Nivolumab in the treatment of advanced renal cell carcinoma: clinical trial evidence and experience

Introduction

Renal cell carcinoma (RCC) accounts for 3% of all cancers with more than 140,000 deaths per year and 350,000 new cases diagnosed in 2013 worldwide [Capitanio and Montorsi, 2016]. RCC has long been recognized as an immunoresponsive tumor with rare spontaneous regressions observed [Freed et al. 1977]. The lack of significant efficacy observed with chemotherapy, together with the recognition of the frequent presence of several immunologic dysfunctions [Porta et al. 2007], led research to focus on immunomodulatory strategies to foster antitumor activity.

For two decades, clinical experience with interferon-alpha (IFN-α) and high-dose interleukin-2 (HD IL-2) has provided proof of principle that immunotherapy can produce durable responses in a small percentage of patients with RCC [McDermott, 2007]. However, the toxicity and the limited efficacy of HD IL-2 have limited its clinical use.

More recently, several new agents targeting the vascular endothelial growth factor (VEGF) (sunitinib, sorafenib, pazopanib, bevacizumab and axitinib) and mammalian target of rapamycin (mTOR) pathways (temsirolimus and everolimus) have been approved for the treatment of metastatic RCC (mRCC) [Escudier et al. 2014]. While the response rates (RRs) with these agents appear higher than those with IL-2, complete responses remain very rare, with large part of the data showing significant improvement in disease control and progression-free survival (PFS), but not in overall survival (OS) [Albiges et al. 2012; Procopio et al. 2013]. Furthermore, resistance to these agents invariably develops, often within the first year of therapy [Motzer et al. 2007; Porta et al. 2012].

In recent years, a renewed interest for immunotherapy in RCC has been observed, because of significant improvement in the understanding of mechanisms of escape to antitumor response. RCC usually presents a prominent immune cell infiltrate that consists of T cells, natural killer (NK) cells, dendritic cells (DCs) and macrophages. Despite being strongly infiltrated, immune dysfunction promotes RCC tumor growth and evasion. The tumor-induced changes in DC differentiation and the induction of anergy-associated genes in T cells can partially explain the impaired antitumor response [Noessner et al. 2012]. The inhibitory mechanisms that mediate the interaction between an evolving tumor and the host immune response provide one explanation for why immunotherapies frequently fail to produce clinically relevant responses. A critical regulator of tumor-induced immune suppression is the programmed death-1 (PD-1) pathway [Keir et al. 2008].

PD-1 (CD279), a 55-kd type I transmembrane protein belonging to the CD28/CTLA-4 family, consists of a N-terminal immunoglobulin (Ig) superfamily domain, a transmembrane domain and a cytoplasmatic tail. Signaling downstream of PD-1 is mediated by the presence of an immunoreceptor tyrosine-based switch motif (ITSM) and an immunoreceptor tyrosine-based inhibitory motif (ITIM) that recruit anti-SRC homology phosphatase (SHP)-1 and SHP-2 protein tyrosine phosphatases to modulate inhibitory signaling [Nirschl et al. 2013; MacFarlane et al. 2014]. PD-1 is inducibly expressed on CD4+ and CD8+ T cells, NK cells, B cells and monocytes within 24 h from their immunological activation [Keir et al. 2008]. Its expression is increased by many cytokines, such as IL-2, IL-7, IL-15 and IL-21. Patients whose tumors contain PD-1(+) tumor-infiltrating lymphocytes (TILs) have more aggressive disease compared with patients with PD-1(-) lymphocytes [Thompson et al. 2007].

Two PD-1 ligands (PD-L1, also called B7-H1, and PD-L2, also known as B7-DC), have been described. While PD-L1 is expressed on several cells, including resting T cells, B cells, macrophages, DC cells, vascular endothelial cells, and pancreatic islet cells, PD-L2 is expressed only on macrophages and DCs [Keir et al. 2007]. PD-L1 and PD-L2 inhibit T-cell proliferation and adhesion, as well as cytokine production. PD-L1 modulates T-cell function in peripheral tissues, whereas PD-L2 regulates immune T-cell activation in lymphoid organs [Ghiotto et al. 2010].

Thompson and colleagues demonstrated that two-thirds of clear cell RCCs had PD-L1 expression and those with high expression had worse cancer-specific survival [Thompson et al. 2004]. They also investigated TILs in primary RCC and observed that TILs also express PD-L1, and high expression on either tumors or TILs was correlated with advanced stage, worse prognosis and rapid metastatic progression [Thompson et al. 2005, 2007]. Further studies and preclinical tumor models have demonstrated that blockade of the PD1/PD-L1 pathway by monoclonal antibodies can improve the antitumor immune response and result in tumor rejection, providing therapy for multiple solid tumors, including RCC [Iwai et al. 2002; Wang et al. 2014; Berman et al. 2015].

This review discusses the clinical efficacy and tolerability of nivolumab in advanced RCC.

Pharmacodynamic and pharmacokinetic properties of nivolumab

Nivolumab (BMS-936558) is a fully human, IgG4 monoclonal antibody that binds to the PD-1 receptor and, by preventing its interaction with its ligands PD-L1 and PD-L2, it disrupts negative signaling to restore T-cell antitumor function [Wang et al. 2014; Berman et al. 2015]. Nivolumab has high affinity and specificity for the PD-1 receptor thus, even if serum concentrations of nivolumab are undetectable, sufficient nivolumab should be able to maintain plateau PD-1 receptor occupancy. Nivolumab may continue to occupy PD-1 receptors for up to 3 months after treatment [Brahmer et al. 2010; Brahmer, 2013].

The pharmacokinetics (PK) of nivolumab is linear in the dose range of 0.1–10 mg/kg. The geometric mean clearance (CL), terminal half-life, and average exposure at steady state at 3 mg/kg every 2 weeks of nivolumab were 9.5 ml/h, 26.7 days, and 75.3 µg/ml, respectively, based on a population PK analysis.

Nivolumab is degraded into small peptides and amino acids via catabolic pathways in the same manner as endogenous IgG. The dosage does not need to be adjusted in patients with mild hepatic disorder or mild/moderate renal impairment. Data from patients with moderate/severe hepatic and severe renal dysfunction are too limited to draw conclusions on this population [European Medicines Agency, 2015].

Therapeutic efficacy of nivolumab

The efficacy of nivolumab in humans was first shown in a phase I, dose-escalation safety trial, in which 39 patients with advanced solid tumors (one single patient with mRCC) were treated with escalating doses of nivolumab [Brahmer et al. 2010]. The patient with mRCC who had received three doses of nivolumab at 10 mg/kg intravenously in 5 months obtained a partial response. He continued to have tumor regression after stopping treatment and reached a complete response after more than 4 years from the last dose of nivolumab [Lipson et al. 2013].

The preliminary evidence of clinical activity along with a favorable safety profile led to the design of a multi-dose trial with nivolumab in 296 patients, confirming its antitumor efficacy in melanoma, RCC and non-small cell lung carcinoma. In this phase Ib study, 34 patients with RCC were treated with nivolumab 1 mg/kg (n = 17) or 10 mg/kg (n = 16). The objective response rate (ORR) was 29.4% while an additional 10% of patients showed prolonged stable disease (SD) and median duration of response was 12.9 months for both doses. Furthermore, median PFS was 7.3 months and 3-year OS was 52% [Topalian et al. 2012].

Recently, the results of a randomized, blinded, dose-ranging phase II trial of nivolumab in 168 mRCC patients were presented. The patients had received at least one VEGF tyrosine-kinase inhibitor (TKI), but no more than three previous systemic therapies, and they were randomized to receive nivolumab at 0.3, 2 or 10 mg/kg until disease progression or intolerable toxicity [Motzer et al. 2015a].

Median PFS was 2.7, 4.0 and 4.2 months for the 0.3, 2 and 10 mg/kg groups respectively. ORR was 20%, 22% and 20% among the three groups of patients, while median OS was 18.2, 25.5 and 24.7 months, respectively. More than 50% of responders with all doses had objective responses lasting more than 12–20 months. Furthermore, Choueiri and colleagues showed that responses to nivolumab were more common in patients with higher PD-L1 tumor expression (⩾5%), but were also seen in patients with PD-L1 (-) RCC [Choueiri et al. 2014].

Based on these promising results, an open-label, phase III study (The CheckMate 025) evaluating nivolumab versus everolimus was designed: in this trial 821 patients with advanced or metastatic clear cell RCC, already treated with antiangiogenic therapy, (no more than three previous treatments) were randomized, in a 1:1 ratio, to receive nivolumab 3 mg/kg every 2 weeks versus everolimus 10 mg/day [Motzer et al. 2015b]. This study has met its primary endpoint: nivolumab has shown an improvement in OS in comparison with everolimus (25 months versus 19.6 months, hazard ratio (HR) 0.73). Nivolumab has also shown an increase in objective RR (25 versus 5%), without differences in terms of PFS (4.6 versus 4.4 months, HR 0.88) (Table 1). Investigators performed an ad hoc sensitivity analysis of PFS in patients who did not progress or die at 6 months (145 patients in the nivolumab arm and 129 patients in the everolimus arm) and showed that patients treated with nivolumab had a PFS of 15.6 months (95% CI 11.8–19.6) versus 11.7 months (95% CI 10.9–14.7) in the everolimus group (HR 0.64; 95% CI 0.47–0.88). Nivolumab has shown clinical benefit independently from PD-L1 expression. The subgroup analysis presented at the 2016 Genitourinary Cancers Symposium has confirmed the advantage of nivolumab irrespective of Memorial Sloan Kettering Cancer Center (MSKCC) prognostic score, number of previous antiangiogenic therapies, sex and region, with the exception of very elderly patients (⩾75 years old). This analysis has also demonstrated that in patients who received prior sunitinib (63%), median OS was 23.6 months with nivolumab versus 19.8 months with everolimus. Patients receiving prior pazopanib (32%), median OS was not estimable in the nivolumab arm while was 17.6 months in the everolimus arm. For the 10% of patients who had prior IL-2, median OS was not estimable for nivolumab while was 17.2 months for everolimus [Motzer et al. 2016]. This study has confirmed that nivolumab is associated with a longer survival in patients with mRCC treated with two or more previous lines of therapy in comparison with everolimus, thus leading to the approval of nivolumab by the US Food and Drug Administration (FDA).

OPEN IN VIEWER

Table 1.

Single agent anti-PD-1 (nivolumab) studies in RCC.

Phase	Total number of patients (with RCC)	Regimen	Results	Reference
I	39 (1)	nivo 10 mg/kg	PR after 3 doses CR after 4 years from the last dose	[Brahmer et al. 2010] and [Lipson et al. 2013]
Ib	296 (34)	nivo 1 mg/kg or 10 “mg/kg q2w	ORR 29.4%, SD 10% mPFS 7.3 months 3-year OS 52%	Topalian et al. [2012]
II	168 (168)	nivo 0.3 mg/kg, 2 mg/kg or 10 mg/kg q3w	ORR 20% versus 22% versus 20% mPFS: 2.7 versus 4.0 versus 4.2 months mOS 18.5 versus 25.5 versus 24.8 months	[Motzer et al. 2015a]
III	821 (821)	nivo 3 mg/kg q2w versus Eve 10 mg/die	ORR 25% versus 5% mPFS 4.6 versus 4.4 months mOS 25 versus 19.6 months	[Motzer et al. 2015b]

CR, complete response; eve, everolimus; mOS, median overall survival; mPFS, median progression-free survival; nivo, nivolumab; ORR, objective response rate; PD-1, programmed death-1; PR, partial response; q2w, every 2 weeks; q3w, every 3 weeks; RCC, renal cell carcinoma; SD, standard deviation.

Nivolumab has shown promising results also in combination phase I trials with other agents, including sunitinib and pazopanib [Amin et al. 2014]. The rationale for combining immune checkpoints inhibitors with TKIs is that these agents can modulate the immune environment by decreasing the number of circulating regulatory T cells (T regs) and myeloid-derived suppressor cells [Ko et al. 2009, 2010]. In this study, nivolumab was administered at a starting dose of 2 mg/kg, which was increased to 5 mg/kg. While the pazopanib combination arm was closed because of dose-limiting liver toxicity, the sunitinib combination arm was dose-escalated to a higher nivolumab dose and enrolled treatment-naïve patients. Overall RR was 52% in the sunitinib arm and 45% in the pazopanib arm. At 6 weeks from the beginning of the study, partial responses were observed in 41% of patients in the sunitinib arm and 56% in the pazopanib arm. PFS at 24 weeks was 78% versus 55% in the two arms, respectively. This study showed efficacy and a manageable safety profile of nivolumab in combination with the two drugs which are the standard of care in a first-line setting and the RRs observed are the best reported in trials of patients with mRCC, even if we have to take into account that the number of treatment-naïve patients was higher in this trial than in the previous ones.

Another interesting phase I trial in RCC patients (CheckMate 016) evaluated the association of nivolumab plus ipilimumab at different doses in 44 mRCC patients with favorable/intermediate MSKCC score [Hammers et al. 2014]. ORR was 48% in the group of patients that received nivolumab 1 mg/kg + ipilimumab 3 mg/kg and 43% in those who received nivolumab 3 mg/kg + ipilimumab 1 mg/kg, with SD reported in 35% and 24% of patients, respectively.

In consideration of the acceptable safety profile and the encouraging efficacy, a phase III trial is evaluating this combination in a first-line setting versus sunitinib with OS as the primary endpoint (CheckMate 214) [ClinicalTrials.gov identifier: NCT02231749] (Table 2).

OPEN IN VIEWER

Table 2.

Nivolumab in combination with angiogenesis inhibitors or other immune checkpoint inhibitors in RCC.

Phase	Total number of patients	Regimen	Results	Reference
I	53	nivo 2 mg/kg with planned escalation to 5 mg/kg q3w plus sun 50 mg daily for 4 weeks then 2 weeks off or pazo 800 mg daily	ORR 52% in sun arm and 45% in pazo arm PFS at 24 weeks 78% and 55%, respectively	[Amin et al. 2014]
I	44	nivo 3 mg/kg + Ipi 1 mg/kg or nivo 1 mg/kg + Ipi 3 mg/kg q3w for 4 doses then nivo 3 mg/kg q2w	ORR 48% in nivo 1 mg/kg + Ipi 3 mg/kg and 43% in nivo 3 mg/kg + Ipi 1 mg/kg SD 35% and 24%, respectively	[Hammers et al. 2014]
III	1070 (estimated enrollment)	nivo 3 mg/kg + Ipi 1 mg/kg q3w for 4 doses then nivo 3 mg/kg q2w or sun 50 mg daily for 4 weeks then 2 weeks off	Ongoing	ClinicalTrials.gov identifier: NCT02231749

Ipi, ipilimumab; nivo, nivolumab; ORR, objective response rate; PD-1, programmed death-1; PFS, progression-free survival; q2w, every 2 weeks; q3w, every 3 weeks; RCC, renal cell carcinoma; SD, standard deviation.

Safety

Nivolumab showed, in most clinical trials, a manageable safety profile in patients with advanced RCC [Brahmer et al. 2010; Topalian et al. 2012; Motzer et al. 2015a, 2015b; Procopio et al. 2016] and was better tolerated than everolimus in the CheckMate 025 trial [Motzer et al. 2015b]. In this phase III study, treatment-related adverse events (AEs) of any grade were reported in 79% of patients treated with nivolumab and 88% of everolimus patients, while treatment-related grade 3 or 4 AEs were reported in 19% and 37% of patients, respectively. The most common AEs of any grade were fatigue (33%), nausea (14%), and pruritus (14%) among patients who received nivolumab, and fatigue (34%), stomatitis (29%), and anemia (24%) among patients who received everolimus. Other AEs were diarrhea (12% for nivolumab versus 1% for everolimus), rash (10% versus 1%), pneumonitis (4% versus 3%). The most common grade 3 or 4 AEs were fatigue (2%) with nivolumab and anemia (8%) with everolimus. No deaths from study-drug toxic effects were reported in the nivolumab group, and two deaths in the everolimus group (one from septic shock and one from acute bowel ischemia). Treatment-related AEs led to treatment discontinuation in 8% of patients treated with nivolumab and in 13% treated with everolimus.

Nivolumab has been associated with immune-mediated adverse reactions, including immune-mediated pneumonitis, colitis, hepatitis, nephritis and renal failure, hypothyroidism and hyperthyroidism [Motzer et al. 2015a; European Medicines Agency, 2015]. Nivolumab has shown a manageable safety profile also in combination strategies. In the phase I study that evaluated the association of nivolumab with a VEGF-TKI [Amin et al. 2014], grade 3–4-related AEs were observed in 82% of patients treated with nivolumab and sunitinib. The most common treatment-related grade 3–4 AEs were elevated alanine aminotransferase (ALT) and hypertension (18% each), hyponatremia (15%) and lymphopenia (15%). Grade 3 pneumonitis occurred in one patient and 30% of patients discontinued treatment due to toxicity. The combination of nivolumab and pazopanib showed grade 3–4 AEs in 70% of patients and the most commonly reported AEs were elevated ALT and aspartate aminotransferase (AST, 20%), diarrhea (20%) and fatigue (15%). Unfortunately, in this arm dose-limiting toxicities leading to closure of the arm included: elevated ALT/AST and fatigue. When associated with ipilimumab [Hammers et al. 2014], grade 3–4 AEs occurred in 43% of patients, mostly reporting elevated lipase and ALT (16% and 11% respectively), diarrhea (9%), colitis (5%) and elevated amylase (5%). No grade 3–4 pneumonitis was seen. Treatment discontinuation was necessary in 16% of patients.

Five years forward view

In summary, the mRCC treatment landscape, still dominated by targeted therapies, could be significantly changed by the approval of novel immunotherapies such as nivolumab. Cancer immune surveillance is considered to be a fundamental host protection process to maintain cellular homeostasis and to inhibit carcinogenesis. Cancer immunoediting includes multiple mechanisms of escape from the immune system able to operate a selective pressure on the tumor microenvironment, leading to malignant progression. The encouraging results obtained by nivolumab, which has shown good tolerability and durable clinical benefit, have shed new light on treatment options for patients with advanced RCC. However, even if the new immunological agents appear to have the potential to reduce the tumor-induced immunosuppression and improve the outcome of mRCC patients, several questions remain to be answered. Further efforts are needed in order to understand the most effective setting for the use of nivolumab and the best drug sequencing. In most clinical trials, nivolumab has been investigated in the second to the fourth line of therapy: the rationale for a sequence of VEGF directed therapy followed by nivolumab is provided by the mechanism of resistance mediated by immune escape. Unfortunately, we still know little about its activity in first-line treatment, but encouraging results have been obtained in the phase I trial investigating the combination of nivolumab with sunitinib in treatment-naïve patients. The results from the ongoing phase III trial (Checkmate 214) which is assessing the combination of nivolumab and ipilimumab, will hopefully clarify its efficacy and safety in this setting. Other immune-checkpoint inhibitors are currently in development in association with anti-VEGFR in treatment-naïve patients with advanced RCC. Some phase I trials are exploring the combination of axitinib plus pembrolizumab [Atkins et al. 2015] or avelumab [ClinicalTrials.gov identifier: NCT02493751], respectively and the association of pembrolizumab and pazopanib [ClinicalTrials.gov identifier: NCT02014636]. Another clinical trial evaluating atezolizumab after nephrectomy in high risk patients will be available as adjuvant therapy, thus, the role of nivolumab might change in the light of the results of these studies. An unexplored field is represented by the use of nivolumab as adjuvant and neoadjuvant strategies. The use of this agent in the adjuvant setting raises important questions, for instance if primary tumor antigen is necessary to recruit and amplify T-cell activation with PD-1 inhibitors. Some evidence suggests that resection of renal primary tumors may result in a decrease of PD-1 and PDL-1 expression, reducing the potential target for this class of agents [MacFarlane et al. 2014; Jilaveanu et al. 2014]. Selection of patients who might benefit from checkpoint blockade is another open issue: considering the results obtained by the use of IFN-α and IL-2 [Flanigan et al. 2001], patients with low tumor burden could represent the best candidates to receive this approach. Moreover, the identification of possible predictive biomarkers of response might guide treatment choice in every single patient in order to avoid toxicity, maximize outcome and reduce costs. The expression of PD-L1 on cancer cells seems to be the most important feature that correlates with worse outcome and response to PD-1 pathway inhibitors [Taube et al. 2014; Choueiri et al. 2014, 2015] although conflicting results compromise the prognostic and predictive role of tumor PD-L1 expression [Fay et al. 2014]. An interesting association between PD-L1 expression on tumor-infiltrating immune cells and response to treatment has been recently shown. For example, elevated levels of pre-existing CD8+ T cells at the invasive tumor margin, along with the presence of immune cells expressing PD-1 and PD-L1, seem to promote PD-1 blockade-mediated tumor regression. The amount of CD8+ T cells suggests that the immune system has already recognized the tumor, and therefore infiltrating immune cells induce higher sensibility to checkpoint inhibitors [Tumeh et al. 2014].

A significant challenge remains the identification of a molecular profiling in order to predict the response or the resistance to immunotherapy. RCC is a genetically unstable disease: the DNA mismatch repair system (MLH1 and MSH2), PBMR1 and BAP1 are some of the genes involved in renal cancer pathogenesis and their loss can lead to an increased mutational load that correlates with sensitivity to checkpoint blockade. It seems that inactivating mutations of PBMR1 and BAP1 are frequently associated with increased PDL1 expression on RCC cells and concomitant PD-1 expression on TILs [Joseph et al. 2014; Piva et al. 2016]. This suggests that these patients may be good candidates for this immunological approach, but it would be necessary to experimentally validate this correlation and, most importantly, clinical evidence would have to be provided to support this hypothesis.

Despite considerable advances, the identification of patients whose cancer is sufficiently immunogenic and whose immune system is sufficiently reactive to inhibit tumor growth is still warranted.

Conclusion

Nivolumab has shown encouraging objective RRs and an OS improvement in mRCC patients, after failure of previous antiangiogenic therapies, with an acceptable safety profile. Combinations with other agents, such as angiogenesis inhibitors and other immune checkpoint inhibitors have demonstrated higher RRs but at the cost of increased toxicity. The selection of patients most likely to benefit from anti-PD-1 therapy is a key point for maximizing outcomes in this setting.