How human iPSC models are revolutionizing gene therapy safety testing: an interview with Saqlain Suleman
In this interview, Saqlain Suleman, a lecturer at Anglia Ruskin University (ARU; Cambridge, UK), dives into how human induced pluripotent stem cell (iPSC) models are revolutionizing gene therapy safety testing by offering an alternative method to traditional animal models. He also discusses how the hInGeTox model, which employs human-derived iPSCs and derived hepatocyte-like cells, works to dissect the factors that contribute towards building a comprehensive picture of the genotoxicity of a vector, ultimately aiming to de-risk vectors for downstream use.
Saqlain Suleman joined ARU in July 2024 and is currently a lecturer in biomedical science. Prior to his current role at ARU, he completed his Bachelor of Science (Hons) and PhD at Brunel University of London (UK). During his Doctorate, in the group of Michael Themis, he specialized in developing the first human-based model for understanding lentiviral vector interactions with the human genome in contributing towards genotoxicity. Since then, Suleman has worked with TestAVec (Berkshire, UK), a gene therapy company, and completed his postdoctoral research on the effects of sphingolipids on Friedreich’s ataxia.
Following his postdoctoral role at TestAVec, Suleman took on a faculty role managing various research projects and facilities before moving to ARU for his lectureship and as an independent group leader in gene therapy. Here, he developed his research in reducing gene therapy bottlenecks further at all stages of the production pipeline, from improving vector production and purification methods, understanding contamination and safety and was recently awarded a UKRI SMART Engineering Biology award to develop the next generation of CAR-T cells for cancer immunotherapy.
Can you explain why predicting AAV integration sites in humans has been challenging with traditional animal models?
AAV integration has been difficult to detect in humans for a few reasons. The first is due to limitations with sequencing technologies. Due to repetitive regions and concatemers forming hairpin regions, it is difficult to sequence insertion sites accurately and to a high degree of confidence. Furthermore, as AAV was thought to be a non-integrative vector researchers were initially not looking for insertions in the human genome, to move forward from the dangers of insertional mutagenesis seen with retro- and lentiviral vectors. The other main factor is that while we have seen an increasing number of papers showing AAV integration, this remains relatively low. Due to the low frequency of AAV integration, these rare events are more technically difficult to quantify.
How does the hInGeTox model function, and what specific advantages do human iPSC-derived hepatocyte-like cells offer for studying vector-genome interactions?
The hInGeTox model works to dissect the factors that contribute towards genotoxicity of a vector. We understand from the vector architecture as well as clinical studies, that viral vectors, while efficient at gene transfer, can cause dysregulation of the host genome contributing towards cancer. There are multiple reasons for this, the main ones of which are analyzed using the hInGeTox model. This works by analyzing the insertion sites of vectors in the human genome as well as looking at any clonal outgrowth of insertion sites over time to associate these with cancer genes. The model also looks at transcriptomic analysis where the gene expression profile of gene therapy treated cells is compared to the transcriptome of a range of cancer cells to identify any shift towards oncogenesis. The RNA sequencing also allows for the identification of fusion transcripts between the vector and host which are known to contribute towards genotoxicity. Beyond DNA and RNA technologies, hInGeTox also employs epigenetic analysis using methylation markers. Through looking at DNA-seq, RNA-seq and epigenetic outputs, the model helps build a comprehensive picture of the genotoxicity profile of a vector with the aim to de-risk vectors for downstream use.
The model employs human derived iPSCs and derived hepatocytes which have been characterized as close to the transcriptome of primary hepatocytes. These liver-like cells also present a non-cancerous transcriptomic background, providing a naive cell state to asses genotoxic risk. This not only provides a human based in vitro platform to provide relevant results for gene therapy products but also provides a versatile platform for target cells given the iPSCs can be differentiated into target derivatives and provides an avenue for more personalized medicine.
What does it mean when AAV is described as “non-integrative,” and why has this been challenged recently in the field?
AAV was initially thought to be an episomal vector, so the DNA transferred by the vector to target cells would remain in the cytoplasm. While this is true of other vectors, such as adenoviruses, this also means that these vectors do not allow permanent gene transfer as the episomal construct is diluted as the cells divide. This was initially thought of AAV. However, recent studies for example Chandler’s canine model of AAV genotoxicity and recently a paper identifying AAV insertion sites in humans, has challenged this and we have seen that AAV can integrate, though at not as high a level as established fully integrative vectors such as HIV-1 lentiviruses. However, the process of integration is still not fully understood.
In your study, it was proposed that AAV integration isn’t random but involves a “tethering” mechanism. Can you explain what that means and what led you and your team to this finding?
This work, in conjunction with Professor Themis and TestAVec, was performed as AAV integration is not fully understood, we sought to understand a potential mechanism for this process. To briefly summarise, here we looked at the predicted transcription factor binding sites (pTFBS) that are present in the AAV genome, particularly the 5’ ITRs, and compared these to the pTFBS identified around the insertion sites after AAV gene transfer. Here we found the majority of sites to be similar and significantly enriched. The tethering model is proposed with other vectors, such as HIV-1 lentiviruses, where transcription factors bind to sites in the viral genome and guide the vector genome to common sites in the host genome. This work suggests a partial mechanism for AAV integration, the first step to elucidating this process.
Disclaimer
The opinions expressed in this interview are those of the interviewee and do not necessarily reflect the views of RegMedNet or Taylor & Francis Group.
