The role of PHD2 inhibition in melanoma progression, metabolic adaptation, and therapy resistance

Therapy-resistant melanoma poses significant challenges due to its diverse phenotypes and limited treatment options. A strong dependency of melanoma cells on EGLN1 , which encodes prolyl hydroxylase domain protein 2 (PHD2)-a key regulator of HIF-alpha degradation- is evident in large-scale dependency datasets, prompting us to investigate EGLN1 / PHD2 as a potential therapeutic vulnerability. Melanoma cell lines were engineered using CRISPR-Cas9 to generate knockout (KO) models of EGLN1 , HIF1A , and HIF1AN . Pharmacologic assays were performed using the PHD inhibitor roxadustat and the FIH inhibitor tool compound DM-NOFD. Proliferation was assessed in short- and long-term assays. Protein expression was analyzed by immunoblotting, and transcriptional changes were evaluated by RT-qPCR. Metabolic consequences were assessed by measuring NADH/NAD⁺ ratios. To model resistance to BRAF + MEK inhibition (BRAFi+MEKi), two systems were generated: (1) TGFbeta1-induced adaptive resistance and (2) drug-acquired resistance following continuous BRAFi+MEKi exposure. Resistance was confirmed by phosphorylated ERK immunoblotting and IC₅₀ analysis. EGLN1 KO reduced melanoma proliferation by 70-80% relative to control cells (p < 0.05). Roxadustat inhibited growth in a HIF-1alpha-dependent manner, with IC₅₀ values of 40-80 µM in wild-type cells versus >100 µM in HIF1A KO cells. Long-term treatment with 10-20 µM roxadustat reduced wild-type proliferation by 30-70%, but only 0-10% in HIF1A KO cells. Disruption of HIF1AN or treatment with DM-NOFD enhanced roxadustat sensitivity, decreasing proliferation by 40-60%, an effect absent in HIF1A KO models. Roxadustat increased PHD2 protein levels and induced EGLN1 transcription in wild-type cells, accompanied by upregulation of canonical HIF-1 targets ( LDHA, PDK1, BNIP3 ). A 20-50% increase in the NADH/NAD⁺ ratio after 4 h of roxadustat treatment indicated a metabolic shift toward reductive stress. In BRAFi+MEKi-resistant models, roxadustat enhanced BRAFi+MEKi potency and inhibited ERK phosphorylation, reversing MAPK pathway reactivation in A375 acquired-resistant cells, but had no effect in A375 HIF1A KO resistant models, demonstrating a requirement for HIF-1alpha. In TGFbeta1-induced adaptive resistance, HIF1A was necessary for the development and maintenance of resistance phenotypes, indicating that HIF-1alpha signaling contributes to early adaptive BRAFi+MEKi tolerance. These studies demonstrate that PHD2 inhibition suppresses melanoma growth through HIF-1alpha stabilization and can be enhanced by concurrent FIH inhibition. The requirement of HIF-1alpha for both adaptive and acquired BRAFi+MEKi resistance and for roxadustat-mediated re-sensitization highlights a functional link between hypoxia signaling and therapeutic tolerance. These findings support continued evaluation of PHD2-directed strategies, including combination approaches, in melanoma.