As only few patients derive benefit from immune checkpoint inhibitors (ICIs), it is crucial to identify reliable predictive biomarkers of response. One important pathway in regulating immune cell reactivity is L-arginine (ARG) metabolism. The present study investigated the impact of baseline plasma ARG levels on clinical outcomes of cancer patients treated with ICIs and found that elevated plasma ARG levels were associated with durable clinical benefits, while low levels were associated with poor outcomes.
Immune checkpoint inhibitors (ICIs) using anti-programmed cell death protein 1 (PD-1)/programmed death-ligand 1 (PD-L1) or anti-cytotoxic T-lymphocyte antigen 4 (CTLA-4) antibodies have yielded considerable clinical benefit in multiple cancer types. However, the benefit is observed in only a fraction of patients, and to date, there are no reliable predictors of response that can accurately identify patients who will benefit from this therapy. Therefore, it is crucial to identify reliable predictive biomarkers of response to ICIs. L-Arginine (ARG) is an amino acid serving as a building block for protein synthesis and precursor for multiple metabolites. ARG metabolism has been shown to regulate immune cell reactivity, being essential in T-cell activation. As a modulator of the antitumour immune response, ARG could then modulate the efficacy of ICIs. The present study aimed to investigate the impact of baseline plasma ARG levels on clinical outcomes of cancer patients treated with ICIs.
To investigate the predictive value of circulating ARG levels, the available serum metabolomics dataset from the CheckMate 025 trial was analysed. This study included 392 patients with advanced renal cell carcinoma and treated with nivolumab. Subsequently, the correlation between ARG levels and clinical ICI activity was assessed by analysing plasma samples obtained before treatment onset in two independent cohorts of patients with advanced cancer included in two institutional molecular profiling programs (BIP, N= 77; PREMIS, N= 296) and from patients in a phase I first-in-human study of budigalimab monotherapy (ABBV-181). Quantitative ARG assessment was performed using a validated ELISA kit. Additionally, the correlation between ARG levels and ICI efficacy in preclinical settings was evaluated using a syngeneic mouse model of colorectal cancer responsive to ICIs. ARG concentration was quantified in plasma collected one day before ICI administration. Finally, to correlate ARG levels with features of circulating immune cells, the authors characterised the profile of matched peripheral blood mononuclear cell (PBMCs) obtained before immunotherapy onset from 66 patients included in the BIP study using multiplexed marker panels and flow cytometry analysis.
First, the available serum metabolomics dataset from CheckMate 025 was analysed. ARG levels were higher in patients with longer survival (median overall survival (mOS) was 38.8 vs. 24.6 months in patients with high and low ARG levels (ARGHigh and (ARGLow), respectively). Subsequently, ARG levels were quantified using an immunoassay in plasma collected at treatment onset from 77 patients treated with ICIs and enrolled in the BIP study (Discovery cohort). ARGLow patients had worse clinical outcomes than ARGHigh patients, with a clinical benefit rate of 19.2% vs. 40%, respectively (p= 0.013). Median progression-free survival (mPFS) was 1.9 vs. 12.1 months in ARGLow and ARGHigh patients, while mOS was 5.7 vs. not reached (NR). To confirm the robustness of these results, the correlation between baseline plasma ARG levels and response to ICIs was evaluated in an independent validation cohort (PREMIS study, validation cohort 1). ARG levels were significantly higher in patients with better clinical outcomes than in those without durable benefits (p= 0.012). ARGHigh patients had a significantly higher overall response rate (ORR, 29.96% vs. 14.28%), longer mPFS (3.83 vs.1.87 months) and mOS (13.2 vs. 4.97 months) than ARGLow patients. Dynamic ARG level analysis under ICI treatment indicated that ARG tends to be consumed more in patients with worse outcomes (p= 0.029), whereas its levels remain stable in patients with durable clinical benefit (DCB) (p= 0.37). Using sera from patients enrolled in a phase I first-in-human study investigating the safety and preliminary efficacy of budigalimab (ABBV-181), a new monoclonal antibody targeting PD-1, the predictive value of circulating ARG levels was analysed. High ARG levels were associated with a marginally longer PFS (mPFs was 1.87 vs. 1.74 months). Moreover, using a syngeneic MC38 colon carcinoma tumour model, it was shown that mice with high baseline ARG levels displayed longer survival than those with lower concentrations. Besides, the rate of full tumour rejection upon anti-PD-1/PD-L1 antibody treatment was significantly higher in mice with elevated ARG levels (75%) than that in the ARGLow group (25%, p= 0.004). Finally, immunophenotyping of PBMCs showed that low ARG levels were significantly associated with increased programmed death-ligand 1 expression in several immune cell subsets from the myeloid lineage.
These results demonstrate that baseline ARG levels predict ICI response in patients with advanced solid tumours. Elevated plasma ARG levels are associated with durable clinical benefits in patients treated with ICIs, while low ARG levels are associated with poor outcomes. Therefore, plasma ARG quantification may represent an attractive biomarker to select patients likely to benefit from combination therapies with ARG inhibitors and ICIs.