In a search for potential drivers of resistance to anti-PD-L1 immunotherapy, Huseni, Wang, Klementowicz, and Yuen et al. identified IL-6 as a correlate of poor response across various tumor types in large clinical trials. To better understand how IL-6 impacts tumor control and responses to immunotherapy, the researchers analyzed several sources of clinical data and performed various in vitro and in vivo experiments to begin to elucidate the mechanism. Their results were recently published in Cell Reports Medicine.

To begin, the researchers evaluated biopsies from IMmotion150 – a phase II trial of first-line atezolizumab (with or without bevacizumab) versus sunitinib in patients with metastatic RCC. This revealed that in atezolizumab-treated patients, expression of IL-6 (among other inflammatory factors) was associated with progressive disease (PD), and in atezolizumab monotherapy-treated patients, high tumor IL-6 also correlated with poor overall survival, even in patients with a strong CD8⁺ T cell signature. Ingenuity pathway analysis also identified IL-6, along with the IL-6 receptor, and the IL-6-induced transcription factor STAT3 as a putative regulators of PD-associated genes. Plasma IL-6, which was higher at baseline in patients compared to healthy controls, was also associated with reduced overall survival in all treatment arms. Similar results were found in atezolizumab-treated patients with different types of cancer in different trials.

To evaluate this pattern in vivo, the team used a mouse model bearing syngeneic EMT6, modeling triple-negative breast cancer. In this setting, the combination of anti-IL6R and anti-PD-L1 antibodies improved tumor control and progression-free survival compared to either monotherapy. Dual blockade of IL6R and PD-L1 also enhanced cytotoxic T lymphocyte (CTL) function and increased the frequency and abundance of CTLs and polyfunctional CTLs. The effects on CD4⁺ T cells, Tregs, and myeloid cells were limited, supporting an increased CD8:CD4 ratio. Similar results were observed in mice bearing MC38 and CT26 tumors. Additionally, inflammatory Ly6C⁺ cancer-associated fibroblasts were identified as the main source of IL-6.

Next, the researchers selected patients from the clinical trials IMmotion150 (RCC) and IMvigor210 (UC) with high or low levels of plasma IL-6. A number of genes were shared among IL-6-high patients and IL-6-low patients across trials. Gene analysis showed that high IL-6 was more associated with a quiescent state, while low IL-6 was more aligned with activation and effector functions, IFN signaling, the MHC-I pathway, and oxidative phosphorylation. Several clusters of CTLs, including clusters of polyfunctional effector cells, terminal effector-like cells, and early activated cells were associated with low IL-6. In contrast, inactive and functionally impaired cells with low expression of IFNG and GZMB were associated with high IL-6.

To better understand the mechanism by which IL-6 restricts CTL effector functions, the researchers turned to in vitro studies. Using SIINFEKL-activated OT-I cells or CD3/CD28-activated CTLs from spleens of WT or IL6R KO mice cultured with IL-6 under various conditions, they showed that IL-6 inhibited expression of IFNγ, TNF, and granzyme B by peptide-activated CTLs; reduced secretion of IFNγ, IFN-responsive chemokines, and GM-CSF; and limited cell division. In contrast, IL-6 also improved cell viability and induced production of IL-10 and factors related to Th2/Th17 responses. IL-6 also repressed the effector functions of both naive and memory CTLs and attenuated their ability to kill peptide-loaded MC38 cells.

While IL-6 inhibited CTL activation in vitro, blockade of IL6R in vivo only showed clear effects in combination with anti-PD-L1, suggesting that the effects of IL-6 may only become apparent in the setting of strong T cell activation. This was confirmed in vitro, as IL-6 had limited effects on suboptimally activated T cells, but inhibited expansion of IFNγ+ CTLs with the addition of activation-enhancing PD-L1 blockade.

Next, the researchers activated CTLs in the presence of IL-6 or anti-IL6R, and found by RNA sequencing that IL-6 suppressed expression of effector genes, chemokines, and transcription factors necessary for effector differentiation, while promoting expression of genes associated with naive and central memory states and transcription factors that oppose effector differentiation. Reactome pathway analysis and mapping of IL-6-regulated genes onto transcriptional modules associated with differentiation stages of CTL revealed further evidence that IL-6 limited T cell activation and effector differentiation. Similar analysis of data from IMmotion150 showed that higher IL- 6 was associated with a reduced effector:naive expression ratio. Investigating how timing plays a role, the researchers found that IL-6 affected differentiation of CTLs primarily during early priming. Looking at the impact of IL-6 preconditioning, the researchers treated CTLs with IL-6 for 1 day before initiating T cell activation, which suppressed effector functions at day 3. This suggests that IL-6 exposure can alter the expression of downstream regulators that limit T cell functions, regardless of ongoing IL-6 signaling.

Comparing CTLs stimulated with IL-6 with or without anti-CD3/CD28 activation, the researchers found that IL-6 induced expression of IFNγ, genes associated with co-inhibitors (CTLA-4, TIGIT) and repressors of inflammatory signaling (IL-10 and SOCS-family members), while suppressing expression of genes associated with co-stimulation (CD28, CD226) and dendritic cell cross-talk (CD40L, XCL1). These effects were also found to be dependent on STAT3 and downstream BATF.

Turning back to mouse models, Huseni, Wang, Klementowicz and Yuen et al. adoptively transferred wild-type or IL6r-deficient OT-I cells into mice bearing OVA-expressing tumors. In combination with OVA vaccination and anti-PD-L1, the IL6r-deficient, but not the wild-type OT-I cells enhanced polyfunctional CTL responses. Next, the researchers generated mice with CTL-restricted IL6R deficiency. In these mice, MC38 tumors grew more slowly, and were better controlled with anti-PD-L1 immunotherapy due to stronger CTL responses. RNAseq showed that anti-PD-L1 treatment induced genes related to IL-2, TGFβ, and VEGF signaling in WT CTLs, whereas it drove genes associated with IFN response and OxPhos in IL6R-deficient CTLs. Overall, the IL6R-deficient CTLs showed enhanced effector differentiation compared to their IL6R-proficient counterparts, which maintained more naive and memory-like phenotypes.

While the role of IL-6 is multifaceted, it is possible that the function of IL-6 in promoting naive/memory-like properties associated with longevity also limits effector differentiation, antitumor immune responses, and the initial efficacy of immune checkpoint blockade. Blocking IL-6 signaling could potentially be useful in enhancing immune priming, and the availability and widespread experience with clinically approved therapies targeting IL-6 and IL6R support the rapid clinical translation of combining IL-6 with immune checkpoint blockade.

Write-up and image by Lauren Hitchings

Meet the researcher

Our first interview of the year features lead author Nathan West and co-first author Mahrukh Huseni.

What was the most surprising finding of this study for you?
IL-6 is a well known pro-inflammatory factor in various pathophysiological settings (such as arthritis, cytokine-release syndrome, etc), due in part to its ability to promote differentiation of inflammatory CD4⁺ T cell phenotypes. Therefore, we were surprised to see such a potent inhibitory effect of IL-6 on classical CD8⁺ cytotoxic T cell function in vitro and in mouse models of tumor immunotherapy, along with a clear association between high IL-6, CD8⁺ T cell dysfunction, and poor outcomes to PD-L1 blockade in patients with cancer. This emphasizes the complex, context-dependent nature of cytokine biology, and tells us that even the most well-studied factors in immunology can still yield surprises.

What is the outlook?
A key outstanding question is the impact of IL-6 or IL6R blockade on CD8⁺ T cell function in humans, particularly in settings that are expected to involve significant antigen-specific CD8⁺ T cell responses, such as cancer or viral infections. It will be very interesting to characterize in detail the phenotype and function of CD8⁺ T cells before and after treatment with anti-IL-6 or anti-IL6R therapies. Because agents targeting IL-6 signaling are well established in the clinic, we anticipate that clinical studies of IL-6 blockade, most likely in combination with other approved immunotherapy agents, can be pursued readily by both academic and industry investigators.

What was the coolest thing you’ve learned (about) recently outside of work?
NW: Over Christmas holiday with my family in British Columbia, I discovered that the native red squirrels can zip around efficiently through deep snow by burrowing tunnels at ground-level and along tree trunks, which I thought was pretty neat! I also learned how uninhibited my six-year-old son becomes when provided with a toboggan and steep hills.
MH: During my recent travels to Costa Rica, I had a chance to meet with a Shaman, and was impressed by their knowledge of both traditional and modern medicine, and their eagerness to blend both forms to treat disease and improve lives. Of course, the eco- and bio-diversity there is also awe-inspiring. I was able to observe various species of monkeys in the wild, and was fascinated by their inter- and intra-clan hierarchical structure and social interactions in their native environment. It is amazing how much we have in common with our distant cousins!