Science Friday - How One Gene Affects Alzheimer’s Risk

Episode Date: February 25, 2026

An estimated 500,000 people are diagnosed with Alzheimer’s disease in the United States each year, but the causes and mechanisms of the condition remain a neurological mystery. A recent study looked at the role of variants in a gene called APOE in Alzheimer’s, and found that while it’s not a simple determinant of developing the disease, that one gene seems to play a significant role in promoting disease risk. Researchers hope work like this could point to new areas to study and even potential treatments. Epidemiologist Dylan Williams joins Host Ira Flatow to explain the findings and discuss the challenges in tracing a complex disease to its roots. Guest: Dr. Dylan Williams is a principal research fellow in molecular and genetic epidemiology at University College London. Transcripts for each episode are available within 1-3 days at sciencefriday.com. Subscribe to this podcast. Plus, to stay updated on all things science, sign up for Science Friday's newsletters.

Hi, it's Ira Flato, and this is Science Friday. Today on the podcast, trying to untangle the genetic factors involved in Alzheimer's disease. Although the exact mechanism of the illness is still a mystery, researchers hope work like this could point to new areas to study and maybe potential treatments. Joining me now is Dr. Dylan Williams, his principal research fellow in molecular and genetic epidemiology at University College London, in the UK. Welcome to Science Friday. Thank you very much. It's a pleasure to be speaking to you. Thank you. You're looking at variations, if I understand it, in one specific gene called APOE. What does that gene do normally? Why are you looking at that? So, its background, this protein apolliproproprotein E, or APOE for short, is a small protein which has a well-established role in shudderick fat around the body. And we also know it has diverse roles in the brain. people can inherit one of three types of the gene, which we refer to as E2, E3, and E4. And so we've known for about 30 years that the risk of Alzheimer's disease differs between

people depending on which type of apoid gene they inherit. E2 is low risk, E3 is medium risk, and E4 is high risk. And this gradient of risk is common knowledge amongst Alzheimer's disease researchers. However, knowledge of this gradient alone doesn't tell us that always full importance. And to grasp that, I think we need to answer what is a face value, a fairly simple question. So what proportion of Alzheimer's disease cases would not occur without the impact of the medium and the high-risk forms of the gene? In other words, if we were to take away effects of E3 and E4, the medium and high-risk versions of the gene from people, with approaches such as gene therapy or other drugs, how much Alzheimer's disease do we think we

would prevent? So you leave the other population of people with that other gene alone and just tackle the medium or high risk factor? That's right. So we're really interested in knowing what this one gene's contribution is. And this is really important because we've got potentially a single target for people to be thinking about in terms of therapeutic development and other strategies. And Apo is a little bit odd in the sense that despite this historical link and these really strong effects, its role in Alzheimer's disease has been overshadowed and certainly I don't think proportionate to its importance. Well, then put this in perspective for us, how big a role does this gene seem to play compared to other risk factors? So we estimated

this proportion, as it were, in four different settings. And we found that around 70 to potentially more than 90% of Alzheimer's disease cases would not have occurred without a contributing role of E3 and E4. Wow, that's a high number all right. Exactly, yeah. And just to stress, this is not the usual case for any particular gene. We couldn't pick one of the many other dozen genes that we know affects the susceptibility of Alzheimer's disease and get anywhere near that type of proportion individually.

So if you're able to find medications or some sort of way to attack these two versions of the gene, that would be really a large contribution to fighting Alzheimer's. I think there's a strong case here for more research activity in funding towards that objective, yes. So where do you go from here? Well, in the case of our own research, I think there are a few key questions to answer. So this research was conducted using samples of individuals who were of European ancestry. We know Apois role in Alzheimer's disease might differ for other ethnic groups, and that's quite critical to understand.

It might tell us what other characteristics that are present in different ethnic groups could be either amplifying or buffering the effects of Apoenae on Alzheimer's. So that's one important thing to investigate. The second will be other genetic and environmental factors, which could also be modifying this risk. And I think, for example, it's worth considering that even the people that inherit the highest risk version of APOE, so if they inherit two copies of E4, they might have a lifetime risk. So the risk of developing Alzheimer's in a typical lifetime of perhaps 60%. And that's reassurance in the sense that many of these individuals, perhaps 40%, will live a full lifetime without developing. dementia due to Alzheimer's disease. So there's clearly other factors at play that are of importance

too. So finding those factors will also give us other ideas for interventions that might help prevent or treat Alzheimer's cases. And the third thing that I think, again, has been overshadowed in Alzheimer's disease research for a long time is what is special about this low risk form of the gene E2? What properties does it have that mean these people have a very low susceptibility of Alzheimer's disease? And the flip side of that is what other health consequences might there be to having E2 relative to having the much more common E3 and E4 versions of the gene? And that can tell us both about the prospects for being able to prevent Alzheimer's disease, but also about the safety of related therapeutics if you were to try and modify apparie related risk for individuals.

Well, do most people have, as you say, from what I hear you correctly, most people don't have the lower risk variant two. They have either the three or the four. Well, that's right. About 15% of the population will have inherited one copy of E2. It's uncommon. It's not exceedingly rare, but uncommon to have inherited two copies of E2. So it's somewhere between about half a percent and one percent of people. So maybe one in 200 to one in every hundred people will have two copies of E2. This is one of the reasons why cities like How's and Nolfo in the sense that you need very big resources to find a large number of this group to be able to compare the risk with everyone else. I see. When we say this gene has a role in Alzheimer's, are we talking about the memory symptoms and dementia, or are you talking about the physical presence of the plaques and the tangles in the brain that have been associated with Alzheimer's? Well, we would probably think of it in terms of both. So we know that they actually build up of

the hallmark characteristics of Alzheimer's in the brain varies by these forms of api. And similarly, we know that api will relate to the clinical symptoms in terms of the problems that we see in dementia, problems with memory and thinking and so on. And obviously, these things are very closely interrelated. It isn't really to sideline this, you know, our study doesn't really suggested we should be sidelining investigations into the more classic we were thinking about how we could prevent Alzheimer's disease, which is to try and remove these characteristic proteins, other proteins that we see build up in people's brains. It's more that progress towards doing that removal of these proteins that build up has been very slow. We don't have that much

to show for 30 years of progress. Things are looking more optimistic now, but I still think there's a case for diversifying how we try and tackle this disease. And as I say, interest in apparelia has probably not been proportionate to any extent to its importance. Yeah, that's what I hear you saying. Let's talk about using this research for possibly treating or preventing Alzheimer's. Would this show you where to target or drug, for example, or would you potentially be looking to gene edit that bad variant of the gene, or would that be the drug itself? Well, I think those are both viable routes to be considering. So obviously with the maturing

technologies in gene therapy, being able to switch genes on and off, or being able to transfer in genes and an individual might not have inherited themselves, these are all emerging technologies and they're being used in the clinic now for certain diseases, including some examples of diseases of the central nervous system. So what's nice about these potential, and I speak about this hypothetically, but what's what's, nice about the potential of these therapies is to some extent we can circumvent some of the uncertainty about what it is that Apple is doing to increase people's risk by so much. You can kind of tackle the problem at its root, so to speak. And then obviously the second approach will be more, in

terms of more conventional pharmacology. So genetic risk really points us to potential pathways by which a gene leads to a disease. So we can think about trying to modify certain aspects of the pathway with more usual therapeutic approaches. At what point does it become, I guess, a good thing to start screening people for this gene earlier in life? Would that be useful or would it just be scaring people too much? I think the latter at this point. I mean, it certainly wouldn't be recommended in the UK for people to seek it out other than fire direct to consumer testing. Because these genes are not going to be destiny, as it were, not.

not determining risk to the point where you're guaranteed to be getting Alzheimer's or not. They don't really help us predict very well who's going to get the disease, but we know that they have some fundamental role in most people who do go on to get it. So, yeah, we would have really recommend people seek their apiary genes because at the moment it wouldn't really give anyone much of actionable information. You don't want the gene to get a reputation as, oh, this is what causes Alzheimer's. Well, not in a singular fashion. I think it's a nuanced thing to understand. I mean, one analogy I use is smoking in lung cancer, where a lifelong heavy smoker,

for example, might have a lifetime risk of lung cancer in the region of about 15%. So again, maybe 85% of long-term smokers could go a full lifetime without getting lung cancer. This is clearly a strong predominant cause of lung cancer. and yet that on its own is not sufficient to produce the disease. Other causes are sort of needed in concert. And I think that's true also of the role of Apoe in Alzheimer's disease. Although other factors will be needed to kind of amplify or buffer Apoor his contribution, ultimately the disease would not occur for most people without its effect.

And equivalently, if we remove that effect, most disease could be preventable. That's a very good analogy. You said this has really been under-investigated over the years. Is it being fully investigated now by you and other people? I think there's more attention for it. One fact that I think is quite telling is that amongst about 140 or so different types of drugs that are in clinical trials for Alzheimer's disease now, only one directly relates to APOA as a form of gene therapy, actually. And so I think it's quite telling after 30 years. years have known this link. This is the only form of therapeutic we have amongst a repertoire of

over 100 others. Very interesting. And thank you for taking time to tell us about it. Oh, you're very welcome. It's good to step out of the ivory tower every now and again, or a 1960s tower block in the case of University College London. Yeah, I get it. Dr. Dylan Williams, he's a principal research fellow in molecular and genetic epidemiology at University College London the UK. This episode was produced by Charles Berkwist. Thanks for listening. And please rate and review us on your favorite podcast platform if you like the show, of course. We'll see you soon.