I had the pleasure of speaking to Dr. Sek Kathiresan, head of obesity research at Eli Lilly and Verve Therapeutics CEO — the pioneer of one and done lipid-lowering therapies. The conversation was fascinating, packed full of information, and well worth a listen — here it is!
This episode is brought to you by my partners at Consensus, the AI agent for literature research and medical questions. Get a free trial with my link!
Sek is a cardiologist and an academic. He made many of the initial discoveries of targets for cardiovascular prevention, based on human genetics — some of which later became medicines, including NPC1L1, the target of ezetimibe, APOC3, the target of olezarsen, and ANGPTL3, the target of evinacumab. He also discovered that HDL is not causally related to heart disease, and that clonal haematopoiesis is a causal risk factor for heart disease.
He started Verve Therapeutics to make one-and-done lipid-lowering medicines, and the company was acquired by Eli Lilly in 2025 for up to $1.3B. The Verve pipeline includes base editors targeting PCSK9, ANGPTL3 and lipoprotein(a).
Sek is now also the head of obesity research at Lilly. So, really at the forefront of many of the most exciting parts of medicine and biotech — and a perfect person to speak to at this time.
Sek:
I consider my career in three phases. Phase one, roughly the first 10 years, was really learning human biology and medicine. So I trained in internal medicine, in cardiology.
The second phase, about another 10-15 years, is research. After finishing clinical training I spent five years as a post-doc fellow in human genetics. This is the early days of human genetics, 2003 to 2008. The human genome sequence had just been completed and I learned human genetics at the Broad Institute and the Framingham Heart Study, to understand the correlation between genotype and phenotype. And really focused on two phenotypes. One is cholesterol levels in the population. The second is heart attack (myocardial infarction, coronary artery disease) in the population, to try to decipher the inherited basis for those two phenotypes.
Then the third phase, which started in 2018, is around entrepreneurship and taking some of the insights from that second phase and trying to make medicines to improve patients’ lives based on those genetic insights.
The ANGPTL3 story is particularly resonant to me because back in 2008, we made the initial observation of people who completely lack this gene. There was a family in St. Lewis, a very large three generation family where LDL cholesterol being very low and triglycerides being very low ran in the family — as if a single gene was responsible for that. In the second generation, multiple people had the phenotype of very low lipids — that we called familial combined hypolipidemia. We ended up taking two family members and sequencing the entire exome, back in 2009.
The reason they had the low lipid levels were they were human knockouts for the ANGPTL3 gene. We went on to show that there are hundreds of people around the world that had this phenotype. They were healthy, they had the low lipids. As a result there have been medicines in development targeting this gene, both a monoclonal antibody and now an siRNA, and then we’re developing at Verve, a gene editing approach as well.
At Verve, the main program is a one-and-done cholesterol-lowering medicine.
The PCSK9 program at Verve is a one-time therapy that turns off the PCSK9 gene in the liver permanently, and leads to LDL reduction hopefully lifelong. And we have that program heading into phase two. The phase one data was just presented a couple of months ago at the European Atherosclerosis Society meeting and a simultaneous publication in the New England Journal of Medicine. So I would urge the listeners to track that paper down. It really nicely summarizes the progress we’ve made with this new technology. This is a base editor. The goal here is to make a single spelling change in the DNA of the liver in the PCSK9 gene to turn the gene off and to really give that protection that some people naturally got by having the gene turned off.
We were able to show in that paper that yes, we can turn off the gene in the liver. The LDL came down between 50 to 60% on average. And it was durable. We showed durability data extending out to 18 months in that study. And I think this kind of editing in the liver is likely to be durable for a long time, probably a lifetime.
The paper Sek mentions is here:
Hepatocytes regenerate throughout life, and some patients also have metabolic liver disease and fibrosis — which are those same patients that have high LDL cholesterol. Does the regeneration or these comorbidities affect the longevity of the therapy?
We’ve looked at this pretty closely and the DNA editing really looks like it’s going to be durable probably for a lifetime. Despite the fact that the liver regenerates, when we first do the therapy all the cells that are responsible for liver regeneration. So when they divide they carry forward the edit. That’s probably what’s happening. And the reason I can say with reasonable confidence that it’s going to be durable for a long time is that there are groups that have shown editing in monkeys is durable for more than seven, eight years now.
In patients the furthest durability is for a different target called TTR by a company called Intellia, and they have now four, almost five years of durability in some of their earliest patients treated. So this really looks quite different actually from viral vector gene therapy which is another form of potential one and done, where there it’s been a little less clear that it’s going to be truly durable for a lifetime — but I think the in vivo gene editing looks quite different than that.
On the competitive landscape of PCSK9, which, like many high conviction targets, we’re hitting from many different angles. Enlicitide, a once daily pill for PCSK9 inhibition; monoclonal antibodies once every two or four weeks; inclisiran, once every six months. How do you view the competitive landscape and how the PCSK9 base editor, Verve 102, might fit in?
Ultimately, different patients will make different choices. For some, I think the idea of a one-time therapy and being done with it after that is going to be particularly appealing. I think particularly people who are younger, who don’t want to face the burden of lifelong therapy. And so we’ve done some surveys of patients, to understand at this present time what the patient preferences are. It turns out when given those four choices, we surveyed about 500 people who have high cholesterol or have had heart disease.
40% choose daily pills (like enlicitide); 5-6% choose twice-monthly injections (like evolocumab); 14% choose twice-yearly injections (like inclisiran); ~33% choose the one-time treatment (Verve-102). People will make different choices.
Adherence is an issue as well, right? Especially for once daily therapies. The adherence to statins drops off to ~50% over the course of a year, whereas of course with a one-and-done therapy, there is really no adherence issue at all.
This is the major insight that led us to start Verve — this idea that lifelong low cholesterol is probably the answer to heart disease. So if everybody had their LDL cholesterol between 35 and 50 mg/dl for their whole life, we may not have any coronary heart disease. Let me say that again. If everybody’s cholesterol was between 35 and 50 lifelong, there may be no heart attack in the world. Now, heart attack is the leading cause of death in the world, so this is a bold statement that I’m making but I think this is true and the conviction behind that insight is what led us to start Verve.
A couple things that people ask when I say this. One is, is it safe to have the cholesterol 35 to 50 lifelong? The average in the United States is like 120, and people think average is normal. Average is not normal. We’re not meant to live at 120. So, what are we meant to live at? What is a healthy range? It turns out, let’s look at a range when we’re born, babies, or range in rural China before industrialization, where there’s actually very little heart disease. Well, it turns out babies’ cholesterol is between 35 and 50. And babies require a lot of cholesterol, because there’s myelinization ongoing in the brain. It’s just in the last several hundred years for industrialization, diet, lifestyle has really shifted that cholesterol curve population-wide to now where the average is 120, but the average is not normal.
What about smoking, diabetes, inflammation, hypertension?
There are a few other drivers for sure, and I think the three major drivers that we’ve seen from human genetics work are cholesterol, blood pressure and weight. Those are probably the main if not the only drivers of atherosclerosis. Ultimately you want to tackle all of them, but really the key driver in my mind, that’s necessary and sufficient to drive atherosclerosis, is cholesterol.
Cardiovascular medicines are typically approved on the basis of surrogate endpoints (lowering LDL cholesterol) without necessarily having to prove a risk a reduction in cardiovascular events — will the same be true of your medicines?
We are anticipating, based on precedent, that our medicines would get approval based on LDL lowering for a couple of reasons. One is the best understood relationship in all of medicine is really the reduction in LDL and its subsequent effect on reduction in hard heart events. In particular for the PCSK9 target, we have now like four different outcomes trials that essentially have proven this and several more ongoing. So I think that’s the precedent with which we anticipate approval, based on LDL lowering.
On the Stripe podcast with Dave Ricks, the CEO of Lilly, he mentioned the need for innovation on pricing for a medicine like this — do we do a one-time high price or actually, what he suggested, is a subscription model where you pay on a monthly basis if the LDL is still lowered — and if it’s not lowered then you don’t pay. Is that the kind of model that you anticipate?
I think it’s a little too early to talk specifically about pricing for our medicines except to make a couple of points. One is, when people think about one-time medicines, they’re thinking automatically a huge price tag. Because the earlier medicines in this space have all focused on rare disease. We’re not focused on rare disease. We’re literally focused on the most common problem in the world, coronary heart disease. So, we have a lot of patients that we need to treat to help. So, it’s not rare. The second is the cost of goods to how much does it cost to make this stuff. Our medicine, unlike the previous gene editing gene therapies, can be made very cost-effectively because it’s mRNA packaged in a lipid nanoparticle and we know that kind of product can be made very cost-effectively. Based on those two features we don’t we anticipate pricing to be way different than the millions of dollars per dose for current one-time therapies.
We think, ultimately, our pricing will be much more like other one-time procedures in cardiovascular medicine. In fact, that’s the way I think about our medicine. Less a drug, more like a molecular surgical procedure. I’m not putting in a catheter to insert a stent, but rather what I’m doing is I’m infusing something that will go to the liver and make a single spelling change in every liver cell. Kind of like a micro scalpel.
The technology focuses on base editing, and has no off switch. Do you anticipate any resistance from patients or prescribers with this kind of approach?
I think that we have to document the safety over time. There are two aspects to the safety of these medicines. As I said, it’s delivered by a peripheral IV. So just a simple IV in the arm, dripped in over about four hours and then you’re done.
In terms of safety, first there’s acute safety, around the infusion. Sometimes the infusion can cause a little bit of challenge to the liver acutely, but we’ve got a good handle on that. We monitor the patients and our current drug product has very minimal effects on the liver. Then there’s the long-term safety. People worry that having the cholesterol low for a long time, and the PCSK9 gene turned off for a long time? That’s one issue. The second is, might you be making a spelling change somewhere else in the genome and causing a problem? For example, in a cancer-causing gene and leading to cancer. Those are the two major issues long term.
There are people walking around who have this gene switched off either one copy or two copies and, lifelong and they’re totally fine. There’s nothing we found wrong with them. And there are millions of patients now who have received the PCSK9 inhibitors for many years and there are no adverse signals that have emerged. So there’s both genetic and pharmacologic validation of the safety of turning off this target, and having the cholesterol low. So that’s very reassuring.
And then on the off-target issue, we’ve looked very carefully to see if we making any spelling changes anywhere else in the genome beyond the intended spot, and we don’t find that. And on top of that, we’ll be following all the patients in our trial for 15 years. This is a requirement by US and ex-US regulators. So, we’ll be able to track them for safety. At the time of approval, hopefully we’ll have at least three to five years of follow-up in a lot of the patients and that should give people reassurance on the safety and then even longer follow-up will come over time. So, I think that we’ll have a pretty good safety database to reassure people.
And then on top of that, there are lots of things we do in medicine that are permanent. And the number one thing is surgery. Every surgery is permanent — we don’t reverse surgeries typically, and patients take it up all the time if you explain the benefits and the risks.
As the head of obesity research at Lily, I’m curious about the problem that you’re aiming to solve and how you’re solving it.
There are several unmet needs that we’re trying to address and I’ll just go through them each. One is: can we get even greater weight loss? The second: can we get even better tolerability — particularly GI tolerability? The third: can you make these things even more durable? The fourth: can we, wherever there’s an injectable option, develop an oral option? And the fifth: can we work around some unwanted effects — particularly muscle loss — that happen sometimes with these agents.
What about brenipatide, Lilly’s GLP-1/GIP receptor agonist? It works in a similar way to tirzepatide but Lilly are focused on neuroscience indications — things like depression, schizophrenia, bipolar disorder. Why is it that that particular drug is focused on neuroscience?
There has been some observations made largely using clinical data in patients receiving these medicines. People have shown that these medicines seem to decrease cravings, cravings for food of course but also alcohol and substance abuse — suggesting that these may have broader neurologic benefits, neuropsychiatric benefits. And this has been confirmed in a series of observational studies. Probably among the best was in the American Veterans Administration System database. These are all observational data, so of course could be due to confounding. The best way to get around and prove that the medicine that it’s cause and not correlation mere correlation is actually test a medicine in a randomized way — and that’s really what Lily is intending to do with brenipatide.
What about retatrutide, Lilly’s triple agonist — the next generation after tirzepatide, that causes a lot of weight loss. Anecdotally when people take it they say they get some tachycardia (fast heart rate). To me it seems like it’s going to be a more specialized, focus drug, for very overweight people, but maybe not for everyone given that it seems like it’s a little uncomfortable to take. Is that a fair assessment?
Reta is in late-phase development right now. And you just saw some of the phase three results this year. The most dramatic finding is the weight loss, the degree of weight loss. We’re getting to bariatric surgery-levels of weight loss with an injection, without having to go through that dramatic procedure of bariatric surgery. So, that’s the most striking finding in my my view. It is an agonist of three hormones GLP1, GIP and glucagon. And it has some of the profile that you mentioned of an increase in heart rate and other items that come from the glucagon component.
When I started my medical career there were no options for obesity. This was a graveyard of medicine development. Every medicine attempted failed and it was pretty hopeless actually.
We now have a sea change in the last five years. Now we have a plethora of approved medicines and ones in development that all seem to work. What’s going to happen, I think, over the years it’s going to be like the blood pressure space or the cholesterol space where a physician is going to have probably a range of mechanisms they could turn to — four or five different mechanisms. It’ll segment the market into people with different needs, and we’ll have different medicines that address those needs. You can see a future, not too far from now, where there will be four hormones that can be agonized for weight loss. One is GLP-1, the other would be GIP, the third would be glucagon, the fourth is amylin. You can imagine all the different permutations of those four — onesies, twosies, threesies, maybe even four.
The other day Novo Nordisk had a big trial so for an IL6 inhibitor, Ziltivekimab, in chronic kidney disease, patients with existing cardiovascular disease and high inflammation. The trial was a negative. Lilly have an NLRP3 inhibitor in trials. I wonder what your thoughts are on the IL6 trial, because it was also based in genetic evidence and then what your thoughts are on NLRP3 and its chances of preventing cardiovascular events, given the failure of IL6 inhibition.
Inflammation has been a very challenging pathway to target. There’s very good observational data that high levels of blood CRP correlate with increased risk of atherosclerotic cardiovascular disease, but one of the biggest challenges in medicine, when it comes to therapeutics, is confusing correlation with cause. This literally is the biggest problem I think, and we’ve gone down blind alleys many many times because we confuse correlation with cause. HDL cholesterol is a great example of this. I think for inflammation here it’s been challenging to come up with therapies that decrease inflammation and also decrease rate of heart attacks.
I think this IL6 inhibitor is a good example — very good pre-specified hypothesis that people with CKD who have high CRP and high IL6 would be at particularly high risk, and you’re targeting specifically the driver of inflammation — and it was a very null result.
So, what happened? It very well could be that inflammation plays a role, but the ability to target it and have a therapeutic benefit, particularly in advanced disease, is just not there. You’re not going to get a big bang for your buck by targeting this pathway, even though it may play a causal role.
And then you have to be a little careful here on the human genetics. It’s typically said that the human genetics supports IL6. It actually supports IL6 receptor and it’s pretty different from IL6. The gene variant that supports IL6 receptor has this pattern: the allele that lowers risk of disease has lower blood CRP level but higher blood IL6 level. How’s that happening? So if you truly wanted to generate a medicine that mimic the genetics, you would need to find a medicine that lowered blood CRP while simultaneously raising blood IL6. I’m being a bit silly, but that’s just to say that we don’t we may not truly understand what that variant is doing, in which tissue, at what time point — to actually get that benefit in patients.
I think that the jury is still out on other targets like NRLP3, and I think it’s worth testing. But I’m fairly sceptical for atherosclerotic cardiovascular disease. Inflammation has been hypothesized to play a role in a range of cardiovascular conditions, not just atherosclerosis, so I’m open that it may play a role in these other conditions beyond atherosclerosis, but I think for atherosclerosis I think the prior probability of success is quite low, given the Zeus results.
Thank you very much to Sek for the fascinating and insightful conversation!
Fantastic and fascinating interview!