Aspartate-Driven Immune Suppression in AAG Metabolic Subtype of Gastric Cancer: A Testable Hypothesis for Combination Immunotherapy
Hypothesis
In gastric cancers exhibiting the alanine-aspartate-glutamate (AAG) metabolic subtype, elevated extracellular aspartate acts as an immunosuppressive signal that binds NMDA‑receptor‑like channels on tumor‑associated macrophages, triggering calcium‑dependent NFAT activation and up‑regulation of PD‑L1. Pharmacologic blockade of aspartate uptake (e.g., via SLC1A3/GLAST inhibitors) will reduce macrophage PD‑L1 expression and enhance the efficacy of anti‑PD‑1 therapy, improving progression‑free survival specifically in AAG‑high patients.
Mechanistic Rationale
Multi‑omics integration has shown that the AAG subgroup derives survival advantage from altered amino‑acid flux, yet the functional consequence of aspartate secretion remains unexplored [2]. Aspartate is a known excitatory neurotransmitter that can activate NMDA‑type receptors on non‑neuronal cells, leading to intracellular calcium spikes and downstream transcriptional programs [1]. Spatial proteomics of colorectal tumors has demonstrated the capacity to map metabolite‑immune interactions at single‑cell resolution, a technique readily adaptable to gastric tissue [3]. We hypothesize that aspartate released by AAG‑high cancer cells engages NMDA‑receptor‑like complexes on CD68⁺ macrophages, causing Ca²⁺ influx, calmodulin‑dependent kinase activation, and NFAT translocation to the nucleus, where it drives CD274 (PD‑L1) transcription. This creates a localized immune‑checkpoint‑rich niche that blunts T‑cell cytotoxicity despite checkpoint blockade.
Experimental Design
- Cohort assembly – Prospectively enroll 200 treatment‑naïve gastric adenocarcinoma patients undergoing curative resection. Perform multi‑omics (RNA‑seq, proteomics, untargeted metabolomics) on pretreatment biopsies to assign AAG/GG metabolic subtype per established criteria [2].
- Baseline spatial mapping – Apply multiplexed ion beam imaging (MIBI) or CODEX on resection specimens to quantify extracellular aspartate (using antibody‑based aspartate sensors), SLC1A3 expression, CD68⁺ macrophage density, and PD‑L1 colocalization at the tumor‑stroma interface [3].
- Intervention arm – Patients with AAG‑high tumors receive standard neoadjuvant chemo‑immunotherapy (fluoropyrimidine/platinum + pembrolizumab) plus oral SLC1A3 inhibitor (e.g., UCPH‑101) at a dose determined from prior pharmacokinetic studies. Control arm receives chemo‑immunotherapy plus placebo.
- Endpoints – Primary endpoint is investigator‑assessed progression‑free survival at 12 months. Secondary objectives include objective response rate, change in macrophage PD‑L1 (by spatial proteomics), and circulating aspartate levels.
- Statistical plan – Power calculation assumes a hazard ratio of 0.6 for PFS in the combination arm among AAG‑high patients (α = 0.05, 80 % power). Interaction term between treatment and metabolic subtype will be tested in a Cox model.
Expected Outcomes and Falsifiability
If aspartate‑mediated macrophage activation drives immune evasion, the AAG‑high subgroup receiving aspartate‑uptake inhibition will show a significant increase in PFS compared with control, accompanied by reduced macrophage PD‑L1 signal and lower aspartate levels in the tumor microenvironment. Conversely, if aspartate blockade fails to improve PFS or does not alter macrophage PD‑L1 despite adequate target inhibition, the hypothesis is falsified, suggesting that aspartate’s role is either redundant or operates through a non‑immune mechanism. This trial design directly tests a mechanistic link derived from multi‑omics, spatial proteomics, and AI‑guided subtype stratification, addressing the translational gap highlighted by the lack of prospective validation of metabolic subtypes for immunotherapy response [1].