Navigating Early Clinical Drug Development Challenges by Dr. Sara Glickstein of Genentech

Good afternoon, everyone, and thank you for joining us today for the Emory Pharma guest speaker series. So my name is Janet Liu. I am the director of Biology here at Emory Pharma. We are a contract research company and we service a lot of early stage pharmaceutical development for both small molecule and large molecule drugs. We're joined today by our speaker, Dr. Sarah Gleckstein. She is a senior early stage clinical scientist from Genentech. Her area of focus includes early clinical trials aimed at respiratory and fibrotic indications. She got her PhD from Cornell University in neurobiology. And before she joined Genentech, she spent a decade working as a medical writer and as a strateg, and she also held faculty positions at Cornell at Columbia University. So without further ado, look forward to your presentation.

Thank you so much, Janet. All right, so really thank you again for having me. And to have you guys here on a Friday afternoon interested is really the statement about your company, which is pretty awesome. So I'm going to talk to you today about early clinical development. And we work a lot with companies like yours in sort of executing these trials. But I work on the clinical side. And what I'm going to talk to you about today is sort of what's our overall strategy in approaching clinical development of a drug? How do we design a clinical trial? And really what we see are the gaps in what we need to do in the future to do this even better than we do now. And so let's start with strategy and clinical development. And so you guys all know this. You've probably seen the funnel, right? Clinical development is lengthy, expensive, and it's risky. And so it can take like 10 years to get from your phase one to your approved drug, usually more. It can be expensive. 2.6 billion is an average. Most drugs are going to fail. Only about 12% entering into the first in human trial actually make it to be ultimately approved. And when I started in clinical development, I thought, well, if you make it through phase one, you're like partway there. You're a third of the way there. That's not the case. Right. If you can make it through phase one, there's no guarantee you're going to continue even if you make it through phase two. Unfortunately, I just put up this is one analysis and I'm not sure of exactly how they did their modeling, but it's just showing you that the overall chance of success is so low when you consider any one sort of gait from phase one to phase two or phase two to phase three. And it really varies by the type of disease you're going after. And it's pretty dismal. But the importance is there's a lot of unmet need out there. And so when we have a new drug coming into early clinical development, we really need to define the strategic context and the target product profile. And that happens from the get go, right out of the gate. And so the strategic context is the foundation for clinical development. And we really have a vision for what this drug is going to be. Is it going to be the best in disease, first in class? What is the differentiator from the standard of care or from the competition that is worth all of this time and investment and this high likelihood of failure? And we have a core strategy, we consider what's the entire standard of care for that indication and the competition? Are there additional indications really? How do we sequence these in our clinical development plan? And what are other features that might be required in order to use this drug in the clinic or might be advantageous based on this molecule? Is there a companion diagnostic or a specific segment of the population? Is there something we really want to see listed in the drug label so that the drug is used in a very specific way and so the target product profile? I don't know if all of you have ever seen one of these. Most of you have. We really define what we want that drug to look like on paper and we speak about the indication. So what's the disease area but within that indication, what's the patient segment? Do they have a specific genotype? Do they have a specific physiological parameter that would qualify them for use of this drug? For efficacy? We really want to pick out approvable endpoints, but also those we want to understand what bar we have to reach to be better than competition or standard of care, what's going to really improve patients lives? Safety. We want it to have an acceptable safety profile. We really think about risk, benefit. That's one of the major discussions we have in terms of safety and tolerability. We also think about our competitors or standard of care. If we have a drug that's just as efficacious but has a reduced patient burden, that's definitely a benefit to patients. And that might be the strategy where we're aiming for strategically in terms of dose, we have to think about the formulation, the strength and the frequency. If we have a drug that's oral and it's these big horse pills and it's for pediatric indication, that'd be problematic. And so our TPP would define what dosing would be appropriate for that population. We talk about route of administration and other sort of variables like health care measures and market access. What sort of features might this drug offer? And so we define what is the minimal profile, and that's a viable profile. It's like if it just got that, okay, that would make it viable. But we also define an acceptable and an aspirational, and that helps us with making key decisions on how to develop that drug. And we use this as a way to really hone in on where our gaps and really make sure we're always sort of getting to where we want to be ultimately. And so we also, very early in development, it's like a game of chess. We form the clinical development plan. So when we start in phase one, we sort of project how we're going to get to phase three. And we don't just consider our drug, we consider any other drug that might be with Genentech's a large company, so we consider what other drugs. When are they going to read out, how's that going to inform decision making, but also external data, when are our key competitors reading out and how is that going to impact our investment? And you know when to sequence in additional indications. And you really are kind of having a lot of these gates and you think about them in advance, so you know exactly when you're going to want to trigger that switch or make that change. And there's a lot of, at a company like Genentech, there's a lot of people involved. Even at smaller companies, you have people with many different types of expertise to track what's going on externally, to really understand how to design these trials, to be efficient and quick. And so let's move on to clinical trial design. And so in early clinical development, what we're moving towards is uncertainty to more confidence in our understanding our data. So that's all across development. But in early clinical, we're kind even this middle region where we're going from uncertainty into a little bit of confidence. Early clinical encompasses phase 1 and phase 2 trials. Phase 1 trials are safety studies, usually not a lot of patients. Phase 2 trials have more patients, still not a huge number of patients. And they're really designed to see if there's any activity or efficacy and to understand the dose. When we're designing these trials to move into that area of confidence, there are many things we have to consider. One is the primary objective and that is really that might change depending on the type of trial. And we'll talk a little more about that, whether it's an FDA Approvable endpoint or it's something we're doing just to understand a little more about the drug and its activity. Study population, Is it healthy volunteer? Is it patients? Are we enriching for any specific feature among these patients? What's the number of subjects? How are we going to randomize them? Is this a controlled study? Are we blinding? What are the doses? Are we testing a dose range? Is there an exposure cap? We need to consider group design. Is this a parallel study or sequential study or crossover? Are patients coming in and being sent into arms or do they change throughout the study? How flexible do we want the study to be? And then our endpoints again, our primary, secondary, what are our key endpoints? What is the health authority guidance for that indication? What is going to be most feasible and efficient to run, which is a big question. What is the required safety data and monitoring required for this trial? What is our desired label? Are we developing a companion diagnostic? And there are some differences between biologics and small molecules in terms of how you think about all of these designs. One of the main ones is whether or not you need a food effect cohort, which you need for oral drugs, or do you need to test for anti drug antibodies, which are really critical for biologics. And so at all levels, when we consider the study population, it's really a risk benefit question in development. You're exposing patients to some unknowns. And so they're questions the patient is going to decisions the patient's going to be making about benefit risks. And ultimately, when you're seeking approval, providers and the FDA will be making similar calculations based on individual patients or groups of patients or the whole population. And one of the first questions we face is patients are healthy volunteers. And it really depends on the drug and on your intent. But you really have to make sure the benefits outweigh the risks. So phase one is for drugs being tested or developed for chronic diseases. Oftentimes we use healthy volunteers because they should be safe enough to test on a healthy volunteer. Whereas drugs for cancer, which are often toxins, you don't want to put a healthy volunteer in a trial like that. But there's some cases when our drugs meant for chronic indications, it's too risky for a healthy volunteer. In terms of the indication, you typically have some preclinical data that is going to direct you. But nowadays we have these drugs which might have benefit in a wide variety of indications. And so you'll also consider what is going to be the most feasible, how would you sequence these and where is the unmet need in terms of eligibility, what will allow for the best evaluation of the primary objective, what is the least restrictive label? And so you really try to have a broad of a patient population as possible in the trials and you try to balance the experiment you want to run with, to test the hypothesis with what's the opportunity for patients. And you know, you determine the number of subjects. And for those of you trained as scientists, you'll be shocked to know that we don't always enroll enough subjects to run the best experiment. And so, you know, these are all questions that we have to determine in terms of dose. You know, we really have to understand the pkpd, the therapeutic index, which is that range between when you have a therapeutic effect and a toxic effect of your drug over a range of doses. And you want to understand like what doses can, if there is an adverse event, can we monitor and mitigate for any safety issue? And there's often you have to think about your dosing objective. In an oncology drug you often want to go for MTD or the maximum dose a patient can take, but for some indications, maybe you just want partial inhibition over time. That would be your objective. And so you really look for benchmarks for molecules that are similar sometimes to understand what's your dosing range you want and you oftentimes dose sentinels to sort of check what you're doing in terms of biomarker strategy. We really want to understand two things. What is target engagement and what is target modulation of the pharmacodynamic effect. And so we have data from our pre clinical work, but you don't know if that drug is inert. That's always a concern when you do these trials. So we want to know that the target's engaged and that we're seeing an activity readout of some sort. Our trials are too small to really see efficacy a lot of times in phase one and phase two, but we want to be able to measure something that shows us that the drug has legs. And so oftentimes in early trials we'll do something called enrichment where we'll enrich first or population that we really think if we don't see that pharmacodynamic effect in those patients, then we'll know that we should be a little concerned. But that oftentimes isn't a great strategy overall because you want your label to be least restrictive. And we often consider companion diagnostics. You really have to have your eye on that from the get go. And these are things that might be coupled with your medication when it gets approved, so that it, it picks out patients who might have an optimal response to that therapeutic and just to explain a companion diagnostic. So one idea is for patients who are most likely to benefit, you can also use it to screen out patients who might be more prone to side effects, and you can use it to monitor patient's response to treatment. But you have to start this work early. And a diagnostic development, really, it's coupled with every stage of drug development. And there's a lot of data from your clinical studies being fed into the diagnostics division as they do these assessments and develop that diagnostic. And so let's talk a little bit about phase one and phase two trials. And so phase one, so that's the primary objective is safety. And the first study we do is called a first in human or a single ascending dose study. And that's where patients get a single dose of the drug and you monitor them for a period of time. And this is showing you just an example of what that might look like. And in this example, we're moving from IV dosing to subcutaneous dosing. And they're really small cohorts. This study, it's eight patients per cohort, so six are treated to a placebo. And I'm going to tell you, when you run a first in human and you know that your patients are going to be dosed the next day, like you, you don't sleep very well the night before. Even though you trust your calculations, you're still concerned, like, what if I hurt someone? What if something was wrong? Phase 1a, you administer a dose, single dose, and you wait for an amount of time and you assess for safety. In pk, the main goal is to evaluate safety and tolerability. And it can be depending on the type of drug, this can go pretty fast or it can go slowly. For a biologic, you have to wait till a certain half life of the drug so they have a longer half life. Can be a month, month or more. And so the goal is to evaluate safety and tolerability, to define what is the maximal dose. And you have to get to a certain dose level to enable repeat doses at a level you need exposure coverage. You want to characterize safety across a range of doses. PKPD relationships, proof of activity. Rarely it's clinical benefit, although in oncology trials, sometimes you get a little more information and you want to confirm or understand preclinical data. And this is not a power trial. So sometimes we'll do these studies with really few patients in a cohort Maybe like three patients, four patients in the drug in oncology. And sometimes people will come in and they'll be like, well, why don't we do more patients? Why don't we do 10 patients per cohort? And the reality is you're not going to see something in 10 that most likely that you don't see in three. You might need to test 100 or 1,000, which you're not going to do in a phase one. Again, as I mentioned, to move from your phase one A to phase one B, you really need to think about exposure coverage. And so we'll typically have to get to a certain dose level, which gives us an exposure, so we can open up a repeat dose. So repeat dose, where patients are getting doses over a number of days. And so the overall objective is to run these trials together so that you're not wasting time, you're not getting to what here would be 80mg sub Q, you are opening up 5mg sub Q as soon as you can and you're kind of sequencing all the cohorts so you can run both studies at the same time. And so the multiple ascending dose study, it really impacts your phase two dose selection and you really are considering what is the risk benefit so that you can understand how long to run those cohorts for. Is it, you know, for a biologic, typically we'll just do one dose, three doses, but depending on the half life, it's very often one dose every month or a dose every two weeks. And you really want to look at steady state pk, terminal pk, half life, you really want to determine that and you really want to elucidate PKPD relationships, activity of the drug and any relationship of drug dosing with safety. And you want to refine that therapeutic index. You really understand what is your range to test this drug? And so the success criteria for a phase one trial are that you. Oops, just close that, sorry. That you have target efficacious exposure. So you really understand, did you get that range where you needed to examine pk, is the drug getting to the right place in terms of pd, Is there evidence of biomarker modulation? Are there dose dependent PD effects that will give you some sense that there's really an activity of your drug? Is the drug safe or if there were any safety issues identified, are they manageable and monitorable and mitigatable? Are there any lab data trends? You know, is the drug tolerable? Is it, you know, when you, when you compare it to what might be competition at the time of launch, would this drug have be useful? And so what do you do next? And this gets into your question a little bit, and it's There are many different things we can do. Most commonly, you'll do a Phase IIA Proof of Activity or Proof of Concept trial, but sometimes it'll look a little different. It'll be a Phase two, but it'll, you know, have some special bells and whistles. Sometimes it'll be a flexible phase two, three, where it's a Phase two study that you might expand to be a Phase three sometimes.