Serotonin-mediated Production of Interstitial Collagenase by Uterine Smooth Muscle Cells Requires Interleukin-1α, but not Interleukin-1β*

The activation of the gene for interstitial collagenase in myometrial smooth muscle cells is absolutely dependent upon the presence of serotonin. Our previous studies investigating the mechanisms of this induction demonstrated that the mRNAs of both interleukin-1 (IL-1) isoforms, IL-1α and IL-1β, are induced by serotonin and that the induction of IL-1 is required for the subsequent induction of collagenase. These data provided compelling evidence that serotonin-induced IL-1 acts via an autocrine loop in activating the collagenase gene. The experiments described here were designed to examine the potential role of each IL-1 isoform in collagenase production by using neutralizing antisera specific to each isoform of the cytokine. The antisera were examined for their ability to inhibit the serotonin-dependent production of the mRNA for collagenase and of the cytokines themselves. Neutralizing antiserum against IL-1α, but not against IL-1β, inhibited the induction of the mRNA for collagenase and of the mRNAs for both IL-1α and IL-1β. Western analysis indicated that detectable levels of IL-1α protein, but not that of IL-1β, are produced at the time of serotonin-dependent collagenase induction. In contrast, significant levels of IL-1β protein are detected only when bacterial lipopolysaccharide is added to the cells. Taken together, the results of our study indicate that IL-1α, but not IL-1β, plays an obligatory role in multiple serotonin-mediated gene regulations in the myometrial smooth muscle cell. In addition, the data suggest that IL-1β production has the potential for modifying myometrial function in pathological settings, particularly that of uterine infection.

After the delivery of its fetal contents is complete, the mammalian uterus undergoes a remarkable involution, in which all of the collagen that has accumulated during pregnancy is rapidly degraded and removed from the tissue (1). The degradation of fibrillar collagens (Types I, II, and III) is mediated by interstitial collagenase, which catalyzes the initial cleavage in the collagen molecule (2). This cleavage allows further catabolism of the protein by other proteases. We have shown that interstitial collagenase is produced by the myometrial smooth mus-cle cells during the post-partum period, and not during pregnancy (3). Given its crucial importance in uterine collagen degradation, regulation of the expression of this enzyme and its activity is of major importance to our understanding of the biology of the uterus.
We have developed an in vitro system of primary myometrial smooth muscle cells to study the induction and regulation of interstitial collagenase (4). In this cell culture system, we have shown that serotonin (5-hydroxytryptamine; 5-HT), acting as a hormone, is an obligate inducer of interstitial collagenase (5), and that this induction represents an activation of the gene for the enzyme (6). This hormonal requirement for serotonin appears to be selective to the smooth muscle cells of the rat and the human (5). Of equal importance, we have shown that serotonin also induces the genes for interleukin-1␣, interleukin-1␤ (7), and of interleukin-6. 1 These serotonin-dependent inductions occur after the cells have been confluent for approximately 13 days, and we refer to these cells as "spontaneously active." Until this time, even though the cells are exposed to a concentration of serotonin (ϳ5 M) sufficient to saturate all known 5-HT receptors, none of these transcripts, nor collagenase protein, can be detected; we refer to these cells as "early cells." Early cells, however, do produce collagenase when bacterial lipopolysaccharide (LPS) 2 is added to the medium, again only in the presence of serotonin (7). One effect of LPS in these early cells is the induction of the mRNAs for both interleukin-1␣ and -␤. This finding led us to investigate the role of these cytokines in spontaneously activated cells. Using the IL-1 receptor antagonist (IL-1ra) in the spontaneously active cell setting, we demonstrated that IL-1 is required for the induction of interstitial collagenase, as well as of its own mRNA (7), and that of both IL-6 mRNA and protein. 1 These results indicate that IL-1, acting via an autocrine/paracrine mechanism, is an obligate intermediate in the serotonin-mediated production of collagenase.
Since IL-1␣ and IL-1␤ mRNAs were induced by serotonin both in early and in spontaneously active cells, we wished to examine whether one or both isoforms of interleukin-1 are involved in these serotonin-dependent inductions. Interleukin-1 (IL-1) is a term used to describe two related polypeptides: IL-1␣ and IL-1␤ (8,9). Both isoforms are synthesized from different genes as 31-kDa precursors. These precursors can then be cleaved by proteases to the mature form of 17 kDa. This cleavage is required for IL-1␤ to have biological activity and occurs by the IL-1 converting enzyme. In contrast, IL-1␣ has been shown to have biological activity both as the precursor and as a "mature" form, apparently resulting from a calpain cleavage (for a recent comprehensive review, see Ref. 10). Even though the two isoforms share a low amino acid homology, ranging between 20 and 30% (8) , both IL-1␣ and IL-1␤ exert  their biological activity via the type I IL-1 receptor (IL-1R), (9). Additionally, the two proteins have been shown to have very similar biological activities in many cell systems (11)(12)(13)(14), one of which has been the ability of both forms of the cytokine to induce interstitial collagenase and other matrix metalloproteinases in a variety of mesenchymal cells (10). Thus, a continuing biological conundrum has existed as to why two isoforms of IL-1 are produced, both of which appear to possess similar if not identical activity. Recently, several investigations have suggested that IL-1␣ is specifically important in a number of biological settings (14 -16). Of particular interest, a study by West-Mays et al. (15), using rabbit corneal fibroblasts reported that the production of collagenase upon stimulation with phorbol esters has also been shown to require the establishment of an IL-1␣ autocrine loop for the induction of interstitial collagenase. In this study, however, the corneal cells do not produce IL-1␤ mRNA or protein, and accordingly the cells have no choice in cytokine isoform usage (17). In our own studies (7), interleukin-1 has been shown to be a critical intermediate, acting by way of an autocrine loop pathway, in the serotonininduced gene activation of interstitial collagenase in myometrial smooth muscle cells. Furthermore, the temporal relationships between the regulation of IL-1␣ and the induction of interstitial collagenase are much more congruent than the analogous relationships between IL-1␤ and collagenase inductions.
In the studies reported here, the differences in the potential of the two IL-1 isoforms to regulate serotonin-mediated induction of IL-1 and collagenase have been examined using isoformspecific antibodies. Mature rat and mouse IL-1␣ and IL-1␤ possess 78 and 89% identity, respectively (8,9). Thus, polyclonal isoform-specific anti-mouse IL-1 antibodies, which were selected for their neutralizing ability, were utilized. The mouse IL-1␤ antibody has been shown to inhibit the pyrogenic activity of rat IL-1␤ in vivo (18), whereas the IL-1␣ antibody, although only used in mouse studies in vivo (19), has been shown to interact with rat IL-1␣ by Western blot analysis. 3 Also, commercially available rabbit polyclonal, isoform-specific antibodies to rat IL-1 were utilized in parallel with the antibodies described above. The results of experiments using both these neutralizing anti-mouse IL-1 antibodies, as well as commercially obtained anti-rat IL-1 antibodies indicate that, although both IL-1 mRNAs are induced in vitro by the myometrial smooth muscle cells, only IL-1␣ is involved in serotonin-dependent collagenase production.

MATERIALS AND METHODS
Cell Cultures-Primary cultures of rat myometrial smooth muscle cells were prepared as described previously (20). Uteri were obtained from 4-day post-partum Sprague-Dawley rats, minced, and washed free of blood. Uterine tissue was then incubated in 0.05% trypsin in Dulbecco's modified Eagle's medium containing 30 mM HEPES, 200 units/ml penicillin, and 200 g/ml streptomycin for 2 h at 37°C. Tissue fragments were dissociated by forcing the tissue fragments through a finebore 10-ml pipette. Tissue was then centrifuged at 50 ϫ g to remove debris. The remaining supernatant was centrifuged at 800 ϫ g to pellet the cells. Cells were resuspended in Dulbecco's modified Eagle's medium containing 10% fetal bovine serum (FBS) that had been adsorbed with dextran-coated charcoal to remove serotonin (CS-FBS (5, 21)), 30 mM HEPES, 200 units/ml penicillin, and 200 g/ml streptomycin. For all experiments, cells were plated at approximately 100,000 cells per square cm of surface area in 75-cm 2 culture flasks and allowed to reach confluence, at which time 5 M serotonin was added to the culture medium until spontaneous up-regulation of collagenase occurred (7). Other additions to medium are described in detail for each experiment.
Neutralizing Antibodies-Isoform-specific neutralizing antisera to mouse interleukin-1 were produced as described previously (18,19). The antibody against mouse IL-1␣ was generated in goats. The antibody against mouse IL-1␤ was produced in rabbits. Isoform-specific rabbit anti-rat antibodies for IL-1␣ and IL-1␤ were obtained from Endogen (Cambridge, MA), as was recombinant rat IL-1␣. Recombinant murine IL-1␣ and IL-1␤ proteins were obtained from R & D Systems (Minneapolis, MN), which also provided the recombinant rat IL-1␤. The isoform specific antisera and antibodies were added directly to cell cultures and allowed to preincubate for 15 min prior to addition of 5 M serotonin. As a control for the antisera, rabbit non-immune serum was utilized. Cells were also cultured in the presence of 1:500 dilution of goat serum. This addition to the cell cultures had no effect on collagenase production. Statistical significance of the results was determined using a single variable ANOVA. The data presented are compiled from experiments in which the dilutions of antisera varied from 1:200 to 1:500.
RNA Isolation and Northern Blot Analysis-Poly(A ϩ ) RNA was isolated from primary cultures of myometrial smooth muscle cells. Cells were lysed and mRNA isolated directly from cell extracts by modification of a commercial kit (FastTrack™ kit, Invitrogen, San Diego, CA). Poly(A ϩ ) RNA samples were denatured with formaldehyde and formamide and electrophoresed on 1.3% agarose gels containing 0.66 M formaldehyde for 1 to 2 h as described by Davis et al. (22). Poly(A ϩ ) RNA was transferred overnight to nitrocellulose filter by capillary blotting in 10 ϫ SSC which contains 1.5 M NaCl, 0.15 M trisodium citrate. The blot is baked at 80°C for 3 h and then prehybridized for 30 min in 4 ϫ SSC, 50% deionized formamide, 20 mM Tris-HCl (pH 8), 10% dextran sulfate, 200 g/ml denatured salmon sperm DNA, and 1 ϫ Denhardt's solution which consists of 0.02% Ficoll, 0.02% polyvinylpyrrolidone, 0.02% bovine serum albumin. cDNA probes were radiolabeled with [␣-32 P]dCTP (3000 Ci/mmol) using the Random Primer Plus Extension Labeling System (NEN Life Science products Inc., Boston, MA). Autoradiographic signals were quantified by video capture and digital image processing using the BioImage system (BioImage/Millipore, Ann Arbor, MI). In all cases, the blots were simultaneously probed with a cDNA to mouse ␤-actin to provide controls for RNA loading. The variation in actin levels was always less than 15%; nevertheless, all values for specific hybridization were normalized to the values for actin.
cDNA Probes-The interleukin-1␣ probe is a 2-kb cDNA to mouse IL-1␣ cloned from mouse DBA/2 P388D1 cells (pmIL1AcDNA; Lomedico et al. (23)). It was obtained through the American Type Culture Collection, Rockville, MD. The interleukin-1␤ probe is a 0.3-kb cDNA to rat IL-1␤ and was a generous gift of Dr. Richard Rohan, University of Maryland School of Medicine. The cDNA probe for rat collagenase is a 2.6-kb clone isolated from a rat osteoblastic cell cDNA library and contains the entire coding region for the active enzyme (UMRcase 54: Quinn et al. (24)).
Collagenase Assay-Conditioned medium harvested from cell cultures was collected, precipitated by ammonium sulfate to 60% saturation, and the precipitate resuspended in 200 l of 0.05 M Tris and 0.01 M calcium chloride and dialyzed overnight against the same buffer. Aliquots (100 l) were then assayed on reconstituted [ 14 C]glycine-labeled guinea pig skin type I collagen fibrils as described previously (25,26). Briefly, the pro-enzyme was activated by treatment with trypsin at 25°C for 10 min. Trypsin action was then terminated by the addition of a 4-fold excess of soybean trypsin inhibitor for 15 min, and the resultant active enzyme was added to the collagen fibrils and incubated at 37°C. After appropriate substrate digestion had occurred, 250 l of 0.05 M Tris and 0.01 M calcium chloride was added and the samples were centrifuged at 10,000 rpm for 10 min. The radioactivity in the supernatants was determined by scintillation counting. The counts from solubilized collagen are proportional to the collagenase concentration in the medium (25).
Western Blot Analysis-For titer determination, Western analysis according to the method of Towbin et al. (27) was employed. Purified IL-1 was analyzed for isoform specificity using goat anti-mouse IL-1␣ antisera and rabbit anti-mouse IL-1␤ antisera. The purified proteins were electrophoresed on 12% SDS-polyacrylamide gels and transferred to nitrocellulose membranes (Bio-Rad). Membranes were incubated overnight in 1:500 dilution of primary antisera. Alkaline phosphataseconjugated rabbit anti-goat IgG (Sigma) or goat anti-rabbit IgG (Sigma) was used as secondary antibody at a working dilution of 1:1000 and developed using an alkaline phosphatase substrate kit (Bio-Rad). For detection of IL-1 in the myometrial cells, cultures were rinsed 3 times 3 I. Otterness, unpublished data.
with phosphate-buffered saline. Cells from a single flask were then scraped into 300 l of phosphate-buffered saline containing 2 mM EDTA and protease inhibitors (Complete, Mini, Boehringer Mannheim, Mannheim, Germany), subjected to 4 freeze-thaw cycles, and centrifuged at 20,000 ϫ g for 10 min at 4°C. Supernatants were recovered and protein concentrations determined by the method of Smith et al. (28). Approximately 50 g of cell lysate protein were electrophoresed on 15% SDSpolyacrylamide gels and transferred to nitrocellulose membranes (Bio-Rad). Membranes were incubated overnight at 4°C with a 1:200 dilution of polyclonal rabbit anti-rat IL-1␣ or polyclonal rabbit anti-rat IL-1␤ (Endogen, Cambridge, MA). A 1:2500 dilution of alkaline phosphatase-conjugated goat anti-rabbit IgG (Sigma) was used as a secondary antibody. Membranes were washed and developed with an alkaline phosphatase substrate kit (Bio-Rad).

RESULTS
Effect of Neutralizing IL-1 Antisera on Collagenase mRNA and Collagenase Activity Levels-In all experiments, rat myometrial smooth muscle cells were grown to confluence in the absence of serotonin. On day 7 of culture, cells were transferred to medium supplemented with 5 M serotonin until spontaneous production of collagenase began, as confirmed by enzymatic analysis (cf. "Materials and Methods"). At that time, cells were transferred to serotonin-free medium for 4 days, at which time interstitial collagenase production was reduced to undetectable levels. The flasks were then randomly divided and treated as shown in Fig. 1, A and B, for 24 h. The data in Fig. 1, A and B, represent those obtained from five separate experiments. In all cases, cells were preincubated with antiserum for 15 min prior to the addition of serotonin. Control cells were preincubated with the appropriate non-immune serum at the same dilution of neutralizing antibody used. After 24 h, cells were harvested for mRNA isolation and Northern analysis (A). Conditioned medium was also assayed for collagenolytic activity (B). Cells cultured in the absence of serotonin produced basal levels of collagenase mRNA and activity. Addition of serotonin resulted in the expected massive increase in the production of interstitial collagenase and its mRNA. The serotonin-dependent induction of collagenase mRNA was essentially completely inhibited by addition of antiserum to IL-1␣. Similar decreases in the enzymatic activity of collagenase in the culture medium were also observed with treatment of the anti-IL-1␣ antiserum. Conversely, antiserum to IL-1␤ did not affect the serotonin-mediated induction of collagenase mRNA or enzymatic activity. Cells cultured in the presence of serotonin and both neutralizing antisera showed a decrease in the levels of collagenase mRNA and in collagenase activity essentially equal to those obtained with anti-IL-1␣ alone. The data presented in Fig. 1, A  and B The single experiment depicted in Fig. 1C represents the time course of the inhibition of collagenase mRNA production by the antisera to the two isoforms of interleukin-1. Cells cultured in the presence of the neutralizing antisera were harvested at 6, 12, and 24 h following incubation with each antiserum. Poly(A) ϩ RNA was isolated, and Northern analysis was performed. The cells that received medium containing serotonin and antiserum to IL-1␣ produced, at the very most, 25% of the maximal levels of serotonin-induced collagenase mRNA. In contrast, the induction of collagenase mRNA by cells cultured in serotonin and antiserum to IL-1␤ produced levels of collagenase mRNA essentially identical to those produced by cells cultured in serotonin only. Thus, a clear specificity for IL-1␣ appears to exist in the induction of collagenase mRNA. The neutralizing ability of anti-IL-1␣ has already been documented (18,19), but the ability of anti-IL-␤ to neutralize its cognate antigen has not been shown. Accordingly, we examined this issue, and the results of a typical experiments are shown in Fig. 2. This figure illustrates the ability of a 1:300 dilution of anti-IL-1␤ to inhibit the induction by IL-1␤ of interstitial collagenase mRNA production by cultured myometrial smooth muscle cells. Nearly complete inhibition was observed, indicating the efficacy of the antibody in neutralizing the activity of the cytokine.
Effect of Neutralizing Antisera on IL-1 mRNA Levels-Uterine smooth muscle cells were isolated from the rat and cultured in serotonin-free medium until confluence was attained. At this time 5 M serotonin was added to the medium, and the cells were cultured until they produced collagenase. At that point, the cells were transferred to serotonin-free medium for 4 days.
Following the depletion of serotonin, cells were treated with isoform-specific neutralizing anti-IL-1 antisera for 24 h. As shown in Fig. 3, cells cultured in the presence of serotonin resulted in the induction of IL-1␣ mRNA, which was completely blocked by incubation with the neutralizing antiserum to IL-1␣ or with the combination of both IL-1␣ and IL-1␤ antisera. Cells cultured in the presence of serotonin and antisera to IL-1␤ displayed only a modest, indeed marginally significant, inhibition of IL-1␣ mRNA induction. Cultures to which preimmune serum was added displayed no observable effect on IL-1␣ mRNA levels.
Essentially identical results were obtained from studies on the ability of isoform-specific antisera to modulate the levels of IL-1␤ mRNA (Fig. 4). Serotonin-induced increases in IL-1␤ mRNA levels were nearly completely inhibited by antiserum to IL-1␣. On the other hand, cells that received serotonin and antiserum to the ␤-isoform produced nearly normal levels of IL-1␤ mRNA. Thus, these studies suggest that IL-1␣ regulates not only the expression of its own mRNA, but that of IL-1␤ as well.
IL-1 Isoform Specificity of Neutralizing Antisera-To ensure the fidelity of antiserum specificity, Western blot analysis was performed, using purified mouse IL-1␣ and IL-1␤ protein and their cognate antibodies (Fig. 5). IL-1␣ amounts were 0.5 (lane detect the IL-1␣ antigen. Interestingly, the IL-1␤ antiserum could detect its homologous antigen at a much higher sensitivity than did the IL-1␣ antibody. Quantitative scanning indicates that this "titer," as determined by Western analysis for the IL-1␤ antiserum is approximately 10-fold higher than that of the IL-1␣ antiserum for its antigen. If the extent of these differences in sensitivity on Western blots is reflected in neutralizing potency, the modest effects of antiserum to IL-1␤ observed in these experiments should be considered as maximal and possibly nonspecific. In an attempt to confirm the results observed with the neutralizing anti-mouse antisera, anti-rat polyclonal antibodies, obtained from a commercial source (Endogen) were utilized. Although the potency of these antibodies was not as high as those of the antisera selected for neutralizing ability, the serotonin-mediated induction of IL-1␣, IL-1␤, and collagenase mRNA and collagenase activity were modestly inhibited by treatment with the anti-IL-1␣ antibody, consistent with the results obtained with the specific neutralizing antibodies, whereas no effect was ever observed with equivalent concentrations of anti-IL-1␤ (data not shown).
Immunochemical Detection of IL-1␣ and IL-1␤ in Cultured Myometrial Smooth Muscle Cells-Myometrial smooth muscle cells were grown to confluence and cultured in serotonin-free media supplemented with 100 ng/ml LPS, 5 M serotonin, or the combination of both. When cells were verified as producing collagenase, cell cultures were rinsed and cells from a single flask were then scraped into 300 l of phosphate-buffered saline containing 2 mM EDTA and protease inhibitors, subjected to 4 freeze-thaw cycles, and centrifuged as described under "Materials and Methods." Supernatants were recovered and approximately 50 g of cell lysate were examined by Western analysis for the presence of either IL-1␣ or IL-1␤ using rat isoform-specific antibodies. The results, gathered in Fig. 6, indicate that in lysates of cells spontaneously producing interstitial collagenase in the presence of serotonin, IL-1␣ could easily be detected, whereas immunoreactive IL-1␤ was present at very low levels, if at all. Interestingly, the form of IL-1␣ detected in these lysates is the unprocessed 31-kDa species, which has been shown to possess full biological activity (for review, see Ref. 10). On the other hand, significant levels of IL-1␤ protein were found to be induced in parallel with its mRNA, but only when LPS (100 ng/ml) was added to spontaneously active cultures of myometrial smooth muscle cells (Fig. 6A). In addition, under these conditions, only the non-processed, biologically inactive, 31-kDa form of IL-1␤ appears to be present and remains cellassociated. No signal was ever observed in the absence of serotonin. In addition, IL-1␤ was not detectable in the cell culture medium, although clearly seen in cell extracts, as shown in Fig. 6B. Thus, although serotonin is required for the induction of both IL-1␣ and IL-1␤ mRNAs, only IL-1␣ protein appears to be translated at detectable levels under conditions in which interstitial collagenase is induced. Furthermore, interleukin-1␤ protein, when induced as a consequence of the presence of bacterial endotoxin, is only observed in an unprocessed, inactive, cell-associated form. Taken together with the immunoneutralization studies, it appears that IL-1␣ is the requisite isoform of the cytokine for the serotonin-mediated induction of interstitial collagenase.

DISCUSSION
The results of this study demonstrate that IL-1␣ is a crucial intermediate for a number of serotonin-dependent inductions which occur in the myometrial smooth muscle cells. Primary cultures of rat myometrial smooth muscle cells incubated with antiserum specifically selected for its neutralizing ability to IL-1␣ resulted in a massive inhibition of collagenase mRNA levels and collagenase protein production. Addition of anti-IL-1␤ antiserum to cells failed to significantly affect collagenase mRNA or protein production. Furthermore, cells cultured in the presence of anti-IL-1␣ antiserum resulted in concomitant decreases in the mRNA levels for both IL-1␣ and IL-1␤. The neutralizing antiserum to IL-1␤, which has been previously documented to inhibit the in vivo activity of rat IL-1␤and confirmed in the present study to exhibit neutralizing capacity (Fig. 2), had, at most, only a slight inhibitory effect on collagenase mRNA or the mRNAs for either isoform of IL-1. In addition, we have shown in this study that the antiserum against IL-1␤ effectively neutralizes the ability of IL-1␤ to induce the production of collagenase by myometrial smooth muscle cells (Fig. 2). In addition, when the titers of these two antisera against their cognate antigens were determined by Western blot analysis, that of the IL-1␤ antiserum was found to be approximately 10 times higher than that of the IL-1␣ antibody. Assuming that the neutralizing capacity of the antisera parallel the titers indicated by Western blot, the effect of the IL-1␣ antiserum are, if anything, underestimated in the experiments described in this study.
In addition to the above studies, we have found that the appearance of IL-1␣ protein (Fig. 6) in our cultured myometrial smooth muscle cells is consistent with both the time course of the induction of interstitial collagenase and the inhibition of collagenase production by the isoform-specific antiserum directed to IL-1␣. Since the IL-1␤ antiserum had little or no effect, and because detection of the rat proteins was limited by the availability of reagents, we initially did not know if IL-1␤ is being produced by the myometrial smooth muscle cells. The immunochemical studies presented here now indicate that, indeed, IL-1␤ protein is not produced in our spontaneously active cell setting, even though mRNA levels are induced in these cells cultured with serotonin. Of considerable interest, however, is our ability to detect significant levels of IL-1␤ protein when LPS is added to the cell culture medium. It appears, however, that even under these conditions in the myometrial cell, IL-1␤ protein is produced but not processed or secreted (Fig. 6, A and B). Although at this point in our studies, no biological role for IL-1␤ can be identified, the notion that IL-1␤ is translated only under conditions of infection is of potential importance to the physiology and pathophysiology of pregnancy. Future studies will concentrate on determining the levels, biological activity of each protein, and to further examine the involvement of each isoform in the myometrial cells.
These data, taken as a whole, indicate that IL-1␣ is a crucial intermediate not only for the serotonin-dependent production of collagenase, but also in the regulation of both IL-1␣ and of IL-1␤ themselves in the myometrial smooth muscle cell. Preliminary evidence recently obtained in this laboratory 4 indicate that several other serotonin-induced gene inductions are similarly mediated by interleukin-1. These include the genes for IL-6 and for two other members of the matrix metalloproteinase gene family: transin (stromelysin) and 92-kDa gelatinase (gelatinase B). Thus, in our overall model, the activity of serotonin, mediated initially by the 5-HT 2A receptor subtype in myometrial cells (29), results in the activation of the gene for IL-1␣, after which the induced IL-1␣ then proceeds to act via either an autocrine or paracrine mechanism to maintain a continuous source of IL-1␣ (7). In our system, this interleukin-1 autocrine loop results in multiple gene inductions. The interleukin-1 autocrine loop was originally observed in human smooth muscle cells (30,31). This feedback mechanism of IL-1 production has now been identified in several other cell types. In a study by Loppnow and Libby (32), for example, IL-1 produced by vascular smooth muscle cells elicited an autocrine/ paracrine effect on IL-1 levels in naive smooth muscle cells, although no other biological end point was identified in this study.
One report, using rabbit corneal fibroblasts has shown an effect possibly analogous to that observed in our myometrial cell system (15). In this study, cells stimulated by either phorbol esters or cytochalasin B require the intermediate action of IL-1␣ in order for the cells to produce interstitial collagenase. A further similarity of this system to ours is that primary cultures of fibroblasts cannot sustain the IL-1␣ loop, whereas fibroblasts which have been passaged in culture do continuously produce IL-1␣ (15). This transition-dependent production is potentially analogous to the time-dependent up-regulation of interstitial collagenase, IL-1, and IL-6 that occurs in our spontaneously active myometrial cells. These studies further suggest that the establishment of one or more autocrine/paracrine mechanisms of IL-1 induction may be regulated by a time-dependent developmental process, as yet undefined, which is required to sustain the induction of the cytokine. Finally, a recent study (33) using human endometrial cells has shown that IL-1␣, but not IL-1␤ or a variety of other cytokines, is responsible for the induction of interstitial collagenase in endometrial fibroblasts. Thus, there appears to be an accumulating body of evidence implicating IL-1␣ as a critical mediator of the induction of a variety of genes in biological systems.
The myometrial smooth muscle system appears to be one of the emerging models in which only one IL-1 isoform, IL-1␣, is utilized to mediate an important cellular process, namely tissue remodeling. It is possible that the myometrial cells when activated with other stimuli, such as bacterial endotoxin in the example described here, may utilize IL-1␤ for a different purpose. In summary, this study suggests that IL-1␣ and IL-1␤ have different roles in myometrial smooth muscle cells stimulated with serotonin and argues to the fact that the two isoforms, which were once thought to possess essentially the same biological activity, will eventually be identified as separately important regulators of cellular activity.