Uncovering mechanisms underlying complement-mediated cancer immune evasion

Main text

The days of old where the complement system was recognized as merely a signaling cascade important for hematological control of bacterial infections are long gone. New, diverse functions are now routinely attributed to the complement system, particularly in the context of cancer immunology. These novel roles for the complement system are often studied on single immune cell subsets and/or in a single signaling pathway; thus, tying these disparate roles for complement signaling in the context of cancer immunology together is critical for a comprehensive understanding of the tumor microenvironment (TME). Enter Luan et al. In their article in this issue of Molecular Therapy,¹ they show how signaling through the complement 5a receptor (C5aR) on tumor-associated macrophages (TAMs) induces a pro-tumorigenic phenotype in these cells and restricts TAM production of the T cell attractant CXCL9, resulting in impairment of CD8⁺ cytotoxic T lymphocyte tumor infiltration and corresponding tumor progression. These findings not only provide a more unified view of the role of the complement system in the TME but pave the way for rational therapeutic targeting of this signaling axis.

The impetus to investigate a role for the complement system in cancer was born from early studies showing that the complement protein C5a, an anaphylatoxin and byproduct of the splitting of C5 into C5a and C5b by C5 convertase, has immunomodulatory properties² and that elevated levels of its receptor (C5aR) are correlated with worse clinical outcomes.^3,4 Multiple discrete roles for C5aR in immune cell regulation have already been identified, including the association of C5aR with an anti-inflammatory, M2-like macrophage phenotype, a link between C5aR and impaired T cell cytotoxic activity,⁵ and elevated CXCL9 levels with C5aR inhibition.⁶ Luan et al. have now performed the critical studies linking these prior findings together, revealing dual roles for C5aR in controlling TAM responses to ovarian carcinoma.

The first step the authors took to define this C5aR-TAM-CXCL9-T cell axis was to uncover a more precise role for C5aR signaling on macrophage subtypes in mice. Luan et al. found that C5aR was preferentially expressed on CD206⁺ immunosuppressive TAMs compared to other macrophage and immune cell subsets.¹ Supporting this association of C5aR with an immunosuppressive TAM phenotype were experiments both showing that tumor-conditioned media induce C5aR expression on macrophages and that macrophages lacking C5aR have improved toxicity toward cancer cells.¹ To begin to tie this association back to human cancer, the authors used data from The Cancer Genome Atlas (TCGA) to show strong correlations of C5AR1 with both M2 TAM levels and markers of M2 TAMs (CD209 and CD163) in the ovarian carcinoma cohort of TCGA.¹ While there are multiple lines of evidence the authors present to link the expression of C5aR with an M2-like macrophage phenotype, precisely how C5aR alters macrophage development and function can still be evaluated further. Specifically, it has become increasingly apparent that macrophages do not adopt these binary M1 or M2 phenotypes but in fact reside on a continuum encompassing many intermediary states.^7,8 Future studies employing high-dimensional single-cell approaches could help to dissect these subtle cell states and the clinical implications this may have when considering therapeutic targeting of C5aR.

The authors further investigated the intrinsic effect of C5aR on macrophages through multiple RNA sequencing studies on both primary murine TAMs and in-vitro-stimulated bone-marrow-derived macrophages (BMDMs) where they found, as expected, that C5aR^−/− macrophages have a less M2-like gene signature.¹ Interestingly, by digging deeper into these data, Luan et al. noted that the “leukocyte migration” pathway was enriched in C5aR^−/− macrophages along with increased expression (at the RNA and protein levels) of multiple chemokines, notably CXCL9.¹ Given CXCL9 is well established as a CD8⁺ T cell chemoattractant, this prompted the authors to hypothesize that C5aR signaling on TAMs may be controlling tumor immune responses via a second mechanism involving CD8⁺ T cell tumor infiltration. Murine studies examining the effect of C5aR^−/− macrophages and tumoral T cell infiltration confirmed that a lack of C5aR expression on macrophages does indeed induce higher levels of CXCL9 and, correspondingly, more CD8⁺ T cells in tumors in a CXCL9-dependent manner.¹ In vivo murine tumor models and prior studies showing that a small-molecule inhibitor of C5aR (PMX-53) also induces CXCL9 from macrophages⁶ lend further credence to this mechanism. The authors took care to perform multiple studies, including clodronate liposome depletion and BMDM adoptive transfers, to verify that altered expression of CXCL9 from macrophages in vivo was responsible for changes in CD8⁺ T cell tumor infiltration; however, future studies employing a conditional knockout of C5aR on macrophages (a model not currently available) would help to remove all possible confounding variables and provide definitive proof.

The authors round out their investigations of C5aR in ovarian carcinoma by performing studies using primary samples from human patients to confirm the association between C5aR (and C5a) with both CXCL9 levels and CD8⁺ T cell tumor infiltration. Their results largely mirror those from their above-mentioned murine studies. They observed a significant negative correlation between CXCL9 (at both the RNA and protein levels) and C5a and C5aR expression in primary human patients with ovarian carcinoma and statistically significant increases in survival in patients with ovarian carcinoma with low expression of C5AR1.¹ Perhaps most enlightening was a correlation analysis from immunohistochemistry samples of primary ovarian carcinoma showing a negative correlation between C5aR and CD8 expression, which provides strong support at the protein level for a C5aR-TAM-CXCL9-T cell axis.

Through the integration of multiple murine cancer models along with studies of primary human ovarian carcinoma samples, Luan et al. fit together multiple pieces of the puzzling relationship between the complement system and immuno-oncology. Much of the literature surrounding C5 and C5aR signaling in cancer suggest that inhibition of this C5aR signaling should have potent anti-tumor activity, which begs the question of why this receptor has not yet been targeted in clinical trials. In fact, the PMX-53 small-molecule compound the authors use here to pharmacologically inhibit C5aR signaling has been subjected to clinical trials; however, it failed a phase 2 trial for rheumatoid arthritis and psoriasis due to bioavailability issues and limited half-life.⁹ A newer small-molecule inhibitor of C5aR (INF904) produced by InflaRx has just recently shown a favorable safety and tolerability profile in a phase 1 study, yet it remains to be seen if this compound will overcome the issues faced by PMX-53. The results from Luan et al. in this issue of Molecular Therapy provide insights on both the mechanism of action of C5aR inhibition and which cell types should be targeted (read: TAMs), which will be invaluable in future work directed toward precision manipulation of C5aR signaling. Given C5aR signaling has established roles across multiple cancer subtypes and inflammatory diseases, the utility of pharmacological manipulation of this receptor cannot be understated.