Slowing of Transcription and Epigenetic Rewiring

Transcript

• 00:00We'll get started. Welcome, everybody.
• 00:03I'm Manuj Pillai. I'm one
• 00:04of the faculty at the
• 00:05Yale Cancer Center, and it
• 00:06is my privilege and honor
• 00:08to introduce today's speaker, doctor
• 00:10Prajwal Bodu, whom I have
• 00:11mentored for the past four
• 00:13years.
• 00:14Many of you know Prajwal
• 00:16as a trainee here. He
• 00:17finished his hematology and oncology
• 00:19fellowship in two thousand twenty
• 00:20two,
• 00:21and, he did his medical
• 00:24education in at the Osmania
• 00:25Medical School in Hyderabad, India,
• 00:28one of the premier schools
• 00:29of of the country, and
• 00:30graduated with, multiple honors,
• 00:33followed by a residency in
• 00:35internal medicine in Chicago,
• 00:37following which he did a
• 00:38two year clinical fellowship
• 00:40in, leukemia at the MD
• 00:42Anderson Cancer Center,
• 00:44where he was, phenomenally productive,
• 00:46with over forty manuscripts, most
• 00:48of them as first author,
• 00:50and he focused on clinical
• 00:51and outcomes research in myeloid
• 00:52malignancies at that time. In
• 00:54twenty eighteen, he matched to
• 00:56Yale's hematology oncology program, and
• 00:59he had decided by then
• 01:00that he would pursue basic
• 01:01research as part of his
• 01:02fellowship training.
• 01:03And I was very lucky
• 01:04that he chose to join
• 01:05our group.
• 01:07His, the main question that
• 01:09he wanted to answer
• 01:11was a large expansive difficult
• 01:12one that has perplexed
• 01:14a lot of us in
• 01:15the RNA
• 01:16field, which is, what are
• 01:18the biological mechanisms that render
• 01:20these recently described mutations in
• 01:23RNA splicing factors
• 01:24to be oncogenic.
• 01:26And, as he will,
• 01:28go over go over in
• 01:30his talk,
• 01:31we should be looking at
• 01:32RNA processing as a whole
• 01:34and not just limit ourselves
• 01:36to RNA splicing.
• 01:38And,
• 01:39his recent paper in molecular
• 01:40cell is a two d
• 01:41force in multiple techniques that
• 01:43he has learned and he
• 01:44himself has developed over the
• 01:46last four years.
• 01:47And, some of it is
• 01:49what he's going to present
• 01:50as,
• 01:51published data and some of
• 01:52it is unpublished ongoing work.
• 01:56So,
• 01:57as this mentor, I can
• 01:59list many superlatives,
• 02:00for Pajal as a trainee,
• 02:02but I will just focus
• 02:03on one, which is this
• 02:04absolute fearlessness in taking on
• 02:06big complex scientific questions
• 02:08and learning the relevant techniques,
• 02:10to complete those projects along
• 02:12the way.
• 02:14Most impressively, his work has
• 02:15been entirely funded by, experimental
• 02:18funding,
• 02:18throughout his, time. He started
• 02:20on his training grant that
• 02:22is led by Roy and
• 02:23Ali Ping Chen, and then
• 02:24he
• 02:25got a young investigator award
• 02:26from the Yvonne's MDS Foundation.
• 02:28And most recently, he's just
• 02:30been funded by, NIH k
• 02:31o eight, which started in
• 02:33July in its first submission
• 02:34with an extraordinary score and
• 02:37which I believe is the
• 02:37first k o eight in
• 02:38the section for over a
• 02:39decade. So,
• 02:42Prajul is currently enrolled in
• 02:43the investigative medicine program, and
• 02:45he's expected to graduate in
• 02:47the coming summer.
• 02:48And, I'm confident that as
• 02:50he transitions from mentee to
• 02:51independent investigator and a collaborator,
• 02:54we can expect more paradigm
• 02:55defining stuff from.
• 02:57So I'm happy to welcome
• 02:59him to the podium. Thank
• 03:00you.
• 03:17Good afternoon, everybody. Thank you
• 03:19all for coming, and, thank
• 03:20you for the very kind
• 03:21introduction, doctor Pillai, and for
• 03:22nominating me to present today.
• 03:24Thank you to the Yale
• 03:25Cancer Center for giving me
• 03:26this wonderful opportunity.
• 03:27As some of you know,
• 03:28I've been a fellow here,
• 03:29and, I'm now a staff
• 03:30physician in the department of
• 03:31hematology
• 03:32and, also an associate research
• 03:34scientist with doctor Pillai. The
• 03:36the topic I'll be presenting
• 03:37today,
• 03:40is ongoing work and work
• 03:41that I have done over
• 03:42the past four years, some
• 03:44of which is published and
• 03:45some which is unpublished.
• 03:46Before I start, I do
• 03:47not have any, any disclosures.
• 03:50I decided to split my
• 03:52presentation into the following topics.
• 03:53I'll start off with a
• 03:54brief introduction on myelodysplastic syndromes
• 03:56and then go go into
• 03:57disease models that we have
• 03:58generated to study splicing factor
• 04:00mutations and how we have
• 04:01used these models to study
• 04:03RNA polymer transcription RNA polymerase
• 04:05transcriptions.
• 04:06And I'll finally end with
• 04:08how we think or envision
• 04:09our findings to be of
• 04:10therapeutic elements.
• 04:13So brief refresher on myelodysplastic
• 04:15syndromes.
• 04:16MDS can be best characterized
• 04:17as a triad of recurrent
• 04:19mutations and genomic instability.
• 04:21These mutations occur in certain
• 04:22progenitor cells resulting in their
• 04:24inappropriate clonal expansion.
• 04:26And because this clonal expansion
• 04:28is inappropriate, the cells do
• 04:29not mature and proliferate properly,
• 04:31and this culminates in ineffective
• 04:33hematopoiesis and bone marrow failure.
• 04:35So we have seen a
• 04:36lot of therapeutic progress made,
• 04:37especially in the lower risk
• 04:38MDS category where we have
• 04:40seen the
• 04:41the development of drugs such
• 04:42as luspatercept
• 04:43and,
• 04:44Imetelstat.
• 04:45However, in higher risk and
• 04:46intermediate risk MDS, there is
• 04:48still a a lot of
• 04:49scope for therapeutic development.
• 04:51The the outcomes
• 04:52after hypomethylating agent failures remains
• 04:54dismal, and the only chance
• 04:56for cure in these patients
• 04:57is transferred.
• 05:00Now sampling of large patient,
• 05:02cohort,
• 05:03datasets have shown that MDH
• 05:04falls in the spectrum of
• 05:05clonal myeloid disorders.
• 05:07On one end of the
• 05:08spectrum is what we call
• 05:09clonal hematopoiesis of indeterminate potential.
• 05:11So So these are patients
• 05:12with normal peripheral blood counts,
• 05:14but they have
• 05:15had a sampling of their
• 05:16bone marrow for some of
• 05:17the other reason. And,
• 05:19it is found that some
• 05:20of the clones harbor mutations
• 05:21in, MDS associated genes.
• 05:24On the other end of
• 05:25the spectrum
• 05:27is acute myeloid leukemia, which
• 05:29is a fatal and potentially
• 05:30life threatening disorder that requires
• 05:31urgent medical attention.
• 05:33MDS falls in the middle
• 05:34of the spectrum. A typical
• 05:36case of MDS is a
• 05:37patient who has low peripheral
• 05:38bed counts in one or
• 05:39more of the cell lineages,
• 05:40either in the white blood
• 05:41cells, platelets, or red blood
• 05:43cells.
• 05:45You can see my pointer.
• 05:46Okay. White blood cells, red
• 05:48blood cells, or platelets.
• 05:49And bone marrow profiling shows
• 05:51that the morphology of these
• 05:52cells are abnormal. You see
• 05:54dyspartic,
• 05:55abnormal looking stem cells, and
• 05:57these cells have mutations in
• 05:58MDS associated genes.
• 06:01Coming to the mutation spectrum
• 06:02in MDS,
• 06:03the majority of cases of
• 06:05MDS have mutations in a
• 06:06group of genes called splicing
• 06:07factors, up to fifty percent
• 06:08of MDS cases.
• 06:10Another forty five percent of
• 06:11patients have mutations in epigenetic
• 06:13regulated genes. And then there's
• 06:15another twenty five percent of
• 06:16patients who have an overlap,
• 06:17a commutation occurrence of a
• 06:19splicing factor gene along with
• 06:20an epigenetic
• 06:21regulated gene.
• 06:23Now MDS shares, a lot
• 06:25of genetic abnormalities with ChIP
• 06:27as well as AML. You
• 06:29can see with ChIP that
• 06:30the majority of mutations are
• 06:31those involving epigenetic regulators, which
• 06:33are shown here in blue,
• 06:34counting over fifty percent of
• 06:36cases. The splicing factor mutations
• 06:38account for a lot less
• 06:39in the order of five
• 06:39to ten percent. And similar
• 06:41is the case with AML
• 06:42where you see a high
• 06:43proportion of cases with mutations
• 06:45in epigenetic regulators and a
• 06:47smaller proportion involving splicing factor
• 06:49genes.
• 06:50MDS stands out in this
• 06:51regard. You can see that
• 06:52compared to the other two,
• 06:53there is a very high,
• 06:55occurrence of splicing factor genes,
• 06:57in MDS sphere. And this
• 06:59speaks to the biological relevance
• 07:01of splicing factor mutations in
• 07:02MDS biology.
• 07:07So the,
• 07:08splicing factor mutations were first
• 07:10described in two thousand eleven
• 07:11by a British group and
• 07:12a Japanese group. And since
• 07:14then,
• 07:15they have been known to
• 07:16occur not just in hematologic
• 07:17malignancies, but also in other
• 07:18cancers.
• 07:19Although you can see that
• 07:20the majority of,
• 07:21or the the highest frequency
• 07:23of these mutations is in
• 07:24minor malignancies,
• 07:25we also see them in
• 07:26chronic lymphocytic leukemia.
• 07:28We also see them in
• 07:28solid cancers, such as melanomas,
• 07:30bladder cancer, pancreatic cancer, and
• 07:32breast cancer.
• 07:34However, as I mentioned, they
• 07:35are very prevalent in myeloma
• 07:36emergencies, and this has been
• 07:37the focus of our work
• 07:38in the lab, which is
• 07:39to understand biochemistry of splicing
• 07:41factor mutations in MDS and
• 07:42AML.
• 07:45Now we are all familiar
• 07:46with the central dogma of
• 07:47molecular biology, which is that
• 07:48the DNA is transcribed into
• 07:50messenger RNA or mRNA, which
• 07:52is then translated into proteins.
• 07:54However, the process of transcription
• 07:56is not a straight straightforward
• 07:58process.
• 07:58The DNA
• 08:00has several noncoding regions, which
• 08:01are called as introns. And
• 08:03during the process of transcription,
• 08:04these intronic,
• 08:06regions from the pre mRNA
• 08:07have to be spliced out.
• 08:09And this process of removing
• 08:10the introns is, which we
• 08:11call splicing, is facilitated by
• 08:13multiple proteins. There are over
• 08:14three hundred RNA binding proteins
• 08:16which facilitate splicing.
• 08:18And these can be they
• 08:19fall under category of what
• 08:21we call splicing factor splicing
• 08:22factor proteins.
• 08:24Now splicing is a very
• 08:25complex process. It's a multistep
• 08:26process, and it involves multiple
• 08:28proteins.
• 08:30However, it is very efficient.
• 08:31It is so efficient that
• 08:32it occurs concurrently with transcription.
• 08:37Now although there are, as
• 08:38I mentioned, close to three
• 08:39hundred RNA binding proteins that
• 08:40facilitate splicing, only a handful
• 08:42of,
• 08:44genes, splicing factor genes are
• 08:45recurrently mutated.
• 08:47These include SF three b
• 08:48one, which is the most
• 08:49commonly mutated gene in MDS.
• 08:51And then there is u
• 08:52two f one and SRS
• 08:54f two.
• 08:55So coming to the s
• 08:56SF three b one, which
• 08:57is actually the focus of
• 08:58our lab work, which is
• 08:59look understanding s f three
• 09:00mutations in MDS, the most
• 09:02common mutational hotspot
• 09:04occurs at the k seven
• 09:05hundred locus
• 09:06followed by the k triple
• 09:08six and the r six
• 09:09twenty five.
• 09:11So there are several unique
• 09:12features to splicing factor mutations.
• 09:14First is that these mutations
• 09:16are nonsynonymous, which means a
• 09:17mutation in the splicing factor
• 09:19gene doesn't result in loss
• 09:21of function or loss of
• 09:22protein expression, but a change
• 09:24in the,
• 09:25amino acid sequence,
• 09:27possibly changing its function.
• 09:29Second, these mutations always occur
• 09:31hit are heterozygous in nature,
• 09:32which means a cell that
• 09:33harbors these splicing factor mutations
• 09:35cannot tolerate
• 09:36mutation in both the alines
• 09:38of the splicing factor gene.
• 09:40Third is these mutations, the
• 09:42splicing factor mutations co occur
• 09:44with other epigenetic modified genes.
• 09:47However, what is very notable
• 09:49is that these, mutations are
• 09:51mutually exclusive to one another.
• 09:53In fact, if you can
• 09:54see here in this this
• 09:55is data from the cBioPortal
• 09:56where you can see that
• 09:57they are largely mutually exclusive
• 09:59to one another. And, generally,
• 10:00when we see this phenomenon
• 10:01of mutual exclusivity,
• 10:03we're talking about biological convergence,
• 10:04which is that,
• 10:06common biological mechanism may be
• 10:08driving the pathogenesis across the
• 10:09splicing factor mutations.
• 10:13Given the preeminent roles in
• 10:14splicing,
• 10:15it has started mutations in
• 10:16the splicing factors.
• 10:18Results in misplicing of downstream
• 10:20target genes, whether it be
• 10:21tumor suppressor genes or oncogenes.
• 10:24And this results in change
• 10:25in protein sequence or protein
• 10:27function of these mispliced genes.
• 10:29This is something we refer
• 10:30to as the single gene
• 10:31model paradigm.
• 10:34And this is just a
• 10:35partial list,
• 10:36where studies have looked at
• 10:37potentially misplaced genes that may
• 10:39be of relevance due to
• 10:40mutation in the upstream splicing
• 10:41factor.
• 10:43Now this model has several
• 10:45limitations and some of which
• 10:46I've outlined here. Firstly, this
• 10:48paradigm doesn't explain the mutual
• 10:49exclusivity
• 10:50that I described in the
• 10:51previous slide.
• 10:53Second, the degree of misplacing
• 10:54that happens in these target
• 10:56genes is relatively small, and
• 10:57it is inconsistent across MDS
• 10:59datasets.
• 11:00Finally, the changes in the
• 11:02RNA isoform ratios that we
• 11:04see doesn't correspond into the
• 11:06doesn't correspond to the protein
• 11:07expression changes.
• 11:09And so given the limitations
• 11:10of this model, we looked
• 11:11at an additional context in
• 11:12which, these splicing factors operate.
• 11:15And that is, of course,
• 11:16transcription splicing.
• 11:18So the process of transcription
• 11:20where the DNA is being
• 11:21made into RNA, it begins
• 11:22with the RNA polymerase or
• 11:23pol two
• 11:24binding to the promoter sequence
• 11:26of the DNA.
• 11:27Once the pol two binds,
• 11:28it then transcribes across the
• 11:30length of the gene making
• 11:31the nascent RNA.
• 11:33And then it finally terminates
• 11:34with the RNA polymerase two,
• 11:37terminating at the transcription site.
• 11:38It is now understood that
• 11:40splicing machinery
• 11:41interacts with the pore
• 11:43two at multiple aspects, including
• 11:44at initiation,
• 11:46elongation, and termination.
• 11:47And so this made us
• 11:48question or hypothesize
• 11:50whether mutations in splicing factors
• 11:52such as in s f
• 11:53three b one affect the
• 11:54way the pore two molecule
• 11:55itself is moving, what we
• 11:56call as RNA transcription kinetics.
• 12:00So to be able to
• 12:01study
• 12:03transcription, we need a suitable
• 12:04disease model. And one of
• 12:06the major limitations in the
• 12:07field has been the lack
• 12:09of a suitable isogenic scalable
• 12:11model system.
• 12:12By isogenic, I mean that
• 12:13the conditions that are being
• 12:15compared, in this case, the
• 12:16SFG one wild type and
• 12:18the SFG mutant, share the
• 12:19same genetic features except for
• 12:21the mutation in question.
• 12:23Scalable, meaning we're able to
• 12:25expand these cells to very
• 12:26high numbers
• 12:27before we express the mutant
• 12:28splicing factor protein.
• 12:30Third is inducible, which means
• 12:31we are able to temporarily
• 12:33regulate the expression of the
• 12:34mutant splicing factor protein.
• 12:35And these challenges come from
• 12:37the fact
• 12:38that these splicing factor mutations,
• 12:41although they promote clonal advantage
• 12:42in the in vivo state,
• 12:44they paradoxically
• 12:45inhibit cell survival and cell
• 12:47growth in the in vitro
• 12:48state in fast dividing cells.
• 12:50And shown here, just to
• 12:52exemplify this, when we overexpressed
• 12:54s f three one mutant
• 12:55in the k for sixty
• 12:56two cells, we saw that
• 12:57there was a dramatic reduction
• 12:58in the cell growth, eventually
• 13:00culminating in growth arrest.
• 13:02This is not an obscene
• 13:03singular observation by us, but
• 13:04it has been reported by
• 13:05multiple other investigators in the
• 13:08field. And so this presents
• 13:09some very unique challenges to
• 13:10be able to use a
• 13:11genome in genome editing system
• 13:13such as CRISPR Cas9
• 13:14to knock in for these
• 13:15mutations.
• 13:16First is that a knock
• 13:18in model system
• 13:20requests, the cellular mechanism called
• 13:22as homologous recombination, which is
• 13:23far less efficient than nonhomologous
• 13:26in joining, which is what
• 13:27is involved in the knockout
• 13:29model systems.
• 13:30Second
• 13:31is that the mutant splicing
• 13:32factor,
• 13:33once you knock in for
• 13:34this mutation, it starts to
• 13:35express right away. The proteins
• 13:36start to express right away.
• 13:38And because these are toxic,
• 13:39the cells fail to expand.
• 13:40So even though we have
• 13:41not been the mutation, because
• 13:42the cells fail to expand,
• 13:44we cannot isolate the clones.
• 13:46Finally, a constitutive expression model
• 13:49system cannot be used to
• 13:50study acute effects of the
• 13:52mutant's pricing factor protein on
• 13:53whole transcription.
• 13:55And so to circumvent these
• 13:56challenges, we developed a novel
• 13:57strategy, which which I'm calling
• 13:59the AV intranetrap CRISPR system.
• 14:01This is a a strategy
• 14:02that we've already published on,
• 14:04and so I won't go
• 14:04into the details of this
• 14:05strategy.
• 14:06But the highlights of this
• 14:07strategy is that the CRISPR
• 14:09Cas9 is what knocks the
• 14:11mutate knocks in the mutation,
• 14:13and the intron trap prevents
• 14:15the mutant allele from expressing.
• 14:18So it keeps it out
• 14:19of frame, basically. And it
• 14:20is only after we use,
• 14:22doxycycline in usable v pre
• 14:23recombinase that we're able to
• 14:24flox out the cassette and
• 14:26put the mutant allele in
• 14:27frame so as to be
• 14:28able to express the mutant
• 14:29splicing factor protein. And this
• 14:30has been a blessing to
• 14:31us because with this system,
• 14:33we are actually able to
• 14:34study acute effects using,
• 14:36an isogenic heterozygous system.
• 14:38We extensively validated this system,
• 14:40and we found that it
• 14:41was at seventy two hours
• 14:42out after expression of after
• 14:44exposure to doxycycline
• 14:45that, there is optimum expression
• 14:47of the mutant splicing factor
• 14:48protein.
• 14:50So now that we have
• 14:51the that we have the
• 14:52disease model, the next step
• 14:53was to understand
• 14:54how these mutations
• 14:56are changing RNA polymerase transcription.
• 15:01So when we talk about
• 15:01nascent RNA and co transcription
• 15:03splicing,
• 15:04we cannot use steady state
• 15:05RNA seq. So steady state
• 15:07RNA seq or bulk RNA
• 15:08seq is where we take
• 15:10the cells, perform a whole,
• 15:12cell, extract,
• 15:13and then extract the RNA
• 15:15to be able to study
• 15:16it. So this RNA, which
• 15:17is a steady state RNA,
• 15:18is fully processed RNA. This
• 15:20is the RNA that has
• 15:21already been made. And so
• 15:22we cannot inform us what
• 15:24is going on at the
• 15:25nascent RNA level.
• 15:26And so to be able
• 15:27to study what is happening
• 15:28at the nascent RNA level,
• 15:30there are a couple ways
• 15:31to study it. So one
• 15:32aspect is looking at the
• 15:33number of poll two molecules
• 15:35in a unit space,
• 15:37which we refer to as
• 15:38poll two density.
• 15:39And the other aspect is
• 15:40looking at the amount of
• 15:41nascent RNA, which is shown
• 15:42here in blue, as it
• 15:44is being made by the
• 15:44pole to molecule.
• 15:46This is what we call
• 15:47nascent RNA synthesis rate.
• 15:51So we first, using our
• 15:52model, we first looked at
• 15:53the pole to density changes.
• 15:55And shown here is ChIP
• 15:56seq where we profile the
• 15:58poll to density changes, and
• 15:59we've and you've see that
• 16:00there is increased poll to
• 16:02density,
• 16:03in the gene body region
• 16:04in the mutant. So here
• 16:05in this metagen plot, the
• 16:07TSS is the transcription start
• 16:08site. This is where the
• 16:10poll to molecule binds and
• 16:11starts to transcribe.
• 16:12The TES is the transcription
• 16:14end site, which is where
• 16:15the poll to molecule terminates.
• 16:18The region in between the
• 16:19TSS and the t TS
• 16:20is the gene body region.
• 16:22And you can clearly see
• 16:23here that there was increased,
• 16:25pool to density
• 16:26in the mutant.
• 16:29Now we collaborated these
• 16:31findings using an, a complimentary
• 16:33assay, which is called the
• 16:34GroSeq.
• 16:35And this is also a
• 16:36pool to density based technique.
• 16:38And, it's a nuclear run
• 16:39on assay. And you can
• 16:40similarly see that as we
• 16:41have seen with ChIPSeq, there
• 16:43is increased gene body density
• 16:44in the s f three
• 16:45one mutant.
• 16:47We then look to see
• 16:47further whether there is a
• 16:49differential
• 16:49portal density based on the
• 16:51region within the gene body.
• 16:53Specifically, we looked at, the
• 16:55intronic regions versus the exonic
• 16:56regions.
• 16:58And you can see that
• 16:58compared to the exonic regions,
• 17:00there is a striking increase
• 17:01of pole to density in
• 17:03the intronic regions.
• 17:04And this makes sense because
• 17:06SFTP one is a part
• 17:07of the uto complex, which
• 17:08bind to the binds to
• 17:09the branch point sequence of
• 17:10the intron.
• 17:13Now the two techniques I
• 17:14did I just described are
• 17:15pole to density based techniques,
• 17:17and so we followed this
• 17:18up with a a technique
• 17:19that looks at the nascent
• 17:20hardness in the system.
• 17:22And for this, we use
• 17:22a technique called TT time
• 17:24lapse sequencing.
• 17:26So in this technique, what
• 17:27we do is we expose
• 17:28the cells to a five
• 17:29minute pulse of fourth diurethane.
• 17:30So fourth diurethane gets incorporated
• 17:32into the nascent RNA during
• 17:33that five minute period. And
• 17:35then what we do is
• 17:36we pull down for the
• 17:36nascent RNA, and we quantify
• 17:38the nascent RNA signal.
• 17:40And you can see here
• 17:41that it is in the
• 17:42s f t m mutant
• 17:42that we see a reduction
• 17:44in the nascent RNA gene
• 17:45body signal. This despite the
• 17:47fact that we see increased
• 17:48pore to density.
• 17:50So strongly suggesting that the
• 17:51findings that we see are
• 17:52consistent with one of a
• 17:54decreased pore to speed within
• 17:55the gene body regions.
• 17:59Now so far, what I've
• 18:00shown you is,
• 18:01RNA pore to speed kinetics.
• 18:03And so our next question
• 18:05was to see how the
• 18:06s f three one mutation
• 18:07changes the core transcription splicing.
• 18:09So this is splicing as
• 18:10it happens on the nascent
• 18:12RNA as the portal is
• 18:13transcribed.
• 18:14And for this, we use
• 18:15the technique called long read
• 18:16sequencing.
• 18:17So this is a technique
• 18:18that was,
• 18:19pioneered by doctor Colin Yugaber,
• 18:21who's one of my co
• 18:22mentors.
• 18:23And what the long read
• 18:24sequencing allows us to do
• 18:26is that unlike the short
• 18:27read sequencing where we fragment
• 18:29the RNA and then perform
• 18:30sequencing, in the long lead
• 18:31sequencing, we are able to
• 18:32sequence the whole,
• 18:34transcript, the whole RNA molecule.
• 18:36And so this is much
• 18:37more accurate in be in
• 18:38quantifying the core transcriptional splicing
• 18:41efficiency.
• 18:42Using this data, we computed
• 18:44something called the CoSC,
• 18:46which is basically the number
• 18:47of spliced reads to the
• 18:48total reads spanning a particular
• 18:50intron. Shown in this illustration,
• 18:52you can see that the
• 18:53spliced reads spliced reads here
• 18:55are shown in black, dense
• 18:56based reads in blue.
• 18:58The total reach is the
• 18:59black plus blue, whereas the
• 19:01splice rates here at the
• 19:02black. So we computed a
• 19:03metric called. We performed this
• 19:05analysis genome wide. And shown
• 19:07on the right, you can
• 19:08see that there is a
• 19:08significant reduction
• 19:10in the core transcription splicing
• 19:12ratio in the mutant.
• 19:14We collaborated our data with
• 19:16a similar analysis on a
• 19:17GroSeq, and we similarly see
• 19:19that there is a reduction
• 19:20in core transcription splicing efficiency
• 19:22in the mutant.
• 19:23Strongly suggesting
• 19:24that the s f three
• 19:25one mutant not only reduces
• 19:27the port to speed in
• 19:28the gene body, but also
• 19:29reduces the efficiency at which
• 19:31splicing is happening at the
• 19:32nascent RNA level.
• 19:34So given this data, we
• 19:35next wanted to understand what
• 19:36might be the mechanism that
• 19:38is driving the portal elongation
• 19:40defect.
• 19:41So we turned our attention
• 19:42to recent structural studies that
• 19:43have looked at how splicing
• 19:45interacts with transcription.
• 19:48One such paper, they they
• 19:49describe a model called the
• 19:50intron loop model. So in
• 19:51this model,
• 19:52shown here is the portal
• 19:54molecule
• 19:55as it is transcribing across
• 19:56the exon.
• 19:58And you can see the
• 19:59nascent RNA shown in red
• 20:00here that is exiting from
• 20:01the pore two exit site,
• 20:02and you can see that
• 20:03it's already capped
• 20:05as it is being transcript.
• 20:06Now as the pore two
• 20:07molecule reaches the five prime
• 20:09splice site, it binds to
• 20:10the human complex. So human
• 20:11complex is the is one
• 20:13of five spliceosome complexes,
• 20:15and it is the first
• 20:16complex that binds to the
• 20:18pre mRNA.
• 20:20The portal molecule, as it
• 20:21transcribes across the intron, remains
• 20:22bound to the u one,
• 20:24and this results in exclusion
• 20:25and looping of the,
• 20:27pre mRNA nascent pre mRNA.
• 20:30It as the portal molecule
• 20:32reaches the three prime splice
• 20:33site, it is then that
• 20:35the uto components,
• 20:36which includes s f three
• 20:37b one, to assemble on
• 20:38the portal surface.
• 20:41And what the model suggests
• 20:42is that it is only
• 20:43after the complete assembly of
• 20:44the uto comp components that
• 20:46the portal is able to
• 20:47liberate itself from the human
• 20:48so as to be able
• 20:49to transcribe into the downstream
• 20:51exon.
• 20:54And so based on all
• 20:55this, we wondered whether the
• 20:56s f three mutation is
• 20:57hampering its association with other
• 20:59u two components to efficiently
• 21:00assemble to form the u
• 21:01two complex.
• 21:03So when we look at
• 21:03the splicing,
• 21:05assembly,
• 21:06the process obviously involves multiple
• 21:08steps, but the binding of
• 21:09s f three b one
• 21:10to the u two, to
• 21:11the pre mRNA occurs relatively
• 21:13early in the splicing process.
• 21:15Recent studies have shown that
• 21:17s f three b one
• 21:18requires to interact with
• 21:20multiple accessory proteins, but the
• 21:21most important being h status
• 21:23f one and d d
• 21:24x forty six.
• 21:26And these interactions are required
• 21:28for the s f three
• 21:28b one to undergo certain
• 21:29confirmation changes
• 21:31so as to be able
• 21:32to bind
• 21:34and assemble the u two
• 21:35complex
• 21:36with the other u two
• 21:36components.
• 21:38And so what we did
• 21:39was we performed an IP
• 21:40where we look to see
• 21:41whether the s f three
• 21:42one mutant is not able
• 21:44to interact properly with these
• 21:46accessory proteins.
• 21:48We specifically looked at the
• 21:49chromatin fraction because this is
• 21:50where
• 21:51the splicing is happening. And
• 21:53what you can see that
• 21:54compared to the s f
• 21:55three one wild type,
• 21:57it is in the s
• 21:57f three one mutant that
• 21:58we see a defective interaction
• 22:00with each status of one,
• 22:01u two f one, and
• 22:02u two f two.
• 22:04We next wondered whether this
• 22:05has a functional significance. And
• 22:07for this, what we did
• 22:08was we performed a a
• 22:09very classically in vitro,
• 22:11biochemical assay called the in
• 22:13vitro splicing assay. So what
• 22:15we do in this is
• 22:16with that we expose
• 22:18radio labeled pre mRNA substrate
• 22:21to nuclear lysates from the
• 22:22wild type condition and the
• 22:23mutant condition.
• 22:26So shown here, the e
• 22:27complex is the early complex,
• 22:29and this is much smaller
• 22:30because there are fewer proteins
• 22:31on the pre mRNA, and
• 22:32so it migrates faster on
• 22:33the gel. Whereas the a
• 22:35complex, which is a later
• 22:36complex, is much larger because
• 22:37of more proteins on the
• 22:38pre mRNA, and it migrates
• 22:40slower.
• 22:41So if there is a
• 22:41defective transition from the e
• 22:43to a complex, it would
• 22:44be it would be reflected
• 22:45on the gel. And in
• 22:46fact, you can see that
• 22:47compared to the sf t
• 22:48one wild type, which is
• 22:49shown here,
• 22:50it is in the sf
• 22:51t one mutant. You see
• 22:52a defective transition to the
• 22:54a complex across the various
• 22:55time points.
• 22:57So this showed that, indeed,
• 22:59the SFM mutation is hampering
• 23:01the early splicing assembly transition.
• 23:06So despite these multiple assays,
• 23:08the reviewer still wanted us
• 23:10to show a direct confirmatory
• 23:11evidence
• 23:12that there is indeed a
• 23:13defective u one poll to,
• 23:15release in the mutant.
• 23:17And so what we what
• 23:18we did was we developed
• 23:20a a novel biochemical assay,
• 23:21which I'm calling as the
• 23:22poll to release assay.
• 23:24And this is inspired from
• 23:25the PTFE release assay that
• 23:26was described in, more than
• 23:28a decade ago. So what
• 23:29we did in this is
• 23:30we exposed cells, h two
• 23:32ninety three c t cells,
• 23:33to alpha amyloidin resistant h
• 23:36attack pore two,
• 23:37expressing plasmid.
• 23:39So,
• 23:40alpha amanitin is a pol
• 23:41two toxin. What it does
• 23:42is it binds to the
• 23:43pol two molecule, prevents it
• 23:44from elongating, and,
• 23:46makes it disengage from the
• 23:47chromatic.
• 23:48However, if you create certain
• 23:49mutations within the pol two
• 23:50molecule, you can actually create
• 23:52resistance to the alpha.
• 23:54What we also did was
• 23:55that we tagged the poldo,
• 23:56and so we are able
• 23:57to distinguish exogenous poldo
• 23:59from the endogenous native poldo.
• 24:02So after transplanting the cells
• 24:03with this with such a
• 24:04plasmid,
• 24:05we expose the cells to
• 24:06alpha mitten for twelve hours.
• 24:08So this would selectively enrich
• 24:09for the h attack portal
• 24:11while causing the native portal
• 24:12to dis to disengage.
• 24:15We then performed,
• 24:18IP to pull down for
• 24:19a protein called FUS. So
• 24:20FUS is a bridge
• 24:22that connects the u one
• 24:23to the pole two.
• 24:26And what we did next
• 24:28was we exposed the chromatin
• 24:29complexes containing this u one
• 24:31pole two to nuclear extents
• 24:32from the wild type and
• 24:33the mutant condition.
• 24:35The expectation from this assay
• 24:36is that if there is
• 24:37a defective interaction of the
• 24:39u one poll two in
• 24:39the mutant, you would see
• 24:41less tag poll two going
• 24:42into the solution.
• 24:45And, consistently, what we found
• 24:47was across the different time
• 24:48points that were analyzed, we
• 24:50found a defective
• 24:51tag portal illusion
• 24:52into the solution.
• 24:54Strongly suggestive that there is
• 24:55indeed a defective release of
• 24:57u one portal, providing a
• 24:58more direct confirmatory evidence.
• 25:01So this is my model,
• 25:02which is that in the
• 25:03normal physiological state, s f
• 25:05three one is able to
• 25:05assemble with the other u
• 25:06two components
• 25:08to,
• 25:09to create the a complex
• 25:10on the portal surface, and
• 25:11this allows the portal to
• 25:12liberate itself from the u
• 25:13one. However, in the mutant
• 25:14condition,
• 25:15because there is defective assembly,
• 25:17the portal is not able
• 25:18to disengage itself from the
• 25:19u one, resulting in it
• 25:21getting stuck within the intronic
• 25:22regions causing a pile up
• 25:23and slowing of transcription speed.
• 25:27So great. We we we
• 25:28find a mechanism that is,
• 25:30that can explain the transcription
• 25:31defect. But what is the
• 25:33what are the physiological consequences
• 25:34of this transcription dysregulation?
• 25:37We know that transcription is
• 25:38a very tightly coordinated process.
• 25:41It has to be very
• 25:41closely coordinated with replication so
• 25:43that they do not conflict
• 25:45with one another.
• 25:46Also, if the poll to
• 25:47speed is too slow, it
• 25:48creates what are called r
• 25:50loops, which are triplex structures.
• 25:52And consistently, what we found
• 25:53was there was increase in
• 25:55r loops,
• 25:56suggestive of a slow pull
• 25:58to speed and also transcription
• 26:00replication conflicts because the slowing
• 26:01of the pull to speed
• 26:03was causing it to conflict
• 26:04with the the replication machinery
• 26:06causing TRCs.
• 26:08Now both our loops and
• 26:09TRCs are genotoxic. They're they're
• 26:11DNA damaging. And And what
• 26:12happens is when these are
• 26:14increased, this results in a
• 26:15growth growth phase arrest. And
• 26:17that is what we found.
• 26:18We found that these cells
• 26:19were eventually going into s
• 26:20phase growth arrest.
• 26:22Now this has been shown
• 26:23by other investigators, but we
• 26:24show for the first time
• 26:25that this is tied to
• 26:26desegregated transcription kinetics.
• 26:30So what about the effects
• 26:31on chromatin?
• 26:32Now we know that,
• 26:34chromatin and transcription are very
• 26:36tightly interlinked. The chromatin needs
• 26:38to be opened so that
• 26:39the poll two can bind
• 26:40to the chromatin and transcribe.
• 26:42And so to understand the
• 26:42chromatin landscape changes, we broadly
• 26:45profiled for the major histone
• 26:46marks and also looked at
• 26:47chromatin accessibility.
• 26:50The the most,
• 26:51notable effects were on the
• 26:53asymmetylation
• 26:54signal where we see that
• 26:55there was a there was
• 26:56a global reduction in the
• 26:57promoter asymmetration,
• 26:59in the SFG one method.
• 27:02It's very interesting that these
• 27:03changes in the h k
• 27:04four trimethylation
• 27:05very closely corresponded to the
• 27:06pole to density changes.
• 27:11So when we talk about
• 27:12chromatin accessibility,
• 27:13the promoter region, which is
• 27:14shown here, is a nucleosome
• 27:15free region. This is because
• 27:17the portal has to be
• 27:18able to bind along with
• 27:19other transcription factors and transcribe.
• 27:21So this region is nucleosome
• 27:22devoid.
• 27:24However, if for any reason
• 27:26the port the, the promoter
• 27:27chromatin is not active, what
• 27:29happens is these nucleosomes reposition
• 27:31into the promoter
• 27:32causing an increased nucleosome density.
• 27:36And so when we looked
• 27:37at nucleosome density changes
• 27:39using attack sequencing, we find
• 27:41that indeed there was increased
• 27:42nucleosome density at the promoter
• 27:43region, which is consistent with
• 27:45the closed chromatin configuration.
• 27:49So given these changes to
• 27:50the promoter chromatin accessibility and
• 27:52to, the issue for trimethylation,
• 27:56we wondered whether this is
• 27:57tied in any way to
• 27:58the port two density changes
• 28:00of the promoter.
• 28:01And for this, we used
• 28:02another technique, which is called
• 28:03MNET seek. So this is
• 28:05an RNA,
• 28:06poll to pull down technique,
• 28:07and this allows us to
• 28:08profile poll to density at
• 28:10single nucleotide resolution.
• 28:12And you can see clearly
• 28:14here that compared to the
• 28:14wild type, which is shown
• 28:15in blue, there is a
• 28:17a dramatic reduction in the
• 28:18poll to density at the
• 28:19promoters in the s f
• 28:20three mutant.
• 28:21Now the poll to density
• 28:23at the promoters is tightly
• 28:24regulated by multiple protein complexes.
• 28:27Perhaps the most important one
• 28:28of them is the PTFB
• 28:30release complex.
• 28:31What PTFB does is it
• 28:33it evicts or releases the
• 28:34pore two from the promoter
• 28:36into the gene body regions.
• 28:37So if there's increased PTFB,
• 28:39you would expect that there
• 28:40there to be decreased pore
• 28:41two density. And, consistently, what
• 28:43we found was when we
• 28:43profiled the PTFB
• 28:45promoter recruitment, it was indeed
• 28:47increased in the mutant. This
• 28:48fits, in line with the
• 28:50with the paradigm where the
• 28:52increased PTFB is causing porter
• 28:54to be released into the
• 28:55gene bodies, causing a decrease
• 28:56in porter density.
• 28:59So this is what we
• 29:00think is happening. In a
• 29:02normal physiological state, the pole
• 29:03two at the promoter region
• 29:05associates with multiple other transcription
• 29:07factors and chromatin remodelers
• 29:09so as to be able
• 29:10to keep the chromatin accessible.
• 29:12However, if there is a
• 29:14loss of pole to density
• 29:15such as is happening in
• 29:17the mutant,
• 29:18it cannot recruit chromatin remodelers
• 29:20and transcription factors. In fact,
• 29:22we exemplified this using
• 29:24a chip, looking at CHD
• 29:26one occupancy, and you can
• 29:27see how that it's dramatically
• 29:28reduced in the mutant.
• 29:31And so what happens is
• 29:32the nucleosomes reposition to the
• 29:34promoter, shutting down the promoter
• 29:35chromatic.
• 29:37Now all of this data
• 29:38that I've shown so far
• 29:39is in in in k
• 29:40phase six two cells. And
• 29:41so we sought to validate
• 29:42these findings in, human MDS
• 29:44cells. So what we did
• 29:45was we obtained patient samples,
• 29:47and we,
• 29:49we flow sorted for CD
• 29:51thirty four cells. And we
• 29:52profiled the CD thirty four
• 29:53cells for polder density changes
• 29:55as well as for chromatin
• 29:56accessible changes.
• 29:58And, for this, we performed
• 29:59special assays which are scalable,
• 30:02and we similarly find the
• 30:03change in polder distribution
• 30:05and increased nucleosome density.
• 30:07We also validate this in
• 30:08a mouse model.
• 30:10Now this model has been
• 30:11described and published, way back
• 30:12in two thousand sixteen. It
• 30:14has a relatively modest metabolic
• 30:16phenotype. However, biochemically, we see
• 30:18that these changes in the
• 30:19promoter bordo density and the
• 30:21nucleosome density
• 30:22occur very early on in
• 30:23the disease process.
• 30:27So, great. I've we have
• 30:28our data so far suggests
• 30:29that the transcription changes are
• 30:31altering
• 30:32chromatin. They're causing DNA damage.
• 30:34But what about the effects
• 30:35on alternate splicing program?
• 30:37So we went on to
• 30:38look to see how the
• 30:39transcription alterations may be changing
• 30:41the alternate splicing program. This
• 30:43data that I've not published
• 30:44yet, but this is ongoing,
• 30:45and we hope to publish
• 30:46soon.
• 30:47So before I go into
• 30:48the data, what is alternative
• 30:49splicing? It is a mechanism
• 30:51of which the cell
• 30:53generates RNA and protein isoform
• 30:54diversity.
• 30:56So in certain situations, the
• 30:57cell may choose to use
• 31:00or may choose to skip
• 31:01or include certain exons, which
• 31:02we we which we call
• 31:03skip text on events.
• 31:05In certain cases, it may
• 31:06choose to retain an intron,
• 31:07which we call intron retention,
• 31:09or it may choose to
• 31:10use alternative five prime or
• 31:11alternative three prime splice sites,
• 31:13which we call a five
• 31:14prime alternative or three prime
• 31:15alternative splice site selection.
• 31:19Now this process of alternative
• 31:20splicing
• 31:21is very tightly coordinated by
• 31:23multiple RNA binding proteins. Perhaps
• 31:24the most well described and
• 31:25best characterized are the SA
• 31:27proteins and the h and
• 31:28r and p proteins.
• 31:30These have been described since
• 31:31the nineteen nineties, and their
• 31:32functional antagonism has been well
• 31:33characterized.
• 31:36So SA proteins can be
• 31:37best categorized or classified as
• 31:38splicing enhancers,
• 31:40and h and rPs can
• 31:41be best categorized or classified
• 31:42as splicing repressors.
• 31:45Now these two
• 31:47master RNA binding protein regulators
• 31:49are in turn regulated by
• 31:50multiple phosphorylation and kinase pathways.
• 31:52Just shown here is a
• 31:53list, an extensive list of
• 31:55all those pathways, and these
• 31:56pathways are very sensitive to
• 31:57cellular stress signals.
• 32:02Now given the the roles
• 32:03in splicing and alternative splicing,
• 32:05a lot of,
• 32:06effort and interest has been
• 32:07looking at how splicing factor
• 32:09mutations change the alternate splicing
• 32:11program. And one such was
• 32:12a group that, published recently
• 32:16on an extensive cohort of
• 32:17close to seventeen hundred splicing
• 32:19factor samples.
• 32:20And, what they found was
• 32:22that the majority of events
• 32:23that were misplaced were those
• 32:24involving skipped exons.
• 32:26So they went on to
• 32:27see further how these skipped
• 32:28exon events are getting altered.
• 32:30And shown on the right
• 32:31is a histogram where you
• 32:32can see that the majority
• 32:33of the misplaced events in
• 32:35the skip dexon category
• 32:36are relatively small in the
• 32:38order of zero to point
• 32:39two.
• 32:40However, there were smaller proportion
• 32:41which were highly misplaced.
• 32:44And so they went on
• 32:45to look further what is
• 32:46happening or what is the
• 32:47overlap of co occurrence of
• 32:49these skip decks on altered
• 32:51events across the splicing factor
• 32:53mutant categories. And you can
• 32:54see here shown in here
• 32:55in blue, you can see
• 32:57that there's hardly any overlap.
• 32:58The the type of splicing
• 33:00skip takes on events that
• 33:01are occurring across the categories
• 33:02show very little overlap.
• 33:04There is one exception to
• 33:06this category, however, and that
• 33:07is the intron retention program.
• 33:08So both this group and
• 33:10another group that published in
• 33:11two thousand eighteen showed that
• 33:13the intron retention program,
• 33:15shows a lot of overlap
• 33:16across the splicing factor categories.
• 33:19So we, we are collaborating
• 33:20with this group, in fact.
• 33:21This is a German group,
• 33:23and they were kind kind
• 33:24enough to provide us with
• 33:24these patient samples, which is
• 33:26close to, in this scale,
• 33:27like, nine hundred samples. And
• 33:29shown here is a clustering
• 33:30heat map where you can
• 33:31clearly see that the intron
• 33:32retention events in the wild
• 33:34type, which is shown here
• 33:35in pink, clusters differently from
• 33:36the mutant, which are shown
• 33:37here in gray.
• 33:38So we decided to look
• 33:39more granularly at the intron
• 33:41retention events.
• 33:42So to classify the intron
• 33:44retention events, we can classify
• 33:45them into loss or gain.
• 33:47Loss means those events where
• 33:49there is excessive splicing, increased
• 33:51splicing happening in the mutant.
• 33:53And you can see clearly
• 33:54here that there is a
• 33:55striking overlap of the shared
• 33:57loss of intron retention events
• 33:58across the three mutation categories.
• 34:01We then looked at gain
• 34:02of intron retention events. So
• 34:03these are events where there
• 34:05is decreased splicing in the
• 34:06mutant
• 34:07causing increased intron retention.
• 34:09And you can similarly see
• 34:10a very high overlap of
• 34:11shared gain of intron retention
• 34:13events.
• 34:14What is perhaps most notable
• 34:16in this data is that
• 34:17there were no discordant events,
• 34:18which means we did not
• 34:19see a single event out
• 34:20of the nine thousand events
• 34:21where there was a gain
• 34:23in one category and loss
• 34:24in the other two or
• 34:25a loss in one and
• 34:26gain the other two.
• 34:28So when we look at
• 34:29when we go back to
• 34:30the, the splicing factors proteins,
• 34:32you can see here that
• 34:33the s f three b
• 34:34one, u two f one,
• 34:35and s r s two
• 34:36bind to very
• 34:37distinct regions on the intronal
• 34:39axon.
• 34:40S f three one binds
• 34:41to the branch point sequence.
• 34:42U two f one binds
• 34:43to the three prime splice
• 34:44site, and s f two
• 34:46binds to the axon splicing
• 34:47enhancer. And so it is
• 34:48highly implausible that
• 34:50mutations in splicing factors that
• 34:52bind to different regions within
• 34:53the exonic or intronic regions
• 34:54would have the same effects
• 34:55on the internal program.
• 34:58So on the one hand,
• 34:59we show that the all
• 35:00the mutations, splicing factor mutations
• 35:02are causing a common phenomenon
• 35:03of replicative stress. On the
• 35:05other hand, we see these
• 35:05large scale concordant introned even
• 35:07changes. And so that made
• 35:09us wonder whether there is
• 35:10a global kinase of phosphorylation
• 35:11pathway that is being dis
• 35:12regulated, which may in turn
• 35:14be changing the function or
• 35:15activity of key or any
• 35:17binding proteins.
• 35:20For this, what we did
• 35:21was we decided to leverage
• 35:22the data from the encode
• 35:23consortium.
• 35:24So this is a large
• 35:25scale,
• 35:26project that was developed by
• 35:27the human, National Human Genomics
• 35:29Research Institute,
• 35:30a subsequent to the human
• 35:31genomics project. And the goal
• 35:33with this database was to
• 35:34be able to provide researchers
• 35:36around the world with functional
• 35:37data on functional elements and
• 35:39their significance on gene expression.
• 35:41So this database has, data
• 35:43available on RNA binding protein
• 35:45knockdown conditions
• 35:46representing close to four hundred
• 35:48RNA binding proteins. So what
• 35:50we,
• 35:50envision to do with this
• 35:52was to be able to
• 35:53see whether
• 35:54the IR pattern changes that
• 35:55we see in the mutant
• 35:56are similar to any of
• 35:58the RNA binding protein knockdown
• 35:59conditions.
• 36:00So we focus specifically on
• 36:02the four thousand or so
• 36:03events that are commonly
• 36:05altered, whether it be gain
• 36:06or loss, and we perform
• 36:08the clustering analysis. So the
• 36:09the expectation with this clustering
• 36:11analysis
• 36:12is that the conditions which
• 36:13are closest
• 36:14to the mutant condition on
• 36:16the heat map are likely
• 36:17to be most functionally linked
• 36:19or functionally similar,
• 36:21to the mutant category.
• 36:25And what we found was
• 36:26the SRSF and knockdown category
• 36:27was the one that was
• 36:28most similar, to the mutation
• 36:30interon retention changes.
• 36:32Whereas h and r and
• 36:33p a zero and a
• 36:34b were among the ones
• 36:36that were,
• 36:37farthest to the mutant IR
• 36:39category.
• 36:41Shown in another way, this
• 36:42is the SRS of a
• 36:43knockdown condition, which correlated the
• 36:45most with the, IR changes
• 36:47seen in the in the
• 36:47mutant category and the h
• 36:49and r and p a
• 36:49zero and a b being
• 36:51the most anticorrelated.
• 36:53And this is interesting because
• 36:54as I mentioned in previous
• 36:55slide, the ASR proteins and
• 36:57h n r n p's
• 36:57are functionally antagonistic.
• 36:59And
• 37:00so that's that got us
• 37:01interested in in the SRS
• 37:02seven protein. So when we
• 37:03look at SRS seven protein,
• 37:05the structure, it has the
• 37:07RRM domain in the n
• 37:08terminal region and the RS
• 37:09domain in the c terminal
• 37:10region.
• 37:11So the RS domain is
• 37:13as you can see here,
• 37:14it is it is very
• 37:15rich in prolines, arginines, and
• 37:17serines.
• 37:18And so it is it
• 37:19is very richly phosphorylated.
• 37:21It is this phosphorylation
• 37:23that governs its activity,
• 37:25localization,
• 37:28and ability to cause alt
• 37:29alterations in splicing.
• 37:33And the phosphorylation of SRSF
• 37:35one at the RS domain
• 37:36is regulated by two,
• 37:38key kinases, which is the
• 37:39CLK one and the SRP
• 37:41k one. And it is
• 37:42only after SRP k one
• 37:43and CLK one bind to
• 37:45SRSF one that it is
• 37:46able to localize into the
• 37:47nucleus,
• 37:48able to bind to the
• 37:49pre mRNA and facilitate alternate
• 37:51splice.
• 37:53And so what we did
• 37:54was we overexpressed the mutant
• 37:55splicing factor proteins,
• 37:57in KFS six two cells,
• 37:59and we look to see
• 38:00whether the mutation was causing
• 38:01a change in the phosphorylated
• 38:03phosphorylation of SRSF one. And,
• 38:05constantly, what we found was
• 38:07that, there was a reduction
• 38:08in the phosphorylation state of
• 38:10SRSF one.
• 38:12We collaborated our findings using
• 38:14patient samples. So this is
• 38:15just a
• 38:16preliminary confirmation where we took
• 38:18healthy control peripheral blood mononuclear
• 38:20cells and compared it to
• 38:22mutant
• 38:23MDS peripheral blood mononuclear cells.
• 38:25And you can see that
• 38:26compared to the wild type
• 38:27condition, there is a reduction
• 38:28in the hyperphosphorylated
• 38:29form of SRSF1.
• 38:31And, you can see the
• 38:32near absence of the hyperphosphory.
• 38:35So,
• 38:36in fact, now we are
• 38:37collaborating with the Yanshang Ling
• 38:39Liu Lab in the department
• 38:41of pharmacology.
• 38:42Their lab has an expertise
• 38:44in
• 38:45mass spec quantitative phosphoproteomics,
• 38:47and the goal with this
• 38:48is for us to be
• 38:49able to globally
• 38:50profile the phosphoproteomic changes that
• 38:53happen in the splicing factor
• 38:54mutants.
• 38:55In fact, we have already
• 38:56performed this analysis on our
• 38:57receptor receptor mutant condition.
• 38:59And when we when we
• 39:00so so the region that
• 39:01was captured by the phosphor
• 39:02proteomics is the n terminal
• 39:04region
• 39:05of the RS domain.
• 39:08And what we found was
• 39:09consistent with what we've seen
• 39:10in the westerns, there is
• 39:11a reduction in the phosphorylation
• 39:13of the SR S1 in
• 39:14the internal region of the
• 39:16RS domain.
• 39:18In ongoing work, we are
• 39:19now performing westerns and phosphoproteomics
• 39:21on, CD thirty four cells
• 39:23derived from primary patient samples.
• 39:25We are, in fact, expanding
• 39:26our efforts to include a
• 39:27larger cohort of patient samples
• 39:29to validate our findings.
• 39:30We are particularly interested in
• 39:32understanding the mechanistics of how
• 39:34DNA damage may be changing
• 39:35phosphorylation of SRS.
• 39:37To this end, what we
• 39:38did was we
• 39:39exposed,
• 39:41CD thirty four cells to
• 39:42agents like camptothecin
• 39:44and etoposide.
• 39:45So camptothecin
• 39:46induces single stranded DNA damage,
• 39:48whereas etoposide creates double stranded
• 39:50DNA breaks. And you can
• 39:51see here that it is
• 39:52only with the, exposure to
• 39:53camptothecin
• 39:54that we see a change
• 39:55in the levels of phosphorylation
• 39:57in SRS one. And this
• 39:58would be consistent with the
• 40:00fact that the the type
• 40:01of damage attack occurs in
• 40:02splicing factor mutants is one
• 40:04of single standard DNA damage
• 40:05because it's an odd loop
• 40:06mediated damage, which creates single
• 40:07standard loops.
• 40:11Finally, I think our phosphoprotemic
• 40:13data gives us ideas on
• 40:14which access or pathways to
• 40:16investigate further.
• 40:18We are particularly interested in
• 40:19exploring the AKT SRP k
• 40:21one SRS open access, and
• 40:23this is guided by our,
• 40:24phosphoproteamid data.
• 40:29In the final section, obviously,
• 40:31you know, as a physician
• 40:32physician scientist,
• 40:33we always like to, translate
• 40:35our findings in the lab
• 40:37to potential therapies.
• 40:39Now there has been a
• 40:40lot of progress that is
• 40:41being made in the MDS
• 40:42sphere. You can see the
• 40:44whole
• 40:45range of different pathways that
• 40:46are being explored using drugs,
• 40:48and all of these drugs
• 40:49are in various stages of
• 40:50clinical development.
• 40:52And these expanding therapeutic paradigms
• 40:54reflect the
• 40:56increasing knowledge of the pathophysiology
• 40:58of MDS. However, I think
• 40:59there's still a lot of
• 41:00scope for improvement.
• 41:02The outcomes after hypomethylating agent
• 41:04failure remains dismal, and these
• 41:06patients have very poor outcomes.
• 41:08In fact,
• 41:10certain therapies such as splices
• 41:11or modulators,
• 41:12so this is an agent
• 41:13called h three b eighty
• 41:15eight hundred, and there's another
• 41:16agent called a seven e
• 41:17seven one zero seven, which
• 41:19are splicing modulators.
• 41:20And they have been tried
• 41:21in MDS, but the outcomes
• 41:23have been,
• 41:24have not been promising. And
• 41:26in hindsight, this makes sense
• 41:27because I think what our
• 41:28data shows is that, splicing
• 41:30factor mutant MDS is not
• 41:31a disease driven by misplicing,
• 41:33but rather a disease that
• 41:34is probably driven by transcription
• 41:36and chromatin
• 41:37changes.
• 41:39And so can we explore
• 41:40this therapeutic paradigm of transcription
• 41:42chromatin changes in MDS? And,
• 41:44in the day data that
• 41:45I'll show in the subsequent
• 41:46slides,
• 41:47it suggests that we can.
• 41:48Right.
• 41:50So as as I mentioned
• 41:51in previous slides, what the
• 41:53mutation does is it slows
• 41:54down the cells, and it
• 41:55cause it causes global chromatin
• 41:57changes. And so are we
• 41:58able to reverse the chromatin
• 42:00defects?
• 42:01So to be able to
• 42:01reverse the chromatin defects and
• 42:03rescue the cells from the
• 42:04growth arrest, what we did
• 42:05was we exposed the mutant
• 42:07cells,
• 42:08the KFI six two cells,
• 42:10to a forward,
• 42:12genetic screen library.
• 42:13So this is a library
• 42:14that includes shRNAs that target
• 42:16close to three hundred and
• 42:17fifty epigenetic regulators.
• 42:19And the goal with this
• 42:20screen is to be able
• 42:21to, rescue these cells from
• 42:23the growth growth arrest. And,
• 42:24indeed, what we were able
• 42:26to do was rescue the
• 42:27cells.
• 42:27Shown here on the right
• 42:29is a waterfall plot where
• 42:30the the circles in green
• 42:32represent factors which when lost
• 42:34improve the cell survival,
• 42:36whereas circles in pink represent
• 42:37factors which when lost worse
• 42:39in the cell survival.
• 42:41We performed a pathway enrichment
• 42:42analysis,
• 42:43and what we found was
• 42:45that these factors that are
• 42:46shown here in green circles
• 42:47are enriched for the sintree
• 42:48H stack complex.
• 42:51Shown in on the right
• 42:52is just, the top targets
• 42:54that we derived from the
• 42:55screen. So ink two, pH
• 42:56of twenty one a, and
• 42:57h tag two are targets
• 42:59which when we knock down
• 43:00improve the cell survival, whereas
• 43:02knocking down for WDF five
• 43:03and EPC two
• 43:05reduced cell survival further.
• 43:07Now what is notable is
• 43:08that all these five proteins
• 43:09are part of the LST
• 43:11one Synthri H type
• 43:13complex. So this is a
• 43:14major transcriptional repressor complex that
• 43:17modulates a s t k
• 43:18four methylation activity in the
• 43:19cell.
• 43:23We perform secondary validation experiments
• 43:25where we, selectively knock down
• 43:27for these top targets.
• 43:28And, we we,
• 43:30found that indeed knocking down
• 43:31for them improve the cell
• 43:32survival. And this translated biochemically
• 43:35to a reversal of the
• 43:37chromatin accessibility to a normal
• 43:39state and also reversal of
• 43:40transcription kinetics.
• 43:43All this data I showed
• 43:44you was in k phase
• 43:44six two cell lines, and
• 43:45so we went on to
• 43:46validate our findings in MDS
• 43:48patient cells,
• 43:49focusing particularly on WDR five
• 43:51into and h stack two.
• 43:54So shown here are chronic
• 43:56growth assays where, here we
• 43:57have knocked down or reduced
• 43:58the activity of WDR five
• 44:00using a drug called OICR
• 44:02nine four two nine. And
• 44:03you can see that when
• 44:03we knocked it down,
• 44:06using this drug, there are
• 44:07selective reduction in chronic growth
• 44:09in the mutant condition as
• 44:10opposed to the wild type
• 44:11condition.
• 44:13Now there are no drugs
• 44:14that selectively inhibit into an
• 44:17HDAC two. And so we
• 44:18used shRNA
• 44:19to knock down for these
• 44:20proteins.
• 44:21And you can see the
• 44:22the increased growth,
• 44:24upon knocking down for these
• 44:25two as opposed to the
• 44:26decreased growth when we knocked
• 44:27down for tiber five. And
• 44:29so these findings strongly validate
• 44:31our
• 44:32genetic screen results.
• 44:35This brings me to my
• 44:37summary slide.
• 44:39So what we show is
• 44:40that we we characterize splicing
• 44:42factor mutants as functionally epigenetic
• 44:44disorders.
• 44:45And the chromatin landscape changes
• 44:46that we see, seem to
• 44:48be driven by transcription elongation
• 44:49defects, which in turn are
• 44:51caused by
• 44:52misassembly of splicing at the
• 44:54u two complex level.
• 44:57We also characterize disrupted co
• 44:59transcription splicing as a new
• 45:00disease paradigm, and this may
• 45:02extend itself not just to
• 45:03splicing factor cancers, but other
• 45:05cancer entities where there may
• 45:06be increased or decreased expression
• 45:08of splicing factors in the
• 45:10u two assembly.
• 45:12We also find that the
• 45:13transcription defects are causing DNA
• 45:15damage response, which has pleiotropic
• 45:16effects. You can see that
• 45:18the DNA damage response causes
• 45:19these cells to have,
• 45:21our lobes,
• 45:22growth arrest,
• 45:24and also, changes in the
• 45:26chromatin level while also changing
• 45:28the alternate splicing,
• 45:29program.
• 45:31Finally, we're able to reverse
• 45:33some of the chromatin defects
• 45:34by knocking down for key
• 45:35targets in the synthetic HVAC
• 45:36complex, particularly ink two and
• 45:38HVAC two and WDR five,
• 45:40and we're able to rescue
• 45:42these cells. And so we
• 45:43are exploring this complex components
• 45:45as a potential therapeutic target.
• 45:47Now I think one of
• 45:48the key findings from our
• 45:49screen results is that this
• 45:51gives us a two pronged
• 45:51approach. So if when we
• 45:53knock down w d r
• 45:53five, we are killing the
• 45:55mutant cells, and so we
• 45:56may be able to eliminate
• 45:57the mutant clone.
• 45:59Whereas by knocking down into
• 46:00our HVAC two, we are
• 46:01able to help the cells,
• 46:02the mutant cells grow better.
• 46:04So we may actually be
• 46:05able to improve the cell
• 46:06counts this way. So this
• 46:07is a two pronged approach
• 46:08that we we are
• 46:09excited by.
• 46:12And I'll conclude by ongoing
• 46:14work in the lab.
• 46:15These are actually part of
• 46:17the aims for my q
• 46:18eight as well. In the
• 46:19first aim,
• 46:21we are looking at how
• 46:22the mutation affects the pool
• 46:23to density, not just at
• 46:25the lower, order chromatin,
• 46:27which I was referring to
• 46:28the chromatin accessibility,
• 46:29but also how it's, affecting
• 46:31higher order chromatin. So these
• 46:32includes
• 46:33changes that enhance a promoter
• 46:35contacts and as well as
• 46:36higher order chromatin, compaction.
• 46:38In the second aim,
• 46:40we are trying to address
• 46:41or taking a stab at
• 46:42a very basic fundamental biochemical
• 46:44question, which is how is
• 46:45splicing interacting with transcription?
• 46:48Is it,
• 46:50are they linked through the
• 46:51SFTP1 status of a node?
• 46:54Finally, we are testing our
• 46:55targets in patient derived xenograft
• 46:57models,
• 46:58in SFTP mutant MDs.
• 47:02With this, I'd like to,
• 47:04end and thank my mentor,
• 47:05doctor Pillai.
• 47:07My
• 47:08colleagues in the lab. Shout
• 47:09out to who's a very
• 47:10bright postgraduate student who's helped
• 47:12me with some of these
• 47:13assays. Some of which require
• 47:14a failed fair deal of
• 47:16optimization.
• 47:17My thesis committee members,
• 47:19doctor,
• 47:20doctor, and doctor Mushen,
• 47:22our collaborators,
• 47:23and,
• 47:26the NIH, t thirty two
• 47:27leadership, doctor Herbst, who's, helped
• 47:30me support my research in
• 47:31the first two years.
• 47:34My
• 47:35funding support from the Evans
• 47:36Foundation, and finally, ongoing support
• 47:38from the NIDDK
• 47:39gateway.
• 47:41With this, I'll end, and
• 47:42I'm happy to take any
• 47:43questions you may have.
• 47:49Questions?
• 48:00Thanks. Well, thank you. That
• 48:01was a really beautiful talk,
• 48:04and,
• 48:05I I think you really
• 48:06brought out this polyfunctional
• 48:08aspects of these mutations,
• 48:11which affect, I think, at
• 48:12least four cancer relevant processes.
• 48:16The spectrum of of proteins
• 48:18that are spliced products,
• 48:21the,
• 48:22effects on
• 48:24transcriptional regulation,
• 48:26the effects on on overall
• 48:28chromatin organization,
• 48:30and, of course, DNA damage
• 48:31responses.
• 48:32What I'm wondering is
• 48:35how you weigh the impact
• 48:37of these various processes,
• 48:39all of which could
• 48:41contribute
• 48:43in in different proportionate ways
• 48:45to the
• 48:46phenotype
• 48:48associated with the disease you're
• 48:49looking at, which is a
• 48:50clonal expansion relative to the
• 48:52to Yes. Counterparts.
• 48:53Yeah. That's a very difficult
• 48:54question I admit. And so,
• 48:56you know, the cell may
• 48:57co opt multiple mechanisms, and
• 48:59that may eventually, at variable
• 49:00levels, contribute to the final
• 49:02tonality that is that we
• 49:03see with the, splice impact
• 49:04mutations.
• 49:05So the the data that
• 49:06I've generated and worked on
• 49:07is using a model, which
• 49:09is an acute
• 49:12inducible model system. So we
• 49:13are looking at transcription changes
• 49:14that happen right after the
• 49:15expression of the mutant protein.
• 49:18What we see in the
• 49:18MDS cells so so these
• 49:20are, cells that have been
• 49:22dividing, proliferating over months, potentially
• 49:24years. And the transcription changes
• 49:26may
• 49:28combine with other effects, like
• 49:29misplicing, and that may result
• 49:31in a final clonal state
• 49:32that it is hard to
• 49:33tease out what is contributing
• 49:34to what. But I can
• 49:35say that, when we profile
• 49:36the c d thirty four
• 49:37cells from the s f
• 49:38three of mutant cells, we
• 49:40looked specifically to see whether
• 49:41we can similarly recapitulate or
• 49:43see those portal density
• 49:45distribution changes and,
• 49:47chromatin accessibility changes. And even
• 49:49though there is a lot
• 49:49of heterogeneity when it comes
• 49:50to patient sample profiling, I
• 49:52think what we found was
• 49:53consistently there was indeed this
• 49:54change of a drop of
• 49:56portal density of the promoter,
• 49:57increased density at the gene
• 49:58body region, and, a a
• 50:00closed chromatin configuration at the
• 50:02the nucleosome. So what we
• 50:03think is maybe transcription is
• 50:05contributing in a large part
• 50:07to what we're seeing, and
• 50:08this is being carried over
• 50:09over multiple generations of,
• 50:11of division. And
• 50:13it also explains one other
• 50:14thing, which is the paradoxical
• 50:15behavior that we see with
• 50:16splice factor mutations. We know
• 50:18that these are they cause
• 50:19total advantage, but also these
• 50:21cells grow much slower. They,
• 50:22in fact, grow very slowly.
• 50:23And the cells that are
• 50:25most
• 50:26susceptible to the DNA damage
• 50:27are fast dividing cells. So
• 50:29cells like k phase two
• 50:30cells. But
• 50:31when we talk about MDS
• 50:32and CD thirty four cells,
• 50:33most of them are quiescent.
• 50:35They are very slowly dividing.
• 50:36And so it may be
• 50:37that the DNA damage that
• 50:39is incurred is relatively small,
• 50:40and the cells are able
• 50:41to escape from it.
• 50:43However, the poll two changes
• 50:44that happen eventually remodel the
• 50:46chromatin landscape, and that contributes
• 50:48to the clonality. And so
• 50:49a large,
• 50:50part of the effort ongoing
• 50:51effort is trying to look
• 50:53how
• 50:54these chromatin changes carry over
• 50:56through multiple generations
• 50:57and to also look at
• 50:58how chromatin is getting altered,
• 50:59not just at the local
• 51:00level, but also how it
• 51:01is changing
• 51:03chromatin compaction and how it
• 51:04is how it stays through
• 51:06the period of disease pathogens.
• 51:09Thank you.
• 51:23Hi. Thanks for a great
• 51:24talk. So early on in
• 51:25the talk, you said about
• 51:27fifty percent of patients,
• 51:29cancers will have these splicing
• 51:30mutations.
• 51:32Do you find that within
• 51:33a given patient,
• 51:34or has it been looked
• 51:35into that, all of the
• 51:37clones typically have the splice
• 51:38mutation? Or Yeah. That's a
• 51:40good question. So in even
• 51:41in those small percentage of
• 51:42patients where we do see
• 51:43a co occurrence, they generally
• 51:45occur in in different clones.
• 51:47In fact, I think it's
• 51:48in the order of less
• 51:49than point five percent that
• 51:50you see a commutation occurrence
• 51:52within the same clone or
• 51:53same cell. So it's very,
• 51:54very small.
• 51:55Thank you. So for a
• 51:57given patient that does have
• 51:58a splice It's typically in
• 51:59different clones. Even if they
• 52:00co occur at the at
• 52:01the bulk level, if you
• 52:02profile and you see that
• 52:03they co occur, it is
• 52:04in different clones. Thanks.
• 52:12We have a couple of
• 52:13questions online.
• 52:15Amar is,
• 52:17attending by Zoom, and he
• 52:18asks, great talk, Prajwal. How
• 52:20does a phase one trial,
• 52:21like,
• 52:23based on this work look
• 52:24like?
• 52:25I think,
• 52:27we do see a lot
• 52:28of encouraging,
• 52:29our date data is encouraging
• 52:30in that, you know, at
• 52:31least when we perform the
• 52:32in vitro corneasys,
• 52:34we do find that the
• 52:35WDFI knockdown is able to
• 52:36kill the mutant cells selectively.
• 52:38Whereas, you know, knocking down
• 52:39into and h type two
• 52:40actually helps them divide better.
• 52:42So we want to explore
• 52:43this further. Obviously, the drug
• 52:45that we use for WDFI
• 52:46inhibition is a drug called
• 52:47OICR nine four two nine.
• 52:49This is the that that
• 52:49is therapeutic in the micromolar
• 52:51concentration, so it's not very
• 52:52potent. And so we are
• 52:54trying other drugs like protac
• 52:55inhibitors
• 52:56to see whether we can
• 52:57actually use something that is
• 52:58more potent and works at
• 52:59the nanomolar
• 53:00range rather than a micromolar
• 53:02range.
• 53:02So that is one thing.
• 53:04The other thing is, I
• 53:05think,
• 53:06we had already started efforts
• 53:07at, you know, using PDX
• 53:08models
• 53:09and mouse models to be
• 53:10able to further validate our
• 53:12findings. And, hopefully, in, you
• 53:13know, the next session when
• 53:14I present, I'll be able
• 53:15to give the data, and
• 53:16then that would give a
• 53:17strong foundation to explore this
• 53:18further in in phase one.
• 53:20Right.
• 53:21Two things. First of all,
• 53:22I wanna say that was
• 53:23just the most wonderful talk
• 53:24and for the fellows here.
• 53:26You know, you were a
• 53:26fellow here. You did the
• 53:27t thirty two. You got
• 53:29funding. You went to the
• 53:29investigative medicine, and it's great
• 53:31to see you doing so
• 53:32well and giving such a
• 53:33wonderful talk. So we at
• 53:34the stage now with MDS
• 53:35where we're gonna be personalizing
• 53:36the therapy, and patients, are
• 53:38they getting fully sequenced? And
• 53:40or do you envision someday
• 53:41talking about phase one that
• 53:42we could have multiple
• 53:44therapies like this, you know,
• 53:46based on their actual Yep.
• 53:48That's the future?
• 53:50Absolutely. So epigenetic modulation has
• 53:52been explored before. I'd just
• 53:54like to cite a few
• 53:55drugs. So one of them
• 53:56is the the vorinostat.
• 53:58So that was so the
• 53:59HSTAC inhibitor, the pan HSTAC
• 54:01inhibitors have been tested, and
• 54:02they have met with,
• 54:03no improvement in outcomes. Now
• 54:05I must mention that we
• 54:06have to take a nuanced
• 54:07approach because the HSTACs, the
• 54:08pan HSTACs inhibit both the
• 54:10HSTAC one and HSTAC two.
• 54:12So these are class one
• 54:13h tags. And so when
• 54:14we when we inhibit both
• 54:15the h four h tag
• 54:16one and h tag two,
• 54:17it actually causes toxicity to
• 54:18the cells. The cells don't
• 54:19do it, whether it be
• 54:21wild type or muted. What
• 54:22we found in our screen
• 54:23is that if we selectively
• 54:24inhibit h tag two, that
• 54:26we're able to rescue the
• 54:27cells and make them grow
• 54:28better.
• 54:29We have been in talks
• 54:30with the,
• 54:32drug development group, and I
• 54:33think one of the major
• 54:34challenges in the field has
• 54:35been able to create a
• 54:37selective HVAC two inhibitors. And
• 54:39the problem with that comes
• 54:40from the fact that they
• 54:40share a lot of homology.
• 54:42And so I think it's
• 54:43maybe HSTAC,
• 54:45two inhibition may not be
• 54:46the way to go, but
• 54:47IN2 inhibition would be something
• 54:48we are excited by because
• 54:49IN2 is a scaffold protein.
• 54:50So it allows the synthetic
• 54:51HSTAC to to assemble
• 54:53on the promat promoter chromatin.
• 54:55And so that is that
• 54:57that is showing promise, and
• 54:58so we probably want to
• 54:59look at that for.
• 55:03Another question. Actually, two questions
• 55:05from Tim Robinson from radiation
• 55:07oncology from Zoom. The first
• 55:09is, given the impact of
• 55:10splicing mutations on single stranded
• 55:12DNA damage, would you expect
• 55:14splicing mutations to sensitize to
• 55:16radiation therapy therapeutically?
• 55:18I would think so. Especially,
• 55:19I I I must say
• 55:20that, ATR inhibitors so single
• 55:22stranded DNA damage activates ATR
• 55:24pathway, the phospho ATR pathway.
• 55:26And ATR inhibitors have been
• 55:28tried, and they are showing
• 55:29some promise.
• 55:30It it could well be
• 55:31that the radiation therapy could
• 55:33synergize with the with,
• 55:35like, ATR inhibitors and further
• 55:37augment the selectively it's synthetic
• 55:39lethality that we envision to
• 55:41see in mutant cells. However,
• 55:43I think we are not
• 55:44there yet. Radiation therapy, you
• 55:45know, it has its own
• 55:46toxicities. You're exposing the cell
• 55:48the
• 55:49the the body to radiation
• 55:51effects, and that has long
• 55:52term effects in itself.
• 55:53So I think we need
• 55:54something which is more selective,
• 55:55something that can more selectively
• 55:57target. And,
• 55:59I think epigenetic modulation is
• 56:01always shown promise, and it
• 56:02has been tested so much.
• 56:03I think we just need
• 56:04to harness it and nuance
• 56:05our approach to be able
• 56:06to better target epigenetics.
• 56:09You Also had another question,
• 56:10I think, which you answered
• 56:11earlier, but, you showed that,
• 56:13mutation splicing mutations were negatively
• 56:15selected in vitro. How do
• 56:17you reconcile with their, positive
• 56:19selection clinically?
• 56:20Yes. I'd like to, I
• 56:22think I probably mentioned this
• 56:26a few, minutes ago, but
• 56:28what we think may be
• 56:29happening is that in the
• 56:31the cells that are growing
• 56:32in the the HSEs that
• 56:34are in the bone marrow,
• 56:35they grow really slowly. They're
• 56:36mostly in the g one
• 56:37phase. They slowly go into
• 56:38the g one phase. The
• 56:39cell cycle,
• 56:40time is very long.
• 56:42So the cells that are
• 56:43most sensitive to DNA damage
• 56:44are fast dividing cells. Those
• 56:45these include cell lines like
• 56:47our k phase two cells
• 56:47where you see a dramatic
• 56:48DNA damage response. But when
• 56:50you talk about quiescent or
• 56:52slowly dividing cells, the DNA
• 56:53damage is not enough to
• 56:54actually kill them. But in
• 56:56fact, probably make them stronger
• 56:57in that they're able to
• 56:59make their way beyond the
• 57:00s phase
• 57:01while also accumulating chromatic landscape
• 57:03changes because of the portal
• 57:05changes. So there is remodeling
• 57:06happening. At the same time,
• 57:08the cells are able to
• 57:08survive through that critical period
• 57:10of DNA damage block.
• 57:14Alright. I don't see any
• 57:15other questions, so we'll end
• 57:17there.