Introduction: Adeno-associated viral (AAV) gene therapy has received approvals for the treatment of severe hemophilia A. Preclinical studies on the natural history of AAV have shown that, although most vectors persist as extrachromosomal episomal forms, integration into the host genome occurs at low frequencies. There is a paucity of data on the relevance of AAV integration in humans and whether this could impact on the long-term safety of gene therapy.

Aims: To characterize integration events following AAV gene therapy for the treatment of hemophilia and to evaluate for evidence of insertional mutagenesis.

Methods: Transjugular liver biopsies were obtained with informed consent from patients with severe hemophilia A (HA) treated using a liver-targeting recombinant AAV vector: AAV-HLP-hFVIII-V3 in the GO-8 Study (ClinicalTrials.gov: NCT03001830 & NCT04817462). Target Factor VIII levels pre-biopsy were >70% and further treatment was administered as per the study protocol in the first 48 hours post-biopsy. Integration site (IS) analysis was performed using target enrichment sequencing (TES) coupled to short-read next-generation sequencing. Common integration sites (CIS) were defined as clustering of ≥2 IS, within a 50kbp window and a significant CIS represented a CIS with frequency ≥5. Clonal expansion was evaluated based on the frequency of individual IS within a sample, with a significant threshold of >30% based on retroviral studies. All IS located ≤100kbp of the transcriptional start site (TSS) of annotated genes were referenced to Catalogue of Somatic Mutations in Cancer (COSMIC) cancer gene database to evaluate for enrichment proximal to known proto-oncogenes.

Results: Liver biopsy samples were obtained from four patients at between 0.1 and 3.1-years post AAV gene therapy. There was no evidence of tumorigenesis on histological assessment. Integration site analysis was performed on two liver biopsies with retrievable genomic DNA after 0.9 and 3.1-years follow-up. Vector-vector (concatemeric/episomal) forms accounted for 56.1-57.4% and vector-genome (integrated) 42.6-43.9% of all sequencing reads. Integration was seen in both liver biopsy samples, with a total of 237 unique integration sites detected (111 and 126 IS/samples). Approximately half of the unique IS were intergenic (54.4%, n=129). Of the IS located within genes (n=108), most were located within introns (86.1%, n=93) and all 15 exonic insertions were unique events. Vector-genome breakpoints were distributed throughout the AAV-HLP-hFVIII-V3 genome. Strong correlation (r=0.79, p <0.001) was seen between chromosome length (base pairs, bp) and the number of unique integration events occurring per chromosome. The frequency of unique integration sites after adjusting for chromosome length (IS/bp per chromosome) demonstrated uniform distribution of integration site counts across all chromosomes (mean IS/BP/chromosome = 7.37e-08, chi-squared= 4.48e-07, p=1). No significant common integration sites with a frequency ≥5 were seen, demonstrating no evidence of clustering of integration events within the human genome. There was no evidence of clonal expansion. The most frequent integration site (n=1) which represented two sequencing reads was located within intron 14 of PTPRM. There was no increase in integration events proximal to known cancer genes. A single IS read was located 13kb proximal to LIM domain only 1 (LMO1) with no evidence of clonal expansion.

Conclusion: Integration was seen in both liver biopsy samples studied at 0.9 and 3.1 years post-AAV gene therapy for the treatment of severe hemophilia A. In these two biopsies, there was no evidence of significant common integration sites, insertional mutagenesis or clonal expansion. These preliminary data provide further support for the long-term safety of AAV gene therapy.

This content is only available as a PDF.
2025
Sign in via your Institution