Interferon signaling restrains renal cell carcinoma heterogeneity

Abstract

Type I interferon (IFN) is central to cancer surveillance as it mediates both direct and immune-mediated oncosuppressive effects. Recent data suggest that the ability of renal cancer cells to tolerate complex karyotypic alterations as elicited by chromosomal instability (CIN) and ultimately acquire full-blown metastatic is also negatively regulated by IFN signaling.

Type I interferon (IFN) is a family of pleiotropic cytokines that play a key role in the control of viral infection and developing neoplasms [1]. Indeed, while weak, indolent and chronic type I IFN responses have been consistently associated with accelerated tumor progression in the context of increased cancer stemness and failing immunosurveillance [2,3], robust, acute and resolving type I IFN signaling not only mediates direct cytostatic or cytotoxic effects on (at least some) malignant cells, but also (1) suppresses tumor-associated neoangiogenesis, and (2) promote both innate and adaptive immunity against neoplastic cells [2,4]. Recent data from Perelli and collaborators demonstrate that intact type I IFN responses prevent renal cell carcinoma (RCC) cells from tolerating complex karyotypic alterations as elicited by chromosomal instability (CIN), de facto curbing their ability to acquire additional heterogeneity and evolve towards a highly aggressive phenotype with full metastatic potential [5] (Figure 1A).

Figure 1. Oncosuppressive activity of type I IFN.

A. Upon disruption of chromosome 4p mouse renal cancer cells acquire elevated levels of chromosomal instability (CIN), which is accompanied by the acquisition of intratumoral heterogeneity. Type I interferon (IFN) signaling limits such process by promoting cell senescence and death via cancer-cell intrinsic mechanisms driven by type I IFN receptors (IFNARs). In line with this notion, loss of chromosome 16q, which encodes for several IFN receptors, is positively selected by evolutionary mechanisms, resulting in accrued cancer cell malignancy and heterogeneity. B. Type I IFN exerts potent oncosuppressive effects via direct and indirect (including immunological) mechanisms. Cancer cells evading such mechanisms, for instance upon IFNAR downregulation or via the establishment of an immunosuppressive tumor microenvironment (TME), are hence advantaged and generate rapidly progressing and heterogenous neoplasms. RCD, regulated cell death.

Building upon results from the TRACERx consortium, identifying NF2, SETD2 and VHL as the most prevalently disrupted tumor suppressor genes in RCC [6], Perelli and colleagues harnesses high-throughput in vivo and ex vivo platforms of somatic mosaic genetically engineered mouse models (SM-GEMMs) leveraging CRISPR-based genome editing in models of RCC to determine that disruption of chromosome 4q (which is homologous to human 9p21.3) as achieved by deletion of Cdkn2a and Cdkn2b (which encode endogenous cell cycle inhibitors) in Nf2^−/−Setd2^−/− or Vhl^−/−Setd2^−/− genetic backgrounds generates rapidly disseminating tumors that overall resemble aggressive human RCC with loss of 9p21.3 [5]. Further genomic analyses revealed a high similarity between these mouse models and human RCC, suggesting a common convergent strategy for the acquisition of metastatic potential involving recurrent loss of chromosome 12 and 16 coupled with chromosome 5 gains [5].

In general, Nf2^−/−Setd2^−/−Cdkn2a^−/−Cdkn2b^−/− tumors stratified into two genomic clusters: (1) a cluster characterized by relatively high chromosomal instability (CIN) and recurrent copy number variations (CNVs) along with the accumulation of micronuclei eliciting cyclic GMP-AMP synthase (CGAS) signaling; and (2) a cluster characterized by limited CIN and inconsistent CNVs. In line with previous data linking CIN with CGAS activation and accrued metastatic potential [7], Nf2^−/−Setd2^−/−Cdkn2a^−/−Cdkn2b^−/− tumors from the former cluster were more aggressive than their counterparts from the latter. Similar results were obtained by stratifying RCC patients from the “The Cancer Genome Atlas” (TCGA) based on 9p21 loss [5], suggesting that the disruption of 4q (in mice) or 9p21 (in humans) contributes to the emergence of aggressive cancer cell populations.

To delineate the molecular pathways underlying this observation, Perelli and colleagues employed single cell RNA sequencing on Nf2^−/−Setd2^−/−Cdkn2a^−/−Cdkn2b^−/− genetically engineered kidney organoids (GEKOs). Computational deconvolution of >80,000 cells from these GEKOs by two independent algorithms linked CIN with an elevated transcriptomic heterogeneity, generating 18 different clusters of cells. In line with data from SM-GEMMs and patient samples, cells from Nf2^−/−Setd2^−/−Cdkn2a^−/−Cdkn2b^−/− GEKOs also expressed genes associated with cell cycle progression, mesenchymal plasticity and sarcomatoid differentiation. Importantly, single cell trajectory analyses and cross-platform structural variant analyses revealed that Nf2^−/−Setd2^−/−Cdkn2a^−/−Cdkn2b^−/− GEKO cells clustered into 2 major subclasses diverging at the loss of chromosome 16, which was in line with previous whole-genome sequencing data on SM-GEMMs. Of these two clusters, cells lacking chromosome 16 exhibited the largest distance from the origin of the route, suggesting this as the endpoint for intratumoral RCC evolution. Moreover, 16q^−/− RCC cells generated more aggressive tumors upon implantation into immunodeficient mice than their 16q^+/+ counterparts, pointing to a major oncosuppressive role for one or more gene locus encoded in chromosome 16 [5].

Importantly, mouse chromosome 16 and human chromosome 21 displayed a considerable degree of homology in a conserved 200-kilobase genomic region coding for multiple IFN receptors (i.e., IFNAR1, IL10RB, IFNAR2, IFNGR2). In line with this notion, 16q^−/− RCC cells exhibited defective type I (as well as type II and III) IFN responses as well as the activation of genetic programs linked to cell cycle progression. Similar findings were obtained in multiple public transcriptomic datasets of RCC samples for which genomic information including metrics of aneuploidy are available, including (but not limited to) TCGA and TRACERx [5]. These results suggest that high levels of CIN enable accelerated RCC evolution that is restrained by IFN receptor-coding genes encoded by chromosome 16q (in mice) and 21 (in humans). The impact of indolent CGAS signaling as driven by CIN [7] on this instance of tumor evolution remains to be formally elucidated.

Finally, Perelli and collaborators went on to test the impact of various pharmacological and genetic strategies targeting type I and II IFN signaling on the aggressiveness of Nf2^−/−Setd2^−/−Cdkn2a^−/−Cdkn2b^−/− cells bearing or not intact 16q. In this setting, the aggressive phenotype associated with the loss of 16q could be phenocopied by the deletion of Ifnar1 or Ifng1 as well as by the administration of a small agent targeting the IFN signal transducer Janus kinase 1 (JAK1). Conversely, administration of recombinant type I or II interferon mediated potent oncosuppressive effects in vivo, at least in partially by promoting a permanent proliferative arrest commonly known as cellular senescence [8] in RCC cells. Finally, the authors harnessed a murine model of Down syndrome with a partial trisomy of chromosome 16 spanning the IFN receptor locus to elegantly demonstrate a gene dosage-dependent inhibitory effects on tumor initiation and progression as elicited by the Nf2^−/−Setd2^−/−Cdkn2a^−/−Cdkn2b^−/− genotype [5].

Taken together, these findings extend the oncosuppressive activity of type I IFN signaling to restrain the evolution of RCC with elevated CIN (Figure 1A). Given the critical role of IFN responses in anticancer immunosurveillance, it will be important to evaluate the impact of functional immunity in this process. This is especially true considering that CGAS signaling driven by CIN may act per se as a barrier against tumor evolution (rather than a driver thereof) in the presence of innate and adaptive immune effector cells, as previously demonstrated in a variety of natural and therapy-driven settings [9,10]. Irrespective of these and outstanding questions, the findings from Perelli and colleagues strengthen the importance of type I IFN signaling for cancer cell-intrinsic oncosuppression (Figure 1B).