138. PI3K Activation Drives Protein Homeostasis Pathology in Venous Malformations

BACKGROUND: Venous malformations (VMs) are congenital slow flow malformations that affect ~1% of the population. They produce a spectrum of disease, from asymptomatic nodules to large painful lesions that cause functional limitation, coagulopathy, or death. Current interventions are palliative and recurrence is common; there are no FDA-approved treatments. Discovery of PIK3CA variants in VMs has identified a role for PI3K/AKT/mTOR hyperactivation in VM pathophysiology and led to the repurposing of drugs that inhibit the PI3K pathway, but a cure has so far been elusive. Further development of therapeutics is limited by an incomplete understanding of the mechanism by which PI3K hyperactivation causes VM pathology. Previous work with endothelial cells isolated from patient VMs (VMECs) has shown proteostasis defects, abnormal cell morphology, and hyperproliferation consistent with the clinical presentation of VMs. We hypothesize that protein homeostasis defects in VMECs are downstream of PI3K hyperactivation and that distinct variants affect phenotypic severity. METHODS: Human umbilical vein endothelial cells (HUVECs) were engineered to express empty plasmid or H1047R, E545K, or E542K pathogenic variants of PIK3CA using a lentiviral system. A GFP-expressing plasmid was used to confirm successful transduction. The resulting cells lines were stained by immunofluorescence for pAKT, CD31, and VECADHERIN and imaged using fluorescence microscopy. Quantification was performed using ImageJ and analyzed with Prism. Statistical analysis was performed using one-way ANOVA with post-hoc Tukey t-test. RESULTS: GPF transduction of HUVECs showed almost 100% efficiency. All three PIK3CA variants significantly increased the levels of activated AKT (ANOVA, p<0.0001) confirming effective transduction and function of the transgenes. All three variants led to significantly increased cell size (ANOVA, p<0.0001). The H1047R variant also showed significantly increased number of multi-nucleated cells (post-hoc Tukey t-test, p<0.0001). Abnormal cytoplasmic accumulation of CD31 and VECADHERIN was seen in all three cell variants. Quantification of protein levels showed that H1047R and E542K variants had significantly increased CD31 expression (post-hoc Tukey t-test, p< 0.05 and p<0.005). Only the H1047R and E545K variants had significantly increased VECADHERIN expression (post-hoc Tukey t-test, p<0.005). These changes in protein expression mimic changes observed in patient derived VMECs. CONCLUSIONS: The introduction of PIK3CA variants that hyperactivate PI3K signaling recapitulated cell morphology and proteostasis defects seen in patient VMECs. The H1047R variant demonstrated the most severe phenotype, which correlates with published clinical data. Disparities between variants may be relevant to understanding both the clinical heterogeneity of VM patients and their differential response to treatment. These data provide compelling evidence that VMEC phenotypes are downstream of PI3K/AKT/mTOR signaling. Future work examining how PI3K hyperactivation causes VMEC dysfunction may uncover novel avenues for therapeutic treatment.