Targeted Therapies Inhibit the HIF Pathway

One of the major downstream effectors of HIF is vascular endothelial growth factor (VEGF), a secreted protein that promotes angiogenesis. Thus, various tyrosine kinase inhibitors (TKI) have been employed to block signaling through the VEGF receptor (VEGFR). Among those used clinically are sunitinib, sorafenib, axitinib, pazopanib, cabozantinib, and lenvatinib (9). In addition, bevacizumab is a monoclonal antibody directed against VEGF itself. Inhibition of VEGF prevents the formation of robust tumor vasculature, leading to nutrient deprivation and reduced growth.

Other drugs target the activity or expression of HIF itself. Inhibitors of mammalian target of rapamycin (mTOR), a kinase central to many pro-oncogenic signaling pathways, are also used in patients with ccRCC. Drugs such as everolimus and temsirolimus are used in treatment-refractory ccRCC to block the upregulation of HIFα translation by mTOR (10). Finally, belzutifan, a recently developed HIF2α inhibitor, reduces activity of the major transcription factor that drives kidney cancer (11, 12). It has been approved as the first ever drug to treat patients with VHL disease (13) but is not yet approved for use in mRCC.

Immunotherapies (Re)Activate the Immune Response to ccRCC

As ccRCC tumors have relatively high levels of infiltrating T cells, another treatment strategy involves reactivation of the immune system against the cancer cells. Years ago, this was accomplished using cytokine therapy, where high-dose interleukin-2 (IL-2) or interferon-α (IFN-α) were administered to slow tumor growth and activate cytotoxic T-cell responses. Although a small number of patients achieved complete remission with high dose IL-2 or IFN-α, low response rates, considerable toxicity, and the advent of targeted therapy led to their disuse (14).

Immune checkpoint inhibitors (ICI), including anti-PD-1 (nivolumab, pembrolizumab), anti-PD-L1 (avelumab, atezolizumab), and anti-CTLA-4 (ipilimumab) drugs are frequently used in the management of ccRCC. These therapies work by neutralizing checkpoint molecules that dampen T-cell function, thus reactivating immune surveillance and fostering the destruction of cancer cells. Anti-PD-1 therapy in combination with a TKI is currently recommended as a first-line treatment for all patients with ccRCC due to improved response rates (15). Interestingly, neither tumor mutational burden nor PD-L1 status accurately predicts ccRCC patient response to ICI (16–18). It was recently shown, however, that the presence of tissue-resident CD8⁺ T cells correlates with improved responses to ICI in ccRCC (19).

Overview of the Type I IFN Pathway

The type I IFN pathway is canonically associated with the host response to viral infection, but its activity has been increasingly implicated in cancer biology. The pathway is activated by several inputs, including DNA damage, replication stress, bacterial or viral nucleic acids, mitochondrial DNA, and micronuclei. On exposure to pathogenic RNA/DNA or mislocalized self-DNA, various cytoplasmic pattern recognition receptors (PRR) sense the abnormal nucleic acids and initiate signaling cascades, including the cGAS-STING, MDA5/RIG-I-MAVS, and NF-κB pathways, to stimulate the phosphorylation of IRF3 and the production of many cytokines involved in the immune response (34) (Fig. 2). Among these cytokines are the type I IFNs, including 13 distinct IFN-α proteins and a singular IFN-β, which signal in both autocrine and paracrine manners. The binding of IFN to the interferon α receptor (IFNAR) stimulates JAK1 and TYK2 to phosphorylate the IFNAR and, in turn, recruit and phosphorylate STAT1 and STAT2. Phosphorylated STATs 1 and 2 bind IRF9, comprising the heterotrimeric ISGF3 transcription factor. ISGF3 transcribes various ISG, which function to put the cell in a state that is antiviral, antiproliferative, proapoptotic, and proimmunogenic (35). In addition, IRF3 and NF-κB, both transcription factors upstream of IFN production, can activate certain ISG (36, 37). Unphosphorylated ISGF3 (U-ISGF3) and the STAT2/IRF9 complex can promote ISG expression as well (38–41).

Open in a separate window

Figure 2.

Diagram showing how the type I interferon pathway is regulated, and how the major cancer genes in clear cell renal cell carcinoma (ccRCC) impact it. The figure was created with BioRender.com. BAP1, BRCA1-associated protein1; HIF, hypoxia inducible factor; IFN, interferon; PBRM1, polybromo1; SETD2, SET domain containing 2; VHL, von Hippel–Lindau tumor suppressor gene.

Virtually all nucleated cell types can produce IFN-β, but IFN-α is mainly produced by dendritic cells (DCs). All cells express the IFNAR, and, as such, all can respond to type I IFN, though the responses vary based on cell type.

Role of the Type I IFN Pathway in Cancer

In cancer, the type I IFN pathway has a predominantly tumor suppressive function. In fact, the success of many cancer treatments, including chemotherapy, radiotherapy, and immunotherapy, depends on an intact IFN pathway (42). The antitumor effects of IFN signaling in cancer are twofold: 1) it directly blocks the growth of cancer cells via reducing proliferation and promoting apoptosis (tumor-intrinsic effects) and 2) it indirectly promotes immune cell activation and infiltration of the tumor to block cancer growth (immune cell-dependent effects; Fig. 3). Type I IFN stimulates DC maturation and antigen cross-presentation to effector T cells, which is crucial for an antigen-driven antitumor immune response (43). IFN also activates natural killer (NK) cells to promote their survival and cytotoxic function against cancer cells (44) and antagonizes the functions of immunosuppressive cells, such as regulatory T cells (Treg) and myeloid-derived suppressor cells (MDSC) (45, 46). Plasmacytoid dendritic cells (pDCs) secrete large quantities of IFN-α on stimulation, which promotes the maturation of B cells into regulatory B cells (Breg) and/or plasmablasts depending on the local IFN-α concentration (47). Breg cells, in turn, stimulate the conversion of effector T cells into Tregs and dampen pDC IFN-α production via IL-10 to control inflammation. Thus, type I IFNs are a key effector in bridging the gap between the innate and adaptive immune systems.

Open in a separate window

Figure 3.

Schematic representation of how ISGF3 can be activated and how that can block clear cell renal cell carcinoma (ccRCC) tumor growth. The figure was made created BioRender.com. DC, dendritic cell; ISGF3, interferon-stimulated gene factor 3; MDSC, myeloid-derived suppressor cells.

Interestingly, the antitumor effects of type I IFN are not exclusively dependent on the immune system in preclinical models of cancer. IFN-β gene therapy reduced the growth of various xenografted human cancer cells in mice with severe immunodeficiencies (48), showing that the functions of T, B, and NK cells can be dispensable for its tumor suppressing effect. This phenomenon was also seen in syngeneic mouse models of melanoma and fibrosarcoma, where endogenous IFN-β suppressed tumor growth in the absence of T, B, and NK cells (49). In one study, cells transduced with IFN-β lost the ability to form tumors in vivo due to their entry into a senescent-like state (50). Moreover, depletion of STAT1, STAT2, or IRF9, the components of the ISGF3 transcription factor, or IFN-β itself significantly increased the growth of ccRCC xenografted tumors in nude mice, which lack T cells (32, 51). This supports the hypothesis that type I IFN exhibits a cancer cell-dependent tumor suppressor function, though it does not exclude the possibility that cells of the immune system play a significant role as well.

On the other hand, the type I IFN pathway has also been shown to confer resistance to cancer treatments, making its activation a double-edged sword. Upregulation of the type I IFN signature has been observed in cell lines and breast tumors resistant to radiation (52, 53). In colon cancer cells, overexpression of IRF9 alone caused resistance to various chemotherapies (54). Upregulation of IRF7, a transcription factor for type I IFN, also conferred resistance to chemotherapy in an ER⁻ breast cancer model (55). Chronic IFN stimulation can permit cancer cell dormancy, allowing tumor cells to escape immunosurveillance, or even permit the emergence of an immunosuppressive tumor microenvironment (56). For example, in a mouse xenograft model of ICI-refractory melanoma, chronic type I IFN activity was found to maintain the treatment resistance program initially established by type II IFN (57). Treatment with ruxolitinib, a JAK1/2 inhibitor, resensitized tumors to ICI therapy by reducing the expression of immune checkpoint receptors and partially reactivating exhausted T cells. Thus, the timing, duration, and context of type I IFN pathway activation dictate its effect on tumor growth. Activating type I IFN as a therapeutic strategy must be carefully considered based on cancer type, and ccRCC may uniquely benefit from it based on the molecular circuitry.

VHL

VHL mutations are observed in 70%–80% of patients with ccRCC, with many cases also showing epigenetic silencing by promoter hypermethylation (58). As described in The Biology of ccRCC Guides Drug Development, loss of VHL is the initiating event in ccRCC that leads to the stabilization of HIF and the hypoxia transcriptional program.

Chronic stimulation of kinases involved in type I IFN signaling has been observed in VHL^−/− ccRCC cell lines. In the absence of VHL, TBK1 was hyperactivated via a HIF-independent mechanism and promoted tumorigenesis in a kidney orthotopic xenograft model (59). Interestingly, TBK1 activation did not lead to IRF3 phosphorylation and IFN production, suggesting this is a novel function of TBK1 in ccRCC that is independent of its role in innate immune signaling.

Loss of VHL in ccRCC has been linked to oncogenic NF-κB signaling in multiple studies. In a meta-analysis of gene expression data from four studies, the NF-κB and type I IFN signatures were significantly upregulated in ccRCC compared with matched normal tissue (60). Both hereditary (VHL disease-associated) and sporadic ccRCC showed increased ISG expression compared with tissue from VHL patients missing only one copy of VHL (VHL^−/+), suggesting this was due to biallelic VHL loss. Immunohistochemistry revealed elevated nuclear staining of STAT1, a component of ISGF3, and RelA/p65, a component of NF-κB, in patients with ccRCC. Mechanistically, VHL stabilized transcriptional regulator ZHX2 in a hydroxylation-dependent manner, permitting its binding to RelA/p65 and activation of NF-κB target genes (61). In addition, VHL loss led to HIF-independent phosphorylation and activation of CARD9, which boosted the NF-κB pathway (62).

Conflicting results have been reported regarding the effect of VHL on ISG expression. In microarray analyses of 786-O VHL^−/− ccRCC cell lines, ISG expression was elevated in a HIF-dependent manner, and restoration of VHL reversed this effect (32) (Fig. 2). Furthermore, VHL re-expression in 786-O and Ren-02 ccRCC cells reduced both IFNB1 transcript levels (the gene encoding IFN-β) and ISGF3 target protein levels (51). On the contrary, another group showed VHL overexpression enhanced type I IFN production and ISG expression in 786-O cells in a STAT1-dependent manner (63). This difference may be attributed to the clonality of the cell populations: Liao et al. and Langbein et al. used polyclonal cultures, whereas Zhu et al. isolated individual VHL-expressing cells to form a monoclonal culture, which may be susceptible to biases. Consistent with the notion that VHL decreases ISG and antiviral activity, it was reported that loss of VHL led to HIF-dependent resistance to vesicular stomatitis virus (VSV) (64).

HIF

Constitutive activation of the HIF pathway is the major oncogenic event in ccRCC. HIF is a heterodimeric transcription factor consisting of an oxygen-sensitive α subunit and a stably expressed β subunit (HIF1β/ARNT). HIF1α and HIF2α are the two major isoforms of HIFα and play different roles in ccRCC. Although HIF2α has been firmly established as an oncogene in ccRCC (4), HIF1α functions as a tumor suppressor (65, 66), though Hif1α was recently implicated in tumorigenesis in an autochthonous mouse model of ccRCC (67).

To decipher the effects of HIF1α and HIF2α on ccRCC development and progression, a ccRCC mouse model was established, whereby Vhl, Trp53, and Rb were conditionally deleted in the kidneys (67). Deletion of Hif2α in this background led to increased expression of genes involved in the response to IFN-β, MHC antigen presentation, and T-cell activation, which led to higher amounts of intratumoral CD69⁺, perforin⁺ activated T cells. On the other hand, deletion of Hif1α in this background led to higher levels of intratumoral B cells. Interestingly, both Hif1α and Hif2α deletions increased the infiltration of CD8⁺ T cells and had no effect on PD-1⁺ cell abundance compared with Vhl^Δ/Δ/Trp53^Δ/Δ/Rb^Δ/Δ mice, suggesting that both HIF isoforms suppress the migration of cytotoxic T cells into the tumor and affect the tumor immune microenvironment (TIME) differently.

In contrast, ISGs were found to be enhanced by HIF2α in ccRCC cell lines (32) (Fig. 2). Suppression of HIF2α, and not HIF1α, reduced ISGF3 target levels in VHL^−/− cells, whereas expression of a stable HIF2α construct increased the IFN signature in VHL^+/+ cells. Mechanistically, HIF2α suppression in both Ren-02 and 786-O ccRCC cells decreased IFNB1 transcript levels, suggesting that HIF activates ISGF3 through IFN-β production (51).

The stimulation of ISG expression by HIF may also involve its regulation of NF-κB. In VHL^−/− ccRCC cell lines, upregulation of HIF1α or HIF2α activated NF-κB through EGFR-PI3K-AKT signaling (68). It remains to be seen whether this signaling downstream of HIF is responsible for ISG expression in ccRCC cells. Moreover, knockdown of HIF1α was shown to increase HLA-I expression in ccRCC cells (69), highlighting the opposing functions of the major HIF isoforms in this disease.

Type I IFN also appears to regulate the expression of HIF. In endothelial cells, IFN-α treatment increased both HIF1α and HIF2α transcripts in an ISGF3-dependent manner (70). HIF1α suppression did not affect ISG levels, suggesting this was not a feed-forward regulatory loop, and the impact of HIF2α suppression was not investigated. IFN-α also induced HIF1α in various cancer cell lines (71). Thus, the regulatory interactions between HIF and IFN appear to be bidirectional.

PBRM1

PBRM1 (BAF180) is the chromatin-targeting component of the SWI/SNF PBAF chromatin remodeling complex. PBRM1 mutations are most commonly found in ccRCC, cholangiocarcinoma, uterine corpus endometrial carcinoma, and stomach adenocarcinoma. In ccRCC, PBRM1 mutations are associated with a better prognosis relative to tumors with SETD2 or BAP1 mutations (28, 72, 73).

PBRM1 and KDM5C were found to regulate similar genes in microarrays performed in isogenic VHL^−/− ccRCC cell lines (32) (Fig. 2). Gene ontology analyses revealed that this common set of genes was enriched in ISG, which are targets of the ISGF3 transcription factor. Suppression of PBRM1 reduced the protein levels of STAT2 and IRF9, two components of ISGF3, and overexpression of STAT2 and IRF9 suppressed the growth of PBRM1-deficient ccRCC xenografts. Moreover, loss of PBRM1 correlated with reductions in nuclear IRF9 and cytoplasmic STAT2 in ccRCC patient samples, suggesting that PBRM1 positively regulates ISGF3 in human ccRCC. The mechanism by which this regulation occurs, however, is not yet known.

In colon cancer cell lines, PBRM1 knockdown increased the expression of ISG and to a lesser extent cGAS, a cytoplasmic dsDNA sensor (74). Although STING expression was unaffected, PBRM1 suppression increased phosphorylation of TBK1 and IRF3 in both unstimulated and viral mimic-treated conditions. Mechanistically, PBRM1 was found to bind within the promoters and first exons of RIGI (encoding RIG-I) and IFIH1 (encoding MDA5), two ISGs that are also upstream of IFN signaling, to repress their transcription. IFN-β treatment, in turn, reduced the transcript levels of PBRM1, indicating this is a feed-forward regulatory loop. This deviates from what is seen in ccRCC—it is possible that different modulating factors in various cancer types can change PBRM1 from an activator to a repressor of the same targets.

The prognostic significance of PBRM1 mutations in the response to ICI remains controversial. One group found that ccRCC tumors harboring PBRM1 mutations tended to be nonimmunogenic and therefore resistant to ICI (75). Another group found that improved responses to ICI were linked to loss of PBRM1 function in patients with ccRCC (76). As a result, additional analyses are needed, and the role of type I IFN in PBRM1’s effect on ICI response merits further investigation.

BAP1

BAP1 is a deubiquitinase that catalyzes the removal of ubiquitin from histone H2A at lysine 119 (H2AK119ub) and other substrates. Mutations in BAP1 are primarily found in mesothelioma, uveal melanoma, cholangiocarcinoma, and ccRCC. Loss of BAP1 function is associated with poor prognosis in various cancer types, including ccRCC (77–79), but better prognosis in mesothelioma (80–82).

Suppression of BAP1 was shown to reduce the protein levels of STAT2 and IRF9 as well as ISGF3 transcriptional targets in VHL^−/− ccRCC cell lines (32) (Fig. 2). Re-expression of BAP1 in BAP1^−/− ccRCC cell lines enhanced STAT1 and STAT2 phosphorylation and ISG expression, which was dependent on its deubiquitinase activity (51). Mechanistically, BAP1 upregulated STING expression and activation, leading to IFN-β production and canonical signaling through the IFNAR (51). In addition, BAP1 expression significantly correlated with cytoplasmic STAT2 and nuclear IRF9 expression in human ccRCC samples across various disease stages and grades.

In malignant peritoneal mesothelioma (PeM), loss of BAP1 expression is associated with significantly reduced STING mRNA levels (83). As the cGAS-STING pathway is a major stimulator of type I IFN production, this may result in suppression of the type I IFN pathway and enhanced tumor growth.

Multiple studies have shown that BAP1 status shapes the TIME. In uveal melanoma, loss of BAP1 was shown to correlate with increased NF-κB activity and higher levels of intratumoral CD8⁺ T cells (84, 85). This may lead to chronic IFN pathway stimulation that promotes tumor cell dormancy, T-cell exhaustion, and resistance to treatment, or may even promote the emergence of an immunosuppressive tumor microenvironment. Accordingly, mutations in BAP1 have been associated with increased presence of immunosuppressive cells in both uveal melanoma and mesothelioma (83, 86). In ccRCC, BAP1 deficiency is associated with a highly inflamed subtype and worse prognosis (87). Considering the aggressive features of cancers associated with BAP1 loss, it will be important to determine whether modulating the type I IFN pathway will benefit these patients.

SETD2

SETD2 is a methyltransferase that catalyzes the trimethylation of histone H3 at lysine 36 (H3K36me3) and methylation of α-tubulin at lysine 40 (88). Mutations in SETD2 are frequently observed in uterine corpus endometrial carcinoma, ccRCC, mesothelioma, and lung adenocarcinoma. Loss of SETD2 function is associated with poor prognosis in ccRCC (79).

In VHL^−/− ccRCC cell lines, suppression of SETD2 was shown to reduce the protein levels of STAT2 and IRF9 as well as ISGF3 transcriptional targets (32) (Fig. 2). Furthermore, immunohistochemical analyses of ccRCC tumor tissue revealed significant correlations between SETD2 expression and nuclear IRF9 or cytoplasmic STAT2 expression, indicating that this connection persists in human ccRCC.

In hepatocellular carcinoma cell lines, SETD2 was shown to regulate the IFN response to hepatitis B virus infection on multiple levels: SETD2 monomethylated STAT1 on K525 to permit its activation and transcriptional function and also trimethylated H3K36 in the promoters of select ISG to enhance transcription (89). Similarly, in mouse embryonic fibroblasts, IFN-β stimulation resulted in the recruitment of SETD2 and H3K36me3 occupancy in the promoters of ISG (90). At present, it is not known whether SETD2 functions though a similar mechanism to regulate ISG in ccRCC, but it is probable based on these findings in two different model systems.

KDM5C

KDM5C is a demethylase that removes methyl groups from histone H3 at lysine 4 (H3K4me2/3). Mutations in KDM5C are found in uterine corpus endometrial carcinoma, stomach adenocarcinoma, ccRCC, and lung adenocarcinoma.

In VHL^−/− ccRCC cell lines, knockdown of KDM5C reduced ISG expression, and though the mechanism is not yet deciphered, KDM5C was shown to physically interact with PBRM1 in HEK293T cells (32) (Fig. 2). In line with this, suppression of KDM5C decreased the IFN signature in unstimulated cultured preadipocytes (91).

In breast cancer cell lines, however, evidence suggests that KDM5C inhibits type I IFN signaling. KDM5C repressed IFN-β and ISG through direct binding and inhibition of TBK1 autophosphorylation (92). This was found to require its enzymatic activity in the cytoplasm. Knockdown of KDM5C increased CD8⁺ T-cell infiltration in an immunocompetent breast cancer xenograft model. Furthermore, KDM5C was shown to repress STING transcription via H3K4 demethylation, leading to reduced STING-TBK1-IRF3 signaling and reduction of the type I IFN signature in breast cancer cell lines harboring cytoplasmic DNA (93). Together, this suggests that the role of KDM5C in the type I IFN pathway is cell-type dependent.

HERV-E

Human endogenous retroviruses (HERVs) are a family of viruses integrated into the human genome as a result of ancestral exogenous retroviral infection of germline cells (94). The expression of HERVs has been silenced by methylation or interrupted by mutations and deletions accumulated throughout mammalian evolution (95). However, recent studies show that some products of HERVs play a pivotal role in modulating the TIME in various cancer types via activating the type I IFN signaling pathway (96, 97).

In ovarian cancer, treatment with DNA methyltransferase inhibitors (DNMTi) induced HERV demethylation, leading to HERV expression, triggering of dsRNA sensing and activation of ISGF3 (96). Similarly, higher levels of HERVs and ISG were detected in an experimental mouse model of mesothelioma compared with nontumor samples (97). Therefore, low type I interferon activity plays important roles such as mediating immune evasion and drug response to ICI in a mouse melanoma model.

KDM5B, another isoform in the KDM5 family, repressed endogenous retroelements such as MMVL30 through recruitment of SETDB1. In the absence of KDM5B, the expressed retroelements activate cGAS-STING and the type I IFN response, which led to tumor rejection and immune memory. In these models, depletion of KDM5B induced strong adaptive immune responses and enhanced the response to ICI (98).

HERV expression was found to be most significant in ccRCC based on a pan-cancer computational workflow that identified more than 3,000 transcriptionally active HERVs (99). HERV expression was associated with RIG-I signaling (99) and response to ICI in ccRCC (100). PBRM1 loss activated HERV expression, which depended on HIF (101). HERV-E, a specific type of HERV, is highly specific to ccRCC, and HERV-E TCR transduced CD8⁺/CD34⁺ T cells are being tested in a clinical trial in patients with mRCC (95, 102, 103). In both small cell lung cancer and RCC, de-repression of HERV expression led to activation of the innate antiviral pathway involving type I IFN (104, 105). The impact of HERVs on ISGF3 merits further investigation in ccRCC, where their expression is prominent.

ISGF3 Expression

Genetic manipulation of ISGF3 has been shown to alter tumor growth in ccRCC nude mice xenograft models. Knockdown of STAT1, STAT2, or IRF9 caused significantly larger tumors to form in nude mice, and IRF9 knockdown reduced CD45⁺ immune cell recruitment into xenografted tumors (32). Furthermore, overexpression of STAT2 and IRF9 reduced the growth of both BAP1- and PBRM1-deficient ccRCC xenografts (51), demonstrating that ISGF3 functions as a tumor suppressor in this context. An important caveat of these studies is that the mice used were deficient in T cells. How ISGF3 activation influences T-cell recruitment, cytotoxicity, and exhaustion in ccRCC warrants further investigation. Nevertheless, an advantage of these strategies is that ISGF3, the main effector downstream of type I IFN, was directly modulated, which may bypass some of the negative consequences of systemic IFN administration. A disadvantage is that intratumoral overexpression of ISGF3 is not plausible in the clinic. Pharmacological reactivation of ISGF3 by a molecule other than IFN itself may improve patient outcomes.

STING Agonists

Currently, several STING agonists are being investigated in the treatment of various cancers. In preclinical models, stimulation of STING leads to IRF3 and NF-κB transcriptional activity, causing the production of type I IFNs, other proinflammatory cytokines, and ISG that slow tumor growth. Small molecule STING agonists fall into two major categories: cyclic dinucleotides (CDN), which mimic 2’-3' cGAMP (the endogenous STING ligand), and noncyclic dinucleotides, which take various forms. A major concern regarding CDN is the frequent requirement for intratumoral injection due to low stability in circulation, which limits their use to accessible tumors (113). Newer STING agonists, including TAK-500, TAK-676, SNX281, and E7766, use a more stable scaffold that permits their systemic administration and are currently being investigated in clinical trials (NCT05070247, NCT04879849, NCT04420884, NCT04609579, and NCT04144140).

STING activation alone or in combination with ICI may help draw immune cells into tumors and reactivate their effector functions. In a syngeneic immunocompetent mouse model, systemic treatment with diABZI, a non-CDN, caused regression of colon cancer tumors in a CD8⁺ T cell-dependent manner (114). Interestingly, diABZI also suppressed the growth of ccRCC BAP1-deficient xenografted tumors in nude mice, which lack T cells (51). Treatment was shown to increase IRF3 phosphorylation, ISG expression, and CD45⁺ immune cell recruitment into tumors. It is possible that the unique molecular circuitry of ccRCC renders T cells dispensable for the efficacy of diABZI in this model. This hypothesis deserves further attention. With any immune-stimulating therapy, however, there is concern regarding possible autoimmune effects. A therapeutic window that exacts the maximal clinical benefit with tolerable side effects will need to be identified.

The mutation frequencies reported here are based on data generated by the TCGA Research Network: https://www.cancer.gov/tcga.