Decoding the Non-coding Genome in Blood Development and Disorders

December 18, 2020

Information

Jian Xu, PhD
Associate Professor, UT Southwestern Medical Center
YCCEH Invited Speaker
December 18, 2020

ID6029

To CiteDCA Citation Guide

00:00Alright everyone, good afternoon,
00:02let's go ahead and get started so
00:06everyone can get on with their day.
00:09It is our great pleasure today to welcome
00:11Doctor Jian Xu as our distinguished speaker
00:14of Hematology for that Yale Cooperative
00:16Center of Excellence in Hematology.
00:19We are in enjoyed very glad to
00:21welcome him not just on behalf of
00:24the Yellow Cooperative Center of
00:26Excellence of Hematology,
00:27but also for all of the Centers of excellence
00:31in hematology across the United States.
00:33As you can see, the title of his talk
00:36is decoding the noncoding genome in
00:39blood development and disorders.
00:40He was trained at UCLA.
00:42He did a postdoc at Boston Children's,
00:45and he's now an associate
00:47professor at you T Southwestern.
00:49His lab studies the molecular
00:50mechanisms that regulate gene
00:52expression and hematopoiesis and cancer,
00:54especially transcriptional enhancers
00:55and epigenetic regulatory regions.
00:56So we're just very excited to have him work.
01:00We're grateful he's allowed
01:01us to record his talk today.
01:03I'm at the end of his talk.
01:05We ask that if you have questions,
01:07you put your questions in the
01:09Q&A or in the chat.
01:10I think I will have difficulty unmuting you,
01:12so if you type your question I will
01:14read it and then he can answer it.
01:16So again,
01:17thank you so much and we were quite
01:19excited about your talk today.
01:21Thank you so much Jeannie for that
01:25really nice introduction and I want
01:27to sort of start up before sinking
01:29Pat Gallagher Anadan across for the
01:32opportunity and a kind invitation to a
01:35virtually visitor cooperative center
01:37of excellence of hematology at year and,
01:39and I'm excited to share some more
01:42recent work related to the study of
01:45noncoding genome in the context of
01:47blood development and disorders.
01:49So just a brief introduction of my lab,
01:53so we started the mechanism that
01:56regulate gene expression during
01:58blood cell differentiation,
01:59and how deregulation these process on the
02:03lines of development of blood disorders,
02:06and more specifically we aim to understand
02:09how we need specific transcription factors.
02:12Epigenetic regulators cooperate with
02:14environmental signals to control
02:17cell identity by acting on a set
02:19of non coding regulatory elements
02:21such as transcriptional enhancers,
02:23which can be the central focus
02:26of my talk today.
02:29So as we all know that a noncoding
02:32genome occupies nearly 99% of the
02:35human genomics space and consistent
02:38various regulatory elements as well
02:40as many of the pulley characters.
02:43Repetitive element an other genomic DNA is,
02:46on the other hand,
02:47from genetic studies we know vast
02:50majority of disease associated genetic
02:52variants are located in a noncoding genome,
02:56but identify the causal mechanism.
02:59Has remained a significant
03:00challenge for the field.
03:02An an.
03:03In select cases we know our careful
03:06dissection of the underlying pathways
03:08can often lead to new insights in Human
03:12Genetics and an even therapeutics,
03:15for example,
03:16by associating common genetic
03:18variations with certain block
03:20trades such as feeding him a globin
03:23expression levels a gene called BCL
03:2611 A was identified more than 10
03:28years ago. OK, I did as opposed to
03:31Federal in Boston Children's with walking
03:34and together with Vijay Sankar Anadem.
03:37Bawa have shown that this genetic
03:40variants that were identified forms
03:43you are status actually do not affect
03:45the coding sequences of BCL 11 A.
03:48But instead that a tissue specific
03:51enhancer for this transcription factor,
03:53so normally as nicely illustrated
03:55by Los Harrison Global Bill,
03:57that PC anime is activated in.
04:00Definitive hematopoietic cells
04:01by this tissue,
04:03especially enhancer to request fitting
04:06hemoglobin expression and impatient when
04:08the adult beta globin gene is mutated,
04:11such as in the context of sickle cell anemia,
04:15and this is permissive for
04:18developping as sickle cell disease.
04:20The GEOHASH variant basically
04:23functions to attenuate enhance
04:25activity leads to less of a PCR.
04:28My expression an and reactivation
04:30of feeling hemoglobin.
04:32Question 2A milli.
04:33Ameliorated disease symptoms.
04:35So this is how the genetics works,
04:38but this fine is not only
04:41established the underlying genetic
04:43basis of hemoglobin switching,
04:45but also raise the possibility that
04:48teach you specific enhancer elements.
04:51Maybe potential therapeutic targets
04:53and this idea actually has led
04:56to an ongoing clinical trials by
04:59Christmas Therapeutics in partnership.
05:01Vertex Pharmaceuticals.
05:02To target PCR 11 Enhancer in sickle
05:05cell disease and beta thalassemia
05:08patients using CAS 9 genome editing.
05:11So when you walk since first
05:14to mobilize our city.
05:16Sorry for positive hematopoietic
05:18stem progenitor cells from affecting
05:20individuals and transducer cells with
05:23rubber nuclear proteins containing
05:25cost 9 and a single godani that
05:27target the PCL away enhancer,
05:29which leads to disruption of
05:31the enhancer function,
05:33and then these edited cells.
05:35We are well reinfused back to the same
05:39patient and observe for disease phenotypes.
05:43As you may know,
05:44the result of the first 2 patient one
05:47sickle cell patient in one Peter Self well,
05:51recently reported in a paper at
05:53publishing Union General Medicine
05:55with more cases recently reported in
05:57the Ash American Society Hematology
05:59Annual Meeting just a couple weeks ago.
06:02So as one example showing here that
06:05here's the result for the first bit of
06:08cell patients that are treated by CAS 9,
06:11editing of the PCL 11 Enhancer.
06:14As you can see on the left graph
06:18before infusion of the edited cells,
06:21the patient had a hemoglobin level
06:23of nine .0 gram per deciliter with
06:27feeding him global level only .3 grams
06:30per deciliter but just a couple months
06:34after infusion of the edited cells,
06:36the feeling hemoglobin level
06:38increased to 6.5 gram per deciliter
06:41an letter 1113.1 gram per deciliter.
06:44With total hemoglobin increased to
06:4714.1 and this is accompanied by nearly
06:51100% F cells or these feeding him
06:54globin positive cells and up to 18
06:57months of follow up so this patient
07:00is clinically cured technically,
07:03therefore this proof of principle
07:06studies are very encouraging not only
07:09to show that therapeutic targeting
07:12disease associated enhancer elements can.
07:14Likely provide a cure for the most
07:17common monogenic disease that was first
07:20described more than a century ago,
07:23but also open up new ideas
07:25and opportunities for people
07:27to understand biology and more important,
07:30it develop approaches to target this disease.
07:33Associated noncoding regulatory elements.
07:36So. First question we ask when
07:39when I started my lap at you T
07:43Southwestern is what are enhancers?
07:45How enhancer was initially discovered?
07:47So Enhancer was first
07:48discovered nearly 40 years ago,
07:50an actually from a viral genome as a shot.
07:54Deer sequences in the SV 40 Gino
07:56that can enhance the expression of
07:59a rabbit bitter globin gene in the
08:01orientation independent manner.
08:03In a transient transfection
08:05assays soon after the 1st century
08:07enhancer was discovered.
08:09From the mouth immuno globin gene.
08:12In a tissue specific manner so therefore
08:15even from the very early days enhancer
08:18at defined as SESAC ingenious sequences
08:22that function at often at a distance to
08:25activate gene transcription in orientation
08:28independent and tissue specific manner.
08:30As such, it has been very
08:33difficult to studying answers,
08:34and an still remains very tickled.
08:37To identify enhancer target
08:38genes and their in vivo function,
08:41largely due to the lack
08:43of experimental tools.
08:44So therefore there are several
08:47important questions enhancer biology
08:48that remain to be addressed Anan.
08:51This is some of the questions that we are
08:53very particularly interesting excited about,
08:56and I would like to share some recent
08:59work that we have done trying to address
09:03some of these important questions.
09:05He has a biology and particularly
09:08how to identify enhances and housing,
09:10has a regular their target gene
09:13expression and how it has it themselves.
09:16Regulated and finally,
09:17how do he has alterations contribute
09:20to diseases,
09:21typically in the context of
09:25hematopoietic malignancy?
09:26So the first question is how
09:28to identify Hazard.
09:30This has been challenging,
09:31but with the advances in next Gen
09:34sequencing technologies coupled with an
09:36analysis of variety of chromatin features,
09:39we can now easily annotate these
09:42noncoding genome using a number of methods
09:45such as chip sequencing as you know,
09:48to examine protein,
09:49DNA interactions,
09:50air taxi or dinner sequencing to
09:52examine chromatin Accessibility as
09:54a surrogate for open property, or.
09:57Active transcription and more recently
09:59chromatin confirmation capture
10:01or three CBEST methods to examine
10:04high order chromatin structures.
10:06So now we can.
10:08Fairly easily to identify putative
10:10enhances other regulatory elements in
10:12a systematic manner and using these
10:15tools are we can use a simple combination.
10:19We enhance associated histone Marks
10:21and chromatin Accessibility to identify
10:23putative active enhancers or other
10:25regulatory elements across the Gina.
10:27So using our favorite local that human
10:30beta globin gene cluster as example.
10:33As you can see from this genome browser,
10:36tracks that we can easily identify
10:39the upstream enhancer.
10:40Pass this on Locus control region
10:42or else are by the presence of
10:45dinners will have sensitivity.
10:48He can 27 simulation and
10:50HCC for modern isolation.
10:51On the other side you can also
10:54identify active promoters by
10:56the presence of translation.
10:58At 3K four there's all done in primary
11:01humour as well progenitor cells.
11:04So using this approach,
11:06we and others have previously compared
11:08enhancer landscapes between human primary,
11:11hematopoietic stem progenitor cells
11:13and differentiated industry sales,
11:15and we notice that initially that
11:18enhances undergoes a pretty progressive
11:20turnover during limits differentiation,
11:23such that about 2/3 of the enhancer
11:26that were found in undifferentiated
11:28City City for positive cell or
11:31loss and replaced by a similar
11:34number of Linux specific enhancers.
11:37In just a few cell divisions,
11:40we then identified transcription
11:42factors and their combinations that
11:44are required for this image and a
11:47developmental stage specific enhance
11:49activities and I should say that similar
11:51results obtained by elegant studies
11:53for pack animals group here at the Yell
11:57and Ross Hardison, and many others,
12:00in different model systems.
12:01And this really has provided a
12:04useful resource for the community
12:06to understand the regulatory.
12:08Basketball hematopoietic
12:09cell differentiation.
12:10So with this increasing availability
12:13of this large genomic data set,
12:15so we wonder whether we could use
12:18this information to understand
12:19how it has a regular,
12:21their target genes blushing.
12:23Anne,
12:24this let us to focus a set of these
12:27linear specific enhancer clusters or
12:29super enhancers as some people call it,
12:32including one showing here that's located
12:35upstream of the gene call SLC 25 S 37,
12:38which is a includes iron transporter
12:41critical for hematopoietic cells.
12:43So as you can see,
12:45these tissue specific enhancer that
12:47are conserved between mouse and human
12:49are contains 3 individual enhancers
12:51that are virtually indistinguishable
12:53in terms of histone Marks and
12:55transcription factor binding.
12:56In this case got one.
12:58Inhuman amounts are extra cells so
13:01therefore the critical question
13:03we had initially was how do these
13:06individual constitute enhances
13:07contribute to the function of
13:09this super enhancer as a whole.
13:11So we decided to use CRISPR
13:13knockout to knockout each individual
13:15enhances and their combinations.
13:17We found that quite surprisingly that
13:20this thing has a seems to be organized
13:24as a functional hierarchy such that
13:27not card enhances 3 E 3 or the most.
13:29This thing has us completely
13:31abolished in has activity,
13:33whereas knockout the other two
13:35neighboring hazards had little effect.
13:37And this was also seen by others in
13:40different hazard clusters and cell types.
13:42So therefore the key question is what
13:45is unique about this enhanced E3?
13:47In fact they are hard to distinguish
13:49based on the chromatin features
13:51and how to identify.
13:53This seems to be functionally more
13:55important or or predominant enhancers in
13:57super enhancer function and more importantly.
14:00This list another important question
14:02that we've been studying is how
14:04enhancer themselves are regulated
14:06or organized in native quality.
14:08In, in, in, in prime itself.
14:11So you attempt to address this question.
14:14We thought that we will ideally be able
14:18to isolating hazard from this native
14:21quality environment and bisector.
14:23Regularly composed compositions.
14:24So we ended up engineering are
14:27inside your capture assay by
14:30leveraging CRISPR CAS 9 technology.
14:32So briefly many of new by combining a
14:36gene specific garden sequence with a
14:39nucleus in activity cost 9 or D CAS 9.
14:43The crisper cast 9 Rd CAS 9 can
14:45be targeted
14:46to the proximity of of any answer
14:49that you might be interested
14:51or other regulatory elements.
14:54We then met further
14:55modifications of the system,
14:57including adding an epitope tag
14:59that can be enabled by terminated,
15:02then using high affinity stood
15:04averaging biotin based application.
15:06We can isolate the enhancer complex and
15:09all associated molecular interactions
15:10and determine has associated proteins.
15:13Amaze and DNA complexes by proteomics
15:17and next generation sequences.
15:20So as a proof of principle this approach
15:23we started with human telomeres,
15:25which contains many copies of repetitive's
15:29at telemetric repeat sequence that
15:31can be targeted by a single God on it.
15:34Upon inside your capture,
15:36we found out that email can be
15:39labeled by image Ng if you fuse,
15:41because now with the GOP molecule
15:44for example also by Q PCR,
15:46we identified chilling your
15:48DNA are highly enriched,
15:49ANAN more important by Western blot and
15:52subsequent in mass spec based pre omics.
15:55We can identify many of the noise,
15:57artillery,
15:58associated proteins and many of the
16:00unknown factors that she annualizing
16:02to how telomere may be regulated.
16:04This is a highly repetitive sequences.
16:07What about a single copy?
16:09Locals in the human genome?
16:11So we choose to focus on our favorite.
16:14Locals are human beta globin gene
16:16cluster as you know that these
16:18cluster contains a set of five beta
16:20like globin genes that undergoes
16:23developmental switching doing.
16:25Insurance cell differentiation and all
16:27these things are controlled by a shield.
16:29Locals control raging or super
16:31enhancer if you wish and look at
16:34it upstream of this gene clusters.
16:37So using a single godani that
16:39are designed to be specific for
16:42the three Prime HS,
16:43one insulator element we found by
16:46chip sequencing of the DCAS 9 capture
16:49DNA that only three primary chest one
16:52and no other regions and across loci,
16:55is highly enriched.
16:56Similarly,
16:56if you use a gardener targeting HPV,
16:59you will see a single peak antagony
17:02HPP promoter regions we know in
17:05the human genome the gamma gene is.
17:08Duplicate it at pgy and beaches.
17:10SVG two, so therefore a single God,
17:13an actual capture both,
17:14and that's exactly what we're seeing,
17:17and then we can simply by design
17:19different God.
17:20Honest, having different regulatory elements,
17:22enhancers for example,
17:23and we can capture each individual enhances.
17:26Using this card on is,
17:28and more importantly,
17:29if we pull all these God honors
17:31together an echo expressing the same
17:33cell that we can capture all five
17:35enhancers suggestion that this system
17:37can be multiplexed to capture many
17:40enhances at the same time in the same cell.
17:44So based on these findings,
17:46but we went ahead,
17:47developed a capture proteomics
17:48to determine decathlon capture,
17:50look locals are specific podium and
17:52identify in this case again using
17:55beta globin gene cluster as example
17:57that we identify manufacturers
17:58that are social with,
18:00for example HS2 enhance at the
18:03LCR region including some familiar
18:05faces like God.
18:06I want a long beyond the four
18:08etc and but also some other new
18:11factors that we decided to follow
18:14up on. Such as a nuclear pore proteins
18:16and you can see some of these factors
18:19also present at the gamma globin gene
18:22promoters and other factors are present in.
18:25Beta globin gene clusters.
18:28So this analysis provides the initial
18:31evidence for the composition based
18:34organization of the beta globin
18:36gene enhancers and promoters.
18:39We all know that the human genome is
18:42organized into a multilayer structure.
18:44Units are organized by Long
18:46range dinner interactions,
18:47or chromatin looping,
18:48so we never saw that.
18:51Can we combine this DCAS 9 capture
18:53with three C analysis to identify local
18:56specific long range thing interactions?
18:58So the way it works is we first using
19:01in vivo bite Internet DCAS 9 to
19:04isolate enhances that you might be
19:07interested and Cromartie is crosslinked.
19:09Followed by enzyme digestion.
19:11Usually we use frequent Cadillac
19:14DPM two and this is followed
19:17by a proximity ligation of the.
19:19Adjusted genomic DNA Anan followed
19:22by fragmentation and then the
19:24enhancer can be directly isolated
19:26using strip having best purification,
19:29ANAN together with all the DF
19:32fragment that will really get it,
19:34and using pale and sequencing we
19:37can identify all the long range
19:40interactions that are social with
19:42this target enhances and more
19:45importantly by combining capture
19:46as we see analysis with capture.
19:49Baryonyx we hope to be able to
19:51identify causality.
19:52What are the factors that might
19:55be responsible for this Sweetie
19:57Genome organization?
19:58This is just showing one of the
20:00example of the data how this works,
20:03so here we're showing capture 3C
20:05analysis for the active HP Gigi.
20:07This was done in Cape RC2 sales.
20:10Now you will see that is active
20:12Edge BGM contains many long range
20:14interactions with other six element
20:16across the low side but not the
20:19nearby repressed genes.
20:20But first I if you capture the
20:22repressions HPB you will see the
20:24HPV forms of predominant shorter
20:26range interactions and downstream
20:28sweeper matches one insulator.
20:30There's no interaction between active
20:32and repressed genes on when we capture this.
20:35We next capture each individual
20:37in hazard within,
20:39else are an interesting and we
20:41found that H S3 has a contains
20:43many more long range interaction
20:45then other nearby enhancers,
20:47including access to and.
20:49We further validated this using
20:51independent God honest and we
20:53notice that H3 has a consistently
20:55contained many more interactions
20:57than the nearby actions to enhance,
21:00and this was interesting to us because
21:02it's just too has been sought to be
21:06the strongest enhancer within LCR.
21:08By transgenic enhancer Reporter,
21:10access YH Three was also shown
21:12to be important,
21:13but only in the context of native quality.
21:16So therefore our findings may help
21:19to explain these findings and
21:21support a model that the hierarchy
21:23organization with a beta globin LCR
21:25in which that itches to my function
21:28to recruit a trans acting factors and
21:30function more like a conventional enhancer,
21:33yhs 3 might might be more important,
21:36immediate and long ranged in actions of.
21:39Formative,
21:39loopy,
21:39and really,
21:40it's a combination posing hazard.
21:42Am I dictating what the locals control?
21:45Regional Super Enhancer will
21:47function in its native quality.
21:50So as a brief summary,
21:52so we think that this decathlon
21:54capture my provider are useful
21:56tool for multi omic dissection
21:59of this regulatory elements from
22:01their endogenous loci and this
22:03biotin scribbling based affinity
22:05capture provides high specificity
22:07and sensitivity and this SGI based
22:10targeting allows for Multiplex
22:12analysis which are going to show
22:14in a minute and more importantly
22:17the simultaneous analysis of local
22:19specific proteome and 3D interactions.
22:21Help to establish causality and
22:24the system should be brought it
22:26up application applicable to other
22:29genomic locals or model systems.
22:31However,
22:32one of the major limitation of
22:34the original capture Mesa was that
22:36it requires large number of cells
22:39and typically 10s of sometime
22:41hundreds of millions of sales for
22:43local specific proteomic studies,
22:45and thus was not applicable to primary
22:48cell types or male cell populations.
22:50So we wonder whether we could improve
22:53the design to increase capture
22:55efficiency so you know original design.
22:57We fuse the biotin tag to the
23:00N terminus of the kastein.
23:02We just close to the nucleotide
23:05protein interacting pocket that
23:06might cause arepa tomaski,
23:08whereas the C terminus of CAS 9
23:11proteins largely exposed an unstructured.
23:13So we have tried tagging the seat
23:16owners instead for by connection.
23:18Moreover, in our original design,
23:20we use the conventional every tag
23:23together with the bacteria puje biotin
23:25ligase for in vivo bike nation.
23:27So this will make a three vector system.
23:31However there are other.
23:32Peptide sequence that can be bought
23:35in an area using endogenous biting
23:38like is that expressing my man in
23:40sales so we have been testing some
23:42of those sequences and more recently
23:45approximately biting ligands such as APEX.
23:47Two was shown to be able to buy
23:49terminate nearby proteins on comity
23:51which may increase the signal to noise
23:54ratio for identification by proteomics.
23:56So we have been trying and
23:59adapting those systems for capture.
24:01So with this improved capture system
24:04with first attempted Multiplex capture,
24:06many answers in a single experiment,
24:08so we started with the LCI enhancer
24:11with 10 individual gardeners to
24:13capture all 5H S enhances and observe
24:16interactions social with all enhancers
24:18in a single experiment and more
24:20importantly you actually can resolve
24:22this data into a single enhancer resolution.
24:25Look at what other enhancer mediated
24:27by each individual enhances,
24:29as you may recall,
24:31the pattern looks almost the same.
24:33As we did previously for the
24:36individual capture,
24:36you negating that multiplicative
24:38multipliers capture.
24:39We tend the native permitting architecture.
24:42So based on this,
24:44we next performed much less capture
24:47of actually super enhancers.
24:49Using this super enhancer calling algorithm,
24:52and in this experiment we use nearly
24:542000 God on his targeting more
24:57than 150 super enhances containing
25:00more than 800 individual enhances.
25:03A very large panel.
25:04And that data shows that we can
25:07capture the vast majority of that
25:10enhances in a single experiment.
25:13So with these high resolution
25:16multiplus capture,
25:16we notice that some interesting features
25:19about this special configuration of
25:22these super enhancers, specifically,
25:23we can often identify one or few
25:26enhances within a super enhancer
25:29that have unusually higher frequency
25:31of interactions compared to other
25:34nearby enhances we call this.
25:36He has a happy hanses.
25:38We further developed a computational
25:40workflow and a fund at about
25:421/4 of the Super enhancers are
25:44organized as a hierarchy structure
25:46containing hopping answers,
25:48and here just showing a snapshot of
25:511 example of a hierarchy of super
25:53enhancer that contains a hopping
25:55answer that you can see that has
25:58more frequent interactions with
26:00other enhancers and promoters
26:02within the same genomic region,
26:04and what I won't show you the
26:06data we actually went ahead using
26:08crisper cast 9 to knockout.
26:11These hopping hazard.
26:13Versus the nearby non hop enhances and
26:17almost all the time that we can see,
26:21knockout hopping has at least a more
26:24profound effect and target gene
26:26expression comparing to knockout
26:29an unhappy hazards suggestion that
26:31this enhancer might be functionally
26:34more potent enhances that.
26:36Functioning within a super has a
26:39gene cluster. So as a brief summary.
26:42So here we showcase our several
26:44potential implications for the D
26:46cast net capture system for high
26:48resolution Multiplex analysis and
26:50local specific quality interactions.
26:52We hope that the capture system can
26:55enable us to determine a special
26:57configuration enhances and their
26:59target genes,
27:00as well as a temporal regulation
27:03during development.
27:04So with this,
27:05we started to one of the next important
27:08question that that would be interested in is.
27:11How do we have to actually control tagger?
27:15Jinx,
27:16brushing?
27:16Inside to it,
27:17particularly doing in people development.
27:20And this has not been trivial.
27:22Trivial as you can imagine,
27:24you can knockout this individual enhancers,
27:26but offering you're not getting answer.
27:29Look at in a cell grows best answer,
27:32you don't really see a difference
27:34is so we thought that to address
27:36this we will need a tool that
27:38allow us to systematically bisector
27:41functional requirement of this
27:43enhances in their native comity.
27:45Ideally doing in people development
27:47and such as human quality limits,
27:49differentiation so.
27:50We recently redesigned the Cast 9
27:53best so called epigenetic editing
27:56system that can efficiently perturb
27:59enhancer activities by modulating
28:01enhance associated chromatin features.
28:04Particularly careful and K 27 simulation
28:07specifically for enhanced activation,
28:09refused ecaster with P300,
28:11which catalyze H3K27
28:13assimilation within adapted,
28:15helping actimel Ms 2 on the
28:18SG and sequence to recruit.
28:21Another activated VP 64 similarly for
28:24the repression that refused the cast.
28:26Now with LC-1 which catalyze
28:28the removal of H3K,
28:30four born monogamous flashing
28:32and then as geometry,
28:34could another repressor correct,
28:36so therefore the main advantage of
28:39this system which we call increased
28:41by or increase for a or enhancer
28:44targeting increased bio quiz for a is
28:47that this system combines effectors
28:49with distinct mechanism for modulating.
28:51Enhancer associated genetic marks.
28:53So we have used this system in
28:56a variety of cell models and
28:59showing that the
29:00new system is superior than the
29:02original crisper I whisper in
29:05system for enhancing perturbations.
29:06However, the main challenges analysis of
29:08enhancer function doing illegal development.
29:11To address this, we are when I
29:13generate a knock in mouse model.
29:1610 D CAS. Nine care app converging
29:18under the docs inducible promoter
29:20in College in one in one low side.
29:24So we then developed by in Vivo Enhancer
29:27perturbation asset you then there's no
29:29key mouse to determine hazard function.
29:32Doing him at opposes so briefly the
29:34way it works is that we reasoned
29:37that by ABBA genetic modulation
29:38of Venus physic enhancers in HCS
29:42followed by pulmonary transplantation.
29:44We could access the HSE deriving
29:46mature cell images.
29:47As a read out for the functional
29:49impact of he has a population
29:52doing HSC Danish differentiation.
29:54So that way you would do it.
29:57You will isolate him out of politics.
29:59Tampa general sales from this
30:01knocking Mouse and then transduced
30:03with God only library with Garnet
30:05design against each individual
30:07enhancer that you might be interested
30:09followed by bone marrow transplant.
30:11Then D Kastner Kara will be induced
30:14by docs admin administration in
30:16the period of 12 to 16 weeks.
30:19Then you want to isolate a
30:22mature cells cells and perform
30:25amplicon sequencing to determine.
30:28The abundance of SG and a in a starting
30:32population before transplant an the
30:34resulting population after transplantation.
30:37So as a proof of principle this we
30:39focus on several key hematopoietic
30:41transcription factors and their
30:43annotating enhances an and.
30:45Design multiplies got an A plus for
30:47in vivo enhancer population squeeze,
30:49so I would like to show you the CBF
30:52and locals as a first example which
30:55contains out for annotating hazards,
30:57look at it in a different upstream
31:00or downstream regions to the
31:01transcription start site by enhance
31:03in billing has a population.
31:05We found that SGML for CDL 4 plus.
31:08So disable.
31:10Enhancer all year for Enhancer in this.
31:17A significantly depleted in mylow
31:19cells but not in B&T cells.
31:21This is interesting,
31:23as previously has been shown that
31:25CBF on gene knockout or the plus 37
31:29Hanson AKA annoying to be required
31:31by more minor cell differentiation
31:34but not for lymphopoiesis,
31:35so our result not only validate
31:38these findings,
31:39but also show that the plus a
31:42keeping hazard is E2 enhancer
31:44showing over here are also important.
31:47For modeling differentiation,
31:48but the other two enhancer seems
31:50to be indispensable,
31:51and none of these enhancer seems to be
31:55important for B&T cell development.
31:57The second example is a SPI one
32:00locus or PU .1 G,
32:02and we observed this gene has a single
32:05enhancer located at 14 KP option.
32:07With this gene we notice that the
32:10three independent guard on it
32:12against its enhancer significantly
32:13depleted in my loiselle and B cells,
32:16but not in T cells, and this result,
32:19again are consistent with the role of SPI.
32:22One Pu .1 for normal Milo and
32:25busier development,
32:25but not for teacher development, so again.
32:28I did that this initial genetic studies.
32:32Using mouse models.
32:34And next time I want to show is that
32:37the wrong Swan Locust so runs well?
32:40Actually have two annotated
32:41transcription start site with three
32:43enhancers at different genomic regions.
32:45Again using this perturbation we find that
32:47none of the enhancer seems to be important.
32:50Actually, in myeloid B&T cells,
32:51but the promoter garden is
32:53somewhat you reached in my low
32:55self but depleted and B&T cells,
32:57so this is one of the first example.
33:00We actually see the opposite
33:01phenotype in different cell images
33:03and this is interesting because
33:05ranks will knock out your mouth.
33:07Has shown to to develop Mylar
33:10preffective phenotype characterized
33:12by myeloid enhanced my lawyer.
33:15Refreshing but defective B&T cell
33:18maturation so therefore our results
33:21also consistent with this and
33:24recapture the phenotype of ranks.
33:26One deficiency showcasing that this
33:29crisp Ohio increased by best epigenetic
33:33editing and can be convenient assay too.
33:37Isaca regulatory elements that
33:39are required for this linear
33:41specific transcription factors.
33:43And finally,
33:44we perform a Multiplex perturbation
33:46by pulling all the God are
33:49made in a single experiment.
33:51Anan by this analysis.
33:52Now you can have a ranking order of
33:55different enhancers and promoters that
33:58are specifically enriched or depleted.
34:00In my lawyer B&T cells,
34:02for example with Dan Ified that
34:04CB back on cartoon has a required
34:07for myeloid differentiation,
34:09while ranks one enhancer promoters
34:11are required for PMT sales.
34:13Therefore, this enhanced CRISPR editing.
34:15I could provide a useful tool for
34:18functional interrogation of SIS
34:20element doing illegal development
34:22and more importantly by combining
34:24these knocking mouse model with
34:26other disease models we might be
34:29able to study enhancer function
34:31doing different biological process
34:33of pathological process and this is
34:36something that we're very excited about
34:38and certainly looking forward to any.
34:41Collaborations whom I simply
34:43store might be useful for their
34:45respective disease models.
34:47So as a brief summary for for this part,
34:51so we've shown that enhance the
34:54control limits, differentiation,
34:56disease,
34:56phenotype and undergoes profound
34:58turnover during development.
34:59We've developed the Constine best capture
35:02tools for multimeric analysis of local,
35:05specific quality interactions.
35:06We also redesign enhancing target,
35:08increase pain response system that
35:10will enable in vivo functional
35:12interrogation of enhancer.
35:14As we have other success element doing in.
35:18We want development and we're
35:19happy to share the tools or the
35:22construct has been deposited action.
35:23And if you are any of your
35:26colleagues are interested,
35:27feel free to reach out.
35:30So in the last few minutes I want
35:32to switch gears a little bit and I
35:35want to discuss some of the recent
35:38effort we trying to address the last
35:41question that is how to pathological
35:43enhance alterations contribute to diseases,
35:45particularly the development block answers.
35:48As we know,
35:49much of our knowledge on cancer driver
35:52mutations is based on alterations
35:54of protein coding sequences and
35:57little is knowing whether man,
35:59how noncoding alterations may contribute
36:02to disease on passive biology,
36:04especially in the developer
36:06hematopoietic malignancies.
36:07So we started this rather
36:09ambitious project several years ago
36:12with the goal to identify leukemia,
36:15associating handsome mutations
36:16by targeted sequencing.
36:18So the way it works is that we first
36:21annotated plot Vinny specific enhancers
36:24based on chip sequencing and a taxi,
36:27just like everybody else is doing that,
36:30I introduced earlier and then we can gather.
36:34This game is almost twenty 2000s
36:36of enhancers that are present
36:39in variety of different normal.
36:42Gmail cell lines that we have gathered
36:44and then we can design target sequencing
36:48panel to specifically resequence
36:50the enhancer sequences in a panel
36:53of human hematopoietic malignancy
36:55is a particularly in an email.
36:58An NDS conditions.
37:00We also included some informal samples
37:03an acute lymphoblastic leukemia and,
37:05importantly, thirty. One of these samples.
37:08We actually have tumor normal
37:10pairs that we can.
37:12Identify somatic mutations in
37:14the in the non coding Gina.
37:17Pen and if you do this and like other
37:20people doing a protein coding sequence
37:22as well you you can easily identify
37:26thousands of recurrently mutated mutations.
37:29In this case we identify almost
37:31slightly over a 4000 frequently
37:34mutated noncoding elements.
37:36We call them mutational Hotspot Ann and
37:39these overlays about 1800 enhances that
37:41we have identified from the initial steps.
37:45So these are the enhancement
37:47carries somatic mutations.
37:48That are frequently mutated in
37:50human hematopoietic malignancy.
37:52The key question is and the key challenges.
37:55How do you know whether they are
37:57functional an what is how to access their
38:01functional roles in cancer pathogenesis?
38:03So Fortunately with the crisper a quiz
38:05by system that we have engineered that
38:08especially for enhances so we could
38:10perform a functional interrogation of
38:13functional perturbation screens using God.
38:15RNAs that are designed to
38:17target this enhances.
38:19Then we perform the screening and
38:21multiple leukemia cell lines.
38:23And by this we can identify hundreds of
38:26enhances that seems to be a putative tumor.
38:29Suppressive or uncle genic.
38:31In other ways,
38:32we use cell growth as a reader so that
38:35that perturbation of this enhances often
38:38can lead to an enhanced or inhibited
38:41cell growth phenotype and this really
38:44have provided a number of candidate
38:47enhances an associated genetic loci.
38:49Subsequence studies.
38:49So I would like to focus on one
38:53of the enhancer that we follow up
38:56with more detailed analysis and
38:58this enhances located about 150
39:00KB upstream of the gene called
39:03carrots that contains several non
39:05coding variants in AML samples,
39:06and by chromatin profiling and 3D
39:09chromatin confirmation capture.
39:10We found this thing has it seems to
39:13be physically interact with Akira's
39:15promoter regions which is located about
39:17150 KB downstream of this enhancer.
39:20And more importantly,
39:21when we use CRISPR CAS 9 to knockout this,
39:25he has a in a leukemia cell line
39:27and we found that Carros expression
39:30was significantly down regulated.
39:32But none of these other genes within
39:34the same neighborhood affected
39:36suggestion at this.
39:37Enhancer is selectively required
39:40for the expression of casting.
39:42So is this was interesting because
39:45unlike any grass care US,
39:48protein coding mutations are rarely
39:50found in human animal patients.
39:53However, high care,
39:54high level of care as expression in
39:57AML is associated with poor survival
40:00using data from the TCG cohorts.
40:03So we thought that Miss May identify
40:07potential functional enhancer that
40:09plays a role in a male biology. So too.
40:13Establish the functionality of whether
40:15not is kerosene has is important for AML.
40:18We generate Caroline Hanson
40:20AKA AML cell line.
40:22This is Daniels is selling called Mkpo
40:25one and observe the significant less
40:28tumor burden in bone marrow cells and
40:31blood of the Xeno graft recipient.
40:33Using two independent enhancer
40:35single Cell developed in Hazen
40:38Okok looms as a control,
40:40we also knockout the carrots
40:42protein coding gene.
40:43And we see that he has a knockout
40:47almost recalculate the care as GM
40:49lockout in this access and just
40:52showing more data showing that this
40:55is reflected by Les Plus constant
40:57prefer block as well as less.
41:00And number of frequency of the
41:03Premier sales in xenograft animals
41:05in the bone marrow and blood.
41:08So this result demonstrator care us
41:10and this semantic enhance associated
41:12enhancer are required for AML cell
41:15grows in virtual and an illegal.
41:17So then we went ahead.
41:19Get the motif analysis and found
41:22this recurrent carers enhancer that
41:24we found the AML patient highly
41:26colocalized with binding site
41:28of interest in nuclear hormone
41:30receptors in particularly par,
41:32gamma and ice are showing over here.
41:35I'm so interested in the cameras
41:37in has are also found in has
41:40mutations are also found in other
41:43cancer types based on TCG and pain.
41:46Cancer pan cancer data set here showing
41:48this is a leukemia reputation that we found,
41:52but hole in has actually carries
41:54many recovery mutations.
41:55Many of these mutations are also
41:58overlap with predicted nuclear
41:59hormone binding sites,
42:01and this was interesting as nuclear
42:03hormones are family of lichen
42:05regulated transcription.
42:06Factors that are activated by hormones,
42:09ligands or growth factors and
42:12usually in the absence of ligands,
42:15PPL comma AXA dimer will recruit
42:18Corey Presser complex immediate
42:20derepression appan liggen binding.
42:22This time are actually cutco
42:25activist coactivators through
42:27activating transcription.
42:28So we wonder whether cameras enhance
42:31it might actually be regulated
42:33by nuclear hormone signaling,
42:35so to validate days we perform chip
42:39sequencing analysis in multiple AML
42:41cell line as well as non email.
42:44Tumor cell lines.
42:45We found a strong PPR gamma and I
42:49saw binding at the very associated
42:52enhancer element and if you zoom in
42:54you can see some of the speaker really.
42:57Directly overlapping the variant that
43:00that was found in the AML samples.
43:04To directly test whether this email
43:06associated non coding variants
43:08indeed modulate nuclear hormone
43:10binding and enhancer function,
43:12we went ahead,
43:13generate a knocking audio and this
43:16was not trivial.
43:18That efficiency is still not very high,
43:21but anyway we achieved by knocking
43:24by Christmas targeting of the either
43:26the white type or the mutant value in
43:29two K562 leukemia cells within screen.
43:32Single cell derived knocking
43:34clones and measure nuclear Homer
43:37by name by chip experiment.
43:39So the data showing on the top right
43:42corner so you will see in the input
43:45dinner you will see this expecting
43:47one to one ratio of the whiter
43:50versus knocking earlier because we
43:52generate the heroes actors knocking
43:54values in in this cell lines.
43:56However, if you look at abundance
43:58of this of the ratio of the
44:01knocking mutually versus Vytorin,
44:02the chipped in and you see the knocking
44:05mutant value are significantly enriched,
44:07and this suggests that the mutant value.
44:10Actually have stronger Association
44:12with the nuclear hormone receptors
44:14by chip experiment and consistent
44:16with this and nuclear agonist,
44:18I can list more enhanced CARROS expression
44:21in a knock in sales compared to the
44:24wild type cells against suggestion that
44:27these mutant assume associated cameras
44:30enhancer are regulated by nuclear
44:32hormone receptors and this recurrent
44:34mutations my function to enhance
44:36nuclear hormone receptor binding.
44:38To trans activate Cara sticks.
44:41Watching you email sales.
44:43So we also valid that is finding other
44:46enhances that I won't have time to show
44:49an including a interesting Lee has a
44:52controlling our security engine code,
44:53PO2 this is all publisher
44:55if you're interested.
44:56You're more than happy to
44:58read about the details so.
45:00Moreover, we are in a global analysis.
45:03We found his nuclear hormone,
45:05our finest artist seems to be frequent
45:08targets of noncoding mutations,
45:10in particular email,
45:11but also other hematological malignancy,
45:13suggestion,
45:13perhaps a more generalizable mechanisms.
45:16So therefore we seek our findings
45:18support model that is pathogenic and non
45:21coding variants are noncoding mutations,
45:23might cooperate with signal
45:25independent transcriptional machinery.
45:26In this particular case nuclear
45:28receptors to rewire signal dependent
45:30gene expression programs.
45:31Then my potentially promote are
45:34functionally that contribute to the
45:37development of hematopoetic malignancies.
45:40So as a final summary.
45:43Is that an explosion of genomic
45:46and epigenomic information?
45:48In recent years we have learned a
45:50great deal of how gene regulation
45:53controls normal development and
45:55how dysregulation of this process
45:58contribute to human diseases.
46:00How what.
46:01We currently know only represents
46:03a very small portion of the
46:06complicated complex.
46:07The human genome,
46:09and in retrospect and maybe
46:11more relevant to our studies.
46:13As we know,
46:15the first documented cases of
46:17sickle cell disease was described
46:19by an James Herrick and in 90.
46:22Early 90 app.
46:24Centuries and 9010 specifically,
46:27which was later dropped as a first
46:30molecular disease by a Linus Pauling.
46:33In 1947 an enhancer was not
46:35discovered until early 1980s
46:37and followed by the completion.
46:39The first draft of human
46:41genome in early 2000s.
46:43Now more than a century after
46:46discovering sickle cell disease and
46:4840 years of discovery of enhances,
46:51we might have a first enhancer targeting
46:54therapist for this molecular disease.
46:56Very soon,
46:57and we certainly hope that by
46:59focusing on your hands as another
47:02non coding regulatory at genomic
47:05elements that we might be able to
47:08identify new mechanism and genetic
47:10pathways that contribute to a normal
47:13blood cell development an animal.
47:15Long term that we might be
47:17able to develop our
47:18enhanced targeting therapeutics
47:20for blood disorders.
47:22So with that most important,
47:24I want to acknowledge all
47:26the people who met his work,
47:28possible the initial study PCL
47:30PCL Evan Hansen was done in
47:33Stockings Lab in collaboration
47:34with Stambaugh and Vicious Ankara,
47:37and profiling work done in primary
47:39hematopoietic cells was done in
47:41collaboration with the formal post or
47:43fellow John Huang from Stalking Slab.
47:46Now he has his own lab in Sherman University.
47:50The development of the cast net capture
47:53was spearheaded by a former poster in
47:56my lap as she knew now has his own
47:59app together with another poster fellow.
48:02Again using the developer of
48:04Enhancer targeting CRISPR,
48:05Aquifer was spearheaded by another
48:07poster fellow Kyle only think is
48:10one of the fellows over here.
48:12If I can move my brows anyway
48:14together with another Postal federal,
48:17most of them are transitioned
48:19to their independent positions.
48:20Together with other collaborators
48:21and we couldn't do this with a
48:24wonderful collaboration from Boston
48:25Children's and our local collaborators
48:27at you T Southwestern UT Dallas.
48:28I will stop here.
48:29I'm happy to take any questions
48:31that you might have.
48:33Thank you very much for your time.