The Bench Brief - Dr. Tommy Gambles

James Mosedale (00:09)
Hello and welcome to The Bench Brief presented by ichorbio, your rapid insight into cutting edge life science research. Today we're joined by Dr. Tommy Gambles from Thera-T Pharmaceutics and the University of Utah. Tommy will be presenting on Advancing a Modular Multispecific Antibody Platform for Hematologic Cancers. Tommy, thank you for joining us.

Tommy (00:34)
Thank you. Thank you, James. This is an amazing opportunity. I'm super excited to share what we've been working on. And so thank you for including me in this seminar series for everyone. I'm Tommy Gambles. I am in Salt Lake City, Utah in the States. And I am the co-inventor of this technology that we'll be talking about today. And I co-founded this company, Thera-T Pharmaceutics, and we're a spin out from the University of Utah. And so we're currently in the incubator space, this technology and getting it ready for IND development.

Thank you for joining. I'm really excited to share with you our technology, this split antibody take for hematological

So just to get started, when we're talking about T cell therapies, there's two major shortcomings involved with these therapies. That is refractory relapse and toxicity. Relapse usually comes from this concept of the

cancer population being very heterogeneous, especially in blood cancers. These are cancers of typically B cell populations that have built-in mechanisms to expand very readily. And so they accumulate different mutations and they become very heterogeneous. So when we apply a therapy to this heterogeneous population, let's say we can attack all the red cells and the red cells die and these orange or brown ones sort of have resistant serendipitous phenotypes. We're sort of selecting for this phenotype in a way.

And these cells will persist and eventually expand. And now the patient will relapse and oftentimes with refractory disease to the original therapy. And this can be lethal obviously to patients. That's a big problem in the field. Secondly, over-activation of our effector cell, our T cells is a big problem. T cells are really good at killing cancer cells. But if they're over-activated, I'm talking CAR-Ts or through biologics like bispecific antibodies.

we can get pro-inflammatory molecules, cytokines released systemically, and this can cause a lot of problems for patients, even up to death. These two shortcomings really limit the therapeutic window and prognoses for patients in the clinic. So we wanted to tackle both of these

I just want to show you the schematic of our We call it multi-antigen T cell hybridizers or MATCH for short. It's a clever acronym that will come into play in a second.

So if we look at a traditional bispecific antibody structure, we can bind to one T cell and one cancer cell antigen. I wanted to rethink this. So I split the antibody into two components, just using the T cell engaging fragment and the cancer cell engaging fragment. And then I functionalize them with complimentary oligonucleotides. oligos are very similar to DNA, but they have this morpholino.

delineation where their backbone has been stabilized so that you can use these in vivo. But they still work like DNA. They want to dimerize and form a helix, kind of self-assemble in solution. And they'll do that spontaneously and very rapidly. So if we functionalize two desired fabs with complementary oligos, if we put these in solution, they will spontaneously hybridize and form a molecule very similar to a bispecific antibody construct. But the beauty of splitting it up into two components is twofold.

first, it becomes very modular. So we can target anything we want on the cancer cell. We can even exchange the effector cell engager and we can create these libraries where we can have a bunch of cancer cell engagers on the shelf. And when a cancer cell comes in, we can look at its profile of what antigens it's expressing and sort of handpick a cocktail specific that matches that antigen expression profile. So this becomes multi-specific.

because any bound fab on the cancer cell surface can then hybridize with an available T cell engager and recruit a T cell. And we'll get to that schematic in a second. The second major advantage is we are no longer stuck in a single molecule. So we can optimize dosing on the cancer cell side separate from dosing on the effector cell side. In a traditional IgG, we're stuck in this one to one ratio. We can increase or decrease the dose of the molecule

but we're always going to have this one-to-one ratio. When we split the two binding events into two distinct components, we can optimize separately. So we can say, do a higher cancer cell engager dose just to bind as much as we can on the cancer cell surface, and then really come in later with a much lower T cell engager dose to try to control toxicity. And so as we've published a number of papers on this technology, and so we validated a ton of

in vitro data, ex vivo patient samples from our cancer clinic here. And we've done a number of in vivo studies and I just wanted to hand pick some of the cool experiments that we've been able to do to validate

this technology. So in general here, this is the schematic of everything that we've studied. We're mostly focused on liquid tumors currently. So leukemia, lymphoma and multiple myeloma. And here's our list of targets that we validated and we're getting into solid tumors that we're pretty excited about.

On the effector cell side, we've worked a lot with T cell engagers. And then we've done a number of studies with NK cell engagers, but what we're really interested lately are these macrophage engagers where we can get phagocytosis. And that's how we're starting to attack solid tumors. And so we have a paper coming out on that later. And what I'm showing in this schematic here is actually four different antigens being bound by our MATCH at the same time. And then...

when the T cell engagers dosed, any available bound fab can then hybridize and then recruit a T cell. And so this becomes multi-specific. can truly target upwards of, you know, two, three, four antigens at the same time while controlling the dose and activation of the T cell side.

So let's demonstrate that with our first kind of major in vitro experiment here. What I had in my mind, what I was picturing was teaching

the same cohort of healthy donor naive T cells, how to kill three different cancer cells. And so just picture a big Petri dish full of, full of these donor T cells. And we're going to spike in different cancer cells over a 72 hour period. So I'm just showing untreated control here, T cells in blue and day one we'll spike in a lymphoma cell line. We can take a small aliquot and go to flow cytometry and see our target cells. And again, we'll do that with multiple myeloma on day two. You can see the target cells day three leukemia. You can see the target cells.

So now what we're going to do is have our treatment group and add a dose of MATCH specific to that cancer. So for our lymphoma cell line, we'll add a CD20 targeting MATCH therapy, and that ablates the lymphoma cell lines. The next day we add our multiple myeloma cell line and a BCMA MATCH therapy, and that ablated the myeloma cell line. And lastly, on day three, we add our leukemia cell line and use a CD38 MATCH dose, and that ablates our leukemia cell line. So this is the same T cells.

being taught how to kill three different cancers. And that really demonstrates how modular we can exchange these targeting motifs. So the next big question for us was, can this self-assembly occur in

vivo? And so we set up a pilot study. This is a human lymphoma mouse model. So we have these Raji cells, which is lymphoma cell line in SCID mice, and we have healthy donor naive T cells. We add them as a

like a co-culture inoculation. And we really wanted to control the ratio of these two. So we dosed on the same day and we're just looking at a single dose of our MATCH therapy. And this is the trick. It's administered in two different steps. So the first IV injection contains our cancer cell engager. And then there's a time lag and our preliminary studies were showing us our half-life is about one hour. So this is about four to five half-lives. That allows free unbound fab to get cleared from plasma.

And then we do a second IV injection of our T cell engager. And that way the cancer cells are primed, pre-coded with cancer cell engager, and now we're just recruiting T cells. So this should be our therapeutic dose. So for this model, untreated mice usually succumb to disease around three to four weeks. And we just tried a bunch of different dose iterations here. But this orange bar here, you can see this is after 75 days, these mice had a complete response.

And this was at a 60 microgram dose, which is a typical standard dose for if you were testing like an IgG. So 60 micrograms of cancer cell engager, five hour lag, which is four to five half lives, and then 60 micrograms of T cell engager. And we observe self-assembly in vivo, successful recruitment of T cells and, survival of the mice. So this was fantastic. so we really wanted to scale this up, but then we were also wondering, we can really play around with the ratios that

traditional bi specific antibodies cannot achieve, right? We can do like a, we can, we can fine tune the dose on the T cell side. And so we were wondering if we could get away with having the same efficacy, but decreasing the dose of T cell engager, and that should help our toxicity issue. So we set up a very similar experiment, same cell line, T cells. can see our cell count here and same dosing regimen. So we, we dose on day zero, one hour post-inoculation same treatment. So 60 micrograms of our cancer cell engager.

our time lag and now each cohort of mice get a different dose of T cell engager. So this would be like a one to one, this would be three to one, 10 to one and 30 to one. So we're just kind of titrating down our T cell engager dose. And I've simplified the survival curve here, but when you look at the mice survival, first of all, this is 150 days. This is five months, just a single dose. We see this pink bar here. This is the six microgram.

T cell Engager dose and we see complete response. And if you look at the day 141, I have this images, they're completely clean of cancer. if you go higher, if you add more T cell Engager, you actually see less survival, not as, not as many mice survived. Same. If you go too low, you're just kind of diluting out your Engager and it's not as effective. So it's kind of this hook effect where we observed at least in our models that a 10 fold reduced dose on the T cell Engager side.

was actually optimal and was better than this blue bar here is our one-to-one. So traditional, this would be kind of like a bispecific antibody formulation. We see long-term survivors and we see increased survival over untreated, but there's something magical about this consecutive dosing that we really like with our technology. So really excited with this. We can get away with less T cell engager. And this has proven to we've demonstrated in vitro that

this significantly reduces cytokine release compared to clinical standard bi specific antibodies. So once we got here, you know, all this has been published. what I've been working on the last year or year and a half, I've been really asking the question, how do we truly approach treating patients, cancer stoichiometrically and making very, very patient specific doses? You know, we can take some blood, analyze their cancer cells for antigen expression levels.

take their tumor burden, their T cell counts, and we should be able to custom make a dose specific for that patient. And as a human, we can kind of intuitively guess what that might look like, but I thought it would be so much better if we could get computers to model this for us. So we have certain inputs, we can plug that into a model and the model will predict a dose for that patient. And we're still trying to patent this, so I have to be a little bit vague in how this works, but I can at least describe the concept.

We take a patient, the actual sample, you know, PBMCs from a cancer clinic, and we can quantify our target receptor, antigen receptors on the cell surface. We can quantify the tumor burden, the T cell counts, all that data into our model that we've built. And our model will then suggest or predict doses, the best and second best doses that should work for that patient. And this has led us to a lead asset that we are developing.

trying to get through IND currently, but just the idea stands where we validated that we can actually get patient samples, plug it into the model and the model predicts doses that are actually the best dose compared to other controls. so we're really excited about this. want

to, I want to show you just a teaser piece of data. So this is the model in schematic form. Basically our inputs are just tumor burden, number of T cells, cancer cell identity. This is just their antigen profile.

And then our system is all the different dose combinations. So we've done bi- specific, tri specific up to four specificities we've modeled in all different ratios and variations. And the model computes based on how many cancer cells die and how many T cells survive. And then it will predict a dose. So I just want to show, I have to keep it kind of vague, but this data is coming out. We're going to submit this manuscript in a couple of weeks, but here's an in vivo model.

This is a late stage lymphomas. We've allowed the disease to disseminate for three days. And you can see untreated mice pretty full of cancer. This is an FDA approved biosimilar that we actually purchased from ichor bio Some efficacy, can see a mouse responded pretty well, less tumors overall, but still pretty prevalent cancer. And this is our AI predicted MATCH formulation. It's just a single dose, multi-specific, and it completely ablated the cancer cells. So we're really, really excited about this model.

And that's what we feel is the next frontier for us. So just in summary, MATCH as a modular, multi-specific T cell therapy. That's very tunable. We can exchange cancer cell binders. can exchange effector cell binders. We can get successful self-assembly in vivo with successful recruitment of effector cells. and very interestingly, we observe that maybe perhaps this one to one ratio that a traditional bi specific antibody is

stuck in isn't the best for all situations. So we're really excited about the versatility of our technology and we're grateful that we could share it with you all today. If you have any questions or if you want to reach out to me, here's all my contact information. And thank you, James for setting this up. This is an awesome seminar and I'm super excited to be a part of it. So thank you guys.

James Mosedale (14:06)
Not at all, so I mean that was really, really interesting and I'm sure our viewers will have lots of questions about your research.

However, to keep things brief, what is the best way that the community can support your work?

Tommy (14:20)
Yeah, currently we're looking for pharma partners to help develop our technology. We're interested in currently lymphoma, but we want to get into multiple myeloma and leukemia. And so we'd love to work with partners on developing this further.

James Mosedale (14:40)
That's great. To our community, you can find all the details and contact information on Tommy's final slide, and we'll make sure to include it in the show notes. Thank you for tuning in to the Bench Brief. You can find all our episodes and transcripts on ichor.bio forward slash the Bench Brief, and please subscribe on YouTube to ensure that you never miss an update. Join us again next time for more essential insights by scientists for scientists...