INTRODUCTION

Interleukin (IL)¹ 6 is a multifunctional cytokine with diverse effects on inflammation, immune responses, and hematopoiesis. This cytokine stimulates the acute phase response to tissue injury or infection by inducing the release of hepatic acute phase-reactive proteins that act systemically to limit tissue damage (1). In adaptive immunity, IL-6 promotes terminal differentiation of B cells into antibody-producing cells (2), co-stimulates the production of IL-2 by T cells in response to mitogen or anti-T cell receptor antibody (3), and with IL-2 promotes the development of cytolytic T cells from thymocytes (4). IL-6 also supports hematopoiesis through its abilities to co-stimulate with IL-3 the proliferation of multipotent hematopoietic progenitors (5) and the differentiation/proliferation of myeloid and erythroid progenitors (6).

Mast cells are the major source of IL-6 in tissues undergoing certain types of allergic inflammation. In biopsy specimens of nasal mucosa from individuals with allergic rhinitis, >90% of the cells that contain IL-6 are mast cells (7). Similarly, in bronchial mucosal biopsies from persons with allergic asthma, nearly 100% of the cells that contain IL-6 are mast cells (8). In these tissues, the expression of IL-6 was specifically associated with mast cells that express the neutral protease tryptase but not chymase or carboxypeptidase A. This protease phenotype is characteristic of mast cells located in mucosal tissues (8).

The culture of mouse bone marrow in medium containing IL-3 results in the differentiation and growth of immature mast cells (BMMC) that produce IL-6 mRNA and protein when activated immunologically through their high affinity Fc receptors (9, 10). Stimulation of BMMC with exogenous cytokines results in generation of leukotriene C₄ (11) and prostaglandin D₂ (12) with kinetics and amounts similar to IgE/antigen-mediated activation. We therefore sought a non-IgE-dependent, cytokine-mediated combination for IL-6 expression. We now describe the induction of expression of the IL-6 gene in BMMC by a particular combination of cytokines: c-kit ligand (KL), IL-10, and IL-1. Whereas the combination of KL + IL-10 is sufficient to elicit transcription of the IL-6 gene, the resulting mRNA is unstable because of a destabilizing protein and produces little IL-6 protein. The introduction of IL-1, either simultaneously or within 3 h of the addition of KL + IL-10, markedly increases the IL-6 mRNA level and the attendant IL-6 protein production. The ability of KL + IL-10 to stimulate IL-6 gene transcription and the capacity of IL-1 to induce a protein that stabilizes IL-6 mRNA together reveal a basis for the synergistic action of these three cytokines in directly eliciting production of the multifunctional cytokine, IL-6, from mouse mast cells.

EXPERIMENTAL PROCEDURES

Mouse recombinant IL-1, IL-4, and transforming growth factor (TGF)-1 (Endogen, Cambridge, MA), nerve growth factor (NGF) (Sigma), and tumor necrosis factor (TNF)- (Genzyme, Cambridge, MA) were purchased. Mouse recombinant KL, IL-3, IL-9, and IL-10 were produced through expression by the infection of Sf9 insect cells with recombinant baculovirus (13).

Bone marrow cells from male BALB/cByJ mice (Jackson Laboratory, Bar Harbor, ME) were cultured for 3-6 weeks in 50% enriched medium (RPMI 1640 medium containing 100 units/ml penicillin, 100 µg/ml streptomycin, 10 µg/ml gentamicin, 2 mM L-glutamine, 0.1 mM nonessential amino acids, 10% fetal calf serum), and 50% WEHI-3 cell (American Type Culture Collection, Rockville, MD) conditioned medium. After 3 weeks, more than 95% of the cells in culture were mast cells as ascertained by metachromatic staining with toluidine blue. The BMMC were washed once with enriched medium, resuspended at a density of 1 × 10⁷ cells/ml in enriched medium supplemented with KL (50 ng/ml) + IL-10 (20 units/ml) with or without various concentrations of IL-1, and cultured up to 48 h. These concentrations of KL and IL-10 were selected because they are within the range of concentrations that modulate the BMMC secretory granule phenotype (14, 15, 16) and stimulate the release of lipid-derived mediators from BMMC (12). At various time points, the cells were centrifuged at 200 × g for 10 min in a Beckman GPR centrifuge, the supernatants were collected for determination of the IL-6 protein concentrations by ELISA with the mouse IL-6 MiniKit (Endogen, Cambridge, MA), and RNA was isolated from the cell pellets. In some cases, the concentrations of IL-6 protein retained in the cell pellets and in the total reaction mixture were determined by ELISA after the samples were sonicated on ice.

Total cellular RNA was isolated according to the manufacturer's instructions with QIAshredder and the RNeasy Total RNA kit (Qiagen, Chatsworth, CA). RNA blots were prepared containing a 5-µg portion of each RNA sample. The blots were sequentially hybridized with ³²P-labeled cDNA probes for mouse IL-6 and 18 S ribosomal RNA and washed as described (13). The amount of radioactivity associated with each band was quantitated with a Betascope 603 Blot Analyzer (Betagen, Waltham, MA). The counts per minute for IL-6 mRNA were divided by the counts per minute for 18 S RNA obtained from the same lane to adjust for differences in sample loading and transfer between lanes. The fold increase in steady-state mRNA for each sample was calculated by dividing the 18 S RNA-adjusted value for cytokine-stimulated cells by that for untreated cells.

To determine mRNA half-lives, actinomycin D (5 µg/ml) was added to cells at the time of peak steady-state IL-6 mRNA accumulation. RNA was isolated from samples of cells taken at several times after the addition of actinomycin D, and RNA blot analysis was performed as described above. The data were graphed as the percent of RNA remaining versus the time after the addition of actinomycin D, and the slopes of the linear portions of the decay curves were calculated by linear regression to determine the mRNA decay constants. mRNA half-lives were calculated with the formula t_1/2 = ln 2/mRNA decay constant (17).

2.5 × 10⁷ cells were stimulated with cytokine combinations for various periods, and a nuclear run-on assay was performed as described previously (18) except that the labeled run-on RNA products were extracted according to the protocol provided with the QIAshredder/RNeasy Total RNA kit for liquid samples. Alkaline-denatured plasmid Bluescript II KS with or without a mouse IL-6 cDNA insert and a pUC vector containing a mouse -actin cDNA were slot-blotted at 5 µg/well onto a nylon membrane for hybridization.

RESULTS

When BMMC, generated in 50% WEHI-3 cell-conditioned medium as a source of IL-3, were cultured for up to 48 h in either KL (50 ng/ml) + IL-10 (20 units/ml) alone or with IL-1 (5 ng/ml), steady-state IL-6 mRNA levels rapidly increased in the cells by 0.5 h, as shown on typical autoradiograms (Fig. 1, A and B) and by the 18 S RNA-adjusted data derived by Betascope analysis from multiple experiments (Fig. 1C). Steady-state IL-6 mRNA levels induced by KL + IL-10 reached a maximal 2-fold increase by 0.5 h, declined by 50% during the next 2.5 h, and returned to base-line level by 5-7 h. The addition of IL-1 to KL + IL-10 resulted in a 4-fold increase in IL-6 mRNA levels by 0.5 h and a maximal 12-fold increase 3-5 h after cytokine addition. The elevation in the steady-state IL-6 mRNA level was still ~3-fold higher than the starting level 24 h after the addition of the cytokines. The readdition of the three cytokines 3 h after their first addition to BMMC did not alter the time course or magnitude of the increase and decrease in steady-state IL-6 mRNA level. The findings with and without cytokine readdition were superimposable as assessed with Betascope analysis in three experiments (data not shown). The fact that IL-1 alone did not stimulate an increase in IL-6 mRNA levels (data not shown) indicates that IL-1 acts synergistically with KL + IL-10 to elicit the increase in IL-6 mRNA.

The augmentation of the steady-state IL-6 mRNA levels by the exogenous cytokines also elicited the release of IL-6 protein by BMMC. BMMC stimulated with KL + IL-10 released ~0.2 ng/10⁶ cells of immunoreactive IL-6 by 5-7 h (Fig. 1D). The addition of IL-1 to KL + IL-10 markedly enhanced IL-6 release, which reached a plateau value of ~9 ng/10⁶ cells between 7 and 24 h. BMMC exposed to each of the three cytokines alone at the same concentrations produced negligible quantities of IL-6 (data not shown, n = 3). The dependence of IL-6 release on concentrations of IL-1 ranging from 0.05 to 10 ng/ml was assessed at 24 h in the presence of the fixed concentrations of KL and IL-10. A dose-related increase in IL-6 release was found that reached a plateau of 9.2 ± 1.4 ng/10⁶ cells (mean ± S.E., n = 6) at 5 ng/ml IL-1 (data not shown).

To determine if IL-6 protein was also retained by cytokine-stimulated BMMC, the amounts of IL-6 protein associated with the cell pellets, supernatants, and whole reaction mixtures were determined 24 h after stimulation of cells with either KL + IL-10 or KL + IL-10 + IL-1. Only 1.12 ± 0.26% (mean ± S.E., n = 7) of IL-6 protein was found in the separated cell pellets as compared to supernatants, and the amount of IL-6 protein in the supernatants was 120 ± 7.3% (mean + S.E., n = 7) of that in the unseparated total reaction mixture. Thus, the amount of IL-6 released by BMMC was considered to represent the amount produced.

The specificity of the synergistic effect between IL-1 and the combination of KL + IL-10 was investigated by substituting other cytokines for IL-1 in cultures containing KL + IL-10 and measuring the IL-6 protein level by ELISA. The addition of IL-3 (1,000 units/ml), IL-4 (10 ng/ml), IL-9 (1,000 units/ml), NGF (10 ng/ml), TGF-1 (10 ng/ml), or TNF- (10,000 units/ml) did not produce more than 3% of the level of IL-6 protein released by KL + IL-10 + IL-1 (data not shown, n = 3).

Effects of Priming with KL + IL-10 before the Addition of IL-1 on the Expression of IL-6 mRNA and Protein

BMMC were cultured either with KL + IL-10 + IL-1 (5 ng/ml) added simultaneously (no priming) or with KL + IL-10 for 0.5 or 3 h before the addition of IL-1; the time at which IL-1 was added was designated as time 0 (Fig. 2). Priming the cells with KL + IL-10 for 0.5 h more than doubled the maximum amount of steady-state IL-6 mRNA achieved 3 h after the addition of IL-1, compared with cells that were not primed (Fig. 2A). When the cells were primed for 3 h, the time required to achieve the maximal level of IL-6 steady-state mRNA was decreased to 1 h, and half-maximal production of IL-6 protein occurred at 2 h as compared to ~4 h for no or 0.5 h of priming (Fig. 2B). Despite the increases in peak steady-state levels of IL-6 mRNA that occurred with the 0.5- and 3-h priming protocols, there were no increases in the maximal amounts of IL-6 protein produced, relative to cells that were exposed to the three cytokines simultaneously.

Effects of priming of BMMC with KL + IL-10 on the time course of IL-6 mRNA and protein induction in response to IL-1.

Effect of KL + IL-10 with and without IL-1 on IL-6 Gene Transcription Assessed by Nuclear Run-on Analysis

There was no detectable transcription of the IL-6 gene in BMMC before treatment with the recombinant cytokines (Fig. 3, lane 1). Exposure of BMMC for 0.5 h to KL + IL-10 alone (lane 2) or combined with IL-1 (lane 4) resulted in approximately the same elevation in the level of IL-6 gene transcription. In BMMC stimulated with KL + IL-10 for 3 h, the point at which steady-state IL-6 mRNA had decreased by 50% (Fig. 1C), IL-6 gene transcription continued unabated (Fig. 3, lane 3). The level of transcription remained unchanged in BMMC exposed to KL + IL-10 + IL-1 from 0.5 to 5 h (Fig. 3, lanes 4-6), a period in which the steady-state IL-6 mRNA level increased and peaked (Fig. 1C). IL-1 alone did not stimulate transcription of the IL-6 gene (data not shown).

Effects of KL + IL-10 alone or with IL-1 on IL-6 gene transcription assessed by nuclear run-on analysis.

Nuclear run-on analyses were also performed with BMMC primed with KL + IL-10 for 0.5 and 3 h before the addition of IL-1. The nuclei were harvested at 3 h (0.5-h priming) or 1 h (3-h priming), the times at which peak steady-state mRNA levels were indicated by RNA blot analysis for these experimental conditions. The level of transcription was not visibly altered by the addition of IL-1 after either duration of priming as compared to the level obtained with KL + IL-10 alone (Fig. 3, lanes 7 and 8 versus lanes 2 and 3).

Effects of KL + IL-10 with and without IL-1 on IL-6 mRNA Stability

BMMC were treated with either KL + IL-10 for 0.5 h or KL + IL-10 + IL-1 for 3 h to establish peak mRNA levels before actinomycin D (5 µg/ml) was added to the cells. At various times thereafter, RNA was isolated from the cells, and IL-6 mRNA was examined by RNA blot analysis, as shown in a typical autoradiogram (Fig. 4, A and B) and by Betascope analysis of 18 S RNA-adjusted data (Fig. 4C). In BMMC treated with KL + IL-10, IL-6 mRNA had a calculated half-life of 0.4 h. This half-life was faster than the IL-6 mRNA decay time of 0.9 h in the absence of actinomycin D. In contrast, the half-life of IL-6 mRNA produced by BMMC treated with KL + IL-10 + IL-1 was 3.2 h, which was the same as the decay time in the absence of actinomycin D. At the protein level, actinomycin D added to BMMC 0.5 h after exposure to KL + IL-10 abrogated IL-6 production (data not shown). However, actinomycin D added to BMMC 3 h after exposure to KL + IL-10 + IL-1 did not reduce the production of IL-6 protein at any subsequent time point studied, as compared with BMMC that were not treated with actinomycin D (Fig. 4D).

In contrast to the results obtained when actinomycin D was added after the cells were treated with KL + IL-10 + IL-1, its addition 10 min before the cells were exposed to KL + IL-10 + IL-1 completely inhibited the appearance of IL-6 mRNA by RNA blot analysis (data not shown). The addition of actinomycin D to BMMC 0.5 h after the addition of the three cytokines resulted within an additional hour in the complete loss of the mRNA that had accumulated just before the actinomycin D was added (Fig. 4E). Thus, the effects of actinomycin D on IL-6 mRNA induced by the three exogenous cytokines were time-dependent.

The stability of IL-6 mRNA induced by 0.5 and 3 h of priming BMMC with KL + IL-10 and the subsequent addition of IL-1 was studied by adding actinomycin D 3 and 1 h later, respectively, at the times of peak IL-6 mRNA levels for each sequence. The calculated half-lives of IL-6 mRNA elicited with the 0.5- and 3-h priming protocols were 0.9 and 1.0 h (Fig. 5), respectively, about twice the value obtained with BMMC stimulated with KL + IL-10 only.

Assessment of the Requirement for Protein Synthesis for IL-6 mRNA Accumulation Induced by KL + IL-10 with and without IL-1

BMMC were preincubated with or without cycloheximide (10 µg/ml) for 0.5 h and then maintained either in medium alone or in medium with cytokines for 3 h before RNA isolation and blot analysis. The addition of cycloheximide to cells maintained in medium alone or in medium with 100 units/ml IL-3 resulted in an ~2-fold increase in the IL-6 mRNA level compared with the levels in cells maintained in the same medium without cycloheximide (shown in a representative experiment in Fig. 6A and analyzed from four experiments in Fig. 6B). KL + IL-10 stimulated a 5-fold increase in IL-6 mRNA relative to cells maintained in medium, and this increase was not changed significantly by preincubation of the BMMC with cycloheximide. In contrast, a 30-fold increase in IL-6 mRNA induced by KL + IL-10 + IL-1 was reduced by 52.8 ± 5.2% (mean ± S.E., n = 4) by preincubation of the cells with the drug.

The effects of cycloheximide on IL-6 mRNA accumulation after gene induction were studied by adding cycloheximide to cells 10 min after the addition of KL + IL-10 (Fig. 7, A and C) and 1 h after the addition of KL + IL-10 + IL-1 (Fig. 7, B and C). In both cases, IL-6 mRNA was superinduced as assessed by the peak mRNA levels achieved in the presence of cycloheximide. To assess the potential role of IL-6 gene transcription in the superinducing effects of cycloheximide on IL-6 mRNA accumulation, nuclear run-on analyses were performed with BMMC cultured in KL + IL-10 + IL-1 without cycloheximide, and with cycloheximide added to cultures simultaneously with or 1 h after the three cytokines. The level of transcription in all three cases was essentially the same (data not shown).

DISCUSSION

We have shown that a particular combination of exogenous cytokines acts synergistically to induce high levels of IL-6 expression in mouse BMMC and that these cytokines act at discrete steps in the pathway of IL-6 generation. Whereas stimulation of BMMC with KL + IL-10 increased steady-state levels of IL-6 mRNA by 2-fold with minimal release of IL-6 protein, the addition of IL-1 to the combination resulted in a 12-22-fold increase in the peak level of steady-state IL-6 mRNA and a 50-100-fold gain in protein to 9 to 30 ng of IL-6/10⁶ BMMC (Figs. 1 and 2). Nuclear run-on analysis showed that approximately the same level of IL-6 gene transcription was induced 0.5 h after BMMC were exposed to KL + IL-10 with or without IL-1 (Fig. 3). Furthermore, the level of transcription was relatively unchanged throughout periods corresponding to the rise, peak, and decline phases of steady-state IL-6 mRNA elicited by these cytokine combinations (Fig. 1C). Thus, KL + IL-10 induces IL-6 gene transcription in BMMC, whereas IL-1 appears to act post-transcriptionally.

As determined with actinomycin D, the half-life of IL-6 mRNA elicited by the three exogenous cytokines was 8-fold greater than that elicited by KL + IL-10 alone (Fig. 4). The increase in the half-life of IL-6 mRNA stimulated by IL-1 is attributed to the induction of a protein that stabilizes IL-6 mRNA, because preventing protein synthesis with cycloheximide 0.5 h before the addition of the three exogenous cytokines inhibited by ~50% the level of steady-state IL-6 mRNA generated, relative to cells not exposed to cycloheximide (Fig. 6). The generation of the stabilizing protein required transcription, because actinomycin D added 0.5 h after the addition of the three exogenous cytokines prevented the accumulation of steady-state IL-6 mRNA (Fig. 4E), even though maximal transcription of the IL-6 gene was already occurring at the time the actinomycin D was added (Fig. 3). One hour after the addition of KL + IL-10 + IL-1 to BMMC, the synthesis of the stabilizing factor appeared to be sufficient to stabilize transcripts, because cycloheximide added at that time did not decrease the level of IL-6 mRNA (Fig. 7, B and C). Thus, a protein that stabilizes IL-6 mRNA is made early after the addition of IL-1 to BMMC in which IL-6 gene transcription is being induced by KL + IL-10. Once made, the stabilizing protein appears to be long-lived, because the addition of actinomycin D to BMMC 3 h after the addition of the three exogenous cytokines did not accelerate the decay of IL-6 mRNA compared with that in cells not exposed to the drug (Fig. 4, A and B). Accordingly, the amount of IL-6 protein released from BMMC was not diminished by the addition of actinomycin D 3 h after the addition of KL + IL-10 + IL-1 (Fig. 4D).

A second effect of KL + IL-10 on IL-6 mRNA was revealed by the finding that the addition of cycloheximide to BMMC after stimulation with KL + IL-10 with or without IL-1 increased the levels of steady-state IL-6 mRNA relative to that of cells not treated with the drug (Fig. 7). These results indicate that in addition to stimulating transcription of the IL-6 gene, the combination of KL + IL-10 induces a protein factor that destabilizes IL-6 mRNA, with no apparent contribution from IL-1. Because the level of IL-6 gene transcription in BMMC treated with KL + IL-10 continued unabated (Fig. 3) during the period when the steady-state levels of IL-6 mRNA decline (Fig. 1C), it is likely that the destabilizing protein induced by KL + IL-10 is the predominant regulator of steady-state IL-6 mRNA levels and hence limits the amount of IL-6 protein released in the absence of IL-1.

Delaying the addition of IL-1 to BMMC exposed to KL + IL-10 by 0.5 and 3 h did not prevent the induction of high levels of steady-state IL-6 mRNA and the release of substantial IL-6 protein. Indeed, the peak IL-6 mRNA levels were ~2-fold greater when the addition of IL-1 was delayed (Fig. 2). The level of IL-6 gene transcription in cells exposed to IL-1 after the other two cytokines was similar to that in cells exposed to the three cytokines simultaneously (Fig. 3), and the half-lives of the steady-state IL-6 mRNA obtained with sequential addition of the cytokines were ~2-fold greater than those of cells exposed to KL + IL-10 alone (Fig. 5). These findings suggest that the greater peak accumulation of steady-state IL-6 mRNA levels that occurs with sequential cytokine addition in part provides the high levels of IL-6 protein released (Fig. 2). Thus, both the peak magnitude and post-peak half-life of steady-state IL-6 mRNA contribute importantly to the amount of IL-6 protein released, and both parameters are increased by the addition of IL-1 to KL + IL-10.

In BMMC, IL-1 also acts synergistically with KL + IL-10 for the induction of prostaglandin-endoperoxide synthase-2 (12) with attendant downstream prostaglandin D₂ release. However, the possibility that IL-1 was acting at a step(s) distinct from that of KL + IL-10 was not examined. As with IL-6 elaboration, IL-1 cannot be replaced by other cytokines to induce prostaglandin-endoperoxide synthase-2 synthesis in combination with KL + IL-10 (19). Thus, the combination of KL + IL-10 + IL-1 appears to be unique in providing an alternative to immunoglobulin activation of mast cells for generation of IL-6 and prostaglandin D₂ through the induction of essential pathways for product formation and stabilization. The fact that KL and IL-10 also both mediate mast cell development (20) is compatible with their being present at sites of mast cell hyperplasia, where these cytokines may additionally provide a stimulus for regulated mast cell activation in conjunction with a proinflammatory cytokine such as IL-1. Indeed, KL is widely expressed on epithelium and stromal cells (21), IL-10 is produced by bronchial epithelial cells (22), and IL-1 is found in airway walls during late asthmatic reactions (23). Thus, this combination of cytokines is likely present at sites of allergic inflammation and may indeed provide an alternative pathway to mast cell activation of cytokine release in vivo.

The steps by which IL-1 species and other exogenous cytokines act synergistically to produce IL-6 have been defined for a limited number of cases. In human fibroblasts, IL-1 and TNF- each stimulate transcription of the IL-6 gene and synergistically induce expression of steady-state IL-6 mRNA and protein (24). Moreover, the half-life of IL-6 mRNA induced by the two exogenous cytokines is ~3 times greater than the half-life of IL-6 mRNA induced by each cytokine alone. IL-1 and TGF-1 synergistically increase transcription of the IL-6 gene, steady-state IL-6 mRNA, and release of IL-6 protein in human fibroblasts (25). IL-1 also elicits IL-6 gene transcription in human fibroblasts, promotes stabilization of the mRNA, and acts synergistically with TNF- to increase steady-state IL-6 mRNA levels and release protein (26). Our studies demonstrate that in mouse BMMC, IL-1 neither elicits transcription of the IL-6 gene nor enhances transcription stimulated by KL + IL-10, but rather acts to stabilize IL-6 mRNA induced by the latter two cytokines. Thus, IL-1 can act at a discrete step from other cytokines with which it acts synergistically so as to produce appreciable release of IL-6 protein.

The AU-rich, 3-untranslated regions of many protooncogene and cytokine mRNA species contain single or multiple copies of the sequence AUUUA, which, in the proper context, target these mRNAs for rapid degradation and may also suppress their translation (27, 28, 29, 30, 31, 32). IL-6 mRNA contains multiple copies of the AUUUA sequence, and factors that bind AUUUA-containing regions so as to regulate other mRNAs have been described. AUF1 is a cytosolic factor that speeds the decay of c-myc mRNA and granulocyte/macrophage-colony stimulating factor (GM-CSF) mRNA in erythroleukemia cells (33), whereas AUBF stablizes GM-CSF mRNA by binding to and protecting the AUUUA sequence element from destabilizing factors (34). Thus, the interplay between these opposing factors determines the degradation rate of the GM-CSF mRNA. Our data indicate that cytokine-induced IL-6 gene expression in BMMC is regulated in a sequential manner. The combination of KL + IL-10 both stimulates IL-6 gene transcription and counter-regulates its effects by inducing a transcript destabilizing factor. IL-1 causes the early induction of an IL-6 transcript stabilizing factor that appears to be long-lived. Because the amount of IL-6 protein produced depends critically on the effects of added IL-1, factors analogous to AUF1 and AUBF that alter IL-6 mRNA levels may be the key regulators of mast cell production of IL-6 in response to cytokine signals from the microenvironment.