Unique Expression of Suppressor of Cytokine Signaling 3 Is Essential for Classical Macrophage Activation in Rodents In Vitro and In Vivo¹

Macrophages have critical roles in almost all aspects of inflammation from initiation to resolution (1, 2, 3). Resident tissue macrophages sense injury and recruit circulation leukocytes. Infiltrating macrophages respond to microenvironmental signals by engaging distinct activation programs that whenever possible restore normal anatomy and function either by destroying microorganisms or by promoting healing, and when this is not possible, by inducing scarring. Knowledge of the range of macrophage activation states and the cues that induce polarization is far from complete, but two broad types of macrophage activation have been characterized in detail: classical and alternative activation. Classically (or M1)⁴ activated macrophages are induced by the combination of IFN-γ and proinflammatory stimuli (such as LPS or TNF-α) and have anti-microbial and cytotoxic properties, whereas alternative (or M2) activated macrophages are anti-inflammatory or reparative. The precise properties of M2 macrophages vary depending on the activating conditions, and they have been separated into M2a (IL-4- or IL-13-induced), M2b (immune complexes in combination with IL-1β or LPS), and M2c (IL-10, TGF-β, or glucocorticoids) subtypes (1, 2, 3).

The number and diversity of receptors on macrophages and the complexity of the gene expression patterns they induce (4) emphasizes the need for stringent control mechanisms. Understanding how macrophages integrate the multiple activating signals they are exposed to on infiltrating damaged tissue is a major challenge. One of the key ways macrophages integrate contrasting signals was revealed by in vitro studies that demonstrate classical activation with IFN-γ and TNF-α or LPS induces unresponsiveness to IL-4 and other alternative activating stimuli; similarly, alternative activation with IL-4 induces unresponsiveness to classical activating stimuli (5, 6). Such programmed unresponsiveness is determined by the macrophage’s first encounter with an activating cytokine and markedly constrains responses to subsequent signals.

Programmed unresponsiveness focuses activation of individual macrophages in a particular direction not only in vitro but also in vivo as demonstrated in models of glomerulonephritis (7, 8). Consequently, it is not surprising that macrophages infiltrating inflamed glomeruli can be activated in different ways, presumably stochastically as a result of their initial cytokine contact in the complex microenvironments within inflamed tissue (8). Understanding the molecular mechanisms for such unresponsiveness could unlock new therapeutic approaches and here we have examined the potential role of one of the prime candidates – suppressors of cytokine signaling (SOCS) proteins.

There are eight SOCS family proteins in mammals, and which family members are expressed varies with the cell type. Macrophages express SOCS1 and SOCS3 (9, 10), and both are rapidly inducible and have potent but distinct inhibitory roles. SOCS1 controls macrophage responses to IFN-γ and signaling through TLR4 and TLR9, and mice genetically deficient in SOCS1 develop severe systemic autoimmune and inflammatory disease (10, 11, 12, 13). SOCS3 attenuates IL-6-induced STAT3 anti-inflammatory effects (14, 15) as well as IL-4-induced gene expression that is mediated through insulin receptor substrate-2 (16, 17). SOCS1 and SOCS3 interrupt cytokine receptor signaling at multiple levels, including preventing STAT binding to cytokine receptors, targeting the signaling complex for ubiquitination and proteasomal degradation, and inhibition of JAK tyrosine kinase activity (11, 18). In addition, they are thought to direct effects on MAPK, PI3K, and NFκB pathways (13, 19, 20). This demonstrates the influence of SOCS1 and SOCS3 on macrophage function and emphasizes their potential role in macrophage activation in disease.

The present study was designed to examine macrophage SOCS protein expression in immune-mediated inflammation in vivo. Using nephrotoxic nephritis (NTN), a robust model of acute Th1-mediated injury, we show that the majority of infiltrating glomerular macrophages are polarized to express a single SOCS protein, predominantly SOCS3. We therefore analyzed the role of SOCS3 in M1 activation and established that its expression in macrophages is a critical determinant of injury in this model. Manipulation of macrophage SOCS3 expression could therefore provide a powerful mechanism to down-regulate inflammation.

Rat bone marrow-derived macrophages (BMDM) were isolated by aspiration of the femur and tibia, as previously described (5), and suspended in culture medium comprising DMEM supplemented with 10% heat inactivated FCS, 100 U/ml penicillin, and 100 μg/ml streptomycin with the addition of 20% L929 conditioned medium produced using a standard protocol (21); this yields a population >95% macrophages. BMDM were stimulated with the following rat cytokines as indicated in the text; IL-4 10 ng/ml, IL-10 20 ng/ml (PeproTech), IFN-γ 2.5 ng/ml, and LPS 100 ng/ml (Sigma-Aldrich). Nephrotoxic serum was produced according to our standard protocol (22).

Normal or SOCS3 knock down BMDM were plated at 1 × 10⁶ cells/well in a 24-well plate and left overnight before stimulating with cytokines as indicated in the text. NO in the culture medium was assessed by nitrite production using a Greiss reaction (5). The concentration of cytokines in the supernatant was determined by capture ELISA using kits specific for TNF-α (BD Biosciences), IL-10, and IL-6 (BioSource International). STAT1 and STAT3 activation was measured with a TransAM ELISA-based kit (Active Motif) performed on nuclear extracts of BMDM according to the manufacturer’s protocol.

Inbred male Sprague–Dawley rats (200–250g) (Harlan) were preimmunized with s.c. injection of 1 mg of rabbit IgG (Sigma-Aldrich) in CFA and injected i.v. 1 wk later with rabbit nephrotoxic serum using our standard protocol (23). To establish high concentrations of IL-4 and IL-10 locally within the inflamed glomerulus, macrophages overexpressing these cytokines were injected directly into the renal artery (23, 24) and severity of injury assessed as previously described (24).

CD68⁺ macrophages were detected in methyl Carnoy’s-fixed tissue by ED1 Ab (AbD Serotec) and alkaline phosphatase-labeled rabbit anti-mouse and mouse alkaline phosphatase anti-alkaline phosphatase and developed using Fast red as a substrate. Double immunohistochemistry for detection of SOCS1 and SOCS3 positive glomerular macrophages was performed in archived methyl Carnoy’s-fixed samples of renal tissue. This material was collected in tissue arrays each containing 150 cores of kidney from groups of rats killed at different time points throughout the course of disease with and without IL-4 or IL-10 treatment. Sections were deparaffinized and dehydrated by a series of xylene and alcohol washes. Following quenching of endogenous peroxidase activity with 3% (vol/vol) H₂O₂ in methanol, tissues were blocked with 1% (wt/vol) horse serum and incubated with goat anti-rat SOCS1 (AbD Serotec; 1/200) or rabbit anti-rat SOCS3 (Abcam; 1/500) followed by peroxidase-conjugated secondary Ab (1/300); peroxidase-stained sections were developed with 3,3′-diaminobenzidine (DAB) and counterstained with hematoxylin (Sigma-Aldrich). Double staining with anti CD68 (1/600) was performed after microwaving SOCS stained sections in citrate buffer. Ab binding was detected by secondary anti-mouse followed by streptavidin-conjugated alkaline phosphatase and developed with liquid permanent red (DakoCytomation) that was detected by immunofluorescence. The number and proportion of SOCS1 and SOCS3 positive macrophages per glomeruli was scored in each case from superimposed images. For triple immunohistochemistry, dewaxed sections were stained for SOCS1 and SOCS3 (as above) and detected with the secondary Abs anti-rabbit tetramethylrhodamine isothiocyanate and anti-goat FITC, respectively, before staining for ED1 with peroxidase-conjugated secondary Ab and DAB substrate.

For SOCS1 and SOCS3 detection on cell preparations, BMDM were grown on coverslips, stimulated with cytokines as indicated in the text, and fixed in 2% paraformaldehyde. After blocking in 1% glycine/PBS, they were incubated with Abs to SOCS1 and SOCS3 (as above) and detected with Alexa Fluor 488-conjugated chicken anti-goat and Alexa Fluor 594-conjugated donkey anti-rabbit secondary Abs, respectively. The nonspecific background was determined by use of an isotype control Ab and the secondary Ab alone. Images were captured with a Leica TCS-4D confocal microscope system with a Kr-Ar laser and a ×40, 1.0-numerical-aperture objective.

BMDM were stimulated with or without cytokine for indicated time periods and total RNA prepared using the Trizol extraction reagent (Invitrogen) followed by RNA clean up using an RNeasy mini kit (Qiagen) according to the manufacturer’s instructions. A total of 5 μg from each sample was reverse transcribed using the first-strand cDNA synthesis kit and oligo (dT) primer as recommended by the manufacturer (Invitrogen) in a 15-μl reaction volume. For each gene, the PCR was performed on 50, 100, and 200 ng of the cDNA using the following primers:

SOCS1 forward 5′-ATGGTAGCACGTAACCAG-3′

SOCS1 reverse 5′-CTCCAGCAGCTCGAAGAG-3′;

SOCS3 forward 5′-ACCAGCGCCACTTCTTCACA-3′

SOCS3 reverse 5′-GTGGAGCATCATACTGGTCC-3′;

Rat iNOS forward 5′-TCTTGGTGAAAGCGGTGTT-3′

Rat iNOS reverse 5′-TGTTGCGTTGGAAGTGTAGC-3′;

Arginase forward 5′-TATCTGCCAAGGACATCGTG-3′

Arginase reverse 5′-GTTCTGTTCGGTTTGCTGT-3′;

IL-10 forward 5′-CTCCACTGCCTTGCTTTTA-3′

IL-10 reverse 5′-CACTGCTATGTTGCCTGCTC-3′;

Mannose receptor forward 5′-AACAAGAATGGTGGGCAGTC-3′

Mannose receptor reverse 5′-AACTCCTCGTCCGTCTGTC-3′; and

GAPDH forward 5′-TGACATCAAGAAGGTGGTGAAG-3′

GAPDH reverse 5′-TCTTACTCCTTGGAGGCCATGT-3′.

The cycling conditions were 28 cycles of 95°C for 40 s, 58°C for 30 s, and 72°C for 40 s, followed by a single 10-min extension at 72°C. To control for genomic DNA contamination, equal amounts of cDNA from each sample were PCR amplified without reverse transcription. RT-PCR products were separated on 1% (w/v) agarose gel and visualized by staining with ethidium bromide. The level of mRNA expression in each sample was determined by densitometric image analysis and standardized against the GAPDH measurement.

Three predesigned SOCS3 siRNA duplexes ([1] ID #192801, [2] ID #192802, and [3] ID #192803) and a functional, nontargeting siRNA sequence negative control containing at least four mismatches to any mouse human or rat gene were purchased from Ambion. BMDM were transfected with 75 pmol siRNA using 5 μl Lipofectamine RNAi Max (Invitrogen) in serum free medium according to the manufacturers’ instructions. Mock transfection using Lipofectamine RNAi Max but no siRNA and no transfection controls were included in each experiment. Cells were incubated at 37°C for 48 h before treatment. SOCS3 siRNA transfection efficiency was 89 ± 9% as determined by counts for the number of macrophages containing fluorescently tagged siRNA. Cell viability was not altered by transfection reagent alone (mock transfection 98.9 ± 1.2%) or transfection with siRNA (99.0 ± 0.7%).

Protein lysates were prepared from BMDM and 10 μg separated by SDS/PAGE for Western blot analysis with specific primary Abs. Immunolabelled proteins were detected by using appropriate HRP-conjugated secondary Abs, followed by visualization with ECL (Amersham Pharmacia Biotech). For flow cytometric analysis, cells were stained with FITC-conjugated anti-CD86 (BD Pharmingen) or FITC-conjugated isotype control.

Results are presented as mean ± SD and differences between groups of cells or animals were determined by Student’s t test or two way ANOVA followed by a multiple range test (Tukey analysis). Values of p < 0.05 were considered significant.

NTN is a robust model of macrophage-dependent injury in which there are reliable methods for quantifying infiltrating leukocytes and characterizing their phenotype (23, 24). It was therefore ideal for analyzing SOCS protein expression in infiltrating macrophages. Neither SOCS1 nor SOCS3 protein could be visualized by immunohistology in normal rat glomeruli, but both proteins were readily detectable after induction of nephritis. As reported previously (25), staining for SOCS3 was more prominent than SOCS1, but both proteins were expressed by intrinsic glomerular cells as well as by infiltrating leukocytes. In normal rats, double staining with Abs to CD68 and either SOCS1 or SOCS3 identified a small number of resident glomerular macrophages that did not express SOCS proteins. In contrast, 24 h after induction of nephritis, there was a 25-fold increase in glomerular macrophages of which a third (36 ± 11%) expressed SOCS1 and two thirds (63 ± 14%) expressed SOCS3 (Fig 1). These proportions remained unchanged over the next 48 h, but by day 7, the proportion of SOCS1 expressing macrophages had increased, although proportions of SOCS3 expressing macrophages still remained predominant.

It was notable that the combined number of glomerular macrophages that stained positive with Abs to SOCS1 or SOCS3 approximated the total number of infiltrating macrophages, suggesting that individual macrophages express either but not both SOCS proteins. We therefore repeated the analysis using triple staining to determine the proportions of macrophages that exclusively expressed either of the SOCS proteins or both of them (Fig. 2). On days 1 and 3 after induction of nephritis, 27 ± 6% of infiltrating glomerular macrophages expressed only SOCS1 and 54 ± 12% exclusively expressed SOCS3; 10% were positive for both SOCS proteins, and 10% expressed neither at day 1 (Table I). The proportion of SOCS1 and SOCS3 double positive macrophages was greater on day 7 when the inflammation is subsiding.

Thus, over 80% of infiltrating macrophages in the acute stage of injury expressed a single SOCS protein, and the infiltrate was heterogeneous with twice as many of the cells expressing only SOCS3 as SOCS1 and only a tenth expressing both proteins. This was surprising because IFN-γ is a prominent activating cytokine in glomeruli in acute nephritis, and in vitro studies have documented that macrophages activated by IFN-γ have increased expression of both SOCS1 and SOCS3 (16, 26). We therefore re-examined macrophage SOCS expression in response to different types of activation in vitro.

We incubated rat BMDM with IFN-γ (0.5–20 ng/ml) and LPS (100 ng/ml) either singly or in combination. As previously reported (16, 26), macrophages incubated with IFN-γ alone at concentrations above 2 ng/ml increased both SOCS1 and SOCS3 mRNA expression (Fig. 3, A and B) as did incubation with 100 ng/ml LPS. However, incubating macrophages with LPS together with IFN-γ had a marked polarizing effect: SOCS1 expression was suppressed, whereas SOCS3 expression increased. Induction of SOCS3 by IFN-γ/LPS was rapid, peaking at 30 min and declining after 2 h to levels that were, however, still above baseline (Fig. 3 C); after this time, there was a corresponding increase in SOCS1 mRNA.

Incubation of BMDM with the M2a activating cytokine, IL-4, increased SOCS1 mRNA but not that of SOCS3 (Fig. 3, A and B). Maximum induction of SOCS1 by IL-4 was slower than that of IFN-γ/LPS induced SOCS3, peaking at 2 h (Fig. 3 C). No induction of SOCS3 by IL-4 was observed over the time course. IL-10 also induces anti-inflammatory effects but these are distinct and IL-10 activated macrophages are designated M2c activated (or deactivated), whereas IL-4 activated macrophages are designated M2a (2). Unlike IL-4, IL-10 did not induce SOCS1 mRNA at any time point but induced a small increase in SOCS3 mRNA.

To ensure that the changes in mRNA were reflected in the amounts of protein expressed, we quantified SOCS protein expression by immunofluorescence. Unlike resident glomerular macrophages, low levels of SOCS1 and SOCS3 proteins were detected in resting BMDM, presumably because the assay was more sensitive than immunohistology of the glomeruli. As reported by others (16, 26), stimulation with either IFN-γ or LPS alone increased staining for both SOCS (Fig. 3, D and E). The results were very different when macrophages were incubated with IFN-γ and LPS together. Again, SOCS3 staining increased to levels comparable to that seen with LPS alone, whereas SOCS1 staining was consistently suppressed to levels below that of resting macrophages (Fig. 3, D and E). By 24 h after stimulation with IFN-γ/LPS, the SOCS1 protein concentration had returned to baseline but no higher, despite the increase in concentration of mRNA that was detected 20 h earlier; these results were confirmed by Western blotting (data not shown). The IL-4-induced increase in SOCS1 protein expression without change in SOCS3 protein faithfully reflected changes in mRNA (Fig. 3, D and E). This was observed not only during the initial up-regulation of SOCS1 at 3 h but also when mRNA levels were declining at 24 h; this contrasts to the sustained SOCS1 protein expression seen after stimulation with IFN-γ/LPS. By contrast, the combination of IL-4 and LPS was not polarizing but had simple additive effects on SOCS1 and SOCS3 staining (data not shown). Incubation with IL-10 did not alter expression of either SOCS1 or SOCS3 protein (data not shown). Thus, M1 and M2 activation resulted in SOCS3 and SOCS1 polarized macrophages, analogous to the polarized populations observed in vivo.

The strongly polarized SOCS protein responses suggested they have a determining role in macrophage properties after activation. Accordingly, we addressed the functional consequences of the rapid and selective up-regulation of SOCS3 in M1 by transducing BMDM with three different predesigned SOCS3 siRNA sequences before incubation with IFN-γ/LPS. All three individual sequence SOCS3 siRNA but not control siRNA or mock-transfected attenuated the endogenous SOCS3 mRNA response (Fig. 4,A) and reduced SOCS3 protein expression (Fig. 4 B).

The functional effectiveness of the SOCS3 knock down was assessed by quantifying STAT3 phosphorylation that is inhibited by SOCS3 (27). Using a specific ELISA, we analyzed STAT3 phosphorylation in normal and knock down BMDM before and after activation with IFN-γ and LPS. Levels of phosphorylated STAT3 in the nucleus were the same in unstimulated BMDM regardless of whether SOCS3 had been knocked down. However, concentrations of phosphorylated STAT3 were significantly greater in IFN-γ/LPS-activated SOCS3 knock down BMDM than in control siRNA-treated cells. The specificity of our siRNA protocol was demonstrated by the absence of any effect on STAT1 phosphorylation (Fig. 4 C). This confirms that our knock down protocol is sufficient to attenuate the inhibitory effects of SOCS3 on STAT3 phosphorylation and so enabled us to examine the role of SOCS3 in M1 by IFN-γ/LPS.

Treatment with IFN-γ/LPS-induced distinctly different gene expression patterns in BMDM transduced with SOCS3 siRNA compared with mock or control siRNA transduced cells. The SOCS3 knock down BMDM expressed significantly more mRNA for the M2a markers, macrophage mannose receptor and arginase (Fig. 4,D) and also for SOCS1, whose expression increased to a similar degree to that which we observed after activation with IL-4. Interestingly, although there was no change in iNOS mRNA expression, the normal IFN-γ/LPS-induced increase in NO was markedly attenuated in the SOCS3 knock down cells, as reflected by their nitrite production (Fig. 4,E). Up-regulation of the costimulatory molecule CD86 and IL-6 secretion were similarly reduced (p < 0.05) in the SOCS3 knock down macrophages, whereas there was no significant change in TNF-α secretion (Fig. 4, E and F). Notably, the IFN-γ/LPS-induced increase of the anti-inflammatory cytokine IL-10 was maintained.

We next examined the role of SOCS3 in determining the unresponsiveness of M1 to M2 activating stimuli (5). As previously shown, prior exposure to IL-4 attenuated IFN-γ/LPS-induced NO generation by 52 ± 6%, whereas treatment with IL-4 after activation with IFN-γ/LPS was relatively ineffective (11 ± 4% reduction) (Fig. 5). IL-4 responsiveness was completely restored in the SOCS3 knock down cells, demonstrating that it was another property of M1 that was dependent on SOCS3. Indeed, these in vitro studies show that activating macrophages with IFN-γ/LPS in the absence of SOCS3 induces similar properties to those of macrophages activated by IL-4.

The next experiments addressed whether IL-4 could exert a similar effect in vivo and if so with what functional consequences. We have previously shown that a high local concentration of IL-4 in the glomerulus, induced by IL-4 over-expressing macrophages, significantly reduced the severity of renal injury as assessed by glomerular macrophage number and albuminuria (24). Kidney sections from these experiments were re-analyzed to determine the total number glomerular macrophages and the number expressing SOCS1 and SOCS3. As reported previously, IL-4 significantly reduced the total number of infiltrating glomerular macrophages to 60% of those found in untreated rats on day 1 after induction of NTN and 35% on day 7 (Fig. 6 A). Despite this, the number of infiltrating macrophages that was uniquely SOCS1 positive, SOCS1/SOCS3 double positive, or SOCS negative were not changed by the IL-4 treatment. By contrast, there was an 88% reduction of glomerular macrophages that uniquely expressed SOCS3 day 1 (p < 0.001) and a 51% reduction on day 7 (p < 0.01). Thus, the reduction in injury in the IL-4-treated rats correlated exclusively with a decrease in SOCS3 single positive macrophages in the glomerular infiltrate, suggesting that these cells are critical for injury.

The importance of SOCS3 single positive macrophages for injury is supported by the results of experiments to increase local concentrations of IL-10 – a treatment that also attenuated injury in NTN (23). This treatment also decreases infiltrating macrophages (Fig. 6 B) and induced a comparable reduction of SOCS3 single positive cells in the infiltrate (82% and 83% reductions on days 1 and 7, respectively). However, the effects of IL-10 were less specific and infiltrating SOCS1 single positive macrophages were also decreased (by 59% and 44% on days 1 and 7, respectively) together with a concomitant increase in SOCS negative macrophages. This is consistent with IL-10’s reported role in macrophage deactivation (2). Taken together, these results demonstrate that expression of SOCS1 and SOCS3 is determined by the activating microenvironment providing a key directional influence for macrophage inflammatory potential in vivo as well as in vitro.

In this study, we demonstrate for the first time that macrophages infiltrating inflamed glomeruli in an acute model of immune-mediated injury are both heterogeneous and polarized regarding SOCS1 and SOCS3 protein expression. We establish unique patterns of SOCS expression in differentially activated macrophages in vitro and show that SOCS3 expression is selectively associated with proinflammatory cells and is essential for maintaining their properties. Thus, in the absence of SOCS3, M1 have an impaired ability to develop proinflammatory properties but instead have enhanced anti-inflammatory characteristics. Furthermore, loss of SOCS3 expressing macrophages in inflamed glomeruli is associated with a reduction in inflammation confirming the functional importance of these cells in vivo.

The most notable feature of our in vivo results is the degree to which macrophages infiltrating this Th1-mediated model of immune injury are polarized with regard to SOCS expression. Nearly 90% of the SOCS-expressing macrophages exclusively express a single SOCS protein and even 7 days after induction of disease only 12% express both SOCS proteins. This is surprising given the large body of evidence that IFN-γ induces both SOCS1 and SOCS3 (16, 26) – a result confirmed by our own experiments. This apparent paradox was resolved by the polarizing effect of LPS and our demonstration that activation of macrophages with both IFN-γ and LPS suppresses expression of SOCS1 and induces the development of macrophages that uniquely expressed SOCS3, similar to those identified in vivo. Interestingly, the suppression of SOCS1 protein is sustained even though mRNA concentrations subsequently increased. This is in contrast to IL-4 where changes in protein simply reflected mRNA with no dissociation. IL-10 did not increase SOCS3 protein despite the increase in SOCS3 mRNA. These results emphasize the distinct differences in M2 as reflected by their polarized SOCS expression. IL-4 M2a have enhanced SOCS1 and suppressed SOCS3, whereas SOCS expression is not altered by IL-10 M2c activation and these differences will have a marked effect on macrophage properties.

The preferential polarization of SOCS1 and SOCS3 in M2a and M1 is analogous to the polarization of SOCS5 and SOCS3 in Th1 and Th2 cells, respectively (28, 29). Although mediated by a different mechanism, this strongly suggests that macrophages, like T cells, show antagonistic regulation of different SOCS proteins. This is confirmed by our SOCS3 knock down studies where a decrease in SOCS3 expression results in enhanced expression of SOCS1. IL-4-induced SOCS1 expression is known to be a potent physiological inhibitor of both IFN-γ and LPS signaling in macrophages and therefore can antagonize IFN-γ/LPS-induced expression of SOCS3 (10). Moreover, the SOCS1 promoter has STAT3 binding sites and IFN-γ/LPS-induced SOCS3 can inhibit STAT3 activity thereby preventing SOCS1 expression (30).

The most striking observation is the functional importance of SOCS proteins in regulating macrophage properties. This is clearly highlighted by our knock down studies where blocking SOCS3 prevents classical activation. The SOCS3 knock down macrophages have enhanced STAT3 activity that is associated with an increase in the M2a activation associated molecules mannose receptor, arginase and indeed SOCS1; decreased expression of the proinflammatory mediators IL-6 and NO and the costimulatory molecule CD86; and restored responsiveness of these cells to the anti-inflammatory effects of IL-4. STAT3 antagonizes proinflammatory signals to limit excessive or prolonged macrophage activation. Thus, SOCS3 inhibits STAT3 activity preserving the natural cytotoxicity of M1 and here we demonstrate that without it, macrophages develop characteristics of M2a activated cells.

Our findings with macrophages have parallels with those using dendritic cells. Dendritic cells genetically deficient in SOCS3 have hyperactivation of STAT3 and a potent tolerogenic phenotype that can suppress the severity of experimental autoimmume encephalomyelitis (31). The importance of SOCS3 for effective innate immunity is emphasized by the observation that both microorganisms and tumors have evolved mechanisms to block SOCS3 expression or activity resulting in enhanced STAT3 activity thereby subverting macrophage cytotoxic functions and promoting their own survival (32, 33). M1 are unresponsive to the anti-inflammatory effects of IL-4 (5) and here we show that the unresponsiveness is critically dependent on SOCS3. Consistent with this, O’Connor et al. (17) demonstrated the importance of SOCS3 in impairing IL-4-induced gene expression in macrophages by its effects on PI3K. Therefore, SOCS3 not only maintains the proinflammatory properties of M1 activated cells but prevents them from becoming anti-inflammatory. The role of SOCS3 as a negative regulator of M2a activation is entirely consistent with previously published data demonstrating that SHIP can prevent M2 macrophage development by inhibiting PI3K (34). Given that SOCS3 prevents IL-4-induced PI3K activity in macrophages (17), it will be important to determine whether SOCS3 and SHIP exert their effects on macrophage activity by similar mechanisms.

A key role for SOCS3 in dictating the proinflammatory properties of macrophages was confirmed in vivo. Mice lacking the SOCS3 gene in macrophages are resistant to acute inflammation as modeled by LPS-induced endotoxin shock (14). In this study, we show that local delivery of IL-4 to inflamed glomeruli has a major effect on reducing the number of SOCS3 expressing glomerular macrophages, and this is reflected by a decrease in the severity of nephritis. IL-4 down-regulates inflammation by multiple mechanisms (1), including a decrease in infiltrating macrophages. However, the specific decrease in SOCS3 expressing macrophages with IL-4 treatment, together with the identification that SOCS3 expressing macrophages are proinflammatory in vitro, strongly supports their essential role in mediating macrophage-dependent injury in NTN. Conversely, reports from models of arthritis show macrophages lacking SOCS3 exhibit enhanced activation (35), although this was expected given that progression of chronic arthritis in this model is mediated by enhanced IL-6/STAT3 therefore SOCS3 would have a beneficial role.

The ability to redirect macrophage activity would provide a novel and potentially highly effective therapeutic approach for treating inflammatory and immune-mediated disease. Our demonstration of a strong association of SOCS3 with proinflammatory macrophages in vitro and in vivo and the expression of characteristic M2a markers on exposure to M1 activation after SOCS3 knockdown suggests the potential of SOCS3 manipulation for this. However, any therapeutic strategy involving SOCS3 manipulation for inflammatory disease would have to be targeted specifically to macrophages as SOCS3 is widely expressed. Indeed loss of SOCS3 in hematopoietic lineages in SOCS3 Vav-Cre mice leads to profound neutrophilia and inflammatory neutrophil infiltrations (36).

In summary, we propose that many of the properties of M1 are critically dependent on the polarizing effect of the combination of IFN-γ with LPS on SOCS3 protein expression. The result is that these macrophages exclusively express SOCS3 that has a major effect on determining the well documented proinflammatory properties of these cells. The marked polarization of macrophages with respect to SOCS1 and SOCS3 is observed in vivo in NTN, a model of injury dominated by a Th1 response suggesting that the analysis of macrophage SOCS protein expression could be a useful biomarker for assessing disease activity.

We acknowledge Anton Jäger for help with illustrations.

The authors have no financial conflict of interest.