Granulocytes Are Unresponsive to IL-6 Due to an Absence of gp130

This article is featured in In This Issue, p.3315

Interleukin-6 is a pleiotropic cytokine that is capable of driving immunity to infection, hematopoiesis, and inflammation (1). Dysregulation of IL-6 is linked to many chronic inflammatory and autoimmune diseases, which has led to its targeted inhibition in the clinic (2). Inhibition of IL-6 responses with tocilizumab (TCZ), a humanized anti–IL-6Rα Ab, has been successfully used to treat patients with rheumatoid arthritis (RA), juvenile idiopathic arthritis, Castleman disease, and cytokine release syndrome (2–6). It also appears to be a promising approach to prevent acute graft-versus-host disease (7). IL-6 mediates its broad biological activities through versatile signaling mechanisms, including classical, trans, and the recently described cluster signaling pathways, that use a receptor system composed of the ligand-specific IL-6Rα and the common gp130 signal transducer (1, 8, 9).

IL-6 classical signaling requires IL-6Rα and gp130 expression by the target cell and is driven by extracellular IL-6 binding to membrane-bound IL-6Rα (mIL-6Rα), inducing gp130 recruitment and homodimerization. Alternatively, IL-6 trans and cluster signaling requires only gp130 expression by the target cell, and IL-6/IL-6Rα complexes are formed externally. During trans signaling, IL-6 binds to soluble (s)IL-6Rα that interacts with gp130 expressed by the target cell. Cluster signaling has been proposed in dendritic cell (DC)–T cell interactions, such that IL-6/IL-6Rα complexes are formed intracellularly by DCs prior to trafficking to the cell surface and subsequent interaction with gp130 expressed by target T cells within the immunological synapse (8, 9).

All IL-6 signaling pathways result in gp130 homodimerization and phosphorylation of STAT3 and STAT1 proteins, which translocate to the nucleus to induce transcriptional modification of target genes (10). Because gp130 is reported to be ubiquitously expressed (11–13), IL-6 trans signaling has been proposed to widen the spectrum of IL-6 to virtually all cells in the body, and a few studies have confirmed the direct ability of human leukocyte populations to respond to IL-6 (14). Moreover, the effects of IL-6 on neutrophil function has been unclear, with conflicting reports regarding its influence on neutrophil apoptosis and downstream signaling (15–20). In this article, we demonstrate that granulocytes (CD66b⁺Gr; neutrophils and eosinophils) are unresponsive to IL-6 classical and trans signaling due to their absence of gp130 expression. Given that granulocytes constitute 50–70% of circulating leukocytes, this demonstrates that the targets of IL-6 signaling are somewhat more restricted than previously thought.

Peripheral blood was collected from healthy adult donors (n = 10) or clinical trial patients (day 60 and 90 after allogeneic stem cell transplantation [allo-SCT], ACTRN12614000266662, n = 28). Apheresis products were collected from G-CSF–mobilized stem cell donors for allo-SCT (G-CSF at 10 μg/kg daily for 4 or 5 d prior to collection). All mouse studies used 8–12-wk-old female C57BL/6 mice that were purchased from the Animal Resources Center (Perth, WA, Australia). All studies were approved by the institutional ethics committees, and all patients signed informed consent.

Abs used in flow cytometry and high-throughput imaging were purchased from BioLegend (mouse IgG1 [MOPC-21], mouse IgG2a [MOPC-173], rat IgG2a [RTK2758], rat IgG2b [RTK4530], anti-human CD45RA [HI100], CD38 [HIT2], CD123 [6H6], CD126 [IL-6Rα, UV4], CD8 [SK1], CD3 [HIT3a], CD4 [RPA-T4], CD14 [HCD14], CD10 [HI10a], anti-mouse CD11b [M1/70], Ly6C [HK1.4], Ly6G [1A8], CD4 [GK1.5], CD3ε [145-2C11], c-kit [CD117, 2B8], and FcγII/IIIR [CD16/32, 93]), BD Biosciences (mouse IgG2a [MOPC-173], and anti-human p-STAT3 [pY705, 4/P-STAT3], p-STAT1 [pY701, 4/A-STAT1], CD34 [581], CD33 [WM53], gp130 [CD130, AM64], CD66b [G10F5], CD45 [HI30], anti-mouse CD8α [53-6.7], and Sca-1 [D7]), eBioscience (anti-mouse IL-6Rα [CD126, D7715A7] and CD34 [Ram34]), and R&D Systems (anti-human gp130 [CD130, 28126] and anti-mouse gp130 [CD130, 125623]). For murine stem cell and progenitor analysis, the following lineage (Lin) Ab mixture was used (BioLegend): anti-mouse CD3ε (145-2C11), CD5 (53-7.3), Ter-119 (TER-119), Gr-1 (RB6-8C5), Mac-1 (M1/70), and B220 (30-F11).

Cells in whole unmanipulated human and mouse peripheral blood (100 μl) were surface stained and stimulated (15 min at room temperature) with recombinant human G-CSF (100 ng/ml; Amgen), human IL-6 (5–50 ng/ml; Life Technologies), mouse IL-6 (10–30 ng/ml; BioLegend), or hyper–IL-6 (H-IL-6; human or mouse, 15–150 ng/ml; R&D Systems) or were left unstimulated. A similar molar ratio of free IL-6:IL-6 within the H-IL-6 construct (free IL-6 molecular mass ∼ 20 kDa versus H-IL-6 construct molecular mass ∼ 59 kDa IL-6/sIL-6R/linker protein complex) was used. For time-course experiments, granulocytes were enriched by Ficoll-Paque density separation and stimulated with recombinant human G-CSF (100 ng/ml) or human IL-6 (50 ng/ml) for the indicated time or were left unstimulated. Erythrocytes were lysed, and cells were fixed with BD Phosflow Lyse/Fix Buffer (BD Biosciences). Cells undergoing intracellular staining were permeabilized with chilled Perm Buffer III (BD Biosciences) for 30 min on ice. Permeabilized cells were stained for p-STAT3 and p-STAT1 or IgG2a isotype controls for 1 h. Samples were washed and analyzed using flow cytometry (BD LSR Fortessa) and FlowJo software (v9.9). Unimodal p-STAT3 and p-STAT1 expression in myeloid cells was displayed as geometric mean fluorescent intensity (GMFI) values minus isotype controls, and bimodal p-STAT3 and p-STAT1 expression in T cells was expressed as the percentage of positive cells after gating on isotype controls. For murine studies, bone marrow (BM) was flushed, splenocytes were isolated by mechanical disruption, and both samples were treated with lysis buffer, to remove contaminating erythrocytes, and surface stained.

For high-throughput imaging analysis, cells were also stained with Hoechst 33342 nuclear dye (0.3 μg/ml), and fluorescent images were visualized and acquired via an ImageStream X Imaging Flow Cytometer (Amnis; EMD Millipore).

To assess gp130 expression on granulocyte progenitor cells, naive mouse BM was isolated, treated with lysis buffer to remove contaminating erythrocytes, and stained for Lin and progenitor cell markers (Lin⁻Sca-1⁺c-kit⁺ [LSK] progenitor cells and within the Lin⁻Sca-1⁻c-kit⁺ gate, megakaryocyte-erythroid progenitor [MEP] cells [CD34⁻FcγII/IIIR⁻], common myeloid progenitor [CMP] cells [CD34⁺FcγII/IIIR^lo], and granulocyte-monocyte progenitor cells [CD34⁺FcγII/IIIR^hi]) (21). To assess gp130 expression on human progenitor cells, CD34⁺ cells were enriched by magnetic depletion of Lin⁺ cells (BD IMag Human Lineage Cell Depletion Set) and stained for progenitor markers (human hematopoietic stem progenitor cells [HSPCs; Lin⁻CD34⁺CD38⁻] and within the Lin⁻CD34⁺CD38⁺CD33⁺CD10⁻ gate, CMP cells [CD123⁺CD45RA⁻] and granulocyte–monocyte progenitor cells [CD123⁺CD45RA⁺]) (22). The responsiveness of mouse and human granulocyte progenitor cells to IL-6 was assessed as above for peripheral blood samples (100 μl).

Peripheral blood was collected from healthy adults, and granulocytes (SSC^hiCD66b⁺CD14⁻), monocytes (CD66b⁻CD14⁺), CD4⁺ T cells (CD3⁺CD4⁺CD8⁻), and CD8⁺ T cells (CD3⁺CD4⁻CD8⁺) were purified for quantitative PCR (qPCR) and immunoblotting analysis. Granulocytes were enriched from lysed whole blood samples (BD Pharm Lyse buffer; BD Biosciences) by magnetic separation (CD66abce MicroBeads Kit human; Miltenyi Biotec) and further purified by FACS (BD FACS Aria, >99% granulocyte purity). Monocytes, CD4⁺ T cells, and CD8⁺ T cells were collected by Ficoll-Paque density gradient of peripheral blood and sort purified (BD FACSAria). In all sort protocols, cells were maintained at 4°C, and dead cells were excluded by 7-aminoactinomycin D (7AAD) staining.

Neutrophils were purified from the peripheral blood of healthy adults by Ficoll-Paque density gradient separation (>97% granulocyte purity), followed by negative immunomagnetic isolation enrichment (EasySep Direct Human Neutrophil Isolation Kit; STEMCELL Technologies, >99.8% granulocyte purity). Five milliliters of peripheral whole blood was diluted in 20 ml of saline, overlaid onto 20 ml of Ficoll-Paque Plus (GE Healthcare), and centrifuged (515 × g, 20 min, room temperature, no brake). After centrifugation, the top layers of plasma, the PBMC interface, and Ficoll-Paque were removed, leaving an RBC and granulocyte pellet. The pellet was resuspended to 5 ml with saline, and neutrophils were enriched by removing the RBCs and remaining contaminating cell types by negative immunomagnetic isolation using an EasySep Direct Human Neutrophil Isolation Kit (STEMCELL Technologies), as per the manufacturer’s protocol.

Neutrophils were purified from human peripheral blood by Ficoll-Paque density separation (followed by negative immunomagnetic isolation), plated at 1 × 10⁶ cells per milliliter in IMDM supplemented with 10% FCS, and stimulated with IL-6 (20 ng/ml) or G-CSF (20 ng/ml) at 37°C, 5% CO₂ for 24 h or were left unstimulated. Neutrophil apoptosis was determined using an Annexin V–Apoptosis Detection Kit (BD Biosciences), as per the manufacturer’s protocol.

For qPCR analysis, sort-purified CD66b⁺Gr, monocytes (CD14⁺Mo), CD4⁺ T cells, and CD8⁺ T cells were transferred to Buffer RLT (QIAGEN) and stored at −80°C. Total RNA was extracted using an RNeasy Micro Kit (QIAGEN), and cDNA was prepared with a Maxima H Minus First Strand cDNA Synthesis Kit (Thermo Fisher). Gene expression was determined by qPCR using TaqMan GE assays (Applied Biosystems) for human IL6ST (Hs00174360_m1) and were run in parallel with the housekeeping gene B2M (Hs00187842_m1). Reactions were run and analyzed on a ViiA 7 Real-Time PCR System (Thermo Fisher). IL6ST gene expression was determined using the comparative Ct method (2^−ΔΔCt) normalized relative to B2M, and one sample within the dataset was used as a reference.

For gp130 protein immunoblotting, protein cell lysates were prepared from sort-purified CD66b⁺Gr, CD14⁺Mo, CD4⁺ T cells, and CD8⁺ T cells using RIPA lysis buffer (20 mmol/l Tris [pH 8], 150 mmol/l NaCl, 10% glycerol, 1% Nonidet P-40) containing 1× cOmplete Protease Inhibitor Cocktail (Roche). For STAT3/p-STAT3 immunoblotting, granulocytes were purified using a Ficoll-Paque density gradient, followed by negative immunomagnetic bead enrichment, and were washed and stimulated with IL-6 (50 ng/ml) or G-CSF (100 ng/ml) for 15 min or were left unstimulated. After stimulation, cells were washed twice with PBS, and protein cell lysates were prepared using RIPA lysis buffer containing 2× cOmplete Protease Inhibitor Cocktail and 1 × PhosSTOP Phosphatase Inhibitor (Roche). Using a Bio-Rad Protein Assay Kit (Bio-Rad, Hercules, CA) and the Bradford method, 20 μg of denatured proteins was separated using 10% SDS-PAGE and transferred to polyvinylidene difluoride membranes. The membranes were blocked for 1 h in 5% skim milk or BSA in TBST (10 mM Tris-HCl, 150 mM NaCl, 0.1% Tween-20). Immunoblotting was performed with primary Abs anti-gp130 (1:500; Santa Cruz Biotechnology), anti-Erk1/2 (1:1,000), total STAT3 (1:2000, clone D3Z2G), or p-STAT3 (Tyr⁷⁰⁵, clone D3A7, 1:1,000) (Cell Signaling Technology) and detected with HRP-conjugated anti-rabbit IgG (1:10,000; Cell Signaling Technology). After applying ECL detection reagents (GE Healthcare or Thermo Fisher Scientific), protein bands were visualized using x-ray film (Fujifilm). The intensity of p-STAT3 and STAT3 bands was scanned from films, and the level of STAT3 activation was calculated using the following formula: percentage of STAT3 activation = intensity of p-STAT3/(intensity of STAT3 + intensity of p-STAT3) × 100%.

Stimulation data sets using IL-6, H-IL-6, and G-CSF were compared with their unstimulated values using one-way ANOVA with the Dunnett multiple-comparisons test. Surface receptor expression was compared individually between cell types using repeated-measures one-way ANOVA, with the Tukey multiple-comparisons test. The Kruskal–Wallis test with the Dunnett multiple-comparisons test was used to evaluate differences in gp130 protein expression and IL6ST and IL6R gene expression across multiple groups. Data are presented as mean ± SEM, and a p value < 0.05 was considered statistically significant. Statistical analyses were performed using Prism 7.01 (GraphPad).

IL-6 signaling was investigated in circulating CD66b⁺Gr, CD14⁺Mo, CD4⁺ T cells, and CD8⁺ T cells in healthy adult peripheral blood without prior manipulation. Cells were stimulated with IL-6 (classical signaling) or rIL-6/IL-6Rα protein complex H-IL-6 (trans signaling), and p-STAT3 and p-STAT1 levels were determined (Figs. 1, 2). In response to IL-6 (Figs. 1A, 1B, 2A–D) and H-IL-6 (Figs. 1C, 1D, 2A–D) stimulation, we observed a significant and dose-dependent upregulation of p-STAT3 and, to a lesser extent, p-STAT1 in CD14⁺Mo, CD4⁺ T cells, and CD8⁺ T cells. Unlike CD14⁺Mo and CD4⁺ T cells, CD8⁺ T cells showed a heterogeneous response to IL-6 and H-IL-6 stimulation, with ∼40% of cells upregulating p-STAT3, demonstrating diversity in IL-6 signaling within this population (Figs. 1, 2C, 2D). Critically, however, CD66b⁺Gr were unable to upregulate p-STAT3 or p-STAT1 in response to multiple concentrations of IL-6 or H-IL-6 (Figs. 1, 2A, 2B) or to upregulate p-STAT3 over multiple time points (Fig. 2E). This inability to respond to IL-6 or H-IL-6 could not be attributed to impaired cell function, because G-CSF stimulation of CD66b⁺Gr induced p-STAT3 expression as expected (Figs. 1A, 1C, 2A).

We next confirmed this unresponsiveness of granulocytes to IL-6 stimulation using immunoblotting. Because low-level impurities in granulocyte preparations can potentially alter experimental outcomes (23), we isolated granulocytes using Ficoll-Paque density separation, followed by negative-immunomagnetic separation, which resulted in granulocyte populations with high purity (>99.8%, Fig. 2F). Minimal p-STAT3 was detected in unstimulated granulocytes and, consistent with flow cytometry data, there was no response to IL-6 stimulation (Fig. 2G, 2H). In contrast, G-CSF stimulation elicited a strong p-STAT3 response. Furthermore, culturing purified neutrophils in vitro with G-CSF, but not IL-6, protected granulocytes from apoptosis, consistent with the latter’s inability to signal (Fig. 2I, 2J). These data suggest that, despite the pleiotropic effects of IL-6, circulating human CD66b⁺Gr are not targets of IL-6 signaling, through classical or trans signaling pathways.

To understand the apparent differential responses to IL-6 stimulation, we examined the key components of the IL-6R system required by classical (gp130 and mIL-6Rα) and trans (gp130) signaling pathways in CD66b⁺Gr, CD14⁺Mo, CD4⁺ T cells, and CD8⁺ T cells. Using flow cytometry (Fig. 3A–C) and high-throughput imaging analysis (Fig. 3E, 3F), we observed gp130 expression on CD14⁺Mo, CD4⁺ T cells, and CD8⁺ T cells. Approximately 40% of CD8⁺ T cells did not express gp130 (Fig. 3A, 3B), reflecting their heterogeneous response to IL-6 stimulation (Figs. 1, 2C, 2D). Surface gp130 was coexpressed with mIL-6Rα on ∼70% of CD14⁺Mo and ∼40% of CD4⁺ T cells, but on a smaller fraction of CD8⁺ T cells (Fig. 3C). In contrast, healthy adult CD66b⁺Gr exhibited a striking absence of gp130 expression (Fig. 3A–C, 3E, 3F), consistent with their inability to respond to IL-6 or H-IL-6 stimulation. Importantly, this absence of surface gp130 protein was not limited to healthy adult CD66b⁺Gr; it was also observed in CD66b⁺Gr 60 and 90 d after allo-SCT (Fig. 3D), a setting in which IL-6 dysregulation is profound, and IL-6 inhibition results in very low rates of acute graft-versus-host disease (7). To exclude possible gp130 internalization or preferential expression of gp130 splice variants in CD66b⁺Gr, we confirmed the absence of gp130 protein in purified (>99%, Fig. 3G) CD66b⁺Gr by Western blot (Fig. 3H, 3I) and low IL6ST gene expression by qPCR (Fig. 3J). Moreover, IL6ST gene expression in granulocytes was significantly lower than in CD14⁺Mo, CD4⁺ T cells, and CD8⁺ T cells in the HemaExplorer public online database (Fig. 3K) (24). Finally, we confirmed the absence of gp130 staining in neutrophils with a second mAb (Fig. 3L). We consistently observed mIL-6Rα expression on CD66b⁺Gr, despite the fact that mIL-6Rα is unable to signal in the absence of gp130 (Fig. 3A–C). Therefore, granulocytes are refractory to IL-6 due to the absence or very low protein expression of the signal transducer gp130.

Given the broad use of murine models to understand IL-6 responses and delineate IL-6 signaling pathways (1, 8, 16, 25, 26), we further investigated IL-6 responses in murine granulocytes. As observed in humans, we identified a striking inability of circulating murine neutrophils (Ly6G⁺) and eosinophils (Siglec-F⁺SSC^hi, data not shown) to upregulate p-STAT3 or p-STAT1 in response to IL-6 classical (IL-6) or trans (H-IL-6) signaling (Fig. 4A–E). This was in contrast to murine monocytes (Ly6C^hiMo) and T cell subsets, for which significant upregulation of p-STAT3 was observed after IL-6 and H-IL-6 stimulation. Importantly, granulocyte function was not inherently defective in these assays, because parallel G-CSF stimulation elicited a strong p-STAT3 response in Ly6G⁺ (Fig. 4A, 4D). Despite gp130 expression in circulating murine monocytes and T cells, the gp130 IL-6 signal transducer was not detected in murine neutrophils by flow cytometry (Fig. 4F). This absence was not restricted to circulating neutrophil populations, because Ly6G⁺ derived from primary lymphoid tissue (spleen) and BM also lacked gp130 expression (Fig. 4G). Together, these data demonstrate that, consistent with human circulating granulocytes, murine neutrophils are refractory to IL-6 stimulation through IL-6 classical and trans signaling pathways due to the absence of gp130.

Because previous studies have reported a significant role for IL-6 signaling in granulopoiesis (27, 28), we next investigated the ability of murine and human granulocyte progenitor cells to respond to IL-6. In this study, we detected gp130 expression in LSK cells, CMP cells, and granulocyte-monocyte progenitor cells in murine BM (Fig. 5A–C), as well as p-STAT3 upregulation when stimulated with IL-6 (Fig. 5D, 5E). In contrast, MEP cells did not express gp130 and, consequently, did not respond with p-STAT3 upregulation when stimulated with IL-6. Next, we investigated whether human HSPCs, CMP cells, and granulocyte-monocyte progenitor cells were also responsive to IL-6. In G-CSF–mobilized leukapheresis products, all three progenitor populations expressed gp130, and IL-6 stimulation resulted in significant upregulation of p-STAT3 (Fig. 5F–J). Therefore, although mature granulocytes are not responsive to IL-6, IL-6 may instead drive granulopoiesis via effects on their progenitors.

IL-6 is a pleotropic cytokine that modulates multiple components of the adaptive and innate immune systems to promote and maintain inflammation (reviewed in Ref. 1). The cells mediating the downstream inflammatory cascades include T and B lymphocytes, in addition to diverse myeloid lineages. Granulocytes represent the largest circulating innate population in humans and are well-described mediators of inflammation in a range of responses to infectious pathogens and autoimmune Ags, in which IL-6 is known to drive pathogenicity via the secretion of cytokines, chemokines, and proteases (29). In this article, we demonstrate that, contrary to current dogma, IL-6 does not directly influence granulocyte function. The role of IL-6 in granulocytes has been unclear, with multiple conflicting studies (15–20). For instance, IL-6 has been reported to promote STAT3 phosphorylation in granulocytes, but not to alter function or apoptosis (16), whereas Dienz et al. (15) suggested that IL-6 promoted the expression of antiapoptotic genes and neutrophil survival but not STAT3 phosphorylation. Other studies have suggested that IL-6 can increase (17) or decrease (18–20) neutrophil apoptosis in vitro. Recent reports have shown that contaminating cell populations in common neutrophil-isolation procedures can affect experimental outcomes, including cytokine expression in cultured neutrophils (23), although impurities have been noted to have little impact on transcription profiles (30). Thus, these conflicting data sets likely reflect the difficulties in interpreting the effects of cytokines on mixed myeloid populations that exhibit poor survival in vitro and the potential for indirect effects of cytokine stimulation on signaling cascades. In support of this, adoptive transfer of labeled neutrophils into patients receiving IL-6R inhibition has recently demonstrated a definitive effect of IL-6 on neutrophil migration (with retention in liver and spleen) in RA patients, but it excluded any effect on apoptosis or function (31). Our data clearly demonstrate that IL-6 does not induce p-STAT3 or p-STAT1 activation in granulocytes. Furthermore, we conclude that this is due to a lack of gp130, which is required for signal transduction of all known forms of IL-6 signaling; therefore, IL-6 does not directly affect granulocyte function or apoptosis. Thus, the clear effects of IL-6R inhibition on neutrophil migration are likely to be indirect effects that are mediated by IL-6 signaling in other cell types, such as endothelial cells or monocytes. Consistent with this, although (usually mild to moderate) neutropenia is observed in a significant fraction (<20%) of RA patients receiving TCZ treatment, there has been no reported temporal link between a decrease in circulating granulocytes and an increase in infection (32, 33).

Interestingly, we consistently observed mIL-6Rα expression on granulocytes, despite the absence of its signal transducer, gp130. sIL-6Rα, which is required for IL-6 trans signaling, is primarily generated by proteolytic cleavage of mIL-6Rα, and neutrophils have been shown to release sIL-6Rα upon apoptosis (34). Moreover, neutrophils have been proposed to play a key role in promoting IL-6 trans signaling, driving the transition from neutrophilic to mononuclear infiltration and resolving inflammation (25, 35). Thus, these data suggest that mIL-6Rα expression by granulocytes may act as an important source of sIL-6Rα in vivo. Alternatively, in concert with IL-6 secretion, mIL-6Rα in mature granulocytes could potentially provide cluster signaling to other gp130-expressing target cells, such as that proposed during DC–T cell interactions (8, 9). However, there is no evidence that cluster signaling can be performed by non-APCs.

Although human peripheral T cells expressed gp130 and responded to IL-6, we did observe differential mIL-6Rα and gp130 expression and the capacity to respond to IL-6. Furthermore, gp130 expression on CD4⁺ and CD8⁺ T cells was much lower in patients after allo-SCT than in healthy adults, suggesting that gp130 is downregulated during inflammation. Previous studies have reported that mouse T cells downregulate mIL-6Rα and gp130 upon TCR or IL-6 stimulation (36, 37). Moreover, this downregulation of gp130 was reported to be associated with acquisition of memory markers, such as CD44, rendering cells unresponsive to IL-6 (36, 38). Therefore, reductions in gp130 expression and responsiveness to IL-6 in human peripheral blood T cells after allo-SCT are likely a reflection of activation in response to alloantigen or pathogens within a lympho-deplete environment; however, this does not account for the lower expression of gp130 in CD8⁺ T cells relative to CD4⁺ T cells in peripheral blood of healthy donors. This is in contrast to murine CD8⁺ T cells from naive mice that expressed high levels of gp130. This disparity is likely to reflect the relatively high abundance of memory T cells in humans and the presence of significant numbers of additional CD8⁺ T cells, such as mucosal associated invariant T cells, which did not express gp130 (data not shown). It is also important to note that, although mIL-6Rα expression on T cells was low, as long as these cells expressed gp130, they responded to IL-6. Moreover, this responsiveness of T cells could not be attributed to the presence of sIL-6Rα (and trans signaling) in our assay, because addition of the IL-6 trans signaling inhibitor sgp130:Fc failed to inhibit IL-6–induced STAT3 phosphorylation (data not shown). Thus, the expression of IL-6Rα is likely to be underreported by current flow cytometry approaches.

Finally, although mature granulocytes are refractory to IL-6 due to the absence of gp130 expression, their progenitors do express gp130, confirming that this receptor subunit is lost during cellular maturation. Thus, IL-6 can still influence granulopoiesis at an early stage of development, which may account for the neutropenia that is sometimes observed in RA patients after TCZ treatment (3, 32, 33, 39, 40). Alternatively, TCZ has been suggested to induce neutropenia by acting on mature neutrophil migration and margination (31); however, by virtue of signaling incompetence within these cells, this is likely due to indirect effects. Importantly, our data confirm that IL-6 inhibitors are unlikely to function via the modulation of mature neutrophil responses; instead, they are likely to mediate effects by inhibiting signaling in T cells, monocytes, and target tissue.

We thank Assoc. Prof. Steven W. Lane for providing reagents and acknowledge the assistance of the Flow Cytometry and Imaging Facility at QIMR Berghofer, Paula Hall, Michael Rist, and Grace Chojnowski.

G.R.H. has received funding from Roche for clinical trials of IL-6 inhibition in transplantation. The other authors have no financial conflicts of interest.