BIO from the BAYOU: Build Your Biotech Network with Key Opinion Leaders and Investors
19 Feb 2025 · 13 min
0:00
In this episode of Bio from the Bayou, host Dr. Elaine Hamm interviews Alexis Dacoti to explore cutting‑edge biomedical research happening at Tulane University and its translational relevance for biotech in the Gulf South. The conversation begins with Dr. Hongwen Deng’s Nature paper that applies StereoSeq spatial‑temporal transcriptomics to human prefrontal cortex tissue from Alzheimer’s patients, creating a “Google Maps” of gene expression that highlights early‑stage molecular changes, mitochondrial dysregulation, and a transcription‑factor‑driven gene network as potential therapeutic targets beyond amyloid‑beta. The discussion then shifts to Dr. Deng’s earlier work on osteoporosis, where a large‑scale analysis of 4,000 participants from the Louisiana Osteoporosis Study uncovered structural genomic variations that explain more heritability than single‑nucleotide mutations, opening new avenues for bone‑health interventions. Next, the podcast highlights Dr. Xu’s chemoinformatics platform that aggregates toxicity data for thousands of small molecules and applies machine‑learning models to predict hepatotoxicity from chemical structure, enabling early safety screening in drug discovery. Dr. Joo’s AI‑driven nanomedicine effort is also covered: a curated nanomaterial database coupled with ML models predicts physicochemical properties and toxicity, accelerating design and optimization of nanoparticle‑based therapeutics. Throughout, the hosts emphasize the importance of open data portals, collaborative contracts with industry, and a commercial‑viability mindset that bridges basic science to product development. The episode concludes with a call to stay curious, leverage these resources, and explore commercialization pathways for the highlighted technologies.
Chapters
From Google Maps of the Alzheimer's brains to using AI to spot toxic drugs before they hit the market, tune in and geek out with us for the future of medicine. You're listening to Bio from the Bayou, empowering biotech leaders with the tools, assets, and expertise to improve healthcare outcomes.
Each episode provides insights and information to expand your network, advance your product line, and help your company thrive.
Serving the entire Gulf South, but recorded in the heart of New Orleans. Hey guys, and welcome back to Bio from the Bayou. I'm Dr. Elaine Hamm, Executive in Residence at Tulane University School of Medicine, and today's guest is Alexis Dacoti. And that can mean one thing, and that is we are going to put our geeky science hats on and talk about some of the interesting research occurring here at Tulane. Welcome back, Alexis.
Hi, great to be here.
As always, when we think about commercial viability, I really love imagining the future potential of a discovery if everything falls into place, because I believe that curiosity naturally leads to the next steps of figuring out what it would take to make it a reality. And so before I dive into, will this work, and putting sort of my know hat on, I like to really think about the genuine discovery that the researchers here at Tulane are creating here. And so let's talk a little bit about some of the researchers here at Tulane. I'm like the first start with Dr. Hongwen Deng's work on, let's start with the Alzheimer's piece, because I think it's such fascinating work that he's doing. He had a Nature article recently, I believe, specifically looking at StereoSeq of Prefrontal Cortex in Aging and Alzheimer's Disease. And that's not necessarily a super sexy title. It's very sciencey, but what I kind of took away from this really interesting paper is I think of it as sort of like Google Maps for the Alzheimer's brain, and it allows you to really see exactly how genes behave in different parts of the brain as we age and if we get Alzheimer's disease. And I kind of think of it as like zooming in on the Google Street View of the brain, as it were, to really see where things are going wrong at microscopic levels. So that's what I find very interesting about this paper. What did you find interesting about it, Alexis?
Well, one of the things that really sets this paper apart from the beginning is a lot of Alzheimer's research is done in mouse models, which are not necessarily representative. They can tell us a lot, but ideally you'd like to see this in humans.
Yeah.
And so they used, they used brains from 6 Alzheimer's disease human patients. And another thing that I think is cool is you describe this like Google Maps of the brain, as if you're kind of seeing what's happening at each place. You can zoom in, like satellite view what's happening. What genes are being more expressed right next to this amyloid plaque or right next to this area of neuronal death? But it also allows a time-based component. So you can look at how this, what's happening, say maybe when the disease is at an earlier stage of progression versus later on, and you're getting not just those spatial changes. It's not just a spatial map, but it's also kind of a temporal map as well.
That's really interesting because Alzheimer's disease can be a very long disease and it can start very early in life. It can start in very late in life as well. And so understanding which genes change early in the Alzheimer's disease process might be able to help scientists develop more targeted or personalized treatments that could stop the disease before it starts causing major damage.
Right. One of the biggest issues that we have in Alzheimer's research is for years, and there was sort of a recent scandal with paper retractions and things, but amyloid beta has been essentially our only target, that and tau tangles. But It's actually a fairly heterogeneous disease, and it manifests differently in different people. And so it would be fantastic if we had a better understanding of potential targets. One thing that I found also interesting about this paper was they identified these larger gene sets that were associated with some of the issues in the neuronal damage and the progression of the disease, and found that they were all potentially regulated by a transcription factor that could act as a a potential target in the future for new treatments that would maybe hit all of these molecular pathways at once, as opposed to just hitting amyloid beta.
The standard, but not always helpful. Yeah. And that's one of the more frustrating— having been in Alzheimer's disease research, it has been incredibly frustrating to— it's a very complicated disease with a heterogeneous patient population. And we've definitely looked a lot at amyloid beta. So it's nice to be able to expand to different targets. But one of the things I also saw in this paper was the study find that looked at specifically changes in mitochondrial genes too, which, as we all, as we might remember, is the powerhouse of the cell from our high school science classes. And that could explain a little bit. I mean, the brain needs a ton of energy, and maybe that could help explain a little bit about why brain function declines in Alzheimer's. So maybe looking at new ways to slow or maybe prevent disease.
Yeah, when you get this sort of mitochondrial dysfunction, everything starts breaking down a little bit, and that can create almost a spiral. They don't have the power and the energy to like keep up the processes. So then more things accumulate, more damage occurs, then the mitochondria get even worse. And so, yeah.
He's also done some additional work in osteoporosis. So growing up, my mom always really hammered home drinking milk, getting calcium. And now that I'm near my menopause age, I mean, getting calcium and bone health is really important. Having strong bones is really tied to elder health. And so let's talk a little bit about osteoporosis and his work in that area.
Yeah. So Dr. Deng actually studied osteoporosis prior to the Alzheimer's work, and that's been the majority of his career. But You know, osteoporosis is characterized by low bone mineral density, essentially higher risk of like fracture and so on. And another thing we know about it is it's highly heritable. So if you have someone in your family or a family history of osteoporosis, you are more likely to get it, right? But—
So you're saying I can blame my mom?
That's great.
Yep.
That's it. Just drink your milk, Elaine.
Okay.
But despite a lot of study, we've only found like a few genes that are actually explain a lot of that heritability, right? And when I say genes, I mean maybe a small mutation in a gene or a collection of genes that is explaining like what's being passed down from parent to child. But that's not the only sort of genetic change that can be passed on. So a recent study where they looked at over 4,000 patients from the Louisiana Osteoporosis Study, they were looking at what's called genomic structural variations. And they're these sort of larger scale changes using sequencing of these patients, they found that those were actually potentially explaining more of this heritability than all the studies that were looking at the individual point mutations did before. And so this can potentially go a long way to finding new targets or at least understanding why this disease is passed down like this.
That's fascinating. That's, well, and okay, so now let's move a little, and we can't talk about technology these days without talking about AI. Let's talk a little bit about Dr. Xu's research specifically and his machine learning-based platform. And I wanna first talk about his work in liver toxicity. I probably think about liver toxicity more than the average person.
I figured you would.
I mean, I think about it, obviously I'm in drug development, liver toxicity is always an issue, but I also own a wine bar. And so I'm always thinking about alcohol consumption as it relates to my ALT and AST levels. But let's talk a little bit about the work that he's doing specifically at tackling liver toxicity.
Yeah, so when you're developing drugs, obviously it's great if the drug works, but if the drug also takes out your liver and causes damage, that's not great. But ideally you'd like to be able to find that out earlier on when you're screening your compounds, when you're— before you start putting them into animals, start putting them into humans and find out, oh, well, actually the efficacious dose, this also is going to cause some significant side effects. So what his lab actually, they focus on what's called chemoinformatics, just trying to understand the properties of chemicals, of materials, and how that relates to potentially medicine.
I like that. Chemoinformatics. That's going to be my word of the day. Chemoinformatics. Awesome.
And so they found all these databases of different compounds and the toxicity of like small molecule compounds and toxicity data from different assays. And they put this all into a big machine learning algorithm. That essentially is relating the potential toxicity of a compound to its structure, right? So the idea is then you could potentially feed in a structure and get some idea without having to do any real experiments. You're just checking this database of what its toxicity might be. And that would allow you to screen a lot of compounds, kind of hone in on the ones that are most likely to both still have an effect therapeutically, but also be less likely to cause liver damage.
And that of course saves time and money and some potential mice lives as well. So we try to think through our SAR campaigns as we're developing new drugs. He's also done some research related to AI-driven design of nanomedicine, which sounds a little bit sci-fi, honestly. Of course, anytime you say nanomedicine, I just kind of think of like little robots crawling through your, through your body. Um, let's talk a little bit about his work there.
Nanomedicines, they're a relatively new class of medicine. It's essentially using nanomaterials in medicine, and they're sometimes used for drug delivery. Maybe you encapsulate a drug that can't normally get into a cell in almost like a, a little droplet in a sense, or it could be incorporating like genetic material. And these can get into the body and maybe get into cells where the drug normally wouldn't be able to reach. So that's one potential use case, right? But because there aren't as many of these as there are, say, like small molecule compounds, there's just less known about Sometimes they also have their own potential toxic effects. They have different structures, things like that, but there aren't as many examples. So it can be hard to predict in advance, again, what the properties of this nanomedicine are going to be or how to optimize. And so what his group has done is, again, made this big database of different nanomedicines, nanoparticles, and materials, and used machine learning methods to essentially create a system whereby you could predict properties of different structures and optimize them before you actually go through the manufacturing process of making that. Nanomedicine. Another thing I actually really enjoy about Dr. Joo's work is that he's very collaborative and he's done contracts with Merck and other pharmaceutical companies, but he also is kind of committed to making sure these resources are available. So for both the toxicity, hepatotoxicity research and the nanomedicine research, he's actually put these on portals online so that if researchers want to use them, they can. Or if they want to set up some sort of contract with him, they're free online, but if they want to set up some sort of contract for deeper research, he's also open to things like that.
That's fantastic. And we will make sure that we put those links in our show notes too. We've covered a lot of ground today. So we've gone everything from the Google Map of the brain to nanomedicine. And if there's one thing that kind of stands out, it's just that there's a lot of different things that we're doing as it relates to drug discovery, drug development. And some of these are basic research projects in many ways, but they could help really inform how we think about new treatments, whether it's for Alzheimer's or personalized medicine. And the key takeaway from these podcasts that we have developed on our geeky hats is to really stay curious and be open to where data is headed and thinking about commercialization in a little bit of a different way. So a huge thank you to Alexis for sharing your insights. And of course, to all of you for tuning in, we really appreciate it. So For now, as always, make sure to check out the informational links posted in our show notes to connect with these incredible investigators and learn more about the work that they're doing at Tulane University School of Medicine. Thank you once again for joining us.
Thanks for joining us for Bio from the Bayou, and we hope you'll join us again.
To never miss the latest updates in biotech, hit the subscribe button now.
We'll catch you on our next episode of Bio from the Bayou.
