International Medical Conference Endometriosis 2025:
Endometriosis 2025: Your Mother Should Know, Your Doctor Should Know Better!
Mice, or microfluidics? How Multi-omics Analysis Guides Us to Build the Most Useful Models of Endometriosis Patient - Linda Griffith, PhD
Okay, welcome to our last session. This is Science and Endometriosis part two. I have the great pleasure of introducing Dr. Linda Griffith. She's a professor of biologic and mechanical engineering at the Massachusetts Institute of Technology and a pioneering researcher in tissue engineering and endometriosis. Her groundbreaking work in bioengineering has advanced the understanding of endometriosis through innovative disease models in organ on a chip technology. In recognition of her significant contributions to endometriosis research and advocacy, Dr. Griffith received the Endometriosis Foundation of America's prestigious Harry Rich Ward in 2011. Welcome, Dr. Griffith.
This was such a great meeting. Thanks for having me. And it's always fun to be here. And the talk I'm going to give today is more to start a conversation with this audience. This is a place where I can really get great discussion about the topic I want to mention today. And so I've subtitled this move over mice. I didn't say put the mice in the garbage disposal or something like that. It's just we have had it come up several times today that we can learn a lot from mice. And some of these things may be very specific for a particular thing we want in drug development. But the world I come from, I work with a lot of pharma companies going back 30 years on how do we build better human tools to do drug development, particularly preclinical models initially for toxicology, a lot of work in liver, but more recently everybody's coming to us.
How do we build human models for preclinical development? So that's what I want to talk with you about today. And it's a conversation. There's no disparagement of anybody in the room for not using a tool we have because they're still in my lab, although some of them are becoming commercial and I don't hate major gel or anything. It's just that we need to get the tools. My place in the ecosystem is to build things that will be useful and I need your input on how to make them useful. So that's what I want to do today. And I have a disclosure. I am working with somatic on the June kinase inhibitor. I'll talk about that. And we also have a patent on the extracellular matrix that I'll describe. That patent is finally being moving into beta testing. So anybody who wants the synthetic matrix OS e or anyone, just email me and I'll put you on the beta test list because you can finally get at my lab.
Could not support everyone who wanted that reagent. It's not like sending you a plasma, it's actually work for the lab to build it and give it to you every time, but it will be commercially available soon. Now, I have for many, many years gotten pulled into pharma because as we know, immunology is very species specific. Some of the phenomenology can be captured in mice or baboons, but when you get to the particulars, particularly how an immunologically focused drug against inflammation or other things works, you often really need to have data in humans because humans are incredibly variable. We also have learned a lot about how it's sex specific, and this is a huge area of focus in our lab right now to build the kind of tools that will let us build a female liver, a male liver, and to do a much better job of our endometriosis in vitro models.
And I'd love to talk to you about what we're doing there and we really need to get away from animals. And just as a little bit of an intro, there's been a movement in pharma for years is why they come to me to get away from animals. I work a lot with Novo Nordisk, I've worked with them for eight years. And here's Lata Nuen who won the last award last year in a beautiful story. Anybody who wants to read this about GLP one, she got them into the clinic talking about how you would never have gone after GLP ones if you started in mice. And then the VP of Novo at an NIH meeting explaining how they don't really want to use animals and cardiometabolic disease drug development anymore because they are too many rabbit holes. Novo built a lab right outside Boston to interact with us around these kind of technologies, not for endometriosis unfortunately, but for cardiometabolic disease and so on.
So I have a big lab of doing these other things. So with Novo, this is an example. We make a little liver and a microfluidic device, the green or blood vessels that you can send immune cells through the purple or aggregates of hepatic cells with non parenchymal cells and hepatocytes. So you can look at immune cell trafficking in a model of type two diabetes. Very sophisticated things. I forgot to put, there's actually, yeah, there's a reference on bio archive if you want to see that. Now this has gotten to be such a tough thing. NIH did a study a few years ago, actually Larry Tabak when he was acting director, I banged on him for years about this when I was on the advisory committee to the director to study how NIS as animals. And then he did that study and they decided they really need to build what are called novel alternative methods or alternatives to animals.
So I was on a working group of the advisory committee to the director together with the head of our division of comparative medicine at MIT who's a veterinarian to really study how do we bring these humanized models, these so-called NAMS into the realm where people like you in the room who'd want to do in vitro studies can use them and how drug companies can use them. And so out of that came a report on combining computational things, everything from genetics to systems biology, et cetera, with the so-called micro physiological systems, NPS, so that we can build better models of humans. And I wrote a lot of that report and it says specifically endometriosis is an area that really needs these models. So there's an RFA out on the street from NIH, and I'm working on a grant with Stacey and Andrew Horn and others to be responsive to that.
How do we build these tools and make them better? So I'm going to talk to you a little bit. Some things we're thinking about, nothing is cast in stone. I'm really here to learn and have a conversation and share with you where going. So some of the questions our group discusses a lot because I have colleagues who study TB and the biggest most ardent supporter of the move over mice movement at MIT and we really have a movement is my colleague who studies TB because mice are terrible models for tb. And so we think a lot about the gaps in using mouse models. And again, this isn't to criticize anyone because there's gaps. I didn't say that they're not useful. And you heard Katie Burns talk, I found her review article incredibly valuable to me because I don't use mice. I've never used mice, I don't know that much.
And there are people in this room who've contributed tremendously to development of very detailed protocols to try to understand more about endometriosis by studying various models of putting them in mice. And here's one from Aaron Greaves and Andrew Horn. And you heard from Katie and I think she really, again, this paper is amazing for someone like me to get oriented to all the facets. And of course you've heard today from others. This is just one paper that I think Andrew Horn's A with Aaron Greaves, a co-author looking at how using mice to look at how macrophage populations that are in the peritoneal cavity are attracted into lesions when a bit of tissue is implanted there. And I'm using this slide to highlight. I have a graph and it may be a little small. There's something called large peritoneal macrophages shown on the left. It looks like those are going into the lesion and they have a model that there's green fluorescent protein so they can count them in the lesions and you can kind of see that.
And then there's something called small peritoneal macrophages and large peritoneal macrophages are characteristic of macrophages that lived in the peritoneal cavities since, went through ony in the peritoneal cavity. So they're so-called tissue resident, whereas the small peritoneal macrophages are recruited from CD 14 positive monocytes into the peritoneal cavity. Okay, so these are very important because they have different programs of gene expression. The LPMs have expressed gata six, you can Rob Taylor had beautiful work on this and I think all of us here, a moment at some point in this meeting to remember Rob, because I miss him every day that I do endometriosis, but that retinoic acid contributes to maturation of macrophages and the peritoneal cavity mice have a bit more retinoic acid, et cetera. So there's a lot of beautiful work in this paper characterizing which macrophages are contributing to these lesions. I'm not going to go through this whole paper, I just wanted to bring up that point.
But if you now go into the immunology literature, there's starting to be a lot more appreciation for differences between human and mouse immunology. And so for example, this is when randolph's work where she characterized differences between mice and humans. And the punchline or headline is that in mice you have almost all tissue resonant macrophages. But in humans you have, those are rare and that's in adults and children. And so I really recommend this paper for people who care about peritoneal cab macrophage biology. And then you can start to look in endometriosis models and there's a paper by Carina Zondervan and Christian Becker using mass cytometry in control and endo patients. And then a mouse model that was recently reported of pain using SCR aeq. And you can see that there's some pretty profound differences in the populations, for example, of T cells and B cells in these two models.
And this is in a comprehensive review, but there are notable differences. So we could ask some questions about the immunology in these mouse models. I'm not here to again pass judgment. I'm just observing things and thinking about it myself. Then there's of course using mouse models for pain. And I think that Pariah Bain made a really good case for the LPS model for a particular question they have about drug development having to do with uptake of a drug that will go into a patient and is it getting to the target not necessarily validating the target. So there's some distinctions. And so there's this paper that just came out, for example, CGRP of course is very prominent throughout the nervous system and a huge look at interactions between nerves and macrophages. But then you see other things coming out about pain mouse models. And I've started to talk to a lot of the pain community about the mouse models and it's complicated.
There's really a great guy, some of you may follow on Twitter, Brian Kim who wrote this really great paper that I just love. He's an itch guy and he studies itch. And when it gets into these models and what you can learn from mice, again, you have to be very specific because there are many, many modulators of macrophage and nerve when they get together. And some of them are very species dominant in terms of the way they translate. So I have come up with questions because so many companies want to develop, they need some kind of pain models. And for now mice are it. So we've started to ask, can we start to capture features of these pain models in R vitro cultures? So you have the tissue level pain can start there from inflammation. You go to the dorsal root ganglion, we know that there's macrophage, there's stuff going on there.
And then you go to the dorsal horn up to the brain and let's just focus for now on the tissue level. And we are thinking about the DRG also, but let's focus on the tissue level. You can get sensitization by inflammation here and central. So there's a lot going on, but let's kind of focus on the tissue level. Now I got interested in the tissue level because of interaction. We identified June kinase. I'm not going to go through this in a clinical study over 10 years ago and trying to look at mechanism based classification. And we were thrilled when Steve Palmer pointed out that Merck was taking these through preclinical models and ultimately licensed it. Unfortunately not a successful model, but he published these two papers and it worked really well in these models. He went to Baylor and you heard from Pariah the fabulous stuff that came of that.
But he told us, and I met him, we were together at A SRM in Hawaii the year that there were 12 inches of rain. We sat in a bar and talked about the future of June kinase inhibitors. He says, well, I'll make the new molecules, but Linda, what you need to do is you need to build a little lesion that interacts with blood vessels and immune cells. Lemme show you how we've done that. So the first thing, you want to take patient cells and build a model of these lesions that you can manipulate and that you think has relevance to the InVivo situation. It's hard to do that. I'm not done, but I'm getting close and we will have some tools to share with you beyond just the matrix. So the first thing we did was develop a replacement for Matrigel and everybody in this room and we figured out how can we co-culture organoids and stromal cells together in a synthetic matrix?
Because when we use matrigel, which is what everyone uses because that's what we know, you can take glands, dissociate them to individual cells and put it in this commercial gel, which comes from a mouse doer. It just has some proteins that support epithelial cells. So everyone uses it. And so I'm dissing it from my job as an engineer to say, oh my god, it's not going to help us have crosstalk. We need to replace it with something that gives us better biology and better reproducibility. It's got all of these problems. So that's what we did that I showed you. And Juan Ko, who's now at Tufts really banged on this during his postdoc to develop this and show that you can get responsive progesterone. And a huge advantage is you can culture cells for weeks and weeks in this and it doesn't fall apart like Matrigel does.
And we can tune that. The other thing that we can do in this, and a good reason to culture for weeks was highlighted by someway grows talk because fibrosis is a huge part of this whole process and in vitro, I'm going to walk you through this slide because we started to use a special kind of microscopy that involves multi photon imaging. So we go beyond confocal, and this is done by an amazing lab. And I forgot to put the reference on there. I'm so sorry. I just put this in at the last minute after some ways talk. So my colleague Wan Yu, she does something called multi photon label-free imaging. So it's sort of like when you have confocal microscopy, you want to excite a floor four that gets excited at 400 and you shoot two 800 nanometer wavelengths and they land at the same time to give it the same energy as a 400 would have.
And that lets you, because it's longer wavelength, you penetrate deeper in the tissue. Okay, so she does this to the max. So there's no label, there's no staining of these samples at all. This is all endogenous signal from laser excitation with this big old microscope she has in her lab and her computational ability to take the light that comes out of the exci tissue and look at it. So over on the left is just an organoid. The endometrium tissue is a fiduciary sample from an early luteal phase, which we've gotten some amazing biology I can talk to people about, I'm not going to talk about that. That's a fiduciary of the tissue. So the green is collagen and the yellow is a meta redox state. But over here is the organoid. The bottom is just the individual channels, but the top is what I want you to look at, that green is collagen that's being deposited by the stromal cells because it's a co-culture.
So this is only a four day culture. If we go out longer, there's more collagen. I just didn't have the image. So this is a very important tool that we can now, if we can culture for a long time, we could start to get collagen deposition and look at how environmental factors will do that and monitor that plus the metabolic state in real time. This isn't something that's going to go to your lab tomorrow, but it could go to your lab in about a year and a half. Stay tuned. The other thing about a synthetic matrix is we can change the microenvironment to make it very stiff. Like if we don't get the collagen deposited in culture, we can actually mimic, this is an adenomyosis lesion with light sheet microscopy. So you see the tentacles, the epithelial cells are invading into the muscle. And if we have a soft synthetic hydrogel, the organoids are beautiful, these are co cultures, and on the right it's invasive.
So there's a lot we can do with these synthetic tools. But mostly what we've been doing lately is saying, how do we put this together so that we've got our nice static cultures, which anyone in this room can do because you can now sign up to be a beta tester for that matrix. And if you are in the endometrium field, I make a special category for its biotechni that's doing it. There's a special category if you work on endometriosis to get that prioritized. So on the left is a microfluidic device on which there's two fluidic channels and the movie is kind of showing now it's slowed down. There's a perfusal micro vasculature in between. So this whole thing, and that device is about 250 microns tall and you inject a fibrinogen thrombin precursor with vascular cells and it makes that. But we adapted that kind of technology microfluidics.
So there's a channel on either side, and we injected this at time zero. So this is a time-lapse movie over six days. We injected it with the microvascular cells, but also with proto lesions, organoids and stroma. So you can see over this movie, the vessels are forming and they're vascular rising, these lesions that are in a microfluidic device. And this device is about, I'd say 18 months away, I think, from being able to leave our lab to a very friendly understanding beta tester. And so we're probably going to put it in Edinburgh first. So we're doing that and we are now working with somatics because we want to ask how can we start to capture some of these features that we really need for immunology in these systems. So we are moving toward that and Somatics is one of those friendly people who will understand that it takes some time to get the protocols to be robust.
One of the worst, most daunting challenges for us as engineers is actually getting that microenvironment just right in terms of controlling the hormone microenvironment. When we put progesterone in a co-culture, it's gone in a New York minute if a New York Minute's, like two and a half hours, but it gets metabolized very quickly. And the metabolism, we know this is from a review by Philips Saunders and co-authors is very complicated. So we're building, we're paying a company that does quantitative assistance, pharmacology a huge amount of money to take this and make a mathematics model. And then we do measurements on metabolism when working with Edinburg for this. So we can build a system that will keep the hormones very tightly controlled over weeks and culture. That's not going to happen overnight. And it's a $4 million project, which we got philanthropic funding for. So the question I have, and we can talk about this over dinner, is what we have right now was what I showed you before this, we have the vessels, we have the immune cells because we have macrophages, we have the lesions.
Do we need to add nociceptive neurons to this? And so we've started a project with the Moore lab at Tulane with the fabulous postdoc, Sam Holt, and we hired Michael Moore's technician. We're starting to work with Ross Siegel's lab at Dana Farber. It's still in the early days, but our synthetic gel does better than Matrigel to get no susceptive neurons to grow from a dorsal root ganglion. There's a lot of issues on stability and so on, but I think we'll make it finally in the last couple of seconds, how do we decide that our in vitro models are capturing in vivo phenotypes? We're accruing lots of transcriptomic data and our mathematical modeling collaborators, including my husband, Doug Berger, has ways to look at these signatures. And it goes beyond PCA plots. There's a lot more to it. But as we get these well stratified patient populations with their data, a lot of math can be done to do that.
So we'll close by saying we're building a core facility at MIT for liver first because liver, everybody wants liver. And liver is actually very important in drug development for any endometriosis drug as Bayer knows because they had a drug vallum phase two due to drug induced liver injury a couple of years ago. So this is happening. We've got this really cool technology that we hope will become turnkey in about a year and a half for a couple of intrepid users to use. And I'll finish there and thank my amazing collaborators, including Dave Trumper, who builds all the hardware, my clinical collaborators and so on. Thanks.
