P-selectin suppresses hepatic inflammation and fibrosis in mice by regulating interferon γ and the IL-13 decoy receptor

Authors

  • Thomas A. Wynn,

    Corresponding author
    1. National Institute of Allergy and Infectious Diseases, National Institutes of Health, Bethesda, MD
    • Immunopathogenesis Section, Laboratory of Parasitic Diseases, National Institute of Allergy and Infectious Diseases, National Institutes of Health, Bldg. 50, Rm. 6154, MSC 8003, Bethesda, MD 20892
    Search for more papers by this author
    • fax: 301-480-5025

  • Matthias Hesse,

    1. National Institute of Allergy and Infectious Diseases, National Institutes of Health, Bethesda, MD
    Current affiliation:
    1. College of Veterinary Medicine, Cornell University, Ithaca, NY
    Search for more papers by this author
  • Netanya G. Sandler,

    1. National Institute of Allergy and Infectious Diseases, National Institutes of Health, Bethesda, MD
    Search for more papers by this author
    • N.G.S. was a Howard Hughes Medical Institute-National Institutes of Health Research Scholar.

  • Mallika Kaviratne,

    1. National Institute of Allergy and Infectious Diseases, National Institutes of Health, Bethesda, MD
    Search for more papers by this author
  • Karl F. Hoffmann,

    1. National Institute of Allergy and Infectious Diseases, National Institutes of Health, Bethesda, MD
    Current affiliation:
    1. Department of Pathology, Microbiology and Parasitology Unit, University of Cambridge, Cambridge, UK
    Search for more papers by this author
  • Monica G. Chiaramonte,

    1. National Institute of Allergy and Infectious Diseases, National Institutes of Health, Bethesda, MD
    Current affiliation:
    1. Division of Immunobiology, The Children's Hospital Research Foundation, Cincinnati, OH
    Search for more papers by this author
  • Rachael Reiman,

    1. National Institute of Allergy and Infectious Diseases, National Institutes of Health, Bethesda, MD
    Search for more papers by this author
  • Allen W. Cheever,

    1. Biomedical Research Institute, Rockville, MD
    Search for more papers by this author
  • Joseph P. Sypek,

    1. Wyeth Research, Cambridge, MA
    Search for more papers by this author
  • Margaret M. Mentink-Kane

    1. National Institute of Allergy and Infectious Diseases, National Institutes of Health, Bethesda, MD
    Search for more papers by this author

Abstract

The selectin family of cell adhesion molecules is widely thought to promote inflammatory reactions by facilitating leukocyte recruitment. However, it was unexpectedly found that mice with targeted deletion of the P-selectin gene (PsKO mice) developed unpolarized type 1/type 2 cytokine responses and severely aggravated liver pathology following infection with the type 2–promoting pathogen Schistosoma mansoni. In fact, liver fibrosis, which is dependent on interleukin 13 (IL-13), increased by a factor of more than 6, despite simultaneous induction of the antifibrotic cytokine interferon gamma (IFN-γ). Inflammation, as measured by granuloma size, also increased significantly in the absence of P-selectin. When infected PsKO mice were treated with neutralizing anti-IFN-γ monoclonal antibodies, however, granuloma size was restored to wild-type levels; this finding revealed the potent proinflammatory role of IFN-γ when expressed concomitantly with IL-13. Untreated PsKO mice also exhibited a significant (sixfold) reduction in decoy IL-13 receptor (IL-13 receptor alpha-2) expression when compared with infected wild-type animals. It is noteworthy, however, that when decoy receptor activity was restored in PsKO mice by treatment with soluble IL-13 receptor alpha-2-Fc, the exacerbated fibrotic response was completely inhibited. Thus, reduced expression of the decoy IL-13 receptor mediated by the elevated type 1 cytokine response probably accounts for the enhanced activity of IL-13 in PsKO mice and for the resultant increase in collagen deposition. In conclusion, the current study has revealed the critical role of P-selectin in the progression of chronic liver disease caused by schistosome parasites. By suppressing IFN-γ and up-regulating the decoy IL-13 receptor, P-selectin dramatically inhibits the pathologic tissue remodeling that results from chronic type 2 cytokine–mediated inflammation. (HEPATOLOGY 2004;39:676–687.)

For many infectious diseases, rapid development of a polarized type 1 or type 2 cytokine response determines whether resistance or susceptibility is observed; however, when the inflammatory agent is not inactivated or removed quickly and the polarized state persists, chronic inflammation may ensue and in some cases may actually become more damaging than the invading pathogen itself. Recent studies have shown that persistent inflammatory responses are damaging to the host, regardless of whether type 1 or type 2 cytokines dominate the response.1 The destructive tissue pathology induced by each type of response, however, appears to be distinct.2, 3 In persistent Th2 responses, fibrosis is a common pathologic sequela, with interleukin 13 (IL-13) playing an indispensable role.4–6 Because fibrosis is a significant complication in a variety of chronic diseases, including asthma, idiopathic pulmonary fibrosis, chronic graft rejection, and hepatic fibrosis, we sought to identify endogenous mechanisms that antagonize this disease pathway.

In schistosomiasis, a chronic inflammatory disease of the liver and gut, type 2 cytokines are necessary for granuloma formation and for development of hepatic fibrosis following infection.1, 7, 8 Several studies have shown that during Schistosoma mansoni infection, type 1 cytokines inhibit fibrogenesis, whereas type 2 cytokines promote fibrogenesis.2 Although Th2-mediated liver pathology is highly dependent on IL-4, IL-4 receptor, and Stat6,8–10 recent studies have identified IL-13 as the central mediator. Mice that are deficient in IL-13 fail to develop significant fibrosis, even from chronic infection.4, 5 Thus, murine schistosomiasis provides an ideal model system for elucidating the mechanisms that regulate IL-13-mediated liver fibrosis.

The selectins are a family of carbohydrate-binding adhesion molecules that regulate the extravasation of leukocytes from blood to sites of inflammation and control the recirculation of naive lymphocytes through secondary lymphoid organs.11, 12 Therefore, regulated expression of selectins and their ligands is critical for efficient recruitment of cells to sites of tissue damage and infection. Although P-selectin originally was observed on activated platelets, expression of both P- and E-selectin was found to be rapidly induced on endothelial cells during inflammatory responses.13, 14 The ligand for P-selectin, P-selectin glycoprotein ligand-1 (PSGL-1), which initially was identified on human and murine myeloid cells, also is expressed on subsets of activated effector T cells and is believed to be essential for the movement of type 1 CD4-positive T cells into inflamed skin.15, 16 Nonetheless, the extent to which selectins regulate the movement of cells to visceral organs and (more specifically) the contribution of selectins to the regulation of chronic type 2 cytokine-dependent liver disease remain relatively unclear.

In the studies presented in the current report, we investigated the function of selectins in schistosomiasis, as complementary DNA3 and oligonucleotide microarray analysis (TA Wynn et al., unpublished observations, Dec. 2000) showed elevated expression of P-, E-, and L-selectin messenger RNA (mRNA) in the liver following infection. Granuloma size, tissue eosinophilia, and fibrosis were examined in detail in three selectin-deficient strains following infection with S. mansoni or following intravenous challenge with eggs in the lungs. It is noteworthy that granuloma size and hepatic fibrosis increased substantially in the absence of P-selectin; these findings suggest that interactions between P-selectin and PSGL-1 are required to down-regulate type 2 cytokine-driven inflammation. Also notable is the fact that no defects were observed in mice with targeted deletion of the P-selectin gene (PsKO mice) exhibiting a polarized type 1 response; this suggests that the anti-inflammatory and antifibrotic effects of P-selectin are restricted to Th2-mediated disease. Thus, interactions between P-selectin and PSGL-1 may prove to be important in a variety of disorders in which type 2 cytokines promote wound healing and/or development of destructive fibrotic pathology.

Abbreviations

IL, interleukin; PSGL-1, P-selectin glycoprotein ligand-1; mRNA, messenger RNA; PsKO mice, mice with targeted deletion of the P-selectin gene; EsKO mice, mice with targeted deletion of the E-selectin gene; SEA, soluble egg antigen; sIL-13Rα2-Fc, soluble interleukin 13 receptor alpha-2-Fc; Con A, concanavalin A; ELISPOT, enzyme-linked immunospot; IL-13Rα2, interleukin 13 receptor alpha-2; RT-PCR, reverse-transcription polymerase chain reaction.

Materials and Methods

Animals, Parasite Infections, and Antigen Preparation.

Female mice age 6–8 weeks that were deficient in P-selectin or E-selectin (EsKO mice) on a C57BL/6 background were obtained from Taconic Farms (Germantown, NY). Age- and sex-matched C57BL/6 mice also were obtained from Taconic Farms. S. mansoni eggs were extracted from the livers of infected mice (Biomedical Research Institute, Rockville, MD). For the induction of pulmonary granuloma, mice were sensitized intraperitoneally, and then groups of 5–6 mice were challenged with 5,000 eggs intravenously and sacrificed on Days 4, 8, and 12. In the infection experiments, mice were infected percutaneously via the tail with 25–30 cercariae of a Puerto Rican strain of S. mansoni (NMRI) that were obtained from infected Biomphalaria glabrata snails (Biomedical Research Institute). In some experiments, mice were sensitized with S. mansoni eggs and treated with recombinant interleukin (rIL)-12 as described previously.7 In a final series of experiments, wild-type and PsKO mice were infected and treated with control immunoglobulin (clone GL113) or anti-IFN-γ monoclonal antibodies (clone XMG1.6) twice weekly (with a 1 mg intraperitoneal injection) from Week 5 through Week 8 postinfection. rIL-12 was provided by Wyeth Genetics Institute (Cambridge, MA). Soluble egg antigen (SEA) was purified from homogenized S. mansoni eggs, as is described elsewhere.17 The soluble IL-13 receptor alpha-2-Fc fusion protein (sIL-13Rα2-Fc) and immunoglobulin G were prepared as described previously5, 6 and were provided by Wyeth.18 In the IL-13 blocking studies, animals were treated with an intraperitoneal injection delivered in 0.5 mL phosphate-buffered saline (200 μg per mouse per day) every other day; the dose administered had been found to be optimal in previous studies.5, 6 All animals underwent perfusion at the time of sacrifice so that worm and tissue egg burdens could be determined, as is described elsewhere.17

Histopathology and Fibrosis.

The sizes and collagen content of pulmonary and hepatic granulomas were determined on histologic sections that were stained with Wright's Giemsa stain or with picrosirius red (Histopath of America, Clinton, MD). The percentages of eosinophils, mast cells, and other types of cells were evaluated in the same sections. In some cases, the frequency of mast cells was evaluated on a scale ranging from 0 to 7, with 0 representing a low frequency and 7 representing a high frequency. The number of schistosome eggs in the liver and gut and the collagen content of the liver, as measured by hydroxyproline levels, were determined as described elsewhere.17 The same individual scored all histologic features without knowledge of the experimental design.

Lymphocyte Culture and Cytokine Detection Using the Enzyme-Linked Immunosorbent Assay (ELISA).

Spleen and mesenteric lymph nodes (infection model) were removed aseptically, and single-cell suspensions were prepared.17 Cultures were incubated at 37°C in a humidified atmosphere of 5% CO2. Cells were stimulated with SEA (20 μg/mL), concanavalin A (Con A; 1 μg/mL), or medium alone. In some experiments, anti-CD4 monoclonal antibodies were included in cultures to block interactions between antigen-presenting cells and T cells. Supernatant fluids were harvested at 72 hours and assayed for cytokine production. IFN-γ, IL-4, and IL-5 levels were measured using two-site ELISA, as is described elsewhere.17 In some experiments, IL-4 levels were determined by assaying the proliferation of CT4S cells and were presented as cpm × 10−3.

Enzyme-Linked Immunospot (ELISPOT) Assay.

Single-cell suspensions of liver granuloma–derived lymphocytes were prepared as previously described19 and were placed in 24-well plates (at a concentration of 4 × 106 cells/mL) and stimulated with SEA (20 μg/mL). After 24 hours, cells were harvested and cultured in 96-well plates (1 × 105, 5 × 104, and 1 × 104 cells per well) that had been coated with different cytokine-specific antibodies (anti-IFN-γ, anti-IL-4, or anti-IL-13) to a final concentration of 10 μg/mL and subsequently blocked with 5% fetal calf serum. Cells were incubated overnight at 37 °C in an atmosphere of 5% CO2. After washing with phosphate-buffered saline/Tween 20 (Sigma-Aldrich Corp., St. Louis, MO), plates were incubated with biotinylated anticytokine antibodies (1:1,000 dilution) for 2 hours at 37°C. Plates were washed and incubated with alkaline phosphatase–avidin (1:5,000 dilution) for 1 hour at room temperature and then developed using 5-bromo-4-chloro-3-indolyl phosphate agarose; spots were counted the following day using an inverted microscope.

RNA Isolation and Purification and the Real-Time Polymerase Chain Reaction.

Liver tissue samples were homogenized in RNA-Stat 60 (Tel-Test Inc., Friendswood, TX), and total RNA was extracted according to the recommendations of the manufacturer. The real-time reverse-transcriptase polymerase chain reaction (ABI Prism 7900 Sequence Detection System; Applied Biosystems, Foster City, CA) was used to determine relative mRNA levels for several cytokine and cytokine receptor genes; the reaction was performed using SYBR Green PCR Master Mix (Applied Biosystems) after reverse transcription of 1 μg RNA. The amount of reaction product was determined using the comparative threshold cycle method, as described by Applied Biosystems for the ABI Prism 7700/7900 sequence detection systems; in this method, mRNA levels for each sample were normalized to hypoxanthine guanine phosphoribosyl transferase mRNA levels and then expressed as relative increases or decreases compared with levels in uninfected controls.

Statistical Methods.

Hepatic fibrosis (adjusted for egg number) decreases with increasing intensity of infection (as measured by total number of liver eggs). Thus, these variables were compared by analysis of covariance, using the logarithm of the total number of liver eggs (the covariate) and the logarithm of hydroxyproline content per egg. Variables that did not change with infection intensity were compared using one-way ANOVA or the Student's t test. Changes in cytokine mRNA expression, ELISPOT assay results, and granuloma size were evaluated using ANOVA. Results were considered significant when P < .05.

Results

Increased Pulmonary Granuloma Size With Unimpaired Tissue Eosinophilia in the Absence of P-selectin.

In initial studies, we examined whether pulmonary granuloma formation, mediated by a predominant Th2 response, was altered in PsKO mice. Mice were sensitized via intraperitoneal injection of 5,000 eggs 2 weeks before the intravenous challenge. Under these conditions, mice exhibit a vigorous anamnestic response in the lungs, with lesion size peaking between Day 7 and Day 8 following the challenge.20 Initial granuloma development was unaltered on Day 4; however, by Day 8, there was a significant increase in pulmonary granuloma size in PsKO mice compared with wild-type mice, and this increase was sustained through Day 12 (Fig. 1A). Despite the difference in inflammatory response, there was no significant difference between mouse strains in terms of granuloma-associated eosinophils at any time point (Fig. 1B).

Figure 1.

Secondary granuloma size increases and tissue eosinophilia is unimpaired in the lungs of PsKO mice. Groups of intraperitoneally egg-sensitized C57BL/6 wild-type and PsKO mice (n = 5–6 mice per group for each time point) were challenged intravenously with 5,000 purified S. mansoni eggs. Mice were sacrificed on Days 4, 8, and 12, and (A) granuloma volumes and (B) eosinophil percentages were determined. Bars (from bottom to top) indicate the 10th, 25th, 50th, 75th, and 90th percentiles, respectively, of the tested samples. Circles denote single outliers. Data are representative of several experiments. The asterisk indicates a significant difference from wild-type at a given time point.

Greater-than-sixfold Increase in Hepatic Fibrosis in Chronically Infected Mice in the Absence of P-selectin.

The results of experiments conducted using the lung model provided the first evidence of the important anti-inflammatory role played by P-selectin. To further characterize the function of P-selectin and to investigate its role in a chronic disease setting, we examined granuloma formation in the livers of wild-type and PsKO mice following infection with S. mansoni. In these experiments, animals were infected percutaneously with 25 S. mansoni cercariae and then examined for several immunologic and parasitologic changes at 8 weeks (acute changes) and 16 weeks (chronic changes) following infection. Mice underwent perfusion at the time of sacrifice, and no differences in total worm burden, number of paired worms, eggs produced per worm pair, or tissue egg burden were revealed; these findings demonstrated that wild-type and PsKO mice were equally susceptible to S. mansoni infection (Table 1).

Table 1. Infection Status of Mice
GroupParasite Recovery wk 8Parasite Recovery wk 16
Worm PairsTotal WormsEggs/Worm PairWorm PairsTotal WormsEggs/Worm Pair
  • (N)

    (N), Number of mice. All data are means for each group ± 1 SEM.

Exp 1
A      
  WT5.42 ± 0.7212.7 ± 1.116.44 ± 0.705.50 ± 0.4312.0 ± 1.29
 (N)(7)(7)(7)(4)(4)(4)
B      
 PsKO4.66 ± 0.8012.5 ± 1.216.00 ± 0.393.50 ± 1.348.75 ± 1.1440.91 ± 10.84
 (N)(6)(6)(6)(4)(4)(4)
Exp 2
A      
 WT6.25 ± 1.2713.6 ± 2.758.25 ± 0.536.40 ± 0.5513.70 ± 0.9918.02 ± 1.83
 (N)(8)(8)(8)(10)(10)(10)
B      
 PsKO4.75 ± 0.9311.2 ± 1.646.57 ± 0.983.13 ± 0.378.12 ± 1.1531.15 ± 4.26
 (N)(8)(8)(8)(8)(9)(9)

Liver sections from infected wild-type and PsKO mice were examined microscopically to measure granuloma size and to determine the average eosinophil and mast cell content in lesions (Fig. 2). It is noteworthy that macroscopic observations at the time of sacrifice provided the first evidence that infected PsKO mice were developing severe disease (Fig. 3A). Granuloma size nearly doubled in PsKO mice, and this increase was maintained throughout the acute (Week 8) and chronic (Week 16) stages of infection (Fig. 2A). Marked inflammation and liver regeneration also were obvious by Week 20 in PsKO mice (Fig. 3A), and these findings were confirmed by comparing liver weights (Fig. 3B). Consistent with findings made in the pulmonary granuloma model, the number of lesion-associated eosinophils was nearly identical in both types of mice (Fig. 2B). However, mast cell counts were decreased in PsKO mice at both time points, suggesting an important role for P-selectin (Fig. 2C). More detailed differential counts revealed additional differences, including a 36%–38% increase in plasma cell levels and a 43%–50% decrease in granuloma-associated fibroblasts in PsKO mice at both time points (Table 2). There was no evidence of liver inflammation in either mouse strain before infection.

Figure 2.

Granuloma size and fibrosis are exacerbated in infected PsKO mice. Groups of 10 C57BL/6 wild-type and PsKO mice were sacrificed 8 or 16 weeks after S. mansoni infection, and (A) liver granuloma size, (B) tissue eosinophilia, (C) mast cell frequency, and (D) hepatic fibrosis (in μmol per worm pair) were evaluated. The experiment shown was repeated twice, with similar results both times. The asterisk indicates a significant difference from wild-type at a given time point.

Figure 3.

Chronically infected PsKO mice exhibit marked liver regeneration. (A) Marked liver regeneration observed in PsKO mice after 20 weeks of infection. Note the growth of the lower lobes of the liver in the PsKO mouse (arrow). The lower panels show low-power views of liver sections stained with picrosirius red. The examples shown are from wild-type and PsKO mice with similar worm and egg burdens. Original magnification ×5. (B) Total liver weight (in grams) as a percentage of total body weight in wild-type and PsKO mice after 20 weeks of infection. Data shown were pooled from two separate experiments (n = 20 per group).

Table 2. Granuloma-Associated Leukocytes
Group LymphsPlasma CellsMacsFibroblastsEosinophilsMast Cells
  1. N, Number of S. mansoni infected mice pooled from two separate experiments. All data are granuloma associated leukocytes and are shown as means ± SD. Values in bold and italics are significantly different from the WT group at that time point, P < 0.05 by Student's t-test.

Week 8       
 AWT (N = 15)4.69 ± 0.475.30 ± 0.5828.44 ± 1.617.72 ± 0.8253.73 ± 2.420.12 ± 0.08
 BPsKO (N = 14)5.94 ± 0.428.26 ± 1.2030.16 ± 1.324.38 ± 0.7651.21 ± 2.440.05 ± 0.03
Week 16       
 AWT (N = 13)7.44 ± 0.558.32 ± 1.1432.92 ± 1.2416.00 ± 1.8235.00 ± 2.330.71 ± 0.19
 BPsKO (N = 12)6.75 ± 0.6413.45 ± 2.2128.38 ± 1.368.16 ± 0.9343.83 ± 2.420.10 ± 0.04

The extent of hepatic fibrosis following infection was assessed by measuring hydroxyproline levels in the liver (Fig. 2D) and by staining tissue sections with the collagen-specific stain picrosirius red (Fig. 3A). It is striking that both techniques demonstrated that PsKO mice were developing uncontrolled liver fibrosis. In fact, the significant twofold increase observed by Week 8 in PsKO mice became a greater-than-sixfold increase by Week 16.

Because E- and P-selectin share many functional activities, and also because both were induced in the liver following infection (data not shown), we conducted additional studies in EsKO mice. However, there was no evidence of increased inflammation or fibrosis in EsKO mice, even after chronic infection (data not shown). Similar findings were made in mice that were deficient in L-selectin21 (data not shown). Although there was a marked reduction in mesenteric lymph node size in L-selectin-deficient mice, this reduction had no effect on liver pathology, even after 16 weeks of infection.

Development of a Nonpolarized Type 1/Type 2 Cytokine Response in the Absence of P-selectin.

Cells isolated from mesenteric lymph node, spleen, and granulomatous liver samples were assayed using either ELISA or ELISPOT. Mesenteric lymph node cells were stimulated with SEA or with Con A, and 72 hours later, individual culture supernatants were assayed for IFN-γ, IL-5, IL-10, and IL-13. As expected, infected wild-type mice developed a polarized type 2 response, with high levels of IL-5, IL-10, and IL-13 and relatively little IFN-γ in response to SEA or Con A at 8 weeks (Fig. 4A) and at 16 weeks (Fig. 4B). In contrast, infected PsKO mice developed a robust and mixed type 1/type 2 response. Although egg-specific IL-10 production was comparable to what was observed in wild-type mice, IL-5 and IL-13 levels were slightly and consistently elevated at both 8 weeks and 16 weeks. Nonetheless, increased production of IFN-γ was the most dramatic change noted, and this change was observed only after infection. PsKO mice exhibited high levels of IFN-γ in response to both SEA and Con A by Week 8 (Fig. 4A), and although these levels decreased somewhat by Week 16, they remained significantly elevated (Fig. 4B). There was no detectable response to SEA before infection. The cytokine response to splenocytes was similar, although in addition to exhibiting elevated IFN-γ production (data not shown), PsKO cells secreted an average of 39.8 ng/mL IL-13, whereas wild-type cultures produced an average of 24.8 ng/mL; these findings provided additional evidence of a nonpolarized cytokine response in infected PsKO mice. IL-12 production also was elevated in PsKO mice (data not shown).

Figure 4.

Severe liver pathology in PsKO mice is characterized by a mixed type 1/type 2 cytokine response in lymphoid tissues. C57BL/6 wild-type and PsKO mice were sacrificed (A) 8 weeks or (B) 16 weeks after S. mansoni infection, and individual mesenteric lymph node cultures were restimulated with SEA (20 μg/mL) or Con A (1 μg/mL) for 72 hours. Culture supernatants were assayed for IFN-γ, IL-13, IL-5, and IL-10. Bars (from bottom to top) indicate the 10th, 25th, 50th, 75th, and 90th percentiles, respectively, of the tested samples. Circles denote single outliers. Data shown are from 6 mesenteric lymph node cultures per group at the 8-week time point and from 4 individual samples at the 16-week time point. Similar data were obtained in a repeat study.

The local cytokine response also was examined with the ELISPOT assay, using leukocytes isolated from granulomatous liver samples. Similar to the ELISA results obtained with mesenteric lymph node cells and splenocytes, the frequency of IFN-γ-producing cells was markedly increased (by a factor of 4–5) in PsKO mice compared with wild-type mice (Fig. 5). It is noteworthy that both groups exhibited similar numbers of IL-4- and IL-13-producing cells, although there were slightly more type 2 producers by Week 10. Despite the increase in IFN-γ levels, PsKO and wild-type mice exhibited relatively normal serum immunoglobulin E (IgE) responses by Week 20 following infection (total IgE: wild-type, 3.59 ± 0.36 mg/mL; PsKO, 4.11 ± 0.75 mg/mL).

Figure 5.

The frequency of IFN-γ-producing cells increases in the livers of infected PsKO mice. Liver granuloma–derived lymphocytes were plated in 24-well plates (4 × 106 cells per mL) and stimulated with SEA (20 μg/mL). ELISPOT assays were performed as described in Materials and Methods, and the frequency of cytokine-producing cells per 106 granuloma cells is shown. Bars denote mean values ± standard errors from three individual samples that were prepared from three pooled livers per group. The asterisk indicates a significant difference from wild-type for a given cytokine (P < .05).

Reduction of Lesion Size to Wild-Type Levels With Concomitant Increased Fibrosis Following IFN-γ Blockade.

The most dramatic change in immune response that may have accounted for the altered pathologic reaction of infected PsKO mice was the marked increase in IFN-γ levels. To directly investigate the function of IFN-γ, we administered neutralizing anti-IFN-γ monoclonal antibodies to infected wild-type and PsKO mice. Consistent with earlier observations, liver granuloma size increased significantly in untreated PsKO mice (Fig. 6A). It is noteworthy, however, that lesion size was restored to wild-type levels in PsKO mice following treatment with anti-IFN-γ monoclonal antibodies. In contrast, fibrosis increased in similar fashion in both wild-type and PsKO mice following anti-IFN-γ treatment (Fig. 6B).

Figure 6.

Granuloma size decreases while fibrosis is enhanced in PsKO mice following IFN-γ monoclonal antibody treatment. Groups of 5 C57BL/6 wild-type and PsKO mice were sacrificed 8 weeks after S. mansoni infection, and (A) liver granuloma size and (B) hepatic fibrosis were evaluated. Mice were treated with either a control anti–beta galactosidase antibody (GL113) or anti-IFN-γ monoclonal antibodies (XMG1.6). The asterisk indicates that the PsKO group is significantly different from the corresponding wild-type group.

Control of Hepatic Fibrosis by Type 2 Cytokines in PsKO Mice.

Data from the pulmonary and liver granuloma models suggest that P-selectin is critical to the polarization of type 2 cytokine responses in vivo. Thus, to determine whether the development of a polarized type 1 response was similarly affected by the absence of P-selectin and whether fibrosis in PsKO mice was dependent on type 2 cytokines, immune deviation studies were conducted. In these experiments, groups of wild-type and PsKO mice were sensitized with schistosome eggs in the presence of Th1-inducing adjuvant IL-12 before infection to generate an egg-specific type 1 response.7 The effects on granuloma size, fibrosis, tissue eosinophilia, mastocytosis, and subsequent Th1/Th2-type cytokine production then were compared with the corresponding effects in unsensitized infected controls. As expected, unsensitized PsKO mice exhibited larger granulomas and worse fibrosis compared with wild-type mice (Fig. 7A). Nonetheless, the development of egg-induced pathology was reduced to a similar extent in both groups of mice following sensitization with schistosome eggs and IL-12. Tissue eosinophil and mast cell responses also were much less pronounced in IL-12-sensitized mice. These findings were in agreement with the observed development of a relatively polarized CD4-positive Th1-type response in both wild-type and PsKO mice (Fig. 7B). In fact, high levels of IFN-γ and low levels of antigen-specific IL-4 were detectable in all sensitized animals, regardless of P-selectin genotype. The data on IL-5 and IL-13 levels were similar to the data on IL-4 levels (data not shown).

Figure 7.

Liver fibrosis is controlled by the type 2 cytokine response. (A) Groups of control (−) and egg/IL-12-sensitized (+) C57BL/6 wild-type and PsKO mice were sacrificed 8 weeks after S. mansoni infection, and liver granuloma size, tissue eosinophilia, mast cell frequency, and hepatic fibrosis were evaluated. Asterisks indicate significant differences from the control wild-type group. (B) Mesenteric lymph node cultures also were prepared at the time of sacrifice, and pooled samples were restimulated with SEA (20 μg/mL) or Con A (1 μg/mL) for 72 hours. Anti-CD4 monoclonal antibodies were included in some cultures to block interactions between T cells and antigen-presenting cells. Culture supernatants were assayed for IFN-γ and IL-4, and average values ± standard deviations are shown. Similar results were obtained in a second experiment.

Reduction of IL-13 Receptor Alpha-2 (IL-13Rα2) Expression in PsKO Mice and Reversal of Exacerbated Fibrotic Response by Restoration of Decoy Receptor Activity With sIL-13Rα2-Fc.

The decoy receptor IL-13 receptor alpha-2 (IL-13Rα2) is critical to the functionality of IL-13, because IL-13Rα2 does not associate with a signaling chain.22 It is noteworthy that binding to this receptor has been shown to block interactions with the signaling IL-4 receptor alpha/IL-13 receptor alpha-1 complex, and recent studies conducted in IL-13Rα2-deficient mice have documented the role of IL-13Rα2 as a functional decoy receptor for IL-13.23, 24 Therefore, we examined the expression of IL-13Rα2 to determine whether the exacerbated disease observed in PsKO mice was associated with changes in decoy receptor expression. In these experiments, the real-time reverse-transcription polymerase chain reaction (RT-PCR) was performed using total RNA isolated from the livers of mice that had been infected for 8 weeks; this RT-PCR assay has been established to be a highly specific and sensitive method for measuring IL-13Rα2 expression.23, 24 IL-13Rα2 mRNA expression was almost undetectable in the livers of uninfected wild-type and PsKO mice, whereas a 30-to-60-fold induction of gene expression was observed in wild-type mice following infection (Fig. 8A). It is striking that the observed increases were minimal in infected PsKO mice and in animals treated with rIL-12, regardless of genotype; these findings indicated that a highly polarized Th2 response was critical for receptor expression.

Figure 8.

sIL-13Rα2-Fc treatment reverses the exacerbated fibrotic response of PsKO mice. (A) Groups of control (−) and egg/IL-12-sensitized (+) C57BL/6 wild-type and PsKO mice were sacrificed 8 weeks after S. mansoni infection, and individual liver tissue samples were assayed by real-time RT-PCR for expression of IL-13Rα2 mRNA. Bars (from bottom to top) indicate the 10th, 25th, 50th, 75th, and 90th percentiles, respectively, of the tested samples. Circles denote single outliers. In separate experiments, infected C57BL/6 and PsKO mice were either left untreated or treated with control immunoglobulin G (cIg) or sIL-13Rα2-Fc (13Rα2) as described in Materials and Methods. Ten mice per group were sacrificed on Week 10 after infection, and (B) granuloma size and (C) liver hydroxyproline levels were measured.

To determine whether restoration of IL-13Rα2 activity would reverse the profibrotic phenotype of PsKO mice, animals were treated with sIL-13Rα2-Fc.18 In these experiments, wild-type and PsKO mice were infected, and beginning on Week 5, some animals received either sIL-13Rα2-Fc or immunoglobulin G treatment for 5 weeks. All animals were sacrificed 10 weeks after infection, and the effects of treatment on egg-induced inflammation (Fig. 8B) and liver fibrosis (Fig. 8C) were examined. Consistent with previous observations,4 sIL-13Rα2-Fc treatment had no effect on the acute granulomatous inflammatory response (Fig. 8B). It is noteworthy, however, that all treated mice experienced a marked and highly significant decrease in fibrosis, as measured by liver hydroxyproline content (Fig. 8C). In fact, although lesion formation remained elevated in PsKO mice (Fig. 8B), fibrogenesis decreased to a similar extent in both groups of mice treated with sIL-13Rα2-Fc. Together, these findings suggest that the elevated level of liver fibrosis observed in PsKO mice was largely due to the reduced decoy IL-13Rα2 response (Fig. 8A) and to the resultant increase in IL-13 effector function (Fig. 8C).

Discussion

The movement of leukocytes from the blood to tissues is mediated in part by interactions between selectins and their ligands. In fact, selectins are widely thought of as critical mediators of innate immunity, because they regulate the initial tethering of leukocytes to vascular surfaces. Consequently, a great deal of research on selectins has focused on the role of these molecules in the initiation of immune responses. However, much less is known about the regulation and role of selectins in chronic inflammation. Of the three known selectins, P-selectin was found to have the most prominent role during infection with S. mansoni. It is surprising, however, that the effects of P-selectin deficiency were most evident in the livers of chronically infected mice and in animals experiencing a secondary immune response in the lung; in these animals, granuloma size increased dramatically in the absence of P-selectin. Thus, although we originally hypothesized that the egg-induced inflammatory response would be reduced,11, 13, 14 PsKO mice actually experienced a much more severe reaction. Thus, the findings presented in the current study have revealed a new role for P-selectin by demonstrating that this molecule also exhibits critical anti-inflammatory and antifibrotic activity. Furthermore, the fact that similar findings were made in both the lung and the liver suggests that such activity may be a global feature of P-selectin, as the activity was not restricted to only one type of tissue. Nonetheless, it remains unclear whether this activity results from the direct effect of P-selectin on leukocyte recruitment or whether it simply is a consequence of the altered immune response of the host.

Also surprising was the finding that P-selectin played no role in eosinophil recruitment, because numerous studies have demonstrated reduced rolling, adhesion, and/or recruitment of eosinophils in mice lacking P-selectin or in animals treated with selectin-blocking antibodies.25–29 In fact, it has been suggested that P-selectin antagonists may be effective in inhibiting eosinophil accumulation at sites of allergic inflammation.30–33 Nonetheless, the current study clearly demonstrated that the blocking of P-selectin/PSGL-1 interactions alone may be insufficient. A possible explanation for these findings is that many of the previous studies that examined the function of P-selectin involved models of acute, rather than chronic, inflammation; a recent study showed that although eosinophil recruitment was markedly reduced in PsKO mice 3 hours after allergen challenge in the lung, the level of suppression decreased dramatically by 24 hours after challenge.34 This observation, combined with our finding of normal eosinophil recruitment, suggests that other mediators contribute to eosinophil recruitment during chronic inflammation.

PSGL-1 also is found on bone marrow–derived mast cells, and cell adhesion assays have demonstrated that mast cells bind P-selectin. Furthermore, PSGL-1 antibodies have been shown to effectively block mast cell binding; this represented the first evidence of a possible role for P-selectin in mast cell recruitment.35 Although mast cells are a minor cellular constituent of schistosome granulomas in infected wild-type mice (Fig. 2C), far fewer mast cells were found in lesions in PsKO mice; this finding provided evidence of P-selectin–dependent recruitment in vivo. The exact role of mast cells in schistosome egg–induced pathology, however, remains unclear, although data from studies on mast cell–deficient animals suggest that the contribution of mast cells is relatively minor.36, 37

In humans, the development of fibrotic liver pathology is the primary cause of chronic morbidity and mortality following infection with S. mansoni, primarily because such pathology contributes to the development of portal hypertension and to the formation of esophageal varices.38 In the murine model of schistosomiasis, the development of fibrosis is almost entirely dependent on IL-13,4, 5 whereas granuloma formation itself requires both IL-4 and IL-13.39, 40 Therefore, elucidation of the regulation of these cytokines and the roles of their cognate receptors may help to clarify how Th2-mediated disease is regulated.41–43 It is striking that although fibrosis and granuloma size increased in the absence of P-selectin, production of the profibrotic cytokine IL-13 did not change dramatically. Instead, increased production of IFN-γ was the most striking change observed in PsKO mice. Thus, knockout mice experienced a mixed type 1/type 2 cytokine response following infection. This was surprising, because much of the in vivo data on P-selectin suggested that this molecule was critical for the recruitment of IFN-γ-producing Th1 cells,15, 16 and a decrease in IFN-γ levels therefore might have been expected. It is noteworthy that the addition of anti-CD4 monoclonal antibodies to our lymph node cultures confirmed that CD4-positive T cells were required for IFN-γ production (Fig. 7B). Thus, our studies demonstrated enhanced, rather than suppressed, recruitment of type 1–producing cells in PsKO mice and, even more unexpectedly, exacerbated IL-13-dependent disease.4, 5, 8, 10

It has been demonstrated that mixed Th1/Th2 responses can trigger more severe disease in models of allergic inflammation and asthma.44–47 One study, which used an adoptive transfer system, examined the ability of Th1 cells to counterbalance the proasthmatic effects of Th2 cells. Not only were Th1 cells incapable of attenuating Th2 cell–induced airway hyperreactivity, but they even caused more severe acute airway disease.44 A related study of eosinophilic airway inflammation suggested that Th1 cells may provide some as yet undefined signal that is required for the effective recruitment of Th2 cells.45 Together, these studies demonstrate that, rather than being counterregulatory, the presence of both Th1 and Th2 cells exacerbates inflammation. To determine whether the increased IFN-γ response observed in PsKO mice was influencing granuloma formation, we treated infected knockout mice with neutralizing IFN-γ monoclonal antibodies. It is noteworthy that granuloma size was restored to wild-type levels by anti-IFN-γ treatment, confirming the important aggravative role of IFN-γ. Nonetheless, there was no evidence that IFN-γ was responsible for the relatively severe pattern of fibrosis observed in PsKO mice. Instead, fibrosis increased to a similar extent in both groups following anti-IFN-γ treatment, thus confirming the well-known antifibrotic activity of IFN-γ in schistosomiasis.48 Together, these findings support and extend the findings of recent studies showing that IFN-γ inhibits certain activities attributed to IL-13 and exacerbates the effects of this cytokine in other situations.46

To determine whether the severe fibrosis observed in infected PsKO mice was due to the type 2 cytokine response, immune deviation studies involving IL-12 were conducted.7 Consistent with previous experiments, PsKO mice that were not treated with IL-12 developed a mixed type 1/type 2 cytokine response (Fig. 7B) and much more severe disease (Fig. 7A). Nonetheless, granuloma size and fibrosis decreased dramatically following sensitization with eggs and IL-12. In fact, both wild-type and PsKO mice exhibited highly polarized type 1 cytokine responses, reduced granuloma size, markedly suppressed liver fibrosis, and decreased tissue eosinophilia and mastocytosis. These findings confirmed the critical profibrotic activity of type 2 cytokines in wild-type mice and infected PsKO mice; more notable, however, was that no significant regulatory role for P-selectin during a chronic type 1 polarized immune response was found.

We previously demonstrated that sIL-13Rα2-Fc was a highly effective inhibitor of IL-13-mediated fibrogenesis.4 The endogenous IL-13Rα2 is not believed to associate with a signaling chain and can be found as a soluble receptor in the serum and urine of mice.22 Recent knockout studies have confirmed that IL-13Rα2 acts as a functional decoy receptor for IL-13, and experiments examining the regulation of IL-13Rα2 have suggested that receptor expression is dependent on IL-4 receptor alpha and Stat6 and is suppressed by IL-12 via an IFN-γ-dependent mechanism.23, 24 Consistent with these observations, IL-13Rα2 mRNA was nearly undetectable in liver samples from uninfected wild-type and PsKO mice but was markedly induced in wild-type mice after the onset of egg laying. It is noteworthy that wild-type mice were developing highly polarized type 2 cytokine responses. In contrast, infected PsKO animals, which were developing mixed type 1/type 2 cytokine responses, showed little evidence of receptor up-regulation following infection. Thus, although production of the profibrotic cytokine IL-13 was nearly identical in both groups of mice, PsKO mice exhibited a markedly reduced decoy receptor response that probably was mediated by the mixed cytokine response. Therefore, decreased expression of the decoy receptor IL-13Rα2 is the likely explanation for the exacerbated pattern of IL-13-dependent fibrosis that was observed. That sIL-13Rα2-Fc treatment was highly effective in ameliorating fibrosis in PsKO animals further supports this conclusion (Fig. 8).

Although selectins commonly are referred to as proinflammatory mediators, the findings presented here suggest that the prevailing paradigm for selectins should be modified to accommodate the novel anti-inflammatory and antifibrotic activities described in the current study. The experiments conducted in the schistosomiasis model suggest a much more complex mechanism of action for P-selectin in vivo, as was reported recently in models of arthritis and glomerulonephritis.49, 50 In the current study, for the first time, P-selectin was shown to play a critical role in the polarization of a type 2 cytokine-driven inflammatory response. The result of the mixed type 1/type 2 cytokine response was a much more severe pathologic reaction to the parasite's eggs; this reaction was mediated by heightened IFN-γ levels, reduced IL-13Rα2 levels, and enhanced IL-13 activity. Therefore, targeting interactions between P-selectin and PSGL-1 may prove to be effective in treating a variety of diseases in which chronic IL-13 expression is known to trigger inflammation and tissue-damaging fibrosis. In summary, the current study has revealed a novel and potent mechanism for attenuating the pathologic changes that accompany chronic type 2 cytokine-driven inflammation, and this revelation may offer new avenues for the treatment of fibrotic liver disease.

Acknowledgements

The authors thank Mary Leusink and Debra Donaldson for their comments and suggestions regarding the article. The authors also thank Fred Lewis (Biomedical Research Institute, Rockville, MD) for providing parasites, Bruce Jacobson for providing RT-PCR primer sequences for IL-13Rα2, and the animal staff at the National Institutes of Health (Bethesda, MD) for their excellent technical assistance.

Ancillary