ASGCT Members Provide Recommendations During FDA Liaison Meeting

ASGCT held its fourth annual liaison meeting with FDA CBER’s Office of Tissues and Advanced Therapies (OTAT) on Nov. 8. In addition to a presentation from OTAT’s director, Wilson Bryan, MD, highlighted on the ASGCT news blog, Society members spoke on three significant topics in the field, followed by comments from OTAT staff.

rAAV Gene Therapy and Peripheral Nervous System Ganglia Toxicity

Presenters: Laurence Whiteley, DVM, PhD, DACVP, Pfizer Inc; and Juliette Hordeaux, DVM, PhD, DECVP, University of Pennsylvania

Dorsal root ganglia (DRG) lesions are recognized as part of the spectrum of AAV-related toxicities observed in preclinical research. In this presentation, Dr. Hordeaux presented on study design considerations and species selection for the evaluation of toxicity in ganglia following rAAV gene therapy administration. Nonclinical findings indicate that transgene expression is important in the induction of neuronal toxicity, but it is unclear the level of expression required to induce toxicity. Because of uneven transduction, collecting samples from several regions is important. Nonclinical findings to date indicate that maximum severity occurs at early time points with subsequent resolution of active neuron injury.

At this time, monkeys are considered most relevant due to the preponderance of studies conducted in this species. Based on current literature, DRG pathology is almost universal after AAV gene therapy in nonhuman primate studies (assuming transgene expression in DRG) and is affected by route of administration, dose, and age at time of injection.

The translatability and significance for humans remains unknown, so more clinical data is needed to understand the relevance of nonclinical findings. Human DRG pathology may not be associated with clinical symptoms, and little understanding exists on background aging-related changes. Clinical monitoring strategies for patients include age-appropriate neurological exam, MRI, nerve conduction velocity testing, and intradermal nerve fiber counts.

ASGCT recommends that the absence of a NOAEL (no observed adverse effect level) for DRG toxicity should not preclude clinical development in the context of appropriate risk/benefit considerations and incidence/severity of DRG findings. Nonclinical risk assessment for AAV DRG pathology when intended to treat disease indications that are considered severely debilitating, life-threatening, or have high unmet medical need may be based on the concepts in ICH S9 Nonclinical Evaluation for Anticancer Pharmaceuticals.

In discussion of this topic, OTAT reviewer, Mercedes Serabian, MS, DABT, noted the shortage of nonhuman primates (NHPs) as a problem, and Dr. Whiteley added the uncertainty that NHP studies are predictive of clinical outcomes as a challenge. Both acknowledged, however, the lack of a good alternative. Dr. Bryan indicated NHP studies are not currently the default, that sponsors need to propose what they think is appropriate in their INDs, which varies by IND, and provide a scientific argument for their product and indication, and OTAT will consider it.

Integration/Insertion Considerations for AAV-Based Gene Therapy Vectors

Presenters: Kevin Eggan, PhD, BioMarin Pharmaceutical and Markus Grompe, MD, Oregon Health & Science University

Dr. Eggan explained that the content of this presentation arose from a roundtable of multi-stakeholder experts in the field that ASGCT convened in August to discuss AAV integration, findings from nonclinical research, and implications for drug development and clinical trials. This talk also considers discussion from FDA’s September Cellular, Tissues, and Gene Therapies Advisory Committee.

While recombinant AAV (rAAV) can integrate into the host genome at low frequencies, the risk of oncogenesis associated with rAAV chromosomal integration in humans is theoretical based on current data and information. Several challenges in assessing potential risk include that estimates of integration frequencies remain highly variable, illustrating a need for better assays to quantitate and characterize integration events in nonclinical and clinical samples. ASGCT finds value at this time in performing integration-site analysis with existing methods on tissue samples from patients, and WGS on tumor samples from nonclinical and clinical studies should they occur. ISA, WGS and gene expression analysis of any malignant tumors should be performed. Limitations to these methods may be overcome by rapid and ongoing improvements in sequencing technologies, such as target enrichment and long-read sequencing methods, which are likely to provide greater insight into concatemeric and highly rearranged integrated vector genomes. Cell-free DNA (cfDNA) should be further researched and validated in the field as a potential safety biomarker.

Regarding the translatability of various models to humans, the use of sensitive mouse models of oncogenesis may provide information about the relative risk of different vector designs. However, the translatability of any murine oncogenic events to humans remains unclear. Recent observations of clonal expansions in canine hepatocytes in the absence of tumor formation warrant additional studies in the field, but they may reflect natural clonal dynamics in aging animals. Further research is needed on the development and potential predictability of in vitro models on integration and associated risks, but limitations in primary hepatocyte culture currently are a substantial barrier to the development of such models.

While additional studies assessing the oncogenesis of strongly transactivating vectors in large animals will be needed to better understand these issues, these studies are impractical as a near-term solution for specific development programs and may require one to two decades of research. ASGCT’s view is that given the current diseases targeted with AAV gene therapy, and the actual documented risks, the risk/benefit ratio would suggest that trials should be able to proceed in parallel with additional investigation and in the post-market setting. This is particularly true for disorders in which the liver is not the target.

Product labeling and risk management plans will be critical elements of communicating risks, as well as communicating the relative risk/benefit equation to patients and their physicians. Monitoring should focus on non-invasive methods such as ultrasound, liquid biopsies, and biomarkers through blood work. Biopsies of neoplastic nodules and surrounding non-tumor tissue should be conducted to further investigate any positive signal picked up in non-invasive methods. Duration of long-term follow-up monitoring should be risk-based and depend on the product-specific risk factors (e.g., promoter, route of administration, dose) and patient population’s background risk, including age and presence of chronic liver regeneration.

During the discussion period, Dr. Bryan noted that because of being so early in the development of gene therapy, sponsors need to make a proposal on risk assessment and justify it, and OTAT will consider it. He reflected that development can go forward with the uncertainty surrounding this topic in serious, life-threatening, rare diseases, but as the field moves forward for diseases that are less severe and more common, the agency may become less comfortable with the uncertainty.

Recommendations on CMC Expectations for Gene and Cell Therapy Products

Presenters: Jan Thirkettle, PhD, chief executive officer, Transine Therapeutics and J. Fraser Wright, PhD, professor of pediatrics, Center for Definitive and Curative Medicine, Stanford University School of Medicine

Due to their greater complexity than small molecule and protein therapeutic products, analytical characterization represents a particular challenge for cell and gene therapy (CGT) products. ASGCT’s position is that rapid innovation in the CGT field warrants a CMC framework that remains flexible, risk-based, and correlated with the extent of clinical experience.

Dr. Thirkettle discussed realistic expectations for specification tests vs. characterization tests. Because the overall objective of product developers is to improve product quality and clinical potency over the course of clinical development, ASGCT suggests:

• Focusing early-stage specifications on the limited number of well-defined CQAs at that stage.

• Allowing characterization assays performed throughout development to inform the product specifications applied in late-stage development.

• Using available data and science-based risk assessment to guide the evolution of specifications at the appropriate point of development (including post-marketing, if so justified).

Developing in vitro potency assays which mimic closely the mechanism of action of CGT products is an important but quite challenging goal. Development of such assays suitable for a QC/GMP environment is a significant undertaking that may take many years. ASGCT recommends FDA allow a risk-based approach in the context of the unmet medical need of the patient population. The development of one strong potency assay addressing the main mechanism of action(s) of the final drug product should be sufficient for product release. For example, FDA could eliminate requirements for AAV infectivity or potency testing of rLV when these vectors are used to create a final genetically modified cell product which has its own potency assay. The Society suggests allowing continuous validation of potency assays during review and post-licensure to refine acceptance criteria for products where high replicate batch data is challenging (e.g., autologous products).

Implementation of some complex characterization assays and validation in a GMP environment is not possible due to compliance challenges. ASGCT suggests that FDA develop broad, high-level guidance for the CQAs that should be the focus of characterization studies for each product class (i.e., rAAV, rLV, CAR-T cells). The focus of performing characterization assays should be to gain a deeper understanding of product CQAs during product development, while the focus of performing quality control assays to ensure lot-to-lot consistency of known product CQAs within acceptable ranges.

Statistical analyses in setting comparability acceptance criteria is often not realistic given the lack of sufficient batch numbers that can realistically be generated during clinical development and insufficient understanding of the clinical impact of biological variation. ASGCT’s position is that it should be acceptable to use characterization assays without acceptance criteria to provide relevant but difficult-to-quantify data to inform the overall assessment of comparability for attributes that are not expected to impact safety.

In the discussion, Dr. Bryan expressed that this is a particularly excellent presentation on what is being looked at as a field, and that OTAT recognizes the difficulty of CMC challenges. He noted challenges include the rapid pace of the science of gene therapy leading to early efficacy (e.g., phase 1), and the inability to use statistics on very small trial populations. He acknowledged that while addressing CMC before getting into the clinic is ideal, academics do not have the resources to do so, and some questions about CMC cannot be answered until a product gets into the clinic. He concluded that OTAT therefore needs to make judgements about risk tolerance.

Related Articles