Background: Recent data in mice demonstrated that transient FVIII expression from plasmid DNA injection resulted in an increased incidence of hepatocellular carcinoma (HCC) through a proposed mechanism of misfolded FVIII inducing an unfolded protein response (Kapelanski-Lamoureux et al., 2022). While development of HCC has not emerged post adeno-associated virus (AAV) gene transfer in hemophilia A (HA) patients, these data raised major safety concerns that FVIII expression may result in HCC post gene transfer. Concern around HCC development is compounded by theoretical safety risks raised by observations of HCC in neonatal mice post systemic AAV vector infusion and by underlying HCC risk factors in the HA population, e.g. viral hepatitides and above background prevalence of hepatic steatosis.
Aims: We sought to investigate tumorigenesis and long-term liver health in mice treated with an AAV8-FVIII vector to express human FVIII-WT and a FVIII variant, FVIII-R336Q/R562Q (FVIII-QQ), at variable FVIII expression ranges relative to untreated wild type (WT) mice.
Methods: 8-12 week old male HA/CD4KO mice (n=62) were treated with AAV8 vectors to express FVIII at plasma concentrations in the range of mild HA (0.05-0.4nM), normal FVIII (0.5-1.5nM) and elevated FVIII (>1.5nM) and followed for 72 weeks post vector administration. AAV treated mice were compared to mock-injected WT/CD4KO mice (n=10) on the same genetic background and fed the same 11% fat diet. Plasma samples were collected every 4 weeks throughout the study. 51 animals survived to 72 weeks and livers were harvested. Left lobe sections were stained with H&E, Masson's Trichrome and Congo red to evaluate for: thrombosis, steatosis, neoplastic changes, fibrosis and amyloidosis. Blinded histopathological evaluation and morphometrical analysis was performed. Harvested livers were also analyzed for markers of cellular stress by RT-qPCR and Western blot. Alanine aminotransferase (ALT) and alpha fetoprotein (AFP) plasma levels were measured at different time points.
Results: 42 of 62 AAV treated animals and 9 of 10 WT mice survived to 72 weeks; the majority of AAV treated animals lost before 72 weeks expressed FVIII in the mild HA range. Analogous to human clinical trial observations, mice with initially elevated FVIII expression lost half of plasma FVIII by week 72 while more modest FVIII expression was stable up to 72 weeks post vector; analysis of liver F8 mRNA and vector copy number (VCN) did not demonstrate evidence of gene silencing while loss of VCN correlated with loss of FVIII expression (manuscript under review). Liver histology demonstrated no evidence of neoplastic or pre-neoplastic lesions or vascular thrombosis, including animals expressing the FVIII-QQ gain-of-function variant. WT controls showed significantly greater micro- and macrovesicular fatty changes consistent with steatosis, compared to AAV treated animals. All animals demonstrated a low degree of inflammatory changes that were significantly lower in animals expressing FVIII in the range of mild HA. Congo red staining did not demonstrate evidence of protein aggregation or amyloid plaque formation in AAV treated animals. Notably, however, animals with initially elevated FVIII expression (>1.5nM) showed upregulation of CHOP, a pro-apoptotic protein associated to the unfolded protein response. Concurrently, mice in the elevated FVIII cohort had higher ALT plasma levels at week 68 relative to the mild HA group, while the normal FVIII cohort did not significantly differ. AFP levels were not significantly elevated in any mice.
Conclusions: Our study provides the first long-term follow up of FVIII expression and liver health, including tumorigenesis, in mice after AAV gene transfer. FVIII expression durability data support that mice recapitulate human observations and may be used to mechanistically define thus far unexplained year-over-year losses of FVIII expression post gene transfer observed in some HA gene therapy trials. There was no evidence of tumorigenesis or HCC post FVIII-mediated gene transfer supporting that hepatocyte expression of FVIII is not an independent risk factor for HCC. The observations of significantly greater steatosis in WT mice versus AAV treated HA/CD4KO mice will be further investigated. Studies investigating evidence of endoplasmic reticulum stress and an unfolded protein response are ongoing.
George:Pfizer: Consultancy; Regeneron: Consultancy; Asklepios BioPharmaceutical: Patents & Royalties; Spark: Consultancy; CSL Behring: Consultancy; Form Bio: Membership on an entity's Board of Directors or advisory committees; Tome Biosciences: Consultancy.
This feature is available to Subscribers Only
Sign In or Create an Account Close Modal