Epigenetic landscape of acute myeloid leukemia

On-demand webinar

Summary:

Acute myeloid leukemia (AML) exhibits significant genetic and transcriptional heterogeneity. Modulating epigenetics shows increasing potential in tackling unmet clinical needs in AML. Get up to speed with recent advances, challenges and future therapeutic directions in the epigenetic landscape of AML.

Epigenetic landscape of AML

Moderator and speakers:

Brian Huntly (University of Cambridge, UK)

Carsten Müller-Tidow (Heidelberg University Hospital, Germany)

Video Transcript

00:00 - 00:17: Hello, and thank you for joining today's virtual conference on the Epigenetic Landscape
00:17 - 00:23: of Acute Myeloid Leukemia, part of Abcam's three-part series on translational and therapeutic
00:23 - 00:27: advances in oncology and epigenetics.
00:27 - 00:31: All participants are automatically muted and their camera switched off.
00:31 - 00:36: However, please submit your questions at any time during today's event using the Q&A box
00:36 - 00:38: found at the bottom of your screen.
00:38 - 00:43: I'm excited to inform you that Abcam is approved as a provider of continuing education programs
00:43 - 00:49: in the clinical laboratory sciences by the ASCLS PACE program.
00:49 - 00:53: PACE credits are available for this live session, and a link to request credits will be sent
00:53 - 00:56: in the chat at the conclusion of this event.
00:56 - 01:02: You can also find more details regarding requested credits in our post-event email.
01:02 - 01:07: I'm now going to share a quick presentation by Abcam.
01:14 - 01:22: Great, so before I hand over to our chairs, I wanted to go through a bit of information
01:22 - 01:26: and what we're doing behind the scenes to help you get those right proteomic tools to
01:26 - 01:29: interrogate the complex pathways in AML.
01:29 - 01:34: And I want to start off by, you know, just giving you a quick idea on the state of play
01:34 - 01:40: today and what we have to address the state of play of AML, and what we're doing to advance
01:40 - 01:44: the tools that we make so that they can be fit for purpose for the challenges that are
01:44 - 01:50: encountered in the complex area of acute myeloid leukemia.
01:50 - 01:55: So, to start off, you probably are all aware of the complexities of epigenetic mechanisms,
01:55 - 02:02: whether it's modifications of DNA or histones, involvement of readers, writers, erasers, chromatin
02:02 - 02:07: remodelers, and then, of course, the emerging and exciting field of epitranscriptomics,
02:07 - 02:10: which you will hear about later today in this talk.
02:10 - 02:14: But I just thought, you know, to make you aware that we have tools to interrogate all
02:14 - 02:20: these different areas, and, you know, the complexities that are involved with these
02:20 - 02:24: areas can be tackled with certain reagents that we have in the catalog, and they could
02:24 - 02:27: include any of these areas here.
02:27 - 02:33: And I will now move on to an area which is of interest to most of you, or some of you
02:33 - 02:39: who have already explored AML, which is the well-known area of histones and DNA modifications.
02:39 - 02:45: As you can see, there are quite a few tools that you can use to test promo domains or
02:45 - 02:52: different areas of histone modifications with these antibodies, and these are already being
02:52 - 02:56: used in the field of AML by different researchers, as you can see by the citations on the right
02:56 - 02:57: of your screen.
02:59 - 03:05: Moving on to the area of chromatin remodelers, of course, there is a lot of interest in this area
03:05 - 03:08: as well, with drug development efforts also here.
03:08 - 03:12: And, you know, there are lots of different ways where you can interrogate these with
03:12 - 03:18: different pathways or different…it could be Polycombs, it could be the Trithorax, or
03:18 - 03:22: it could even be the SWI/SNF complexes.
03:22 - 03:25: And, you know, we've got a whole range of reagents to address those.
03:26 - 03:33: And finally, the emerging area of RNA epigenetics, or commonly known today as epitranscriptomics,
03:33 - 03:36: which is where you see modifications of RNA.
03:36 - 03:40: You know, METTL3, of course, is being investigated as a potential drug development
03:41 - 03:45: candidate, and, of course, there are also other areas there of interest to researchers.
03:47 - 03:51: And I think the exciting part about what we do here at Abcam is to innovate.
03:51 - 03:58: And, you know, we are aware of the heterogeneity of AML and the somatic mutations that do occur.
03:58 - 04:04: And here is an example of how we are innovating to tackle certain targets that are heavily
04:04 - 04:07: mutated in hematological malignancies.
04:07 - 04:13: Of course, EZH2 isn't widely yet implicated as a mutated target in AML, or maybe there's
04:13 - 04:17: recent research showing that's the case.
04:17 - 04:21: However, in follicular lymphoma, it has been known to be mutated.
04:21 - 04:26: And, of course, the DNMT3A and the IDH genes are well known to be mutated in acute myeloid
04:26 - 04:27: leukemia.
04:27 - 04:34: So, to address that, we have made tools that can detect mutations and help you in the immunophenotyping
04:34 - 04:38: of these mutations in tissues or biopsies.
04:38 - 04:41: So, do feel free to check these out on our website.
04:42 - 04:48: And in addition to making the right tools, we also developed the right techniques that
04:48 - 04:48: you can use.
04:48 - 04:54: Again, in the field of epigenetics, we are aware of emerging next-generation technologies
04:54 - 04:59: like CUT&RUN and CUT&Tag, and we are addressing those with the right offerings
04:59 - 05:00: in our portfolio.
05:00 - 05:02: So, feel free to check those out as well.
05:03 - 05:07: And finally, the last thing I do want to highlight is the innovation we're doing in terms of
05:07 - 05:12: giving you highly sensitive reagents in terms of antibodies.
05:12 - 05:19: And this is one where we won an award recently by Citeab, and it is recombinant multiclonals.
05:19 - 05:25: And what that simply means, it's a combination of monoclonals in the same vial, all to different
05:25 - 05:27: epitopes of that protein.
05:27 - 05:33: So, you get that sensitivity of a polyclonal with a monoclonal, but in this case, multiple
05:33 - 05:35: monoclonals in the same vial.
05:36 - 05:38: Well, that's it for me.
05:38 - 05:42: If you have any questions at any point, you can email us directly by reaching out to us
05:42 - 05:43: at this email.
05:43 - 05:47: I'm now going to hand over to our meeting chair, Professor Brian Huntley.
05:47 - 05:53: Professor Huntley is the head of the Department of Hematology at the University of Cambridge.
05:54 - 05:59: The interest of his laboratory is in understanding how normal stem cells and progenitor cells
05:59 - 06:04: function and are supported during the stepwise evolution of hematological malignancies,
06:05 - 06:07: particularly AML.
06:07 - 06:08: So, over to you, Brian.
06:08 - 06:09: Thanks very much.
06:10 - 06:11: Thanks very much, Vinit.
06:11 - 06:13: And so, welcome from me also.
06:13 - 06:17: This is a very exciting series of talks that we have for you.
06:17 - 06:22: And it covers, as Vinit says, the sort of gamut of the epigenetics of acute myeloid
06:22 - 06:28: leukemia, all the way from, you know, the very emerging field of epitranscriptomics
06:28 - 06:36: through histone writers, these are H3K27, and then to the ATP-dependent chromatin remodelers.
06:36 - 06:41: And I think there's quite a lot for everyone in terms of basic science through to translational
06:41 - 06:43: and with line of sight to the clinic.
06:43 - 06:48: And Sydney is going to tell us today about EZH2 inhibition and its induction of blast
06:48 - 06:50: differentiation in AML.
06:50 - 06:52: Sydney, over to you.
06:53 - 06:55: Hi, I'll just share my screen.
06:56 - 06:59: Thank you, Dr. Huntley, for that introduction.
06:59 - 07:04: And thank you to the organizers for inviting me to talk today.
07:04 - 07:05: I'm Sydney Flaubert.
07:05 - 07:08: I'm an MSTP student at Ohio State.
07:08 - 07:13: And I'm currently working in the Leukemia Drug Development Laboratory at the University
07:13 - 07:18: of Cincinnati, under the mentorship of Dr. John Byrd and Dr. Erin Hartline.
07:18 - 07:23: And today, I would like to talk about our project, looking at how the inhibition of
07:23 - 07:29: the Enhancer of Zeste Homolog 2 or EZH2 can induce blast differentiation in AML.
07:30 - 07:32: I don't have any disclosures.
07:32 - 07:37: So, real briefly, I just want to run through a quick background of AML.
07:38 - 07:43: So, it is a prevalent proliferation of immature myeloid cells known as blasts.
07:43 - 07:47: And so, in a healthy person's blood and bone marrow, they have lots of different types
07:47 - 07:53: of white blood cells, neutrophils, macrophages, monocytes that fight infection.
07:53 - 07:59: And when a patient develops AML, these immature white blood cells start to accumulate and
07:59 - 08:02: they start to take over the blood and the bone marrow.
08:02 - 08:05: And so, these are non-functional white blood cells.
08:05 - 08:06: So, they can't fight infection.
08:06 - 08:09: They reduce the number of healthy white blood cells present.
08:09 - 08:11: They reduce the space for the red blood cells.
08:11 - 08:16: So, patients are anemic and patients are very sick and typically need treatment as soon
08:16 - 08:17: as possible.
08:18 - 08:25: AML is the most common acute leukemia of adults with a median age of diagnosis of 65 to 70 years
08:25 - 08:26: of age.
08:26 - 08:28: So, it is a disease of the elderly.
08:29 - 08:33: And overall, prognosis with AML is pretty poor.
08:33 - 08:37: The average overall survival when patients are treated with chemotherapy is about six
08:37 - 08:37: months.
08:37 - 08:41: So, there's definitely a need in the field to continue to improve treatments.
08:43 - 08:49: The genetic heterogeneity in AML makes it pretty complex because it's characterized
08:49 - 08:52: by the sequential acquisition of mutations.
08:52 - 08:58: So, in the schematic is just an example of what could happen in a patient.
08:58 - 09:02: So, at diagnosis, you could have this clone one, which is the dominant clone that's made
09:02 - 09:07: up of mutation A, but you could also have this subclone that has mutation B.
09:07 - 09:14: And as the patient undergoes treatment, you can see typically some reduction if they respond
09:14 - 09:18: in the number of blasts, but you could also start to see this second clone starting to
09:18 - 09:19: grow out.
09:19 - 09:24: And so, then by the time the patient relapses, this second clone is starting to take over.
09:24 - 09:29: You can also have a third clone start to develop, which has mutation C.
09:29 - 09:34: And so, it's really tricky to treat because a patient's disease changes and evolves and
09:34 - 09:39: gains and loses mutations based on the treatment that a patient is given and the selection
09:39 - 09:40: pressure.
09:40 - 09:45: So, what a patient relapses with is not typically what they're diagnosed with, with
09:45 - 09:47: regards to mutational status.
09:48 - 09:52: There are a couple of different classes of gene mutations in AML.
09:52 - 09:54: These are some of the most common ones.
09:54 - 09:58: You have your signal transduction genes, which include FLT3 and NRAS.
09:58 - 10:04: You have your chromatin modifiers, which include ASXL1 and EZH2.
10:04 - 10:10: And you also have DNA modification genes, such as DNMT3A, TET2, and IDH1 and 2.
10:10 - 10:15: MPM1 is also commonly mutated in AML, and that's typically considered a class of its
10:15 - 10:16: own.
10:18 - 10:24: So, traditionally, many patients actually still receive 7 plus 3 induction consolidation
10:24 - 10:27: chemotherapy if they're fit enough to withstand treatment.
10:27 - 10:33: This is a very intensive treatment regimen, and patients typically have to be hospitalized
10:33 - 10:34: to undergo this therapy.
10:34 - 10:41: And so, because AML is a disease for the elderly, most patients can't undergo this type of
10:41
- 10:42: intensive treatment.
10:43 - 10:49: And so, the field of AML within the past 5 or 10 years has really evolved with newly
10:49 - 10:51: FDA-approved targeted therapies.
10:52 - 10:57: And so, the idea behind using targeted therapies, if we go back to the schematic, is that you
10:57 - 11:02: can give different drugs to target different mutations that a patient has.
11:02 - 11:08: So, you could give drug A, which would target clone 1, drug B would target clone 2, and
11:08 - 11:14: drug C would target clone 3, with the idea that as a patient's disease evolves, you can
11:14 - 11:19: give them therapy that's specific to their blasts with the hopes of limiting the effect
11:19 - 11:25: on normal white blood cells that are still hanging around, but also being specific to
11:26 - 11:28: what is providing a survival advantage in the blast cell.
11:29 - 11:35: And so, with regards to the newly approved therapies, venetoclax and azacitidine have
11:35 - 11:36: been used a lot.
11:36 - 11:40: It's a BCL-2 inhibitor, as well as a combination hypomethylating agent.
11:41 - 11:45: There's mitostorin and gilteritinib, which are FLT3 inhibitors, and then there are
11:45 - 11:50: ivacaftor for IDH1-mutated patients and enasidenib for IDH2-mutated patients.
11:53 - 11:58: So, the protein that we're specifically interested in is EZH2, which is the catalytic
11:58 - 12:05: core subunit of the polycomb repressive complex 2. This PRC2 complex, as you can see here,
12:05 - 12:11: is made up of a couple of different proteins that work together, and EZH2 is the actual protein
12:11 - 12:15: that catalyzes the trimethylation of lysine 27 on histone H3.
12:16 - 12:21: And this H3K27 trimethylation is associated with transcriptional repression.
12:22 - 12:28: EZH2 has been associated with a couple of different types of cancers, and it has been
12:28 - 12:33: described as overexpressed in breast cancer, prostate cancer, as well as lymphomas.
12:35 - 12:43: So, EZH2's role in AML is pretty complex. There are EZH2 mutations in AML. They're
12:43 - 12:47: pretty rare, about 4% of patients, but they do tend to be loss of function
12:49 - 12:55: and associate with poor prognosis. So, based on the literature, EZH2 mutations tend to be
12:55 - 13:02: early events in AML development. Also, it's important to know that EZH2 is located on
13:02 - 13:09: the Q arm of chromosome 7, and so its expression is also reduced in patients with monosomy 7
13:09 - 13:16: and deletion or loss of 7Q. And so, the obvious question is, why would we want to inhibit EZH2
13:16 - 13:23: if EZH2 mutations in AML are already loss of function? And I think this goes back to EZH2's
13:23 - 13:28: role in AML changes is not the same during AML development and maintenance, and there's some
13:28 - 13:37: literature supporting this. And so, it's kind of complex. It has been shown that EZH2 can
13:37 - 13:45: regulate granulocyte-macrophage progenitor cells in an MLAF9 model of leukemia. So, essentially,
13:45 - 13:52: if you knock out EZH2, they see differentiation of these macrophage markers. A paper came
13:52 - 13:59: out from Dr. Huntley's group a couple of years ago that really elegantly showed that EZH2 can
13:59 - 14:04: act as a tumor suppressor or an oncogene at different stages during AML development and
14:04 - 14:09: maintenance. And I would just like to highlight a couple of their findings, because I think Dr.
14:09 - 14:16: Huntley and his group really nicely showed how EZH2 is not regulating the same set of genes during
14:16 - 14:24: AML development as it is during maintenance. They used an EZH2 knockout mouse model, and what
14:24 - 14:32: they did is for this first mouse model is they knocked out EZH2 in the mouse using poly I:C and
14:32 - 14:37: then isolated bone marrow cells from these mice. And then what they did is they used…they
14:38 - 14:46: transfected the mice…sorry, the bone marrow with MLAF9 or AML1-ETO, which are fusion proteins
14:46 - 14:53: that can drive leukemia development. They took…once these bone marrow cells were
14:53 - 14:58: essentially AML cells, they put them into a primary transplant and monitored for survival.
14:58 - 15:04: Essentially, what this has done is it has knocked out EZH2 prior to AML development,
15:04 - 15:11: so they're seeing how EZH2 is functioning in the development of AML. They found that in
15:11 - 15:16: these mice with EZH2 knockout, they actually had a shorter overall survival compared to wild-type
15:16 - 15:23: EZH2, suggesting that EZH2 is functioning as a tumor suppressor during and prior to AML development.
15:26 - 15:32: However, in this second mouse model, they knocked out EZH2 after AML development. So, they,
15:32 - 15:39: again, isolated bone marrow and transfected with these AML fusion proteins, put it into a
15:39 - 15:44: primary transplant, allowed AML to fully develop, and then took splenocytes from these mice and
15:44 - 15:50: transplanted them into a second mouse and then knocked out EZH2. Essentially, this
15:50 - 15:56: model is looking at what happens when you inhibit EZH2 once AML is fully developed.
15:57 - 16:02: They found that knocking out EZH2 after AML development allowed the mice to live longer,
16:02 - 16:07: which is really suggesting that EZH2 is functioning more as an oncogene during AML
16:07 - 16:14: maintenance. This really suggested to us that if we use an EZH2 inhibitor in an AML
16:14 - 16:22: patient who has already transformed to AML, this could be a potentially surprising therapeutic
16:22 - 16:26: benefit. It might not be. I don't think it would be a great option in a patient with MDS who doesn't
16:26 - 16:31: have AML, but in a patient who has a high-level blast burden, we think this could be a good
16:33 - 16:40: treatment option therapy. The objective of our study was to understand the impact of EZH2
16:40 - 16:47: inhibition in AML blasts, which is based on the hypothesis that in AML patients without an EZH2
16:47 - 16:52: mutation, loss of EZH2 function may produce a phenotype that would allow for therapeutic
16:52 - 17:01: targeting without influencing normal hematopoiesis. To do this, we used EPZ011989
17:01 - 17:09: to inhibit EZH2 activity. EPZ is in the same family as tazemetostat, which is an FDA-approved
17:09 - 17:16: EZH2 inhibitor for relapsed refractory follicular lymphoma with an EZH2 mutation. So, EZH2 mutations
17:16 - 17:21: in lymphoma tend to be gain-of-function, so it makes sense to use an EZH2 inhibitor in these
17:21 - 17:26: patients. Tazemetostat is also approved for epithelial sarcoma with metastasis for local
17:26 - 17:33: advancement. The way EPZ works is it's an S-adenosylmethionine, or SAM, competitive
17:34 - 17:41: inhibitor. SAM is the protein that donates the methyl group to EZH2, and EPZ
17:41 - 17:48: competes with SAM in that SAM binding pocket, so only EPZ or SAM can bind. This allows
17:48 - 17:54: for EPZ to be specific to EZH2 function and not affect other histone methyltransferases.
17:57 - 18:03: So, we wanted to first start by looking at if EPZ works in our hands to inhibit EZH2
18:03 - 18:11: function in the WOM13 cell line. We treated WOM13 AML cells for 48, 72, and 96 hours,
18:11 - 18:19: and we see a nice reduction in H3K27 trimethylation. We could tell that we were hitting our
18:19 - 18:25: target, so we next wanted to look at proliferation when these cells were treated with the EZH2
18:25 - 18:31: inhibitor. We did a 7-day MTS assay using metabolic activity as a surrogate for proliferation,
18:31 - 18:37: and we can see at these higher concentrations of EPZ, we are getting a reduction in proliferation
18:37 - 18:44: with about an EC50, I'm sorry, at about 5 micromolar. It's a little higher than we
18:44 - 18:51: would like, but we are getting some reduction in proliferation. Because EZH2 inhibition is
18:51 - 18:58: associated with…I'm sorry, EZH2 is associated with differentiation, we wanted to look at CFU
18:58 - 19:04: assays in the WOM13 cells. The advantage of a CFU assay is it can tell us two different things.
19:04 - 19:10: It can tell us the differentiation state of the cells by looking at the total number of colonies
19:10 - 19:16: with a higher number of colonies associated with an increased stemness or a de-differentiated state,
19:16 - 19:21: while a fewer number of colonies is more indicative of a reduced stemness in a more
19:21 - 19:26: differentiated state. We can also look at the size of the colonies as kind of a surrogate
19:26 - 19:31: for proliferation capacity as well, because if a cell is able to divide more, the colony will be
19:31 - 19:38: bigger and more dense. This is a 14-day CFU assay in the WOM13 cells, and we see a modest
19:38 - 19:43: reduction in the total number of colonies, which is not surprising because WOM13 cells are
19:43 - 19:49: immortalized and difficult to differentiate. What was really interesting is when we look at
19:49 - 19:54: the size of the colonies. In our DMSO-treated samples, we have really nice, dense colonies
19:54 - 20:00: with lots of cells, relatively speaking.
20:00 - 20:06: When you look at the 1-micromolar EPZ-treated condition, these colonies are smaller. They're
20:06 - 20:10: more diffuse, suggesting that we are reducing proliferation a little bit, which is in line
20:10 - 20:20: with what we saw in our MTS assays. That led us to next ask if the inhibition of EZH2 in vivo
20:20 - 20:26: associates with improved survival. We saw modest effects with the WOM13 cells in vitro,
20:26 - 20:33: so we did expect a modest increase in survival in the WOM13 luciferase model, especially because
20:33 - 20:39: this is a very aggressive model. What we did is we engrafted the mice on day minus three,
20:40 - 20:46: and then on day zero, they were randomized to receive vehicular 80 mg per kg EPZ. This dose
20:46 - 20:52: was increased to 300 mg per kg on day 10, and they were gavaged daily until endpoint. We did
20:52 - 20:57: have a subset of mice that we did weekly bioluminescence imaging to look at disease burden
20:57 - 21:02: as well. We can see that we get a modest survival advantage in the EPZ-treated mice,
21:02 - 21:06: about three days, which is aligned with what we're seeing in our in vitro study,
21:07 - 21:12: but does show that we can slow down the engraftment of the WOM13 cells with an EZH2 inhibitor.
21:15 - 21:20: When we look at disease burden on day 21, we can visually see a reduction in the amount
21:20 - 21:25: of disease in the vehicle compared to the EPZ-treated group, with the caveat that we
21:25 - 21:31: did lose one mouse in our imaging cohort. So, this suggests to us that there are
21:31 - 21:35: some advantages to inhibiting EZH2 in the WOM13 cells.
21:37 - 21:40: Because WOM13 cells are immortalized and difficult to differentiate,
21:41 - 21:47: we wanted to look at the effect of inhibiting EZH2 in primary AML samples. The advantage of using
21:47 - 21:53: this is, of course, it's more related to actual patient samples, but they also have the capabilities
21:53 - 22:02: to differentiate. So, we started by checking to see that the EZH2 inhibitor worked in primary
22:02 - 22:07: AML samples. This is a representative Western blot after seven days of treatment with EPZ,
22:07 - 22:13: and we can see a nice reduction in the H3K27 trimethylation levels. I also included
22:14 - 22:20: EPZ-treated WOM13 lysates on this blot as a positive and negative control for our antibodies.
22:20 - 22:26: So, we feel confident that we are hitting our target. When we then looked at proliferation
22:26 - 22:31: by an MTS assay, we don't really see that reduction in proliferation that we saw at the
22:31 - 22:37: higher concentration in the WOM13 cells. There was this one patient that looks like it potentially
22:37 - 22:42: had a little bit of a differentiation burst, but in general, we're not killing the cells.
22:44 - 22:51: When we look at colony formation assays in primary AML samples compared to normal
22:51 - 22:58: CD34-positive donor cells, we do start to see a difference. These normal donor CD34-positive
22:58 - 23:08: cells were isolated from bone marrow samples that were donated, and then we selected CD34-positive
23:08 - 23:12: cells from these healthy donors. The advantage of CD34 is that it is an immature white blood
23:12 - 23:18: cell marker, so this is essentially looking at what happens when we inhibit EZH2 in a healthy,
23:18 - 23:24: immature white blood cell. This is a 14-day CFU assay, and in our AML blast samples,
23:24 - 23:30: we are seeing a reduction in the colony-forming capabilities of the AML blasts at the 1-micromolar
23:30 - 23:37: EPZ condition. When we look at the CD34-positive cells, we don't see that reduction
23:37 - 23:44: in the colony count, which again suggests that EZH2 is regulating different sets of genes
23:46 - 23:52: in different cell types. Potentially, using an EZH2 inhibitor in an AML blast
23:52 - 23:55: sample would be more specific to the blast cell.
23:58 - 24:03: When we look at the colonies visually, again, we see diffuse, large, dense colonies in the
24:03 - 24:08: DMSO-treated sample that we don't see in the 1-micromolar EPZ-treated sample,
24:08 - 24:13: suggesting that we are inducing some differentiation and potentially reducing
24:13 - 24:21: proliferation a little bit. CFU assays are a good indication of differentiation, but we also wanted
24:21 - 24:27: to see if we could see any cell surface marker changes. We treated primary AML samples for
24:27 - 24:34: seven days with 1-micromolar EPZ, and then looked at CD11B, which is a pan-white blood cell
24:34 - 24:40: marker, as well as CD14 and CD16 expression, which are markers of monocytes. Overall,
24:40 - 24:46: we see a modest, it's very modest upregulation in the cell surface marker expressions, about 30%,
24:47 - 24:52: but we do see some differences there. When we take the cells out to day 14,
24:52 - 24:57: we can start to see more morphological changes of differentiation. In the DMSO-treated sample,
24:57 - 25:04: we have a blast-like phenotype for the cell with a large nucleus and a scant cytoplasm, and in the
25:04 - 25:10: 1-micromolar EPZ-treated condition, we have a more macrophage/monocyte-type phenotype with a large
25:10 - 25:19: cytoplasm filled with vacuoles and a small nucleus. We do see morphological changes of differentiation,
25:19 - 25:24: which kind of back up the CFU assays. What was particularly interesting to us is that we saw EZH2
25:34 - 25:42: inhibition promoted the upregulation of CD38, and CD38 is a marker that is on monocytes, but it can
25:42 - 25:47: also be seen on B and T cells and plasma cells, and it was a really interesting target for us because
25:47 - 25:54: there is daratumumab, which is an anti-CD38 monoclonal antibody, and so this presents a potential for
25:54 - 25:59: therapeutic combination. To go through what's in this graph, it's from seven days of EPZ-treated
25:59 - 26:09: primary AML samples, and we see this nice upregulation of CD38. Because this has the potential for
26:10 - 26:15: therapeutic combination, we have some preliminary data looking at combinations. What we have done
26:15 - 26:24: is treated a couple of primary AML samples with DMSO or 1 micromolar EPZ for 14 days, and what we did
26:25 - 26:35: is a couple times throughout the treatment process, we spun down the cells and re-added media and drug,
26:36 - 26:42: and then on day 14, the cells were spun down again and then were resuspended in media containing 20
26:43 - 26:48: micrograms per mL of daratumumab, and the cells were incubated with daratumumab for 24 hours,
26:48 - 26:56: and then we stained for F-actin, and the idea behind this experiment is that potentially
26:56 - 27:02: we can induce fratricide. So, in the DMSO-treated cells, the cells are going to
27:02 - 27:08: maintain more of a blast phenotype. They're not going to be able to express CD38,
27:08 - 27:12: so daratumumab isn't going to be able to bind, and the cells will be able
27:12 - 27:17: to induce cell death. However, in the one micromolar EPZ-treated condition, the idea
27:17 - 27:23: is that we will have an upregulation of CD38, as well as some differentiated AML cells that
27:23 - 27:30: when you add daratumumab, they can bind to induce fratricide, and F-actin is one of the proteins
27:30 - 27:36: that help stabilize this formation, and so what we are looking for is F-actin staining between
27:36 - 27:42: two cells to suggest that at least conjugates are being formed as kind of preliminary data.
27:43 - 27:49: And so, we've done this a couple of times, so this is an example of a DMSO-treated sample
27:49 - 27:55: and a one micromolar EPZ-treated sample, and you can see by these white arrows, an increased binding
27:55 - 28:00: in the EPZ-treated samples compared to the DMSO, so this is really encouraging for us.
28:01 - 28:08: So, in conclusion, we looked at EZH2 inhibition in MOM13 cells and didn't find any resulting cell
28:08 - 28:15: death, but I did see a reduction in colony formation, especially size. In vivo treatment
28:15 - 28:21: with EPZ prolonged survival in the MOM13 luciferase xenograft model. We've also looked
28:21 - 28:28: at EZH2 inhibition in primary AML samples and seen a modest induction of differentiation,
28:28 - 28:35: and we also saw EZH2 inhibition in primary AML samples upregulate CD38 expression
28:35 - 28:42: for potential combination therapy. So, kind of our future directions is we want to pre-treat
28:42 - 28:47: primary AML samples with EPZ prior to engraftment to look at differences in engraftment capabilities,
28:47 - 28:53: and we think this would be a really great addition, to what we've been working on. So,
28:53 - 28:58: the idea is that if we treat a primary AML sample that we know can engraft into an EPZ-treated
28:58 - 29:04: NRGS mouse, in vitro with DMSO, you're going to maintain this blast-like phenotype that's going
29:04 - 29:09: to be capable of engrafting in an NRGS mouse, and they're going to meet ERC as expected.
29:10 - 29:15: However, if you treat the same sample with EPZ, you're going to get some differentiation,
29:15 - 29:20: so there will be fewer blasts, decreased stemness, so it's going to engraft into a mouse slower,
29:20 - 29:24: and so we would expect these mice to have a prolonged survival.
29:24 - 29:29: We're also looking to increase our number of anti-CD38 conjugation formation experiments. We
29:29 - 29:34: know that's an NF2, so we definitely want to get that up, and if that looks as expected, which we
29:34 - 29:41: would expect based on our flow data, we want to combine EPZ and daratumumab in vivo to look at its
29:41 - 29:48: effects on overall survival. So, I'd like to thank my mentors, Dr. Birtlein and Dr. Bird,
29:48 - 29:52: as well as everyone in the Leukemia Drug Development Lab at the University of Cincinnati.
29:55 - 29:59: The Experimental Hematology Lab at Ohio State, as well as the patients and their families for
29:59 - 30:04: donating the samples that were used for this work. Thank you for your time, and I'd be happy
30:04 - 30:10: to take any questions. Great. Thank you, Sydney. That was a really, really interesting talk,
30:10 - 30:15: and thank you for being gracious with some of the preliminary data that directed your study as well.
30:16 - 30:20: We have a couple of questions coming in. A lot of them are very kind of therapeutic,
30:20 - 30:25: and I know that you're a student and you're not an experienced clinician,
30:25 - 30:29: so I might try and help you out with some of these. So, Finn Weiland has asked,
30:30 - 30:33: this relates very much to the other parts of your talk when you were setting the scene,
30:33 - 30:37: although the side effects might be more significant, is there a positive impact
30:37 - 30:43: of giving several targeted therapeutics at once? Is this to prevent some mutated clones from surviving,
30:43 - 30:50: or is this irrelevant? So, I think that depends on yes and no, depending on what
30:50 - 30:57: drugs you're giving. So, in this case, because it's not targeting a specific mutation, it's more
30:57 - 31:03: to induce cell killing, because we don't see a ton of apoptosis with an EZH2 inhibitor. However,
31:03 - 31:08: I think that idea in the field of giving multiple drugs to prevent some mutated clones from
31:08 - 31:12: surviving is kind of the goal with targeted therapies and combination therapies as well.
31:12 - 31:18: I completely agree. I've delivered some of these epigenetic therapies in the clinic,
31:18 - 31:25: and they are, in fact, nothing as good as a monotherapy in AML, basically. So, I would
31:25 - 31:31: completely concur. Again, something of a philosophical question, how does the
31:31 - 31:36: differentiation that's induced with the EZH2 inhibitor compare to other differentiation
31:36 - 31:40: therapies? So, I suppose a more specific question would be, have you been able to look at it by
31:40 - 31:46: comparison with ATRA or IDH inhibitors, or other things that do the same thing?
31:47 - 31:54: Yeah, we've done ATRA and DHODH inhibitors, brecanar, and we do see the similar upregulation
31:54 - 32:01: of CD38. We actually got the idea based on Dr. Shishila Tridentapani's paper,
32:01 - 32:06: looking at ATRA and daratumumab. So, we have used that as our positive control in our primary
32:06 - 32:12: samples, and we do see that as well. Okay, a question from Jiangping Li,
32:12 - 32:18: nice talk. I have two questions. Do the AML blasts used for our EPZ treatment have EZH2 mutations?
32:18 - 32:25: And the second one is, since loss of function mutations of EZH2 contribute to AML,
32:25 - 32:29: how can you reconcile this if they've got EZH2 mutations?
32:30 - 32:35: So, the blasts we used in our samples did not have EZH2 mutations. We specifically
32:35 - 32:42: screened for that. So, in our studies, no. And I think the reason why inhibiting EZH2
32:42 - 32:47: doesn't aggravate AML progression is I think it's regulating, and I think Dr. Huntley,
32:47 - 32:53: your group showed this really nicely, different sets of genes in a fully developed AML cell as
32:53 - 33:00: compared to a pre-AML development of more of an MDS phenotype. So, I think because we used AML
33:00 - 33:05: blasts, and these were like apheresis samples and bone marrow samples from patients with AML,
33:06 - 33:13: these were mainly made up of fully transformed blast cells. And so, I think if you did a more,
33:13 - 33:21: if you, yeah. So, again, I would agree. I think that we can come back to this. This is a question
33:21 - 33:26: I have. It might be different for different AML genotypes, because I think they're wired
33:26 - 33:33: in different ways. I mean, I know that some people have been looking at EZH1, EZH2 inhibitors,
33:33 - 33:38: and there is a school of thought for a lot of these loss of function mutations that,
33:38 - 33:43: because they're often in epigenetic regulators that are utterly required, and you can tell that
33:43 - 33:51: from mouse experiments where germline loss is embryonic lethal, that targeting residual function
33:51 - 33:59: of these loss of function mutations might be, you know, a ploy. Can I ask, does your inhibitor
33:59 - 34:05: have any effects against EZH1, or is it highly specific to EZH2? It should be specific to EZH2.
34:06 - 34:14: Okay. Yes. Okay. So, just coming on to the strategy with anti-CD38,
34:15 - 34:19: we've got someone who's very bashful, who won't give their name. Anonymous attendee,
34:19 - 34:25: does the combination therapy require an anti-CD38 and monoclonal antibody, or would
34:25 - 34:31: you think inhibition of CD38 would work as well? So, do you think it's just a target for these
34:31 - 34:36: cells as a, you know, perhaps a readout of differentiation, or do you think it's the
34:36 - 34:44: function of CD38 that is required? I think it's more a target. I don't, haven't seen
34:44 - 34:51: serotubamate, or not serotubamate, sorry, many CD38 inhibitors induced, or knockout,
34:51 - 34:55: CD38 knockout cells induced cell death. So, I would think it's more of a target,
34:55 - 35:00: and the idea being if we can get some slightly functional white blood cells present,
35:00 - 35:07: we can actually induce killing of the blast itself, if that answers the question.
35:07 - 35:10: So, perhaps I could come back to genotype, because I think that, you know,
35:11 - 35:16: one of the fascinating things, but one of the sort of infuriating things about AML is the
35:16 - 35:21: genetic heterogeneity, and just, well, just everything you look at is heterogeneous.
35:22 - 35:31: Do you have a feel for any potentially sensitive genotypes to the inhibitors? And second part is,
35:31 - 35:40: I saw the upregulation of CD38, but it was quite broad. Again, does the induction of CD38
35:42 - 35:46: relate in any way to genotype? Now, you might not have looked at enough samples to,
35:47 - 35:50: but I'm putting you on the spot, sorry. Yes, unfortunately, the answer is we've
35:50 - 35:55: not looked at enough samples yet. We probably, I think it's like an N of 12 that we've looked at,
35:55 - 35:59: so I don't, I don't think we can make any conclusions, because as you mentioned,
35:59 - 36:05: it's very broad, and so looking at four or five samples, that has to have a bigger change,
36:05 - 36:09: I don't think. It's just, it's, it's, I think it's a really important point that we can come
36:09 - 36:15: back to in the discussion, because I do think that choosing the right patients for these inhibitors,
36:15 - 36:21: and particularly if you're suggesting, as you are, you know, a combination, that might be even more
36:22 - 36:29: horses for courses. Great, any other questions coming in for Sydney at all?
36:31 - 36:37: Okay, can I, can I just ask, I mean, people have inferred non-catalytic functions to EZH2,
36:38 - 36:43: and certainly for some of the epigenetic regulators, such as, you know, CREB-BP that I work
36:43 - 36:50: with, we can show that non-catalytic functions can be quite important. Have you been able to
36:50 - 36:57: look at inhibition versus maybe a genetic ablation, in terms of the degree of effect
36:57 - 37:04: that you see, and do you think it's all catalytic? I have not had a chance to look at that,
37:06 - 37:12: we don't have, we haven't had a chance to do that yet. If I had to guess, I would think,
37:12 - 37:18: because this is an inhibitor specifically that competes with the methyl group, I don't know if
37:18 - 37:23: I might not be answering your question, but I would think it'd be more related to the, like,
37:23 - 37:31: the DNA methylation and transcriptional bound effects than EZH2 non-canonical function.
37:31 - 37:36: Okay, okay, fantastic. Okay, well, thank you very much, Sydney, that was a really nice talk,
37:36 - 37:42: and you handled the questions very well. So we'll move on now to our last speaker,
37:42 - 37:52: who is Richard Sintori from Foghorn Therapeutics. Richard is a principal scientist, and he is the
37:52 - 38:01: biology lead for the molecule FHD286 that he's going to talk about. So over to Richard now,
38:01 - 38:05: and the title of his talk is Discovery of Novel BAF Inhibitors for the Treatment of
38:05 - 38:09: Transcription Factor-Driven Cancers. Richard, over to you.
38:10 - 38:14: Thank you very much, Dr. Huntley, and yeah, thank you to Vinit and
38:15 - 38:19: Sarah and everyone else at Abcam who helped make this event possible today.
38:23 - 38:32: All right, so yeah, so here are my disclosures. Okay, so at Foghorn, we're interested in
38:32 - 38:40: gene control networks, and specifically how we can overcome chromatin and transcriptional
38:40 - 38:47: dysregulation to treat serious diseases like cancer. And one major regulator of chromatin
38:48 - 38:52: structure in gene expression is the mammalian switch-SNF family of chromatin remodeling
38:52 - 38:59: complexes, and these are also referred to as BAF complexes. And these complexes are large
38:59 - 39:05: molecular machines which are modularly and combinatorially assembled using the products of
39:05 - 39:10: about 29 different genes to make these three final form complexes, which we refer to as canonical
39:10 - 39:16: BAF, polybromo-associated BAF, and non-canonical BAF. And importantly, each of these three
39:16 - 39:22: complexes incorporates one of two mutually exclusive ATPase subunits called SMARCA4 or
39:22 - 39:29: SMARCA2. These are also referred to as BRG1 or BRM. And these provide the catalytic activity
39:29 - 39:36: required for nucleosome remodeling. Now, work from our scientific founders, Seagal Kadosh and
39:36 - 39:43: Jerry Crabtree, among others, have shown that these complexes are highly mutated and dysregulated in
39:43 - 39:50: cancer, with more than 20% of all cancers harboring a mutation in at least one subunit of these BAF
39:50 - 39:59: complexes. And this provides unique therapeutic opportunities, such as subunit dependencies that
39:59 - 39:59: are conferred.
40:00 - 40:05: by specific mutations, synthetic lethal interactions between different paralogs,
40:05 - 40:14: like SMARCA4 and SMARCA2, or ARID1A and ARID1B. And further, BAF complexes interact with
40:16 - 40:20: master oncogenic transcription factors to drive their transcriptional programs.
40:20 - 40:24: And we think this provides an additional opportunity for therapeutic intervention.
40:25 - 40:29: So at Foghorn, we've built a platform to produce these complexes, as well as these
40:29 - 40:37: interacting transcription factors. And we have several programs targeting either specific BAF
40:40 - 40:46: subunits or these transcription factor interactions through either traditional
40:46 - 40:52: small molecule catalytic inhibitors, protein-protein interaction disruptors,
40:52 - 40:58: or targeted therapeutics, sorry, targeted protein degradation mechanisms. And so today,
40:58 - 41:04: I'd like to tell you a bit about our work targeting the ATPase components,
41:04 - 41:08: SMARCA4 and SMARCA2, as a new approach to treating oncogenic transcription.
41:10 - 41:17: So we started this effort by screening the full-length SMARCA2 enzyme for molecules which
41:17 - 41:23: could inhibit its ATPase activity in vitro. And we initially identified this compound, FHT185,
41:24 - 41:30: which was a dual inhibitor of both SMARCA4 and SMARCA2. And this was highly selective against
41:30 - 41:39: other SNF2 family ATPases, such as CHT4, shown here. So this molecule was rapidly advanced
41:40 - 41:47: to this molecule that I'm showing here, FHT1015, which is another in vitro tool compound with
41:47 - 41:53: enhanced potency in the single-digit nanomolar, single-digit nanomolar potency in our biochemical
41:53 - 42:00: assay, maintaining very strong selectivity against CHT4. And same is true in our cell-based assay,
42:00 - 42:06: which is a transcriptional reporter assay for SMARCA4 or SMARCA2 activity. And we can see nice
42:06 - 42:16: potent dose-dependent inhibition, which is on target and specific. So these provided nice
42:16 - 42:20: tool molecules for us to interrogate the biology of these BAF complexes.
42:21 - 42:26: And so one of the first things we did was, since these are, since we're targeting chromatin
42:26 - 42:33: remodeling complexes here, we first looked at global changes to the chromatin accessibility
42:33 - 42:40: landscape across multiple cell lines that represent multiple different lineages. And so what you can
42:40 - 42:47: appreciate here is that when we inhibit BAF complexes with FHT1015, our tool molecule,
42:48 - 42:54: and measure chromatin accessibility using ATAC-Seq, we can see that only a subset of sites
42:54 - 42:59: are affected globally. And this is typically associated more with closing of chromatin or
42:59 - 43:06: loss of chromatin accessibility, as opposed to gain of accessibility. I'm not showing it here,
43:06 - 43:12: but we also characterized the genomic loci of these loss of accessibility sites and found that
43:12 - 43:19: these are largely not present at promoters. So we think they're mostly, this molecule is mostly
43:19 - 43:24: closing chromatin at enhancer sites. And when we look at the transcription factor motifs associated
43:24 - 43:31: with these closing or loss of accessibility sites, we commonly find among the top three
43:31 - 43:38: most enriched motifs, these master transcription factors, which represent major dependencies in
43:38 - 43:46: each particular lineage. So for example, in AML, we see strong enrichment of this P1 or PU1 binding
43:46 - 43:54: motif. And this is well known to be a strong dependency in AML subtypes compared to other
43:54 - 44:00: cell types. Same is true if we go into a prostate cancer cell line, we can see enrichment of the
44:00 - 44:08: FOXA1 binding motif, which is, FOXA1 is a strong selective dependency in prostate cancer cells,
44:08 - 44:15: and so on for other cell types. MyoD1 in this muscle cancer, rhabdomyosarcoma, and SOX10 in
44:15 - 44:21: melanoma. And so we think that this tells us that BAF complexes really collaborate with these
44:21 - 44:30: lineage-specific transcription factors and help these factors access their regulatory motifs
44:30 - 44:41: in the genome. So next to look at the impact of FHT1015 or BAF ATPase inhibition across different
44:41 - 44:47: cancer cell types, we just performed a simple cell titer glow experiment across a panel of about 60
44:47 - 44:56: cell lines representing a variety of tumor lineages. And at this early time point, day three
44:56 - 45:00: of treatment, what we could observe was that there were certain tumor types which seemed to
45:00 - 45:06: be exquisitely sensitive at this early time. And this included AML, as well as this other disease
45:06 - 45:10: here called uveal melanoma. And I'll talk about both of these diseases today, starting with uveal
45:10 - 45:19: melanoma. And so FHT1015, this is in a uveal melanoma cell line here called 92.1. And we saw
45:19 - 45:25: very potent activity in terms of growth inhibition, which was improved compared to other molecules
45:25 - 45:34: that have been used to treat this disease in the clinic. And we could also detect signs of
45:34 - 45:39: apoptotic activity through measuring caspase-3,7 activity with FHT1015.
45:44 - 45:49: So since uveal melanoma cells were extremely sensitive to this compound, we used this model
45:50 - 45:58: to perform some resistance studies to make sure that the activity of the compound was on target.
45:58 - 46:05: And so what we did was we grew this uveal melanoma cell line called MP41 in the presence of very high
46:05 - 46:12: concentrations of FHT1015, about IC90 concentrations, for several months and waited until
46:12 - 46:16: resistant clones could emerge. And you can see that these resistant cells here can tolerate
46:16 - 46:20: growth in the presence of the compound, whereas the parental cells cannot.
46:22 - 46:29: And this effect, this resistance seems to be specific to FHT1015 and not just to any generic
46:29 - 46:34: agent. So we tested a few others here and there's no difference in the growth inhibition of other
46:34 - 46:43: agents to these resistant cells. So we took these clones and we performed whole exome sequencing.
46:44 - 46:52: And remarkably, we identified a single point mutation in the target SMARCA4. And so we
46:52 - 46:56: hypothesized that this was what was causing resistance. And to test this, we introduced
46:56 - 47:02: this mutation into recombinant SMARCA4. And we also introduced the corresponding mutation in
47:02 - 47:12: SMARCA2 and measured the ability of FHT1015 to inhibit these enzymes in an in vitro assay.
47:12 - 47:19: And you can see that while the wild type enzymes could be completely and potently inhibited by FHT1015,
47:19 - 47:24: the mutant enzymes showed a decreased ability to be inhibited by the compound,
47:24 - 47:28: suggesting that this mutation was indeed what was causing resistance to these cells.
47:30 - 47:37: Further, we also took another compound, which we know binds and inhibits SMARCA4 and actually
47:37 - 47:42: binds to a different pocket, comes from a different series. And this compound was still
47:42 - 47:47: able to inhibit the growth of these resistant cells, just as it could in the parental cells.
47:48 - 47:54: So altogether, these data confirm that the activity of our compound is indeed on target.
47:57 - 48:04: Okay, so what about the mechanism? So to further understand that, we took the 921
48:04 - 48:10: uveal melanoma cell line and first performed ATAC-Seq. And again, what you can see here in
48:10 - 48:17: the uveal melanoma cells is that accessibility was lost in a subset of genomic loci,
48:17 - 48:22: and these seem to be highly enriched for SMARCA4 binding at the baseline.
48:23 - 48:30: And so what we did was we also did ChIP-Seq of H3K27 acetyl to map enhancers because
48:31 - 48:36: BAF complexes have been reported to function at both promoters and enhancers. And we asked whether
48:36 - 48:42: there was a difference in the ability of or in the inhibition of BAF complexes
48:44 - 48:49: to keep chromatin open at either of these genomic loci. So what you can see here is that while there
48:49 - 48:56: was a minimal impact of FHT1015 on chromatin accessibility at promoters, this effect was
48:56 - 49:07: much, was pretty robust at enhancers. We also looked at the transcription factor motifs that
49:07 - 49:11: were enriched at these loss of accessibility sites in uveal melanoma cells. And again,
49:11 - 49:18: we found this master transcriptional regulator SOX10, which is important in melanocytes.
49:19 - 49:28: We also found TF82-alpha to be the second most enriched motif. And these two, along with another
49:28 - 49:33: protein called MIDF, another transcription factor, these three transcription factors have been
49:33 - 49:40: reported to cooperate with one another to promote melanocytic gene expression in cutaneous melanoma
49:40 - 49:45: cells and primary melanocytes. So we thought these three were pretty important and decided
49:45 - 49:51: to focus more on these. And so we looked by ChIP-seq at the occupancy of these three
49:51 - 49:58: transcription factors in response to FHT1015 treatment and found that the occupancy of all
49:58 - 50:04: three of these transcription factors was markedly reduced upon BAF ATPase inhibition. So not only
50:04 - 50:09: are we closing chromatin at the binding sites of these important transcription factors, but we're
50:09 - 50:16: also preventing the transcription factors from actually accessing their sites on the genome.
50:20 - 50:26: So using H3K27 acetyl ChIP, again, we mapped super enhancers in this uveal melanoma cell line.
50:27 - 50:33: And by ChIP, again, for these other factors, we found that these different transcription factors,
50:33 - 50:38: along with SMARCA4, were highly enriched at super enhancers in this cell type. And
50:39 - 50:43: in these cells, super enhancers were associated with genes involved in
50:43 - 50:48: important melanocyte processes such as pigmentation and melanosome organization.
50:50 - 50:57: And so when we performed gene expression profiling after inhibition of BAF ATPases with our compound,
50:59 - 51:03: perhaps not surprisingly, these are the same types of gene sets that we found to be affected.
51:04 - 51:11: So SMARCA4 targets were downregulated along with genes involved in melanoma and pigmentation.
51:14 - 51:20: So like many important transcriptional, master transcriptional regulators,
51:20 - 51:26: SOX10 is known to drive its own expression through a feedback loop. And indeed,
51:26 - 51:34: we could observe very strong downregulation of SOX10 when we inhibit BAF complexes with this
51:34 - 51:40: compound in uveal melanoma cells. And this was detected both by bulk RNA-seq as well as by
51:40 - 51:47: looking at nascent transcripts at both the SOX10 gene body and the SOX10 enhancer element by ProSeq.
51:47 - 52:00: So since SOX10 was very strongly downregulated, and SOX10 also represents a major dependency
52:00 - 52:06: in these cells from the DEPMAP database at the Broad Institute, and SOX10 really stood out here
52:06 - 52:12: as a gene that is downregulated and critical for the survival of these cells. And so we tested the
52:12 - 52:17: contribution of SOX10 downregulation to the phenotype of the compound by overexpressing
52:17 - 52:23: SOX10 from a BAF-independent promoter. And you can appreciate that this did result in
52:23 - 52:28: at least a partial rescue of the growth inhibition phenotype induced by FHT1015.
52:33 - 52:40: So to test the effects of BAF ATPase inhibition in vivo, we turned to another tool molecule called
52:40 - 52:48: FHT2344, which is of the same series and is also a dual inhibitor of both BRG1 and BRM,
52:48 - 52:54: or SMARCA4, SMARCA2. And it's also selective against CHT4. And with this compound, when we
52:54 - 53:02: dose daily and orally into animals harboring a 92.1 uveal melanoma xenograft model, we could
53:02 - 53:08: observe a nice dose-dependent tumor growth inhibition with regression actually achieved
53:08 - 53:13: at the highest dose here. And importantly, all of these doses were very well tolerated by the
53:13 - 53:20: animals. We could also see using measuring SOX10 transcription as a pharmacodynamic marker,
53:20 - 53:25: we could see nice dose-dependent inhibition of SOX10 transcription,
53:25 - 53:28: which correlated well with the exposure of the molecule in the plasma.
53:28 - 53:40: So this preclinical data package was compelling for us and led us to the development of FHD286,
53:40 - 53:44: which we have nominated as our clinical candidate and have taken now into phase one studies.
53:45 - 53:49: And so this is a highly potent, I won't get into all the details of this table here,
53:49 - 53:54: but it's a highly potent selective oral first-in-class inhibitor of the SMARCA4 and SMARCA2
53:54 - 54:02: ATPases. It has nice in vivo properties, good oral bioavailability, which allows for daily
54:02 - 54:09: oral dosing, and nice clean safety and selectivity panels where it has been assessed.
54:11 - 54:18: And so just to show this molecule in the same uveal melanoma models, right, it performs very
54:18 - 54:22: similarly to what we saw with our two molecules where we have dose-dependent tumor growth
54:22 - 54:30: inhibition and also can achieve regressions in this 92.1 model. And importantly, this all occurs
54:30 - 54:36: at doses that are very well tolerated by the animals. So as I mentioned, we've taken this
54:36 - 54:45: molecule into phase one studies in metastatic uveal melanoma and also in AML. And so uveal melanoma
54:45 - 54:51: is a rare cancer of the eye. And although it's rare, it's actually the most common intraocular
54:51 - 54:59: malignancy. It's generally well controlled in the early stages through radiation or enucleation.
54:59 - 55:04: However, unfortunately, more than half of these patients ultimately develop metastases.
55:05 - 55:10: And by that point, it's quite a dismal prognosis for these patients with survival of only 4 to
55:10 - 55:17: 15 months. These patients have pretty limited treatment options, and clinical trials have
55:17 - 55:22: been the preferred option for the metastatic disease. Although very recently, this molecule
55:22 - 55:28: tebentafusp has recently been approved for a subset of these patients. And unlike cutaneous
55:28 - 55:33: melanomas, which have a very high mutational burden, uveal melanomas do not. And most of
55:33 - 55:38: these cases are driven by activating point mutations in these G-alpha proteins, which
55:38 - 55:47: signal through protein kinase C and MAP kinase pathways. All right. So mechanistically, we think
55:47 - 55:53: that, you know, BAF complexes are typically keeping chromatin in an open environment at
55:53 - 55:58: enhancers where this allows for these master transcription factors to access their binding
55:58 - 56:04: sites and drive lineage or disease-specific gene expression programs. However, when we inhibit the
56:04 - 56:10: ATPase components of the BAF complexes, this results in closing of chromatin and eviction of
56:10 - 56:15: these transcription factors, resulting in loss of their transcriptional programs and ultimately an
56:15 - 56:20: impact on cell survival. And we think this mechanism could serve, could function pretty
56:20 - 56:27: broadly across different tumor types, which have a very strong lineage transcription program,
56:28 - 56:36: including AML, right, where we see SPI1 transcription factor motifs are closed upon
56:36 - 56:43: BAF ATPase inhibition. So switching gears to AML, right, SPI1 seems to be an important
56:44 - 56:52: transcription factor in this disease. And in the dependency map database, right, many AML
56:52 - 56:57: cell lines have high expression of SPI1 and are highly dependent on SPI1 activity for survival.
56:58 - 57:05: And we can observe a nice growth inhibition of a variety of AML cell line models with our
57:05 - 57:17: tool compounds. Like what we've observed in uveal melanoma, we think that the activity in AML is
57:17 - 57:24: also driven, at least in part, through an apoptotic mechanism, where at least in MV411 here
57:24 - 57:29: and also in other cell types, we can see a dose-dependent induction in the sub-G1 population
57:30 - 57:36: and also, you know, evidence of apoptosis by annexin V and PI double staining.
57:39 - 57:45: Mechanistically, again, we think this is quite similar, where inhibition of these BAF complexes
57:45 - 57:50: results in closing of chromatin at a subset of sites. These sites are highly enriched,
57:50 - 57:55: again, for SPI1 binding motifs. And if we take this one step further, when we look by ChIP-seq,
57:55 - 58:02: we can also see that SPI1 occupancy is reduced dramatically at these sites.
58:08 - 58:15: If we treat a variety of AML xenograft models with, this is actually with our clinical compound,
58:16 - 58:22: FHD286, we can see a nice dose-dependent tumor growth inhibition across a variety of models.
58:23 - 58:30: And this seems to be mutation agnostic, as far as we can tell so far, and occurs, you know,
58:30 - 58:35: regardless of the mutation status of the model. And again, this is all happening at doses that
58:36 - 58:42: are well-tolerated by these animals. We've also done this in a disseminated model of AML,
58:42 - 58:50: MV411, where you can see, you know, severe tumor burden in the control animals by the end of the
58:50 - 58:57: study, whereas when these animals are treated with FHD286, either in the end of the study or
58:57 - 59:04: intermittent schedule, we can see a nice reduction of tumor burden and an impact,
59:04 - 59:13: a survival advantage for these animals. So, as has been mentioned in some of the earlier talks,
59:14 - 59:19: AML is a heterogeneous disease, and it is very hard to treat. And so, combinations are going to
59:19 - 59:24: be very important for any agent, really. And so, we've started to investigate some potential
59:24 - 59:31: combination partners for FHD286, here showing that in the Notable Labs platform, looking at
59:31 - 59:41: primary patient samples, there is nice activity on growth of these primary cells in combination
59:41 - 59:48: with cytarabine. And we've also tested this in an in vivo model using the OCI AML2 model,
59:48 - 59:55: where you can see an enhanced tumor growth delay in the combination of FHD286 with cytarabine,
59:55 - 59:59: compared to either single agent alone.
60:00 - 60:26: We think there is potential beyond cytarabine as well, and we've been collaborating with Dr. Kapil Bala at MD Anderson, and he's generated some data showing that we can at least have, in an ex vivo model of a primary patient sample with an inversion 3 monosomy 7, we can see nice synergy with venetoclax in this case.
60:27 - 60:30: And so we're actively investigating additional combinations as well.
60:35 - 60:46: We also use this Notable Labs platform, again, to try to tease out whether different, you know, to tease out how broad the response is in primary patient samples.
60:47 - 60:59: And we do see a pretty deep response across the majority of these patient samples. And this is irrespective, again, of the disease status, prior treatment, as well as mutation status of these samples.
61:00 - 61:06: One thing we did notice, however, in a subset of samples, was this, what appears to be a pro-differentiation effect.
61:07 - 61:19: So this is something we're really interested in understanding further, and we're currently collaborating with Stéphane de Botton at the Institut Gustave Roussy to further understand the impact of FHD286 on blast differentiation.
61:20 - 61:39: And I talked a bit about SPI-1's role in AML cooperating with BAF complexes, but we actually think this may go beyond SPI-1 as well, as different transcription factors are implicated in different stages of hematopoietic development.
61:40 - 61:52: And we think that perhaps, depending on the stage at which the cancer has developed, there could be additional transcription factors that may cooperate with BAF complexes to drive these oncogenic transcriptional programs.
61:56 - 62:04: And then finally, you know, just to let you know, beyond uveal melanoma and AML, we think there could also be additional opportunities for FHD286.
62:05 - 62:09: So I showed you those two indications were highly sensitive at early time points.
62:10 - 62:17: However, if we treat cells for longer periods of time, we can see growth inhibition in many different cancer cell types.
62:18 - 62:21: So we do think there could be additional opportunities for this drug.
62:24 - 62:32: All right, so in conclusion, I've shown you that we've identified a novel series of potent and selective inhibitors of the BAF ATPases.
62:35 - 62:44: BAF ATPase inhibition affects the proliferation of many different cancer cell types, but has rapid and strong effects on uveal melanoma and AML cell lines.
62:46 - 62:57: And ultimately, this affects the lineage-specific enhancer accessibility, causes reduced occupancy of these master transcription factors, and impacts their gene expression programs and survival.
62:58 - 63:05: We've developed FHD286 as a first-in-class selective oral allosteric inhibitor of these BAF ATPases.
63:06 - 63:17: And this compound has nice efficacy in multiple xenograft and also PDX models, which I didn't have a chance to show you today, in both uveal melanoma and AML at well-tolerated doses.
63:19 - 63:25: Phase I studies in AML and metastatic uveal melanoma were initiated last year, and these are currently underway.
63:26 - 63:28: And we do think there could be additional opportunities for this drug.
63:30 - 63:34: So, yeah, thank you very much for your attention. I'd be happy to take any questions.
63:34 - 63:54: Great, thank you very much, Richard. I'll just remind you to pop your questions in the Q&A. Wonderful talk, really very interesting, and interesting to see some of the differences and the similarities with the two diseases that you talked about.
63:55 - 64:17: So, we have a question from Sengey Kelechi. Forgive me if I got your name a little bit muddled there. I was wondering whether you compared FHD286 with ABCI1, which is a well-known PROTAC for SMARCA2 or SMARCA4. So I suppose a question about catalytic inhibition versus degradation.
64:18 - 64:37: Yeah, great question. So we have not directly compared the two molecules. However, there was a recent publication from Arul Shainam's lab using these dual degrader molecules of SMARCA2, SMARCA4 ATPases in prostate cancer.
64:38 - 64:53: And they see very similar things in terms of closing of chromatin at enhancers and loss of lineage-specific transcriptional programs affecting androgen receptor and FOXA1 in that system.
64:53 - 65:06: So I think the two mechanisms are very comparable. And I don't, although we haven't tested it directly, I don't think there's really any difference between enzymatic inhibition and protein degradation.
65:07 - 65:19: Okay. And just in terms of, oh, another question come in. Constantine Mylanos, have you looked at motifs over regions that display increased chromatin accessibility upon BAF inhibition?
65:20 - 65:36: Yeah, that's not something that we have really spent much effort looking into. As you saw, most of the sites affected are closed and very few sites are actually opening. So I'm sorry, we haven't really looked into it.
65:36 - 65:55: Okay. In terms of many of the, I mean, across the number of diseases that you talked about, for the motifs that you showed, I noticed that a number of them were either well-known or reported to have pioneering activity.
65:56 - 66:09: Do you think this works particularly well with pioneering transcription factors, or do you think it would work with, you know, standard regulators that bind predominantly at open regions already?
66:10 - 66:32: Yeah, that's an interesting question. So, you know, I guess it sort of makes you wonder what is a pioneering transcription factor. And although these have been known to open chromatin on their own, it's possible that in a more complex cell environment, the reason these are pioneering factors is because they're cooperating with chromatin remodelers.
66:33 - 66:38: So I do think, you know, there probably is some specificity there.
66:38 - 66:49: I would agree. I don't think pioneers are special in any way other than they can bind to compacted chromatin. I think that the mechanisms of keeping it open or opening it are probably very similar.
66:50 - 67:03: Okay, so a question from Eric Conway. Why do you think uveal melanoma and AML are so sensitive to your compounds? Do they share mutations that specifically sensitize cells to BAF inhibition?
67:04 - 67:09: Yeah, that's a great question and something that we're trying to understand.
67:09 - 67:19: As far as we can tell, there are not any specific mutations that are predictive of sensitivity in those particular indications.
67:19 - 67:24: Uveal melanoma does have fairly high expression of SMARCF4.
67:24 - 67:29: So that is one thing that we're looking into.
67:30 - 67:42: It may also come down to the actual phenotypic response that we're eliciting. So I've shown you that we think that these two indications do respond through an apoptotic mechanism.
67:42 - 67:55: And depending on, you know, what's actually happening in other disease types, it could be different and could just result in a slower effect. But that's not something we've really figured out just yet.
67:55 - 68:11: So just to kind of expand on that a little bit, I mean, one of the things I, you know, I've read lots of reviews from Gerald and from Sigal about how common the mutations in chromatin-dependent remodelers are.
68:11 - 68:30: In hematological malignancies, they're actually relatively rare and in AML, they're not very common at all. So do you think it's, do you have any evidence to suggest that they're more active in tumors that have no mutations of ATP-dependent chromatin remodelers?
68:30 - 68:47: And the second thing, from my limited knowledge of the comparisons between the genetic mutations, my memory is that BAP1 mutations are quite common in uveal melanomas, and also splicing mutations.
68:47 - 69:06: So, you know, splicing mutations are uncommon in AML, but other things that affect that sort of ASXL1 to, you know, PRC2 axis are. Is there any suggestion that these similarities are perhaps an explanation?
69:06 - 69:10: Yeah, that's, that's a really intriguing thought.
69:11 - 69:30: It's definitely possible. What I can say is, in uveal melanoma we have compared, you know, BAP1 models, BAP1 positive and BAP1 negative models. Same for SF3B1, and we don't really see any difference, as far as we can tell in the response.
69:30 - 69:35: So unfortunately it's probably not that simple.
69:35 - 69:52: Nothing ever is, is it? Okay, a question from Paul Mueller. Are you observing resistance in your in vivo and clinical studies similar to your in vitro model where you find the I1173M resistance mutation associated change?
69:52 - 70:07: Yeah, so we have not really dosed our animal models to the levels and for the periods of time, I guess, where you would expect to start to see resistance observed.
70:07 - 70:12: But it's a great question and something that would be very interesting to do.
70:13 - 70:31: Any more questions? I wonder if I could ask one which is, you know, quite a sort of just theoretical epigenetics. I was very interested to see that the changes were really very, very specific to distal regulatory elements rather than promoters.
70:32 - 70:47: I mean, do you think that the mechanisms for the maintenance, I mean, the nice papers from Dirk Schübeler and Stefan Kubiak last year about the constant requirement for activity of ATP-dependent chromatin remodelers to keep things open.
70:48 - 71:07: But my memory was that it wasn't, that wasn't quite as enhancer-specific. Do you think in these cancer cell line models or in cancer in general, there are different mechanisms keeping distal enhancer elements open from proximal promoter elements?
71:07 - 71:21: Yeah, so I think if I remember the Kubiak paper that you mentioned, they actually did comment on the importance of chromatin remodeling complexes at super enhancers.
71:21 - 71:39: I think, yeah, there is something here about really trying to drive those master transcriptional programs through these super enhancers. And that seems to be where at least the canonical BAF complex seems to be really important.
71:39 - 71:52: I suppose that might contribute to the selectivity and the therapeutic window as well.
71:52 - 71:54: Yeah.
71:55 - 71:58: Sorry, not sure exactly what you mean there.
71:58 - 72:08: Just that, you know, cancer cells seem to be much more addicted to these super enhancers in terms of, you know, by comparisons to their normal tissue equivalents.
72:08 - 72:10: Exactly. Yep.
72:10 - 72:29: Okay. So any further specific questions for Richard? Or should we move on now to the more panel discussion? So, Sydney, if you could come back, please, that would be great.
72:30 - 72:43: I was sort of thinking about, you know, ways to sort of tie these things together and it's a shame that Carsten couldn't join us, but obviously, I think that we're targeting the epigenetic machinery, which is really cool.
72:44 - 72:56: But we've both been, both the talks have been talking about the importance of the transcription factor networks within maintaining these sort of gene programs.
72:57 - 73:12: Do you think there's a way that we can try and target specific transcription factor chromatin regulator interactions or nodes of these transcriptional programs?
73:12 - 73:24: I'm trying to think about ways that we might cut down on more global toxicity and try and make these a little bit more specific to the abnormalities that we see in cancer.
73:24 - 73:31: That's a broad question, and I suppose we'll give it to the older of the two of you. So I'll let you start, Richard.
73:31 - 73:47: Okay, sure. Yeah, so actually it's a really interesting idea. And this is something at Foghorn that we are actually trying to tackle, specifically through these BAF interacting transcription factors.
73:47 - 73:53: You know, potentially as a way to have a more specific effect on particular lineages.
73:53 - 74:01: And so we have a whole platform trying to develop inhibitors of these BAF transcription factor interactions.
74:01 - 74:19: It's interesting, in an unpublished study that we have ongoing where we're looking at CRISPR screening of chromatin factors, of which there are a number of ATP-dependent chromatin remodelers, and this is predominantly in normal hematopoiesis, but also some in malignant.
74:20 - 74:43: We see very different dependencies for BAF complex members at different stages of hematopoiesis. And I thought that your idea of looking maybe at the differentiation level of the AML might have quite a bit of resonance, given what we're finding.
74:43 - 74:45: Yeah, yeah.
74:45 - 75:03: Sydney, in terms of, you know, the EZH2 inhibition that you were talking about, do you think that there will be specific sort of transcriptional networks that are governed by particular transcription factors that might either help to improve efficacy or the other thing,
75:03 - 75:14: the difficulty that we talked about in selecting the right patients for, you know, for the inhibitor? What are your thoughts on that?
75:14 - 75:31: We haven't looked at that specifically, but I would think, yes, that we can, and I think with epigenetic modifiers, it's important to look at what's happening when you inhibit both an abnormal hematopoietic stem cell as well as an AML blast and kind of compare
75:31 - 75:48: with what is changing, and so you can identify those pathways. So that way, when you, with the idea of, we focus on combination therapies, but in other inhibitors as well, that you can specifically target the downstream mechanisms that are specific to a blast
75:48 - 76:05: and hopes to not influence the hematopoietic stem cell as much, but I think that is a massive problem with epigenetic inhibitors is the effects on normal HSCs, and I know that's a thing with chemo as well, but it's definitely something to be cautious of.
76:05 - 76:31: Yeah, I mean, I think it's also a reflection of what we know about, we know a lot about individual transcription factors, we know a lot about certain chromatin factors and chromatin regulators, but these things work together in gene regulatory complexes, and these differ throughout normal hematopoietic differentiation and are aberrant in AML and other malignancies.
76:31 - 76:50: So, you know, I personally think that we need to actually put together these in complexes and look at the function of complexes as much as individual regulators. I think that kind of granularity will give us a lot better insight into where to deploy them and what might be potential toxicities.
76:50 - 77:06: So let's talk a little bit about toxicities. We've all acknowledged the fact that we ain't going to win against AML with monotherapies, but it's brilliant to have multiple agents that might work, potentially in combination.
77:06 - 77:30: So, if I can ask each of you in turn, where would you see your inhibitor working? And by that, you know, would it be up front in high disease bulk sick patients with some form of significant cytoreduction? You both talked about educating these cells.
77:30 - 77:58: You know, it sounds like perhaps yours, Richard, is a little bit more cytotoxic, but obviously, it doesn't look like the EZH2 is, but both of them are probably having a differentiation effect. Do you think that they would be best deployed up front, or could they be safely deployed up front, or should they be phased in a more sort of consolidation, maintenance phase? I'll let you go first, Sydney.
78:00 - 78:16: I think definitely for the EZH2 inhibitor, more of an upfront setting, more of a blast, having more, I think, blasts available, especially because we know EZH2 inhibition prior to AML development is probably not where we want to be, so we want to make sure there's lots of blasts there.
78:16 - 78:35: And then the idea would be combination therapy, so what I'm envisioning in a perfect world would be a combination therapy, and it can be given for a short period of time, and we can get a strong reduction in the disease burden, and then we can give consolidation therapy based on remaining cells.
78:36 - 78:46: If there's, I don't know, for example, an EZH2 mutation, that residual giving an acid inhibitor or something like that is kind of how I'm envisioning this working.
78:46 - 78:49: And Richard?
78:50 - 79:05: It's an important question. I think we're really trying to figure out how best to use our molecule currently, given that we can see both sets of toxicity and signs of differentiation in our preclinical models.
79:05 - 79:21: Yeah, I think it probably, there's a chance you can use it in both settings, right, to treat up front or, you know, to sort of prime those samples towards some particular phenotype or combination.
79:22 - 79:49: Another thing that I think about a lot is, you know, mechanisms of resistance caused by other targeted therapies, and how this can sometimes happen through a transcriptional rewiring mechanism, and thinking about, you know, trying to target or prevent resistance by actually inhibiting important transcriptional regulators.
79:49 - 79:58: I think that that's the mantra, isn't it? As soon as you get something that looks good, you've got to already think about mechanisms of resistance.
79:58 - 80:00: And perhaps one final thing before I go.
80:00 - 80:06: And again, it's slightly linked to how we might deploy these and where would be the best place to use them.
80:06 - 80:14: Do you have a feeling, either of you, that your drug is, and the mechanism that underlies it,
80:14 - 80:21: is most prevalent in leukemic stem cells within the sort of progeny thereof, or do you think it's
80:21 - 80:32: a process throughout AML? Is it more selectively targeting one area of the abnormal hierarchy within
80:32 - 80:43: the disease? Who wants to go first? I can start. Yeah, so I think that's, again,
80:43 - 80:50: an important question that we're trying to tease apart right now, looking at different stages of
80:50 - 80:58: development to try to see if the BAF inhibitors have any differential effects at different stages
80:58 - 81:03: of development. I think, you know, one thing that you're sort of touching on here, and that you've
81:03 - 81:10: both sort of touched on already, is that context is really important for epigenetic, you know,
81:10 - 81:17: therapeutics, right? Sydney, you talked about the, you know, EZH2 as a tumor suppressor or an oncogene,
81:17 - 81:23: and, you know, same can be said for SMARCA4. So I think for all of these epigenetic
81:24 - 81:31: targets, like, just understanding the context could, you know, could have really differential
81:32 - 81:38: impacts on the biology that we're looking at. And probably on limiting toxicity as well.
81:39 - 81:47: Right. Sydney, I'll give you the last word. I would second that the context is super
81:47 - 81:54: important for whether these epigenetic inhibitors are going to be beneficial or
81:54 - 82:00: detrimental, and understanding that is important before going into people. But we have not looked
82:00 - 82:08: at if there is a difference in leukemic stem cells versus progenitor cells. I think our kind of, our
82:08 - 82:12: planned PDX with the pretreatment will really kind of help tease that out a little bit, because if we
82:12 - 82:17: can see evidence of differentiation prior to going to a mouse, but they still meet ERC at about
82:17 - 82:23: the same time, then we would expect it's not affecting the leukemic stem cells, but we haven't
82:23 - 82:29: looked for sure. But I think, again, just reiterating that these epigenetic modifiers
82:29 - 82:34: are very context-specific and change even in a person throughout their disease evolution,
82:35 - 82:41: I think is really important. Okay, well, poor chairmanship have let us run over, so I'm going
82:41 - 82:47: to bring the session to a close. But I'd like to thank all of our speakers, Carsten in his absence,
82:47 - 82:54: and Sydney and Richard for three really stimulating talks. And it's good to know that AML,
82:54 - 83:00: although, albeit still an unmet medical need now, has a number of different options to treat
83:00 - 83:04: patients with. So thank you for attending the session, and thank you to Abcam for putting
83:04 - 83:09: it on. I'll hand over to you, Vinit, if you wanted to say anything to close.
83:11 - 83:19: Thank you.

Epigenetic landscape of acute myeloid leukemia

On-demand webinar

Summary:

Moderator and speakers:

Video Transcript

You may be interested in...