Should patients with pre-existing anti-AAV antibodies be excluded from receiving intrathecally delivered AAV vectors?

Main text

Pre-existing neutralizing antibodies (NAbs) to adeno-associated virus (AAV) remain one of the largest practical barriers to limit the availability of gene therapy treatments to patients.^1,2 Most AAV-based gene therapy clinical trials exclude patients with pre-existing anti-AAV NAbs. Although this is highly justified for AAV-based treatments delivered intravenously (i.v.), there has been an outstanding question of whether this exclusion should be in place for other routes of administration, such as direct intraparenchymal injection, subretinal injection, or intrathecal (IT) administration. In this issue of Molecular Therapy Methods & Clinical Development, Vono et al. rigorously explored this question for IT administration of AAV9 in non-human primates (NHPs).³ At a high level, the results from Vono et al. confirmed what other past studies had at least suggested: when AAV9 is delivered intrathecally, pre-existing anti-AAV antibodies do not inhibit central nervous system (CNS) biodistribution or create any major safety concerns. The significance of this should be clear—that it does not appear necessary to exclude patients from a CNS-directed gene therapy based on pre-existing antibodies to AAV. If peripheral tissue targeting is also desired, the considerations are nuanced.

Initial NHP studies from over a decade ago suggested that animals with low-level pre-existing anti-AAV9 NAbs could be injected IT with AAV9 vectors without any measurable inhibition of transduction.⁴ This contrasted with accepted dogma around i.v.-delivered AAV, namely that there was a clear anti-AAV antibody threshold that would substantially inhibit AAV vector biodistribution, and thus patients with pre-existing antibodies should be excluded from clinical trials. Clinical trials utilizing other routes of administration would oftentimes use similar exclusion criteria of seropositive subjects as i.v.-based clinical trials, either out of an abundance of caution or to simply follow these other established precedents. However, other trials and Food and Drug Administration (FDA)-approved gene therapy drugs have not excluded seropositive patients. The FDA-approved AAV2 vector injected subretinally for Leber’s congenital amaurosis, Luxturna, does not list AAV2 seroprevalence as a contraindication (https://www.fda.gov/media/109906/download). The FDA-approved AAV2 vector injected intraparenchymally into the putamen for aromatic l-amino acid decarboxylase (AADC) deficiency, Kebilidi, likewise does not list seroprevalence to AAV2 as a contraindication (https://www.fda.gov/media/183530/download). However, the Kebilidi drug label notes that “there are insufficient data to determine the effect of anti-AAV2 antibody on the pharmacokinetics, pharmacodynamics, safety, or effectiveness.” IT clinical trials are very mixed, with some excluding seropositive subjects and others accepting those subjects. The first IT gene therapy trial, which tested IT administration of an AAV9/GAN vector for giant axonal neuropathy (GAN), did not exclude seropositive subjects.⁵ While the overall number of trial participants was limited (n = 14), 43% of the enrolled subjects were seropositive at baseline, and there was no apparent difference in safety or efficacy of the treatment correlating with baseline anti-AAV9 seroprevalence. In one subject that was seropositive at baseline, postmortem analysis showed CNS vector DNA biodistribution that was consistent with preclinical studies, although biodistribution to most peripheral organs was extremely low (<0.01 vg copies per cell) except to the spleen. This very limited clinical experience suggests that IT administration of AAV (at least AAV9) can be utilized safely and effectively to target the CNS.

Vono et al. aimed to address the impact of pre-existing serum AAV9 antibodies by conducting an AAV9/mCherry biodistribution study in 3 groups of NHPs, with each NHP receiving a dose of 1 × 10¹³ vg per animal: low seropositive (total anti-AAV9 antibody [TAb] ≤1:20), medium seropositive (TAb 1:160–1:640), and high seropositive (TAb 1:40,960–1:327,680).³ The animals did not receive any immunosuppression. The low and medium seropositive animals had undetectable TAbs in their cerebrospinal fluid (CSF), whereas the high seropositive animals had CSF TAb titers ranging from 1:160 to 1:640. In terms of vector biodistribution, there were no statistically significant differences for brain and spinal cord across any of the groups. In peripheral organs and dorsal root ganglia, vector biodistribution consistently trended lower in the medium group and was significantly lower in the high seropositive group. The notable exception to this was the spleen, where biodistribution increased correlating with higher pre-existing TAbs, similar to the n = 1 experience in the GAN clinical trial. Overall, safety assessment including clinical evaluations, blood chemistry, and histopathology did not reveal any major safety concerns associated with higher amounts of pre-existing TAbs. It is noteworthy that there were trends toward increased severity of mononuclear cell infiltration in the meninges and/or perivascular region in the neuropil in animals from the high seropositivity group, but the changes were asymptomatic and determined to be non-adverse. In terms of immune responses, there were some differences associated with the high seropositivity group compared to the low and medium groups. The high seropositivity group had an anti-AAV9 TAb response that was faster and at a greater magnitude. Likewise, various CD4⁺ T cell populations were elevated to a greater extent in the high group, and there was a trend toward greater CD8⁺ T cell elevation in the high seropositivity group. There were no differences in cytokine activation across the groups except for transient increases of IP-10, I-TAC, MCP-1, and IL-RA in the high seropositivity group. There was no difference in complement activation across the groups.

Many patients will potentially be excluded from receiving IT-delivered gene therapy treatments if pre-existing AAV seroprevalence is an exclusion and/or listed contraindication. AAV9 NAb seroprevalence, for example, is approximately 35% in the general human population.² As the number of IT-delivered AAV clinical trials is increasing, the impact of patient exclusion is growing, and this question needs a clear answer. Vono et al. provided the most comprehensive preclinical assessment of this to date, with an overall conclusion that these seroprevalent patients can safely receive these treatments with an expectation of CNS-directed benefits. Some caveats and key unanswered questions remain. Considering that vector biodistribution to peripheral organs is restricted in seroprevalent individuals, if peripheral organ (or peripheral nervous system) transduction is mediating any part of the overall treatment efficacy, then the overall benefit may be dampened in those individuals. Another issue is that variability in NHP studies is typically high, and the publication by Vono et al. is no exception. The study concluded no major inhibition of CNS transduction in seropositive animals (i.e., no statistically significant differences), but the study was not sufficiently powered to address whether more modest (i.e., 2-fold) reductions in CNS biodistribution occurred. The next logical question is whether subjects could be re-dosed by an IT route of administration. Vono et al. still achieved substantial AAV9 biodistribution to the CNS in animals with TAb serum titers of >1:40,960, which might reasonably represent the level of TAbs following IT administration of AAV9. Thus, the tentative answer to the redosing question suggested by this study is yes, with the very large caveat that modest but meaningful reductions in CNS biodistribution might occur, which this study was not sufficiently powered to detect, and that a redosing paradigm should be tested directly in future preclinical studies. Overall, even with these caveats and unanswered questions, the results of this study represent good news for patients that stand to benefit from IT-directed AAV gene therapy treatments.

Declaration of interests

The author declares no competing interests.