Pancreatitis-Associated Protein I Suppresses NF-κB Activation through a JAK/STAT-Mediated Mechanism in Epithelial Cells¹

Pancreatitis-associated protein I (PAP I)³ is a C-type lectin secreted in the pancreas during the acute phase of pancreatitis (1). It accounts for ∼5% of the protein secreted during the acute phase of inflammation and returns to undetectable levels when the pancreas has totally recovered (2). Two other proteins with similar structure have been identified and named PAP II and PAP III (3, 4). All these proteins show significant homology to lithostathine/regIα, a protein that is also overexpressed during the course of pancreatitis (5).

Although PAP I was originally characterized in pancreas, its expression has been observed in a variety of tissues, including epithelial cells of the small intestine (6), kidney (7), and uterus (8). PAP I is also up-regulated in a number of diseases, such as ulcerative colitis, human colorectal carcinoma, and hepatocellular carcinoma. The physiological role of PAP I remains unclear, but several functions have been suggested. Some data support its involvement in tissue regeneration and cell proliferation (9, 10), whereas overexpression of PAP I in pancreatic acinar cells also increases resistance to apoptosis induced by oxidative stress (11) and TNF-α (12).

The first report suggesting an anti-inflammatory effect of PAP I was based on an ex vivo model of isolated lung. In this model, PAP I administration reduced the edema and synthesis of thromboxane A₂ induced by fMLP (13). In an effort to understand the protective effects of PAP I, we recently investigated the effect of the absence of this protein on the inflammatory response in an experimental model of acute pancreatitis (14). We reported that infusion of anti-PAP I Abs into rats increased neutrophil infiltration in pancreatic tissue. In the same work we observed that PAP I treatment reduced the synthesis of IL-6 and TNF-α in activated macrophages and in the pancreatic acinar cell line, AR42J. A recent article confirmed the protective function of PAP I in acute pancreatitis using an antisense PAP oligonucleotide to block the expression of all three isoforms of PAP (15). Under these conditions, the severity of the pancreatitis was increased significantly. In contrast, administration of high doses of PAP I was shown to induce inflammation in the lung (16). Together, these features are reminiscent of some cytokines that induce severe inflammation and shock at higher concentrations (17).

An anti-inflammatory activity of PAP I would be consistent with its strong induction observed in the course of inflammatory diseases such as pancreatitis, Crohn’s disease, and ulcerous colitis (18). It is noteworthy that the anti-inflammatory responses of PAP I and IL-10 share several features. Both inhibit the activation of neutrophils and macrophages (19), inhibit IL-6 and TNF-α synthesis, and block translocation of NF-κB (14). However, IL-10 is principally produced by macrophages (20), whereas PAP I is synthesized in epithelial cells. Hence, PAP I synthesis could participate in the protection of epithelial cells in response to excessive inflammation, whereas IL-10 has a similar function in macrophages. Finally, the demonstration that IL-22, an IL-10-related cytokine, is able to activate the expression of PAP I in epithelial cells (21) suggests a link between PAP I and the IL-10 family of cytokines. Consequently, we have investigated the parallelisms between PAP I and IL-10 anti-inflammatory mechanisms.

The molecular mechanism through which PAP I inhibits the inflammatory process remains to be elucidated. To gain some insight into this process, we took advantage of the functional similarities between PAP I and IL-10, whose signaling pathway, leading to inhibition of inflammation, has been well described. Interaction of IL-10 with its receptor induces the recruitment of receptor-associated JAKs. JAKs then activate STAT1, -3, and -5 by tyrosine phosphorylation, resulting in their translocation to the nucleus, in which they will activate the expression of several genes (22). Among the genes activated is suppressor of cytokine signaling 3 (SOCS3), whose induction occurs via a JAK/STAT3-dependent pathway (23). Once synthesized, SOCS proteins bind the regulatory tyrosine located in the activation loop of JAK, through Src homology 2 domains, blocking access to the active site and preventing further phosphorylation (24). The negative feedback mediated by SOCS3, which inhibits activation of the JAK/STAT cascade, is a key regulatory element of the IL-10 pathway. However, it is not the only one, because IL-10 also interferes with the NF-κB-mediated proinflammatory pathway by blocking nuclear translocation of the transcription factor (25) and, for the fraction of NF-κB already present in the nucleus, its binding to the DNA (26).

In this study we have assessed whether IL-10 and PAP I could reciprocally regulate their own expression and whether the JAK/STAT and NF-κB signaling pathways are also involved in the suppression of inflammation mediated by PAP I. The study was performed in vitro using AR42J, a rat pancreatic acinar cell line known both to synthesize PAP I and to be responsive to PAP I and IL-10. Our results show that 1) PAP I mRNA expression is induced in AR42J cells by both IL-10 and PAP I; 2) the anti-inflammatory role of PAP I requires de novo protein synthesis; and 3) PAP I-induced NF-κB inhibition is prevented when STAT3 signaling pathway is blocked, showing a cross-talk between both signaling pathways.

All culture supplies were obtained from Invitrogen Life Technologies. Reagents for SDS-PAGE and nitrocellulose membranes were obtained from Amersham Biosciences. Anti-phosphotyrosine-STAT3 (Tyr⁷⁰⁵) Ab was purchased from Cell Signaling Technology. Ab against β-actin and the secondary Ab linked to HRP were obtained from Sigma-Aldrich. Abs against STAT3, p65, histone H1, and SOCS3 were obtained from Santa Cruz Biotechnology. Alexa fluor 488-conjugated anti-goat Ab, used as a secondary Ab in immunofluorescence experiments, was purchased from Molecular Probes. Recombinant TNF-α was obtained from BioSource International. Recombinant IL-10 was obtained from BD Biosciences. Tyrphostin AG490 was purchased from Calbiochem. PAP I was purified from rat pancreatic juice over an immunoaffinity column and eluted in 200 mM glycine-HCl buffer (pH 2.8) as previously described (27). The following reagents were obtained from Sigma-Aldrich: cycloheximide, NaVO₃, sodium fluoride, Nonidet P-40, and ethidium bromide.

AR42J cells were grown in monolayers until 70–80% confluent in DMEM supplemented with 10% FCS Gold (PAA Laboratories), l-glutamine (2 mM), penicillin (100 U/ml), and streptomycin (100 μg/ml) at 37°C in a humidified atmosphere of 95% air and 5% CO₂. Cultures were split every 3 days by trypsinization with 0.1% trypsin in Ca²⁺/Mg²⁺-free PBS containing 0.9 mM EDTA. Experiments were performed on cells plated in round tissue culture plates (10-cm diameter).

Analysis expression of PAP I, IL-10, TNF-α, SOCS3, and IL-6 was conducted using a semiquantitative RT-PCR method. Total RNA from cells was extracted using the TRIzol reagent (Invitrogen Life Technologies). One microgram of total RNA was used for amplification using the Invitrogen One Step RT-PCR System according to the manufacturer’s instructions.

The following primers were used: rat PAP I: forward, 5′-TGACAAGCTGCCACAGAATC-3′; and reverse, 5′-GCTCCTACTGCTATGCCCTG-3′; IL-10: forward, 5′-ATAACTGCACCCACTTCCCA-3′; and reverse, 5′-TTCTCACAGGGGAGAAATCG-3′; TNF-α: forward, 5′-ACTGAACTTCGGGGTGATTG-3′; and reverse, 5′-GTGGGTGAGGAGCACGTAGT-3′; SOCS3: forward, 5′-CCTTTGAGGTTCAGGAGCAG-3′; and reverse, 5′-CGTTGACAGTCTTCCGACAA-3′; and IL-6: forward, 5′-CCGGAGAGGAGACTTCACAG-3′; and reverse, 5′-GAGCATTGGAAGTTGGGGTA-3′. Fragments were amplified using 25–30 cycles of PCR; each cycle consisted of 15 s at 94°C, 30 s at 55°C, and 1 min at 72°C. The resulting RT-PCR products were electrophoresed on 2% agarose gels with DNA markers, stained with ethidium bromide, and visualized under UV light. β-Actin was used as an internal control for stable expression (housekeeping gene) in all experiments. The forward primer was 5′-TCATGAAGTGTGACGTTGACATCCGT-3′, and the reverse primer was 5′-CCTAGAAGCATTTGCGGTGCACGATG-3′.

Cells were lysed using the Nuclear Extract Kit from Active Motif under conditions for preparation of whole cell or nuclear and cytoplasmatic extracts. SDS-PAGE was performed on 10 or 12% acrylamide gels. Proteins were electrotransferred to nitrocellulose membrane and probed with primary Ab (anti-phosphotyrosine-STAT3, 1/1000; anti-histone H1, 1/500; anti-SOCS3, 1/200; anti-p65, 1/400; anti-β-actin, 1/400). The membranes were incubated with the corresponding peroxidase-conjugated secondary Ab, washed, and then incubated with ECL reagents (Amersham Biosciences) before exposure to high performance chemiluminescence films. Gels were calibrated using Bio-Rad standard proteins with markers covering a 7- to 240-kDa range.

For the EMSA, the oligonucleotide used was a 30-bp double-stranded construct (5′-CCTGTGCTCCGGGAATTTCCCTGGCCTGGA-3′) corresponding to a sequence (−72 to −42) in the CINC proximal promoter region containing the NF-κB motif (underlined). Digoxigenin 3′-end labeling was accomplished by using a DIG Gel Shift Kit (Roche). Samples were separated by electrophoresis on an 8% polyacrylamide gel. The gel was transferred onto Hybond-N⁺ nylon hybridization transfer membrane (Amersham Biosciences) using a Hoefer TE70 semidry transfer unit (Amersham Biosciences). Membranes were exposed to X-OMAT AR film (Eastman Kodak).

To monitor STAT3 and NF-κB translocations, cells were incubated in coverslips overnight at 37°C in a humidified atmosphere of 95% air and 5% CO₂. After treatments, cells were fixed with 3.5% formaldehyde for 5 min at room temperature. The cells were incubated with anti-STAT3 or p65 Abs and Alexa fluor 488-conjugated anti-goat secondary Ab. Nuclear localization was examined by fluorescence microscopy.

ImageJ 1.32 software (obtained from 〈http://rsbweb.nih.gov/ij/download.html〉) was used to quantify the intensities of the bands obtained in Western blots and RT-PCR experiments.

Statistical evaluation was performed by unpaired Student’s t test or, when multiple comparisons were made, by ANOVA. A value of p < 0.05 was considered statistically significant.

After 4 h of incubation with increasing concentrations of immunopurified rat PAP I added to the culture medium, AR42J acinar cells displayed a dose-dependent induction of PAP I gene expression. Incubation with 1.25, 6, and 30 nM PAP I (20, 100, and 500 ng/ml) resulted in 1×, 2×, and 10× increases in PAP I expression, respectively (Fig. 1). Analysis of the kinetics of this induction revealed that the level of PAP I mRNA was already increased 15 min after addition of 30 nM PAP I and reached a plateau at 30 min that was maintained until the end of the experiment (4 h; Fig. 2 A).

We also analyzed the effect of PAP I addition on IL-10 mRNA expression. PAP I (1.25–30 nM) induced no significant change in IL-10 gene expression after 4 h of incubation. In a similar experiment, we monitored the influence of 1 nM IL-10 (20 ng/ml) on PAP I and IL-10 gene expression (Fig. 2 B). We observed a strong increase in mRNA levels of PAP I after a 30-min exposure to the cytokine. Again, PAP I mRNA levels remained elevated for the remainder of the experiment. Induction of IL-10 mRNA expression, although significant, was much weaker.

One of the important features of the anti-inflammatory effect of PAP I is that it counteracts the activation of genes encoding inflammatory mediators, such as TNF-α and IL-6. This property was verified in AR42J cells, using the TNF-α autocrine loop as a model. Strong expression of TNF-α was observed in AR42J upon addition of TNF-α to the culture medium. This induction was inhibited by addition of 30 nM PAP I (Fig. 3). To verify that the inhibition of TNF-α was a consequence of intracellular regulation through a pathway controlled by PAP I, cycloheximide (10 μg/ml) was added to the culture medium with PAP I and TNF-α to prevent any de novo protein synthesis. Under these conditions, induction of TNF-α gene expression was restored, indicating that PAP I-mediated inhibition of TNF-α up-regulation requires the activity of newly synthesized proteins. Interestingly, the TNF-α pathway does not have the same requirements, because the amplitude of TNF-α-induced up-regulation was unchanged or even stronger in the presence of cycloheximide. This effect was not restricted to TNF-α, because similar results were obtained when evaluating the expression of IL-6.

Because it has been reported that inhibition of inflammation by IL-10 is mediated in part through SOCS3 induction, we assessed whether that protein could also be involved in the PAP I pathway. RT-PCR analysis revealed extensive and rapid accumulation of SOCS3 mRNA in AR42J cells after incubation with 30 nM PAP I (Fig. 4). These findings were confirmed by Western blot analysis; SOCS3 protein could be detected after exposure of the cells to PAP I.

Because the induction of SOCS3 by IL-10 is mainly mediated through STAT3 activation, we examined the effect of PAP I on this pathway in AR42J cells. Western blot analysis with Abs specific for the phosphorylated (activated) form of STAT3 revealed that PAP I induced tyrosine phosphorylation of STAT3 (Fig. 5). This activation was transient, reaching a maximum 30 min after addition of PAP I to the medium. Because STAT3 activation is known to result in its translocation to the nucleus, we analyzed the change in STAT3 intracellular localization by immunofluorescence in AR42J cells after PAP I addition. As shown in Fig. 6, unstimulated cells showed STAT3 labeling localized in the cytosol, whereas 15 min after PAP I stimulation, staining was mostly transferred to the nucleus. This translocation was similar to that observed in response to IL-10 treatment.

To establish the role of the STAT3 pathway in NF-κB suppression by PAP I, we examined the effect of JAK/STAT3 signaling pathway inhibition on NF-κB activation (Fig. 7). On that basis, AR42J cells were preincubated with a JAK-specific inhibitor, AG490, at a concentration of 30 μM. Immunofluorescence analysis with an specific Ab for the p65 subunit of NF-κB showed that PAP I prevented NF-κB translocation to the nucleus induced by TNF-α. This inhibition was restored by pretreatment with the JAK antagonist, AG490. These findings were confirmed by Western blot when probing p65 translocation and by the EMSA showing NF-κB-binding activity in nuclear extracts from AR42J cells. Finally, RT-PCR showed the same results when measuring the expression of TNF-α, an NF-κB-dependent gene. Our results highlight the interaction that occurs between the JAK/STAT3 and the NF-κB signaling pathways during PAP I treatment.

In this study we have provided evidence that in the pancreatic acinar cell line AR42J, PAP I modulates two signaling pathways, STAT3/SOCS3 and NF-κB. Various functions have been attributed to PAP family members, including inhibition of apoptosis (11, 12) and promotion of cell regeneration and proliferation, e.g., of motoneurons (27). However, none of them can reasonably account for the strong up-regulation observed during the acute phase of pancreatitis. We have recently reported an anti-inflammatory role for PAP I, similar to that described for IL-10. To test this hypothesis, we used AR42J cells, a rat pancreatic cell line known to synthesize significant amounts of PAP I in response to various stress-inducing stimuli (28). We observed a strong induction of PAP I gene expression in response to both IL-10 and, to our surprise, PAP I itself, suggesting the existence of positive feedback mechanisms that enhance the effects of PAP I and IL-10 in the pancreas during acute pancreatitis (Figs. 1 and 2). Interestingly, these results suggest a direct connection between the stress response, which triggers PAP I expression, and anti-inflammatory pathways in the exocrine pancreas.

Given that the anti-inflammatory properties of PAP I and IL-10 are very similar, the observation that IL-10 could induce PAP I expression suggested that PAP I could similarly induce the expression of IL-10. We monitored IL-10 gene expression after addition of various amounts of PAP I to the culture medium and incubation for up to 4 h (Fig. 2). No significant induction of IL-10 mRNA expression was observed, ruling out the possibility that the anti-inflammatory effects of PAP I are due to secondary IL-10 synthesis. This is consistent with the fact that the principal source of IL-10 is leukocytes rather than epithelial cells. Of more significance, however, was the observation that IL-10 treatment resulted in a strong and sustained induction of PAP I mRNA expression, in agreement with previous reports of PAP I induction by IL-22, an IL-10-related cytokine (21). Thus, PAP I could act as a relay for the anti-inflammatory effects of IL-10, allowing amplification of the effects and broadening of their cellular targets. The molecular mechanism by which IL-10 induces PAP I gene expression was not studied by us, but several findings suggest the involvement of the STAT3 pathway. This pathway is activated by IL-10 upon receptor binding. The promoter region of the PAP I gene contains at least two functional STAT3 response elements (29). Moreover, PAP I gene activation in motoneurons in response to ciliary neurotrophic factor (27) as well as that in PC12 cells in response to leptin (30) involve STAT3 activation. In addition, it has been reported that islet neogenesis-associated protein, a protein related to the PAP I family, also induces the phosphorylation of STAT3 (31). The positive autoregulation of PAP I could be controlled by the same pathway, because PAP I addition to the culture medium of AR42J cells induced the phosphorylation and subsequent nuclear translocation of STAT3 (Figs. 5 and 6).

In macrophages, IL-10-induced activation of STAT3 results in increased synthesis of several factors that mediate the anti-inflammatory action of the cytokine (32). To determine whether PAP I-induced activation of STAT3 in AR42J cells also involves newly synthesized proteins, we treated TNF-α-stimulated cells with cycloheximide to block protein synthesis before addition of PAP I to the culture medium. Induction of TNF-α gene expression was completely restored, demonstrating that in these cells, STAT3-mediated PAP I inhibition of inflammation requires de novo protein synthesis (Fig. 3).

Recent findings have highlighted the fact that the main effector of STAT3-mediated IL-10 signaling is SOCS3, the major feedback inhibitor of the JAK/STAT signaling pathway, which inhibits the inflammatory response induced by several cytokines and bacterial products (33). We have demonstrated that SOCS3 is also induced in response to PAP I (Fig. 4), increasing the similarities with IL-10.

The close correlation between PAP I expression and the extent of inflammation reported in several diseases, including pancreatitis and Crohn’s disease, suggests that PAP I could be a major anti-inflammatory factor in epithelial cells. This is especially true for processes involving NF-κB activation, which is known to be inhibited by PAP I in alveolar macrophages as well as in pancreatic epithelial cells (14). To establish the involvement of the STAT3 pathway in NF-κB inhibition by PAP I, we examined the effect of JAK/STAT3 signaling pathway inhibition on NF-κB activation (Fig. 7). We found that STAT3 signaling pathway blockage completely reversed PAP I-induced NF-κB inhibition. This has been shown directly by measuring the nuclear translocation of p65 subunit by Western blot, immunofluorescence, and EMSA. An indirect result that points in this direction was obtained by evaluating the expression of an NF-κB-dependent gene. These findings indicate that the anti-inflammatory effect of PAP I is mainly mediated through STAT3 activation and point to cross-talk between STAT3 and NF-κB signaling pathways.

In summary, this study indicates that PAP I inhibits the inflammatory response through a mechanism dependent on STAT3 activation. These findings suggest that PAP I, whose synthesis is strongly induced in the early stages of several inflammatory diseases, is an important anti-inflammatory factor in epithelial cells. Important functional similarities with the anti-inflammatory cytokine IL-10 suggest that PAP I could be considered an IL-10 counterpart in epithelial cells.

We thank Robin Rycroft (Language Advisory Service, University of Barcelona, Barcelona, Spain) for help in editing the manuscript.

The authors have no financial conflict of interest.