Most patients given immune checkpoint inhibitors don't benefit, and the tissue-based biomarkers used to pick them — PD-L1, tumor mutational burden — are imperfect and hard to sample repeatedly. This study asked whether a single amino acid measured in blood could do better. L-arginine fuels T-cell activation, and across two institutional cohorts plus a first-in-human trial, patients with high baseline plasma arginine lived longer on checkpoint blockade, independent of standard prognostic factors. A mouse model and immune-cell profiling backed the link. Quantified with a simple ELISA, circulating arginine emerges as an accessible predictive biomarker — and a rationale for combining arginine-pathway drugs with immunotherapy.
This original article in Annals of Oncology, co-first-authored by Dr. Florent Peyraud and Dr. Jean-Philippe Guégan (both Explicyte) and led by Prof. Antoine Italiano, tested whether baseline circulating L-arginine predicts benefit from immune checkpoint inhibitors. The team drew on two institutional molecular-profiling programs — BIP at Institut Bergonié (discovery) and PREMIS at Gustave Roussy (validation) — plus re-analysis of the phase 3 CheckMate 025 trial and sera from a first-in-human study of the anti-PD-1 antibody budigalimab, complemented by a syngeneic MC38 mouse model and blood immune-cell profiling. The work was supported by Institut Bergonié and Explicyte Immuno-Oncology. Explicyte anchored the biomarker and profiling program: quantitative arginine measurement by validated ELISA, Ventana Discovery / Akoya multiplex immunohistochemistry with TLS and CD8 quantification, high-parameter flow-cytometry immunophenotyping, and the in vitro and in vivo functional work.
Rather than relying on tissue, Explicyte measured baseline circulating L-arginine using a validated ImmuSmol ELISA, applied to plasma from the BIP discovery cohort (77 patients) and PREMIS validation cohort (~296 patients), with serum used for CheckMate 025 (392 renal-cancer patients) and the budigalimab trial. A maximally-selected-rank-statistics cutoff of 42.3 µM separated arginine-high from arginine-low patients, and the team showed serum reads higher than plasma — so sample type must be standardized.
In the BIP discovery cohort, arginine-low patients had a lower clinical benefit rate (19.2% vs 40%) and much shorter survival (median PFS 1.9 vs 12.1 months, HR 2.43, P = 0.002; median OS 5.7 months vs not reached, HR 2.55, P = 0.007). PREMIS reproduced this — arginine-high patients had a higher objective response rate (30% vs 14%, P = 0.017), longer PFS (3.83 vs 1.87 months, P < 0.001) and OS (13.2 vs 4.97 months, P < 0.001) — and arginine remained independent on multivariate analysis. A rising arginine level on treatment also tracked with better outcomes.
In the PD-1/PD-L1-responsive MC38 colon model, Explicyte quantified plasma arginine one day before treatment. Mice with high baseline arginine (>168 µM) had markedly higher full tumor-rejection rates on checkpoint blockade than arginine-low mice (about 86% vs 24%, or 75% vs 25% by group, P = 0.004) and longer survival (P = 0.02), moving arginine from correlation toward a driver of response.
High-parameter flow cytometry of matched pre-treatment PBMCs from 66 BIP patients showed that low arginine didn’t change immune-cell abundance but was associated with higher PD-L1 expression across myeloid subsets — monocytic, BDCA3+, plasmacytoid and CD1c+ dendritic cells, NK cells, and myeloid-derived suppressor cells. In vitro, arginine concentration dose-dependently drove the nivolumab-enhanced, anti-CD3–induced interferon-γ response, confirming T cells need arginine to function.
The study turns a metabolic insight into a practical, blood-based predictive biomarker — and a patient-selection strategy for the arginine-pathway drugs now entering the clinic.
For oncologists and trialists, plasma arginine is an attractive alternative to tissue biomarkers — cheap, repeatable, quantified by a simple immunoassay, and independent of PD-L1 and tumor type. The strongest translational hook is patient selection: arginase inhibitors such as INCB001158 and other arginine-pathway agents are in clinical development, and this work argues those combination trials should enrich for patients with low baseline arginine rather than treating all comers. Because serum and plasma give different readings, any clinical implementation needs a standardized sample type.
