) at −5, 0, 2, 5, and 7 min after stimulus injection. Samples were immediately centrifuged, and supernatants were collected and stored at −20°C. Mouse insulin was determined by ELISA (Kamiyama Biomedical, Seattle, WA). In addition, liver was homogenized in acid/ethanol (0.1 N HCl in 100% ethanol) and stored overnight at −80°C. The homogenates were centrifuged at 14,000g for 7 min, and the amount of insulin was measured in the supernatants by ELISA as described above.

A group of islet recipients was killed 24 h after PIT. The liver was harvested, fixed in Bouin’s solution, embedded in paraffin, sectioned (4-μm thick), and immunostained following deparaffinization and rehydration. Islet cell death was detected by an enzymatic in situ terminal TdT-mediated dUTP nick-end labeling assay (TUNEL) detection kit (Roche Molecular Biochemicals) according to the manufacturer’s instructions. Staining was achieved using 3,3′-diaminobenzidinetetrahydrochloride (Sigma). Subsequent to TUNEL, sections were stained with polyclonal anti-insulin antibody (diluted 1:150; Dako, Carpinteria, CA), developed with a broad spectrum Histostatin kit (Zymed Laboratories, San Francisco, CA). A minimum of 15 islets was examined per animal (n = 5), and an apoptotic index (percentage of TUNEL-positive cells per islet) was determined (32).

Islet recipients were killed 30 days after the transplant, and the total insulin content was evaluated as described (32) Briefly, the whole liver was extracted and washed in ice-cold PBS, minced, passed through a 21-gauge needle, resuspended in 5 ml of acid/ethanol (0.18m HCl in 70% ethanol), sonicated for 30 s, and extracted at −20°C. Tubes were centrifuged at 3,000 rpm for 10 min at 4°C, and the supernatant was stored at −20°C until insulin determination. After neutralization, the insulin content in the extracts was measured by ELISA as described above.

All data are expressed as means ± SE. Comparison between groups was performed by Student’s t test using SPSS 10.0 for Windows (ACITS; The University of Texas at Austin, Austin TX). Statistical significance was established at P < 0.05.

A series of islet isografts were performed to evaluate the effects of mAPC on PIT. As demonstrated in Fig. 1A, diabetes was reverted within 3 days in 100% of the recipients that were transplanted with 500 IEQ (optimal islet dose) and treated with either vehicle or mAPC. However, lower nonfasting glucose levels were demonstrated early after the transplant in animals given mAPC compared with controls. These differences were more pronounced in recipients given a suboptimal islet dose (Figs. 1B–D). Only 25% of the recipients given vehicle and 250 IEQ presented nonfasting glucose <200 mg/dl after the transplant, compared with 75% of animals given mAPC in combination with the same islet dose (P < 0.05). Moreover, none of the recipients treated with vehicle, and 150 IEQ presented euglycemia compared with 37.5% of mAPC-treated recipients (P < 0.05). Interestingly, although all diabetic recipients transplanted with an optimal dose achieved excellent glucose control, higher glucose disposal rates were demonstrated 10 days after the transplant in animals given mAPC (3.2 ± 0.4% glucose · mg⁻¹ · dl⁻¹ · min⁻¹) compared with vehicle-treated controls (2.3 ± 0.3% glucose · mg⁻¹ · dl⁻¹ · min⁻¹, P < 0.05) (Table 1). To further evaluate the islet grafts, FIM was indirectly evaluated 10 days after the transplant by AIR_arginine (Fig. 2). Significantly higher AIR_arginine was demonstrated in all recipients given mAPC compared with vehicle after transplantation independently of the islet dose infused. Moreover, higher total insulin content assessed 30 days after the transplant was demonstrated in livers harvested from animals given mAPC (500 IEQ and insulin, 377 ± 88 ng) versus animals given vehicle (172 ± 32 ng, P < 0.05) (data not shown). Proinsulin levels, an established marker of islet injury after islet exposure to blood (12), were significantly reduced in animals given mAPC compared with vehicle-treated controls (Fig. 3A). In correlation with proinsulin levels, animals treated with mAPC presented lower islet apoptosis assessed 24 h posttransplantation compared with vehicle-treated controls (Figs. 3B and C). Taken together, these results demonstrate that mAPC significantly decreases loss of FIM after transplantation.

Intrahepatic fibrin deposition was measured in recipients given 500 IEQ (Fig. 4). Minimal levels of fibrin deposition were demonstrated in control recipients given media without islets. A significant 20-fold increase in fibrin deposition was shown in vehicle-treated recipients that received 500 IEQ, and fibrin deposition was markedly reduced in animals given mAPC. Intrahepatic immunohistochemical analysis performed posttransplantation demonstrated that, in vehicle-treated recipients, a capsule of fibrin infiltrated by leukocytes and monocytes surrounding the islets, and these effects were abrogated by the administration of mAPC (Fig. 5). In accordance with the results obtained after the evaluation of total insulin content by biochemical methods, we found a higher islet insulin content in mAPC-treated animals as demonstrated by immunohistochemistry (Figs. 5E and F). These results demonstrate that activation of the coagulation cascade, leukocyte infiltration, and fibrin deposition occur after intraportal infusion of syngeneic islets and that these effects are abrogated by the administration of mAPC.

Previous studies (14,16–18) demonstrated nonspecific activation of PIC after intraportal PIT, an effect that cannot be prevented using conventional immunosuppressive drugs. PICs have profound effects on islet dysfunction and death in diabetes and in PIT (33). Because mAPC inhibits PIC production in different clinical and experimental conditions (20,25,27–29,34), we evaluated the possibility that mAPC would prevent the loss of FIM by inhibiting PIC production. Preliminary experiments demonstrated that the peak of maximal liver mRNA TNF-α and IL-1β expression after intraportal infusion of islets occurred at 6 h (data not shown). Therefore, we chose this time point for the analysis of the effect of mAPC on PIC expression (Figs. 6A and B). No difference in liver mRNA TNF-α and IL-1β expression was observed between normal animals compared with diabetic animals that received an intraportal infusion of media without islets. In contrast, a very large increase in PIC mRNA expression was demonstrated after PIT (500 IEQ/mouse), and this effect was significantly reduced by the administration of mAPC (P < 0.001). These results demonstrate potent anti-inflammatory effects of mAPC given shortly before the islet infusion.

Nonspecific activation and dysfunction of intrahepatic endothelial cells represents a key pathophysiological component in the early loss of islet grafts (14,15,19). Based on the importance of mAPC on endothelial cell physiology (25,35), we evaluated the possibility that mAPC might directly modulate intrahepatic endothelial function after intraportal PIT. Soluble forms of ICAM-1, E-selectin, and P-selectin arise in plasma from proteolytic cleavage from the cell membrane (36) in reactions following cell activation (37–39). When plasma sICAM-1, soluble E-selectin, and soluble P-selectin were evaluated serially after syngeneic intraportal infusion of an optimal islet dose, we found peak plasma levels at 6 h (data not shown). As demonstrated in Figs. 7A–C, no significant increases in plasma sICAM-1, soluble E-selectin, and soluble P-selectin were found between normal animals and control diabetic animals given media without islets. In contrast, four- to fivefold increases in sICAM-1, soluble E-selectin, and soluble P-selectin were shown after PIT in animals treated with vehicle. Strikingly, these increases were almost completely abolished by the administration of mAPC (P < 0.001). To evaluate the role of mAPC on endothelial cell dysfunction after intraportal PIT (14,15,40), we evaluated hyaluronic acid clearance in syngeneic islet recipients given 500 IEQ. Hyaluronic acid clearance, indicating intrahepatic endothelial dysfunction, was delayed in diabetic recipients treated with vehicle (30 min = 288.2 ± 39 mg/dl) compared with mAPC-treated recipients (140.4 ± 31 mg/dl, P < 0.05. Overall, these results demonstrate that mAPC decreased intrahepatic endothelial activation and dysfunction after intraportal PIT.

We evaluated the effects of mAPC on activation of PVECs elicited by coculture with islets in vitro as reflected by the nuclear activity of NF-κB, a key transcription factor involved in endothelial cell activation (35). Within 3 h after PVEC incubation with syngeneic islets, we observed significant increase in PVEC NF-κB nuclear translocation (data not shown), with maximal activity at 6 h (Fig. 8A). mAPC caused a 67% reduction in nuclear activity of NF-κB in the cocultures of treated PVECs and islets compared with vehicle-treated controls (P < 0.001). mAPC also caused a significant reduction in TNF-α production in culture supernatants evaluated 24 h after PVEC exposure to islets (Fig. 8B). Overall, these results demonstrate that mAPC decreased nuclear translocation of NF-κB and, consequently, activation of PVECs elicited by pancreatic islets.

Thrombosis and inflammation are linked in many clinical conditions such as sepsis, atherosclerosis, and as recently described, PIT (7,10,20,34,41). Mechanisms for communication between coagulation and inflammation pathways have been uncovered, including reactions in which proinflammatory mediators may regulate coagulation activation and products of the clotting cascade may affect inflammation (34,42–44). Intrahepatic infusion of isolated islets is the preferred site for clinical PIT. However, instant coagulation/inflammation reaction (IBMIR) and activation of endothelial cells occur regularly during this procedure, even without clinical signs of intraportal thrombosis (7,14,15,19). The final outcome of these events is islet injury and, therefore, reduction in FIM. In the present study, we demonstrated that mAPC is a potent inhibitor of IBMIR and intrahepatic activation/dysfunction of endothelial cells elicited by intraportal infusion of isolated islets in mice. Notably, diabetic recipients that received an optimal or suboptimal islet dose presented higher FIM when pretreated with mAPC, as demonstrated by better glucose control, higher capacity to dispose glucose, and reduced proinsulin levels assessed early after the transplant. In contrast to other antithrombotic agents (e.g., heparin or tissue plasminogen activator) that may predispose to bleeding, clinical and experimental animal studies report that APC generally does not cause bleeding, with the possible exception of severe sepsis patients (20,22,29,41,45). Consistent with the well-known anticoagulation effects of APC (20), we demonstrated a significant reduction in intrahepatic fibrin deposition after PIT in diabetic mice given mAPC. More specifically, we demonstrated significant reduction in fibrin and infiltration of leukocytes surrounding the intraportally infused islets.

Islets constitutively express tissue factor (7), which, in combination with factor VIIa, is the major activator of the extrinsic coagulation system, leading to thrombin formation and generation of fibrin clots (8,9). Thrombin also potentates a prothrombic state when it activates the thrombin-activated fibrinolysis inhibitor. At high concentrations, thrombin may be a potent proinflammatory and proapoptotic mediator (46,47). Accordingly, thrombin inhibition abrogates IBMIR during islet exposure to blood (48). Consistent with these observations, mAPC administration downregulates the expression of tissue factor, decreases thrombin generation, and promotes fibrinolysis (20,48,49). In addition to islet-borne tissue factor, coagulation in PIT can occur indirectly through the release of cytokines that induce coagulation, such as TNF-α (50), which inherently arises from the islet transplant procedure (14,18,51).

It is becoming increasingly clear that anticoagulants can limit coagulation-induced increases in the inflammatory response by mechanisms that involve not only inhibition of coagulation proteases but also direct interactions with cells that generate inflammatory substances following cell activation (34,50). Herein, in addition to decreased fibrin deposition, we demonstrated a significant reduction in liver PIC expression after intraportal PIT in animals treated with mAPC. Novel mechanisms contributing to anti-inflammatory effects of mAPC involve both direct and indirect effects (20,34,50). Indirect effects may be based on downregulation of coagulation pathways and consequent reduction in generation of thrombin and factor Xa, which can each, in turn, activate protease-activated receptors on cells to generate PICs and adhesion molecules. The direct effects may include downregulation of expression of various NF-κB–dependent PIC genes (20,25,27,34,50). Consistent with these concepts, we demonstrated significant downregulation of NF-κB nuclear translocation in mAPC-treated PVECs cocultured with islets and found decreased PVEC activation, demonstrated by reduction in TNF-α production in culture supernatants. These concepts were supported by in vivo studies showing that plasma levels of sICAM-1, soluble E-selectin, and soluble P-selectin, as well as intrahepatic endothelial cell function (evaluated by hyaluronic acid clearance) were improved by mAPC treatment. Overall, the results show that intrahepatic endothelial dysfunction elicited by embolism of isolated islets infused into the portal vein was significantly reduced in recipients treated with mAPC compared with controls. Therefore, mAPC treatment prevents activation and dysfunction of endothelial cells elicited by intraportal infusion of islets and represents a novel approach to preserve FIM after transplantation.

When islet grafts are implanted into the liver, cellular components other than endothelial cells serve as critical mediators of nonspecific inflammatory events, including resident macrophages (Kupffer cells). Previous studies (14,15) demonstrated that prevention of macrophage activation and the consequent release of inflammatory mediators preserve islet mass after transplantation, an effect also potentially beneficial from the immunological point of view because immunogenicity of the transplanted islets is not amplified by the presence of inflammatory cytokines or by efficient antigen presentation by macrophages. The tissue factor/factor VIIa complex induces proinflammatory effects in macrophages (52), whereas binding of APC to specific receptors on mononuclear phagocytes blocks responses of macrophages induced by lipopolysaccharide and PICs, including monocyte-dependent T-cell proliferation (28). Studies are underway in our laboratory to evaluate the hypothesis that APC protective effects also involve direct effects on Kupffer cells during PIT.

The beneficial effects of mAPC in our islet transplantation model could also be related to its antiapoptotic activity (22–26). In the present study, we demonstrated a significant decrease in intraislet apoptotic cells after the transplant in animals given mAPC compared with controls. Moreover, studies in our laboratory suggest that recombinant human APC induces significant protection against PIC-mediated human isolated islet cell death (53). Modulation of several genes in the islet apoptosis pathway, including the Bcl-2 homolog protein and the inhibitor of apoptosis protein, seems to be critical in this effect (J.L.C., C.E., C.A.S., G.B., M.V., S.E.R., C.Y., J.A.T., J.A.F., J.H.G., D.E.E., unpublished data).

In summary, we have demonstrated that mAPC can significantly reduce loss of FIM after intraportal islet transplantation in diabetic mice by downregulation of coagulation, inflammation, and probably apoptotic pathways. The protective effects of APC are associated with decreased activation of intrahepatic endothelial cells elicited by the intraportal embolism of isolated islets. These results suggest that APC therapy may enhance the therapeutic efficacy of islet transplantation in humans.

This work was supported in part by National Institutes of Health Grants R37HL52246 (to J.H.G.) and RO3DK64827-2 (to J.L.C.), an Fujisawa/American Society of Transplant Surgeons Faculty Development Award (to J.L.C.), and the Diabetes Trust Fund (to J.L.C.).

We gratefully acknowledge the excellent technical assistance of Xiao Xu for preparation of the mAPC.