Immune checkpoint blockade in renal cell carcinoma

Abstract

Immune checkpoint blockade (ICB) is the foundation of current first-line therapies in patients with metastatic renal cell carcinoma (mRCC) with the potential for eliciting long-lasting remissions. With the expanding arsenal of ICB-based therapies, biomarkers of response are urgently needed to guide optimal therapeutic selection. We review the data behind ICB therapy in RCC, emerging biomarkers of response, and the evolving role of surgery in patients with mRCC.

1 |. INTRODUCTION

Renal cell carcinoma (RCC) is the eighth most common malignancy in the United States with an anticipated 73,750 new cases diagnosed in 2020 and 14,830 estimated deaths.¹ The 2016 World Health Organization classification defines 16 types of renal cell tumors but the most common subtypes include clear cell (~75%), papillary (~15%), chromophobe (5%), Xp11.2 translocation (~2%), collecting duct (~1%), medullary (~1%), and unclassified (5%–10%).² Surgical resection remains curative for the majority of patients with localized disease although between 20% and 40% of these patients will experience disease relapse.³ The overall risk of recurrence does not differ between patients undergoing partial nephrectomy (for tumors 4–7 cm in diameter) and those undergoing radical nephrectomy.⁴ Approximately one-third of patients will have metastatic disease at the time of diagnosis.⁵ The mainstay of treatment for unresectable or metastatic RCC is a systemic therapy. Initial attempts with hormonal therapy and chemotherapy were largely unsuccessful.⁶ Host defenses were long suspected to play a role in the RCC disease process based on observations including occasional spontaneous regression of disease, prolonged periods of disease stability without disease progression, and very late relapses after nephrectomy.^7,8 Cytokine therapy was the first successful systemic treatment for RCC, further demonstrating the immunogenic nature of RCC. A Cochrane meta-analysis reported a pooled response rate of 12.5% for interferon-α (IFNα) therapy from four randomized trials with a hazard ratio of 0.56 (95% confidence interval [CI], 0.40–0.77) for overall mortality at 1 year, compared with non-immunotherapy controls.⁹ Cytoreductive nephrectomy followed by IFNα therapy was also found to produce superior outcomes compared to IFNα alone with a 31% decrease in the overall risk of death (p = .002) according to a pooled analysis of two randomized trials.¹⁰ A pooled analysis of seven phase-2 clinical trials employing interleukin-2 (IL-2) therapy demonstrated an overall objective response rate of 15% (7% complete response rate) with a median response duration among responders of 54 months (range 3–107+ months).¹¹ Overall, these original immunotherapies were capable of producing remarkably durable responses in a subset of patients with metastatic renal cell carcinoma (mRCC), however, were also associated with significant toxicities often necessitating intensive care monitoring, especially with the use of high-dose IL-2. Cytokine therapies were largely replaced by antiangiogenic agents primarily targeting the vascular endothelial growth factor (VEGF) receptor and mammalian target of rapamycin inhibitors. One of the main limitations of these therapies is the ultimate development of resistance in the majority of patients.¹²

The recent development of immune checkpoint blockade (ICB) agents has signaled a revolution in the treatment of mRCC capable of producing long-lasting remissions that had previously only been seen in a small fraction of patients receiving high-dose IL-2. Immune checkpoints are regulatory mechanisms that inhibit full T-cell activation that normally function to protect the host from autoimmune responses. T cell activation and effector functions require two signals, the first of which is mediated by antigen presented via major histocompatibility complex (MHC) machinery followed by costimulatory and/or coinhibitory signals, the net effect of which determines the ultimate functional state of the T cell¹³ (Figure 1A). The ICB agents currently approved by the Food and Drugs Administration (FDA) for RCC include those that block coinhibitory molecules such as cytotoxic T-lymphocyte activating protein-4 (CTLA-4), programmed death-1 (PD-1), and programmed death ligand-1 (PD-L1) thus facilitating T cell effector function and antitumor response (Figure 1B). The therapeutic efficacy and safety of monoclonal antibodies (mAbs) targeting these checkpoints have been demonstrated in patients with mRCC as monotherapy, in combination with other ICB agents, and in combination with VEGF inhibitors. As a result of these trials, ICB-based therapies are now the standard of care for patients with advanced RCC. However, the field is currently limited by a lack of biomarkers to guide optimal therapy selection. Lastly, the role of surgery in the management of patients with advanced RCC has evolved over the past two decades yet is poorly defined in the context of patients receiving ICB therapy.

FIGURE 1.

T cell activation and effector functions in the context of MHC-mediated TCR stimulation are subject to several regulatory checkpoints. (A) In the peripheral lymphoid organs, cytotoxic T-lymphocyte-associated protein-4 (CTLA-4) can bind CD80/86 and block the costimulatory signal mediated by CD28-CD80/86 interactions thus inhibiting the priming or activation of T cells (bottom panel). Within the tumor, programmed death ligand-1 (PD-L1) expression on tumor cells can inhibit T cell effector functions through interactions with programmed death-1 (PD-1) (top panel). (B) Anti-CTLA-4 mAbs can facilitate the T-cell priming phase in lymphoid organs allowing increased CD28-CD80/86 interactions (bottom panel). Anti-PD-1/PD-L1 mAbs can block the inhibitory actions of PD-1-PD-L1 interactions in the tumor stroma facilitating T cell effector functions. APC, antigen-presenting cell; ERK, extracellular signal-regulated kinases; ITIM, immunoreceptor tyrosine-based inhibitory motif; ITSM, immunoreceptor tyrosine-based switch motif; MHC, major histocompatibility complex; PI3K, phosphoinositide 3- kinase; TCR, T-cell receptor; SHP, Src homology region; ZAP-70, zeta-chain-associated protein kinase 70. Created with Biorender.com

2 |. ICB CLINICAL TRIALS IN RCC

2.1 |. ICB monotherapy

The safety and efficacy of several ICB agents have been studied as monotherapies, in combination with other ICB agents, and most recently, in combination with tyrosine kinase inhibitors. A summary of the phase-3 trials utilizing ICB therapy in RCC is provided in Table 1. Ipilimumab, (anti-CTLA-4 mAb) was the first ICB agent tested in patients with mRCC.¹⁴ In this phase-2 trial (NCT00057889), modest overall response rates, determined by Response Evaluation Criteria in Solid Tumors (RECIST) criteria, of 5%, in the cohort receiving a 3 mg/kg loading dose followed by 1 mg/kg every 3 weeks, and 12.5%, in the cohort receiving 3 mg/kg for all doses, were seen. Toxicities were immune-mediated in nature with grade ≥ III adverse events (AEs) occurring in 43% of the high-dose group and, from most to least common, included enteritis/colitis, hypopituitarism, dermatitis, arthralgia, adrenal insufficiency. Interestingly, there was a 30% overall response rate among patients that experience immune-related toxicity and a 0% response rate among those without immune-related toxicity from the ipilimumab. Although not all studies have detected this correlation, a recent meta-analysis of trials utilizing ICB agents across a variety of solid malignancies found a significant association between patients experience immune-related adverse events and improved overall survival (OS) and progression-free survival (PFS).¹⁵ Though the initial response rates seen in the higher dose group were somewhat promising, the severe nature of the immune toxicities seen (14 grade-III, 2 grade-IV, and 1 grade-V) precluded the further study of ipilimumab as monotherapy in RCC.

TABLE 1.

Summary of phase‐3 trials evaluating ICB in RCC

Year	Trial	Treatment arms	No. of patients	Setting	PD-L1 assay	Cell components quantified	% PD-L1 positive (threshold,%)	ORR	Median PFS (months)	OS	Incidence of grade ≥3 adverse events
2015	CheckMate 025 (NCT01668784)	Nivolumab versus everolimus	821	Secnd line	28–8 pharm-Dx	TC	24 (≥1)	25% versus 5%	4.6 versus 4.4	Median: 25 versus 19.6	19% versus 37% Nivolumab: 33% fatigue, 14% nausea, 14% pruritis
2018	CheckMate 214 (NCT01472081)	Ipilimumab and nivolumab versus sunitinib	1096	First line	28–8 pharm-Dx	TC	24 (≥1)	42% versus 27%	11.6 versus 8.4	At 18 months: 75% versus 60%	46% versus 63% Ipilimumab/nivolumab: 10% elevated lipase, 4% fatigue, 4% diarrhea
2019	Immotion 151 (NCT02420821)	Atezolizumab and bevacizumab versus sunitinib	915	First line	Ventana SP142	IC	40 (≥1)	37% versus 33%	11.2 versus 8.4	Median: 33.6 versus 34.9	40% versus 54% Atezolizumab/bevacizumab: 14% HTN, 3% proteinuria, 2% diarrhea, 2% arthralgia
2019	KEYNOTE-426 (NCT02853331)	Pembrolizumab and axitinib versus sunitinib	1062	First line	22C3 pharm-Dx	TC+IC	60.5 (≥1)	59.3% versus 35.7%	15.1 versus 11.1	At 18 months: 82.3% versus 72.1%	75.8% versus 70.6% Pembrolizumab/axitinib: 22.1% HTN, 13.3% elevated ALT, 9% diarrhea, 7% elevated AST
2019	JAVELIN Renal 101 (NCT02684006)	Avelumab and axitinib versus sunitinib	886	First line	Ventana SP263	IC	63.2 (≥1)	51.4% versus 25.7%	13.8 versus 8.4	At 12 months: 85.7% versus 83.1%	71.2% versus 71.5% Avelumab/axitinib: 25.6% HTN, 6.7% diarrhea, 6% elevated ALT, 3.9% elevated AST

Abbreviations: IC, immune cells; ORR, objective response rate; OS, overall survival; PD-L1, programmed death ligand-1; PFS, progression-free survival; TC, tumor cells.

Results of the first phase-1 trial (CheckMate 003; NCT00730639) of nivolumab (anti-PD-1 mAb) monotherapy (1, 3, or 10 mg/kg every 2 weeks) were reported in 2012 and included patients with metastatic melanoma, castrate-resistant prostate cancer, colorectal cancer, and RCC that had received prior systemic therapy (non-ICB agents).¹⁶ For the patients with mRCC, the objective response rate (ORR) was 27% (9 of 33) and grade ≥ III toxicities were seen in 14% of patients. CheckMate 009 (NCT01358721), was a phase-1b biomarker study of 91 patients with metastatic ccRCC (74% had received prior systemic therapy) treated with 0.3, 2, or 10 mg/kg nivolumab every 3 weeks. Pre- and posttreatment tumor biopsies and blood samples were analyzed using immunohistochemical assay (IHC), RNA microarray, and chemokine multiplex immunoassay. Nivolumab therapy elicited an ORR of 15% in this population and significant increases in intratumoral counts of T helper cells (CD4⁺) and cytotoxic T cells (CD8⁺). Collectively, the increased intratumoral T cells increased TCR transcripts in tumors, decreased TCR transcripts in blood, and increased CXCL9 and CXCL10 serum levels and intratumoral transcripts with nivolumab treatment suggested anti-PD-1 therapy reversed T cell exhaustion and increased T cell recruitment to tumor sites via myeloid-derived chemokines. Subsequently, nivolumab was evaluated in a randomized dose-ranging phase-2 trial (CheckMate 010; NCT01354431) recruiting 168 patients with metastatic ccRCC who had received prior VEGF therapy.¹⁷ ORRs ranged from 20% to 22% across the 0.3, 2, and 10 mg/kg doses without any significant dose–response effects seen, as was also the case for PFS and AEs. As an exploratory endpoint, they observed higher response rates in patients with PD-L1 positive (≥5% tumor cells) tumors (ORR 31%) compared to patients with PD-L1 negative (<5%) tumors (ORR 18%). On the basis of these encouraging results, a randomized phase-3 trial (CheckMate 025; NCT01668784) comparing the efficacy of nivolumab (3 mg/kg every 2 weeks) to everolimus (10 mg daily) was conducted in 821 patients with metastatic ccRCC and prior treatment with VEGF therapy.¹⁸ Nivolumab therapy was associated with superiority and elicited a significantly improved median OS compared to everolimus (25 vs. 19.6 months, respectively (HR, 0.73; 98.5% CI, 0.57–0.93; p = .002). This survival advantage was notably associated with fewer severe (grade ≥ 3) AEs in patients treated with nivolumab (19%) than everolimus (37%) and higher quality of life (QOL) scores on the Functional Assessment of Cancer Therapy Kidney Symptom Index-Disease-Related Symptoms (FKSI-DRS) scoring algorithm. On the basis of these results, nivolumab was FDA-approved as a second-line treatment for patients with mRCC after TKI failure. Recently, updated results from this trial population with a minimum of 5 years of follow-up have established long-term favorable safety and efficacy for nivolumab monotherapy.¹⁹

The safety and efficacy of pembrolizumab (anti-PD-1 mAb) monotherapy in treatment-naïve patients with metastatic ccRCC and nonclear cell RCC (nccRCC) are also currently being evaluated in a phase-2 clinical trial (KEYNOTE-427; NCT02853344). Preliminary analysis of the cohort enrolling 110 patients with ccRCC revealed an ORR of 36.4% with a median follow-up of 12 months at the time of data cutoff.²⁰ The cohort of patients with nccRCC (n = 165), comprised of 72% papillary RCC, 13% chromophobe RCC, and 16% unclassified RCC histology, exhibited a collective ORR of 24.8% and subtype-specific ORRs of 34.6% for unclassified RCC, 25.4% for papillary RCC, and 9.5% for chromophobe RCC.²¹ The anti-PD-L1 mAb, atezolizumab, was found to have a favorable safety profile but modest efficacy in a phase 1a trial (NCT10375842)²² and a randomized phase-2 trial (IMmotion150; NCT01984242).²³ The IMmotion150 trial randomized 305 treatment-naïve patients to atezolizumab, atezolizumab+bevacizumab, or sunitinib and found an ORR of 25% for atezolizumab and 29% for sunitinib. Median PFS was also comparable between atezolizumab (6.1 months) and sunitinib (8.4 months) with an HR of 1.19 (95% CI, 0.82–1.71).²³

2.2 |. ICB combination therapy

Despite concerns over the autoimmune toxicities seen with ipilimumab monotherapy in RCC,¹⁴ the additive efficacy and maintained safety profile of combining nivolumab and ipilimumab in melanoma and lung cancer prompted further evaluation of this combination in RCC.^24,25 Hammers et al.²⁶ (CheckMate 016; NCT01472081) reported on the safety and efficacy of nivolumab and ipilimumab followed by nivolumab monotherapy from a phase-1 dose-escalation study enrolling 194 patients with metastatic ccRCC.²⁶ The two dosing regimens reported included nivolumab 3 mg/kg plus ipilimumab 1 mg/kg (N3I1) and nivolumab 1 mg/kg plus ipilimumab 3 mg/kg (N1I3). Both of the dosing arms reported very promising ORRs of 40.4%, however, there was a higher rate of grade ≥ 3 AEs in the N1I3 arm (61.7%) compared to the N3I1 arm (38.3%). Given the comparable efficacy and more favorable safety profile, the N3I1 dose was chosen for use in the CheckMate 214 trial (NCT02231749); a randomized phase-3 clinical trial comparing nivolumab plus ipilimumab to sunitinib in 1096 treatment-naïve patients with metastatic ccRCC.²⁷ A significant OS advantage was seen in intermediate- and poor-risk patients receiving nivolumab plus ipilimumab with 75% still alive at 18 months compared to 60% of patients receiving sunitinib with an HR of 0.63 (99.8% CI, 0.44–0.89; p < .001). The ORR observed with nivolumab plus ipilimumab was 42% versus 27% with sunitinib (p < .001) and complete responses were seen in 9% versus 1%, respectively. Grade ≥ 3 AEs were seen in 46% of patients in the ICB arm and 63% of patients in the sunitinib arm. Furthermore, the Functional Assessment of Cancer Therapy-Kidney Symposium Index (FKSI-19) revealed significantly higher QOL scores for patients in the ICB arm compared to their baseline scores and compared to those in the sunitinib arm. Within the favorable risk group of patients, initial results favored sunitinib over ICB in terms of ORR and PFS. However, these data need to be interpreted with some caution for several reasons: (1) Favorable risk patients constituted only 23% of the study population and thus were part of an exploratory analysis, (2) updated analysis with a minimum of 30 months of follow-up showed no difference in ORR, PFS, and OS between groups in favorable risk patients,²⁸ and (3) complete responses were seen in 8% of favorable risk patients in the ICB arm versus 4% in sunitinib arm with 90% of the ICB complete responders exhibiting durability at 32 months compared to 40% of the complete responders in the sunitinib group.²⁹ Thus, though the survival benefits of nivolumab plus ipilimumab were primarily driven by responses in intermediate-poor-risk patients, long-term disease control was also demonstrated in select favorable risk patients.

2.3 |. ICB and VEGF inhibitor therapy

Aside from the well-known antiangiogenic effects of VEGF inhibitors, a variety of immunomodulatory properties have been reported in the literature.³⁰ In a preclinical model of ovarian cancer, VEGF knock-down in tumor cells decreased myeloid-derived suppressor cell (MDSC) infiltration, which express VEGF receptors (VEGFR) 1 and 2, and increased CD8⁺ infiltration.³¹ In addition to MDSCs, VEFG receptors are also expressed on and modulate the function of dendritic cells,³² macrophages,³³ cytotoxic CD8⁺ cells,³⁴ and Tregs.³⁵ Collectively, these studies suggest that tumor VEGF production, aside from promoting tumor angiogenesis, exerts immunosuppressive effects on the tumor microenvironment and that blockade of VEGFR through use of small-molecule tyrosine kinase inhibitors (TKIs) or anti-VEGF mAb (i.e., bevacizumab) promotes activation of tumor immunity.

Initial attempts combining VEGF blockade with immunotherapy using the combination of INFα plus bevacizumab had met with success demonstrating improved PFS compared to IFNα alone.³⁶ Bevacizumab was also tested in combination with atezolizumab, an anti-PD-L1 mAb, against sunitinib in the phase-3 IMmotion 151 trial.³⁷ In the PD-L1 positive population, those receiving atezolizumab plus bevacizumab experienced prolonged PFS compared to those receiving sunitinib (11.2 vs. 7.7 months) with an HR of 0.74 (95% CI, 0.57–0.96; p = .0217). There was no difference in OS between groups at the time of data cutoff.

CheckMate 016 (NCT01472081) was the first phase-1 trial to combine ICB and TKI agents in patients with previously treated metastatic clear cell and nonclear cell forms of RCC.³⁸ Patients received nivolumab (2 mg/kg every 3 weeks) plus sunitinib (50 mg daily) or pazopanib (800 mg daily). Although the ORRs of 55% for nivolumab plus sunitinib and 45% for nivolumab plus pazopanib were very promising, there was a high rate of high-grade AEs (grade ≥ 3 82% and 70%, respectively) that precluded further study. Pembrolizumab combined with axitinib was better tolerated, with Atkins et al.’s³⁹ phase-1b study (NCT02853331) of treatment-naïve metastatic ccRCC patients reporting a 65% rate of grade ≥ 3 AEs, and an impressive 73% ORR and 8% CR rate. These encouraging results prompted KEYNOTE-426 (NCT02853331), a landmark phase-3 trial randomizing 1062 patients with treatment-naïve metastatic RCC to pembrolizumab (200 mg every 3 weeks) plus standard doses of axitinib or sunitinib.⁴⁰ The study met its primary endpoint demonstrating significantly improved OS and PFS in patients receiving pembrolizumab plus axitinib versus sunitinib with a 47% overall lower risk of death or progression. Notably, these benefits were observed across favorable-, intermediate-, and poor-risk groups. Updated results from this trial were recently presented with a median follow-up of 27 months showing combination therapy with pembrolizumab plus axitinib elicited durable responses that remained superior to sunitinib monotherapy.⁴¹ These results prompted an accelerated FDA approval for the use of pembrolizumab and axitinib as first-line therapy for patients with metastatic clear cell RCC (ccRCC). In the same issue of the New England Journal of Medicine, Motzer et al.⁴² reported the results of JAVELIN Renal 101, a randomized phase-3 trial of avelumab (10 mg/kg every 2 weeks) plus axitinib versus sunitinib as first-line therapy in 886 patients with metastatic ccRCC. Though the OS data were not mature at the time of reporting, the combination treatment did significantly prolong median PFS compared to sunitinib (13 vs. 8.4 months, respectively) with an HR of 0.61 (95% CI, 0.47–0.79; p < .001). ORRs were also higher in patients receiving avelumab plus axitinib at 55.2% versus 25.5% in those receiving sunitinib. Overall, the rates of high-grade AEs were equivalent between groups at 71.2% in the combination group and 71.5% in the sunitinib arm. Findings from phase-3 CheckMate 9ER trial comparing nivolumab plus cabozantinib to sunitinib in 651 treatment-naïve patients were recently reported in abstract form.⁴³ The combination regimen significantly improved PFS (HR 0.51; 95% CI, 0.41–0.64; p < .0001) and OS (HR 0.60; 95% CI 0.40–0.89; p = .001).

Combinations of ICB (i.e., nivolumab plus ipilimumab) and combination ICB and VEGF inhibitor (i.e., axitinib plus pembrolizumab) therapies represent the current first-line therapy in intermediate-poor and good risk patients with advanced/metastatic RCC, respectively.⁴⁴ However, the selection of therapy can be challenging due to lack of direct comparison trials and imperfect biomarkers of response.

3 |. PROGNOSTIC BIOMARKERS FOR ICB IN RCC

3.1 |. PD-1/PD-L1

Given the biologic rationale for ICB therapy, it would be expected that tumor cells (TC) highly expressing PD-L1 and/or tumors with a high proportion of PD-1 expressing infiltrating immune cells (IC) would exhibit greater responses to anti-PD-1/PD-L1 therapy. Several FDA-approved ICB therapies for other solid cancers have PD-L1 expression thresholds incorporated as a companion diagnostic. However, the data from ICB trials in RCC have yielded conflicting results, thus at this time the four FDA-approved ICB treatments for advanced RCC do not require PD-1/PD-L1 status.

With regard to nivolumab monotherapy, data from the phase-2 study, CheckMate 025, patients with 1% or more of TC expressing PD-L1 (24% of the study population) had overall poorer survival compared to those with PD-L1-negative tumors, but PD-L1 status was not predictive of response to nivolumab.¹⁸ A similar prevalence was observed in the phase-3 CheckMate 214 trial where 24% of the patient population were PD-L1 positive (≥1% PD-L1 positive TC).²⁷ Among patients with PD-L1 positive tumors, those receiving nivolumab plus ipilimumab experienced greater PFS compared to those receiving sunitinib however, longer ORR and OS were seen in patients across tumor PD-L1 status compared to sunitinib.

In contrast, the IMmotion 150 and 151 trials quantified PD-L1 expression on tumor-infiltrating ICs with a threshold of 1% or more used to define PD-L1 positivity.^23,37 With this method of quantification, 40% of the study population in the IMmotion 151 trial was PD-L1 positive. Patients with PD-L1 positive tumors experienced greater PFS benefit with combination therapy compared to sunitinib (HR 0.73; 95% CI 0.57–0.94), but those with PD-L1 negative tumors had PFS in response to combo therapy equivalent to sunitinib (HR 0.93; 95% CI 0.75–1.15).

The JAVELIN Renal 101 trial comparing avelumab plus axitinib to sunitinib reported PD-L1 expression as the percent of PD-L1 positive tumor-infiltrating ICs with a threshold of 1% or more being considered PD-L1 positive.⁴² With this definition, 63.2% of the study participants had PD-L1 positive tumors. Superior ORR and PFS were observed across tumor PD-L1 status in patients receiving avelumab plus axitinib compared to sunitinib. Although PD-L1 expression was quantified differently than JAVELIN Renal 101, similar findings were reported from the CheckMate-426 trial comparing pembrolizumab plus axitinib against sunitinib. Rini et al.⁴⁰ quantified expression as a combined positive score, which incorporates PD-L1 expression on TCs and ICs with a threshold score of 1% or more, considered PD-L1 positive. Overall, 60.5% of the study population had PD-L1 positive tumors using this definition. However, PD-L1 positive and negative patients both experienced significantly improved OS and PFS from combination therapy compared to sunitinib. Exploratory biomarker analysis of the phase-1b trial (NCT02133742)³⁹ cohort treated with pembrolizumab plus axitinib found that TC PD-L1 expression also did not predict ORR or PFS in patients receiving combination therapy.⁴⁵

There are at least several possible explanations for why PD-L1 status has limited predictive value in RCC. PD-L1 expression was found to be discordant between the primary tumor and metastatic sites as well within the primary tumor itself.⁴⁶ As shown above, the prevalence of PD-L1 positive tumors in RCC ICB trials ranged from 24% to 63.2% depending on whether TC, tumor-infiltrating IC, or both were incorporated into the metric (Table 1). Furthermore, it is possible that PD-L1 expression may have opposing prognostic value to anti-PD-1/PD-L1 mAb versus anti-CTLA mAb versus VEGF inhibitors and thus confound the ultimate clinical effect seen in combination regimens. A recent exploratory biomarker analysis of the JAVELIN RENAL 101 trial directly compared different methods and thresholds of PD-L1 expression.⁴⁷ Tumor IHC was quantified separately as the percentage of total infiltrating ICs and the percentage of tumor cells (TC) expressing PD-L1, then, several thresholds for PD-L1 positivity were analyzed. Higher PD-L1 expression using IC or TC was associated with decreased PFS in patients receiving sunitinib, however, no method of quantification or threshold for PD-L1 positivity was associated with response to avelumab plus axitinib. In summary, PD-L1 expression in the RCC tumor microenvironment appears to explain some but not the majority of the clinical benefit derived from ICB.

3.2 |. Tumor-infiltrating ICs

Among solid tumors, RCC is known to have one of the highest degrees of IC infiltration.⁴⁸ The positive association of denser CD8⁺ T cells in tumors and more favorable clinical outcomes has been shown in several malignancies including RCC.^49,50 A detailed characterization of the T-cell infiltrate within RCC tumors (tumor-infiltrating lymphocytes [TILs]) using multiparametric flow cytometry revealed three dominant immune profiles: “Immune-regulated,” composed of Tregs and poorly cytotoxic T cells expressing several inhibitory checkpoint molecules including PD-1, Tim-3, and Lag-3; “immune-activated,” characterized by oligoclonal cytotoxic T cells; and “immune-silent,” enriched in TILs that resembled lymphocytes found in adjacent normal renal parenchyma.⁵¹ Out of these three patterns, only the immune-regulated was associated with a high risk of recurrence following nephrectomy and aggressive histology. The exact role of tumor-infiltrating cytotoxic CD8⁺ cells is controversial. Some early studies of RCC tumor specimens found an association between increased T cell infiltration and worse OS⁵² and higher recurrence rates after nephrectomy.⁵³ However, Weiss et al.⁵⁴ reported tumors with greater CD8⁺ cell infiltrate were associated with longer OS following nephrectomy. A multivariate analysis of tumor stage, grade, tumor CD4⁺ count, and tumor CD8⁺ count revealed CD4⁺ but not CD8⁺ cell infiltrates were associated with cancer-specific survival following nephrectomy.⁵⁵ In the context of metastatic RCC, increased PD-L1 expression and intratumoral CD8⁺ counts were associated with shorter OS in patients receiving TKIs.⁵⁶ However, in the context of PD-1 inhibitor therapy, a tumor heavily infiltrated with CD8⁺ cells would conceptually be primed to respond better to ICB. A biomarker analysis of the phase-1b trial of pembrolizumab plus axitinib found a trend towards an association between higher CD8⁺ cell density and longer PFS, though it did not reach the threshold for significance.⁴⁵ Braun et al.⁵⁷ examined the association between CD8⁺ cell infiltration density and patterns in pretreatment biopsy specimens from patients enrolled in CheckMate 010 and CheckMate 025 before treatment with nivolumab or everolimus.⁵⁷ Tumors were categorized as (i) immune-infiltrated, those with a high-density T cell infiltrate in the tumor center, (ii) immune-excluded, which have a high density of T cells along the tumor margin with a relatively sparse infiltrate in the tumor center, and (iii) immune-desert, which have an overall low T cell infiltrate. They found no association between these CD8⁺ infiltration patterns and PFS or OS. Genomic analyses of this cohort found that the immune-infiltrated tumors (in theory expected to respond better to PD-1 blockade) were also enriched in for 9p21.3 deletions, which were associated with worse outcomes with nivolumab suggesting that, although a clear biomarker is yet to emerge, there may be a complex interplay of immunologic and genomic factors in determining responses to ICB.⁵⁷

3.3 |. Tumor mutational burden

A hallmark of neoplastic cells is the accumulation of a variety of somatic genomic alterations. The tumor mutation burden (TMB) is quantified as the number of nonsynonymous tumor exome mutations per megabase (Mb).⁵⁸ Across a variety of solid tumors, TMB has been positively associated with response rates to PD-1 blockade.⁵⁹ Recently, pembrolizumab was approved for use in adult and pediatric patients with unresectable or metastatic TMB-high (TMB-H; i.e., ≥10 mutations/Mb) solid tumors based on results from the basket trial KEYNOTE-158 (NCT02628067).^60,61 RCC has a relatively low TMB with a median of ~1.1 mutations/Mb, and yet is responsive to ICB.^{18,20,29,40,42,62} In an analysis of the pan-cancer TCGA data set, Turajlic et al.⁶³ reported that RCC harbors the highest proportion of small insertions and deletions (indels), which can create frameshifts producing novel open reading frames which in turn have a predilection for generating neoantigens. Neoantigens on tumor cells can in turn promote tumor-immune responses and the overall degree of tumor neoantigen load has been correlated with clinical response to PD-1 inhibitors in melanoma and lung cancer.^64–67 However, in patients with RCC, there was no association between TMB and clinical response to nivolumab,⁵⁷ atezolizumab,²³ or avelumab plus axitinib.⁴⁷ Neoantigen load also had no relationship with clinical responses to nivolumab⁵⁷ and atezolizumab.²³ Despite the success of TMB measures predicting responses to systemic therapies in other malignancies, it’s role as a solitary prognosticator in RCC appears very limited.

3.4 |. Microsatellite instability (MSI)

Scattered throughout the human genome are repetitive tracts of DNA with loci 10–60 base pairs in length containing multiple repeats of 1–5 base pairs that are termed microsatellites. These regions have an especially high susceptibility to mutations, which are normally repaired and maintained by the cellular mismatch repair (MMR) system.⁶⁸ Clinically, MMR-deficient (dMMR) tumors are classified using IHC to assess for loss of one or more of the four key MMR pathway proteins; melanocyte-stimulating hormone 2 and 6 (MSH2 and MSH6), MutL homolog 1 (MLH1), and postmeiotic segregation increased 2 protein (PMS2). A deficient DNA MMR pathway results in numerous somatic mutations especially within microsatellite loci, the accumulation of which is referred to as MSI. Polymerase chain reaction, IHC, and next-generation targeted sequencing can all be utilized to assess MSI will relatively high concordance.^69–72 In 2017, the FDA granted the first-ever tissue-agnostic approval to pembrolizumab for adult and pediatric patients with MSI-H or MMR-deficient (dMMR) solid tumors. According to NCCN guidelines, routine clinical testing for dMMR/MSI at this time is only indicated for colorectal and endometrial carcinomas. A large-scale whole-exome analysis of MSI prevalence across 39 cancer types reported a 1.47% rate of MSI-high tumors in 339 ccRCC samples.⁷³ However, a recent retrospective analysis of 229 patients with metastatic ccRCC identified a 19% rate of DNA damage repair gene alterations, which include MMR genes, that were associated with superior OS in patients treated with ICB agents but not those receiving TKIs.⁷⁴ The low incidence and prognostic utility of MSI status, currently provides little clinical relevance in the management of RCC.

3.5 |. Transcriptomic analysis

Next-generation sequencing techniques applied to the transcriptome of RCC using microarray or RNA-sequencing (RNAseq) has provided a high-throughput tool to explore prognostic gene signatures from multiple cell populations comprising the tumor microenvironment (TME), such as tumor angiogenesis,⁷⁵ effector T cells,⁷⁶ myeloid inflammation, and NK cells.⁷⁷ Information about the various populations of ICs within the TME can be inferred from bulk transcriptome data using computational methods of IC deconvolution employing regression-based approaches like CIBERSORT⁷⁸ and gene-signature based approaches such as GSEA.⁷⁹ In a biomarker analysis of the IMmotion 150 trial, McDermott and colleagues evaluated the ability of angiogenesis, effector T cell (T_eff), and myeloid cell-gene signatures derived from pretreatment tumor biopsy specimens to predict responses to atezolizumab, atezolizumab plus bevacizumab, and sunitinib.²³ Tumors with high-level angiogenesis gene transcripts (Angio^hi) were associated with greater ORR (46% vs. 9% in Angio^lo tumors) and PFS (HR 0.31; 95% CI, 0.18–0.55) in response to sunitinib. On the contrary, patients with tumors containing high-T_eff (T_eff^hi) gene signatures experienced superior ORR (49% vs. 16% in T_eff^lo) and PFS (HR 0.50; 95% CI, 0.30–0.86) in response to atezolizumab plus bevacizumab. High-myeloid signatures (Myeloid^hi) were specifically associated with reduced PFS in response to atezolizumab monotherapy. When T_eff^hi) tumors were stratified by myeloid signature scores, PFS was superior with atezolizumab plus bevacizumab compared to atezolizumab monotherapy in T_eff^hiMyeloid^hi tumors but equivalent PFS was seen in T_eff^hiMyeloid^lo tumors. These data suggest that the addition of anti-VEGF agents to anti-PD-L1 therapy may help overcome immunosuppressive effects of the innate immune system. In support of this, evidence from preclinical models suggests that anti-VEGF therapies can augment myeloid-derived suppressor cell (MDSC) and dendritic cell function and promote antitumor immunity.³⁰ Braun et al. found that T_eff and Myeloid gene signatures alone did not predict clinical response to nivolumab monotherapy.⁵⁷ In the biomarker analysis of the JAVELIN Renal 101 trial, neither T_eff, Myeloid, and combined T_eff^hiMyeloid^hi composite signatures did not reliably predict response to avelumab plus axitinib therapy.⁴⁷ However, Angio^hi tumors were again found to have superior PFS in response to sunitinib (HR 0.64; 95% CI, 0.48–0.85). Motzer et al.,⁴⁷ therefore, derived a novel gene signature comprised of 26 genes involving elements of both the adaptive and innate immune system, termed “JAVELIN Renal 101 Immuno signature” of which high expression levels predicted greater PFS and OS in response to avelumab plus axitinib but not sunitinib. Martini et al.⁴⁵ also derived a novel gene signature comprised of 12 genes involved in T-cell receptor signaling and T cell activation, differentiation, and proliferation that was able to predict clinical responses to pembrolizumab plus axitinib.⁴⁵

To date, angiogenesis gene signatures have consistently predicted response to TKI monotherapies such as pazopanib and sunitinib in the metastatic setting. Also, though the signature gene compositions used have varied across studies, those derived from either lymphoid and/or myeloid cellular elements appear to predict responses to combined ICB and anti-VEGF therapies. Though future prospective clinical trials will be needed, the use of transcriptome-based analyses to guide individual patient systemic therapy selection is a promising notion.

3.6 |. Gut microbiome

Recently, the gut microbiome had gained attention as an important factor in modulating responses to ICB in several malignancies including RCC.⁸⁰ Routy et al.⁸¹ found that antibiotic use within 2 months before or 1 month following initiation of anti-PD-1 therapy in patients with RCC, urothelial carcinoma, and non-small cell lung carcinoma was associated with significantly decreased PFS and OS.⁸¹ Using quantitative metagenomics to characterize the stool microbiota of these patients, they discovered stool-richness-predicted clinical response and more specifically that the presence of Akkermansia mucinophila was the most significant predictor of response to anti-PD1 therapy. This interesting finding was further confirmed in a preclinical model of RCC where fecal microbiota transplants (FMT), into gut microbiome-depleted mice, from patients responding to and resistant to anti-PD-1 therapy. After RENCA cell orthotopic transplantation, anti-PD-1/CTLA-4 therapy had no effect on mice receiving FMT from nonresponders, whereas those receiving FMT from responders experienced significant reductions in tumor size. These findings were subsequently validated in a separate cohort of 121 patients with RCC receiving anti-PD-1 therapy.⁸² The patients receiving antibiotics within 30 days of starting PD-1 inhibitors (13% of the cohort) had an increased risk of progressive disease (75% vs. 22%, p < .01), shorter PFS (median 1.9 vs. 7.4 months; HR 3.1, 95% CI 1.4–6.9; p < .01), and shorter OS (median 17.3 vs 30.6 months; HR 3.5, 95% CI 1.1–10.8; p = .03). More recently, Akkermansia mucinophila was again found to be enriched in a separate cohort of RCC patients responding to nivolumab.⁸³ Furthermore, using orthotopic RENCA cell allografts in mice as a preclinical model of RCC, oral supplementation with Akkermansia mucinophila was able to reverse the compromised responses to anti-PD-1/CTLA-4 therapy in gut microbiome-depleted mice.⁸³ Prospective clinical trials combining probiotic supplementation with PD-1 inhibitors in patients with metastatic RCC are underway.^84,85

4 |. ROLE OF SURGERY IN ADVANCED RCC

Cytoreductive nephrectomy is utilized in the metastatic setting to achieve local and systemic benefits. However, drawbacks of surgery include perioperative complications and delaying effective systemic therapy.⁸⁶ These drawbacks have become increasingly pertinent to the development of iteratively better systemic therapies. Two seminal randomized trials established the role for integrating upfront cytoreductive nephrectomy with cytokine-directed therapy, with improved OS of approximately five months^87,88 (median OS: 13.6 months with nephrectomy plus IFNα vs. 7.7 months for IFNα alone; p = .002¹⁰). The improved survival from cytoreductive nephrectomy was reflected in the Motzer prognostic model for risk-stratifying patients receiving IFNα, where the absence of a nephrectomy was a risk factor for adverse survival.⁸⁹

The introduction of targeted therapies necessitated the reevaluation of cytoreductive nephrectomy. CARMENA, an 8-year prospective trial, randomized 450 MSKCC-intermediate- and -poor-risk patients to receive either sunitinib with nephrectomy or sunitinib alone. The trial ended early due to under-accrual and reported that sunitinib alone was noninferior to cytoreductive nephrectomy and sunitinib. OS outcomes in CARMENA were comparable with the cytokine era, with a median survival of 18.4 months for sunitinib alone and 13.9 months for cytoreductive nephrectomy plus sunitinib. The survival outcomes were inferior to comparably treated patients described in larger databases.⁹⁰ This motivated an evaluation of whether the trial’s population reflected real-world cases,⁹¹ with limitations identified including operations performed at low-volume centers, exclusion of patients with low metastatic burden, and inclusion of a high percentage of poor-risk patients. Further, the trial only enrolled patients requiring immediate systemic treatment, thereby excluding patients that may have been able to be safely surveilled for a period of time after surgery.⁹²

The SURTIME randomized clinical trial compared integrating cytoreductive nephrectomy upfront or after sunitinib.⁹³ SURTIME also ended prematurely due to under-recruitment, with the trial reporting that compared with upfront nephrectomy, a deferred nephrectomy was a safe strategy with good OS. Large retrospective trials have also explored the sequence of systemic therapy and cytoreductive nephrectomy, reporting that the integrated strategy may yield the greatest benefits when surgery is performed as a consolidative treatment, rather than upfront.⁹⁴ Although retrospective studies that report a survival benefit from cytoreductive nephrectomy are criticized for selection bias, it is notable that only 9% and 20% of primary tumors in IMDC poor- and intermediate-risk patients respond to VEGF-targeted therapy and, therefore, cytoreductive nephrectomy can remove the bulkiest and resistant disease sites.⁹⁵

To date, no randomized trials have reported the role of cytoreductive nephrectomy in the IO era. Notably, ICB combinations that do not include VEGF inhibitors could be administered closer to surgery, as they portend less risk to wound healing. The utility of ICB agents in the neoadjuvant setting is currently an active area of investigation. These trial designs broadly fall into two categories including patients with mRCC responding to ICB that then undergo a deferred cytoreductive nephrectomy and patients with high-risk localized RCC that receive ICB therapy before nephrectomy. Aside from determining the effects of these interventions on OS and PFS, these trials may also yield insight into whether neoadjuvant ICB is able to render more tumors amenable to partial rather than radical nephrectomy. Retrospective studies have demonstrated progressive survival improvement for patients receiving cytoreductive nephrectomy across the cytokine, targeted, and IO eras.⁹⁶ In the largest retrospective study of ICB-treated patients to date, Singla et al.⁹⁷ highlighted that patients receiving cytoreductive nephrectomy plus immune checkpoint blockade had improved survival compared with ICB. An abstract of a prospective randomized single-institution study of nivolumab alone, or in combination with either bevacizumab or ipilimumab before deferred cytoreductive nephrectomy reported comparable safety and excellent PFS.⁹⁸ After 24 months of follow-up, median survival had not been reached for any of the study arms, however, preliminary data appeared to show excellent overall response rates and PFS for patients receiving nivolumab alone. The NORDIC-SUN trial (NCT03977571)⁹⁹ and PROBE trial (NCT04510597)¹⁰⁰ will assess the effect of deferred cytoreductive nephrectomy on OS in patients receiving ipilimumab and nivolumab. The ability of neoadjuvant ICB in preventing disease recurrence after nephrectomy is also being investigated in patients with high-risk localized RCC. The PROSPER trial (NCT02575222) is a phase-3 randomized trial currently recruiting patients with localized RCC to receive nivolumab before surgery followed by maintenance therapy versus surgery alone.¹⁰¹ Treatment of high-risk patients with localized RCC with the combination of avelumab and axitinib before surgical resection will be evaluated in the phase-2 NEOAVAX trial (NCT03341845).¹⁰² The rapid development of newer systemic therapies has outpaced trials exploring optimal integration of multimodal care, whereby the recent cytoreductive nephrectomy trials were reported after sunitinib ceased to be the standard of care. Therefore, the clinical trials that are currently accruing will be critical to inform the role of surgery for advanced RCC in the ICB era.

5 |. CONCLUSIONS

Modern immunotherapy in RCC utilizing checkpoint inhibitors has marked a new era in the treatment of advanced disease resulting in increased survival and the potential for long-term remissions reminiscent of those seen in a small fraction of patients receiving IL-1 during the cytokine era and with much less toxicity. Currently, combinations of ICB agents (nivolumab plus ipilimumab) and of ICB and VEGF inhibitors (pembrolizumab plus axitinib) are standard first-line therapy for patients with mRCC. Promising data is emerging from other combination therapies including pembrolizumab plus lenvatinib¹⁰³ and nivolumab plus cabozantinib.⁴³ However, as the armamentarium of therapeutics grows, biomarkers of response are urgently needed to guide clinicians in choosing the best possible treatment for a given patient. Recently, molecular profiling of the RCC tumor microenvironment with RNA sequencing revealed gene signatures associated with responses to ICB agents and VEGF inhibitors. However, further prospective studies are needed to refine and validate these signatures along with a means to apply them to routine clinical practice in a cost- and time-effective manner. Lastly, ICB therapy may redefine and reinvigorate the historically controversial role of surgery in the management of patients with mRCC. Ideally, future studies will develop biomarker-driven algorithms to guide optimal timing and combinations of medical and surgical care in RCC.

DATA AVAILABILITY STATEMENT

Data sharing is not applicable to this article as no new data were created or analyzed in this manuscript.