AAV and hepatitis: Cause or coincidence?

Adeno-associated viral (AAV) vectors are widely used in clinical gene therapy and, in particular, in applications based on in vivo gene transfer. These recombinant vectors are derived from wild-type AAVs, which are replication incompetent viruses that depend on other viruses, such as adenoviruses (AdVs) or herpes viruses, to complete their lytic replication cycle. Various serotypes, including AAV2, infect the majority of people around the world and are thought to be part of self-limiting respiratory infections without chronic morbidity or mortality. Because of this benign clinical profile, AAV vectors are generally referred to as nonpathogenic viruses. After initially having been associated with a subset of hepatocellular carcinomas (HCCs), two recent clinical investigations further challenge its nonpathogenic nature and report the surprising finding that all explanted livers from children with acute severe hepatitis contained high levels of AAV2. This raises the question of whether AAV could contribute to hepatitis during co-infection with a helper virus.

The nonpathogenic profile of AAV infection was first questioned by the discovery of AAV viral integrations in putative oncogenes in a small subset of HCC. Further investigations showed that nontumor liver tissue from over 4% of these individuals contained AAV episomes, many of which were transcriptionally active.¹ Beyond its likely involvement in HCC development, no other liver pathology was observed in livers with transcriptionally active episomes. Yet, these data illustrate that AAV infection can chronically persist in the livers of a surprisingly large number of people.

Now, a second observation raises questions about the nonpathogenic role of AAV2 infection. In late 2021 and early 2022, after social distancing due to coronavirus 2019 (COVID-19) eased, there has been an uptick in children around the age of 3–4 years old who presented with acute severe hepatitis. Most cases were initially reported in the United Kingdom and since then have been found around the world, including widely across the United States. The large majority recovered with supportive measures, yet some went on to require liver transplantation or died. Because this acute hepatitis occurred mostly in previously healthy children, an infectious origin has been widely suspected. Common viral etiologies such as hepatitis A, B, C, and E viruses and herpes family viruses were ruled out, as well as other noninfectious causes of childhood hepatitis. Further investigations quickly identified AdV-F41 as a possible etiology since ∼90% of children with acute hepatitis tested positive.^2,3 AdV-F41 has been a known cause of pediatric gastrointestinal infections for decades, and many of the children with acute hepatitis reported diarrhea and vomiting in the weeks prior to presenting with hepatitis. However, AdV-F41 was not previously associated with severe hepatitis in immunocompetent children. Furthermore, histopathology did not show AdV inclusions, and AdV DNA was only detected in a small subset (2/15) of livers. These observations raised the question of whether this was truly the cause of this new acute severe hepatitis. Now, two reports from Scotland⁴ and England⁵ have been posted in pre-print format that both applied next-generation, including metagenomic, sequencing to search for infectious etiologies in children with acute severe hepatitis. Both groups identified high levels of AAV2 in all (9/9) explanted livers and in the blood of 15/16 children that recovered without liver transplantation. Minimal to no AAV2 could be detected in various control groups, including in AdV-F41-positive children with normal liver transaminases. These epidemiological observations make a case, but do not prove, that AAV2 is somehow causally involved in this acute severe hepatitis syndrome. Alternatively, the presence of AAV may merely correlate, rather than contribute, to viral hepatitis. In addition, many other questions remain, including whether this was a previously existing hepatitis syndrome that became more prevalent after 2 years of relative underexposure during the COVID-19 period, or whether viral variants have created a new clinical syndrome.

While causality between AAV2 and acute hepatitis needs to be formally proven, these striking findings raise further doubt on a universal nonpathogenic role of AAV2 infection. Some individuals may be more susceptible to respond adversely to AAV2 co-infection. One intriguing first lead may come from the small Scottish cohort, which was strongly enriched for a major histocompatibility complex (MHC) class II allele previously associated with autoimmune hepatitis.⁴ Further understanding how host HLA genetics affect responses to AAV2 infection may also benefit AAV gene therapy. While many participants in liver-directed AAV gene therapy trials show acute hepatitis, clinical predictors as to who mounts liver-damaging immune responses have been lacking. Similarly, it has been unclear why CD8⁺ T cell responses against viral capsid can cause a rise in liver enzyme levels months after gene transfer, a phenomenon that has not been replicated in pre-clinical studies in animals. It is increasingly being suspected that natural AAV2 infection may contribute to pathogenicity in a small subset of individuals. A better understanding of the AAV2 life cycle, its dependence on helper viruses, and immune responses in humans is increasingly becoming pertinent.