STAT3 determines IL-4 signalling outcomes in naïve T cells

IL-4 production is associated with low-avidity, poorly cytotoxic T cell induction that contributes to viral immune evasion and the failure of T cell-based vaccines. Yet, the precise mechanisms that regulate IL-4 signalling in T cells remain elusive. Mounting evidence indicates that cells can dynamically alter their IL-4/IL-13 receptor signature to modulate downstream immune outcomes upon pathogen encounter. Here, we describe how naïve (CD62L+CD44lo–mid) CD4 and CD8 T cells distinctly engage both STAT6 and STAT3 in response to IL-4. We further show that IL-4R⍺ expression is both time- and IL-4 concentration-dependent. Remarkably, our findings reveal that STAT3 inhibition can ablate IL-4R⍺ and affect transcriptional expression of other Stat and Jak family members. By extension, the loss of STAT3 lead to aberrant STAT6 phosphorylation, revealing an inter-regulatory relationship between the two transcription factors. Moreover, IL-4 stimulation down-regulated TGF-β1 and IFN-γR1 expression on naïve T cells, possibly signifying the broad regulatory implications of IL-4 in conditioning lineage commitment decisions during early infection. Surprisingly, naïve T cells were unresponsive to IL-13 stimulation, unlike dendritic cells. Collectively, these findings could be exploited to inform more efficacious vaccines, as well as design treatments against IL-4/IL-13-associated disease conditions.

Since the discovery of interleukin (IL)-4, its multifunctional implications in mediating adaptive immunity have been well-studied 1,2 . In particular, IL-4 promotes humoral responses whilst dampening cell-mediated activity, including via type 2 helper T cell lineage commitment and maintenance 1,3,4 . Consequently, its production and signalling has been implicated in many different pathologies, including infection, cancer, autoimmunity, immunodeficiency and allergy [5][6][7] .
Studies have also shown the importance of IL-4 in determining vaccination outcomes. Kienzle et al. were among the first to demonstrate that IL-4 availability promotes a noncytolytic CD8 low immune phenotype 8 . Interestingly, some viruses (such as HIV) exploit IL-4 production to evade cytotoxic T cell-mediated elimination 9 . Indeed, endogenous IL-4 production serves as an attractive explanation as to why cytotoxic T cell-based approaches to HIV-1 vaccination design have thus far yielded limited protective efficacy 10,11 . Following decades of disappointing outcomes with many HIV-1 vaccine candidates, including the recent HVTN 701 trial 12 , as well as the ongoing SARS-CoV2 pandemic, nuanced approaches to vaccination design is of great importance. Studies in our laboratory have shown that a pox viral vector-based prime-boost vaccine strategy that co-expresses HIV antigens together with an IL-4 receptor antagonist can drastically improve avidity, poly-functionality and cytotoxicity of mucosal and systemic HIV-specific CD4 and CD8 T cells, associated with protection in both mice and macaques 10,[13][14][15] . While this is a hopeful step towards establishing protective immunity against HIV-1, how different T cells govern and regulate their response to different IL-4 conditions at the molecular level still remains unclear.
IL-4 is known to signal via two distinct receptor complexes, signalling predominantly via the JAK1/STAT6 pathway. The high-affinity Type I IL-4 receptor complex is an IL-4R⍺:γc heterodimer that docks JAK1 and JAK3. In contrast, the low-affinity Type II IL-4 receptor is an IL-4R⍺:IL-13R⍺1 complex, capable of additionally binding IL-13. The Type II complex is believed to further dock JAK2 and engage STAT3 under certain conditions, although the function and context under which this alternative pathway is recruited is unknown [16][17][18][19][20] . Finally, IL-13R⍺2, the high-affinity cognate receptor for IL-13, interacts with the IL-4R⍺ cytoplasmic domain in a mutually antagonistic relationship that ultimately restricts IL-4 signalling 21 .
Studies in our laboratory and elsewhere have shown that differential IL-4/IL-13 receptor representation can have drastic implications on immune outcomes 20,[22][23][24][25] . Overexpression of IL-4R⍺ and engagement of STAT6 in M2 macrophage have been linked to liver fibrosis and heightened hepatic inflammation 26 . Similarly,

Results
Time-and concentration-dependent IL-4R⍺ regulation observed in both splenic CD4 and CD8 T cells following IL-4 stimulation. In this study, whole splenocytes were stimulated for 24 h with varying concentrations of IL-4 to mimic the cytokine environment during infection/vaccination 14,25,35 . IL-4R⍺ expression profiles on both CD4 and CD8 T cells were evaluated by flow cytometry (Fig. 1a). In both T cell subsets, the proportion of cells expressing IL-4R⍺ significantly increased with 1, 10 and 50 ng/mL IL-4, unlike the lower concentrations (Fig. 1b,c). These trends were also reflected in the fold-change mean fluorescent intensity values (MFI), normalised to the respective non-stimulated control (Sup. Fig. 1a-c). Interestingly and consistent with previous findings, IL-4-induced IL-4R⍺ up-regulation was not universal across immune cell types 25 , as typified by splenic MCH-II + CD11c + antigen presenting cells (Sup. Fig. 1d,e). Furthermore, IL-13 stimulation did not impact T cell IL-4R⍺ expression, despite the shared receptor system (Sup. Fig. 1f-h).
Data also revealed that the IL-4R⍺ expression was differentially regulated over time following IL-4 stimulation (Fig. 1d,e), as the splenic CD4 and CD8 T cells dramatically down-regulated receptor expression at 1 h, which was subsequently up-regulated at ~ 12 h, and maintained for the remainder of the timecourse. Similarly, significant up-regulation of the Il4ra transcript was observed in sorted CD3 + CD4 + CD8 -IL-4R⍺ + splenic T cells post IL-4 stimulation, including at 1 h (p < 0.01 for all) (Fig. 1f). In contrast, expression of γc and IL-13R⍺1 was mainly down-regulated upon IL-4 encounter (Sup. Fig. 2). Moreover, intracellular and extracellular IL-4R⍺ expression profiles revealed that down-regulation of IL-4R⍺ at 1 h was associated with receptor internalisation (Fig. 1g,h), as no change in the total IL-4R⍺ expression was observed on both T cell subsets (p = 0.9815 [CD4]; p = 0.4295 [CD8]). This may explain the observed downward trend in γc and IL-13R⍺1 expression following cytokine stimulation; that is, whole receptor complexes were being internalised.
To further understand the implications of T cell activation in IL-4R⍺ regulation on T cells, splenocytes were exposed to ⍺CD3ε and ⍺CD28 (Fig. 2d). Interestingly, while ⍺CD3ε and ⍺CD28 failed to elicit changes in IL-4R⍺ expression compared to non-stimulated cells, in combination with IL-4, the receptor was modestly down-regulated (p < 0.0001) (Fig. 2e,f). Interestingly, when the above experiment was performed using sorted naïve CD4 T cells, IL-4R⍺ was found to be upregulated independent of concurrent TCR stimulation (Sup. Fig. 4). This indicated that secondary cytokines (e.g. TNF-⍺ and/or IFN-γ) expressed by other functional cell types, such as effector/memory T cells, could also be associated with IL-4R⍺ regulation on naïve T cells.

IL-4 promotes not only STAT6 but also STAT3 activation in naïve T cells. Our recent studies have
shown that lungs DCs respond to different IL-13 conditions by modulating IL-13 receptors, STAT3/STAT6 phosphorylation and TGF-β1 expression 20 . Knowing IL-4 and IL-13 share a common receptor system, we next evaluated the phosphorylation status of both STAT6 and STAT3 on splenic CD4 and CD8 T cells following 15 min (previously shown to cause an optimum fold-change in pSTAT6) 38 and 24 h IL-4 stimulation (Fig. 3a). At both time points, highly significant increases in STAT6 phosphorylation (pSTAT6) (Fig. 3b,c), as well as pSTAT3, was detected in both T cell subsets (p < 0.0001 for all) (Fig. 3d-f). ImageStream analysis was performed to assess the subcellular localisation of IL-4R⍺, pSTAT3 and pSTAT6 in splenic T cells, comparing IL-4 stimulation conditions (Fig. 3g). The stimulated group significantly increased the expression of these makers, as shown by the elevated fluorescence intensity values between the two groups (p < 0.0001; n < 1200 cells) (Fig. 3h-j). Interestingly, IL-4 stimulated cells showed a higher degree of pSTAT6 nuclear localisation when compared to the 7-AAD nuclear dye control (Fig. 3g). However, surprisingly, pSTAT3 appeared to be less densely trafficked to the nucleus compared to pSTAT6 (Fig. 3g) www.nature.com/scientificreports/ www.nature.com/scientificreports/ may be linked to an active regulation of the two STAT molecules according to the IL-4 environment, and thus warrant further investigation. Our findings also revealed that the IL-4-induced phosphorylation of STAT3 and STAT6 were also linked to the T cell effector status (CD62L expression) (Fig. 4), similar to what was observed with IL-4R⍺ regulation (Fig. 2). Specifically, change in pSTAT6 MFI was ~ tenfold greater in CD62L + CD4 and CD8 T cells compared to the CD62Lpopulation ( Fig. 4a-c). Similarly, significant enhancement of pSTAT3 MFI was only detected in the CD62L + but not the CD62Lsubsets ( Fig. 4d-f). It is also noteworthy that the upregulation of pSTAT6 MFI was found to be IL-4 concentration dependent, with phosphorylation detected from 1 to 50 ng/mL (Sup. Fig. 5a,b). However, regulation of pSTAT3 MFI required much greater IL-4 levels, first being detected at 10 ng/ mL and maximally observed at 50 ng/mL (Sup. Fig. 5c-d). Thus, for optimal STAT3 phosphorylation, higher IL-4 concentration was used throughout the study. It is noteworthy that t-tests reveal no significant difference in the IL-4Rα MFI in cells with 10 or 50 ng/mL concentrations (CD4: p = 0.153, CD8: p = 0.958) (Sup. Fig. 1b-c). Interestingly, T cells stimulated with IL-13 (10 ng/mL) did not show any STAT6 or STAT3 phosphorylation (Sup. Fig. 5). Moreover, when splenic T cells were stimulated with 50 ng/mL IL-4, CD62L + CD44 lo-mid CD4 T cells down-regulated TGF-β1 expression, compared to the unstimulated control (p = 0.0018) (Sup. Fig. 6). Interestingly, the CD62L -CD4 T cell subsets did not down-regulate TGF-β1, which is consistent with the STAT3 and STAT6 phosphorylation profiles observed with CD62L + and CD62L -T cell subsets (Sup. Fig. 6; Sup. Fig. 5).

STAT3 inhibition down-regulates IL-4R⍺ and STAT6 phosphorylation. Knowing STAT6 or STAT3
were phosphorylated following IL-4 stimulation, their impact on IL-4R⍺ regulation was evaluated using smallmolecule inhibitors as described previously (Fig. 5a) 20 . STAT6 inhibition partially down-regulated (halved) the IL-4R⍺ expression on IL-4-stimulated CD4 and CD8 T cells, compared to the uninhibited control (p < 0.0001) (Fig. 5b,c). However, and remarkably, drastic down-regulation of IL-4R⍺ was observed on both subsets following STAT3 inhibition (p < 0.0001); combined STAT6/STAT3 inhibition showed a similar response to STAT3 only inhibition (Fig. 5b,c). Next, phosphorylation analyses following IL-4 stimulation and STAT3 inhibition revealed www.nature.com/scientificreports/ proportion of (e) CD4 and (g) CD8 T cells that were deemed pSTAT3 + . Data are represented as mean + SD, with dots representing biological replicates from a single experiment. The experiments were repeated three times. Two-way ANOVA combined with Tukey's post-hoc multiple comparison test was conducted to compare groups. (g) Whole spleen homogenate derived from naïve 6-8-week-old female BALB/c mice were stimulated with 50 ng/mL IL-4 in vitro for 24 h prior to ImageStream analysis. Representative images of pre-gated total T cells. Note that cells were fixed, allowing 7-AAD to be used as a nuclear stain. The fluorescent intensity of (h) IL-4Ra, (i) pSTAT6 and (j) pSTAT3 were compared between treatment groups. Data are represented as mean + SEM, where n > 1200 from a single experiment. The experiment was repeated three times. Student's t-tests were conducted to compare groups. p-value denotation: '****'p < 0.0001. www.nature.com/scientificreports/ differential IL-4-inducible STAT3 and STAT6 phosphorylation (p = 0.0001) (Fig. 5d,e; Sup. Fig. 6e,f). Specifically, STAT3 inhibition limited STAT6 phosphorylation by ~ 50% in both CD4 and CD8 T cell subsets (Fig. 5f).

STAT3 inhibition also modulates other biomarkers associated with IL-4/IL-13 signalling on
CD4 + IL-4R⍺ + T cells. Fluidigm 48.48 Biomark analysis was performed on single CD3 + CD4 + CD8 -IL-4R⍺ + T cells to compare the transcriptional profiles of IL-4/IL-13 signalling-associated markers following STAT3/ STAT6 inhibition and IL-4 stimulation, as described in the methods. Transcriptional expression of 42 genes were evaluated (Sup. table 1). Subsequent analyses were only performed on 16 genes that met the inclusion criterion: at least 15% of all cells expressing a given transcript (Sup. Fig. 7a). Dichotomised transcript expression analysis of the select genes showed unique expression profiles according to the treatment ( Fig. 6a; Sup. Fig. 7a-c). For example, the proportion of cells expressing Il4ra increased upon IL-4 stimulation, which was down-regulated following STAT6 inhibition and was completely ablated upon STAT3 inhibition (p = 0.0132, Fisher's exact test) (Fig. 6a). These findings clearly reflected the trends observed at the protein level (Fig. 5c). STAT3 inhibition also downregulated the expression of Stat3 and Stat6, although Stat5a remained unaffected (Fig. 6a,b). IL-4 stimulation was associated with reduced expression of Ifngr1 and Cd28, which were recovered following STAT6 inhibition. Interestingly, Pcna was predominantly detected on STAT3 inhibited cells, which was expected since STAT3 inhibits T cell proliferation 39 . 100 sorted CD3 + CD4 + CD8 -IL-4R⍺ + T cell RT-qPCR further validated the trends of select markers (Sup. Fig. 7d-k). Principle component analysis (PCA) was conducted on the 16 select genes, where expression was normalised to the L32 housekeeping gene control, as described previously 40 . The first four PCs accounted for over 83% of the variance, suggesting strong correlations between biomarkers. PC1 accounted for the majority of the variance (54%), with treatment playing a significant role (ANOVA: p = 0.00066). Evaluation of PC1 scores revealed www.nature.com/scientificreports/ that the STAT3i-treated cells clustered away from the other treatment groups (non-stimulated: p = 0.0016, IL-4 stimulated: p = 0.0199; and IL-4 stimulated, STAT6-inhibited: p = 0.0011) (Fig. 6c). When considering the loadings of these PCs (Sup. Fig. 8a-b), it further corroborated with the trends inferred by the dichotomised data. The transcriptional profile of the STAT6-inhibited cells was less severely affected compared to the STAT3-inhibited cells; this may be a consequence of the cells being IL-4R⍺ + , which are likely naïve T cells (Fig. 2) and thus do not have access to GATA3-mediated transcriptional activity. Spearman's rank correlation test was performed to evaluate similarity in the expression profiles between markers ( Fig. 6d; Sup. Fig. 8c). The major cluster of positively correlated genes consisted of Cd28, Jak1, Jak2, Ifngr1, Stat3, Il2rg, Smad2 and Stat6, as reflected by the rank correlation-derived dendrogram (Sup. Fig. 8d). Notably, the expression of these markers was typically downregulated upon STAT3 inhibition, according to the dichotomised outputs (Sup. Fig. 7a). Inversely correlated to this cluster were Stat5a and Pcna, both of which were highly expressed upon STAT3 inhibition.

Discussion
Mechanisms by which IL-4 is regulated in naïve, effector and memory T cell subsets still remain elusive. In the current study, we identify that antigen-naïve (CD62L + CD44 lo-mid ) CD4 and CD8 T cells can significantly upregulate IL-4R⍺ expression in an IL-4 concentration-and time-dependent manner. Moreover, STAT6 inhibition partially reduced the proportion of CD4 and CD8 T cells expressing IL-4R⍺, whilst STAT3 and/or combined  Fig. 7a,b for loadings). (d) For genes expressed by at least 15% of cells, Spearman's rank correlation analysis was performed to evaluate correspondence of genes pairwise. According to the similarity in transcriptional profiles, a major cluster of positively correlated genes were identified and highlighted, along with a negatively correlated cluster (refer Sup. Fig. 7a for dendrogram). The experiment was repeated twice, pooling the data. n ~ 15 cells per group.  20 . Notably, previous studies have also described IL-4-associated STAT3 phosphorylation in lymphocytes 43 , despite IL-4 signalling conventionally being linked to the JAK1/STAT6 pathway 4,42 . Jointly, these findings indicated that STAT3 could be the master regulator of IL-4 signalling for naïve T cells under high IL-4 conditions (Fig. 8). This is similar to that observed in lung DCs, where STAT3 maintains IL-13/IL-13R⍺2 homeostasis 20 .
Interestingly, responsiveness to IL-4 was mainly observed in naïve CD4 and CD8 T cells (CD62L + CD44 lo-mid ), compared to CD62L -(effector and memory) T cell subsets. This may seem intuitive, given the role of IL-4 in determining naïve T cell fate 44,45 . Despite this study exploring the effects of IL-4 in vitro, the observations correspond with vaccination studies performed by Wijesundara et al., where IL-4R⍺ was down-regulated on viral-specific T cells upon infection/vaccination 25 . Here, we further demonstrated that the CD62Lpopulation was unable to phosphorylate STAT3 upon low or high IL-4 stimulation and, consequently, failed to upregulate IL-4R⍺, compared to naïve T cells. Importantly, the IL-4 environment can also significantly impact the quality/ avidity of T cell immunity, where IL-4R⍺ expression was inversely related to IFN-γ and TNF expression on effector T cells following infection/vaccination 11,13,25 . Since we have shown that IL-4 signalling regulation occurs predominantly on naïve T cells, it may indicate that the early IL-4 environment during infection/vaccination may have significant implications in determining T cell fate.
In addition, naïve CD4 T cells also down-regulated the expression of both IFN-γR1 and TGF-β1 upon IL-4 stimulation. These findings may reflect how the cytokine environment upon pathogen encounter (specifically rapid IL-4 secretion by innate immune cells, including eosinophils, basophils, ILCs and macrophages 14,46-50 ) modulates downstream immune outcomes. Our findings may suggest that early IL-4 conditions naïve T cell www.nature.com/scientificreports/ fate. This may explain why elevated serum IL-4 can be detected following certain parasitic, helminthic and viral infections, including SARS-CoV-2 14,46-51 . Moreover, this study revealed that splenic and lung-derived T cells regulated IL-4R⍺ in a similar manner, even though the expression of IL-13R⍺2 was vastly different between the two compartments. Briefly, the lung T cells showed significantly greater baseline IL-13R⍺2 expression (similar to lung DC) that was uniquely downregulated upon IL-4 exposure. Interestingly, IL-13 did not impact IL-4/IL-13 receptor regulation in T cells. Knowing that IL-13R⍺2 can inhibit IL-4 signalling by interacting with the cytoplasmic domain of IL-4R⍺ to prevent JAK1 activity 33 , we propose that site-specific expression of IL-13R⍺2 may facilitate tissue-specific IL-4 signalling outcomes. This may explain how, despite sharing a common receptor system, IL-4 and IL-13 have both distinct and overlapping functions in different compartments (mucosal versus systemic). Specifically, this study supports the emerging dichotomy that IL-4 predominantly regulates adaptive cells, whilst IL-13 regulates innate cells at the first line of defence 20 .
Interestingly, our previous studies have demonstrated that 24 h post viral vector-based vaccination, ILC2derived IL-4/IL-13 uniquely regulated DC activity at the vaccination site, modulating downstream immune outcomes 20,30,36,37,52,53 . Mucosal vaccination can induce lower ILC2-derived IL-4/IL-13 at the vaccination site compared to systemic delivery, with mucosal delivery characterised by high avidity/poly-functional T cells 11,13 . Knowing that IL-4R-antagonising HIV vaccination improves the avidity, poly-functionality and cytotoxicity of CD4 and CD8 T cells in mice and non-human primates 10,14,15,35 , current findings indicate that adjuvants that transiently modulate the STAT3/STAT6/IL-4R⍺ axis may prove useful in fine-tunning pathogen-specific vaccine outcomes. Thus, these IL-4 regulatory mechanisms, particularly the role of STAT3, warrants further investigation in the context of infection/vaccination.
Notably, several alternative IL-4 signalling pathways have long been reported in some cell types [54][55][56][57][58][59] , however, under which circumstances these get activated in T cells is not well-defined. In the current study, while the genes encoding many of these alternative pathways were investigated (Irs1, Irs2, Stat5a and Stat1), Stat5a was the only Figure 8. STAT3 co-ordinates downstream IL-4 signalling outcomes in T cells. Studies have shown, for the first time, that STAT3 signalling capability is critical to regulating T cell supply of IL-4R⍺ and, by extension, responsiveness to the IL-4 environment. Subsequently, inhibition of STAT3 resulted in ablated STAT6 engagement. We suspect this affects IL-4 signalling outcomes, including the production of TGF-β1, GATA3 and IFN-γR1. Collectively, this molecular mechanism may have profound implications for T cell fate and function during infection/vaccination. Created using BioRender. www.nature.com/scientificreports/ biomarker that was significantly up-regulated in response to IL-4, independent of STAT3 and STAT6. Knowing that IL-4 can trigger the phosphorylation of STAT5 via γc-docked JAK1 55,60 , our findings may infer that STAT5 may be important for IL-4 signalling in T cells, specifically, given that Stat5a expression was maintained under STAT3 inhibition. Moreover, our findings also revealed reduced CD28 expression on IL-4R⍺ + cells, suggesting a pro-anergic phenotype. Similarly, down-regulated Pcna expression was observed following STAT3 inhibition, indicating that STAT3 may also be associated with T cell proliferation, which is consistent with previous findings 39 .
In conclusion, here we provide new insight into how different T cell subsets regulate IL-4 activity under different IL-4 conditions (Fig. 8). We show that naïve T cells directly respond to IL-4 (but not IL-13) in a STAT3/ STAT6 dependent manner. Collectively, our findings clearly demonstrate that, similar to IL-13/IL-13R⍺2 signalling in lung cDCs, STAT3 appears to be the master regulator of IL-4/IL-4R⍺ signalling in naïve T cells, allowing them to differentially regulate IL-4R⍺ and modulate downstream immune outcomes. This may implicate T cell fate and, by extension, pathogen immune evasion. We believe our findings have significant potential to inform the design of improved vaccine adjuvants against chronic viral pathogens, as well as therapies against IL-4/IL-13-associated diseases.

Methods
Mice and ethics statement. In this study, 6-8-week-old female naïve Charles River WT BALB/c mice were, purchased from the Australian Phenomics Facility at the Australian National University (ANU). WT BALB/c mice were chosen mainly because all our previous in-vivo infection/vaccination studies related to this study were conducted in this strain of mice 10,11,13,15,20 . All animals were maintained, monitored daily, and experiments were performed in accordance with the Australian National Health and Medical Research Council (NHMRC) guidelines within the Australian Code of Practice for the Care and Use of Animals for Scientific Purposes. Animal ethics were approved by the Australian National University's Animal Experimentation and Ethics Committee (AEEC), ethics protocol number A2017/15. Animals were maintained and used in accordance with the latest ARRIVE guidelines 61 . In vitro cytokine and TCR stimulation. Whole spleen or lung single-cell suspensions was prepared, as previously described 30,52,62 . In stimulation assays, whole spleen suspensions or two-step magnetic-activated cell sorted (MACS) naïve CD4 T cells were used. The MACS sorting was performed according to the manufacturer's instructions (Miltenyi Biotec), where firstly CD4 T cells were negatively selected and subsequently CD62L + cells were positively selected. 1 × 10 6 cells were stimulated with IL-4 (0.001-50 ng/mL) or 0.1-10 ng/mL of IL-13 was used to stimulate cells for 15 min and/or 24 h. Cells were incubated at 37 °C under 5% CO 2 . To dissect the role of TCR and co-stimulation in regulating IL-4 responses, single cell suspensions were seeded in plates coated with anti-mouse CD3ε (clone: 145-2C11, Biolegend), incubating 3 μg/mL for 18 h at 4 °C. Cells were subsequently seeded along with suspended anti-mouse CD28 (clone: 37.51, Biolegend) to a final concentration of 0.5 µg/mL. STAT6 and STAT3 inhibition assays. In order to evaluate the role of STAT3 and STAT6 in the IL-4/ IL-13 signalling in T cells, small-molecule inhibitors were used. Briefly, cell suspensions were pre-incubated with 20 μM Stattic (STAT3i) (Axon Medchem) and/or 100 nM AS161749 (STAT6i) (Axon Medchem) in complete RMPI media for 3 h (37 °C, 5% CO 2 ), which correspond to biological concentrations with ~ 85% inhibition 63,64 . Both inhibitor concentrations have previously shown to prevent cytokine-induced phosphorylation, with limited cross-talk among other members of the STAT family, as well as being non-toxic to cells 63,64 . Cells were subsequently stimulated with IL-4 as described above.
Cell sorting for Fluidigm 48.48 Biomark and qPCR assays. Following cytokine stimulation cells were suspended in LIVE/DEAD Fixable Blue (Thermo Fisher Scientific), according to the manufacturer's instructions. Following immunostaining, live CD3 + CD4 + CD8 -IL-4R⍺ + cells were sorted (BD FACS Aria II) directly into preamplification mixture. This mixture contained 2 × reaction buffer, SuperScript III RT/Platinum Taq Mix, 0.2X pooled assays (described in Sup. Table 1), SUPERase•In RNase Inhibitor and DEPC-treated water. Sorted cell mixture was centrifuged at 1454×g to break the cells and release the mRNA 40 . cDNA was subsequently synthesised according to the thermo-cycling program: 1 × 50 °C for 15 min, 1 × 95 °C for 2 min and 14-20× (95 °C for 15 s and 60 °C for 4 min) (single or 100 cells).

Real-time qPCR analysis.
To validate the primer probes qPCR was performed on 100 cells using TaqMan qPCR mix, which contained 1 µL of gene expression assay (primer probes in Sup. Table 1), 5 µL of 2 × TaqMan Universal PCR master mix, 1 µL cDNA and 4.5 µL DEPC-treated water. Targets were quantified using the 7900HT thermocycler, according to the following program: 1 × 50 °C for 2 min, 1 × 95 °C for 10 min and 40 x (95 °C for 15 s and 60 °C for 1 min). FAM fluorescence was normalised to ROX (6-carboxy-X-rhodamine). SDS v 2.4. was used to obtain cycle threshold (Ct) values. Fluidigm 48.48 Biomark gene expression assay. Fluidigm 48.48 Biomark gene expression assays were performed, as described previously 20,40 . Briefly, the integrated fluidic chip (IFC) (Fluidigm) was primed using the IFC Controller MX, according to the manufacturer's instructions. cDNA was diluted 1:1 with DEPC-treated water and 20× GE Sample Loading reagent was diluted 1:9 in TaqMan PCR Master Mix. Diluted cDNA and loading reagent were combined 1:1 and loaded onto the IFC chip. Primer assays were diluted 1:1 with 2X GE Assay Loading reagent and loaded into the IFC. The sample and assay were subsequently distributed on the chip via the IFC Controller MX loading program and the gene expression assay was performed and analysed using the GE 48.48 Standard.pcl on the Fluidim Biomark. Ct values were acquired from the Fluidigm Biomark, being normalised to ROX. Statistical analysis. Marker-positive cells (calculated as a percentage of the parent population) and MFI values were compared between treatment groups by ANOVA combined with post-hoc, unpaired and parametric multiple comparison tests. Expression of a given marker was represented as either the percentage markerpositive of the parent or as the geometric mean fluorescent intensity (MFI). Fold-change MFI was calculated as described elsewhere 25 and compared via paired Student's t test. Paired tests were similarly used to compared fold-change in transcript expression (2 -∆∆Ct values; normalised to L32, the endogenous control). Dichotomised Fluidigm Biomark outputs were compared between treatments by performing Fisher's exact test. For genes expressed by at least 15% of cells, L32-normalised Ct values were evaluated by performing a Spearman's rank correlation to establish association between gene expression values. This was subsequently used to perform a PCA, as described elsewhere 20,40 .