Suppressor of Cytokine Signaling (SOCS)1 Regulates Interleukin-4 (IL-4)-activated Insulin Receptor Substrate (IRS)-2 Tyrosine Phosphorylation in Monocytes and Macrophages via the Proteasome*

Allergic asthma is a chronic lung disease initiated and driven by Th2 cytokines IL-4/-13. In macrophages, IL-4/-13 bind IL-4 receptors, which signal through insulin receptor substrate (IRS)-2, inducing M2 macrophage differentiation. M2 macrophages correlate with disease severity and poor lung function, although the mechanisms that regulate M2 polarization are not understood. Following IL-4 exposure, suppressor of cytokine signaling (SOCS)1 is highly induced in human monocytes. We found that siRNA knockdown of SOCS1 prolonged IRS-2 tyrosine phosphorylation and enhanced M2 differentiation, although siRNA knockdown of SOCS3 did not affect either. By co-immunoprecipitation, we found that SOCS1 complexes with IRS-2 at baseline, and this association increased after IL-4 stimulation. Because SOCS1 is an E3 ubiquitin ligase, we examined the effect of proteasome inhibitors on IL-4-induced IRS-2 phosphorylation. Proteasomal inhibition prolonged IRS-2 tyrosine phosphorylation, increased ubiquitination of IRS-2, and enhanced M2 gene expression. siRNA knockdown of SOCS1 inhibited ubiquitin accumulation on IRS-2, although siRNA knockdown of SOCS3 had no effect on ubiquitination of IRS-2. Monocytes from healthy and allergic individuals revealed that SOCS1 is induced by IL-4 in healthy monocytes but not allergic cells, whereas SOCS3 is highly induced in allergic monocytes. Healthy monocytes displayed greater ubiquitination of IRS-2 and lower M2 polarization than allergic monocytes in response to IL-4 stimulation. Here, we identify SOCS1 as a key negative regulator of IL-4-induced IRS-2 signaling and M2 differentiation. Our findings provide novel insight into how dysregulated expression of SOCS increases IL-4 responses in allergic monocytes, and this may represent a new therapeutic avenue for managing allergic disease.

Allergic asthma is an immune disorder characterized by elevation of total and specific IgE and infiltration of monocytes, lymphocytes, mast cells, eosinophils, and basophils in the lungs that causes inflammation and wheezing, cough, and dyspnea (1)(2)(3)(4). A complex interplay of genetic and environmental factors contributes to the onset and maintenance of these diseases. Mechanistically, it is known that cytokines secreted from Th2 cells, such as interleukin (IL)-4, IL-5, IL-9, and IL-13, have a pivotal role in dictating the pathology of allergic disease (1,2,5,6). The pathways by which IL-4 and IL-13 exert their biological effects have been a major focus of research and development of therapeutics to block their action through type I and II IL-4 receptors. Previously, we showed that in macrophages, IL-4 engagement of the type I IL-4 receptor resulted in robust tyrosine phosphorylation of insulin receptor substrate (IRS)-2, recruitment of p85 regulatory subunit of PI3K and GRB2, and strong induction of a subset of hallmark M2, also known as M(IL-4) (7), macrophage genes (8,9). In contrast, IL-13 binding to the type II receptor resulted in only modest IRS-2 phosphorylation. Increasing the concentration of IL-13 did not stimulate IRS-2 phosphorylation nor M2 gene expression equal to that elicited by IL-4 in macrophages, indicating IL-4 is the more potent M2 macrophage polarizing cytokine. Numerous studies in mouse models of allergic lung inflammation and in humans with asthma have correlated the presence of M2 macrophages in the lungs and airways with the severity of allergic lung inflammation (10 -13) and poor lung function (14 -17). Understanding how to diminish or prevent M2 macrophage polarization will be a critical step in reducing the severity of allergic lung inflammation.
Therefore, we sought to understand how the activation of IRS-2 by IL-4 was regulated. There are several mechanisms already described that inhibit IL-4 signaling at the level of the cytokine receptors, JAKs, cytoplasmic signaling molecules, or transcription factors within the nucleus. These negative regulators include nuclear and cytoplasmic phosphatases (18 -21), the PIAS (protein inhibitors of activated STATs) (22), the SOCS 3 (suppressors of cytokine signaling) family of proteins (23)(24)(25), and another class of potential JAK-STAT inhibitors consisting of APS, SH2-B, and Lnk adaptor proteins (26,27). The SOCS family of proteins are rapidly induced by JAK-STAT activation, and they inhibit multiple components of the signaling cascade in a negative feedback loop. Eight SOCS proteins, SOCS1-7 and CIS, are reported (28,29). All these proteins contain a central SH2 domain and a C-terminal SOCS box domain (30), which interacts with elongin C and B, Cullins, and RING finger proteins to form an E3 ubiquitin ligase. The active E3 ligase catalyzes the ubiquitination of bound substrate signaling proteins and tags them for degradation (31)(32)(33)(34). Previous studies showed that the two most potent suppressors of signaling, SOCS1 and SOCS3, contain an additional short motif, upstream of the SH2 domain, known as the kinase inhibitory region (KIR). The KIR domain allows SOCS1 and -3 to suppress signaling by direct inhibition of the catalytic activity of the JAKs by acting as a pseudo-substrate to block the active site (35,36).
The amount of expression and activity of different SOCS family members can reciprocally regulate the outcome of polarization of macrophages, T-cells, and dendritic cells (37)(38)(39)(40)(41). SOCS3 regulates development of M1 cells (42)(43)(44), and SOCS1 regulates polarization to the M2 phenotype (45). Furthermore, overexpression studies in HEK293 and Fao cells have suggested that both SOCS1 and SOCS3 can negatively regulate IRS-2 signaling following insulin stimulation (46,47). Based on these studies and our earlier data correlating robust induction of tyrosine-phosphorylated IRS-2 and activation of M2 genes by IL-4, we hypothesized that IL-4-activated IRS-2 may be a target of SOCS action. We investigated how SOCS proteins regulate IL-4-induced IRS-2 signaling and M2 polarization in both human monocytes and mouse macrophages. We found the following: (i) knockdown of SOCS1 prolonged IRS-2 signaling and enhanced M2 gene expression in human monocytes; (ii) association of SOCS1 with IRS-2 increased following IL-4 stimulation; and (iii) IL-4-activated phospho-IRS-2 was subject to degradation by the proteasome. Furthermore, prolonging the activity of IRS-2 by siRNA knockdown of SOCS1 or by proteasomal inhibition enhanced M2 macrophage gene expression. Finally, we show that SOCS1 expression in response to IL-4 is decreased in monocytes from allergic individuals compared with healthy controls. We have provided a deeper understanding of the mechanisms by which IL-4-activated IRS-2 and M2 gene expressions are regulated in monocyte-macrophages. These findings suggest that SOCS1 acting as a negative regulator of IL-4-induced IRS-2 signaling, and M2 polarization may provide new therapeutic avenues in the treatment of Th2-/M2-macrophage inflammatory diseases, such as asthma and allergies.  Stimulation-To examine the kinetics of IL-4 and IL-13 signal transduction in monocytes, we stimulated U937 human monocytes, which express both the type I and type II IL-4 receptor (8), with IL-4 and IL-13. Tyrosine phosphorylation of IRS-2 and STAT6, and phosphorylation of AKT on serine 473, downstream of IRS-2, was assessed by Western blotting (Fig.  1A). We chose a concentration of 10 ng/ml IL-4 based on our earlier work because this concentration elicits maximal IRS-2 tyrosine phosphorylation (8). We found that IL-4 rapidly stim-ulated IRS-2 and STAT6 tyrosine phosphorylation, which peaked by 30 min and declined substantially by 120 min (Fig. 1, B and C). In contrast, IL-13 stimulated only weak IRS-2 tyrosine phosphorylation that was delayed compared with IL-4, despite robust STAT6 tyrosine phosphorylation in agreement with our earlier work (8). Both IL-4 and IL-13 stimulated similar levels of AKT serine 473 phosphorylation, which peaked by 30 min and remained phosphorylated throughout all time points examined (Fig. 1D).

Kinetics of IRS-2 Tyrosine Phosphorylation in Response to
Induction of SOCS Family Member in Response to IL-4 and IL-13 Stimulation-Previous studies found that the SOCS family of proteins is critical for regulating M2 development (42,43,45,48,49) and IRS-1 signaling in response to insulin (47,50,51). Based on our findings that IL-4-induced IRS-2 signaling is rapidly turned on within 30 min and turned off by 180 min, we sought to define whether SOCS family members are involved in this negative regulatory process. To this end, we examined the kinetics of SOCS gene induction in a human U937 monocytic cell line, as well as mouse BMMs. Cells were stimulated with IL-4 or IL-13 and changes in SOCS1, SOCS2, SOCS3, SOCS5, and CIS gene expression were measured by qPCR. In human U937 monocytes, SOCS1 and CIS gene expression was induced by IL-4 (3.4and 1.9-fold, respectively, over unstimulated) within 30 min, peaked by 2 h (10.3-and 4.3-fold, respectively), declined slightly by 3 h but remained elevated until 6 h post-stimulation (Fig. 1E). SOCS1 and CIS gene expression was also induced to a lesser extent by IL-13 (1.8-and 1.5-fold over unstimulated) within 30 min, peaked by 2 h (8.6-and 5.6-fold), and declined gradually to baseline by 6 h (Fig. 1E). SOCS3 was also transiently induced in response to IL-4 (2.2-fold over unstimulated) and IL-13 (3.7-fold over unstimulated), although induction was less robust than SOCS1 and CIS. No increase in SOCS2 or SOCS5 could be detected at any time point. In primary mouse BMMs, we found that Cis gene expression was strongly induced in response to both IL-4 and IL-13 within 30 min, peaked at 4 h (5.4-and 2.4-fold over unstimulated), and remained elevated throughout the 6 h examined (Fig.  1F). Socs1 was also induced in response to IL-4 and IL-13 (5-and 2.5-fold, respectively), although it was delayed (peak at 5 h) compared with Cis gene induction (peaked at 3 h). No increase in Socs3, Socs2, or Socs5 was observed.
Because expression of SOCS1 expression can also be translationally regulated (52), human U937 monocytes were stimulated with IL-4 or IL-13, and SOCS1 and SOCS3 protein expression levels were examined by Western blotting. Expression of SOCS proteins is tightly regulated (52,53), and detection of endogenous SOCS proteins has proven to be exceptionally difficult (54). To detect endogenous SOCS1, U937 human monocytes were treated with 26S proteosome inhibitor, MG-132, at the time of IL-4 stimulation to allow for SOCS protein detection. SOCS1 protein was induced in response to IL-4 stimulation of U937 cells (Fig. 1, G and H). SOCS3 protein was very weakly induced ( Fig. 1, G and H), and CIS could not be detected in whole cell lysates. Taken together, these findings suggest that select members of the SOCS family of proteins, SOCS1, SOCS3, and CIS, are induced in response to IL-4 and IL-13 signaling in human monocytes and mouse macrophages to different degrees. Based on the similar kinetics of SOCS induction and the down-regulation of IRS-2 signaling, we hypothesized that SOCS1, CIS, and SOCS3 may be involved in down-regulating IRS-2 tyrosine phosphorylation in these cell types.
Silencing SOCS1 Prolongs IRS-2 Tyrosine Phosphorylation and Enhances M2 Gene Expression-Because SOCS1, CIS, and SOCS3 were most significantly up-regulated in response to Human monocytic U937 cells were serumstarved for 2 h and stimulated with 20 ng/ml IL-4 or IL-13 for the indicated times. A, cell lysates were prepared and used for immunoprecipitation of IRS-2, and analysis of tyrosine phosphorylation and total IRS-2 was carried out by Western blotting. Whole cell lysates were used for analysis of STAT6 tyrosine 641 phosphorylation and AKT serine 473 phosphorylation by Western blotting. Blots were scanned, and densitometric analysis of phospho (pY)-IRS-2/IRS-2 (B), pY641-STAT6:STAT6 (C), and pS473-AKT/AKT (D) was performed using the ImageJ software (National Institutes of Health) by dividing the band density of the phosphorylated protein by the total protein band density. Data are representative of 2-4 independent experiments. *, p Ͻ 0.05; #, p Ͻ 0.001. U937 (E) and BMM (F) were stimulated with 10 -50 ng/ml IL-4 or IL-13, and induction in SOCS family mRNA was assessed by qPCR and expressed as RFC over unstimulated, n ϭ 2-3, p Ͻ 0.05. * indicates a significant increase in SOCS1; $ indicates a significant increase in SOCS3; and ‡ indicates a significant increase in CIS over unstimulated control. G, in separate experiments, U937 human monocytes were treated with the 26S proteasome inhibitor, MG-132, (5 M) and simultaneously stimulated with 20 ng/ml IL-4. SOCS1 and SOCS3 protein induction was assessed by Western blotting of cell lysates. Data are representative of three independent experiments. H, densitometric analysis of SOCS1 and SOCS3 normalized to ␤-actin. n ϭ 3.

SOCS1 Regulates IL-4-induced IRS-2 Signaling in Human Monocytes
IL-4, we sought to elucidate the regulatory effects SOCS1, CIS, and SOCS3 may have on IL-4-induced IRS-2 phosphorylation, cell signaling, and subsequent M2 polarization. To this end, we used an in vitro siRNA gene silencing system to specifically knockdown expression of the different SOCS in human U937 monocytes and BMM. Nucleofection of siRNAs into U937s reduced IL-4induced mRNA for SOCS1, SOCS3, and CIS by 40, 48, and 56%, respectively ( Fig. 2A). Transfection of siRNAs into BMM inhibited IL-4-induced mRNA expression of Socs1, Socs3, and Cis by 80, 30, and 90% respectively (Fig. 2B). The siRNAs used were highly specific as we did not detect off-target effects of the specific siRNAs against the other SOCS family members. Knockdown of SOCS1 and SOCS3 was validated by Western blotting of whole cell lysates from U937 cells (Fig. 2C). Protein knockdown was more efficient than mRNA knockdown with approximately a 70% decrease in SOCS1 and SOCS3 protein expression.
We then evaluated the effect of SOCS gene knockdown on tyrosine phosphorylation of IRS-2 in response to IL-4 stimula-FIGURE 2. SOCS1 knockdown prolongs the duration of IRS-2 tyrosine phosphorylation and enhances M2 gene expression in response to IL-4. U937 human monocytes were nucleofected (A) and BMM were transfected (B) with siRNAs targeting SOCS1, SOCS3, CIS, or a negative control siRNA (siControl) and then stimulated with 20 ng/ml IL-4 for 24 h. SOCS gene knockdown was assessed by qPCR and expressed as a % of expression compared with the amount in IL-4-stimulated siControl-transfected cells. C, U937 human monocytes were nucleofected, and protein knockdown was assessed by Western blotting of whole cell lysates. Data are mean Ϯ S.E., n ϭ 3, and representative Western blots are shown. D, tyrosine phosphorylation of IRS-2 was analyzed by IP and Western blot as described in Fig. 1A. E, densitometric analysis of phospho-IRS-2 (pY-IRS-2). Data are mean Ϯ S.E. from three independent experiments, presented as a % of IL-4-stimulated value at 60 min, which was set at 100%. * indicates a significant increase in the amount of phospho-IRS-2 in cells transfected with each siSOCS1 (p Ͻ 0.05) compared with the amount of phospho-IRS-2 in siControl-transfected cells. F, cell lysates were prepared, and analysis of STAT6 and AKT phosphorylation was carried out by Western blotting as described in Fig. 1A. G, densitometric analysis of phospho-STAT6 and phospho-AKT, presented as % of the IL-4-stimulated value at 60 min, which was set at 100%. Data are mean Ϯ S.E. from three independent experiments. U937 human monocytes were nucleofected (H), and BMMs were transfected (I) with siRNAs targeting SOCS1, SOCS3, CIS or siControl and stimulated with 20 ng/ml IL-4 for 24 h. Induction of mRNA for the indicated M2 macrophage genes was assessed by qPCR and expressed as RFC over unstimulated. Data are mean Ϯ S.E., n ϭ 3, *, p Ͻ 0.05; **, p Ͻ 0.005. tion in U937 human monocytes. Cells nucleofected with control siRNA displayed the typical kinetics of IRS-2 tyrosine phosphorylation (Fig. 2D, top panel). Knockdown of SOCS1 resulted in significantly prolonged tyrosine phosphorylation of IRS-2 at all time points examined (Fig. 2, D, 2nd panel, and E, upper graph). We did not detect any change in the amount of total IRS-2. In contrast, knocking down SOCS3 and CIS did not prolong tyrosine phosphorylation of IRS-2 ( Fig. 2, D, 3rd and 4th panels, and E, lower graph). Because we did not observe changes in total IRS-2 in the SOCS1 knocked down cells and because SOCS1 is known to regulate JAK activity and STAT signaling (55), we hypothesized that knocking down SOCS1 may have resulted in increased JAK kinase activity, leading to increased phospho-IRS-2 in the SOCS1 knocked down cells. To test this, we measured tyrosine phosphorylation of STAT6, another JAK substrate, to validate that changes in IRS-2 tyrosine phosphorylation were not a result of increased JAK activity. There was no change in the duration or amount of IL-4-induced STAT6 tyrosine phosphorylation in the SOCS1 knocked down cells, compared with control siRNA-treated cells (Fig. 2, F and G), suggesting SOCS1 knockdown is not significantly affecting the activity of JAKs upstream of IRS-2. Knockdown of SOCS1 did, however, modestly prolong AKT phosphorylation (Fig. 2, F and G), a signaling molecule downstream of IRS-2, although this difference was not statistically significant.
Based on findings from our earlier in vitro studies that robust IRS-2 tyrosine phosphorylation corresponded with enhanced M2 gene expression (8), we hypothesized that the prolonged IRS-2 tyrosine phosphorylation observed in the SOCS1 knocked down cells would enhance IL-4-induced M2 gene expression. To address this, we knocked down SOCS1, SOCS3, and CIS in U937s (Fig. 2H) and BMM (Fig. 2I) using the siRNA approach described above and stimulated cells with IL-4 for 24 h. Changes in M2 gene expression were evaluated by qPCR and normalized to expression in the IL-4-stimulated siControl ( Fig. 2H, open bars), set as 100%. Gene knockdown of SOCS1 significantly enhanced the expression of all three hallmark human M2 genes, CD200R (3.51-fold increase), MMP12 (2.53fold increase), and TGM2 (2.7-fold increase), in U937 human monocytes (Fig. 2H, black bars). Similarly, gene knockdown of Socs1 in BMM significantly enhanced expression of all three hallmark mouse M2 genes, Arg1 (2.87-fold increase), Ym1 (3.55-fold increase), and Fizz1 (3.07-fold increase, Fig. 2I, black bars). These data suggest that SOCS1 is a key negative regulator of the IL-4-induced M2 macrophage gene expression program in human monocytes and mouse macrophages. We did observe some increased expression of MMP12 and TGM2 in U937 human monocytes and Arg1 and Fizz1 gene expression in BMM following knockdown of SOCS3 and CIS. Therefore, SOCS3 and CIS may also be involved in the regulation of expression of some M2 genes. However, this regulation either does not involve the JAKs or is downstream of IRS-2, AKT, or STAT6, because siRNA knockdown had no significant effect on the duration or magnitude of phosphorylation of these proteins.
Taken together, these findings suggest that SOCS1 regulates IL-4-induced IRS-2 tyrosine phosphorylation without impacting upstream targets such as the JAKs. Furthermore, the enhanced IRS-2 phosphorylation resulted in enhanced M2 gene expression in both human U937 monocytes and mouse macrophages following IL-4 stimulation.
SOCS1 Interacts with IRS-2 in Human Monocytes-SOCS1 co-immunoprecipitated with both IRS-1 and -2 proteins in response to insulin-stimulated HEK293 cells overexpressing the proteins (46). To address whether SOCS1 and IRS-2 were interacting in IL-4-stimulated U937 human monocytes, we immunoprecipitated IRS-2 and then Western blotted for SOCS1. Indeed, we found SOCS1 associated with IRS-2, and this interaction increased significantly upon IL-4 stimulation (Fig. 3, A and B). This finding is in agreement with the data in the HEK293 overexpression system where the two proteins interact in unstimulated cells, and the interaction is increased upon insulin stimulation (46). There was no immunoprecipitation of SOCS1 with the isotype control antibody. In contrast, no SOCS3 could be detected co-immunoprecipitating with IRS-2 at any time point (Fig. 3A). The reverse co-immunoprecipitation revealed that IRS-2 complexed with SOCS1 at all time points, and more IRS-2 was detected after 30 min of IL-4 stimulation (Fig. 3A). These reciprocal co-immunoprecipitation experiments show that the association of IRS-2 and SOCS1 is increased on IL-4 stimulation in human U937 monocytes.

Down-regulation of Tyrosine-phosphorylated IRS-2 Following IL-4 Stimulation Is Controlled by Ubiquitination-SOCS1
recruits elongin C/B, Cullins, and RBX proteins by virtue of its SOCS box to form an E3 ubiquitin ligase complex. The E3 ligase tags substrate proteins with ubiquitin for proteasomal degradation, thereby controlling target protein expression levels and activation (31,33,56). Previous studies have shown that downregulation of IRS-1 after insulin signaling involves SOCS-directed ubiquitination of the molecule dependent on the SOCS box motif, followed by proteasomal degradation (42,46). Having shown that IRS-2 and SOCS1 co-immunoprecipitate in both monocytes and macrophages, we sought to determine the role of ubiquitination and proteasomal degradation in downregulation of IL-4-induced phosphorylated IRS-2. Human U937 monocytes were either treated with the 26S proteasome inhibitor, MG-132, the 20S proteasome inhibitor, lactacystin, or vehicle control for 30 min and then stimulated with IL-4. Both MG-132 (Fig. 3C) and lactacystin (data not shown) treatment prolonged tyrosine phosphorylation of IRS-2, without changing the amount of total IRS-2. Treatment with MG-132 significantly prolonged tyrosine phosphorylation of IRS-2, even after 180 min (Fig. 3D). We next investigated the degree of ubiquitination of IRS-2 following IL-4 stimulation. We found that poly-ubiquitinated IRS-2 increased following IL-4 stimulation, peaking at 120 min and returning to baseline between 180 and 240 min. The increase in ubiquitination of IRS-2 corresponded with the observed decline in tyrosine-phosphorylated IRS-2 (Fig. 1, A and B). In the presence of MG-132, the amount of ubiquitinated IRS-2 was increased at 240 min compared with vehicle control (Fig. 3, E and F). Because mono-ubiquitinated IRS-2 has been linked to enhanced insulin like-growth factor signaling and mitogenic activity (57), we also examined the mono-ubiquitination status of IRS-2 following IL-4 stimulation. No mono-ubiquitinated IRS-2 could be detected at any time point (data not shown). These findings suggest that following IL-4 stimulation, IRS-2 is poly-but not mono-ubiquiti-

SOCS1 Regulates IL-4-induced IRS-2 Signaling in Human Monocytes
nated to allow for proteasomal targeting and IRS-2 degradation to effectively blunt IL-4 signaling.
Because inhibiting the proteasome prolonged IRS-2 tyrosine phosphorylation, we hypothesized that blocking degradation of the phospho-IRS-2 signal would also enhance M2 gene expres-sion. To this end, we pre-treated human U937 monocytes and BMM with MG-132 for 30 min and then stimulated the cells with IL-4 for 6 and 24 h. Changes in human and mouse M2 gene expression were evaluated by qPCR and normalized to unstimulated controls. Treatment of U937s with MG-132 sig- nificantly enhanced expression of MMP12 (5-fold over vehicle control) and TGM2 (7-fold) after 24 h of IL-4 stimulation (Fig.  3G). Similarly, treatment of BMM with MG-132 significantly enhanced expression of Fizz1 after 6 h (4-fold) and Arg1 (3-fold), Fizz1 (38-fold), and Ym1 (5-fold) after 24 h of IL-4 stimulation (Fig. 3H).
To clearly define the role of SOCS1 in the regulation of IRS-2 signaling by ubiquitination/proteasomal degradation in response to IL-4 stimulation, we examined the poly-ubiquitination status of IRS-2 in U937 human monocytes following siRNA gene knockdown of SOCS1, SOCS3, and CIS. As expected, knockdown of SOCS1 decreased the amount of ubiquitin detected on IRS-2 (Fig. 3, I and J). In contrast, knockdown of SOCS3 did not alter the amount of ubiquitin on IRS-2 at 120 min following IL-4 stimulation. Taken together, these findings suggest that ubiquitination and proteasomal activity are mechanisms that down-regulate phospho-IRS-2 signaling and M2 gene expression in response to IL-4 and that SOCS1 promotes ubiquitination of IRS-2 in IL-4-stimulated cells.
Monocytes from Allergic Donors Failed to Up-regulate SOCS1 in Response to IL-4 Stimulation-Given the role we have defined for SOCS1 in regulating IL-4 signaling and M2 polarization, we hypothesized that there may be differences in SOCS1 expression in monocytes from allergic individuals, compared with cells from healthy controls. We isolated PBMCs from the peripheral blood of healthy and allergic study participants drawn simultaneously and processed in parallel on the same day. PBMCs were plated for 2 h, extensively washed to remove all non-adherent cells, and then stimulated with 20 ng/ml IL-4 for the indicated times. We harvested RNA and protein lysates and examined the kinetics of induction of expression of SOCS family members by qPCR and Western blotting. We found that monocytes from healthy individuals rapidly up-regulated SOCS1 and CIS mRNA within 30 min of IL-4 stimulation (Fig. 4A, circles). Furthermore, the amount of SOCS1 and CIS mRNA induced by IL-4 was significantly higher in healthy monocytes (Fig. 4A, circles), compared with the allergic monocytes (Fig. 4A, squares). Over the 6 h examined, expression of SOCS1 remained significantly elevated (p Ͻ 0.001 with a 95% confidence interval by mixed-effects linear regression analysis) and CIS (p Ͻ 0.001 with a 95% confidence interval by mixed-effects linear regression analysis) in healthy individuals compared with allergic monocytes. This pattern of expression and kinetics mirrors our findings in human U937 monocytes (Fig. 1E). Interestingly, monocytes from allergic individuals up-regulated SOCS3 much more robustly in response to IL-4 stimulation compared with cells from healthy individuals. Because SOCS1 and SOCS3 have previously been shown to reciprocally regulate one another (58), we measured expression of SOCS1 and SOCS3 protein in healthy and allergic monocytes. We saw a significant increase in IL-4-induced expression of SOCS1 protein in the monocytes from healthy donors, although SOCS1 in allergic cells was almost undetectable (Fig. 4, B and C). In agreement with our mRNA expression data, we saw only a small amount of SOCS3 protein in healthy monocytes compared with abundant SOCS3 protein in allergic individuals (Fig. 4, B and C).
To determine whether SOCS1 was interacting with IRS-2 to regulate IL-4 signaling in healthy monocytes but not in allergic monocytes, we immunoprecipitated IRS-2 from IL-4-stimulated healthy and allergic monocytes and then probed for SOCS1. More SOCS1 was detected in the immunoprecipitated lysates from the healthy monocytes compared with the IRS-2 immunoprecipitates from the allergic cells (Fig. 4D). No SOCS3 was detected co-immunoprecipitating with IRS-2 in either healthy or allergic monocytes (data not shown), in line with our findings in U937 human monocytes (Fig. 3A). Because we observed that SOCS1 was regulating ubiquitination of IRS-2 in U937 monocytes, we examined the ubiquitination status of IRS-2 in healthy and allergic monocytes. We found significantly more ubiquitinated IRS-2 in healthy compared with allergic monocytes (Fig. 4, E and F). To determine whether the observed changes in SOCS expression influenced M2 polarization, we measured M2 gene expression in monocytes from healthy and allergic study participants stimulated with IL-4. We found that compared with healthy controls, there was an overall trend to higher expression of M2 genes in the allergic monocytes compared with the healthy monocytes (Fig. 4G). Taken together with our mechanistic findings in the monocytic cell line, these data suggest that monocytes from allergic individuals have diminished regulation of the IL-4-activated IRS-2 signaling pathway and M2 macrophage polarization due to lack of induction of SOCS1 expression.

Discussion
Type 2 inflammation is the driving force of allergic disease and asthma. IL-4 and IL-13 produced by Th2 cells, natural killer cells, basophils, and mast cells have potent effects on lung structural cells, as well as monocyte-macrophages. IL-4 and IL-13 have distinct biological functions, with IL-4 being involved in amplifying the Th2 inflammatory axis through the priming of Th2 cells, whereas IL-13 triggers physiological changes leading to airway hyper-reactivity, mucus hypersecretion, and lung remodeling (59). In humans and mouse models of allergic asthma, M2 macrophages are associated with the severity of allergic asthma and decreased lung function (12,13). In this study, we aimed to define the regulatory mechanisms acting on IL-4 signaling and M2 development. In line with previous findings, we showed that IL-4 stimulated rapid and robust phosphorylation of IRS-2, STAT6, and AKT, whereas IL-13 stimulated only mild IRS-2 tyrosine phosphorylation despite robust phosphorylation of STAT6 and AKT. Taken together with previous reports that IL-4 induces more robust M2 polarization than IL-13 (8, 60, 61), we hypothesized that regulation of IL-4stimulated IRS-2 signaling may control M2 polarization in monocyte-macrophages. To understand the mechanisms that regulate IL-4-stimulated IRS-2 activation, we examined the expression of members of the SOCS family and identified that SOCS1, CIS, and SOCS3 are induced in response to IL-4 in monocytes and macrophages. On a technical note, detection of endogenous SOCS proteins has proven very difficult (54,62), leading many groups to use overexpression systems and recombinant fluorescent or FLAG-tagged SOCS proteins for mechanistic studies. Commercially available SOCS1 antibodies (4H1) could not detect endogenous protein, but rather detected a 37-kDa protein, possibly CIS. Our study is unique in the detection of endogenous SOCS1 and SOCS3, notably in human monocytes without recombinant systems. Targeted gene knockdown revealed that only SOCS1 negatively regulated the duration of IRS-2 activation (tyrosine phosphorylation) in response to IL-4, without significant effects on STAT6 activation in these cells. This suggested that SOCS1 was not affecting the activity of JAKs. AKT activation is downstream of IRS-2, and deletion of SOCS1 modestly enhanced AKT phosphoryla-tion in the U937 cells, consistent with removal of the negative regulator of IRS-2 signaling upstream. Our findings differ from similar SOCS1 knockdown experiments carried out in rat BMM, where the authors report a decrease in phospho-Akt with SOCS1 knockdown, although this decrease was not reported as significant (45). This was attributed to an increase in SOCS3 expression when SOCS1 was knocked down. We did not see increased SOCS3 and decreased AKT phosphorylation in SOCS1 knocked down human U937 cells in our experiments.

. SOCS1 induction is decreased whereas SOCS3 induction is increased in response to IL-4 in human monocytes from allergic individuals compared with monocytes from healthy controls.
Monocytes were isolated from the peripheral blood of healthy and allergic study participants, serum-starved for 2 h, and stimulated with 20 ng/ml IL-4 for the indicated times. A, RNA was harvested, and expression of the mRNA for SOCS1, SOCS3, and CIS was quantified by qPCR. Data are presented as % of IL-4-stimulated cells at 30 min, which was set at 100%. Data are mean Ϯ S.E., n ϭ 3 pairs. *, p Ͻ 0.05; **, p Ͻ 0.005. B, whole cell lysates were prepared, and the amount of SOCS1 and SOCS3 was detected by Western blotting. Representative Western blots from six independent experiments are shown. C, densitometric analysis of SOCS1 and SOCS3 protein, normalized to ␤-actin, showing the ratio of SOCS1/SOCS3 in healthy and allergic monocytes following IL-4 simulation. Data are mean Ϯ S.E. n ϭ 6 healthy, n ϭ 6 allergic. D, in separate experiments, human monocytes were prepared as above, and cell lysates were used for immunoprecipitation of IRS-2, and co-immunoprecipitated SOCS1 was detected by Western blotting. Blots are from one matched patient pair (one allergic and one healthy) and are representative of four independent experiments. E, cell lysates were prepared and used for immunoprecipitation of IRS-2 and analysis of ubiquitination by Western blotting. Blots shown are representative of four sets of participants collected independently. F, densitometry of ubiquitin was normalized to ␤-actin using ImageJ software, and the ratio of ubiquitin (Ub)/IRS-2 in healthy and allergic monocytes was calculated. Data are mean Ϯ S.E. n ϭ 4 healthy, n ϭ 4 allergic, *, p Ͻ 0.05. G, monocytes were stimulated with 20 ng/ml IL-4 for 24 h, and expression of the mRNA for the indicated M2 genes was quantified by qPCR. Data are presented as a % of the healthy control IL-4-stimulated value, which was set at 100%. SEPTEMBER 23, 2016 • VOLUME 291 • NUMBER 39

SOCS1 Regulates IL-4-induced IRS-2 Signaling in Human Monocytes
In addition, species-specific differences in SOCS function may account for our differing results (37). Overall, our findings were further supported by the observation that knocking down SOCS1 expression resulted in an increase in expression in all the M2-associated genes we measured in human monocytes, as well as mouse macrophages. Silencing SOCS3 and CIS caused an increase in only a subset of M2 genes. These findings support the hypothesis that IL-4-stimulated IRS-2 signaling regulated by SOCS1 is a critical determinant of M2 macrophage polarization. Most importantly, we report decreased expression of SOCS1 and increased SOCS3 expression in monocytes from allergic individuals, which may contribute to the enhanced polarization of these cells into M2 macrophages in the setting of allergic inflammation (10 -17). Our findings agree with an earlier study showing monocytes from atopic patients expressed more SOCS3 and more readily differentiated into M2 macrophages (102).
SOCS proteins are known to regulate cytokine signaling through several mechanisms, including steric hindrance of signaling protein-binding sites on cytokine receptor tails, JAK inhibition, as well as targeting substrate proteins for proteasomal degradation by ubiquitination. We found that IRS-2 and SOCS1 co-immunoprecipitate suggesting SOCS1 forms a complex with IRS-2 to regulate IL-4 signaling and M2 polarization. Earlier studies examining insulin-triggered IRS-1 and IRS-2 signaling identified that SOCS1 and SOCS3 are required to limit insulin signaling through ubiquitination and proteasomal degradation of IRS-1/-2. However, many previous studies examining the regulation of IRS-1/-2 by SOCS1/3 relied on overexpression systems. In our experience, overexpression of both the IRS-2 and SOCS proteins yielded non-specific, non-cytokineinducible associations. Our study is uniquely able to dissect apart the role of SOCS1 as a negative regulator of IL-4-induced IRS-2 signaling, whereas no association between SOCS3 and IRS-2 could be detected in any of the cell types we examined. Our findings in primary human monocytes support previous observations that SOCS1 and SOCS3 are reciprocally regulated (42,45,58). Our study also supports the finding that SOCS1 preferentially regulates IL-4-induced IRS-2 signaling (47), whereas SOCS3 may preferentially regulate IRS-1 signaling, as has been observed in response to insulin signaling (51,63,64). In our study, SOCS1-mediated ubiquitination and proteasomal degradation appeared to be one mechanism by which SOCS1 regulates the duration of IL-4-induced phospho-IRS-2 signaling in monocytes and macrophages. Our findings are summarized in Fig. 5.
Dysregulation of SOCS expression has been implicated in several inflammatory diseases, including models of multiple sclerosis (65)(66)(67)(68)(69), rheumatoid arthritis (70,71), systemic lupus erythematosus (72,73), psoriasis (74,75), type 1 diabetes (51, 76 -79), sepsis (80,81), and also in human allergic disease (82)(83)(84). Furthermore, deficiencies in SOCS1 expression have been observed in the smooth muscle in airways (85), and epithelial cells (86) of asthmatic patients and functional SOCS1 polymorphism have been correlated with the development of adult onset asthma and increased IgE levels (87,88). We observed a similar lack of induction of SOCS1 mRNA and protein expression in monocytes from allergic individuals. The poor expres-sion of SOCS1 mRNA may be a result of the Ϫ820G single nucleotide polymorphism in the SOCS1 promoter that binds the transcriptional repressor, YY1 (87), preventing transcription in allergic monocytes or other promoter polymorphisms (88). In mouse models of asthma, decreases in SOCS1 correlated with an increase in SOCS3, suggesting reciprocal regulation of SOCS1 and SOCS3 in allergic disease (89). Indeed, increased SOCS3 expression was noted in peripheral T-cells (82) and eosinophils (90) from asthmatics and in the skin of atopic dermatitis patients (91). We now report a similar increase in SOCS3 expression in human peripheral blood monocytes from allergic individuals. The increased SOCS3 expression that we observed in the allergic monocytes may therefore be a result of diminished SOCS1 expression or vice versa. Because SOCS3 cannot inhibit the activity of JAK3 (34) and therefore type I IL-4 receptor/IRS-2 signaling, the elevated SOCS3 expression in the monocytes from allergic donors would not help control this signaling cascade. From studies of the role of SOCS1 and -3 in rodents (42,43,45,93), the SOCS1/ SOCS3 balance appears to be a critical determinant in the macrophage polarization. Our findings in human monocytes presented here suggest that a balance between SOCS1 and SOCS3 expression exists in these cells. The upset of this balance in monocytes from allergic individuals could promote polarization to M2 macrophages. Consistent with our observations, previous reports have indicated that monocytes from allergic/ In human monocytes and mouse macrophages, IL-4 stimulation triggers rapid STAT6 and IRS-2 tyrosine phosphorylation (within 30 min) leading to nuclear translocation of STAT6 and intracellular signaling through PI3K and AKT, respectively. Increased gene expression of SOCS1, SOCS3, and CIS can be detected within 30 min, peaks at 2 h, and is immediately followed by protein expression. On IL-4 stimulation, the association between SOCS1 and IRS-2 increases, and poly-ubiquitination of phospho-IRS-2 targets it for proteasomal degradation. Blocking the proteasome or reducing SOCS1 expression by siRNA knockdown prolongs the duration of phospho-IRS-2 signaling in response to IL-4. Degradation of phospho-IRS-2 limits the IL-4-triggered tyrosine phosphorylation signaling and M2 gene expression.

SOCS1 Regulates IL-4-induced IRS-2 Signaling in Human Monocytes
asthmatic donors express marker genes of M2 polarization more highly than monocytes from healthy individuals (94,95). Therefore, dysregulation of these negative regulators of Th2 cytokine signaling represents another mechanism of pathogenesis that contributes to allergic inflammation.
Altering the SOCS1/SOCS3 balance or the activity of these negative regulators would be an attractive therapeutic approach and is under investigation (96). A therapeutic SOCS1 KIR mimetic peptide has shown great promise in preventing the development of multiple sclerosis in a mouse experimental autoimmune encephalomyelitis model and even ameliorates clinical symptoms of lupus in a rodent model (97). From our studies presented here, a therapeutic SOCS1 mimetic peptide may have a profound effect on IL-4-initiated IRS-2 signaling and M2 macrophage polarization and the severity of allergic disease by interrupting signaling pathways downstream of IL-4. Indeed, therapies targeting IL-4/IL-4 receptor interaction, namely antibodies targeting the IL-4R␣ as well as cytokineneutralizing antibodies, show promise in preclinical studies; however, they fail to improve both symptom severity and physiological indicators of asthma such as pulmonary function and inflammation in human disease (98 -101). As such, targeting events downstream of the IL-4 receptor may prove to be a more effective way to disrupt the Th2 inflammatory axis. In summary, we have provided insights into the possibility of targeting IRS-2 signaling by SOCS1 to inhibit IL-4-induced M2 macrophage polarization. We have also demonstrated that there is dysregulation of the SOCS1/SOCS3 balance in monocytes from humans with allergic disease, decreased association of IRS-2 with SOCS1, and ubiquitination of IRS-2 and enhanced M2 gene expression. Correcting this imbalance may be a useful new strategy to prevent the differentiation of M2 macrophages, which are correlated with disease severity and asthma exacerbations in humans.
Patient Peripheral Blood Mononuclear Cells-Healthy adult volunteers and allergic volunteers were recruited under the guidelines of the Johns Hopkins Institutional Review Board. All subjects had a positive medical history of allergy. All allergic study participants reported allergy to multiple agents, experienced mild-to-moderate disease, and had experienced symptoms requiring medication within the preceding 12 months (Table 1). Age-, sex-, and race-matched subjects were recruited in pairs, one healthy and one allergic, for blood collection and simultaneous PBMC isolation to minimize day-to-day experimental variation. Data presented in the figures represent matched comparisons between samples collected, processed, and run concurrently. The study was approved by the Johns Hopkins Human Research Ethics Committees, and written informed consent was obtained from each subject. PBMCs were isolated with a Lympholyte H separation protocol according to the manufacturer's recommendations (Cedarlane Laboratories, Hornby, Ontario, Canada). Freshly isolated mononuclear cells were plated in serum-free media for 2 h, and all non-adherent cells were removed by multiple washes in sterile PBS. The remaining adherent cells are typically Ͼ90% purity monocytes. Adherent monocytes were immediately used for cell signaling and gene expression experiments.
Cell Culture-The human monocytic U937 cell line was purchased from the ATCC (Manassas, VA). U937 cells were cultured in RPMI 1640 medium (Gibco) supplemented with 10% FBS, 100 units/ml penicillin plus 100 mg/ml streptomycin (Gibco), and 2 mM L-glutamine. Mouse BMMs were generated from WT C57BL/6 (Charles River, Germantown, MD) mice as described previously (8). Briefly, bone marrow was isolated from femurs and tibias of 4 -6-week old male mice. To deplete adherent stromal cells, the harvested bone marrow was cultured overnight in ␣-minimal essential medium (BioWhittaker Lonza, Walkersville, MD) supplemented with 10% FBS, 100 units/ml penicillin plus 100 mg/ml streptomycin, and 2 mM L-glutamine. Non-adherent mononuclear cells were collected; red blood cell were lysed with ACK lysis buffer, and cells were cultured for 7-10 days in the presence of rmM-CSF (20 ng/ml) to generate BMMs.
Transfection of siRNAs-Pre-designed Silencer Select siRNAs against human and mouse SOCS1, SOCS3, and CIS were purchased from Life Technologies, Inc. Non-targeting siRNA sequence (Control 1) or siRNA targeting GAPDH was used as a negative control. For transient gene knockdown in human U937 cells, a modified version of the Amaxa kit C (Lonza, Cologne, Germany) optimized for U937s was used. Briefly, siRNAs were diluted in Opti-MEM (Gibco Life Technologies, Inc.) to a final concentration of 1 M and were delivered into 3-5 ϫ 10 6 U937 cells by nucleofection using the Amaxa Nucleofector device-II (Amaxa, Cologne, Germany). Following nucleofection, the cells were immediately mixed Real Time RT-PCR-Total RNA was isolated using the RNeasy kit (Qiagen, Valencia, CA) according to the manufacturer's protocol, and complementary DNA (cDNA) was generated with the SuperScript III First Strand Synthesis System (Invitrogen). Quantitative real time RT-PCR analyses were conducted using specific PCR primers in Table 2 and using the ABI 7500 Fast SYBR Green system (Applied Biosystems) using standard conditions as described previously (8,92). Quantitative values were obtained by the cycle number, normalizing genes of interest to HPRT or GAPDH, and determining the relative fold change (RFC) between experimental and control samples (2 Ϫ⌬⌬Ct method). Relative fold change values were compared using t test analyses with significance reached at p Ͻ 0.05. Relative fold change in gene expression in the treatment groups was normalized to expression in the unstimulated or time 0 controls before being compared with the control.
Signaling Experiments-Signaling experiments were performed as described previously (8) with the following modifications. Cells were deprived of serum for 2 h before the addition of cytokine stimuli for the indicated times. Nonidet P-40 lysis buffer was composed of 50 mM HEPES (pH 8.0), 50 mM NaCl, 1% Nonidet P-40, 5 mM EDTA, 10 mM sodium pyrophosphate, 50 mM NaF, 0.25% sodium deoxycholate, 1 mM sodium orthovanadate, 1 mM phenylmethylsulfonyl fluoride (PMSF), and a mixture of 10 different protease inhibitors. Protein G-agarose beads (Invitrogen) were used for immunoprecipitations. For detection of proteins on Western blots, we used horseradish peroxidase-conjugated secondary antibodies (Bio-Rad), and Pierce Clean Blot IP detection reagent (Thermo Fisher, Rockford, IL) or mouse TrueBlot Ultra IP detection re-agent (Rockland Antibodies and Assays, Limerick, PA) for immunoprecipitation experiments. HyGlo (Denville Scientific, Metuchen, NJ) or Clarity Western ECL Substrate (Bio-Rad) was used as the chemiluminescent substrate for visualization of membrane-bound proteins. Blots with darker protein bands were chosen for reproduction quality figures.
Analysis of IRS-2 Ubiquitination-Following immunoprecipitation of IRS-2 for analysis of tyrosine phosphorylation as described (8), protein G-agarose beads were re-boiled in Laemmli buffer, and proteins were resolved by SDS-PAGE. Ubiquitination of IRS-2 was detected following IP by Western blotting using HRP-conjugated mono-ubiquitinated (clone P4D1, Cell Signaling Technologies, Danvers, MA) or poly-ubiquitinated antibodies (clone FK2, Enzo Life Sciences, Farmingdale, NY).
Densitometric Analysis-Shorter exposures of films were chosen for densitometric analysis, so that band intensities were in the linear range of sensitivity of the film. Films were scanned, and the density of the bands on the captured image was analyzed with the Image software, ImageJ (version 1.63f, National Institutes of Health). The amount of phosphoprotein was calculated as a ratio of the density of the band of the phosphorylated protein divided by the density of the band corresponding to the total protein for normalization from films from identical gels run side-by-side.
Statistical Analysis-Average data are expressed as the means Ϯ S.E. from three or more independent experiments. Statistical analysis of U937 human monocytes, BMM, and blot densitometry was performed using a Student's t test. Statistical significance was reached when p Ͻ 0.05. Statistical analysis of SOCS induction in human PBMCs was carried out by general linear model accounting for repeated measures (time) assuming a 95% confidence interval with STATA version 13.1 Soft-  -TGC ACC ACC AAC TGC TTA GC-3Ј  5Ј-GGC ATG GAC TGT GGT CAT GAG-3Ј   Suppressor of cytokine  signaling 1   SOCS1  5Ј-GGA ACT GCT TTT TCG CCC TTA-3Ј  5Ј-AGC AGC TCG AAG AGG CAG TC-