AAV Integration and Cancer Risk: From Theoretical Risk to Molecular Evidence

Author: Dr. Aaron Zhang-Chen, CQA, PhD, CTO at GeneGoCell on November 9, 2026

AAV IntegrationAAV Gene TherapyInsertional OncogenesisGenome Editing INtegration AnalysisIntegration Site AnalysisTransgene IntegrityG-Int℠Cell And Gene Therapy

What the FDA and Novartis disclosed

On November 5, 2026, through SMA Europe, Novartis confirmed nine reports of tumors in patients previously treated with Zolgensma (onasemnogene abeparvovec), an AAV9 gene replacement therapy for spinal muscular atrophy given intravenously under 2 years of age, approved May 24, 2019 and used in over 6,000 patients. Of the nine, “two cases to date have shown evidence of vector DNA integration into the DNA of the patients.”

On November 6, 2026, FDA announced it is investigating a serious risk of malignancy linked to Zolgensma, including confirmed product-related insertional oncogenesis involving the GLI1 oncogene. Three cases involved GLI1-associated tumors; in two, molecular analysis of tumor tissue showed the promoter region of the construct inserted upstream of or within GLI1. FDA determined the risk applies to both onasemnogene abeparvovec products, namely Zolgensma and Itvisma, the intrathecal product approved November 24, 2025 for patients 2 years and older, because they share the same vector and construct. No neoplasia has been reported after Itvisma to date. FDA states the benefits of both continue to outweigh their risks; the current label describes tumorigenicity from vector integration as a theoretical risk.

The scope should be stated precisely. 6,000+ patients treated. Nine tumors reported and two with demonstrated vector integration at GLI1 are different statements.

Why AAV integration-site analysis matters

A theoretical risk cannot be managed until it is measured. AAV remains broadly non-integrating across thousands of recipients; what these reports change is not the frequency estimate but the standard of evidence. Integration analysis is now what distinguishes a coincidental tumor from insertional oncogenesis caused by the product. That requires highly sensitive measurement of the potential insertion site, including those with just one end but not the other, as well as more in-depth analysis of insertion site to evaluate the biological consequence (risk) of the integration. “Not detected” is interpretable only against an assay’s validated sensitivity of the test.

Evidence from Zolgensma, FBX-101, and RGX-111

Zolgensma: promoter insertion at a developmental oncogene. The FDA language is specific and worth reading closely. What was found in the two confirmed cases was not simply vector DNA in tumor tissue but the promoter region of the construct inserted upstream of or within GLI1, a transcription factor that is oncogenic when overexpressed or abnormally activated. That is a mechanistic finding, produced by molecular analysis of tumor tissue, and it is the reason FDA moved its description from theoretical risk to confirmed product-related insertional oncogenesis.

FBX-101: a promoter-only fragment with no transgene. Vo et al. (N Engl J Med, October 15, 2026) described a boy with infantile-onset Krabbe disease who received intravenous FBX-101, an AAVrh.10 vector with AAV2 ITRs carrying GALC under a CMV-enhancer/chicken β-actin (CAG) promoter. Surveillance MRI found an incidental paraspinal mass, diagnosed as a GLI1-altered neoplasm; long-read sequencing identified a clonal 1,038-bp insertion in GLI1 intron 1, present in 44% of reads. The integrant was subgenomic: a partial AAV2 ITR, CMV enhancer, chicken β-actin promoter, exon 1, and a splice donor, no GALC transgene nor polyadenylation signal. The splice donor generated a chimeric β-actin exon 1–GLI1 transcript, and allele-specific hypomethylation established the integration as the proximate cause of GLI1 activation.

RGX-111: a rearranged cassette at a second oncogene. Ahrens-Nicklas et al. (N Engl J Med, June 4, 2026) reported a neuroepithelial tumor after intracisternal magna delivery of an AAV9 vector carrying IDUA. A rearranged cassette integrated at PLAG1 through the chicken β-actin splice donor to PLAG1 exon 5, with PLAG1 expression far above other CNS tumors in the institutional cohort. The integrant carried HEK293-derived chromosome 10 sequence and, unlike FBX-101, retained IDUA fragments.

The shared mechanism is the promoter/enhancer-driven oncogene activation, not therapeutic transgene expression. Both integrated fragments were truncated subgenomic structures retaining a strong promoter-enhancer and a splice donor: integration near an oncogene, promoter-driven transcription, splice-donor fusion to a downstream exon, chimeric transcript, overexpression, and clonal expansion. The therapeutic transgene coding region was not required for the observed oncogene activation.

Two meanings of “backbone” should not be conflated. The vector cassette backbone, including ITR, promoter and enhancer, splice donor, is integrated. Plasmid (bacterial) backbone outside the ITRs is a separate and real manufacturing impurity, but it is not the published tumor driver; the extra DNA in the RGX-111 integrant was HEK293 producer-cell sequence, not bacterial plasmid backbone.

The analytical consequence is direct. An assay that targets only the therapeutic transgenes would have missed the tumor driver entirely, because that fragment contained no coding sequences.

What AAV integration-site testing should assess

How GeneGoCell’s G-Int℠ workflow supports integration analysis

GeneGoCell’s G-Int℠ integration-analysis workflow, introduced August 17, 2026: addresses these questions through G-Int-Site℠ (ligation-mediated PCR with semi-specific target-amplification sequencing and UMIs), G-Int-Efficiency℠ (integration efficiency at a given site), and G-Int-Integrity℠ (long-range PCR with PacBio long-read sequencing to characterize integrated sequence structure). The workflow predates the November FDA safety communication.

One detail is worth noting. The FBX-101 integrated fragment contains a partial AAV2 ITR, which is a shared vector element. A fragment of this type is detectable with the G-Int℠ assay, at our industry-leading sensitivity of 0.01% or better; G-Int-Site℠ can reach sensitivity even better than 0.01% when a program requires it.

GeneGoCell: info@genegocell.com | 858 222-1069 | www.genegocell.com

References

  1. FDA. Investigating Serious Risk of Malignancy Linked to Zolgensma Gene Therapy (November 6, 2026). FDA safety communication

  2. SMA Europe. Information regarding recent Zolgensma safety concerns (Novartis statement, November 5, 2026). SMA Europe statement

  3. Vo JN, Wu YM, Heider A, et al. Integration-Associated Paraspinal Neoplasm after AAV Gene Therapy. N Engl J Med. 2026;395(15):1555–1558. DOI: 10.1056/NEJMc2607927

  4. Ahrens-Nicklas RC, Kotch C, Roche AM, et al. Neuroepithelial Tumor with AAV Integration after Intracisternal Magna Vector Delivery. N Engl J Med. 2026;394(21):2126–2133. DOI: 10.1056/NEJMoa2601608

  5. GeneGoCell. Detection and Quantification of Integration using G-Int℠ (August 17, 2026). G-Int℠ Insights article | G-Int℠ service page

Start today!

Click here to start a conversation with our expert team.