Receptors for interleukin (IL)-10 and IL-6-type cytokines use similar signaling mechanisms for inducing transcription through IL-6 response elements.

The cytoplasmic domain of the receptor for interleukin 10 (IL-10R) contains two box 3 sequence motifs that have been identified in the signal-transducing receptor subunits for IL-6-type cytokines and noted to be required for activating STAT3 and inducing transcription through IL-6-responsive elements. To determine whether the IL-10R has signaling functions similar to IL-6R in cells normally expressing these receptors, leukocytes of the B-, T-, and NK-cell lineages were treated with either cytokine. Both cytokines activated factors that bound to the sis-inducible element and included STAT1 and STAT3. The cell response to IL-10 characteristically differed from that to IL-2/IL-15, IL-4, and interferon gamma. The signaling capabilities of the IL-10R for activating specific STAT proteins and inducing gene transcription were defined by reconstitution of receptor functions in transfected tissue culture cells. COS-1 cells, co-expressing the human IL-10R and individual STAT proteins, confirmed a preference of the IL-10R for STAT3 and STAT1. Unlike many hematopoietin receptors, the IL-10R did not detectably activate STAT5. The IL-10R, together with reporter gene constructs containing different IL-6-responsive gene elements, reconstituted in hepatoma cells an induction of transcription by IL-10 that was comparable to that by IL-6. This regulation could not be appreciably modified by enhanced expression of STAT proteins. The similar actions of IL-10R and IL-6R on the induction of endogenous IL-6-responsive genes were demonstrated in hepatoma cells stably expressing the IL-10R. These receptor functions required the presence of the box 3 motifs, as shown by the analysis of the mouse IL-10R constructs containing progressively truncated cytoplasmic domains. The data demonstrate that the IL-10R, unlike other members of the interferon receptor family, is highly effective in recruiting the signaling pathways of IL-6-type cytokine receptors.

The cytoplasmic domain of the receptor for interleukin 10 (IL-10R) contains two box 3 sequence motifs that have been identified in the signal-transducing receptor subunits for IL-6-type cytokines and noted to be required for activating STAT3 and inducing transcription through IL-6-responsive elements. To determine whether the IL-10R has signaling functions similar to IL-6R in cells normally expressing these receptors, leukocytes of the B-, T-, and NK-cell lineages were treated with either cytokine. Both cytokines activated factors that bound to the sis-inducible element and included STAT1 and STAT3. The cell response to IL-10 characteristically differed from that to IL-2/IL-15, IL-4, and interferon ␥. The signaling capabilities of the IL-10R for activating specific STAT proteins and inducing gene transcription were defined by reconstitution of receptor functions in transfected tissue culture cells. COS-1 cells, co-expressing the human IL-10R and individual STAT proteins, confirmed a preference of the IL-10R for STAT3 and STAT1. Unlike many hematopoietin receptors, the IL-10R did not detectably activate STAT5. The IL-10R, together with reporter gene constructs containing different IL-6-responsive gene elements, reconstituted in hepatoma cells an induction of transcription by IL-10 that was comparable to that by IL-6. This regulation could not be appreciably modified by enhanced expression of STAT proteins. The similar actions of IL-10R and IL-6R on the induction of endogenous IL-6responsive genes were demonstrated in hepatoma cells stably expressing the IL-10R. These receptor functions required the presence of the box 3 motifs, as shown by the analysis of the mouse IL-10R constructs containing progressively truncated cytoplasmic domains. The data demonstrate that the IL-10R, unlike other members of the interferon receptor family, is highly effective in recruiting the signaling pathways of IL-6-type cytokine receptors.
Interleukin 10 (IL-10), 1 like IL-4, is classified as an antiinflammatory cytokine based on its suppressing action on the production of "pro-inflammatory mediators" and its contribution to resolve inflammatory reactions (1,2). Although IL-10 and IL-4 are involved in the control of the same physiological process and have common target cells such as lymphocytes and monocytes, their mechanism of actions appears to be substantially different. The IL-10R is structurally related to IFN receptors (3,4), whereas the IL-4R belongs to the family of hematopoietin receptors (5). Upon ligand binding, the IL-10R forms oligomeric complexes (6). As noted for the IFN␣/␤R, IL-10R action involves the stimulation of JAK1 and TYK2, phosphorylation of STAT1 and STAT3, and activation of the STATs to bind to DNA sequences such as IFN␥ activator site (activating sequence) (7)(8)(9)(10)(11). Regulation of cell proliferation and differentiated gene functions by IL-10 required the presence of specific regions of the cytoplasmic domain of the IL-10R (11). By contrast, the functional IL-4R consists of the heteromeric complex of IL-4R␣ and IL-2R␥, signals via JAK1 and JAK3, and preferentially activates STAT6 (12). These differences in signal transduction may partially explain the difference in regulation of target genes seen in IL-10-and IL-4responsive cells (13,14).
In this study, we compared the signaling of the IL-10R with that of IL-6R and characterized the gene-regulatory action of the IL-10R by its reconstitution in heterologous cell systems. We demonstrated that the IL-10R has a strong preference for STAT3 and exerts a transcriptional induction via gene regulatory elements that is highly similar to that mediated by the IL-6R or other box 3 motif-containing hematopoietin receptors.

EXPERIMENTAL PROCEDURES
Cells-HepG2 and COS-1 cells were cultured as described (15,16). Highly purified (Ͼ95% by flow cytometric analysis) T-cells were prepared from fresh human blood using immunomagnetic bead depletion as described (17). CLL cells were obtained from blood of patients with a proven diagnosis of CLL (median 93% CD5ϩCD19ϩ by flow cytometric analysis) and then stored. Upon thawing, six specimens had 100% viability by vital dye exclusion, and one specimen had 75% viability. The NK-92 cell line was kindly provided by Dr. Hans G. Klingermann (Vancouver, British Columbia) and was maintained in RPMI 1640 medium supplemented with 10% fetal calf serum and 1 ng/ml IL-2 (Hoffmann-La Roche, Nutley, NJ; specific activity 1.53 ϫ 10 7 units/mg). Leukocytes were maintained for 2 h, and the cells from established lines for 16 h, in cytokine-and serum-free medium. Depending upon the experimental settings, the cells were then treated for 15 min to 24 h with serum-free medium containing 100 ng/ml purified recombinant human IL-10 (Schering-Plough Research Institute), IL-6 (Genetics Institute), oncostatin M, LIF, G-CSF, IL-4, IL-15 (Immunex Corp.), IL-3 (Sandoz, Basel), IFN␥ (Genentech), or IL-2 (Cetus), or 5 ng/ml IL-1␤ (Immunex Corp.). The cytokine concentrations and length of treatments have been determined previously to yield maximal response, i.e. activation of STATs, induction of genes, and stimulation of plasma protein secretion (see Fig. 6, C and D as an example).
Transfection-COS-1 cells were transfected by the DEAE-dextran method (27) with DNA mixtures containing 1 g/ml expression vector for receptors and 3 g/ml expression for STATs. Transfected cultures were subdivided and used to determine the activation of STAT isoforms by 15 min cytokine treatment as described (21,23). HepG2 cells were transfected by the calcium phosphate method (28) using DNA mixture containing CAT reporter gene construct (15 g/ml), expression vector for receptors (1 g/ml) and for STATs (1-5 g/ml) (15), and pIE-MUP (2 g/ml) or pSV-␤-galactosidase (1 g/ml) as internal normalization markers. Subcultures were treated for 24 h with cytokines, extracted, and analyzed for CAT activity following standard procedures (16,21,23). The normalized values for CAT activity were expressed relative to the untreated control in each experimental series (ϭ1.0). Cell cultures transfected with bacterial ␤-galactosidase gene were stained, and the percentage of strongly positive cells was determined in randomly chosen areas on the culture dish (containing between 250 ando 390 cells). The transient transfection protocol for HepG2 cells yielded high expressing transfectants that amounted to 2.1 Ϯ 2.4% of the cells in the culture (mean Ϯ S.D., n ϭ 50).
Generation of HepG2 Cells Stably Expressing the Human IL-10R-A retroviral expression vector for the human IL-10R (MSCV-hIL-10R) was constructed by inserting the 2200-bp XhoI fragment of the hIL-10R cDNA (4) into the XhoI site of MSCVneoEB carrying the bacterial neo gene (29). Replication-defective recombinant MSCV-hIL-10R virus with amphotropic host range was produced from stably transduced PA317 packaging cells (Ref. 30; American Type Culture Collection CRL 9078) generated essentially as described (31) by infection with virus harvested from transiently transfected GPϩE-86 ecotropic packaging cells (32). Cells were maintained in Dulbecco's modified Eagle medium with 4.5 g/liter glucose supplemented with 10% fetal bovine serum (Life Technologies, Inc.). Virus-containing supernatant was collected from confluent cultures 24 h after medium change and filtered through a 0.45-m membrane. The virus titer on NIH 3T3 fibroblasts was 1 ϫ 10 7 G418-resistant colony-forming units/ml. The virus-containing medium was added to monolayers of HepG2 cells in the presence of 8 g/ml Polybrene (Sigma) for 12 h. The infected cells were cultured in medium containing 1 mg/ml G418 (Life Technologies, Inc.). The viral infection was repeated twice on the same cell population after 3 and 6 days. The cells with vector integration were selected by increasing G418 in the culture medium to 2 mg/ml. The proliferating cells recovered after 4 weeks of selection (termed IL-10R-HepG2 cells) were used for the analysis of IL-10R functions.

RESULTS AND DISCUSSION
IL-10 Activates STAT Proteins in Leukocytes-The primary structure of the IL-10R indicated the presence of two box 3 sequence motifs in the cytoplasmic domain at positions 446 -449 (YXRQ) and 496 -499 (YXKQ) in the human receptor (4) and at positions 427-430 (YXKQ) and 477-480 (YXKQ) in the mouse receptor (11). Such box 3 motifs have been identified in gp130 and found necessary for activating STAT1 and STAT3 and for inducing transcription via IL-6RE (23,36). Hence, we wished to determine whether the IL-10R exerted a signaling reaction with the specificity of the IL-6R. As the first step, we compared the STAT activation patterns in cells constitutively expressing both receptors (37) and used lymphocytes isolated from a B-cell CLL patient. Treatment with various cytokines revealed the receptor-specific recruitment of latent DNA binding factors (Fig. 1A, Second Preparation). The extracts from IL-10-and IL-6-treated cells produced complexes that comigrated with SIF-A, -B, and -C which, based on the previous analyses (33,38), consisted of SIE bound to STAT3 homodimer, STAT3 and STAT1 heterodimer, and STAT1 homodimer, respectively. The contribution of STAT1 and STAT3 to the SIF patterns was also detected by antibody supershift assay (Figs. 1C and 2B). The cells responded to IL-4 and IFN␥ by activating appropriately STAT6 and STAT1, respectively. These factors yielded the characteristic EMSA patterns with SIE and TB-2 ( Fig. 1A). IL-2 and IL-15 were ineffective; thus, a receptor activation of STAT5 could not be detected in these cells.
The STAT activation pattern in the freshly isolated lymphocyte preparation was comparable to that in cells collected from the same patient one year earlier (Fig. 1A, First Preparation). Differences appeared to be quantitative and included a more effective activation of the SIF-C complex by IL-6 and a lower IL-4 response. The ability of IL-10 and IL-6 to activate SIEbinding factors was also observed in CLL cells collected from different patients (Fig. 1B). In each case, IL-10 was highly effective, whereas IL-6 action was more variable. By antibody supershift assay, we determined that STAT3 and, to a lesser extent, STAT1 contributed to the complex SIF pattern as noted in the example of the IL-10-or IL-6-treated CLL cells from patient A (Fig. 1C). The EMSA patterns obtained with extracts from CLL cells were highly similar to the patterns from IL-6treated HepG2 cells, whose STAT1 and STAT3 compositions ( Fig. 1C) were in agreement with published data (33,38). The minor response to IL-6 of some of the CLL cell preparations (e.g. Patient 6 in Fig. 1B) may in part be due to a lower level of effective IL-6R subunits in these cells. Nevertheless, despite the rather drastic differences in intensities, the SIF patterns produced by IL-10 and IL-6 in the individual cell preparations showed comparable qualitative compositions (Fig. 1, C and D).
To identify whether the pattern of STAT activation by the IL-10R in CCL cells was representative for other IL-10R-bearing cells, we analyzed freshly isolated resting T-cells, and cells from the established NK-92 line ( Fig. 2A). Based on the SIF pattern, IL-10 prominently activated STAT3 and to a lesser extent STAT1, which could be verified by antibody supershift assay (Fig. 2B). IL-6 was only effective in T-cells and elicited a strong activation of STAT1 and -3.
The IL-2 or IL-15 response in T-cells and NK-92 cells demonstrated that these cells, unlike CLL cells, express and activate STAT5 that produced a prominent complex with TB-2 ( Fig. 2A). Although a relatively minor TB-2-bound complex was also noted in IL-10-treated T-cells, this complex was not recognized by anti-STAT5 in supershift EMSA (data not shown) and may represent complexes including STAT1 and/or STAT3 (22). 2 Specificity of STAT Activation by the Human IL-10R-Since each of the lymphocyte types analyzed in Figs. 1 and 2 contained a characteristic and complex mixture of DNA-binding proteins, a precise assessment of the IL-10R's capability to activate individual STAT proteins and to compare its action with that of the IL-6R was difficult. Therefore, we defined the signaling potential of the IL-10R by reconstitution in the IL-10R-deficient COS-1 cells using overexpressed proteins (Fig. 3). To grade the specificity and efficacy of the signaling, we tested in parallel transfected G-CSFR. The G-CSFR was selected because like the IL-10R, it is a homo-oligomeric receptor (39), contains one box 3 motif, and is able to activate with equal efficiency STAT1, STAT3 (Fig. 3A, patterns with SIE), and STAT5B (Fig. 3A, patterns with TB-2). The transfected IL-10R activated STAT1 and STAT3. By comparing the intensity of the SIF complexes, the IL-10R, unlike the G-CSFR, appeared to have a higher preference for STAT3 than STAT1. The STAT3 preference was also noted in the presence of overexpressed STAT1. One half of the radioactivity was confined to a heterodimeric complex that included transfected STAT1 and COS-1 STAT3 and migrated slower than the rat STAT1 homodimer (SIF-C in Fig. 3A). The activation of endogenous COS-1 STAT3 by transfected IL-10R (Fig. 3A, major band in the lane labeled control ϩ 10 is COS-1 SIF-A) was particularly notable because transfection of various hematopoietin receptors, including the G-CSFR and those for IL-6-type cytokines, had yielded primarily activation of COS-1 STAT1 (15,16,21,23). No detectable effect of the IL-10R on STAT5B or -6 was evident. The failure of STAT5B activation appeared to distinguish the IL-10R from many hematopoietin receptors including the G-CSFR (Fig. 3A, TB-2 panel), which have shown the potential to activate STAT5 (40,41). That the reconstituted COS-1 cells provided abundant amounts of STAT5 proteins was verified by Western blot analysis (Fig. 3B).
The results indicated that signaling functions initiated by the IL-10R in the normal cell environment (Figs. 1 and 2) could be accurately reproduced by transfection into heterologous cells (Fig. 3). The effects of the IL-10R, with the exception of activating STAT5, were similar to those of box 3-containing hematopoietin receptors, but clearly differed from IFN␣/␤R or IFN␥R, even though, like the IL-10R, the latter two receptors utilized JAK1 and TYK2 as signal-communicating kinases (10,11,41). Signaling by the IL-10R may involve reactions that have been predicted for gp130 (31,42) and may entail the phosphorylation of the tyrosine residue in box 3 that in turn serves as a high affinity docking site for STAT3, and to a lesser degree, for STAT1. The activation of STAT function is conceivably due to the action of IL-10R-associated JAK1 and/or TYK2 (10, 11).

IL-10R Induces Transcription of IL-6-responsive Genes-To
assess whether the IL-10R executed a gp130-like effect not only on STAT activation but also on gene transcription that may or may not be caused by STATs, we needed a suitable experimental assay system. Since we had no adequate molecular tools available to study gene induction by IL-6 or IL-10 in leukocytes, we resorted to the application of hepatoma cells as surrogate system. Hepatoma cells were attractive because the IL-6R-mediated gene induction had been well characterized in these cells and the signaling process could experimentally be manipulated. We transfected the IL-10R together with characteristic IL-6-responsive CAT gene constructs, pHRRE-CAT and pIL-6RE-CAT, into HepG2 cells. We had verified that HepG2 cells do not express endogenous IL-10R mRNA nor respond to IL-10 treatment (see Fig. 6, A, C, and D). IL-10 treatment of the transfected cells stimulated the expression of both CAT gene constructs to a level achieved by IL-6 ( Fig. 4, left panel), indicating that the ectopically expressed IL-10R was able to engage the cellular signaling mechanisms like the endogenous IL-6R.
The CAT constructs represented artificial and optimized response systems whose regulation might substantially differ from that of natural gene promoters. Therefore, we also tested CAT reporter gene constructs that contained promoter and 5Ј flanking regions of representative APP genes (Fig. 4, right  panel). The promoter of the type 2 rat APP genes, ␣ 2 -MG and SPI-3, both containing an APRE-related sequence, responded equally to the IL-10R and IL-6R signal.
A hallmark of the IL-6 action is not only the specific induction of type 2 APP genes but also the synergism with IL-1 on the expression of type 1 APP genes such as human CRP (26). A 219-bp CRP promoter region containing the elements for both STAT3 and C/EBP transactivation (26) reproduced this synergism in transfected HepG2 cells (Fig. 4, right panel). With transfected IL-10R, the gene construct was induced by IL-10 alone and further enhanced by IL-1. Of note was that the IL-10 induction of the CRP-CAT construct exceeded the IL-6 effect by severalfold.
Gene Induction by IL-10 Is Unaffected by Overexpressed STATs-Activation of the DNA binding activity of STAT proteins by cytokines has often be considered to be causal to induction of transcription of genes containing STAT binding motifs in their promoter regions (41,42). The preference of hematopoietin receptors for certain STAT proteins has been employed for defining the gene regulatory role of the individual STATs (41). One approach is to enhance the stimulatory effect on transcription by ectopic overexpression of STAT proteins. By testing combinations of hematopoietin receptors and STAT proteins, we have recently identified that the regulation via IL-6RE was sensitive to STAT3 but not to STAT5B, and the regulation via HRRE was sensitive to STAT5 but not to STAT3 (21,23).
The preference of IL-10R for STAT3 (Fig. 3) suggested that IL-10R signaling to the IL-6RE-CAT gene might potentially be enhanced in the presence of higher amounts of STAT3. Therefore, we determined gene regulation by IL-10R in the presence of overexpressed STAT3. Overexpressed STAT5B served as a control because it was not expected to be activated by IL-10R. The specificity of receptor action and the effects of overexpressed STATs in these experiments were verified by comparing the responses mediated by the IL-10R with that of the endogenous IL-6R and co-transfected IL-3R, respectively (Fig.  5, A and B). The stimulation of both HRRE-CAT and IL-6RE-CAT by IL-10, like that by IL-6, appeared to be already maximal through the endogenous signaling mechanisms of HepG2 cells. Co-transfected STAT3 or STAT5B proved to be ineffective in significantly enhancing the IL-10 or IL-6 response (Fig. 5A), suggesting that the concentration of STAT proteins provided by HepG2 cells did not limit signaling by either IL-10R or IL-6R. In contrast, the amount of STAT proteins were insufficient for full signaling by the box 3-deficient IL-3R. We achieved, however, the expected regulatory action by the IL-3R through STAT3 on IL-6RE and through STAT5B on HRRE by overexpression. This result also documented that the STAT expression vectors used for our analysis yielded an enhanced level of functional STAT proteins in the transfected HepG2 cells (Fig. 5A).
The use of oligomerized STAT binding sites in reporter gene constructs often exaggerates the specific STAT-induced transcriptional induction by cytokine receptor signals (15,25,43). While the observed induction proves the signaling capability and specificity of the receptors and the cellular factors, it does not necessarily explain the mode of regulation that might occur via natural promoters containing only one or a few cognate STAT binding sites such as ␣ 2 -MG (22,24,35) or SPI-3 (25). We, therefore, analyzed the influence of overexpressed STATs on the IL-10 regulation of these promoters (Fig. 5B). The results indicated that the IL-10R elicited a signaling that was comparable to that of the IL-6R and again was not appreciably influenced by overexpressed STAT3 or STAT5B. This receptor specificity was highlighted by the comparison with the effect of the IL-3R. The signaling of the IL-3R to the APRE-containing ␣ 2 -MG was more effective than IL-10R or IL-6R, whereas the signaling to the SPI-3 promoter required enhanced expression of either STAT3 or STAT5B (44).
IL-10R, like many hematopoietin receptors and IFN␥R, is  15, 16, 21, and 23). This signaling appeared to be independent of STAT3 but could in part be reproduced by activated STAT5B (23). Although IL-10R failed to detectably activate overexpressed STAT5B in COS-1 cells (Fig. 3), this receptor, as well as the other receptors, could conceivably engage the low amount of endogenous STAT5 in HepG2 to induce HRRE-CAT. To assess this possibility, we introduced STAT5B⌬40C that, by virtue of lacking 41 carboxyl-terminal amino acid residues (22), exerted a dominant negative action on the STAT5 pathway (Fig. 5C, bottom). Neither the IL-10R nor the IL-6R effect on HRRE was significantly impaired, suggesting that the induction process by these receptors did not involve STAT5 as a critical mediator. The fact that the regulatory action by the IL-3R was substantially reduced by STAT5B⌬40C attested to a contributing role of STAT5 to the high activity of this receptor.
IL-10R Activates Endogenous APP Gene-The results from the transfection experiments (Figs. 4 and 5) indicated that FIG. 6. IL-10R activates STAT3 and stimulates expression of endogenous APP genes in IL-10R-HepG2 cells. A, total cellular RNAs were extracted from parental HepG2 cells, IL-10R-HepG2 cells, and B-CLL cells from two separate patients and analyzed by Northern blot hybridization for IL-10R mRNA (left panel). Equal RNA loading is shown by the ethidium bromide (EtBr) staining of the gel-separated RNA, including the 28S rRNA band (right panel). B, IL-10R-HepG2 cells were transiently transfected with IL-6RE-CAT. Subcultures were treated for 24 h with the cytokines indicated at the bottom. The media were analyzed by immunoelectrophoresis for haptoglobin and ␣ 1 -anti-chymotrypsin using a mixture of the respective antibodies (upper panel). CAT activity was determined and expressed relative to the control-treated cells (lower panel). C, confluent monolayers of IL-10R-HepG2 cells in six-well cluster plates were treated for 15 min with increasing concentrations of IL-10 or IL-6, or with 100 ng/ml cytokines for various lengths of time. EMSA pattern with SIE and TB-2 as probes were prepared. Equal aliquots of the extracts from cells treated for 15 min with 100 ng/ml IL-10 or IL-6 were reacted with C20 anti-STAT3 antibodies prior to EMSA (two lanes at the right). For comparison, extracts of paternal HepG2 cells after 15 min of treatment with 100 ng/ml IL-10 or IL-6 were included. All samples were simultaneously analyzed. The autoradiographic images after identical exposure time (5 h) are shown. D, confluent monolayers of IL-10R-HepG2 cells and paternal HepG2 cells in 24-well culture plates were treated for 24 h with increasing concentrations of IL-10 or IL-6. The amount of secreted haptoglobin (HP), ␣ 1 -acid glycoprotein (AGP), ␣ 1 -anti-chymotrypsin (ACH), and fibrinogen (FB) were determined by immunoelectrophoresis and expressed relative to the control treated cultures in each group. The results of one experimental series are shown. Virtually identical relative cytokine responses were recorded in two additional, independent experimental series. The low magnitude of fibrinogen regulation is due to the high basal level production of that protein by HepG2 cells.
IL-10R exerted IL-6R-like functions. We expected that this regulatory effect was not restricted to co-transfected reporter genes but also applied to the expression of endogenous IL-6regulated genes. The transient transfection procedure yielded a relatively low number of transfected HepG2 cells per culture (see "Experimental Procedures") and, therefore, rendered the analysis of IL-10R effects on endogenous genes in those cells impractical. Hence, we generated HepG2 cells stably expressing the IL-10R by applying a retroviral vector for both the human IL-10R cDNA and neo gene. After 4 weeks of selection, we obtained a cell population (termed IL-10R-HepG2 cells) that expressed the expected 4.5-kilobase fusion MSCV-IL-10R mRNA detectable by Northern blot hybridization (Fig. 6A). The level of IL-10R mRNA appeared to be in the range of that detected in CLL-cells. Lacking reagents to identify IL-10R protein expression on the cell surface, we assessed the presence of IL-10R by its signaling functions that were predicted from the transfection experiments. First, we transiently introduced the IL-6RE-CAT construct into IL-10R-HepG2 cells and compared its regulation by IL-10-and IL-6-type cytokines (Fig. 6B). IL-10 stimulated CAT expression to a level close to that of IL-6. In agreement with the previously reported data on the paternal HepG2 cells (45,46), oncostatin M exerted in IL-10R-HepG2 cells a stimulation of the CAT expression that exceeded the IL-6 action, whereas the LIF response was substantially below. The relative inducing effects of the cytokines on the transfected CAT gene were essentially identical to those on the endogenous type 2 APPs, as shown by the increase in the production of ␣ 1 -anti-chymotrypsin and haptoglobin (Fig. 6B, upper panel).
To identify the similarity of receptor action at the level of STAT activation, IL-10R-HepG2 cells were subjected to treatment with increasing cytokine concentrations or for increasing lengths of time and then analyzed by EMSA for the factors binding to SIE or TB-2 (Fig. 6C). Maximal response was observed after 15 min of treatment with 100 ng/ml IL-10. Primarily the activation of the SIF-A complex by IL-10 was detected, and this complex was recognized by anti-STAT3 in supershift assay. The time course of STAT3 activation was comparable to that mediated by IL-6. Peak activity was measured at 15 min, followed by a reduction with a nadir at 1 h, and a renewed increase to submaximal level over the subsequent 24 h. Notable differences between IL-10 and IL-6 were that IL-6 was effective at 10 times lower concentrations, was capable of activating STAT1 during the initial 30-min treatment period, and activated factors that together with STAT3 contributed to the observed TB-2 binding activity (Fig. 6C). Although IL-10R HepG2 cells are phenotypically different from leukocytes, the patterns of STAT protein activation by the IL-10R and IL-6R in these cell types show remarkable resemblance (compare Fig.  6C with Figs. 1 and 2).
To confirm that the similarity of IL-10 and IL-6 action also applied to the induction of endogenous genes in IL-10R-HepG2 cells, we tested the cytokine dose-dependent stimulation of APP production (Fig. 6D). We selected ␣ 1 -anti-chymotrypsin, haptoglobin, ␣ 1 -acid glycoprotein, and fibrinogen as representative markers. IL-10 increased all four proteins in a dose-dependent fashion. Although the maximal production levels attained were below those stimulated by IL-6, the relative effects of the two cytokines on these gene products were strikingly similar. The dose response indicated that half-maximal stimulation of the APPs was achieved with 2-5 ng/ml IL-10 and with 0.5-2 ng/ml IL-6. The regulation of the APPs proved to be essentially indistinguishable from the paternal HepG2 cells (Fig. 6C), suggesting that retroviral infection, selection in G418, and IL-10R expression did not detectably modify the phenotype of the HepG2 cell population.
Signaling Functions of the IL-10R Requires the Presence of the Box 3 Motifs-The comparative analyses of leukocytes and transfected cells suggested that the IL-10R and IL-6R engage common intracellular signal transduction pathways. The IL-6R-like action of the IL-10R, in particular the activation of STAT3 and induction of IL-6RE-containing genes, was believed to be dependent upon the box 3 motifs in the cytoplasmic domain. By applying a series of deletion mutants of the mouse IL-10R, Ho et al. (11) had already shown that the carboxylterminal region containing the two box 3 sequences is critical for regulating differentiated cell function in the pre B-cell line Ba/F3. We utilized the same expression vectors encoding the mouse IL-10R with progressive carboxyl-terminal deletions (Fig. 7A) for testing the ability of the receptor forms to induce IL-6RE and HRRE-CAT gene expression in transfected HepG2 cells (Fig. 7B) and to activate STAT3 in COS-1 cells (Fig. 7C). Ho et al. had established that each expression vector directed the synthesis and surface display of the comparable levels of IL-10R molecules in transfected cells (11). The cytoplasmic domain region with both box 3 motifs was required for full IL-10R signaling in both cell systems. Removal of the distal box 3 (⌬433-559) resulted in a drastic loss of gene induction and in a reduced STAT3 activation. Truncation past the proximal box 3 motif (⌬402-559) eliminated all signaling activity. These results highlighted that the regulation of STAT3 and gene induction in our cell systems followed the specificity of IL-10R action as defined by Ho et al. in Ba/F3 cells for the expression of cell surface markers (see Fig. 5 in Ref. 11). The similarity suggested that the box 3-containing region of the IL-10R probably exerts equivalent functions in these distinct cell types and that the activation of STAT3 is one of the common processes. The results of Fig. 7 also revealed a significant difference between the IL-10R and hematopoietin receptors such as gp130 or G-CSFR. The membrane-proximal and box 3-deficient intracellular domain of the IL-10R (as present in ⌬402-559) was unable to trigger an induction of the HRRE-CAT construct, whereas the corresponding membrane proximal domain of the G-CSFR on gp130 (containing only box 1 and 2 motifs), while ineffective toward IL-6RE-CAT gene could still mediate a maximal induction of the HRRE-CAT gene (15). We concluded from this observation that the IL-10R must accomplish initiation of signaling by a mechanism that is not identical to that of gp130. The signaling by both receptors, however, converge onto common signal transducing molecules and identical genetic targets. The precise enzymatic processes that underlie these regulatory pathways remains to be fully characterized (10,11,41).
The results (Figs. 4 -7) demonstrate that the transcriptioninducing action of IL-10R is similar to IL-6-type cytokines but is clearly different from IFN␣/␤ or IFN␥ (36,41,42). Ectopically expressed IL-10R, like hematopoietin receptors, promoted proliferation of hematopoietic cells (3,4,11). In normal differentiated cells, such as monocytes and macrophages, IL-10R function has traditionally been considered to be anti-inflammatory by virtue of suppressing expression of genes for cytokines (13,47,48) or nitric-oxide synthetase (49). However, IL-10 has also been noted to induce gene expression such as arginase in macrophages (50) or enhance immunoglobulin expression in IL-2-treated B cells (51). The mode of IL-10R signaling may follow the IL-6R-like mechanisms that we have defined in HepG2 cells. Hence, we would also predict that IL-6 will act like IL-10 in B-cells, e.g. by activating some STAT proteins ( Figs. 1 and 2). Although similar effects of IL-6 and IL-10 have been noted, the mode of action does not appear to be identical, such as in CD40-activated B-CLL cells (52) or in myeloma cells (53).