Ketogenic Diet Rewires Melanoma Metabolism by Boosting Pro-Death Lipids

Melanoma is one of the most metabolically flexible cancers, capable of switching between glycolysis, oxidative phosphorylation, glutaminolysis, and lipid oxidation to fuel growth and evade therapy. This metabolic plasticity makes it a compelling target for dietary interventions like the ketogenic diet, which forces a systemic shift from glucose to fat-derived ketone bodies. Understanding exactly how the KD reshapes tumor biology at the molecular level is critical for developing rational combination strategies.

The study used three genetically distinct human melanoma xenograft models—BRAF/NRAS wild-type (MeWo), BRAF mutant (A375), and NRAS mutant (Mel-Juso)—implanted in athymic nude mice. Animals were fed either a standard diet or a KD, and tumor tissues were profiled by targeted metabolomics (measuring amino acids, organic acids, acylcarnitines, and lipids including sphingomyelins and ceramides) alongside bulk RNA sequencing. Data integration was performed using VIP score-based methods and supervised latent variable modeling via the mixOmics DIABLO framework, enabling cross-omics pathway identification.

All three models showed delayed tumor growth on the KD. At the individual gene or metabolite level, however, there was strikingly little overlap across models—highlighting the well-known transcriptomic heterogeneity of melanoma. Despite this, pathway-level analysis uncovered convergent signatures: the KD consistently upregulated sphingomyelin and ceramide levels while inducing transcriptional programs favoring ceramide synthesis (e.g., upregulation of serine palmitoyltransferase and ceramide synthase genes) and suppressing ceramide degradation. Ceramides are bioactive lipids with established pro-apoptotic and anti-proliferative functions, making this finding mechanistically significant. The KD also downregulated mRNA levels encoding PI3K catalytic subunits, AKT isoforms, HIF-1α, MEK, and ERK—central nodes in oncogenic signaling—across models, suggesting a diet-mediated attenuation of survival and proliferation cascades.

The implications are notable for both cancer biology and clinical translation. The convergence of anti-tumor effects at the pathway level, despite molecular heterogeneity, argues that the KD may have broad applicability across melanoma subtypes rather than being effective only in specific mutation contexts. The sphingolipid axis in particular represents a tractable therapeutic node: drugs targeting ceramide metabolism are under active investigation, and the KD could potentiate their effects. Similarly, the observed transcriptional suppression of PI3K-AKT and MAPK components suggests potential synergy with targeted inhibitors already used clinically.

Several important caveats temper these conclusions. The xenograft models used athymic (immunodeficient) mice, meaning the immune-modulatory effects of the KD—potentially an important anti-tumor mechanism in immunocompetent hosts—were not captured. The metabolomics panel was targeted rather than untargeted, so novel or unexpected metabolic shifts may have been missed. Additionally, translating dietary interventions from mouse models to human patients involves substantial challenges in adherence, systemic metabolic context, and tumor microenvironment complexity.