Understanding the Many Questions Gene Therapies Raise

Gene therapies have emerged as an important and growing area of medicine, but various players in the healthcare continuum are trying to understand the unique development, regulatory, and other issues surrounding this emerging modality. Avery McIntosh and Alex Sverdlov, both biostatisticians, have edited the new book Development of Gene Therapies: Strategic, Scientific, Regulatory, and Access Considerations, a reflection of their efforts to understand the complex of considerations the advent of these therapies raise. We spoke to McIntosh, director at Pfizer, and Sverdlov, senior director of statistical analysis at Novartis, about their new book, how a pair of biostatisticians view the challenges of gene therapy development, and why these therapies don’t easily fit into existing models.

Daniel Levine: Alex, Avery, thanks for joining us.

Avery McIntosh: Thank you, Danny. Great to be here.

Alex Sverdlov: Thank you, Danny.

Daniel Levine: We’re going to talk about gene therapy development, the challenges around developing what’s still a relatively new modality, and your new book, Development of Gene Therapies, Strategic Scientific, Regulatory and Access Considerations. We talk to a lot of executives and scientists on this show, but we don’t get a chance to talk often to biostatisticians. For listeners who are not familiar with the role biostatisticians play in drug development, perhaps we can start there. Can you explain what biostatisticians do?

Avery McIntosh: Sure, Danny. So biostatisticians are trained in a few domains, the first of which is design of experiments. So that means setting up experiments that operate within certain constraints and operate efficiently to arrive at a scientific conclusion. We’re also experts in data science, data visualization and presentation. And that means we help coordinate the collection of data and the validation of data and interpretation and presentation of data from human clinical trials. And that crosses over into multiple audiences, which might be internal sponsor governance boards, medical societies, presentations for the investor community and official sponsor communications. We also co-lead interactions with data monitoring committees, also known as data safety monitoring boards, that monitor the safety of products during ongoing trials. And we also bring strategic expertise to the overall drug development process by co-leading the clinical development plan for investigational assets and using statistical and scientific principles to help move products forward efficiently.

Daniel Levine: It’s interesting to see this book as an effort led by two biostatisticians. How did the project come about and what led the two of you to put this book together?

Alex Sverdlov: The idea to develop a book came to us in 2021, and at that time, Avery and I worked together in the same company and we witnessed the development and advancement of the field of gene therapies and we acknowledged the challenges and opportunities that this area brings to us. So we are biostatisticians, and by writing this book, we wanted to highlight the pivotal role of quantitative scientists very generally, and specifically biostatisticians in the development of gene therapies. And as Avery said, it is a multi-faceted work. The biostatisticians design clinical trials, they select and perform the most appropriate data analysis. They partner with various line functions in the process to enable science-driven and evidence-based decision making. In fact, Avery and I wrote together a paper on gene therapies that was published in Clinical Pharmacology and Therapeutics in 2021. So we thought that it would be really great to expand it into a full length book on the subject. And so the idea was to develop a book that would provide a useful reference and a navigation tool for scientists, drug developers, healthcare commissioners, educators, and many other stakeholders. And we chose the format of the edited volume to have various subject matter experts work together on specific topics and chapters on this book. So, it was a very fascinating team effort and experience, and we learned a lot ourselves in the process and we expanded our professional networks. It took three years in total, but it was definitely worth it.

Daniel Levine: For rare and ultra-rare genetic diseases gene therapy holds much promise, but these conditions, these therapies can present development challenges like any modality. How unique are these issues when it comes to gene therapies?

Avery McIntosh: Well, gene therapies are generally for rare diseases, and there’s a couple of investigational gene therapies right now under development for more common indications. There’s at least one for Alzheimer’s disease and another for age-related macular degeneration. So the development of a gene therapy product does proceed under some of the same constraints imposed by rare disease drug development generally, which I’m sure many of your listeners are familiar with. And just to highlight some of that, there’s small patient populations from which to draw for clinical studies and sometimes for real world evidence. The patient populations can be geographically or maybe organizationally dispersed, and there’s often a lack of robust natural history to characterize the patient treatment journey. And so then when you’re running a clinical trial, devising the right clinical endpoints to meet the regulatory and payer and patient expectations can be a major hurdle. Obviously from the financial perspective, rare diseases are challenging to develop as by definition there’s a smaller patient population who can be treated by a drug. But what makes gene therapies more unique, I think, in this already unique setting of rare diseases is there are some issues that have to do with the therapy itself, the gene therapy that make it particularly complex. And we can get into some of these maybe more detailed, but just in a nutshell, some of them revolve around dose finding and the challenging nature of finding the correct dose for the right patients when you have a product that’s potentially immunogenic and can usually only be dosed or administered to a patient one time. And there’s also complexity around blinding and randomizing subjects in a controlled trial setting and especially safety monitoring of these products. Those things really make it unique.

Daniel Levine: So as biostatisticians, how do you think through those issues?

Alex Sverdlov: Yeah, that’s a great question. So we actually have one chapter in the book, chapter 10, where we outline seven hallmarks of clinical development of gene therapies, and I can give you some exposition to that. So first of all, the gene therapies are disease modifying. They target the root cause of the disease, not just alleviate the symptoms, and what this means is that they should be given as early as possible when the disease is diagnosed and identified, ideally in the childhood. Gene therapies are usually for rare diseases, and that means that drug development is a bit more challenging compared to more common diseases. And by the way, by rare disease, there’s no universal definition. And although rare diseases are by themselves rare, but collectively they’re not, there’s so many of them. Then gene therapies are administered only once. So many of these drugs are actually just, well, they need to get administered at the right dose and make sure that we have only one chance for the patient. That means that it’s more like a solid organ transplant rather than a pill that a patient may take every day. And that also means that studies of healthy volunteers are not run with gene therapies because the drug is only given to the patients. So there’s no consensus on what endpoints would measure pharmacologic activity. That really depends on the specifics of the gene therapy, but what it means in practice is that a lot of work is required to measure what is happening in the body and what happens when the drug is given to a patient. The safety monitoring is quite challenging because different components of the drug can lead to different potential safety issues, so monitoring is very important. And then challenges related to the ethical issues. So, for example, the use of placebo in the trial to compare gene therapy versus standard of care or the placebo, that’s quite difficult to justify. And so, because the diseases are usually severe and rare, patients may not be participating in the trial, and that might not eventually lead them to receive the effective treatment. And last but not least, the long-term safety and efficacy is still difficult to predict. So that’s why health authorities like FDA and European Medicine Agency in Europe, they require that patients who are given the gene therapy, they must be followed up for a period of five to 15 years to assess the long-term safety of these drugs.

Daniel Levine: The book, it really takes a broad view and does bring in experts to really give a three dimensional sense of everything from scientific to regulatory to ethical and issues of access. The two of you are both contributors to several pieces throughout the book and you collaborated on a chapter together that has to do with statistical design and analysis for gene therapies. You make the point that development of gene therapies is distinct from other modalities. And Avery, you touched on this a moment ago, but why doesn’t it work to develop a gene therapy the way you would develop other medicines?

Alex Sverdlov: Yeah, that’s an excellent question. So I hope the seven hallmarks that I just alluded to give some initial perspective. So I would add one more item. The gene therapies, they offer a more personalized approach to treatment by addressing specific genetic variations in each patient, whereas more standard conventional treatment might have a broader application, but like this individual tailoring. So it makes it very special and unique. And yeah, the toolkit that is used in the development of gene therapies may be very special.

Avery McIntosh: I could just add a small point to that as well. This has just been on my mind lately. So, the clinical development of these products is often quite condensed compared to traditional pharmaceuticals, low molecular weight, and also even antibody drugs where you might have a phase 1 trial in healthy volunteers, and then a phase 2 proof-of-concept in patients, and then proper dose finding. And then a large pivotal, adequate, and well controlled phase 3 trial for the first three approved, at least three approved gene therapies that came in 2017 and 2018, that would be Kymriah, which some people call a cell therapy, but we’ll just put it under the umbrella of gene therapy, and Luxturna, and Zolgensma. These all led to accelerated approvals based on early phase studies, and there were later phase studies that were required by regulators, but that’s not something that you typically see with other disease modalities. So when it comes to preparing the evidence package for regulators and getting ready for the access issues that come with gene therapies, all of that has to be really front loaded in the program in a way that’s pretty unique compared to other drug modalities.

Daniel Levine: You do, in the book, propose strategies and approaches for gene therapy development. What role do you see real world evidence playing in this?

Alex Sverdlov: Yeah, that’s a great question. So real world evidence is playing an increasingly important role in general in drug development, and it suddenly has a pivotal role in gene therapy development. So, the gene therapies are intended to delay, slow down, or hold the disease progression. It means that the investigators should have a reliable disease progression model that would be based on the natural history data in the untreated target population of patients. This model would help investigators enable accurate prediction of individual disease trajectory accounting for patient-specific careers and uncertainties. And then in a randomized controlled trial, there should be evidence that experimental gene therapy significantly alters the course of the disease and has a long-term beneficial impact on the disease progression compared to the control group. So this evidence is actually crucial for regulatory approval and for clinicians to justify the use of a particular treatment for that patient.

Daniel Levine: I think in a lot of the gene therapy concerns, we’ve seen the patient community and even the developers have been focused on the durability of these therapies, but the reality is there’s also a concern about the fact that you are potentially making a long-term or permanent change. How do safety concerns around this affect trial designs?

Avery McIntosh: Yeah, yeah, that’s a great point. They do, and gene therapies, as you said, make potentially permanent changes to human gene expression, and that’ll be mediated depending on the type of gene therapy you might have, if it’s a gene insertion or it’s CRISPR gene editing or what have you. And unlike these low molecular weight and other biologic drugs, these products are not metabolized by organ or cell mediated clearance mechanisms. So that means, as I said, you really have to get the dose right. And if you underdose a patient in a first in human trial, that patient may not be able to be dosed from the same drug or a similar gene therapy in the future due to the immunogenic nature of the products. And if you overdose a subject accidentally during the dose finding portion of an investigational phase, that subject could have permanent overexpression of a therapeutic protein, which could be toxic. And there’s also the risk of an overactive immune response there. So the safety monitoring and assessment of these products in the early stages is really important and has to proceed in a very stepwise and very careful manner. And biostatisticians in particular can aid in the safety monitoring in these settings by helping to benchmark the rate of expected adverse events from the specific gene therapy modality and using different statistical techniques to help adjudicate if a particular adverse event is drug related or not. Sometimes that’s obvious, but sometimes it’s not. And one of the unique aspects of gene therapy drug development is that sponsors are required to follow patients administered a gene therapy for between five and 15 years, depending on the risk profile of the specific product. And that’s really an unprecedented duration of engagement with patients and caregivers and sponsors. And it can be challenging for sponsors, but there are some really interesting and unique—really opportunities for safety monitoring because of that mandate. I’m thinking of a town hall the FDA had last year where the World Federation of Hemophilia gene therapy registry was discussed as a model for safety assessments across pharmaceutical sponsors. And other industry analysts have called for the advancement of platform trial methods in this area, and that could lead to the ability to answer very unique and difficult-to-answer scientific and safety questions around these products. So developing platform trial infrastructure and guidance for long-term follow-up studies for gene therapies has been a particular interest of ours. And I and colleagues at Pfizer recently published a paper on this topic in the Journal of Clinical Pharmacology and Therapeutics. So there’s definitely going to be movement in this area in the coming months and years.

Daniel Levine: The small size of many rare diseases, ultra-rare diseases, not only add to the challenges of developing a gene therapy, but really raise questions about the commercial viability. Do you think we’re at a point where we need new models for developing these types of therapies for certain diseases with small populations?

Avery McIntosh: Well, I think there’s room for industry to pursue rare disease drug development, but it really requires a concerted effort from a multi-stakeholder perspective. So there’s already some good examples of collaboration between academic partners and industry and nonprofits and creating the infrastructure to develop gene therapies for rare and ultra-rare diseases can be done. But of course, it requires the support of the patient advocacy community and really crucially, regulators like the FDA and EMA.

Daniel Levine: Another issue with gene therapies that you touched on early, particularly with the use of a viral vector, is to create an immune response. This makes the issue of dose escalation challenging because these are often patients that will not be able to receive a second dose, they’re participating in a clinical trial. You want to see the patient get benefit, but if you’re not giving a high enough dose, they’re not going to, if you’re giving too high a dose, you’re risking an immune response. How do you balance these issues when it comes to determining the proper dose of a clinical trial of an untested gene therapy?

Avery McIntosh: Yeah, you really hit on a pivotal issue. First and foremost, what this means is that the preclinical models and the translation from concept to human dose potential has to be extremely rigorous, both for safety and for efficacy. So the right cells have to be targeted at the right dose for the right patients at the right time, and the disease course, as Alex mentioned. So this is a really delicate balance, and it is a quantitative exercise, right? So when decisions are made based on data, biostatisticians should be there and other quantitative experts such as pharmacometrics and clinical pharmacologists. We do have a chapter on this topic in the book with some experts, Page Bouchard, Emily Messick, Deepa Chand, Francis Tukov, and others. That’s the preclinical side. And on the clinical side, there’s some really innovative and advanced methods to perform quantitative dose escalation that are used pretty routinely now in the oncology space. And we hope that these methods can be retrofitted and maybe adapted a little bit to the gene therapy space with just a few tweaks to how these models operate. So to that end, I have a paper under preparation on this topic with some colleagues, and Alex and myself are both members of the American Statistical Association Cell and Gene Therapy working group. And that group also has a paper on this topic in a forthcoming special issue of The Journal of Biopharmaceutical Research. So more will be revealed on this. So stay tuned.

Daniel Levine: Well, the ultimate question for biostatisticians in rare disease drug development, it’s always an issue, but I think with the gene therapies we’ve seen, which tend to be even smaller patient populations, how do you power it? So, there’s what would be called statistical significance, and does that term have meaning within the context of a rare disease gene therapy or an ultra-rare disease gene therapy?

Alex Sverdlov: Yeah, and that’s another excellent question. So the general rule is that we cannot sacrifice quality and standards in rare disease development. We need to be really creative and innovative. So there are various options that one can pursue in consultation with health authorities, of course, but there are different novel adaptive designs, there are approaches to analysis. So there are some developments ongoing on development of endpoints that capture the disease progression. And great deal of collaboration is actually ongoing between industry health authorities and academia where we are trying to develop novel clinical endpoints that may involve both patient reported outcomes, quality of life metrics, and other markers that can be captured to demonstrate the benefit, the development of the digital biomarkers, real world digital biomarkers derived from biosensor data, wearable devices such as active devices. They can be quite useful in this space. But I think, I guess, it’s not just one method, one approach. So we really have to be clever and creative and use state-of-the-art technology tools and state-of-the-art statistical tools to develop these methods. And last but not least, of course, the development of statistical methodology research that holds promise and must be used in practice very soon.

Daniel Levine: You have a chapter on statistical innovation for gene therapy development and propose a roadmap. You use a hypothetical case of a gene therapy for a specific form of the neurodegenerative disease ALS, but could you use that example to walk through the elements that are applicable to other gene therapies as well?

Avery McIntosh: Sure, Danny. So, we spent some time discussing this disease, and this is SOD1 mutation ALS. So amyotrophic lateral sclerosis is one of the more well-known rare genetic neurological diseases, and it involves degeneration of the upper and lower motor neurons and a diagnosis it can present in the arms or facial region or the legs. And it’s, on average, fatal within two to four years of diagnosis. And most cases of ALS are of unknown cause, but between four and 10 percent are thought to be caused by single gene mutations, which make them really attractive targets for gene therapy interventions. And among the genetically defined ALS population is ALS attributable to mutations in the SOD1 gene, which is on chromosome 21. And this has been, as I said, a target for gene therapy, but also for antisense oligonucleotides from a number of sponsors. And that’s really due to the well-defined nature of the disorder and that it’s uniformly fatal and very fast progressing, unfortunately. So, for the chapter, we imagine a hypothetical gene transfer or knock-down or entity modality, and we keep all of that and the root of administration a bit unspecified in general, but just walk through the elements of what a program like this would look like. And of course, for something like this, you have to have really years of substantial nonclinical testing involving chemistry, manufacturing controls, and in vivo and in vitro testing being performed to arrive at a lead candidate that is the most promising for clinical development. And from there, we delve into some of the preclinical translational studies that would be needed for a product like this that really focuses on synthesizing the preclinical evidence into human dose translation and helping the broader clinical pharmacology team characterize what we think the biodistribution profile from animal studies can tell us about what we might see in humans. Right. So, for ALS, this affects, as I said, both the upper and lower motor neurons. And a drug would have to really transduce both of those tissues in order to be broadly effective and in the absence of adequate transduction to those specific tissues, it would really be important for the clinical team to understand and try to model what kind of symptom relief might be expected from an imperfect drug because gene therapies are often touted as kind of a cure, but that may or may not be possible depending on the nature of the disease and the scientific tools that we have available. For this disease, even a moderate relief of symptoms and survival benefit could really be enough to pursue a drug candidate, given that there’s really so many limited treatment options for ALS. So from there we talk about use of real world evidence and clinical design and lifecycle management. I know I’ve talked for a while, Alex, do you want to add anything to that?

Alex Sverdlov: I think it’s a great summary. I think in general, an important point to consider is that we need a systematic approach for the design and optimization of clinical programs, gene therapies and the modern computing tools. Nowadays we have a way to evaluate using computer simulations or in silico different clinical development options under various experimental scenarios, under what/if scenarios that may include cases when the drug is ineffective and the cases when there’s great efficacy. And there are different translational steps, of course, in between that we need to bear in mind. But this in silico optimization, it can help us obtain development options that lead to very robust and powerful clinical plans for a given study. And so this is really a great opportunity for statisticians and other quantitative scientists to add value to the clinical development team. So even before going into the experiments, we just test the concept using computer simulations and then recommend the best options.

Daniel Levine: The book is Development of Gene Therapy, Strategic Scientific, Regulatory, and Access Considerations. Avery McIntosh, book co-editor and director at Pfizer, and Alex Sverdlov, book, co-editor and senior director of statistical analysis that Novartis. I’ve got to say, as academic as the book may appear, it’s extremely accessible. And it was a great read and very informative and very thoughtful. Alex Avery, thanks so much for your time today.

Avery McIntosh: Thank you, Danny. This has been great.

Alex Sverdlov: Thank you, Danny.

This book has been lightly edited for clarity and readability.

The RARECast podcast is made possible through support from the Global Genes’ Corporate Alliance. The members of the Corporate Alliance support Global Genes’ mission and programs, work to meet the vital needs of people with rare diseases, and address inequities they face. To learn more about the Corporate Alliance or how your organization can become a member, click here.