When Watson and Crick Get linked: DNA Interstrand Crosslink Repair and Human Disease

March 08, 2023

Yale Cancer Center Grand Rounds | March 7, 2023

Presentation from: Dr. Agata Smogorzewska

Information

ID: 9615
To Cite: DCA Citation Guide

Download Transcript

00:00The Cancer Center grand rounds.
00:03I'm Barbara Burtness,
00:05and I'm really extraordinarily
00:07thrilled to be introducing agato
00:10smuggler Jessica as our speaker today.
00:13She is an associate professor of genome
00:16maintenance at the Rockefeller University.
00:18She received her PhD from Rockefeller and
00:21her MD from Wild Cornell Medical College.
00:25Following a residency in
00:27clinical pathology at MGH,
00:29she joined Harvard Medical School as a
00:31postdoctoral fellow in Stephen Elledge's lab.
00:34She's the recipient of numerous awards,
00:36including the Irma T Herschel Research Award,
00:38the Rita Allen Foundation
00:40Scholars Program Grant,
00:41the Burroughs welcome Fund Career Award,
00:43AV Foundation Translational Research Award,
00:46and the Herald Weintraub graduate student.
00:49Her lab focuses on DNA repair mechanisms,
00:52particularly those involved with
00:55interstrand crosslink repair pathways,
00:57and she's interested in identifying
00:59new genes that are involved in this
01:01repair process in order to reveal the
01:03mechanisms at play in human diseases
01:05that result from deficiencies in
01:07interest strand cross link repair.
01:09This includes Fanconi anemia and
01:12kerio megalith interstitial nephritis.
01:14To this end,
01:15she successfully identified SLX 4,
01:18red 51 and U BE2T is interstrand
01:21crosslink repair factors,
01:23as well as developed a mouse model of
01:25Carrie Magalog interstitial nephritis.
01:27And recently she's begun studies
01:29revolving around how replication
01:31stress is addressed by cells
01:33to promote genome stability.
01:34She also has a deep interest in
01:36those cancers which arise in
01:38patients who have DNA repair.
01:40Defects.
01:40And it's because of this that I've
01:42had the extraordinary pleasure
01:43of beginning to work with her
01:45on the stand up to cancer Grant,
01:46which looks at Fanconi anemia,
01:48associated head neck cancer.
01:51So.
01:51Like I said,
01:53if you wanna come up.
01:55This plaque is to thank you for for
01:57coming to give grand rounds in person,
01:59which many people have not
02:01been willing to do recently.
02:05And we're really looking
02:06forward to your talk.
02:07Thank you so much. Yes, thank you.
02:10Well, thank you Barbara for first
02:12of all for inviting me all of you
02:15for coming to hear the lecture and
02:17all of you on zoom for joining in.
02:20And it's really my pleasure to to
02:23be here and it's been fantastic to
02:26work with Barbara and and learning
02:29from her more about the the clinical
02:32aspects of what we do and hopefully
02:35bringing some new therapies to
02:37these patients eventually, OK, so.
02:41I will talk about DNA interstrand crosslinks
02:44and I'll just remind you though that oh,
02:48these are my disclosures.
02:50I'll just remind you that DNA repair.
02:55It's really essential for a lot of
02:57aspects of of life and depending on what
03:01kind of flavor of DNA damage you have,
03:04whether it's from the outside,
03:07from the UV light creating these dimers,
03:10or from within the cell,
03:11something that I'm quite interested in,
03:14for example, creating mismatches.
03:17There is a repair pathway that's dedicated
03:21to these and there are a number of genetic.
03:24Diseases associated with inability
03:27to repair these different lesions
03:30and they come again in many different
03:34phenotypes resulting in developmental
03:37abnormalities and the generate the
03:40generative diseases and a lot of
03:43them obviously are associated with
03:46cancer prone due to mutagenic nature
03:49of these lesions if the lesions
03:52are not properly repaired.
03:54So my favorite lesion is this interest
03:57in Cross link and I'll give you a little
04:00bit of introduction about these lesions.
04:03This is a covalent linkage
04:04of two strands of the DNA.
04:07It's repaired by the Franco anemia
04:10DNA repair pathway and I'll give you
04:12more of a of the mechanism of how this
04:16is done in in a bit in later slide,
04:18but the whole purpose of this pathway
04:21which is activated during DNA replication.
04:25Is to create 2 pristine.
04:27Double stranded DNA that can be then
04:31moved to dollars trends during during
04:34the cell division and I'll just
04:37mention I won't talk much about it
04:40but this pathway is also activated
04:42at difficult to replicate regions.
04:45So repetitive DNA are loops and there
04:48is growing evidence that at any time
04:52there is stalling of replication forks
04:54the pathway can come to to rescue.
04:57Those forks so that this this intern
05:02crossing is really our model of
05:05of how pathways activated.
05:07And the outcomes of abnormality in
05:10pathway activation or function are
05:13these genomic instability that you
05:15can heat see here in this metaphase
05:17spread where you see number of
05:20these abnormal radial chromosomes.
05:22So we see that in cells from Franconia
05:26anemia patients when they are treated
05:30with external internal crosslink causing.
05:35Chemicals like mydomain sincere splatsin you
05:38see number of abnormalities including gaps,
05:41breaks and these radio chromosomes
05:43and the radial chromosomes occur
05:46because of two double strand breaks
05:48that are that find each other and are
05:52joined together forming these radios.
05:54And obviously these radios come radio
05:57chromosomes can cause instability but
05:59can also cause death of cells when they
06:03are trying to the cell tries to divide.
06:07So. The reason why we know that this
06:09pathway is particularly important in
06:11stem cells are all these phenotypes that
06:14we see in patients with Fanconi anemia.
06:17There are many developmental
06:20phenotypes including these skeletal
06:22abnormalities that you see here,
06:25but pretty much any,
06:26any system can be affected and that's seen in
06:31patients who are truly null for this pathway.
06:35Majority of patients will present with Pence,
06:37Cytopenia.
06:38Around age, median age of 10,
06:41and even though it's called anemia,
06:43all of the.
06:46All of the hematopoiesis can be
06:49affected and and actually platelets
06:52are usually the first to to draw.
06:55And then there is very this very
06:57interesting phenotype that will spend
06:59a lot of time talking about the cancer
07:02predisposition MSDS and AML occurs in
07:05the setting of pancytopenia and squamous
07:08cell carcinoma occurs later in life
07:12but still in in at young age of average 31.
07:16We have very young patients as
07:18I'll I'll show you in in future
07:21slides and then depending on the
07:23actual path of gene that's mutated.
07:25We can also have breast cancer,
07:27medulloblastoma and and
07:29other embryonal tumors.
07:31And there are other phenotypes that are
07:34still fairly poorly understood like
07:36infertility and endocrine abnormalities.
07:38So this is a great model in my
07:42mind for studying DNA repair.
07:46So today, I'll tell you about the first part,
07:49very short part about the mechanism
07:51of interstrand crosslink repair.
07:53Secondly,
07:53I'll talk about the identification
07:55of endogenous sources,
07:57sources of interstrand crosslink.
07:59And today I'll talk about the
08:02bone marrow side.
08:04But we are very much interested in
08:06those in the keratinocytes as well.
08:08And this is an ongoing work in my lab.
08:11And then I'll talk about the mechanism
08:14of tumorigenesis and Fanconi anemia.
08:16Especially in the squamous cell carcinoma.
08:19So just to give you a flavor of how
08:23this pathway actually functions,
08:27this is a very regulated process of
08:30repair of these of these lesions.
08:33So we start with the ACL that you
08:36that I've shown you before and this
08:39is the structure that we know forms
08:42when the replication forks stall
08:44at the lesion and some of this work
08:46has been done in in human cells,
08:48some of it.
08:49This work was done in Johannes Walters
08:51lab in the Xenopus egg extract system.
08:53So biochemically,
08:55it's a fairly understood pathway.
08:58What's essential in this pathway,
08:59and I'll show you the proteins
09:01that are involved in a moment,
09:02but what has to happen is that
09:05the the DNA needs to be broken,
09:07but in a very regulated way that then
09:10gets repaired through translesion
09:12synthesis and then through homologous
09:15recombination that requires
09:17BRC 2 and other proteins.
09:19And this eventually leads to this full
09:23DNA repaired repaired DNA with with
09:27two strands that are fully repaired.
09:31So the proteins that are participating
09:34in this pathway really has been
09:38identified through studies of
09:41Franco anemia patients and.
09:43Just go to go back for a moment
09:45to the history of this disease.
09:48Fanconi anemia has been identified in
09:501927 by Guido Fanconi and described
09:54as hereditary pancytopenia with
09:56short stature and hyperpigmentation.
10:00The first gene fancy was identified 92,
10:02but even then it was known that more
10:05than one gene will be causative
10:07for this mutation.
10:09And this was Buckwald lab who
10:11identified the first gene.
10:13And now we know that there are 22 genes
10:16that can be mutated in Fanconi anemia
10:18and we actually are working on on #23,
10:22which fits very nicely in this pathway.
10:26And the this data is actually showing
10:30the percentage of of patients with
10:33different mutations in different genes
10:35and this is from our International
10:39Franconia registry that I,
10:41I run at at the Rockefeller University.
10:45And you can see that Frank a mutations franc
10:48and Frank G mutations are the most common.
10:51More majority of the diseases of this
10:54disease is autosomal recessive we have.
10:56Language which is excellent and
10:58then uh thank our that we've
11:01discovered right 51 mutations.
11:03These are always the Novo dominant
11:05and we have a number of patients now
11:08with different mutations and and
11:09rat 51 and all of them have the same
11:13dominant function and they're all
11:15denovo presumably because right 51
11:17is necessary during biosis and we
11:19would never recover any any patients,
11:22any transmission really through
11:25through the germline.
11:27OK.
11:27And this registry that I've mentioned
11:29has been really instrumental for
11:31our work and it was started in
11:341982 by Arlene Auerbach who's still
11:37participates in the lab meetings
11:39and has a lot to say about about
11:42the disease and has been really
11:44wonderful in in in helping us through
11:47this through some of these studies.
11:50So now I'll spend really few just few
11:54minutes on the proteins themselves.
11:56Just to give you a gift,
11:58again a flavor for the for this pathway.
12:01So majority of the proteins whose
12:04genes are mutated in Fanconi patients
12:07form this large core complex.
12:10All of the colored proteins are
12:12actually those that have mutations
12:14in Fanconi in Fanconi patients.
12:17And you can see that this core complex
12:20is situated that this cross link as
12:22if it was identifying the the damage.
12:26It's actually unclear.
12:27How that's exactly done?
12:30The key to this core complex is that
12:34it has this E3 ubiquitin ligase fancl
12:37and with Frank T, which is the E2,
12:41they the whole purpose of this core
12:43we we understand now that the whole
12:46purpose is to ubiquitinated Frankie
12:49and Frankie 2 and that process
12:51through beautiful structural studies
12:53that have been done in the past.
12:55Semoran Pablo Ditches lab
12:57shows that the Fanki and Frank.
13:00Who forms a clamp around the the crosslink?
13:04Actually there are more,
13:05probably more clamps around as
13:07shown by the Dean slab that sort
13:10of decorate this cross link.
13:12Eventually,
13:13and this is again not quite worked out,
13:17the proteins that are important for
13:20in making these incisions come in
13:23and one of the important proteins
13:25here is this SLX 4 which acts as a
13:29scaffold for three different nucleases.
13:31And it's quite amazing how the three
13:35nucleases are associated with one
13:37protein and frank P is or a cell.
13:40X4 is essential not only for scaffolding.
13:43But for the function of the of these
13:46other nucleases without this slex
13:484 none of the nucleus is actually
13:51are functional within the cell.
13:53They have full in vitro function
13:55but not not cellular function.
13:57So it's it's fascinating but in
13:59this case XPF is the one that's
14:03making the incisions.
14:04I'll also mention the protein that
14:06that or disease that Barbara mentioned
14:09the carrier Magali Constitution
14:11nephritis which has mutations.
14:13In fan one and Fan 1 gene codes
14:17for this another nuclease and this
14:20system which can also unhook,
14:22but it can unhook this this cross link
14:25outside of S phase and we're still
14:28very much interested in knowing how
14:31how that affects the the the function
14:34especially in the kidney and and the liver.
14:39Eventually,
14:39after these, uh,
14:41these incisions are made,
14:43you bring in translesion polymerases
14:46and they are used to to replicate
14:50across this unhooked lesion.
14:53And that's actually probably
14:56not a mutagenic process.
14:59It's mostly a Posada that that can
15:03that can repair this this lesion.
15:06And eventually, as I mentioned,
15:08there's a whole homologous.
15:10Combination pathway that is involved
15:13in a repair of of the double strand
15:16break and here all of your proteins,
15:18favorite proteins that are
15:20associated with breast and ovarian
15:22cancer when they're mutated when
15:24the genus mutated in one copy.
15:26So BRC 2, power B2B,
15:30RC1,
15:30rad 51 are are functioning in this in this
15:35pathway and eventually we go back to SLX.
15:39Or and associated Mercedes one and
15:43the selects one where the repair needs
15:47to be completed using these these
15:51nucleases or a bloom healer case.
15:54I will also mention that there is
15:56another another pathway that's and
15:58the reason why the RCA two and Route
16:0051 are necessary at these cross
16:02links and that's to protect these
16:04cross links from from nucleases and
16:07our lab has shown that it's there.
16:10There is a protection against DNA to
16:13and Warner and other labs before us
16:17have shown for especially the Jason
16:20lab that there is also protection
16:22against MRE 11 at stalled.
16:24Forks for example,
16:26after hydroxyurea and other other damage.
16:29So I this is this is all I I have
16:32to say about the mechanism of
16:34of function of these proteins.
16:36There are still a lot of details that
16:39need to be that need to be filled in,
16:42but we have an idea of how
16:44how this pathway functions.
16:46And the only other thing that I say
16:48I'll say about this is that it's
16:51really a very well regulated process,
16:53so you make double strand breaks.
16:55But they are immediately being
16:57shuttled to the proper repair.
16:59OK, so we'll talk about what happens
17:01in the in the third part of my talk,
17:04we'll, we'll talk about what
17:05happens when this repair is abnormal
17:07and the brakes are still made,
17:08but they are inappropriately repaired.
17:13So in the second part,
17:15I'll talk about the endogenous
17:17sources of interest and cross links.
17:19So I showed you that the Franconia pathway
17:23is necessary for repair of I CL's,
17:26but you may ask, well,
17:27where are these IC's coming from?
17:30Obviously in the in the
17:32clinic we are very much.
17:37Aware of cisplatin and mitomycin C and
17:40interest rate crosslinks occur after
17:43treatment with these with these agents,
17:46although there are other
17:47lesions that occur as well,
17:48right intrastrand crosslinks and others.
17:51The exogenous damage also can come from
17:55bacterial metabolites, and there are.
17:58There are E coli strains that are
18:02making ICL inducing. Chemicals.
18:04It can come from acetyl,
18:06aldehyde, from ethanol.
18:08It can come from tobacco smoke,
18:11which is full of formaldehyde,
18:12acrolein and other mutagens.
18:15But formaldehyde and acrolein are
18:17very good into strong cross linkers.
18:20But we are particularly interested
18:23in sources of endogenous DNA damage
18:26and really beautiful work that I'll
18:28describe a little bit in the two
18:31slides that are coming up have shown
18:34that endogenous toxic metabolites
18:37formaldehyde and acetaldehyde are
18:41particularly important in the bone
18:44marrow and definitely contribute to bone
18:47marrow failure in Franconia patients.
18:50And in mouse models,
18:52so let me just introduce them.
18:56This work was done mostly by
18:58KJ Patel's group in the UK and
19:02they've defined LDH 2 and ADH 5.
19:05These are alcohol and aldehyde
19:07dehydrogenases as being important
19:09for this first tier of protection.
19:11So they remove the toxins from
19:14the cell and that prevents I CL
19:18DNA damage and of course the.
19:21Frankonia pathway is a second
19:23tier of protection.
19:24Whatever has been incorporated as
19:27cross links will be removed as well.
19:30So let me just give you a a shore
19:33segue and talk about a LDH 2 and ADH 5.
19:37So a LDH 2 is necessary for removal
19:40of acetyl aldehyde which is which can
19:43come from ethanol from the outside,
19:46but also from metabolism,
19:49intracellular metabolism.
19:50And LDH 2 is responsible for
19:54detoxifying it to acetate.
19:57And a KG Patels group has described
20:01LDH 2 knockouts when combined
20:04with Fangy 2 knockouts.
20:06The mouse for the first time
20:08really developed bone marrow fell
20:10failure and leukemia.
20:12So the mouse model without just fancd
20:152 mouse model does not develop.
20:19Bank bone marrow failure or leukemia.
20:21It has mild hematopoietic dysfunction,
20:24but it's nothing that you consider
20:27a bone marrow failure.
20:29In humans combination of FANK
20:32mutations and usually it's frank,
20:34a mutation and a LDH 2 star 2
20:38which is the dysfunctional.
20:40Variant of LDH 2 causes early onset
20:43of Fanconi anemia phenotype and there
20:47is a really beautiful paper from
20:50the Takata group that have that has
20:54shown that and that was published
20:57in 2013 which really solidified
20:59our thinking about this this as a
21:03as a modifier of Fanconi anemia.
21:05And I also will mention that together
21:09with Chris Vakoc's lab we showed that.
21:11Somatic Aldh 2 mutation silencing
21:16in AML's results in in dependency
21:19on the Fanconi anemia.
21:22So when when Chris's lab and they
21:25are at Cold Spring Harbor did a
21:28screen and showed in number of
21:31AML's dependency on Fanconi anemia.
21:34He called us up and and asked what
21:36what should I look at why are they dying?
21:38And I said well just look at LDH 2 I
21:41bet it's. Yeah, 2 and that's what
21:43that's what it turns out to be.
21:45We just helped him a little bit but
21:47some of some of AML's depend on on
21:50Franconia because of the of the
21:52silencing of LH2 which I think is quite
21:56interesting and has implications for
21:59for therapies for these for these AML.
22:01So then ADH five came.
22:05And that that was also from KJ Patel's
22:09work that ADH five is necessary for.
22:13I mean this is this has been known
22:15that has been known that ADH five
22:18is necessary for detoxification of
22:20formaldehyde through this through this
22:23pathway and again they create a DH,
22:26five -, 22 negative mice and they
22:28had bone marrow failure but also
22:30had gromel glomerular damage.
22:32That's actually poorly understood
22:34still and there is this.
22:36Uh new digenic human disease of
22:40LDH 2 ADH 5 double knockouts or
22:44the dysfunctional alleles which
22:47have bone marrow failure,
22:48myelodysplastic syndrome,
22:49foot for some reason,
22:51foot skeletal abnormalities
22:54and also neurodegeneration.
22:56So now you have phenotypes that are
22:59associated with ACL's and creation of
23:02I CL's in multiple different tissues.
23:05Including in the brain and we if somebody
23:08is interested in knowing why the brain,
23:10we can discuss that later.
23:12But there that really shows
23:14us that there is a lot of.
23:18Well,
23:19the cells have a lot of different
23:21pathways for as this first tier
23:24of protection and my lab has been
23:27interested in identifying other
23:29pathways in different cells as cell
23:32types and we started with with cells
23:34of jurkat cells which are T cell
23:37leukemia cells to figure this out.
23:39So our hypothesis was that there
23:41would be more of these detoxification
23:44path pathways in the cells and
23:47by understanding what that is.
23:49They might actually be equipped
23:51with understanding of what can we
23:54increase to have some preventive
23:57measures in in Fanconi anemia.
23:59So this is work from Munjung Jung
24:02who's now a a assistant professor
24:05at Hopkins and she was a clinical
24:08scholar in my lab and she did
24:11a metabolism focus screen.
24:13So this is crisper screen or that she.
24:18Performed in jurkat cells.
24:20And she took Frankie to positive
24:23and negative jurkat cells.
24:25And the whole purpose was to
24:27identify the genes that are dropping
24:30out from fangy to negative cells.
24:33So these are essential,
24:35would be essential for faculty to cells.
24:38And you can see that there are
24:40lots of genes that that dropped
24:42out and some of them were actually
24:45consistent with what we know.
24:47So these are these cells, SLC 7A.
24:5111 and Assoc 3A2.
24:54They form a complex that is responsible
24:57for cysteine movement across the cell
25:01membrane and that feeds into glutathione,
25:04which is important for
25:06formaldehyde detoxification,
25:07so that fit.
25:08There was also a we did get a
25:11DH5 in the screen,
25:13but the gene that we concentrate
25:16on concentrated on was a LH9A1.
25:20Eight out of 10 guides scored in
25:24this in this assay and the rest
25:27of the work was really based on a
25:29LH9A1. This is a aldehyde dehydrogenases
25:33in metabolizes amino aldehydes,
25:35at least in vitro and it's highly
25:38expressed in liver, muscle and kidney.
25:40And the idea here is that it will there
25:44will be tax toxic metabolite that is
25:47now detoxified by AL DH9A1 and this.
25:50If you don't have aldh 91,
25:52the toxic metabolite will create
25:54DNA damage and you don't have if
25:56you don't have Franconia pathway,
25:58these cells would die and get transformed.
26:00But if you have Franconia pathway,
26:02you would have cell survival.
26:05So umm munjung has validated
26:09this in multiple assays,
26:11and there is a bio archives paper that
26:14we've we put out there competition.
26:17Assays cells that didn't have a
26:21two and a LDH 9A1 did much more
26:25poorly and growth assays.
26:28She also could show that there were
26:31increased numbers of apoptotic cells,
26:33increased DNA damage through gamma H2AX.
26:37And increased chromosome breakage.
26:38So this is where we can look at
26:42without exogenous DNA damage.
26:44Look at numbers of breaks that are
26:47occurring in the double knockouts
26:50in different different clones and we
26:52can see that there is a increase of
26:55of chromosome breakage if we don't
26:58have a LDH 9A1 Infinity 2 cells.
27:02She also used human hematopoietic
27:04stem cells depleted Frank A with SH.
27:08RNA's did a knockout with of LDH 9A1
27:11and those could make fewer colonies.
27:14But what happened?
27:15We made a mouse expecting some
27:18level of of bone marrow failure
27:21and we've seen phenotypes so these
27:24mice are born smaller than fanki.
27:27Negative mice which are already
27:28small have increased number
27:30of eye abnormalities at birth,
27:32which has been seen as a DNA.
27:33Damage outcome and they have increased
27:37number of variant tumors in aged mice,
27:40but really very mild hematopoietic defect.
27:43So that tells us that there will be
27:46differences also between mice and
27:48men in how the these detoxification
27:51pathways are working.
27:52And This is why we when we are modeling
27:55things in the mouse with in this in
27:57this pathway I think it's important
27:59to to do things and in parallel
28:02in human system and in the mouse.
28:04And and compare and contrast
28:06because there will be differences
28:08and whatever we model in the mouse
28:11actually might not be as as important
28:14in humans and and vice versa.
28:18So the last question that
28:20we wanted to ask is, well,
28:22we've identified the the enzyme,
28:25but what's really the,
28:27the problem, what's the source,
28:30what's the toxic metabolite that these
28:32cells are dealing with so much junk
28:35did a suppressor screen as we are,
28:37we are geneticists at heart.
28:39So she took the double negative cells
28:43frankly to a LH9A1 double knockouts and.
28:47I redid the screen the the same method
28:51with using the same metabolism library,
28:54but now wanted to see the cells
28:56that are actually growing better.
28:58So now we are knocking out some other
29:02gene that is necessary for production
29:05of our of our toxic metabolite.
29:10And the Omni gene that we've
29:14identified is this ATP 13A3.
29:16We've actually identified a LDH 9A1 itself,
29:20but that's we think is actually
29:22through reversion mutations.
29:24So these are now we are creating a mutant
29:28and a LDH 91 that reverts the function.
29:31So we in a way it shows that
29:34our screen worked in both ways,
29:37but this gene is quite interesting
29:39because this is now.
29:41A protein that's necessary codes for a
29:44protein that's necessary for polyamine,
29:47for example spermine.
29:51Movement through the from between
29:53different membranes and if we
29:56have high level of spermine,
29:58we know at least this is
30:00what we think would happen.
30:02We would get high levels of I mean
30:05the proper now and as I mentioned I'll
30:08ADH 9A1 is involved in detoxification
30:10of amino propanol and propanol
30:13can lead to acrolein and cells.
30:16We actually haven't yet shown that this
30:18these are this is what happens in the system.
30:20But everything that we have is consistent
30:23with this, with this hypothesis, so.
30:28If we have triple mutants, these cells,
30:32these cells now can grow much better,
30:36presumably because we are sequestering the
30:40polyamines outside of the away from the DNA,
30:44not creating these these crosslinks.
30:49So I think this is a a an example
30:53of how we are.
30:55Identifying endogenous types of DNA
30:58damage that are necessary for crosslink
31:01repair for that are necessary to be
31:04repaired by franklinia pathway and
31:07other pathways of crosslink repair.
31:10And really adding to this idea of
31:15how the first tier of protection is
31:19necessary to to protect the the genome,
31:22not creating interstrand crosslinks.
31:25And that eventually obviously is necessary
31:28for normal cell and organ function.
31:31So with that,
31:32I'll move to the last topic,
31:35which is cancer and Fanconi anemia,
31:38children and young adults.
31:42And when you think about Franco
31:45anemia pathway,
31:46there are,
31:47I already mentioned that there are
31:50different types of of tumors that
31:53can form and one of the tumors that
31:56we think about are these embryonal
31:58tumors and AML that are forming when
32:03homology directed repair is absent.
32:05So these are patients who
32:08have beers biallelic BRC 2.
32:11Or probably 2 mutations and they are
32:15identified early age you know they
32:17they have they developed these tumors
32:20within the first five years of of
32:23their life and we actually have a
32:26mouse model of of medulloblastoma
32:28that is quite interesting but
32:30not ready for prime time.
32:32But the the reason why we are
32:35interested again in these tumors
32:37is because we want to understand
32:39why the granule progenitor.
32:42Cells or the cells that lead to
32:44neuroblastoma or tulips tumors,
32:46why do they really need BRC 2 function?
32:50What's so special about these cells
32:52that require BRC 2 function or probably
32:552 function homologous recombination.
32:58So that's something that we are
33:00we are continuing to develop.
33:03But outside of homologous recombination
33:05when the ICL repair is abnormal,
33:08so these are patients with mutations in
33:11any of the core complex fanki FANGY 2 SLX 4.
33:15We don't have yet patients
33:17who have tumors with XPF,
33:18but that's more of a problem of the PF
33:23being necessary also in other tissues.
33:26These patients develop AML's and also
33:29develop squamous cell carcinomas
33:32and today we'll concentrate on
33:34squamous cell carcinomas.
33:36So this is a knowledgeable audience
33:39about squamous cell carcinomas and
33:42we'll concentrate on head and neck.
33:45And in sporadic cases,
33:47you have two types of of head
33:50and neck cancers.
33:51One of them is HPV associated
33:56that we want to spend too
33:58much time talking about.
33:59The other one is a HPV negative
34:03and that those tumors are
34:06associated with carcinogens that
34:08are present in tobacco and alcohol.
34:10And I would also claim that endogenous.
34:15Aldehydes will play a role in here as well.
34:19And maybe for franconi patients who might
34:21not be exposed to alcohol and tobacco
34:24as much as the general population are,
34:27those indulgence aldehydes will be important.
34:32The the truth about head and neck cancer
34:35is that the patients are diagnosed late,
34:38the survival is still poor and it's not
34:41really decreasing and if it's decreasing
34:44it's it's just because there's more,
34:46there are more HPV positive cancers
34:50and again the treatment can be quite
34:56horrendous for for the patients and
34:59really learning from from Barbara.
35:01The the need is really to stratify these
35:05patients to also to identify novel
35:08therapeutics that might not be DNA damaging.
35:11So we, my lab got interested in these
35:16tumors because of our registry and tumors
35:20and Fanconi anemia patients just observing
35:23the patients who are becoming young adults,
35:27especially bone after bone marrow transplant.
35:29But even without the bone marrow transplant,
35:32many of these patients have developed
35:35cancer and in 2003 there was a paper
35:38from the registry when I wasn't.
35:41Yet there but showing that patients with
35:45Fanconi anemia had 700 at least 700 fold
35:50increase of of tumorigenesis in in the
35:53head and neck cancer in head and neck
35:56area and for vulvar cancer and for for
36:01cervical cancer and anal cancer there
36:04were thousandfold increases in these.
36:07So these are cancers that are squamous cell
36:11carcinomas in Fanconi anemia patients.
36:14And they are very there.
36:17There are present in the tongue
36:21gingiva and buccal mucosa,
36:23some pharynx and larynx,
36:25but a lot of cancers in in the oral cavity.
36:30We do have some esophageal cancers as
36:32well and those are quite interesting,
36:35but all of them have similar similar
36:39genetic or molecular characteristics.
36:41So here is our and this is now I'll describe
36:46work and I'll go fairly quickly through it.
36:49Work through that was recently published.
36:53This is our cohort of patients who were
36:56whose tumors were sequenced and you
36:58can see that the agent diagnosis is.
37:00On median of 31 years old,
37:03but we have some 13 year old 16 year
37:06olds with head and neck cancer that's
37:09extremely aggressive and extremely
37:11difficult to treat since we cannot use
37:14this platin for for these patients
37:16and they are radiation sensitive as
37:19well as as patients although not that
37:22radiation can still be used carefully
37:24you can see that these patients don't
37:27do well at all much worse than the.
37:30Sporadic cancers and patients who are
37:34here and survive long usually are
37:38patients who had successful surgical
37:41resection of their of their primary tumor.
37:46Or occasionally radiation therapy
37:48that actually was was successful.
37:51OK,
37:51so we were very much interested
37:54in understanding the molecular
37:57pathogenesis in these tumors and we
38:00showed that majority of these tumors
38:03were HPV negative instead majority of
38:06them had P53 mutations and the P53
38:10mutations were of variety, nonsense,
38:13missense, frame shifts and deletions.
38:17But P53 was pretty much the only Gina gene
38:21that was mutated through point mutations,
38:24OK.
38:25All the other changes were actually
38:28changes that were associated
38:30with structural variants.
38:32So here you can see that these are
38:35number of mutations in across different
38:38cancers through TCG data from TCG
38:40data and I inserted our Fanconi
38:43tumors and they have point mutations,
38:45the number of point mutations.
38:46It's pretty low and these tumors are lower
38:49than in sporadic head and neck cancers.
38:52And if we look at the what kind
38:55of signatures are present,
38:57there is no homology,
38:59directed repair or smoking signature and a
39:03lot of these are signatures of cell division.
39:07So the reason why we think that
39:09they have smaller, fewer.
39:11Mutations is because they are
39:15present in younger patients.
39:17And instead what we see is this
39:20huge genomic instability where this
39:22is a circus plot and all of these,
39:25all of these.
39:27Lines indicate translocations and
39:30structural variants that are present
39:33in in these tumors and if we look at
39:36number of structural variants across
39:39these tumors the there is about 2 to
39:43threefold increase when we compare
39:45it to HPV negative tumors and HPV
39:48positive tumors have very few of
39:50these since they already mutated
39:52P53 and RB and really don't have to
39:56rely on these structural variants.
39:59And this is comparable to the
40:02structural variants that we see in
40:04BRC 2 or BRC 1 tumors which have
40:07high levels of structural variation.
40:12As far as types of type of of structural
40:15variance, we see number of them,
40:16we see deletions, translocations,
40:19inversions, all of them are increased.
40:21And if you take proportion
40:23of structural variance,
40:25actually the on the there is a slight
40:28increase of tandem duplication.
40:30So pretty much across all structural
40:33variants there is an increase.
40:36Which is consistent with breaks being
40:39made when the DNA when Fanconi pathway
40:43cannot function and these breaks being
40:46being repaired inappropriately and
40:49it's and and to show you the extent
40:52of how inappropriate this repair
40:54is we used number of techniques of
40:58long read sequencing and you can see
41:01that pieces of chromosomes are from
41:03different chromosomes chromosome 8.
41:06For 1511,
41:06they're all sort of strung together
41:09in the in in these tumors,
41:12creating these completely
41:15abnormal abnormal structures.
41:18So we also asked how these
41:21DNA molecules come together,
41:23whether there's something special
41:26about them and it's many of you know
41:29if you have double strand break you,
41:31it can be repaired through
41:33three different mechanism,
41:34non homologous end joining,
41:37microhomology mediated repair
41:39and single strand annealing.
41:41And all of these,
41:44all of these should be functional
41:48in Franconia negative cells.
41:50And indeed when we look at the junctions
41:53in franconi associated squamous
41:55cell carcinoma that are shown on
41:58the left and sporadic HPV negative,
42:01we see the same repair mechanism
42:04being active mostly on homologous non
42:07homologous end joining and MJ and that's the.
42:12The percentages are are identical.
42:14So the interpretation of this is
42:16that whenever there is a break
42:18that's formed in these tumors,
42:20it's being repaired just by sticking
42:23the the ends together and random
42:26fashion that results in these high
42:29number of structural variants.
42:31We also asked where these breaks
42:34occur and specifically we were
42:36interested to in in knowing whether
42:39they occur at repetitive sites.
42:42And there is some increase that's
42:46statistically significant in
42:48sign elements, so I'll repeat.
42:51Had higher number of of breaks,
42:55but when we look at the whole genome
42:59we can see that the brakes and here
43:02I'm showing where all of this breaks
43:05out and structural variant ends are.
43:10I think when they're being placed in the in
43:12the genome across all of the chromosomes,
43:15you can see that there is the brakes
43:18are occurring throughout the genome.
43:20But you do see that parts of some of
43:24the genes are being hit multiple times.
43:27And that's really the the way we
43:30think about it is that there is
43:32mutagenesis across and then on top
43:35of that there's a selection of
43:37particular genes that help the tumors.
43:40Growth for example EGFR here is is being
43:44amplified and so are other sites you know,
43:47but one here.
43:48This is the EGFR and and also
43:52others the outcome of this.
43:56Is that these tumors have very high copy
44:00number variance and this is just one tumor,
44:03this is Ascot plot.
44:04So this is a little specific
44:06copy number where total alleles
44:08obviously in the cells should be 2,
44:11the alternative allele should be one.
44:13But you can see that there are
44:16number of amplifications and their
44:18amplifications and genes that we
44:21all know from from tumor pathways.
44:24Pick three CA make cycling.
44:26One that's that's amplified in many
44:29of these tumors and then there are
44:32deletions and tumor suppressors
44:34like CDK and two way and P53.
44:37So this is a pretty representative look
44:42at the at the tumor of Fanconi anemia
44:46patients and this is a different way
44:49of representing it where all of the.
44:52The rows are the the genes that are
44:56mutated and the columns are single tumor
44:59that we've sequenced from Fanconi patients.
45:01And you can see that orange or
45:04amplifications and deletions in
45:06blue and multiple of these pathways
45:08are mutated in each of the genes.
45:11And if you look at pick three CA and
45:14Mick Amplifications 54% of these
45:17tumors have Co Co amplification
45:20of these two of these two.
45:23Oncogenes, so this is a very,
45:25this might explain part of why
45:27these tumors are aggressive but
45:29there are other reasons they they
45:31they might be aggressive as well.
45:33So to just summarize what
45:36happens in franconi tumors?
45:39We have.
45:41Franconia pathway that protects cells
45:43from creating DNA breaks when DNA
45:47interstrand crosslinks are present and
45:49if you have DNA breaks the structural
45:53variants follow and eventually all
45:55of them lead to high copy number
45:58variation in these in these tumors.
46:01And there is also I wanted to mention
46:03there is a paper that recently
46:05came out from Jean Souliers Lab
46:07who looked at AML's and and AML's.
46:11They also see structural variant formation,
46:15fewer structural variants,
46:17and they also see P53 pathway being abnormal.
46:22But actually P53 itself is not mutated,
46:25it's through the MDM 4 pathway that
46:27that there is suppression of P53.
46:30So in the bone marrow similar
46:35events are happening.
46:36Maybe it's at a slightly smaller,
46:39smaller scale, which I don't.
46:41Alright, understand.
46:44And might have to do with the timing of
46:47these tumors and time to evolution of
46:49of and presentation of these tumors.
46:52So knowing what we know
46:55about frankonia tumors,
46:57we then turned to sporadic cancers
47:01and we were wondering whether the.
47:05Structural variants that are present
47:08in sporadic HPV negative tumors.
47:10I showed you that there are still
47:12some of them and they're shown
47:14here on the right hand side.
47:16These are all HPV negative
47:18TCG head and neck cancers.
47:19You can see that there's
47:21plenty of orange and blue.
47:22So there are quite a lot
47:24of copy number variants,
47:25not as many as in in Fanconi tumors,
47:27but still.
47:30A high number.
47:31So we were wondering whether what
47:34we find in Fanconi tumors might
47:37apply to to head and neck cancer
47:41and sporadic and sporadic cases.
47:44And what we decided to do is to look at
47:49a copy number variants and stratify tumors.
47:53So now these are the TCG TCG data
47:56HPV negative tumors and stratify
48:00them into top quartile.
48:02Tumors with a high number
48:04of copy number variants,
48:06so top quartile,
48:07and then compare them to the low quartile,
48:10the top 1/4 of tumors with
48:14the low copy number variants,
48:17and we could correlate the
48:20number of pack years with the
48:23level of copy number variants.
48:25So the higher smoking history,
48:28the higher copy number variants.
48:31And when we looked at signatures that
48:34were present and in these two we could
48:37also show that the top quartile.
48:39Had about twofold increase in signatures
48:45of ID 38SBS4 and others and these
48:49are smoking associated in Dells,
48:50so that's not surprising,
48:52right?
48:52That's correlates with the pack history
48:56but also non homologous end joining
48:59and benzopyrene and acetaldehyde.
49:01That suggests something we we already
49:04know from epidemiological studies that
49:07smoking and drinking go together.
49:10So we so these patients probably
49:13are also enriched for for
49:16increased alcohol exposure.
49:18So taking all of this together,
49:22we come up with this hypothesis
49:25that in sporadic cancers,
49:28the tobacco and alcohol exposure creates
49:33DNA inter interesting crosslinks.
49:36Obviously it creates other mutagenic events,
49:39including P53,
49:40which is an early event that then
49:42allows structural variants to occur.
49:45And Franconia pathway even though
49:48that it's present in these in
49:51these tumors and maybe it might
49:54be somatically decreased or maybe
49:56there are some genetic factors
49:58that that are involved here,
50:01but frankonia pathway which
50:03is largely efficient.
50:05There isn't just enough of it and
50:08it's over is overwhelmed and unable to
50:11repair all the DNA into crosslinks.
50:14It's also important that P53 is is
50:16mutated in these so you cannot depend
50:19on P53 pathway to to increase the Franconia.
50:25Green expression kind of going in
50:28gene expression patterns and all
50:29of this results in DNA breaks,
50:31structural variants and the same
50:34pathway leading to copy number alterations.
50:38So taken together,
50:39you know when we look at the
50:42FRANKLINIA pathway function or if we
50:45have patients with low pathway function,
50:49they have very high levels of head
50:51and neck cancers and other squamous
50:53cell carcinomas, maybe not 100%,
50:55but if they Long live long enough,
50:58very high levels and if we have
51:02full function we are protected.
51:04However, there are a number of
51:06genetic modifiers whether it's a.
51:08The
51:11H2CTA4DC GSTM one and others,
51:13as well as these environmental
51:16modifiers of alcohol tobacco that
51:19is making this pathway functionally
51:22inefficient leading to to higher
51:24probability of head and neck cancer.
51:28So I'll finish and um,
51:31I'll just acknowledge my lab.
51:34This is Arlene, who's been
51:36instrumental in starting the registry.
51:39Munjung Jung has worked on a LH9A1.
51:44They work on head and neck cancers,
51:46was hugely collaborative work,
51:48and we were helped by Matthias Sanders.
51:53Working with Peter Campbell and Andrew
51:56Webster drove this work in my lab.
51:58And we had.
52:00A lot of collaborators,
52:02clinical and patient collaborators
52:06and and also other bioinformaticians
52:10who've who've helped us.
52:13And I'd like to thank my funding agencies,
52:18tons of collaborators and
52:21especially patients and families.
52:23And these are some of the young adults
52:25few years ago before the pandemic and
52:27I know that at least five of these.
52:30Adults are have passed away with with
52:33head and neck cancer and during this time.
52:36So I'll take questions.
52:40Thank you.
52:49OK. That was wonderful.
52:50For those of you who are online,
52:52please use the Q&amp;A function to to bring your
52:55questions and I think Jeff. Jeff, yeah.
52:59Ohh uh you can have a microphone.
53:01I think it's here.
53:04And and maybe I'll just start
53:06with Nadia, Dimitrova asked.
53:08Do you see evidence of extrachromosomal
53:10DNA circles and FA tumors to
53:13explain the amplifications and
53:14we haven't really looked at that,
53:17but that would be a good idea.
53:19We do see a lot of tandem duplications
53:22in the in the especially Mick and
53:25other areas that would explain
53:27the amplifications as well. Jeff
53:29yeah. So that was a great talk,
53:30incredibly clear and insightful.
53:32I really appreciate it.
53:33I really like that last diagram.
53:35I know it was meant diagrammatically
53:36at least the way you showed it,
53:38but it can't.
53:39That doesn't stop me from
53:41asking do you have a way to
53:44quantify Fanconi anemia pathway
53:46function on that X axis or not.
53:48That's a really,
53:49that's a really good question.
53:51So we we have some way of doing that
53:54within the Franconia population,
53:57so something that I haven't talked about.
53:59We have. The registry has been
54:01amazing and we sequence now
54:04pretty much everybody who we had.
54:06The DNA for to look for phenotype,
54:10genotype correlations and
54:11there are a lot of phenotype,
54:14genotype correlations that we can tease
54:16out and we've started to describe them.
54:20And the function we can at the lower
54:23end it's hard to tease it out but
54:25you can definitely do that with
54:28number of foci with ubiquitination
54:30of fancd 2 and Frank I it's not great
54:33I think we need a better assay for
54:36for doing that on the top end I think
54:39when we have like what we see what
54:42we think fully functional pathway.
54:45It's really difficult to say
54:47whether it's 95 or or 90%.
54:51So we haven't really been able to do that.
54:55But as you saw from the second
54:57part of my talk, you know,
54:59even a little bit of DNA damage
55:01can lead to chromosome breakage
55:03like the ones that we induce when
55:05we don't have a LDH 9A1.
55:07So I think there might be ways of of
55:10doing better assay development for this.
55:16So.
55:21So if I can
55:23ask a question. So one of the things
55:25we're noticing in the virtual tumor
55:26board that we do for these young adults
55:29with head neck cancer is that they are
55:31really not at all responsive to immune
55:34checkpoint inhibition which is often
55:35been brought forward as as something
55:37that might not be so DNA damaging.
55:39And I wonder if you could speculate
55:42about that both from the perspective of?
55:45Sort of this global picture you've been
55:47talking about with structural variants,
55:48not some so many single nucleotide variants.
55:51But also, is there anything that came out
55:53in your gene signatures that would speak to?
55:56I saw NSDD one which sometimes is
55:58associated with immune exclusion.
56:00Are you getting a picture from the mutations
56:02of what the immune status might be?
56:05So from the mutations we we
56:08do we do see deletions in.
56:13A message. Will say.
56:18But frankly the the genome instability
56:22probably has a big part in the in the
56:27inability to have immune response right.
56:29So we are very much interested in
56:32that and we are starting to to look
56:34at that the tumor microenvironment
56:36in these tumors in particular
56:38thinking that it's it will be a good
56:41model for for understanding that.
56:44So we'll, we'll, we'll see.
56:45But it definitely looks like it
56:47would be too more suppressive, yeah.
56:52Tommy.
57:05Yes. So that's an interesting question and
57:08there is some some data in the literature
57:12that implicates mitochondrial DNA damage and
57:14Fanconi function within the mitochondria.
57:17I'm not sure how the fund Fanconi
57:20proteins would get into the mitochondria,
57:23but it's something that we have not.
57:26Explored. It would be so that,
57:29I'm sorry, the question was about the
57:31mitochondrial damage and whether there
57:33is increased mitochondrial damage.
57:35We we haven't really explored that.
57:39Um, and there would be.
57:42Well, it's it, it would be doable, right.
57:45I mean you just we would have to
57:47look at mitochondria specifically.
57:50The audio. That's right.
57:53A lot of these, yes.
57:55A lot of these are the
57:57detoxification enzymes.
57:58Not all of them,
57:59but many of them are within the mitochondria.
58:04The spermine? Yep.
58:11Right. Right.
58:18Right. Yeah. So there is a lot,
58:20a lot that we could explore in in this
58:23disease and mitochondrial function
58:24definitely has been brought up as as one
58:28of the places where where we should,
58:30we should look more.
58:34There was one question,
58:36is there a bias toward either focal or
58:40ARM level copy number alterations or
58:43both equally in FA tumors when compared
58:46to HPV negative head and neck cancers.
58:48So we see, we see both,
58:50I don't know if if we really compare them
58:54so we can say whether it's equal or not,
58:57but we see, we see we see some ARM
59:01level copy number alterations as well.
59:04But majority I would say,
59:06are the focal type.
59:14OK, yes. It seems that.
59:22Particular these metabolic enzymes.
59:27Essentially necessary to collect cells of the
59:31dehydrogenases ID. Also very common.
59:36Biosynthesis inhibitors.
59:42Do you think, well, what are your insights
59:44on a particular reason why blood
59:46cells would be more susceptible?
59:51Right. So one possibility is that
59:53they are really not more susceptible,
59:56they are just the there are
59:58fewer of them and they exhaust
60:00faster than any other stem cells,
01:00:02right, so that that's let's
01:00:04leave that as a possibility.
01:00:06The other is we know for formaldehyde.
01:00:11That. Very high levels of formal
01:00:16formaldehyde can be formed close to the
01:00:19DNA during the differentiation process.
01:00:22So one hypothesis is that you're
01:00:27having differentiation that happens
01:00:29during very few replication cycles,
01:00:33and that vulnerability of two
01:00:36formaldehyde is very high.
01:00:39Umm.
01:00:41So that's a possibility in coratina sites
01:00:44when we do an assay for for example
01:00:47whether ADH two or ADH 5 or important
01:00:50in with human characteristics sides,
01:00:53there doesn't seem to be that dependency
01:00:56within the fungi 2 or frank a negative.
01:00:59But we when we start playing with
01:01:02lipids metabolism they show up.
01:01:04And then it makes sense because it's the the,
01:01:08the lipid metabolism is very important in the
01:01:11in the as a barrier in correcting the sites.
01:01:14So some of it,
01:01:15I I do believe that some of it will come to
01:01:20specific functions or a particular stem cell.
01:01:23Um, and I think just more
01:01:25needs to be done to understand,
01:01:27to understand that one of
01:01:29the challenges is that.
01:01:31It's very hard to identify how much
01:01:34of these aldehydes are present and
01:01:37especially within the DNA it's just
01:01:40very difficult to to quantify these.
01:01:44But it's an I think it's
01:01:46a very interesting area.
01:01:50OK, Ryan. Yeah, great talk about it.
01:01:54This morning, your copy number,
01:01:56your copy number, application
01:01:57analysis. Have you ever
01:01:59have you looked at the sequences
01:02:00of those copy numbers?
01:02:01Are they mutating in each copy number?
01:02:03And what do you think
01:02:04driving that copy number,
01:02:05especially the amplification?
01:02:07Do you think it's like a DNA repair
01:02:09pathway, like pull Theta
01:02:11or something like that?
01:02:12Yes. So let me just,
01:02:14I don't know if I left.
01:02:16Oh, I can't escape, OK.
01:02:19Well, I won't try to find slides for this.
01:02:23So we as far as you mean like
01:02:27point mutations within the within
01:02:30the amplified regions,
01:02:32we don't really see any kind of.
01:02:36Aggregation of mutations
01:02:38like at the junctions.
01:02:40Let's say we've looked at that.
01:02:42Um, and the way I some of these
01:02:46amplifications occur through
01:02:48probably or are consistent with
01:02:50let's say that with stalling at the,
01:02:54at the at the cross link or
01:02:57some other lesion and reversal?
01:03:00So we do see these inversions that
01:03:03that are associated would be expected
01:03:06to be associated with replication.
01:03:09Some of them are tandem duplications
01:03:12and I think a nice system to figure
01:03:15this out is actually in a system
01:03:18where you have tester sequences
01:03:21and stalling through tester that
01:03:25that would be one place where you
01:03:27could model some of these, I think.
01:03:29I guess I was thinking of all these
01:03:31reversion mutations that arise
01:03:32like in Fanconi anemia perhaps
01:03:34you know the copy
01:03:35numbers. They're sampling different
01:03:36genetic states so they can, you
01:03:38know, sort of revert to a.
01:03:40I have a lot to say about the,
01:03:42the copy, not the the reversions
01:03:44in Fanconi because we have a whole
01:03:47another paper that's coming soon on
01:03:49mosaicism in the blood and a lot of
01:03:52these mutations are actually somatic
01:03:54recombination events that are UMPD,
01:03:58their uniparental disomy events.
01:04:01So there are, yes,
01:04:02there are point mutations and there
01:04:04are other things happening you know,
01:04:06splice site.
01:04:08New splice site mutations
01:04:10and things like that,
01:04:11but a lot of them are are full uniparental
01:04:15disomy through mitotic recombination.
01:04:18Which I think is interesting.
01:04:20Thank you.
01:04:21Thank you very much.
01:04:23Thank you.
01:04:27Um, for the fellows I think you're.