Pancreatitis-associated protein I (PAP I), an acute phase protein induced by cytokines. Identification of two functional interleukin-6 response elements in the rat PAP I promoter region.

Expression of the pancreatitis-associated protein I (PAP I), an exocrine pancreatic protein, increases rapidly and strongly in acinar cells during the acute phase of pancreatitis. This is reminiscent of the response to stress of acute phase proteins. We have previously demonstrated that serum factors from rats with acute pancreatitis, but not from healthy rats, could induce endogenous PAP I gene expression in the acinar cell line AR-42J (Dusetti, N., Mallo, G., Dagorn, J.-C., Iovanna, J. L. (1994) Biochem. Biophys. Res. Commun. 204, 238-243). In the present work, we have evaluated the influence of several mediators of inflammation on rat PAP I gene transcription in these cells. Tumor necrosis factor α induced an increase in PAP I mRNA expression, and interferon caused an even greater increase in PAP I mRNA level. These stimulations were antagonized by dexamethasone. Interleukin (IL)-1, IL-6, or dexamethasone alone were ineffective. Combinations of IL-1 with IL-6 or dexamethasone were also ineffective. IL-6 and dexamethasone together induced a marked stimulation of PAP I gene transcription, and this effect was slightly attenuated by IL-1. To analyze the cis-regulatory elements responsible for the induction of transcription, we fused a 1.2-kilobase segment of the rat PAP I promoter to the chloramphenicol acetyltransferase (CAT) gene as reporter. The resultant chimeric DNA was transfected into AR-42J cells. Addition of IL-6 or dexamethasone was ineffective, whereas their mixture increased the CAT activity 12 times. Progressive deletions of the PAP I promoter were then fused to the CAT gene, and the constructs were transfected to AR-42J cells. A 12-fold increase in CAT activity was seen upon IL-6/dexamethasone treatment with constructs containing more than 274 base pairs upstream from the cap site. In that region, two sequences are similar to the canonical IL-6 response element. Site-directed mutagenesis of these regions strongly decreased induction, showing that they were functional. PAP I should therefore be classified among acute phase proteins of class 2, whose expression is increased by IL-6 acting in combination with glucocorticoids.

The acute phase of pancreatitis is characterized by a pattern of changes in the expression of secretory proteins (1). Whereas expression of most pancreatic enzymes decreases, mRNA levels of the rat pancreatitis-associated protein (PAP) 1 increase dramatically. Recently, we have described two other PAP-related mRNAs and named the corresponding proteins PAP II and PAP III (2,3). In consequence, the original PAP became PAP I. Like PAP I, PAP II and III are induced in pancreas during the acute phase of pancreatitis. The sequences of the genes encoding rat PAP I, II, and III have been recently determined (2,4,5). All three genes are organized in six exons, and similarities observed in their coding sequences extend to their 5Ј-flanking regions. In addition, the three genes have been located to the same position on chromosome 4q33-34 (6), suggesting that they derived from the same ancestral gene by gene duplication.
In fact PAP I was not detectable in the pancreas of healthy animals. It could be evidenced in pancreatic juice 6 h after induction of an experimental acute pancreatitis, reached a maximum during the acute phase (12-48 h), and disappeared during recovery (7). The rapid and strong induction of the PAPs is unique among secretory proteins and reminiscent of the response to stress of acute phase proteins. Recently, we have demonstrated the presence of factors in serum from rats with acute pancreatitis, but not from healthy rats, capable of inducing PAP I gene expression in the pancreatic acinar cell line AR-42J. In addition, the cis-acting element was localized within the 1.2 kilobases upstream region of the transcription start site (8). It has long been known that the changes occurring in the liver and in other organs during the acute phase response are coordinated by signals generated at the site of injury, among which several cytokines have been well characterized, including IL-1, IL-6, TNF␣, IFN␥, leukemia inhibitory factor, IL-11, and oncostatin M. These proteins are locally produced by the tissue and by circulating mononuclear cells in response to prototype inflammatory stimuli and can elicit the diverse biological effects characteristic of the acute phase response. Interestingly, during the acute phase of pancreatitis, levels of cytokines are strongly increased in serum (9). In the current study, we have evaluated the respective contributions of several cytokines and of dexamethasone to the transcriptional induction of the rat PAP I gene in vitro, using a rat pancreatic acinar cell line.

EXPERIMENTAL PROCEDURES
AR-42J Stimulation-AR-42J pancreatic acinar cells were obtained from Dr. A. Estival (INSERM U151, Toulouse, France) and used after 42-49 passages. The cells were routinely cultivated at 37°C in a 5% * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18  CO 2 , 95% air atmosphere in Dulbecco's modified Eagle's medium containing 10% (v/v) fetal calf serum (Life Technologies, Inc.), 4 mM Lglutamine, 50 units/ml penicillin, and 50 mg/ml streptomycin. The cells were seeded at 3 ϫ 10 6 /100-mm Petri dish. When cells reached 80 -90% confluence (which took approximately 1 week), they were dissociated with 0.05% trypsin and 0.02% EDTA in Puck's saline A and replated.  For the stimulation experiments, cultures were incubated with either  control medium, IL-1, IL-6, IFN␥, TNF␣, dexamethasone, or a combination thereof. After 48 h, the medium was removed and replaced with fresh medium containing the indicated amount of the stimulant. Twenty-four hours later, total cellular RNA was prepared by the acidic guanidinium thiocyanate-phenol-chloroform extraction method (10). Total cellular RNA (15 g/lane) was fractionated by electrophoresis on a 1.0% agarose-formaldehyde gel and transferred onto nylon filters (Hybond). Filters were then hybridized with 32 P-labeled probes for rat PAP I (7) and ␤-actin (8). Filters were then washed extensively and autoradiographed.
CAT Reporter Gene Constructs-All DNA constructs were generated by polymerase chain reaction using the plasmid P/P as a template (4). That plasmid is a pBluescript KS ϩ in which was subcloned a 2859-base pair PstI-PstI genomic DNA fragment containing the PAP I gene, including 1253 nucleotides of the promoter. Accuracy of polymerase chain reaction was increased by using low dNTP concentrations (11), 100 ng of DNA plasmid as template, and only eight cycles of DNA amplification. Amplification was performed in 1 ϫ polymerase chain reaction buffer (50 mM KCl, 10 mM Tris-HCl, pH 8.3, 2 mM MgCl 2 , and 0.01% gelatin) containing 2 M dNTP, 1% Me 2 SO, 25 pmol of each primer, and 2.5 units of Taq polymerase in a final volume of 50 l. The reaction times were as follows: first cycle, denaturation at 94°C for 2 min, annealing at 55°C for 2 min, and extension at 74°C for 2 min; for the next 6 cycles, denaturation at 94°C for 10 s, annealing at 55°C for 2 min, and extension at 74°C for 2 min; for the last cycle, denaturation at 94°C for 10 s, annealing at 55°C for 2 min, and extension at 74°C for 10 min. The product was blunt-ended with Klenow polymerase, kinased, and ligated into the SalI site of the promoterless vector pCAT-Basic (Promega) to generate the plasmids pϪ1253/ϩ10PAPI-CAT, pϪ926/ϩ10PAPI-CAT, pϪ685/ϩ10PAPI-CAT, pϪ444/ϩ10PAPI-CAT, pϪ379/ϩ10PAPI-CAT, pϪ317/ϩ10PAPI-CAT, pϪ274/ϩ10PAPI-CAT, pϪ180/ϩ10PAPI-CAT, pϪ118/ϩ10PAPI-CAT, pϪ61/ϩ10PAPI-CAT, and pϩ10/-1253PAPI-CAT. Mutant plasmids were also constructed, to monitor the function of two sequences in the PAP I promoter corresponding to potential IL-6 response elements. Plasmids pmut1-274/ ϩ10PAPI-CAT, pmut2-274/ϩ10PAPI-CAT, and pmut3-274/ϩ10PAPI-CAT were generated, in which IL-6RE-1, IL-6RE-2 or both were modified, respectively (Fig. 5). The following oligonucleotides were used: 5Ј-CCTTGTGTCGTTAGAAACAGAGTATCTGGAAAAGGGTG-TGGAGGGTTCAAAC-3Ј, 5Ј-CCTTGTGTTTCCCAGAACAGAGTAT-AGTAGGCAGGGTGTGGAGGGTTCAAAC-3Ј, and 5Ј-CCTTGTGTC-GTTAGAAACAGAGTATAGTAGGCAGGGTGTGGAGGGTTCAAAC-3Ј, the underlined sequences corresponding to modified regions. The oligonucleotide used in opposite orientation was 5Ј-TGGATGGTTTGTGAG-GACAGA-3Ј in all cases. Numbers in plasmid names refer to the positions of first and last nucleotides of the insert in the PAP I gene. Plasmid DNA was purified with the Qiagen plasmid Kit (Diagen, Hilden, Germany) and the DNA concentration measured spectrophotometrically. Sequences were verified by the chain termination method using the T7 sequencing kit (Pharmacia Biotech Inc.). Plasmids were also checked for purity, concentration, supercoiling, and restriction pattern by agarose gel electrophoresis.
Cell Transfection and CAT Assays-Fifty to 60% confluent AR-42J (100-mm Petri dish) were transfected using the calcium phosphate DNA coprecipitation method (12). The transfection mixture contained 20 g of test plasmid and 4 g of pCMV/␤-gal (Promega). Twelve hours after the addition of the DNA, cells were subjected to 20% (v/v) glycerol for 2 min, and the cells were washed with serum-free medium and transferred to serum-containing medium. Thirty-six hours later each dish was treated with a different stimulant for 24 h. Each construct was monitored in triplicate. In all cases, at least two separate plasmid preparations were tested in the transfection experiments. Cells were then harvested, and extracts were prepared. CAT activity was determined using a phase extraction procedure (13).

RESULTS
Effect of IL-1, IL-6, IFN␥, TNF␣, and Dexamethasone on AR-42J PAP I mRNA Levels-AR-42J is a pancreatic tumor cell line derived from an azaserine tumor of the rat exocrine pancreas (14), which has retained most characteristics of the aci-nar cells. In these cells, the basal level of the PAP I transcript is extremely low, as in normal pancreas (4). We have monitored PAP I mRNA expression following treatment with several cytokines and dexamethasone. As shown in Fig. 1, treatments with IL-1 (50 and 500 units/ml), IL-6 (100 and 1000 units/ml), or dexamethasone (10 and 100 nM) alone were ineffective. Treatment with TNF␣ (500, 1000, and 5000 units/ml) or IFN␥ (100, 500, and 1000 units/ml) induced weak PAP I mRNA expression by comparison with untreated cells. However, dexamethasone clearly inhibited their stimulatory effects. Combinations of IL-1 with IL-6 or dexamethasone were also ineffective. By contrast, combination of IL-6 with dexamethasone induced a strong stimulation of PAP I gene transcription and subsequent mRNA accumulation. Surprisingly, when IL-1 was added together with IL-6 and dexamethasone, the induction was partially inhibited. The extremely low basal level of PAP I expression in AR-42J cells and the strong signal observed in Northern blots upon stimulation suggest that PAP I gene expression in these cells is primarily controlled by transcriptional regulation. Differences in ␤-actin mRNA concentrations were mainly due to differences in total RNA on the filters, estimated from methylene blue coloration.
Analysis by Progessive Deletion of Sequences Required for IL-6/Dexamethasone Induction of PAP I Gene-The presence of a functional promoter and tissue-specific elements in the 5Ј-flanking region of the rat PAP I gene was tested by transient expression assays. A 1253-base pair fragment containing the PAP I 5Ј region was fused to the bacterial gene coding for chloramphenicol acetyltransferase (CAT) whose expression can be easily monitored in transfected cells. Transfection experiments showed that pϪ1253/ϩ10PAPI-CAT was able to promote basal transcription in AR-42J cells (8). Fig. 2 shows the comparison of CAT activity in extracts from AR-42J cells transfected with the above constructs, then stimulated with IL-6, dexamethasone, or IL-6 and dexamethasone. Treatments with IL-6/dexamethasone increased CAT activity 12 times, whereas IL-6 or dexamethasone alone were ineffective. In order to identify the regions necessary for IL-6/dexamethasone induction, progressive deletions of the PAP I promoter were performed, and the resulting constructs were transfected into AR-42J (see "Experimental Procedures"). As shown on Fig. 3, deletions in the 5Ј to 3Ј direction resulted in a stepwise decrease of CAT gene expression in the AR-42J cell line. Deletion down to position Ϫ926 did not alter significantly the expression of the reporter gene. Deletion to nucleotide Ϫ685 resulted in about 30% decrease in expression. Progressive deletion of the next 368 nucleotide (to position Ϫ317) did not alter CAT activity further. An additional deletion to nucleotide Ϫ274 caused a decrease to 40% of the control. Extending deletion to nucleotide Ϫ180 caused a decrease to 20% of the control, and a further deletion of 62 base pairs (to position Ϫ118) resulted in a reduction of the CAT activity to about 10% of control. Finally, deletion down to nucleotide Ϫ61 further reduced activity about 3 times, although it remained slightly above background.
A 12-fold increase in CAT activity was seen upon IL-6/dexamethasone treatment of cells transfected with constructs containing more than 274 base pairs of 5Ј-flanking sequence (Fig.  4). Deletion to position Ϫ180 led to a 3-4-fold drop in induction. Finally, a 2-fold induction was observed when we transfected with pϪ118/ϩ10PAPI-CAT and pϪ61/ϩ10PAPI-CAT constructs but not with pϩ10/-1253PAPI-CAT.

Mutation Analysis Reveals Two Functional IL-6 Response Elements (IL-6RE) in the PAP I Promoter Region-Computer-
assisted search for sequences similar to previously described IL-6RE showed two potential regions within the PAP I promoter, at positions Ϫ266 to Ϫ260 (in antisense orientation) and Ϫ249 to Ϫ243 (in sense orientation). To analyze these sequences, site-directed mutagenesis of the regions from Ϫ266 to Ϫ260 and Ϫ249 to Ϫ243 was conducted within the CAT construct pϪ274/ϩ10PAPI-CAT and the influence of the mutations on the amplitude of CAT induction by IL-6/dexamethasone was monitored. With mutant pmut1-274/ϩ10PAPI-CAT, in which the sequence from Ϫ266 to Ϫ260 was modified, induc-  Fig. 3. Thirty-six hours after transfection, AR-42J cells were incubated with medium alone (control) or with IL-6 (100 units/ml) in association with dexamethasone (100 nM). Specific induction by IL-6/dexamethasone was calculated as the ratio of the values from induced and control cells. Values represent the means (Ϯ standard error) of four to seven independent transfection experiments. tion was reduced to 22% of control (Fig. 5). Similarly, replacement of the sequence from Ϫ249 to Ϫ243 (mutant pmut2-274/ ϩ10PAPI-CAT) reduced the induction to 35% of control. Induction with mutant pmut3-274/ϩ10PAPI-CAT, which carried both mutations, was further reduced compared to pmut1-274/ϩ10PAPI-CAT or pmut2-274/ϩ10PAPI-CAT constructs. Hence, hexanucleotides TTCCCAG and CTGGAAA are actually involved in the response of the promoter to IL-6/dexamethasone. However, stimulation could not be completely abolished by mutation of the two IL-6REs. The remaining induction (about 2 times) suggests that additional active cis-regulatory elements are present within the Ϫ274/ϩ10 region of the PAP I promoter.

DISCUSSION
Induction of an experimental pancreatitis causes a more than 200-fold increase in PAP I mRNA expression during the acute phase of pancreatitis (7). PAP I mRNA accumulation reaches a maximum 6 h after induction, with a kinetics probably controlled by the cascade of events taking place during the acute phase. That cascade includes the activation of monocytes and macrophages and the synthesis and secretion of inflammatory mediators eventually transported to the target cells. In support of that hypothesis, we have recently demonstrated the presence of factors in serum from rats with pancreatitis, but not from healthy rats, capable to induce PAP I gene expression (8). The present work was carried out primarily to localize the cis-regulatory elements in the PAP I gene and to characterize their response to several acute phase mediators including IL-1, IL-6, IFN␥, TNF␣, and dexamethasone.
The cytokines tested in this study had very different effects on PAP I gene expression in AR-42J cells (Fig. 1). The most striking result was the strong stimulation of the association IL-6/dexamethasone, and the limited stimulation by IFN␥ or TNF␣, compared to the absence of effects of IL-1 or IL-6. Another intriguing finding was the inhibition by IL-1 of IL-6/ dexamethasone stimulation. However, a growing number of reports show that expression of acute phase protein genes is not always mediated by single cytokines but by combinations of several cytokines (15)(16)(17) or by cytokines in association with cofactors such as glucocorticoids (16). It was also shown that one cytokine may modulate the effect of other cytokines (17,18). These findings suggest that specific responses of a cell to various inflammatory stimuli are mediated by specific combinations of cytokines and/or glucocorticoids. Although an important regulatory function during the acute phase reaction has been attributed to glucocorticoids and IL-6 (17), at the utilized doses, dexamethasone and IL-6 alone were unable to induce PAP I gene expression in AR-42J cells (Figs. 1 and 2). Then, two mechanisms may account for the synergy between IL-6 and dexamethasone. First, glucocorticoid and IL-6 response elements might be localized in close vicinity on the PAP I promoter. In that instance, interaction of the nuclear factors binding the two transcription activator elements might enhance individual responses that would be otherwise too weak to be observed. Such a situation was reported for the ␣ 1 -acid glycoprotein (19). Second, dexamethasone could stimulate IL-6 receptor synthesis in AR-42J cells, resulting in an increased number of IL-6 receptors at the cell surface. This was already observed in hepatic cells (20,21). That mechanism might apply to the PAP I gene since, in the hepatic carcinoma cell line HepG2 transfected with pϪ1253/ϩ10PAPI-CAT, CAT activity could be strongly induced (25-fold) by IL-6 alone (data not shown). Hence, the IL-6 enhancer element of the PAP I promoter does not require glucocorticoids to be active. More studies are, however, necessary to understand the synergistic effect of IL-6 and dexamethasone on the PAP I gene induction in AR-42J cells.
The mechanism by which IL-1 down-regulates the stimulation by IL-6 associated with dexamethasone is also unknown. IL-1 has already been shown to inhibit IL-6 induction of the endogenous T kininogen in rat primary hepatocytes (22), but IL-1 and IL-6 can also act independently (additive effect) or synergistically in the regulation of other acute phase genes such as ␣ 1 -acid glycoprotein, haptoglobin, hemopexin, complement C 3 , and serum amyloid A (17). Therefore, relative positions of the different enhancer sequences are likely to influence the effect of cytokines acting in combination. A mechanism involving interaction of IL-1 with expression of the IL-6 receptor, as suggested for dexamethasone, cannot be ruled out, although such regulation has never been reported in other systems. However, inhibition by interaction of IL-1 with the IL-6 receptor is unlikely because the two cytokines have their own specific membrane receptors.
PAP I gene expression was significantly induced by IFN␥ or TNF␣, although 100-fold less than with IL-6 and dexamethasone. Other genes induced by TNF␣ are also induced by IFN␥ (23)(24)(25). This may be due to the ability of TNF␣ and IFN␥ to activate the same transcription factors, such as interferon regulatory factors 1 and 2 (26,27). A similar PAP I mRNA induction was obtained with 100, 500, or 1000 units/ml IFN␥, but an inhibitory effect was observed when we incubated the cells in presence of more than 500 units/ml TNF␣ (Fig. 1), suggesting a toxic effect of this cytokine. Addition of dexamethasone to these cytokines inhibited induction, as already reported for other genes (28 -34). Again, the opposite effect of dexamethasone on IL-6 and TNF␣ or IFN␥ underscores that understanding the mechanism of effector action requires a detailed topological analysis of promoter sequences.
We have chosen to address in this study the molecular mechanism of IL-6 and dexamethasone stimulation of the PAP I promoter. Analysis of the promoter sequence revealed the presence in two positions of the potential IL-6 response element of type 2 (CTGGGA), previously identified in several acute-phase genes (17,18,35) and shown to be functional by mutation analysis (36). Demonstration that the two IL-6REs identified in the PAP I promoter were indeed functional was obtained by mutation and transfection assays (Fig. 5). However, these are not the only cis-elements involved in the IL-6/dexamethasone response of PAP I. Another cis-cytokine response element, localized between Ϫ61 and ϩ10, is responsible for a 2-fold induction. This has been shown previously in other IL-6-activated cellular genes (37). For instance, Baumann et al. (38) have demonstrated for several acute phase proteins that IL-6 acts directly through an IL-6RE but also indirectly by increasing expression of C/EBPs, which in turn stimulates acute phase proteins gene expression.
Acute phase proteins have been divided into two subclasses according to their pattern of regulation by cytokines (17). The synthesis of class 1 acute phase proteins (e.g. ␣ 1 -acid glycoprotein, C-reactive protein, haptoglobin, and serum amyloid A) is induced by IL-1 or combinations of IL-1 and IL-6, whereas the genes for class 2 acute phase proteins (e.g. ␣ 2 -macroglobulin, ␣ 1 -antichymotrypsin, and fibrinogen) are mainly regulated by IL-6 and glucocorticoids. PAP I is therefore an additional member of the second group of acute phase proteins, with the original feature of being a secretory protein.
Finally, it is interesting to note that whereas the PAP I gene is expressed as an acute phase protein in pancreas, it is constitutively expressed by the epithelial cells of the intestinal tract (39,40). The PAP I promoter is therefore complex. It confers to the gene the capacity of being regulated along several pathways, the switch between pathways being possibly under the control of tissue-specific elements.