Interleukin-10 Receptor Signaling through the JAK-STAT Pathway

Interleukin-10 (IL-10) is a cytokine that has pleiotropic effects on a variety of different cell types. Although many of the biologic responses induced by IL-10 are also induced by other cytokines, such as IL-6, IL-10 is relatively unique in its ability to potently inhibit production of pro-inflammatory cytokines in macrophages. In this study, we have used gain-of-function and loss-of-function genetic approaches to define the intracellular components involved in the different biologic actions of IL-10. Herein, we demonstrate that the ability of IL-10 to inhibit tumor necrosis factor α (TNFα) production in lipopolysaccharide-stimulated macrophages requires the presence of Stat3, Jak1, and two distinct regions of the IL-10 receptor intracellular domain. Macrophages deficient in Stat3 or Jak1 were unable to inhibit lipopolysaccharide-induced TNFα production following treatment with murine IL-10. Structure-function analysis of the intracellular domain of the IL-10 receptor α chain showed that whereas two redundant Stat3 recruitment sites (427YQKQ430 and477YLKQ480) were required for all IL-10-dependent effects on either B cells or macrophages, expression of IL-10-dependent anti-inflammatory function required the presence on the intracellular domain of the IL-10 receptor of a carboxyl-terminal sequence containing at least one functionally critical serine. These results thus demonstrate that IL-10-induced inhibition of TNFα production requires two distinct regions of the IL-10 receptor intracellular domain and thereby establish a distinctive molecular basis for the developmental versus the anti-inflammatory actions of IL-10.

Interleukin-10 (IL-10) is a cytokine that has pleiotropic effects on a variety of different cell types. Although many of the biologic responses induced by IL-10 are also induced by other cytokines, such as IL-6, IL-10 is relatively unique in its ability to potently inhibit production of pro-inflammatory cytokines in macrophages. In this study, we have used gain-of-function and loss-of-function genetic approaches to define the intracellular components involved in the different biologic actions of IL-10. Herein, we demonstrate that the ability of IL-10 to inhibit tumor necrosis factor ␣ (TNF␣) production in lipopolysaccharide-stimulated macrophages requires the presence of Stat3, Jak1, and two distinct regions of the IL-10 receptor intracellular domain. Macrophages deficient in Stat3 or Jak1 were unable to inhibit lipopolysaccharide-induced TNF␣ production following treatment with murine IL-10. Structure-function analysis of the intracellular domain of the IL-10 receptor ␣ chain showed that whereas two redundant Stat3 recruitment sites ( 427 YQKQ 430 and 477 YLKQ 480 ) were required for all IL-10-dependent effects on either B cells or macrophages, expression of IL-10-dependent anti-inflammatory function required the presence on the intracellular domain of the IL-10 receptor of a carboxyl-terminal sequence containing at least one functionally critical serine. These results thus demonstrate that IL-10-induced inhibition of TNF␣ production requires two distinct regions of the IL-10 receptor intracellular domain and thereby establish a distinctive molecular basis for the developmental versus the anti-inflammatory actions of IL-10.
IL-10 exerts its biologic effects on cells by interacting with a specific cell surface receptor (1,11). Functionally active IL-10 receptors are composed of two distinct subunits. Both subunits belong to the class II cytokine receptor family that also contains the receptors for IFN␣ and IFN␥ (12). The IL-10 receptor ␣ chain is a 110-kDa polypeptide that plays the dominant role in mediating high affinity ligand binding and signal transduction (1,11). The IL-10 receptor ␤ subunit (also known as CRF2-4) is predicted to be a 40-kDa polypeptide that is largely required only for signaling (13,14).
Engagement of the IL-10 receptor has been shown to activate the JAK-STAT signaling pathway. Specifically, IL-10 effects the activation of Jak1 (associated with the IL-10 receptor ␣ chain) and Tyk2 (associated with the IL-10 receptor ␤ chain) and induces the activation of Stat1, Stat3, and, in some cells, Stat5 (15)(16)(17)(18)(19). Previous work from our laboratory using cells from mice with disrupted genes for Jak1 and Stat1 has revealed that the characteristic ability of IL-10 to inhibit TNF␣ production in LPS-stimulated macrophages (i.e. anti-inflammatory actions of IL-10) displays an obligate dependence on Jak1 but does not require the presence of Stat1 (20,21). Thus, whereas the Janus kinases are clearly required for promoting the anti-inflammatory effects of IL-10, it remains uncertain which STAT protein, if any, is involved in mediating these unique IL-10-induced responses.
Other studies have suggested that Stat3 participates in manifesting at least some of the biologic effects of IL-10 on B cells and macrophages. First, Stat3 is directly recruited to two redundant YXXQ sequences in the intracellular domain of the IL-10 receptor following ligand binding (18,22). Receptor mutants lacking these two sequences fail to activate Stat3 and fail to promote a variety of IL-10-dependent responses when expressed in the Ba/F3 pro-B cell line (16,18). Second, overexpression of a dominant-negative Stat3 mutant protein in the J774 murine macrophage cell line inhibited IL-10-dependent antiproliferative responses and partially blocked IL-10-induced expression of the CD32/16 Fc␥ receptor (23). Importantly, these latter developmental responses are also effected by other cytokines that activate Stat3, such as IL-6. However, these same studies failed to find a blocking effect of mutant Stat3 proteins on the anti-inflammatory actions of IL-10. Thus, the existing data suggest either that the IL-10 receptor utilizes different signaling mechanisms to manifest anti-inflammatory versus developmental effects or that different IL-10-induced biologic responses in cells display differential requirements for Stat3 activation.
Herein we demonstrate that macrophages derived from mice engineered to express a genetic Stat3 deficiency in the myeloid cell compartment fail to respond to IL-10 and secrete high levels of TNF␣ upon stimulation with IL-10 plus LPS. These results thus unequivocally establish the requirement of Stat3 for the anti-inflammatory functions of IL-10 in primary cells. Using gain-of-function and loss-of-function receptor mutants, we also define the functionally critical regions on the IL-10 receptor that regulate developmental versus anti-inflammatory functions of this cytokine. The developmental functions map to IL-10 receptor regions that are critical for Stat3 recruitment and that are shared by receptors for other cytokines that activate Stat3, such as IL-6. In contrast, the anti-inflammatory function of IL-10 displays the additional requirement for a carboxyl-terminal 30-amino acid sequence in the intracellular domain of the receptor that contains at least one functionally important serine residue. Thus, the unique anti-inflammatory functions of IL-10 can be explained by distinctive receptor intracellular domain sequences that are not shared by other Stat3 activating cytokine receptors.
Antibodies-GIR-208 is a murine IgG1 monoclonal antibody specific for the human IFN␥ receptor (25), and 9E10 is a murine IgG1 monoclonal antibody specific for a 13-amino acid peptide tag derived from the human c-Myc protein (SMEQKLISEEDLN) (26). These two antibodies were purified and conjugated to biotin using the Enzo biotinylating reagent (Enzo Biochem, Inc.) as described (25).
Primers-The primers listed below were synthesized on an Oligo 1000 DNA synthesizer (Beckman, Fullerton, CA) and were based on the nucleotide sequence of either the human IFN␥ cDNA (27)  Plasmid Construction-Plasmids were constructed using standard procedures (28). The hgr␣/IL-10R␣ chimeric receptor (hgr␣1-434aa/ muIL-10R␣420 -559aa) was created through the use of a two-step polymerase chain reaction as described previously (29). Primers 100741 and 100141 were used as the 5Ј-and 3Ј-primers, whereas primers 100341 and 100242 were the internal primers used to generate the chimeric receptor. The chimeric receptor point mutants (hgr␣/IL-10R␣Y427F, hgr␣/IL-10R␣Y477F, and hgr␣/IL-10R␣Y427F/Y477F) were also generated through the use of a two-step polymerase chain reaction. All chimeric receptor point mutants utilized primers 100741 and 100141 as the 5Ј-and 3Ј-primers. hgr␣/IL-10R␣Y427F was constructed using primers 5393 and 5394 as the internal primers. hgr␣/ IL-10R␣Y477F was created using internal primers 5395 and 5396. In both cases, a Bluescript plasmid (Stratagene, La Jolla, CA) that contained the wild type hgr␣/IL-10R␣ chimeric receptor was used as template. hgr␣/IL-10R␣Y427F/Y477F was generated using internal primers 5395 and 5396. A Bluescript plasmid that contained hgr␣/IL-10R␣Y427F was used as template. We generated a series of hgr␣/IL-10R␣ chimeric receptor truncation mutants by serially truncating 15 amino acids from the carboxyl terminus of the wild type hgr␣/IL-10R␣ chimeric receptor. All of the receptor truncation mutants were generated using primer 100741 as the 5Ј-primer. The 3Ј-primers used to construct hgr␣/IL-10R␣⌬1-4 were 38436, 38437, 38438, and 38439 respectively. The chimeric receptor serine to alanine mutant (hgr␣/IL-10R␣S541A/S544A/S553A/S554A) was generated using primers 100741 and 373356 as the 5Ј-and 3Ј-primers, respectively. All previously described constructs were digested with SmaI/HindIII and subcloned into a Bluescript plasmid that contained the wild type human IFN␥ receptor ␣ chain digested with the same enzymes. All constructs were then subcloned into the pSR␣ expression vector. The accuracy of all polymerase chain reaction-generated DNA was confirmed by automated sequencing (Perkin-Elmer Corp.).
Cells and Cell Culture-RAW264.7 cells, a murine monocyte-macrophage cell line, were obtained from American Type Culture Collection (Manassas, VA) and maintained in Dulbecco's modified Eagle's medium supplemented with 10% fetal calf serum, 2% L-glutamine, 1 mM sodium pyruvate, 50 units/ml penicillin, and 50 g/ml streptomycin. Ba/F3 cells, a murine pro-B cell line, were generously provided by Kevin Moore (DNAX, Palo Alto, CA) and maintained in RPMI 1640 medium supplemented with 10% fetal calf serum, 1% L-glutamine, 1 mM sodium pyruvate, 50 units/ml penicillin, 50 g/ml streptomycin, and 10 ng/ml muIL-3. Primary macrophages derived from fetal liver or bone marrow were prepared using CSF-1 as described (21). Cells were then placed in culture for two days in D-10 supplemented with 5% heat inactivated horse serum. Bone marrow macrophages derived from Stat3-deficient mice were Ͼ99% positive for MAC-1 and Fc␥ receptor as determined by flow cytometry.
DNA Transfection-Cells (1 ϫ 10 7 ) were transfected with 50 g of plasmid by electroporation at 320 V and 960 F (RAW264.7 cells) or at 400 V and 960 F (Ba/F3 cells) on a Bio-Rad gene pulser. Cotransfections were carried out using 5 g of the pMon1118 plasmid (hygromycin resistance) and 45 g of the expression plasmid. Selection with G418 and hygromycin B was begun 48 h after transfection. After selection was completed, cells were sorted based on their expression of both the hgr␣/IL-10R␣ chimeric receptor and the human IFN␥ receptor ␤ chain. Cell lines were subsequently cloned by limiting dilution. Similar results were obtained with bulk-transfected populations.
Both RAW264.7 and Ba/F3 cells were stably transfected with a plasmid (pSR␣, which confers neomycin resistance) (30) encoding the wild type human IFN␥ receptor ␤ chain (tagged at the amino terminus with a 13-amino acid peptide derived from c-Myc) (30). These cells were then stably transfected with a plasmid encoding either wild type or mutant forms of the hgr␣/IL-10R␣ chimeric receptor and the plasmid pMON1118, which confers hygromycin resistance (30). RAW264.7 cells transfected with different forms of the chimeric receptor and the human IFN␥ receptor ␤ chain were selected on G418 (Life Technologies, Inc., 1.0 mg/ml active compound) and 0.5 mg/ml of hygromycin B (Calbiochem). Ba/F3 cells transfected with receptor constructs were selected using G418 (1.0 mg/ml active compound) and 1.3 mg/ml hygromycin B.
Demonstration of Receptor Expression in Transfected Cell Lines-Flow cytometry for the hgr␣/IL-10R␣ chimeric receptor and receptor mutants or the human IFN␥ receptor ␤ chain was conducted using the biotinylated forms of GIR-208 (anti-human IFN␥ receptor) and 9E10 (anti-c-Myc peptide) respectively and streptavidin-phycoerythrin conjugate (Chromoprobe, Redwood City, CA) as described (31). Cells were analyzed on a Becton Dickinson FACScan.
Electrophoretic Mobility Shift Assay-Cells (5 ϫ 10 6 ) were washed once and then resuspended in 500 l of medium. 1000 units/ml murine or human IFN␥ or 200 ng/ml muIL-10 was added to the cells and incubated for 7 min at 37°C. The cells were analyzed by electrophoretic mobility shift assay as described previously (32) using an 18-base pair oligonucleotide probe that contained the gamma response region of the Fc␥RI gene.
Ba/F3 Proliferation Assay-Ba/F3 cells were washed twice in supplemented RPMI 1640 medium that lacked muIL-3. The cells were seeded in a 96-well plate at a density of 2 ϫ 10 4 cells/well and rested for 3 h in the absence of muIL-3. The cells were then incubated with varying amounts of huIFN␥ or muIL-10 in a total volume of 150 l of medium. Cells were incubated for 48 h at 37°C and an 3-(4,5-dimethyl thiazol-2-yl)-2,5-diphenyl tetrazolium bromide assay was performed as described previously (33). Experiments were performed at least three times using multiple clones of each transfected cell line. Similar results were obtained using transfected bulk populations.
RAW264.7 Bioassay-RAW264.7 cells were washed once, seeded in 96-well plates at a density of 5 ϫ 10 4 cells/well, and incubated with varying amounts of huIFN␥, muIL-6, or muIL-10 in a total volume of 150 l supplemented Dulbecco's modified Eagle's medium. Cells were incubated for 1 h at 37°C and then treated with 10 ng of LPS/well. The cells were placed at 37°C for 24 h, at which point the culture supernatants were harvested, and TNF␣ levels were quantitated via a TNF␣ ELISA (34). Experiments were performed at least three times using multiple clones of each transfected cell line. Similar results were ob-tained using transfected bulk populations.
Peritoneal exudate cells (5 ϫ 10 4 ) derived from naive female BALB/c ByJ mice were plated in each well of a 96-well tissue culture plate. Cells were allowed to adhere for 3 h. Adherent peritoneal exudate cells were then washed twice with warm medium and treated as described above.
Macrophages derived from Stat3-, Stat1-, and Jak1-deficient mice were cultured as described above. Cells were incubated with different doses of muIL-10 for 12 h. LPS was added at a final concentration of 2 g/ml, and the cells were cultured for and additional 24 h. Culture supernatants were harvested, and TNF␣ levels were quantitated via a TNF␣ ELISA.

RESULTS
Stat3 Is Necessary but Not Sufficient for IL-10-dependent Inhibition of TNF␣ Production by Macrophages-To determine whether Stat3 is required for the anti-inflammatory actions of IL-10, we monitored the ability of this cytokine to inhibit LPSinduced TNF␣ production in macrophages derived from mice with a genetic deficiency of Stat3 targeted to myeloid cells (35). IL-10 treatment of wild type macrophages resulted in a dosedependent inhibition of LPS-induced TNF␣ production that reached maximal levels at a 1 ng/ml dose of IL-10 ( Fig. 1). In contrast, macrophages derived from mice with a myeloid cell Stat3 deficiency were unable to inhibit LPS-induced TNF␣ production at any dose of IL-10 used. As additional controls, macrophages from Jak1 Ϫ/Ϫ and Stat1 Ϫ/Ϫ mice were also tested in these experiments for IL-10 sensitivity (21,36). In agreement with our previous reports (20,21), Jak1 Ϫ/Ϫ macrophages were unresponsive to IL-10, whereas the IL-10 response in Stat1 Ϫ/Ϫ macrophages was indistinguishable from that of wild type mice. Thus, Stat3, together with Jak1, is obligatorily required for IL-10-dependent inhibition of TNF␣ production in LPS-stimulated macrophages.
Whereas these results demonstrated that Stat3 was required for mediating the anti-inflammatory actions of IL-10, they did not reveal whether it was sufficient to induce these functions. To address this issue, we compared the anti-inflammatory action of IL-10 to that of IL-6, a cytokine that utilizes a distinct receptor system but that also activates Stat3 and Jak1. Pretreatment of RAW264.7 cells with increasing doses of IL-10 resulted in a dose-dependent inhibition of TNF␣ production, reaching a maximal level of 75% at a dose of 200 ng of the cytokine ( Fig. 2A). In contrast, RAW264.7 cells pretreated with IL-6 were only slightly inhibited in their ability to produce TNF␣ (maximal level of inhibition, 30% at a 200-ng dose of IL-6). We also compared the anti-inflammatory effects of IL-10 and IL-6 on primary murine macrophages. Here again, whereas IL-10 displayed potent inhibitory activity on TNF␣ production, IL-6 did not (Fig. 2B). These functional differences could not be ascribed to different levels of Stat3 activation because the kinetics and ultimate magnitude of Stat3 activated by IL-10 and IL-6 in either RAW264.7 cells or primary macrophages was identical (data not shown). Thus, the JAK-STAT pathway is required but not sufficient for IL-10-dependent inhibition of LPS-induced TNF␣ production.
Demonstration That IL-10-dependent Biologic Responses in Macrophages and B Cells Can Be Manifest by a Gain of Function hgr␣/IL-10R␣ Chimeric Receptor-To identify functionally critical amino acid residues within the intracellular domain of the muIL-10 receptor ligand binding chain involved in mediating the anti-inflammatory versus developmental effects FIG. 1. Stat3 and Jak1 but not Stat1 are required for IL-10-dependent inhibition of TNF␣ production by macrophages. Macrophages derived from wild type, Stat1 Ϫ/Ϫ , Stat3 Ϫ/Ϫ , and Jak1 Ϫ/Ϫ mice were pretreated with varying concentrations of muIL-10 for 12 h. The cells were treated with 2 g/ml LPS and incubated for an additional 24 h. Supernatants were harvested, and TNF␣ levels were quantitated via ELISA. Percentage of inhibition of TNF␣ production upon IL-10 stimulation is indicated.
FIG. 2. The JAK/STAT pathway is necessary but not sufficient for the anti-inflammatory actions of IL-10. A, RAW264.7 cells were pretreated with varying doses of either muIL-10 or muIL-6 for 1 h. Subsequently, the cells were treated with 60 ng/ml LPS and incubated for 24 h. Supernatants were harvested, and TNF␣ levels were quantitated by ELISA. The graph indicates percentage of inhibition of TNF␣ production. B, peritoneal exudate cells were isolated from naive female BALB/c ByJ mice. Cells were plated and allowed to adhere for 4 h. The cells were then washed once and pretreated for 1 h with various doses of either muIL-10 or muIL-6. The cells were stimulated with LPS (60 ng/ml) for 24 h. TNF␣ levels in the supernatant were quantitated by ELISA.
of IL-10, we generated a chimeric receptor that consisted of the human IFN␥ receptor ␣ chain extracellular and transmembrane domains and the first 184 amino acids of the intracellular domain (i.e. truncated just above the Stat1 docking site) attached to a 140-amino acid region of the murine IL-10 receptor carboxyl terminus, which contains the Stat3 docking sites (Fig. 3). The resulting chimeric receptor thus consists of the human IFN␥R␣ chain that retained the Jak1 binding site and lacked the Stat1 docking site but now contained the two redundant IL-10 receptor Stat3 docking sites. The chimeric polypeptide was then expressed in murine RAW264.7 macrophages or Ba/F3 pro-B cells that had been engineered to also stably express the human IFN␥ receptor ␤ chain (Table I).
Engagement of the endogenous IL-10 receptor on RAW264.7 cells with muIL-10 led to a dose-dependent inhibition of TNF␣ production in LPS-treated cells (Fig. 4A). HuIFN␥ has no effect on these cells because of the strict species specificity that governs the interaction of IFN␥ with its receptor. Treatment of RAW264.7 cells expressing the hgr␣/IL-10R␣ chimeric receptor with human IFN␥ induced a dose-dependent inhibition of LPSdependent TNF␣ production in a manner that was quantitatively identical to that induced by the activated, endogenous IL-10 receptor. In contrast, treatment of wild type or transfected RAW264.7 cells with murine IFN␥ caused a dose-dependent increase of TNF␣ production (data not shown). In a similar manner, Ba/F3 engineered to express the wild type muIL-10 receptor proliferated in the presence of muIL-10 ( Fig.  4B) but did not respond to human IFN␥. In contrast, Ba/F3 cells bearing the hgr␣/IL-10R␣ chimeric receptor proliferated upon exposure to human IFN␥. Thus, transfer of IL-10 receptor sequences containing the Stat3 recruitment sites to the IFN␥ receptor results in the generation of a modified IFN␥ receptor that promotes IL-10-dependent rather than IFN␥-dependent responses in murine macrophages and B cells exposed to human IFN␥.

Tyrosine Residues 427 and 477 within the Intracellular Domain of the Chimeric hgr␣/IL-10R␣ Are Redundantly Required for the Development of IL-10-like Effects on Macrophages and B
Cells-In a previous report, we showed that tyrosine residues at positions 427 and 477 in the intact murine IL-10 receptor played redundant roles in recruiting Stat3 to the activated receptor (18). In addition, we found that Stat3 recruitment was required for development of IL-10-induced proliferative responses (i.e. developmental responses) in B cells. To examine whether the anti-inflammatory functions of IL-10 also show the same requirement for these Stat3 recruitment sites and to test whether the functionally redundant tyrosines act in a similar manner in the chimeric receptor, we generated a set of tyrosine to phenylalanine mutant chimeric receptors, expressed them in macrophages and B cells (Table I) and examined their capacity to promote IL-10 like biologic effects on the transfected cells.
Similar results were obtained when the three mutant chimeric receptor proteins were expressed in Ba/F3 cells. Ba/F3 cells expressing either chimeric receptor point mutant proliferated in response to huIFN␥ (Fig. 4B) and activated Stat3 (data not shown) in a manner that was comparable to or greater than that of the unaltered chimeric receptor. In contrast cells bearing the double point mutant chimeric receptor failed to manifest proliferative responses to huIFN␥ and failed to activate Stat3. Thus, the IL-10 receptor residues Tyr-427 and Tyr-477 function in a redundant manner to induce both IL-10 like anti-inflammatory responses in macrophages as well as developmental responses in B cells. These results demonstrate that the chimeric receptor faithfully recapitulates the known functions of the full-length IL-10 receptor and thereby validates the use of the chimeric receptor as a suitable experimental model system.
Distinct Residues within the Carboxyl Terminus of the IL-10 Receptor Are Required to Inhibit LPS-induced TNF␣ Production but Not for the Induction of an IL-10 Proliferative Re-sponse-The data presented this far show that the JAK-STAT signaling pathway and specifically Stat3 and Jak1 are required for induction of IL-10-dependent anti-inflammatory and developmental responses in macrophages and B cells. However, they did not reveal a mechanism by which IL-10 could uniquely manifest its anti-inflammatory effects. Thus, at least one signal in addition to IL-10 must arise from the activated IL-10 receptor in stimulated macrophages. To explore this possibility, we generated a family of progressive deletion mutants of the hgr␣/IL-10R␣ chimeric receptor such that each sequential mutant was 15 amino acids shorter at the carboxyl terminus compared with its predecessor. Each receptor was expressed in RAW264.7 and Ba/F3 cells (Table I and Fig. 3). Expression of each receptor in the resulting stably transfected cell populations was confirmed by FACS analysis, and several clones of each transfected cell line were chosen for further analysis based on their expression of comparable levels of receptor. HuIFN␥ treatment of RAW264.7 cells bearing either intact hgr␣/IL-10R␣ or the chimeric receptor lacking the most carboxyl-terminal 15 residues (hgr␣/IL-10R␣⌬1) produced a dose-dependent inhibition of LPS-stimulated TNF␣ production (Fig.  5A). However, RAW264.7 cells bearing receptor mutants lacking additional carboxyl-terminal sequence (hgr␣/IL-10R␣⌬2-hgr␣/IL-10R␣⌬4) were unable to mediate this response. In contrast, all of the receptor carboxyl-terminal deletion mutants were able to support IL-10-dependent proliferative responses in Ba/F3 cells (Fig. 5B). Thus, additional residues within the carboxyl terminus are used by the IL-10 receptor to inhibit TNF␣ production in macrophages but not proliferative responses in B cells.
The basis for the activity differences observed with the various deletion mutants in RAW264.7 versus Ba/F3 cells was not due to differences in Stat3 activation. As determined by electrophoretic mobility shift assay analysis, huIFN␥ induced comparable levels of Stat3 homodimers (slower migrating band) and Stat1:Stat3 heterodimers (faster migrating band) in each cell type regardless of the presence of the carboxyl-terminal 60 amino acids of the chimeric receptor (Fig. 6, A and B).
Serine Residues within the Carboxyl Terminus of the muIL-10 Receptor Are Required for Inhibition of TNF␣ Production by Macrophages-In an attempt to identify specific amino acids in the carboxyl-terminal region of the IL-10 receptor that are required for inhibition of TNF␣ production, we generated one final hgr␣/IL-10R␣ chimeric receptor mutant in which all of the carboxyl-terminal serine residues (i.e. those residing at positions 541, 544, 553, and 554) were mutated to alanines and stably expressed it in RAW264.7 cells (Table I). HuIFN␥ treatment of RAW264.7 cells bearing this mutant failed to inhibit LPS-mediated TNF␣ production. The minimal levels of TNF␣ inhibition manifest by this mutant were indistinguishable from those produced by IL-6 ( Fig. 7A). Electrophoretic mobility shift assay analysis showed comparable Stat3 activation by huIFN␥ treated RAW264.7 cells bearing either full-length chimeric receptor or the receptor lacking the carboxyl-terminal serines (Fig. 7B). Thus, IL-10-dependent inhibition of LPS-induced TNF␣ production in macrophages requires the concomitant presence of Stat3 docking sites on the receptor and a restricted carboxyl-terminal portion of the receptor that contains at least one functionally important serine residue. DISCUSSION The focus of this study was to characterize the mechanism that underlies the selective ability of IL-10 to inhibit TNF␣ production by LPS-stimulated macrophages. Using cells from mice with a genetic Stat3 deficiency in the myeloid cell compartment, we demonstrated that Stat3 is obligatorily required for expression of the inhibitory functions of IL-10 on TNF␣ production. Taking into account our previous study (21), which showed an obligate role for Jak1 in this process, we can now unequivocally conclude that the JAK-STAT signaling pathway is required for expression of the anti-inflammatory actions of IL-10. In addition, we find that the developmental versus antiinflammatory actions of IL-10 are distinguishable by their re-quirements for different regions of the IL-10 receptor ␣ chain intracellular domain. For the former, the functionally critical receptor region maps selectively to a redundant set of Stat3 docking sites residing at positions 427-430 and 477-480. In contrast, for the latter, the additional presence on the receptor of a carboxyl-terminal sequence that contains at least one

FIG. 4. Demonstration that IL-10-dependent biologic responses in macrophages and B cells can be manifest by a gain of function hgr␣/IL-10R␣ chimeric receptor.
A, RAW264.7 cells bearing the wild type IL-10 receptor, the intact chimeric receptor, or distinct chimeric receptor intracellular domain mutants were pretreated with various doses of IL-10 or huIFN␥ for 1 h and subsequently treated with 60 ng/ml of LPS for 24 h. Again, TNF␣ concentrations were determined by ELISA. The graph depicts percentage of inhibition of LPS-induced TNF␣ production. B, Ba/F3 cells bearing the wild type IL-10 receptor, the intact chimeric receptor, or distinct chimeric receptor intracellular domain mutants were were rested in medium lacking muIL-3 for 3 h. The cells were then treated with increasing doses of either muIL-10 or huIFN␥, respectively. The cells were incubated for 48 h, and subsequently, 50 l of 3-(4,5-dimethyl thiazol-2-yl)-2,5-diphenyl tetrazolium bromide at a concentration of 2.5 mg/ml was added to the cells. The cells were then incubated at 37°C for 2 h. The dark blue formazan crystals were dissolved by adding 100 l of lysis solution. The figure represents viable cells as determined by A 570 .
FIG. 5. Distinct residues within the carboxyl terminus of the IL-10 receptor are required to inhibit LPS-induced TNF␣ production but not for the induction of an IL-10 proliferative response. A, RAW264.7 cells bearing the wild type muIL-10 receptor, the intact hgr␣/IL-10R␣ chimeric receptor, or the chimeric receptor carboxyl-terminal truncation mutants were pretreated with various doses of muIL-10 or huIFN␥ for 1 h. The cells were then treated with 60 ng/ml LPS and incubated for an additional 24 h. TNF␣ concentrations were quantified via ELISA. This panel depicts percentage of inhibition of LPS-induced TNF␣ production. B, Ba/F3 cells bearing the wild type muIL-10 receptor, the intact hgr␣/IL-10R␣ chimeric receptor, or the chimeric receptor carboxyl-terminal truncation mutants were rested in medium lacking muIL-3 for 3 h. The cells were then treated with increasing doses of either muIL-10 or huIFN␥. The cells were incubated for 48 h, and subsequently, 50 l of 3-(4,5-dimethyl thiazol-2-yl)-2,5-diphenyl tetrazolium bromide at a concentration of 2.5 mg/ml was added to the cells. The cells were then incubated at 37°C for 2 h. The dark blue formazan crystals were dissolved by adding 100 l of lysis solution. The figure represents viable cells as determined by A 570 . functionally important serine residue is also required. Our results thus provide a molecular explanation for the unique biologic actions of IL-10 that are not expressed by other cytokine receptors, such as the IL-6 receptor, that activate similar patterns of JAKs and STATs.
Based on previous work from our laboratory and others (15,16,18), the IL-10 receptor was known to activate Stat3, Jak1 and Tyk2. At least two of these signaling components (i.e. Stat3 and Jak1) are also activated by other cytokine receptors, such as those that belong to the IL-6 receptor family. In fact, IL-6 induces many of the same biologic responses in cells as are induced by IL-10. Specifically, both induce proliferative responses in B cells and antiproliferative responses in macrophages as well as inducing expression of a variety of cell surface markers on these cells, such as HSA and CD32/16, respectively (37)(38)(39)(40)(41)(42). Induction of these functions has been shown to require Stat3 recruitment and activation by the ligand assembled IL-6 and IL-10 receptors. In contrast, as shown in the current study, the ability of IL-10 to inhibit LPS-induced TNF␣ production in macrophages is not significantly shared by IL-6. Thus the latter represents a biologic response that is IL-10-specific.
Based on this argument, we considered two possibilities that could explain the distinctive actions of IL-10. The first is that the JAK-STAT signaling pathway is required but not sufficient for expression of IL-10-dependent anti-inflammatory actions and that at least one additional signal must be concomitantly delivered to the cell through the ligated IL-10 receptor. The second is that the anti-inflammatory actions of this cytokine are induced through a different signaling pathway. The latter possibility has been ruled out by the current study, which shows that the ability of IL-10 to inhibit LPS-induced TNF␣ production in macrophages requires the concomitant presence in cells of Stat3 and Jak1. Moreover, as presented elsewhere (35), mice lacking Stat3 in the myeloid compartment display enhanced spontaneous inflammatory responses similar to those observed in mice lacking either IL-10 (43) or the CRF2-4 component of the IL-10 receptor (13). These mice develop chronic enterocolitis and exhibit exaggerated Th1 responses characterized by a 4 -5-fold increase in IFN␥ production as compared with control animals. Thus, taken together, these results show that Stat3 is required for expression of IL-10 anti-inflammatory functions both in vitro and in vivo. In addition, the data presented in the current manuscript directly support the possibility that an accessory signaling pathway in addition to the JAK-STAT pathway is required for expression of the anti-inflammatory actions of IL-10. Specifically we show that a second region of the IL-10 receptor intracellular domain is required in addition to the Stat3 recruitment sites on the receptor to mediate the capacity of IL-10 to inhibit LPS-induced TNF␣ production in macrophages.
This conclusion is thus in conflict with that of O'Farrell et al. (25), who reported that IL-10-dependent inhibition of TNF␣ production occurs in a Stat3-independent manner. The conclusions reached in the latter study were derived from two types of experiments. First, macrophages were engineered to express a tagged form of Stat3 containing the bacterial gyrase B protein (gyrB-Stat3) that could be inducibly dimerized within the cell following addition of the antibiotic coumermycin (44). Whereas this manipulation led to inhibition of macrophage proliferation, it did not effect LPS-induced TNF␣ production. However, based on the data presented in our study, we now know that Stat3 is required but not sufficient for induction of the anti-inflammatory actions of IL-10, and thus the coumermycin experiments delivered only one of the two signals needed to inhibit TNF␣ production. Second, J774 macrophages were engineered to overexpress a dominant negative mutant form of Stat3 lacking the carboxyl terminus. Whereas these cells became either totally or partially insensitive to many of the actions of IL-10, they still displayed IL-10-induced inhibition of LPS-dependent TNF␣ production. However, the cells that overexpressed mutant Stat3 were not completely blocked in their capacity to activate (i.e. phosphorylate) their own endogenously expressed Stat3. Thus, the lack of effect of the dominant negative Stat3 mutant on the anti-inflammatory functions of IL-10 may reflect the incomplete blockade of Stat3 activation and different threshold requirements for activated Stat3 for induction of various biologic responses. It is also possible that for induction of IL-10 anti-inflammatory responses, Stat3 may function as an adapter protein that facilitates recruitment of another signaling component to the activated receptor. In this case, a Stat3 mutant that could not be tyrosine-phosphorylated maintained a functional SH2 domain but could, through its ability to bind to the activated IL-10 receptor, still be able to fulfill a role as an adapter protein.
A key approach used in our present study was the employment of a gain-of-function chimeric cytokine receptor in which the murine IL-10 receptor Stat3 recruitment sites and carboxyl-terminal region were substituted for the Stat1 recruitment site on the human IFN␥ receptor. We specifically chose the human IFN␥ receptor ␣ chain as the recipient of the IL-10 FIG. 6. Carboxyl-terminal chimeric receptor truncation mutants maintain huIFN␥-induced STAT activation. A, RAW264.7 cells bearing the wild type murine IL-10 receptor, the intact chimeric receptor or chimeric receptor truncation mutants were stimulated with medium, 200 ng/ml muIL-10, or 25 ng/ml huIFN␥ for 7 min at 37°C. The cells were then lysed, and the nuclear extracts were generated. 5 g of extract was incubated with the gamma response region of the Fc␥RI gene probe for 30 min. The protein/DNA binding complexes were imaged by audioradiography. B, Ba/F3 cells bearing the wild type murine IL-10 receptor, the intact chimeric receptor, or chimeric receptor truncation mutants were treated as described above, and the protein/DNA binding complexes were imaged by audioradiography. receptor sequence for three reasons. First, the ␣ chains of the IFN␥ receptor and the IL-10 receptor belong to the same class II cytokine receptor family. Second, although the polypeptides belong to the same receptor family, they differ not only in the STATs they recruit but also in the biologic responses they induce. Third, the use of the huIFN␥ receptor extracellular domain allowed us to take advantage of the strict speciesspecific receptor binding activity exhibited by human and murine IFN␥, thus providing us with a mechanism to selectively monitor the activity of the chimeric receptor expressed on murine cells through the addition of huIFN␥.
Based on the data presented herein, we envision two possible models by which the different regions of the IL-10 receptor may function. In the first, the receptor may deliver two distinct signals that arise independently through interaction of different proteins with the two functionally important regions of the receptor intracellular domain. In this model, one of these signals is Stat3, which binds to the Stat3 docking sites on the IL-10 receptor, becomes tyrosine-phosphorylated, dimerizes, and then translocates to the nucleus, where it functions as a transcriptional activator. The second signal is then derived from an additional protein that docks at the carboxyl terminus of the IL-10 receptor, possibly by binding to serine or phosphoserine residues. In the second model, the association of Stat3 with the ligand-induced Stat3 docking site on the receptor may result in the formation of a binding site for another protein that also requires interaction with the carboxyl-terminal sequence of the IL-10 receptor. In this model, Stat3 would function as an adapter protein. An adapter role for Stat3 has been previously reported for signaling by the IFN␣/␤ receptor, where Stat3 binding to the IFNAR1 provides a binding site on the receptor for phosphatidylinositol 3-kinase. Our current efforts are focused on defining the nature of the second signal needed for induction of IL-10-dependent anti-inflammatory responses and in determining the mechanism by which it is recruited to the activated IL-10 receptor. FIG. 7. Serine residues within the carboxyl terminus of the muIL-10 receptor are required for inhibition of TNF␣ production by macrophages. A, RAW264.7 cells bearing the wild type murine IL-10 receptor, the intact chimeric receptor, or chimeric receptor serine to alanine mutant were pretreated with various doses of muIL-10, muIL-6, or huIFN␥. The cells were then treated with 60 ng/ml LPS and incubated for an additional 24 h. TNF␣ concentrations were quantified via ELISA. This panel depicts percentage of inhibition of LPS-induced TNF␣ production. B, an electrophoretic mobility shift assay was performed on these cells. 5 g of extract was incubated with the gamma response region of the Fc␥RI gene probe for 30 min. The protein-DNA binding complexes were imaged by audioradiography.