Drugging the epitranscriptome as an effective anti-cancer strategy

Video Transcript

• 00:00 - 00:13: Hello, everyone. I would like to thank Abcam for giving me this great opportunity to present
• 00:13 - 00:19: this webinar. My name is Kostas Tzelepis, and today I’m going to talk to you about drugging
• 00:19 - 00:24: the epitranscriptome as an effective strategy against devastating cancers such as acute myeloid
• 00:24 - 00:32: leukemia. Our lab is based at the Milner Therapeutics Institute in Cambridge. The lab’s profile
• 00:32 - 00:37: is very translational, aiming to identify novel therapeutic targets, characterize newly
• 00:37 - 00:43: developed drugs, repurpose existing therapies, as well as identify drug resistance and synthetic
• 00:43 - 00:49: lethality mechanisms. Many of these, we are going to do it by collaborations, or we are
• 00:49 - 00:54: currently doing by collaborations with other labs and industrial partners. So for the
• 00:54 - 01:00: above reasons, we have developed several CRISPR-Cas9 platforms. We have generated sophisticated
• 01:00 - 01:08: cancer models and a plethora of other in vitro, ex vivo, and in vivo platforms. Initially, we
• 01:08 - 01:13: developed one of the first, if not the first, genome-wide CRISPR screening platform, truly
• 01:13 - 01:20: focused on one disease, acute myeloid leukemia. Using a panel of AML cell lines and two other
• 01:20 - 01:26: non-leukemic cancer cell lines, we cataloged approximately 500 leukemia essential genes,
• 01:26 - 01:32: and as you can see on the right plot, many of those were belonging to druggable targets.
• 01:32 - 01:37: We also have experience in choosing and following up promising targets from large screens. In
• 01:37 - 01:45: this slide, I present you an example of repurposing of SRPK1 that was discovered by one of our
• 01:45 - 01:51: recent screening efforts. In collaboration with Exonate and GSK, we showed that SRPK1
• 01:51 - 01:57: is not only a druggable target against retinal neovascular disease, but also against MLL
• 01:57 - 02:05: rearranged leukemias, an aggressive AML subgroup presented predominantly in pediatric cases.
• 02:05 - 02:10: The main mechanism was the differential usage of splicing isoforms of key epigenetic regulators,
• 02:10 - 02:16: including BRD4. We also went on to show that there is a strong synergistic effect after
• 02:16 - 02:23: co-treatment with BAT inhibitors. Finally, SRPK1 inhibitors are entering clinical trials
• 02:23 - 02:31: in 2021, and let’s see. We are currently focusing on the role of particular RNA modifications
• 02:31 - 02:37: in cancer. In general, while there are over 140 RNA modifications reported, it is fair
• 02:37 - 02:44: to say that this is a largely underexplored field with very few detected so far on mRNA
• 02:44 - 02:51: or non-coding RNA. However, as you can see from the plot of this slide, recent scientific
• 02:51 - 02:56: highlights have strongly connected cancer initiation and maintenance with a number of
• 02:56 - 03:02: RNA modifications through the function of the writers, readers, and erasers. This association
• 03:02 - 03:08: is also elegantly discussed by Barbieri and Kouzarides in a recently published review.
• 03:08 - 03:15: We are currently going to focus on m6A RNA modification, which is one of the most common
• 03:15 - 03:21: and abundant modifications in eukaryotes, perhaps the most abundant on mRNA. The modification
• 03:21 - 03:30: is catalyzed by a few RNA methyltransferases, predominantly METTL3, METTL16, and the recently
• 03:30 - 03:37: identified METTL5. The modification is specifically recognized by particular m6A readers, mainly
• 03:37 - 03:44: belonging to the YTH domain-containing family, and is erased by ALKBH5 and FTO demethylases,
• 03:44 - 03:47: all of them connected to cancer one way or another.
• 03:47 - 03:54: Pertinent to the m6A RNA modification, we recently published a study whereby using ex vivo
• 03:54 - 04:00: genome-wide and focused CRISPR screens, we identified several RNA-modifying enzymes important
• 04:00 - 04:07: for AML growth, and we went on to characterize METTL3 as a promising anti-AML target.
• 04:07 - 04:13: In principle, we revealed a novel leukemogenic mechanism where METTL3 is recruited on the
• 04:13 - 04:19: promoters of a specific gene subset via the transcription factor CEBPZ. The mRNA originated
• 04:19 - 04:26: from the METTL3-bound genes are m6A modified within their coding sequence, which is pivotal
• 04:26 - 04:32: for their mRNA translation, and loss of this m6A modification within these coding regions
• 04:32 - 04:39: prevents their translation by enhancing ribosomal stalling. Independently, an elegant study
• 04:39 - 04:47: by Vu et al. showed similar effects on the mRNA translation upon loss of METTL3 in AML.
• 04:47 - 04:54: However, all the published data are based on genetic knockdown or knockout of METTL3.
• 04:54 - 04:59: We don’t know the true molecular and cellular impact of blocking the enzymatic activity
• 04:59 - 05:03: of METTL3. There is a lack of agents that specifically block the catalytic activity
• 05:03 - 05:10: of this enzyme without affecting the protein levels, and therefore, the therapeutic potential
• 05:10 - 05:17: of METTL3 cannot be truly investigated. Towards that way, and in a great collaboration
• 05:17 - 05:22: with Storm Therapeutics in Cambridge, UK, we have been able to develop and characterize
• 05:22 - 05:28: the first bioavailable small molecule against METTL3. In this study, I’m showing you details
• 05:28 - 05:34: of two different compounds in Molm-13 cell line, compound 1 and compound 2, but all the
• 05:34 - 05:41: data presented today are related to compound 1. Compound 1, so excellent drug-like properties
• 05:41 - 05:48: with very good METTL3 activity in a biochemical assay on the left side of this slide, and
• 05:48 - 05:54: a good anti-proliferative effect in vitro. We also, as is the target engagement, where
• 05:54 - 05:59: we observed significant reduction of m6A in total and poly(A) RNA, as well as reduction
• 05:59 - 06:06: of the known METTL3 bio-marker SP1, as you can see on the right part of this slide.
• 06:06 - 06:12: Furthermore, compound 1 was well tolerated in vivo with excellent pharmacokinetic and
• 06:12 - 06:17: selectivity profile. Regarding the selectivity, as you can see from the bottom right part
• 06:17 - 06:22: of the slide, our compound was a thousand times more selective for METTL3 compared
• 06:22 - 06:29: to another 48 DNA, RNA, protein, and metabolic-related methyltransferases. We also got an excellent
• 06:29 - 06:35: profile using Eurofins Safety 47 panel against a wide range of kinases, ion channels, etc.
• 06:35 - 06:41: that we are not sharing with you at the moment. Treatment of a panel of human AML cell lines
• 06:41 - 06:48: and primary murine AMLs, here I’m just illustrating a few of them, showed an enhanced myeloid
• 06:48 - 06:55: differentiation followed by elevation of apoptosis. Interestingly, as you can see on the right
• 06:55 - 07:00: part of the slide, the non-leukemic hematopoietic stem and progenitor cell model that we use
• 07:00 - 07:07: harboring only the FLT3-ITD mutation didn’t show any differentiation or apoptotic phenotype
• 07:07 - 07:14: indicating AML specificity. We also looked at the m6A levels of known METTL3 substrates
• 07:14 - 07:23: using m6A RNA immunoprecipitation and qPCR. Here I’m only illustrating SP1 and BRD4.
• 07:23 - 07:27: You can see from the slide, from the upper panel of this slide, pharmacological inhibition
• 07:27 - 07:35: of METTL3 decreased the m6A signal in 24 and 48 hours, but notably two known m6A sites
• 07:35 - 07:41: that are not METTL3 dependent, so absolutely no difference upon treatment further validating
• 07:41 - 07:45: the specificity and the selectivity of our compounds.
• 07:45 - 07:51: We next wanted to examine if the isolated inhibition of the METTL3 catalytic activity
• 07:51 - 07:59: using our compound could have an impact on mRNA translation, similar to what we and another
• 07:59 - 08:04: group from New York have recently shown using genetic inhibition of METTL3. We therefore
• 08:04 - 08:10: went on and performed polysome profiling using Molm-13 cells treated with either vehicle or
• 08:10 - 08:15: METTL3 compound for 48 hours. Looking at the left part of the slide, you will see that
• 08:15 - 08:21: pharmacological inhibition of METTL3 in red led to significant reduction of the high molecular
• 08:21 - 08:29: weight polysome fraction, suggesting a strong effect on mRNA translation. We further confirmed
• 08:29 - 08:36: that by performing validation of several METTL3 targets, here I’m just illustrating SP1, we
• 08:36 - 08:44: observed that the dramatic reduction of SP1 in the high molecular weight polysomes, and an
• 08:44 - 08:51: increase in the low molecular weight polysomes strongly suggested a particular specific mRNA
• 08:51 - 08:56: translational defect. At the same time, as you can see on the further right part of this slide,
• 08:56 - 09:02: but the overall transcriptional level of SP1 didn’t change, further mirroring the genetic
• 09:02 - 09:10: results with METTL3 knockout or knockdown. As a next step, we then went on to perform an
• 09:10 - 09:16: initial in vivo characterization of the METTL3 compound using two cohorts of mice, one vehicle,
• 09:16 - 09:22: one treated, both transplanted with primary murine AML cells harboring an MLL-AF9 fusion,
• 09:22 - 09:29: a FLT3-ITD mutation, as well as a YFP reporter. As this is usually a very aggressive
• 09:29 - 09:37: model, we confirmed that establishment of AML on day 17 via elevated white blood count,
• 09:37 - 09:42: and then we started a daily treatment of 30 milligrams per kilogram, intraperitoneally, for
• 09:42 - 09:47: five days in total. We then harvested the spleen and the bone marrow of those mice,
• 09:47 - 09:55: and we went to characterize them with downstream experiments. As you can see here, observing on
• 09:55 - 10:01: the left side of this slide, you could see that the treated cohort had a significant reduction
• 10:01 - 10:08: of spleen size, illustrating strong anti-leukemic effect in vivo. Also, on the right top part of the
• 10:08 - 10:14: slide, we present analysis of the bone marrow of both cohorts, where we observed significant
• 10:14 - 10:20: representation of the AML cells in the treated cohort. Finally, we confirmed target engagement
• 10:20 - 10:26: in vivo by performing Western blot for the METTL3 biomarker SP1, where there was a protein
• 10:26 - 10:34: loss only at the treated cohort, as you can see at the bottom right part of the slide.
• 10:35 - 10:40: Finally, being very confident about our results, we performed PDX studies using a daily treatment
• 10:40 - 10:47: of 50 milligrams per kilogram intraperitoneally for two weeks in total. Here I show one example
• 10:47 - 10:54: using an NPM1 mutant PDX model, where we achieved considerable reduction of AML expansion in vivo,
• 10:54 - 10:59: as well as significant prolongation of the lifespan of the treated cohort, as you can see
• 10:59 - 11:06: on the right part of the slide. We also further confirmed reduction of the patient-derived cells
• 11:06 - 11:12: in the bone marrow using flow cytometric analysis for human CD45. Importantly, we confirmed target
• 11:12 - 11:19: engagement with SP1 levels appearing significantly and selectively lower in the treated cohort
• 11:19 - 11:27: compared to the vehicle. To summarize, today, I’ve shown you the first study demonstrating in vivo
• 11:27 - 11:33: activity of inhibitors of an RNA methyltransferase. This is the ultimate proof of concept that RNA
• 11:33 - 11:39: modifying enzyme represents a new promising target class for cancer therapeutics. METTL3 inhibitors
• 11:39 - 11:45: have excellent drug-like properties and consistent activity in mechanistic and functional assays,
• 11:45 - 11:49: but most importantly, pharmacological inhibition of METTL3 is efficacious
• 11:49 - 11:56: in primary murine AML and human PDX models. And at this moment, I would like to thank very much
• 11:56 - 12:03: Dr. Ellie Yankova, Dr. Etienne de Braekeleer, Demetrios Aspris, and Professor Tony Kouzarides
• 12:03 - 12:09: for this lovely collaboration, especially Professor Tony Kouzarides for the great mentorship
• 12:09 - 12:14: and co-supervision of this study. I would like to thank Milner Therapeutics Institute and the fantastic team,
• 12:15 - 12:22: Storm Therapeutics for this lovely collaboration, and in particular Wesley, Oliver, Mark, Richard,
• 12:22 - 12:28: Alan, and Byron, Professor George Vasilliou and his lab, in particular Justyna and Gonia,
• 12:29 - 12:34: Irmela Jeremia’s lab in Munich, Germany, as well as my collaborators in Cambridge, Brian,
• 12:34 - 12:40: and Nerea. I would also like to thank my funders at Wellcome Trust, and finally,
• 12:40 - 12:45: I would like to thank very much you for your attention as well as for joining this webinar.
• 12:45 - 12:47: Have a good day and take care.