Aspirin Response in PIK3CA-Mutated Colorectal Cancer

Aspirin has been investigated as an additional treatment after colorectal cancer surgery, but average effects can conceal differences between tumor subgroups. Two independent studies now examine that variation at different levels: a preplanned analysis of participants in a randomized trial and a laboratory study of PIK3CA-mutated colorectal cancer cells. Together they sharpen a biomarker hypothesis without resolving it.

Evidence

The ASCOLT trial had compared adjuvant aspirin with placebo in 1,587 people with colorectal cancer and found no significant improvement in disease-free survival overall. Its newly reported translational study examined whether tumor PIK3CA mutations, PTEN mutations, or COX-2 overexpression separated participants more likely to benefit. Of 778 participants who started study medication, tumors from 289 underwent targeted next-generation sequencing for PIK3CA and PTEN, while another 108 were assessed for PIK3CA exon 9 or 20 mutations by Sanger sequencing. COX-2 expression was measured in 450 tumors.

Among 397 participants with assessable PIK3CA status, 69 had a mutation in any exon. After five years, that subgroup had eight disease-free-survival events in each treatment arm, corresponding to a hazard ratio of 0.93 for aspirin versus placebo, with a wide 95% confidence interval of 0.35 to 2.47. The exon 9-or-20 subgroup included 45 participants and had seven events with placebo versus four with aspirin; its hazard ratio was 0.72, again with a wide interval of 0.21 to 2.46. Combining PIK3CA or PTEN mutations did not strengthen the signal: the hazard ratio was 1.23. Nor did COX-2 overexpression, found in 307 of 450 assessed tumors, identify a benefit; its hazard ratio was 0.99.

The same paper combined three published randomized trials for participants with PIK3CA exon 9 or 20 mutations. That meta-analysis produced a disease-free-survival hazard ratio of 0.61, with a 95% confidence interval of 0.39 to 0.96. This aggregate result points toward fewer recurrence or death events, but it does not erase the null, highly imprecise ASCOLT subgroup result.

The independent laboratory study asked what aspirin does in PIK3CA-mutated cells under hypoxia, a low-oxygen condition used to model one feature of solid tumors. In isogenic SW48 colorectal cancer cells exposed to 1% oxygen, aspirin significantly enriched amino-acid transport gene sets in PIK3CA-mutated cells but not their wild-type counterparts, using a false-discovery-rate threshold below 0.05. Targeted metabolomics in HCT116 cells then showed that aspirin and hypoxia additively increased intracellular glutamine in the mutant cells; that additive pattern was not seen in wild-type cells.

Researchers perturbed glutamine handling with V-9302, an inhibitor of the ASCT2 glutamine transporter, and L-methionine sulfoximine, an inhibitor of glutamine synthesis. Both reduced intracellular glutamine and countered its aspirin-associated rise. However, only V-9302 combined with aspirin produced more growth inhibition than the single treatments in mutant HCT116 and DLD-1 cells, with colony-formation assays supporting the result in HCT116 cells. These experiments used three independent biological replicates and linked higher reactive-oxygen measurements with lower viability, but did not prove that oxidative stress caused cell death.

Analysis — A biomarker meets metabolic context

The cross-study pattern is convergence with an important contradiction. PIK3CA mutation identifies the same biological neighborhood in both studies: a possible patient-level aspirin response and a mutation-dependent change in amino-acid transport and glutamine accumulation under hypoxia. That makes tumor metabolism a plausible context modifier, rather than treating PIK3CA as a simple on-off marker. Yet the strongest direct evidence in the new patient analysis did not show a statistically significant benefit for any PIK3CA category, and COX-2 overexpression was essentially neutral. The favorable estimate appeared only after combining trials. The cell findings therefore help explain why effects might vary, but cannot rescue an underpowered clinical subgroup or establish a treatment mechanism in people. A productive research direction would test genotype together with metabolic and microenvironmental measurements in prospectively defined trial groups. That is an analysis of how the studies connect, not evidence that glutamine transport predicts recurrence or that blocking it would improve outcomes.

Limitations

The clinical analysis used tumor tissue from only a subset of ASCOLT participants. Its mutation-positive groups and event counts were small, confidence intervals were wide, and PTEN was exploratory. The pooled estimate combines trials but is not a second independent primary study in this briefing. Full text for the new ASCOLT report was not available through the licensed ingestion route, so claims here are constrained to its PubMed abstract.

The mechanistic study was conducted in cell lines, not people or animal tumors. Its 2 millimolar aspirin exposure exceeded concentrations typically achieved in human plasma. The models omitted immune and stromal interactions, the metabolomics and viability experiments had three biological replicates, and the study did not genetically confirm that ASCT2 alone caused the V-9302 result. Transporter evidence was mainly at the RNA level, reactive oxygen was not normalized to post-treatment cell number, and no antioxidant rescue experiment established causality. These findings define a testable mechanism; they do not establish clinical efficacy, safety, or a basis for medical decisions.