Genome-Wide Analysis of Recombinant AAV Integration Reveals a Low Risk of Insertional Oncogenesis in the Human Liver

 

Introduction

Recombinant adeno-associated virus (rAAV) vectors have become the leading platform for in vivo gene therapy, transforming the treatment of numerous inherited diseases, including spinal muscular atrophy (SMA) and hemophilia. Hundreds of AAV-based therapies are currently in clinical development or have reached clinical practice because of their ability to achieve long-term gene expression with an excellent overall safety profile. As the number of treated patients continues to increase, however, rare adverse events are becoming more apparent, emphasizing the need for a comprehensive understanding of long-term vector safety.

Among potential long-term concerns, insertional oncogenesis has received increasing attention. Although recombinant AAV genomes remain predominantly episomal, a small fraction integrate into the host genome. Given the exceptionally high vector genome copy numbers that can accumulate in hepatocytes following systemic administration, even low-frequency integration events can translate into millions of genomic integrations across the liver. Animal studies have demonstrated that AAV integration can promote hepatocellular carcinoma (HCC), but the relevance of these findings to humans remains uncertain because human cells are substantially more resistant to malignant transformation than murine cells.

Human clinical experience has been reassuring. Despite hundreds of AAV clinical trials and long-term follow-up of patients receiving liver-directed gene therapy, only isolated cases of malignancy have been reported, and convincing evidence linking recombinant AAV integration directly to cancer remains lacking. Nevertheless, observations from wild-type AAV2 infections, where integration near liver cancer driver genes has occasionally been detected, underscore the importance of evaluating recombinant AAV integration patterns at genome-wide resolution.

To address this question, a recent study published online in the journal Molecular Therapy analyzed AAV integration in primary human hepatocytes xenografted into FRG mice and in the livers of cynomolgus macaques, generating one of the largest AAV integration datasets reported to date.

Genome-Wide Distribution of AAV Integration Sites

Using targeted enrichment of host-vector junctions followed by next-generation sequencing, the investigators identified approximately 1.54 million unique integration sites in human hepatocytes and 1.32 million integration sites in macaque liver. These datasets were generated from multiple independent liver samples and sequencing libraries, providing exceptional statistical power for evaluating genome-wide integration behavior.

Integration sites were distributed throughout the genome with an average density of approximately 0.5 integration sites per kilobase. All autosomes exhibited similar integration frequencies, while the X and Y chromosomes showed comparable integration densities after correcting for chromosome copy number and assembly differences. Within the vector genome, integration junctions were observed throughout the recombinant AAV sequence, with the inverted terminal repeats (ITRs) representing the most common breakpoints.

These findings demonstrate that recombinant AAV integrates broadly throughout both the human and macaque genomes, providing an extensive foundation for evaluating insertional mutagenesis risk.

Common Integration Sites in Human Hepatocytes

To identify genomic regions exhibiting preferential integration, the investigators searched for common integration sites (CIS). Regions with integration frequencies exceeding the genome-wide average by more than 20-fold, 5-fold, or 2-fold were classified as Tier 1, Tier 2, and Tier 3 CIS, respectively.

Only 45 CIS were identified in human hepatocytes, including 5 high-density, 4 intermediate-density, and 36 low-density sites. Importantly, none of these CIS overlapped with genes known to drive hepatocellular carcinoma or confer HCC susceptibility. Only a single low-density CIS intersected ARHGAP26, a gene implicated in several non-hepatic cancers but not liver cancer.

The investigators also specifically examined genes previously associated with wild-type AAV-mediated HCC, including PLAG1 and MEG8, the human ortholog of the murine Rian locus. Although isolated integration events were detected within these genes, none demonstrated enrichment above background integration density.

These observations indicate that recombinant AAV does not preferentially integrate near established human liver cancer genes despite the enormous number of integration events analyzed.

Comparison of Human and Macaque Integration Hotspots

Comparison of human and macaque datasets revealed striking species differences. Although the total number of integration events was similar, the macaque genome contained 510 common integration sites, more than ten times the number identified in human hepatocytes.

Remarkably, only two genes—albumin (ALB) and CHSY1—contained CIS in both species. Furthermore, none of the macaque CIS occurred within genes whose human orthologs are known cancer-associated genes.

These findings indicate that although overall integration behavior is similar between humans and non-human primates, specific integration hotspots are largely species-specific. Consequently, caution should be exercised when extrapolating non-human primate integration data directly to human oncogenic risk.

Integration Preferentially Occurs in Transcriptionally Active Chromatin

Genome-wide analysis demonstrated that recombinant AAV integration is not completely random. Instead, integration preferentially occurred within actively transcribed genes, particularly genes highly expressed in the liver.

Within genes, integration favored intronic sequences, whereas exons were slightly underrepresented in human hepatocytes. Integration was also enriched within CpG islands, promoter regions, and within approximately 10 kb of transcription start sites.

Correlation with publicly available epigenomic datasets revealed significant enrichment of integration within regions marked by H3K27ac, H3K4me1, H3K4me3, and H3K36me3, all hallmarks of transcriptionally active chromatin. In contrast, integration was depleted from regions marked by H3K9me3 and H3K27me3, which represent transcriptionally silent chromatin.

Similarly, integration preferentially occurred within open chromatin identified by ATAC-seq and DNase I hypersensitivity assays and near binding sites for multiple liver-specific transcription factors.

The macaque dataset exhibited nearly identical trends, although it showed a somewhat stronger preference for exonic integration.

Collectively, these findings indicate that chromatin accessibility and transcriptional activity are major determinants of recombinant AAV integration.

Unexpected Integration at the AAVS1 Locus

One unexpected observation was the identification of a common integration site within PPP1R12C, the gene containing the well-characterized AAVS1 locus, which serves as the site-specific integration target for wild-type AAV mediated by the viral Rep68/Rep78 proteins.

Repeated integration at AAVS1 was observed in all three human donor samples but was absent in macaques. Because recombinant AAV vectors do not encode Rep proteins, the investigators explored potential explanations.

PCR analysis detected wild-type AAV rep sequences in one human donor, suggesting latent wild-type infection as one possible source of Rep activity. Deep sequencing of the vector preparation also identified trace amounts of rep DNA contamination, although no replication-competent AAV particles were detected.

These results suggest that low-level Rep activity, derived either from latent wild-type AAV infection or minimal vector preparation contamination, may explain the observed AAVS1 targeting rather than representing an intrinsic property of recombinant AAV vectors.

Vector Genome Biodistribution Across the Liver

The availability of cynomolgus macaque liver tissue also enabled assessment of vector genome biodistribution across the four hepatic lobes.

This analysis addresses an important practical question: whether a biopsy obtained from a single location accurately reflects whole-liver vector distribution. Uniform vector biodistribution would support the use of localized biopsies for measuring vector copy number, transgene expression, and integration frequency, whereas heterogeneous distribution could complicate interpretation of both preclinical and clinical studies.

Understanding liver-wide biodistribution therefore represents an important component of both efficacy assessment and long-term safety evaluation.

Implications for AAV-Mediated Oncogenic Risk

The most important conclusion of this work is that recombinant AAV demonstrates no preferential integration into genes associated with hepatocellular carcinoma despite analysis of nearly three million integration events across two species.

Although integration preferentially occurs within transcriptionally active chromatin, this bias reflects chromatin accessibility rather than selective targeting of oncogenic loci. The absence of enrichment near HCC-associated genes, together with extensive long-term clinical follow-up showing no increase in liver malignancies after liver-directed AAV gene therapy, strongly supports the conclusion that insertional oncogenesis represents a very low-probability event.

Species differences observed between humans and macaques further emphasize that human integration data should serve as the primary basis for clinical risk assessment. Future studies analyzing liver biopsies obtained directly from treated patients will further refine our understanding of long-term integration behavior and determine whether patient-specific factors such as age, sex, vector dose, or underlying liver disease influence integration patterns.

Conclusions

This study represents one of the most comprehensive analyses of recombinant AAV integration performed to date. By analyzing over 2.8 million unique integration sites, the investigators demonstrated that AAV integration is widespread but exhibits a reproducible preference for transcriptionally active genes and open chromatin. Importantly, this integration bias does not translate into preferential targeting of known liver cancer genes or established oncogenic hotspots.

The findings provide strong evidence that recombinant AAV-mediated insertional oncogenesis in the human liver is likely to be rare, supporting the favorable safety profile observed in clinical trials. Nevertheless, continued long-term surveillance of treated patients, combined with direct analysis of integration events in human liver tissue, will remain essential for fully defining the long-term genomic safety of AAV-based gene therapies.

Reference

Scott S, Hallwirth CV, Sasaki N, Cunningham SC, Zhu E, Khandekar N, Kizana E, Baruteau J, Gissen P, Perocheau DP, Bauer DC, Wilson LOW, Lisowski L, Alexander IE, Identification of AAV Vector Integration Sites in Primary Human and Macaque Hepatocytes in vivo, and Analysis of Oncogenic Risk, Molecular Therapy (2026), doi: https://doi.org/10.1016/j.ymthe.2026.07.039

Image credit: Portions of the figure in this article were generated using ChatGPT (OpenAI) or Google Gemini.

Disclaimer: This blog post is intended solely for educational and scientific informational purposes. Any mention of therapeutic drug names, including FDA-approved medications, is for the purpose of accurate reporting and discussion of biomedical research and does not constitute medical advice, endorsement, or promotion. Readers should not interpret the content as a recommendation for any specific treatment. Always consult a qualified healthcare professional for medical advice or treatment decisions.

About Marin Biologic Laboratories

Our Recent Publication/Meeting Presentation on Gene Therapy

1. Development of VNX-101, an Adeno-Associated Virus with Less Immunogenicity and Efficient Long-Term Expression of a CD19 T-Cell Engager. Molecular Therapy Methods & Clinical Development, published online July 24, 2025.

3. Cell-Based Potency Assay for Anti-CD3-Anti-CD19 Diabody. Journal of Immunological Methods. 2025. 545-114004.

3. Development of a Pharmacokinetic (PK) Mouse Serum GLP ELISA for an Anti–CD19–AntiCD3 Diabody
bioRxiv 2025.03.19.644217; doi: https://doi.org/10.1101/2025.03.19.644217.

4. American Society of Hematology (ASH) Annual Meeting 2024.
Abstract link: Using Gene Therapy to Solve Challenges with CAR-T Cell Immunotherapy: Lead Selection and Preclinical Development of an Adeno-Associated Virus with Reduced Immunogenicity Exhibiting Efficient and Long-Term Expression of an Anti-CD19 T-Cell Engager.

Comprehensive Assay Solutions for In Vitro and Cell Based Potency Assays and Pharmacokinetics Studies- Our Expertise

With 30 years of expertise in cell culture, cell-based assays, and preclinical/clinical PK/PD analysis, we specialize in offering assay services essential for a wide variety of therapeutic drug development programs, preclinical studies, IND/BLA applications, and commercialization. Our comprehensive services include both preclinical non-GLP and GLP assays, as well as non-GMP and GMP assays, providing critical support throughout the entire development pipeline.

Watch the following video and explore our latest presentation on the development and validation of potency and pharmacokinetic (PK) assays for AAV vectors, highlighting innovative methodologies and industry-leading expertise.

 

 

Download the full presentation: Development of Custom Cell Based and In vitro Potency and Pharmacokinetics (PK) Assays for AAV vectors- Marin biologic Laboratories

 

Development of Cell-Based Potency Assays: Case Studies and Blogs from Marin Biologic Laboratories (MarinBio)

 

Drug Discovery & Development Assays Offered by Marin Biologic Laboratories (MarinBio)