From Basic Science to Clinical Translation in Kidney Cancer: A Report from the Second Kidney Cancer Research Summit

Kidney cancer is one of the most common cancers and it is estimated that in 2020, more than 75,000 cases were diagnosed, the majority being renal cell carcinomas (RCC), causing over 13,000 deaths in the United States (1). Recent years have seen remarkable advances in the understanding of genomic drivers and pathogenesis of RCC (2, 3), especially within the most common histologic subtype, clear-cell RCC (ccRCC). These insights paved the way for the introduction of many systemic therapies for advanced/metastatic disease, most importantly angiogenesis inhibitors, including tyrosine kinase inhibitors (TKI; ref. 4) and anti-VEGF mAbs (5), as well as immune checkpoint inhibitors (ICI; ref. 6).

The standard of care for the first-line treatment of metastatic RCC (mRCC) is currently based on TKI-ICI or dual ICI combinations (7). While these combination regimens have yielded substantial survival benefits and durable responses, complete responses remain relatively rare and most patients eventually progress. Further improving the outcomes of patients with RCC will only be possible through novel target discovery, therapeutic development, and improved biomarker-based therapeutic allocation. These clinical and translational challenges require advances in the laboratory, where more representative RCC models are actively being developed. These challenges are even more pronounced for variant histologic subtypes of RCC, management of which lags that of ccRCC. Recently, an improved understanding of the molecular drivers of these diseases has begun to lead to improved treatment options for patients with these tumors (8–10).

Therapies targeting VEGF, mTOR, PD-1, and CTLA-4 signaling have become part of the standard of care for RCC as either monotherapies or combinations, with clinical trials aiming to further refine regimens currently ongoing [e.g., COSMIC-313 (NCT03937219); PDIGREE (NCT03793166); (NCT04736706)]. While these therapies have yielded prolonged remissions in up to 30% of patients (12–14), it is likely that novel therapeutic targets need to be explored to achieve sustained remissions more frequently.

HIF2α, a transcription factor that is a key driver of ccRCC, is one such novel target that is currently in clinical development (2). HIF2α was long thought to be undruggable, but structure-based drug discovery efforts provided small molecule inhibitors that prevent its activity by blocking its dimerization with HIF1β (2, 15). HIF2α inhibitors have demonstrated promising efficacy in both sporadic and Von Hippel–Lindau (VHL) syndrome–associated ccRCC in clinical trials (16–19). The HIF2α inhibitor belzutifan was recently approved for the treatment of malignancies associated with VHL syndrome and is currently being tested as a single agent in a phase III trial for the treatment of sporadic ccRCC (NCT04195750; ref. 20) or in combination with immune checkpoint inhibitors and/or TKIs [(NCT04736706), (NCT04586231); ref. 21].

Metabolic targets in RCC are in earlier stages of validation (Table 2). These include MESH1, a cytosolic NADPH phosphatase which has been shown to induce cystine deprivation leading to ferroptosis (i.e., iron-dependent cell death; ref. 22), a cell state associated with programmed necrosis in VHL-deficient RCC (23). Likewise, SLC7A11 encodes a cystine transporter that suppresses ferroptosis and is negatively regulated by BAP1 (24, 25). As a result, BAP1 mutations lead to a downregulation of ferroptosis and consequent tumor growth (25). As BAP1 is mutated in 15% of ccRCC and is associated with poor prognosis (26), this finding is particularly relevant. Cystine homeostasis mediated through SLC7A11 may prove to be a targetable pathway in BAP1-deficient ccRCC (27).

Profilin 1 (Pfn1), a protein that plays an important regulatory role in actin polymerization, is overexpressed in ccRCC and associated with more aggressive disease with poor prognosis (28, 29). In vitro experiments have shown that Pfn1 is a major regulator of tumor cell proliferation and migration in ccRCC (30). Small-molecule antagonists of the Pfn1-actin interaction appeared to significantly reduce tumor cell proliferation in ccRCC (30). TANK-binding kinase 1 (TBK1), a member of the IkB kinase family, is implicated in cell survival, autophagy, mTOR-AKT signaling, and KRAS-driven tumorigenesis (31, 32). In ccRCC, VHL loss has been shown to increase TBK1 activity through a hypoxia-independent mechanism, thereby identifying a novel function of VHL not related to HIF (33, 34). Pharmacologic inhibition of TBK1 resulted in substantially reduced proliferation of VHL-null RCC models, suggesting therapeutic vulnerability in kidney cancer (33, 34). Another promising target in RCC is IGF1R. Decreased expression of methylthioadenosine phosphorylase (MTAP), a frequent metabolic aberration in ccRCC, leads to IGF1R upregulation and is associated with worse outcomes (35). Linsitinib, a selective IGF1R inhibitor, displayed potent inhibition of tumor cell migration and invasion in preclinical models (35), and hence the contribution of IGF1R signaling in RCC progression merits further exploration.

Cyclin-dependent kinases 4 and 6 (CDK4/6), which control progression through cell cycle from G₁ to S (36), were recently identified in preclinical models as being synthetically lethal with VHL inactivation in a HIF-independent fashion (37, 38). CDK4/6 inhibitors are being studied preclinically in combination regimens with either HIF2α inhibitors or ICIs (refs. 37, 39; APART study, NCT pending) and in phase I/II clinical trials with VEGF inhibitors (NCT03905889).

As multiple cellular or metabolic pathways are explored, it will be crucial to find agents that do not potentiate each other's toxicities while optimizing antitumor efficacy through synergy, additivity, or independent action (40).

RCC is among the most immune-responsive solid malignancies despite having only a moderate tumor mutational burden (TMB; ref. 41). However, a large proportion of patients do not respond robustly to current immunotherapy-based regimens (6, 12, 14). The next breakthroughs in immunotherapy must address diverse tumor immune profiles and their dynamics to overcome intrinsic and acquired resistance to treatment and achieve better outcomes.

Current immunotherapies used in RCC target negative regulators PD-(L)1 and CTLA-4, and while combination regimens employing these agents have been studied, the search for new immune targets has been pursued in parallel. Other immune checkpoints under investigation include TIM-3, VISTA, LAG-3, PVRIG, and TIGIT (42–46). Recently, therapeutic agents directed against these molecules have been evaluated in phase I/II clinical trials with anti-PD-(L)1 inhibitors in different cancer types, yielding promising results (47–49). Some of these agents are currently being investigated in patients with mRCC in early phase I/II clinical trials, such as eftilagimod alpha (IMP321; NCT00351949), a soluble dimeric recombinant form of LAG-3, sabatolimab (MBG453; NCT02608268), an anti-TIM-3 mAb, and XmAb22841 (NCT03849469), a bispecific antibody targeting LAG-3 and CTLA-4 (Table 2).

Beyond immune checkpoints, another strategy is to inflame the RCC tumor microenvironment through the stimulation of innate immunity, such as by agonism of pattern recognition receptors such as Toll-like receptors (TLR) and/or the STING pathway (50, 51). This requires intratumoral injection of pro-inflammatory agonists, but could be avoided by developing tumor-targeting conjugates that can be delivered systemically in the form of prodrugs. TLR agonists are starting to be evaluated in RCC (NCT02650635, NCT03435640), among other solid malignancies, in phase I/II trials, while STING agonists in RCC remain to be tested in the clinical setting (Table 2). Oncolytic adenoviruses carrying immunostimulatory payloads honed to the kidney are also under preclinical investigation (52).

Personalized immunotherapy is likely to have an important place in future RCC treatment algorithms. Human endogenous retroviruses (hERV) are transcriptionally silent remnants of past retroviral infections, many of which become aberrantly expressed under hypoxic conditions in RCC (53). hERVs are potentially actionable drug targets: hERV-E was found to be antigenic and to elicit reactive antitumor CD8⁺ T cells in a patient with mRCC following allogeneic stem cell transplantation (54), and an autologous T-cell therapy engineered with a T-cell receptor (TCR) targeting hERV-E is in phase I development (NCT03354390). In addition, the hypomethylating agent decitabine has been shown to increase hERV expression and activate immune signaling in ccRCC cells, and could also be used to indirectly increase immunogenicity of RCC (55). A more patient-specific approach is personalized cancer vaccines, which are composed of antigens identified from sequencing a patient's TCR repertoire (56). Preliminary efficacy has been demonstrated in melanoma and glioblastoma (57, 58) and studies are currently underway in RCC (NCT02950766).

While sequencing-based analyses of the immune microenvironment can describe the types and activation states of infiltrating immune cells, the spatial interplay between tumor, endothelial, and immune cells—immune infiltration phenotypes—can be evaluated by techniques that preserve the spatial architecture of tumor tissue, such as multiplex immunofluorescence. A potential method involves three-dimensional (3D) reconstruction from multiplex immunofluorescence to elucidate the spatial distribution of different cell types in the tumor microenvironment (59). Spatial heterogeneity in tumor samples can be explored in relation to outcomes, and it has been shown that spatial clustering of CD68⁺ tumor-associated macrophages (TAM) was associated with worse overall survival in patients with mRCC (60). Another possibility uses model systems: recently, a 3D vascularized flow-directed system was developed using cells obtained from patients with ccRCC (61). This model is highly relevant as it recapitulates in vitro the major signaling interactions of RCC, providing a useful platform for investigating novel agents (61). In addition, such models can help to map the spatial distribution of immune cells, and explore cytokines and chemokines involved in immune cell migration. While spatial exploration of RCC cannot be employed in clinical practice, it represents an important modality that could help better unravel the patterns of therapeutic resistance, tumor growth, and tumor–immune interactions.

At this juncture, despite promising advances in immunotherapy of RCC, there are many unresolved questions. If TMB is low, which tumor antigens are driving response to immunotherapy? Why does CD8⁺ T-cell infiltration not correlate with better outcomes? Emerging methods, such as single-cell RNA sequencing (scRNA-seq) and single-cell TCR sequencing (scTCR-seq), which help determine the molecular underpinnings of each specific cell type within the tumor, are starting to be applied to kidney cancer and could help answer some of these questions through the high-resolution offered (62–64), potentially unraveling determinants of response to PD-1 blockade (65).

Mutations and copy-number alterations associated with response or resistance to anti–PD-1 monotherapy in RCC have been described recently (66–68). In some reports, PBRM1 loss-of-function (LOF) mutations have been linked to response to PD-1 blockade in metastatic ccRCC resistant to VEGF TKIs (67, 68). While other studies did not fully validate these findings (69, 70), subsequent investigations demonstrated the same previously described association (71, 72), arguing for the potential role of PBRM1 LOF mutations in predicting outcomes specifically and perhaps exclusively among patients with VEGF-resistant ccRCC treated with single-agent ICIs. Focal loss of 10q23.31 also appeared to correlate with better survival outcomes with ICIs in ccRCC (66). Recent analyses of clinical trials investigating ICIs in combination with anti-angiogenesis agents in patients with mRCC (i.e., JAVELIN Renal 101, IMmotion150) have identified transcriptomic signatures associated with better responses to immunotherapy-based regimens (73, 74). These signatures included multiple genes related to functions of the tumor immune microenvironment, such as TCR signaling, T-cell activation and proliferation, and myeloid inflammation, providing a clue for the development of biomarkers (73, 74). In addition, analysis of RNA-seq data from IMmotion151, the phase III trial of atezolizumab + bevacizumab versus sunitinib in advanced RCC, classified patients into biologically relevant clusters in an effort to predict responses to immune checkpoint blockade (75). Certain hERVs have been also characterized and associated with response to checkpoint blockade (76). In a pan-cancer analysis, RCC displayed the highest hERV expression, which was further shown to be strongly associated with immune checkpoint activation (76). With many available systemic agents and multiple therapeutic combinations approved for patients with mRCC, the identification of predictive biomarkers to guide treatment selection remains a high priority for the field. More recently, scRNA-seq analyses in RCC have investigated cell-type specific patterns of response to immunotherapy, and characterized a pro-inflammatory phenotype of TAMs and T-cell subpopulations in responders (63). In addition, tumor programs linked to better outcomes with ICIs were described using scRNA-seq and validated in bulk RNA-seq cohorts (63, 64).

As shown in multiple studies, immunotherapy appears to modify the clonal dynamics of the adaptive immune system (77, 78). Importantly, specific features of TCR and B-cell receptor (BCR) repertoires appear to correlate with response to ICIs across different cancer types. In melanoma, higher posttreatment clonality was associated with better survival outcomes (79). In addition, the expansion of CD8⁺ memory effector cytotoxic T cells was seen in responders during treatment (80). In non–small cell lung cancer, decreased TCR and BCR diversity after treatment with ICIs was seen among responders in the EGFR/ALK wild-type group (81). The relationship of TCR and BCR repertoires and response to ICIs in RCC remains to be further evaluated.

Among patients with ccRCC, African Americans (AA) have been shown to present with higher cancer-specific mortality (82). At the genomic level, these patients were less likely to harbor VHL mutations and exhibited lower expression of HIF- and VEGF-associated pathways as compared with Caucasians (83). In addition, comprehensive genomic and transcriptomic analyses showed distinct DNA methylation and mRNA expression profiles in the AA population (84). More studies are still needed to confirm these findings and to analyze the mutational landscape of RCC across ethnicities. Moreover, as genomic racial differences in patients with lung cancer have been recently linked to immunotherapy response (85), the association between race and clinical benefit from VEGF-targeted therapies and ICIs in patients with RCC should be explored.

Cell-free tumor DNA (cfDNA) can help monitor RCC progression and detect minimal residual disease following surgery (86). In other cancer types, it has also been found to be associated with survival endpoints (87), and identify responders to specific therapies (88). Cell-free methylated DNA immunoprecipitation sequencing (cfMeDIP-seq) is a novel, noninvasive technology that was recently validated as a promising approach to detect RCC from blood and urine samples (89) with a much higher sensitivity than cfDNA variant analysis (90). Future efforts could evaluate a potential role of cfMeDIP-seq in predicting responses to commonly used therapeutic regimens in kidney cancer.

While some proposed biomarkers are promising, they will need to be evaluated and validated in prospectively designed clinical trials before they can be utilized in clinical practice (91). A pioneering example of such a study in RCC is the phase II BIONIKK prospective trial, which employed a 35-gene expression mRNA signature for treatment arm allocation (92) and found that such signatures may help guide treatment decision-making (93). However, owing to the number of patients required and the costs of sample collection and analysis, these trials are difficult, time consuming, and expensive to run. Moreover, biomarkers specific to one systemic therapy regimen might not translate to another, which poses a particular challenge in RCC with so many approved agents. The main challenge for the field in the coming years will be to incorporate the most promising biomarkers (80, 94, 95) that are most likely to be incorporated into clinical practice due to ease of use, such as circulating biomarkers (cfDNA, cfMeDIP-seq, TCR/BCR repertoire), in the design of prospective kidney cancer trials, allowing for optimal selection of patients.

Variant RCC histologies (also called non-ccRCC) represent a heterogeneous group of cancers with unique genomic alterations and often dismal prognosis, including some variants with a higher proportion of pediatric cases than is seen in ccRCC (3). While discussion of each individual histologic variant is beyond the scope of this perspective, unifying features that limit research include pathologic misclassification at diagnosis, unclear disease etiologies, and a dearth of cell line and animal models. Recently, our understanding of the drivers in some of these diseases has improved.

Multi-region genomic analysis of papillary RCC (pRCC) recently revealed lower intratumor heterogeneity as compared with ccRCC (96). Moreover, many somatic copy-number alterations were identified as clonal, helping to better delineate the genomic features of pRCC (96). MET, which encodes a receptor tyrosine kinase involved in proliferation and angiogenesis (97), is frequently activated (through mutation or copy-number gain) in pRCC (98). Savolitinib, a selective MET inhibitor, was recently evaluated in MET-driven pRCC, and demonstrated encouraging efficacy (9), outlining the potential for molecularly targeted therapies based on genomic characterization of non-ccRCC tumors.

Chromophobe RCC (chRCC) represents less than 5% of all kidney tumors and appears to be resistant to ICIs (99). A known metabolic alteration underlying the pathogenesis of chRCC is the impairment of gamma-glutamyl transferase 1 (GGT1; ref. 100). Defects in the gamma-glutamyl cycle might have an important role in chRCC pathogenesis, potentially caused by cells with accumulating dysfunctional mitochondria. In addition, chRCC may be more sensitive to oxidative stress, a vulnerability that could represent a future avenue of investigation (100).

Renal medullary carcinoma is a rare disease which disproportionately afflicts young patients of African descent often expressing sickle cell trait (101). Recent molecular characterization of this tumor has shown that it is driven by inactivation of the tumor suppressor gene SMARCB1 and subsequent gain of MYC during disease progression (102). In addition, DNA damage repair pathways have been characterized as a therapeutic vulnerability with a high sensitivity to PARP inhibitors (i.e., olaparib) identified in preclinical models that should be further explored clinically (102).

Translocation RCC (tRCC) is a rare RCC subtype which nevertheless represents more than 50% of all pediatric RCCs, and is characterized by gene fusions involving the MiT/TFE gene family (TFE3, TFEB, or MITF; ref. 103). tRCC tumors have a “quiet” mutational landscape as compared with other subtypes of RCC (104, 105), and an increased expression of genes related to regulation of apoptosis, lysosomal function (mTORC1 signaling), and antioxidant stress response (105–107). tRCC display elevated NRF2 activity without harboring any somatic alterations in the associated pathway, which could be driving resistance to targeted therapies (106). These insights into the drivers of this disease could lead to novel molecularly targeted therapies.

Recently, pembrolizumab monotherapy and combination therapy with nivolumab and cabozantinib demonstrated promising antitumor efficacy in patients with non-ccRCC (108, 109). Other ICIs, as well as novel immunotherapy drugs, are currently under investigation in variant histologies of RCC (NCT03075423, NCT04413123, NCT04704219). Similar to ccRCC, immune-based combinations with other targeted agents in these rarer subtypes of kidney cancer, such as MET inhibitors in MET-driven pRCC (110), are being evaluated (NCT02819596, NCT05043090) and could offer better outcomes for patients (111).

From basic to clinical research, KCRS 2020 represented a unique opportunity for researchers and physicians to exchange ideas and tackle the biggest challenges in kidney cancer research. While much progress has been made in recent years in the management of patients with RCC (Table 1; ref. 11), much work remains to be done to obtain more durable remissions in a larger proportion of patients with mRCC.

T.K. Choueiri reports grants, personal fees, non-financial support, and other support from Merck, BMS, Exelixis, EMD Serono, and Pfizer, as well as other support from KidneyCAN during the conduct of the study; T.K. Choueiri also reports other support from NCCN panel, GUSC of the NCI, as well as personal fees from Up-to-Date outside the submitted work. L. Albiges reports other support from BMS, Novartis, Pfizer, IPSEN, Astellas, MSD, Merck & Co, Janssen, AstraZeneca, and Eisai outside the submitted work. M.B. Atkins reports grants and personal fees from BMS, as well as personal fees from Merck, Genentech-Roche, Novartis, Eisai, Pfizer, Exelixis, Werewolf, Pyxis Oncololgy, Asher Bio, Simcha, Takeda, SeaGen, AstraZeneca, Calithera, SAB Bio, X4 Pharma, Aveo, Iovance, Allogene, and Neoleukin outside the submitted work; M.B. Atkins also reports stock ownership in Pyxis Oncology and Werewolf. Z. Bakouny reports non-financial support from Bristol Myers Squibb, grants from Genentech/imCORE, and personal fees from UpToDate outside the submitted work. G. Bratslavsky reports other support from Merck and J&J outside the submitted work. D.A. Braun reports grants and personal fees from Exelixis; personal fees from AVEO, MDedge, CancerNetwork, CancerExpert Now, LM Education and Exchange Services, Trinity Group, and Insight Strategy; and non-financial support from Bristol Myers Squibb outside the submitted work. J.B.A.G. Haanen reports other support from Achilles Tx, Iovance BioTx, Instil Bio, Ipsen, Molecular Partners, Merck Serono, PokeAcel, Pfizer, Roche, Sanofi, and T-Knife; grants from Asher Bio, Amgen, and BMS; grants and other support from BioNTech, MSD, and Novartis; and grants and personal fees from Neogene Tx outside the submitted work. A.A. Hakimi reports personal fees from Merck outside the submitted work. M.A.S. Jewett reports Board Member and Chair, MAB of the International Kidney Cancer Coalition (advocacy volunteer group). E. Jonasch reports personal fees from Aveo, Calithera, Eisai, Exelixis, Ipsen, and Pfizer; grants and personal fees from Aravive, Merck, NiKang, and Novartis; and grants from Arrowhead during the conduct of the study. C. Labaki reports grants from Roche/Genentech outside the submitted work. D.F. McDermott reports personal fees from BMS, Merck, and Pfizer outside the submitted work. S.K. Pal reports personal fees from F. Hoffman-La Roche outside the submitted work, as well as research funding to institution from Eisai, Genentech, Roche, Exelixis, and Pfizer. T. Powles reports personal fees from AstraZeneca, BMS, Exelixis, Incyte, Ipsen, Merck, MSD, Novartis, Pfizer, Seattle Genetics, Merck Serono, Astellas, Johnson & Johnson, Eisai, and Roche; grants from AstraZeneca, Roche, BMS, Exelixis, Ipsen, Merck, MSD, Novartis, Pfizer, Seattle Genetics, Merck Serono, Astellas, Johnson & Johnson, and Eisai; and other support from Roche, Pfizer, MSD, AstraZeneca, and Ipsen outside the submitted work. W.K. Rathmell reports grants from Incyte, as well as personal fees from Sitryx and Caribou Bioscience outside the submitted work. B.I. Rini reports research funding to institution from Pfizer, F. Hoffman-La Roche, Incyte, AstraZeneca, Seattle Genetics, Arrowhead Pharmaceuticals, Immunomedics, BMS, Mirati Therapeutics, Merck, Surface Oncology, Dragonfly Therapeutics, Aravive, Exelixis, and Janssen; consulting with BMS, Pfizer, GNE/Roche, Aveo, Synthorx, Compugen, Merck, Corvus, Surface Oncology, 3DMedicines, Aravive, Alkermes, Arrowhead, GSK, Shionogi, Eisai, and Nikang Therapeutics; and stock ownership with PTC therapeutics. S. Signoretti reports grants and personal fees from Bristol Myers Squibb and AstraZeneca; grants from Exelixis and Novartis; and personal fees from CRISPR Therapeutics, Merck, NCI, and AACR outside the submitted work. In addition, S. Signoretti has a patent for Biogenex with royalties paid. N.M. Tannir reports grants and personal fees from Bristol Myers Squibb, Nektar, Calithera Biosciences, Novartis, and Eisai; personal fees from Pfizer, Exelixis, Eisai Medical Research, Eli Lilly, Oncorena, Surface Oncology, Ipsen, and Merck Sharp & Dohme; and grants from Arrowhead Pharmaceuticals outside the submitted work. H.J. Hammers reports personal fees and other support from BMS, Merck, and Surface; personal fees from Aveo, Pfizer, and Oncology; and other support from Aravive outside the submitted work. No disclosures were reported by the other authors.

The authors thank the speakers and attendees of the symposium; the Department of Defense Kidney Cancer Research Program, on which many of the projects presented at this conference were based; and KidneyCAN and their volunteers for organizational support.