Synthetic Lethal Therapy for Head and Neck Squamous Cancer

June 18, 2024

Information

Yale Cancer Center Grand Rounds | June 14, 2024

ID11807

To CiteDCA Citation Guide

00:00Barbara Burkness,
00:01who will be speaking today about synthetic
00:04lethality and head and neck cancer,
00:07is the Anthony Brady Professor
00:09of Medicine and Medical Oncology,
00:11the chief translational research
00:13officer in the Cancer Center,
00:15and the associate director
00:17for translational research
00:19in the Cancer Center as well,
00:22and leads the Head and Neck Cancer program.
00:26So Barbara has many accolades.
00:32She when I was coming to Yale,
00:35I was told by my,
00:37I think very capable colleague at
00:40the Inner Barber who read the Head
00:42and neck cancer program that I was
00:44moving to the institution with the
00:46best head and neck oncologist in the country.
00:48And I think it's fair to say that
00:51Barbara has has contributed widely to
00:54the field of head and neck cancer,
00:57both in terms of laboratory work and
01:00clinical trials and plain old education.
01:05Barbara went to Bryn Mawr College,
01:08went to medical school at Stony Brook,
01:11and then actually had the unfortunate
01:14experience of of coming across me because
01:18she was an intern when I was chief
01:21resident here at this fine institution.
01:24And I can tell you that if you're looking
01:27for an example of someone who fundamentally
01:30hasn't changed over the course of many years,
01:34you can look at Barbara because at
01:37that time she was incredibly bright,
01:41incredibly insightful, capable of
01:44doing almost anything and very direct.
01:48And I am,
01:50I am that wasn't that was pretty gentle.
01:54I said, I, I said there was no roast,
01:56just toast.
01:59And, and it's really a pleasure to
02:02have the opportunity to come back
02:04to yell and work with Barbara.
02:06I should mention that after yell,
02:08Barbara went on to a fellowship at at
02:11Memorial Sloan Kettering and then before
02:13finding her home in head and neck cancer,
02:16actually was briefly interested in breast
02:19cancer and GI cancer and spend time here
02:23at Yale and at Fox Chase and then back at,
02:25at Yale again.
02:26So we're thrilled to have you and
02:29we look forward to your talk.
02:38So it was not my misfortune to
02:39have run across Eric as an intern.
02:41He was the most beloved and the most
02:45inspiring chief resident and actually
02:48leader I think I've ever worked with.
02:50And so it's, it's wonderful that he's back.
02:53So I'm going to talk about synthetic
02:55lethal therapy for head and neck cancer,
02:57but I'm going to start out by
02:59just laying a little groundwork.
03:01So why am I interested in HPV negative
03:03head and neck cancer specifically?
03:05And why in the age of immunotherapy
03:07am I still studying targeted therapy?
03:10So just as as sort of some background,
03:13head and neck cancers,
03:14the 7th most common cancer globally,
03:17nearly half a million people die of
03:20this cancer worldwide every year.
03:22For HPV negative cancer,
03:23which globally is the most common form,
03:26the risk factors are age,
03:27nutritional status, alcohol exposure,
03:30tobacco, other oral carcinogen exposure.
03:34And then we have particularly in North
03:38America and in Europe a rising burden
03:41of HPV associated head and neck cancer.
03:43For both kinds,
03:44the definitive therapy whether it's
03:46surgical or radiation or multi modality
03:49can lead to lifelong morbidity And
03:51both geographically across the Gulf
03:54globe and within the United States,
03:56this is a disease where we
03:58see enormous disparities.
04:02So treatment resistance is a big
04:03problem in head and neck cancer.
04:05We know for multi modality therapy of HPV
04:08associated cancer that we cure about 80%
04:11and in HPV negative we cure about 50%.
04:14And for some years,
04:16we thought that we were curing
04:17more people by intensifying our
04:19chemo radiation approaches.
04:21But it turned out what was happening was
04:22we were seeing a rising incidence of
04:24HPV associated cancer and we hadn't yet
04:26figured out that that had a better prognosis.
04:28So for HPV negative cancers,
04:32the cure rate for definitive
04:34therapy has not really budged.
04:36Historically,
04:36we knew that combination chemotherapy
04:38was active but not curative.
04:40And actually the standard regimen
04:42was something that had Yale roots.
04:44We now know and I'll show you
04:45a little bit about this,
04:46that immunotherapy can provide
04:47long term survival benefit for
04:49a proportion of patients with
04:51recurrent metastatic disease.
04:53And very frustratingly,
04:544 large phase three studies later,
04:56we still do not know how to integrate
05:00immunotherapy with curative therapy.
05:02And so we we have a substantial
05:04number of patients who don't benefit
05:07from our existing therapies.
05:09And This is why we've been sort
05:11of interested in targeted therapy.
05:13But advances in targeted therapy
05:15and head neck cancer have been slow.
05:17And I think that part of this is due to
05:20the fact that the predominant genomic
05:22drivers in HPV negative cancer are
05:25really loss of tumor suppressor function,
05:27not not necessarily that easily targetable.
05:30So what's the evidence that HPV associated
05:33cancer's really a different disease?
05:35The the first observation I think
05:36came from a small ECOG trial,
05:38but this is from a large RTOG study
05:41showing that with the same treatment,
05:43HPV associated cancer has a
05:45much higher cure rate than
05:47HPV negative cancer.
05:49And we looked at that our committee
05:51at ECOG looked at this graph and said
05:53these are two different diseases and
05:55we're treating everybody in the top
05:57curve with a treatment that was designed
05:59for the people on the bottom curve.
06:00Maybe we could de intensify and we did
06:03the first actually trial of treatment
06:05the intensification in head neck cancer.
06:07We reasoned that responsiveness to
06:09chemotherapy would be a hallmark
06:11of the the favourable prognosis
06:14HPV associated cancers,
06:15that it would be a hallmark
06:17of radiosensitivity.
06:17And so we gave everybody induction
06:20chemotherapy and those who responded got
06:22a reduced dose of radiation and those
06:24are the patients in the dark curve.
06:26So you can see that despite
06:28getting less radiation,
06:29they had a higher cure rate.
06:30And we also from this dissected out some
06:33clinical risk factors for poor behavior.
06:35So bulky disease and more lifetime
06:38tobacco exposure of more than
06:4110 pack years were unfavorable.
06:44And this work has subsequently been been
06:46confirmed in a number of other studies.
06:48We also wanted to know if you could
06:50de intensify post operative therapy.
06:51And so this is a large study that
06:53we did in ECOG for patients who
06:56were undergoing robotic surgery
06:57for their head and neck cancers.
06:59The all the surgeons were credentialed,
07:02everybody had good quality surgery.
07:04And then based on the pathology we
07:06assigned them to a risk category.
07:08And if they were in the intermediate
07:10risk category,
07:10we tried to reduce their radiation.
07:12And I just presented at ASCO last two
07:14weeks ago whenever that was the 54
07:17month PFS for the strategy as a whole,
07:19it was 90% and for reducing
07:23radiation from 60 to 50 Gray,
07:26there was no significant difference
07:28in progression free survival.
07:30Interestingly, in this study,
07:32having a history of tobacco
07:35exposure was not deleterious.
07:36And so this kind of highlights
07:38the fact that relying on clinical
07:42criteria like what's your T stage,
07:45how many pack years of smoking do you
07:47tell your treatment team that you've
07:49you've had is a little bit rough.
07:53And so a major focus of one of the
07:56projects in our SPORE or Head Neck
07:59Sport grant that we have is to try
08:02to figure out whether or not we
08:04could get a molecular signature
08:05for more favourable prognosis.
08:07So this is work of Del Yarbrough
08:10and Natalia Isayeva at UNC working
08:12together with a young surgeon,
08:14scientist Travis Schrank,
08:15who's had a lot of support from the spore.
08:18And So what they began with the
08:21observation that the T3 CYLD mutated
08:23subset of head and neck cancer
08:25had a more favorable prognosis.
08:27These patients were less likely
08:28to be smokers.
08:29The HPV was more likely to be
08:31episomal rather than integrated.
08:33The cancers were less methylated,
08:36had a lower APOBEC mutational signature and
08:40expressed a lower level of HPV oncogenes.
08:43And so this LED Dell's lab to
08:46explore demethylation therapy
08:47and a lot of preclinical models.
08:49I won't show you that,
08:50but this has LED us to two clinical trials.
08:53First, a small pilot trial of five
08:56asicitidine monotherapy before
08:57patients went to the OR for HPV
08:59associated head neck cancer.
09:01These the photo photo micrographs on the
09:06on the left showing immunofluorescence
09:08for CD4 and CD8 cells or the from
09:11David Rim's group and on the left
09:13you see two biopsies pre therapy.
09:16On the right you see two biopsies
09:18post therapy.
09:18I think you can see the extensive
09:21infiltration of CD4 and CD8 cells
09:23as well as the formation of some
09:26tertiary lymphoid structures.
09:27So we then took this forward with
09:30a a new randomized trial which is
09:33currently accruing 5 azacitidine
09:35alone immune checkpoint inhibitor with
09:37nivolumab alone or the combination.
09:39And we're seeing very dramatic
09:42deep pathologic responses in
09:44the combination patients.
09:46I I'm showing you here some pictures from
09:48a patient who had a 3.5 centimeter tumor,
09:5017 days later went to the OR had a 2
09:541/2 millimeter tumor with extensive,
09:56as you can see their cholesterol clefts,
09:58evidence of immune mediated cell death
10:00and very extensive T cell infiltration.
10:03The brown stain for P-16 shows
10:06the residual cancer.
10:08So really very few viable cancer
10:11cells left and we are intensively
10:14interrogating the tissue from from this
10:17trial as part of our SPORE project.
10:19So I just also wanted to say a word
10:23about immunotherapy because I think it's,
10:25although it's a genuine breakthrough
10:27for patients with head and neck cancer,
10:29it's because of how disappointing
10:30it's been that we still need to
10:32think about other ways forward.
10:33So this is from Keynote 01/2.
10:35This was the first head and neck cancer
10:38specific study of immune checkpoint
10:39inhibition. Yale was part of it.
10:41I led the head neck cohort when
10:43I was at Fox Chase.
10:45We accrued both HPV positive
10:47and HPV negative patients.
10:49You can see that there were responses
10:52for for both groups and some
10:54responses were deep and durable,
10:57but not all.
10:58And the overall response rate was about 17%.
11:01This actually led to the FDA approval of
11:03pembrolizumab for head and neck cancer.
11:05But the real,
11:06and this was performed in people
11:07who were all platinum refractory,
11:09but the real question was whether
11:11we could get this into first line.
11:13And this was problematic 'cause we
11:16were taking a 17% drug and we were
11:18putting it up against a combination
11:20of chemotherapy and cetuximab.
11:22As I said,
11:23that was a regimen we had developed
11:25years ago when I was here at Yale
11:27then had a 35% response rate.
11:29But there were data from both from
11:33Keynote 01/2 and from another
11:35platinum refractory study led by
11:37Ezra Cohen that had shown that
11:39responsiveness to pembrolizumab in
11:41head neck cancer correlated very
11:43well with expression of PDL 1,
11:45the ligand for PD1 on tumor and immune cells.
11:48So this we sort of had two plans for
11:52how we could dissect out who was
11:54going to benefit from pembrolizumab.
11:56One was to do hierarchical testing
11:58for pembrolizumab monotherapy against
12:00the chemotherapy setuximab combination.
12:02So we started looking at
12:04those patients with CPS 20.
12:05If that was positive,
12:06we could look at CPS one and then we
12:08could look at the total population.
12:10And indeed pembrolizumab with its
12:12puny 17% response rate led to better
12:15overall survival than chemotherapy
12:17and setuximab in anybody who had PDL
12:20one expression and pembrolizumab plus.
12:22And the other strategy we had was
12:24this pembrolizumab plus chemotherapy
12:26regimen that had a very similar response
12:29rate to chemotherapy plus setuximab,
12:31but LED to better overall survival
12:34for the entire population.
12:35The, the way that hierarchical testing works,
12:38you, you look at all the people with
12:40ACPS 20 and then the next analysis you
12:42look at everybody with the CPS one.
12:44So you're including the people
12:45you already analyzed.
12:46I think when I was in medical school,
12:48the New England Journal didn't
12:49take papers that did that,
12:50but now it's pretty common.
12:51But anyway,
12:52we wanted to go back and look and you
12:54can see that if we break down the
12:56PDL 1 low expressors into those with
12:58no expression or those with one to
13:0119 expression, there's no benefit.
13:03In fact,
13:03there's probably a disadvantage to
13:05using pembrolizumab monotherapy in
13:07the patients who are not PDL 1 expressing,
13:10whereas for the one to nineteens
13:13there's a benefit particularly
13:15for pembrolizumab chemotherapy.
13:17So,
13:18but I think what I'm showing you across
13:20these is that at the end of five years,
13:23only 20% of people are still alive
13:25who've had immune checkpoint inhibitor
13:26based therapy with head and neck cancer.
13:29So what are we going to do for the others?
13:30So this is what has taken me back to,
13:34to targeted therapy.
13:35I I think one very interesting
13:37observation about head and neck
13:38cancer is that we have three sort
13:40of main buckets of how people
13:41*** **** and neck cancer.
13:43I mentioned environmental exposures,
13:45I mentioned HPV.
13:46We're also recognizing that in adult
13:49survivors of Franconia anemia,
13:51whether or not they've had
13:52bone marrow transplantation,
13:53there's a astronomical rate of
13:56head neck cancer about 1000 fold.
13:59What, what you would expect.
14:01And if you look at the genome
14:03profiles of these,
14:04these three separate buckets, the same,
14:07the same pathways are targeted.
14:08So you've lost AP 53 function,
14:10whether it's a viral oncoprotein degrading
14:13P53 or it's a loss of function mutation,
14:15you have CDKN 2A mutation or loss PIC
14:18three CA mutation and amplification.
14:20And we're starting to recognize
14:21that FANK genes can be mutated
14:23in sporadic head and neck cancer.
14:25And so the,
14:26I think the question for me was
14:28do these insights point to shared
14:31vulnerabilities and that that would
14:34point us towards synthetic lethal
14:36strategy centered on P53 loss of function.
14:39So this is from over 500 patients
14:41profiled by the broad.
14:43We could also look at TCGA.
14:45We could also look at Karas or Foundation.
14:48This has been published a number
14:50of Times Now,
14:51but about 85% of HPV negative
14:54head neck cancers have
14:56pathogenic mutations in P53.
14:59Direct restoration of this function is,
15:02as you probably all know has not
15:04translated to the clinic yet.
15:06Loss of P53 is associated with loss
15:10of the G1 SDNA damage checkpoint
15:14making these cells much more reliant
15:18on G2 M and and also leading
15:20to accumulation of DNA damage,
15:22which might indicate the role
15:25for immunotherapy that I was
15:27just telling you about.
15:28So a long time ago in 1993,
15:30ECOG started a study of assessing P53
15:34status in optimally resected patients.
15:37So this was over 500 patients,
15:40all of them had margin negative surgery,
15:42all of them got contemporary best
15:46post operative management and then
15:48their P53 status was catalogue P53.
15:52Disruptive mutation in this study was
15:54defined as either a mutation in the DNA
15:57binding domain or a truncation mutation.
15:59And you can see that if you
16:01had a disruptive mutation,
16:02your hazard ratio for death was 1.7.
16:04And this actually held even
16:06for stage 1 patients.
16:08So we wanted to look at whether or not we
16:12could intensify therapy for these patients.
16:15But the first question was this
16:17rule that had been designed for
16:20the study in 1993 for what was
16:22a significant P53 mutation,
16:24was that really the best?
16:25And in the meantime,
16:26a number of computational and experimental
16:29algorithms had come out for calling
16:31the significance of P53 mutation.
16:33So we looked at all of these and compared
16:36them to the rules in our in our prior paper,
16:38which were called the POETA rules.
16:40And although the POETA rules outperformed
16:43all the other algorithms that we looked at,
16:46we could make them better by adding
16:48information about splice variants.
16:50And so this has gone forward as an
16:52integral biomarker actually for one
16:54of our current adjuvant trials.
16:56I'm working together with Erica Golemis,
16:59my long term collaborator at Fox
17:01Chase as well as Karis Biosciences.
17:05We've looked at over a 1000 HPV
17:09negative head neck cancers profiled
17:11in their system to see to,
17:13to reassure ourselves that these
17:15P53 mutations are actually having
17:17an impact on DNA repair.
17:19And so we looked at the proportion
17:21of patients that had a tumor mutation
17:24burden of greater than 15 per mega base.
17:27And we looked at P53 mutation,
17:29any CDK into a abnormality,
17:31some mutation or loss and then Co
17:33occurrence of those two mutations,
17:35which is actually a pretty common
17:37thing in head and neck cancer.
17:38And depending on how we called
17:40the P53 mutation,
17:43there was a different level of significance.
17:45But with the exception of the
17:47gain of function mutations,
17:48all the P53 mutation calls showed
17:51a significant increase in TMB,
17:52all the CDKN 2A mutation calls did.
17:55And if you saw the two together,
17:57which is as I mentioned quite common,
18:00that was the most dramatically significant.
18:02So we so we do think that this is
18:04evidence that P53 mutation leads
18:08to genomic instability.
18:10So we then wanted to explore what
18:12are the target,
18:13what are the the regulators of
18:17G2M that for which
18:21molecular therapies might be
18:23available or becoming available.
18:24And we focused in on mitotic
18:28entry kinase Aurora A and
18:30mitotic checkpoint kinase rewind.
18:33So Aurora A is a serine,
18:35the serine threonine kinase.
18:37Its canonical roles are in maturation
18:40of the spindle assembly complex and
18:43it's known to be regulated by P53.
18:46Its content is also controlled by a couple
18:48of binding partners that stabilize it,
18:51TPX 2 and Ned 9.
18:54And then we one is a mitotic
18:56checkpoint kinase and it exerts its
18:59activity by putting an inhibitory
19:01phosphorylation on CDK one.
19:03And for Aurora A,
19:04when we started this work,
19:06there was a, a drug called alisertib
19:08that had been in development.
19:10And for WE one,
19:11the lead compound for a long
19:13time had been adavasertib and,
19:15and we were able to to work
19:17with both of those.
19:18So I mentioned that P50
19:21threes regulates Aurora.
19:22So when you have loss of P53 you
19:25have increased Aurora levels.
19:27And we,
19:29we were able to develop an Aqua assay
19:33using insight to fluorescence to
19:38localize Aurora within the tumor compartment,
19:42which was labeled green for cytokeratin.
19:45And we did this on ATMA that
19:47we assembled at at Fox Chase.
19:49I don't know if you can see
19:50it with the lights on here,
19:51but the, the red,
19:53which is the Aurora kinase in,
19:55in these in the high expressors
19:57on top is localized to the nucleus.
19:59And that was something that we
20:01initially found very confusing.
20:02But hold that thought because I'm
20:04going to talk more about it.
20:06This is work from my lab here now
20:08looking at a number of either P53
20:11mutated or P53 null workhorse HPV
20:14negative cell lines that people
20:16use to study head neck cancer and
20:19all of them overexpress Aurora
20:21normal tracheobronchial epithelial
20:23cells and fibroblasts do not.
20:26And then we obtained as a kind gift
20:29from Jeff Myers at MD Anderson
20:31some isagenic cells for P53 either
20:33null gain of function or loss
20:36of function and looked at all
20:38of those regulators of G2M that
20:40that were potentially targetable.
20:42And as you can see Aurora A expression
20:45was far more abnormal than any
20:48of the other potential targets.
20:50This shows you the the survival
20:53curves from our Fox chase TMA.
20:55So for the cases that were HPV negative
20:59and overexpressed nuclear Aurora,
21:01a far worse survival,
21:05very statistically significant and
21:07that was independent of what kind of
21:10therapy the patients had received.
21:12At the same time,
21:12we were doing that work to explore
21:14whether or not Aurora would be a good target.
21:16The late Eddie Mendez's group at the
21:18Hutch had done a functional kinome
21:20analysis looking for potential
21:21hits in head neck cancer.
21:23Actually, if you read the paper,
21:24Aurora was a very strong hit in that paper.
21:26But the,
21:27the one that they pursued was we won
21:29and they were able to demonstrate
21:31that it enabled G2M arrest in P53
21:34null cells after DNA damage in the
21:36model they used with cisplatin.
21:38So that was around the time that
21:43my lamp here at, at Yale got going.
21:45And so we explored the combination
21:48of Aurora and we won inhibition.
21:50And this is very predominantly the work
21:53of John Woo Lee, who's here in the room.
21:56And I'll, I'll be mentioning
21:58his name over and over again,
22:00I think as well as some of the
22:02other great people in the lab.
22:03So anyway, the first thing that we
22:05did was we compared a davasortib,
22:08which is the wee one inhibitor,
22:09alasortib, which is the Aurora
22:11inhibitor of the combination.
22:12You can see that with the davasortib,
22:14as you might expect,
22:15if you're losing the mitotic checkpoint,
22:18there's a little bit more
22:20chromatin disaggregation,
22:21so a little bit more early mitosis alacertib,
22:25which effects spindle assembly.
22:27We saw these multipolar spindle
22:30cells and when we put them together,
22:32we saw micronuclei cell desegregation
22:36consistent with mitotic catastrophe.
22:38And Janaki Parameshwaran,
22:40who was one of the fellows in the lab,
22:42counted thousands of,
22:44of mitotic figures and demonstrated
22:46that with combination therapy,
22:48we essentially did not see normal mitosis.
22:52They we also used a,
22:54a flow for a Nexon 5 showing
22:56increase in the proportion of
22:58cells that were apoptotic as well
23:00as an increase in cleave part.
23:03So we started to try to dissect
23:05out how this was happening.
23:07And I,
23:08I'm going to try to walk you
23:10through this western which
23:11maybe is a little bit busy,
23:12but you can see either a Daviser to Balone,
23:15Aliser to Balone or combination.
23:17And if you focus on CDK one,
23:20which is that inhibitory,
23:21the cell cycle inhibitory phosphorylation
23:24that's mediated by WE one,
23:25you can see that after treatment with
23:28alacertib either at 8 hours or 24 hours,
23:30that is markedly up regulated.
23:32And if you add the WE one inhibitor
23:35that CDK one phosphorylation goes away.
23:40We validated this in animal models
23:43whether used high or low dose adavacertib.
23:46This was highly synergistic with
23:49significant improvement in animal survival.
23:51There weren't really any
23:53significant laboratory abnormalities
23:55except for mild cytopenia with
23:58adavasir TIB at the higher dose.
24:00We harvested tumor from those animals
24:03and looked at this under the microscope
24:06both for proliferation with Ki 67 which
24:09was markedly decreased with combination
24:11as well as cleaved castpase which was up.
24:14And then we used Aqua again to look for
24:18phospho CDK one in the tumor leading edge.
24:21So you can see here that the amount of of
24:25phospho CDK one goes down dramatically
24:28with combination therapy relative
24:31to either Aurora or untreated cells.
24:33And I don't know if again with the light,
24:37if you can appreciate,
24:38but you can also start to see some,
24:40some multi nucleated giant cells
24:42there in the combination arm.
24:44So around that time,
24:46Al Assertive disappeared from
24:47clinical development.
24:48Intriguingly, it's back now.
24:50So Puma Pharmaceuticals has picked it up,
24:53but we went looking for a second
24:56generation Aurora inhibitor and
24:58landed on what's now called Vic 1911.
25:01This is more strongly selective
25:03for Aurora A over Aurora B than
25:06the parent compound hip,
25:08good preclinical effectiveness and
25:09has been in phase one testing both
25:12his monotherapy and together with
25:14paclitaxel where it was found to
25:17be both tolerable and effective.
25:19So a lot of the,
25:20the next work that I'm going to do
25:22that I'm going to show you is was done
25:24with the second generation inhibitor.
25:26So we looked across head,
25:28neck cancer, lung cancer,
25:30pancreatic cancer cell lines and 12 of
25:33the 13 cell lines that we looked at,
25:35we saw a significant synergy and
25:38good effects on clonagenic survival.
25:40I'll talk to you a little bit
25:43later about SCC 61 and what may
25:45be different about that.
25:49We were able to demonstrate that in
25:53platinum resistance cells the effect
25:55of Aurora inhibition is maintained
25:57including in conogenic survival assays.
26:01And this just shows you the the
26:05unfocal microscopy to emphasize the
26:07the reliability with which we develop.
26:10We see these multipolar spindle
26:12cells forming with Aurora inhibition,
26:14so suggesting that the canonical
26:17activity on spindle assembly
26:19complex is is being achieved.
26:21And then the micronuclei that we
26:23see during mitotic catastrophe
26:24and using live cell imaging,
26:26you can see that with combination
26:28the micronuclei start to appear
26:30by about 27 hours.
26:34There's dose dependent effects and
26:36xenograft and you can see that at the
26:39highest doses which are doses that have
26:41been used in preclinical studies for
26:43monotherapy development of these agents.
26:45Actually the we see
26:48outstanding tumor control.
26:50There's no effect on normal,
26:53normal organs of these animals.
26:58We've developed a couple of
27:00HPV negative PDXS here at Yale.
27:02This is one that interestingly
27:04does not have AP 53 mutation but
27:07has loss of FANK A and MRE 11.
27:10And here you can see outstanding
27:13synergy of combination therapy.
27:15No effects on animal body weight.
27:17We've also taken this forward
27:19in P53 mutant lung cancer,
27:21where we see very similar effects
27:23both in xenografts and PDXS.
27:25And thanks to Katie Politi
27:27for sharing the PDXS.
27:30We've wanted to see whether or not our
27:33original hypothesis about P53 was holding.
27:35And so you can see that we do see
27:37synergy across a wild type and,
27:39and both loss of function and
27:41and gain of function mutations.
27:43But the effect is most strong with P53
27:47null or loss of function mutations.
27:49And so this led to a model that we
27:52that we had of how this was working
27:55in that we gave Aurora inhibition.
27:57This led to the expected effects on
28:00mitotic entry with the multipolar
28:03spindle formation.
28:04But that this led to cell cycle
28:06arrest and that we could abrogate
28:08that cell cycle arrest by giving a
28:11Wee 1 inhibitor and the cell with
28:12its four poles could not pull into
28:15two normal daughter cells when
28:17it was accelerated into mitosis.
28:20And this led to apoptosis and
28:22mitotic catastrophe.
28:23The only thing that was not 100% clear
28:26about this story was why we won.
28:29Was there to be inhibited in the 1st place?
28:31And I had shown you the data from
28:33Eddie Mendez in that functional kinome
28:35analysis that suggested that if you
28:37enhanced replication stress with platinum,
28:39maybe you would see we won.
28:41We hypothesized maybe it was hypoxia
28:44or the P53 mutation background was
28:47leading to elevated replication stress.
28:50I'm not gonna show you this,
28:51but the to the limited extent
28:53that we've tried replicating
28:54this under hypoxic conditions,
28:56it doesn't seem that much more dramatic.
29:00And so around this time we read a paper
29:02that indicated that there were some
29:05non canonical activities of Aurora A
29:08that perhaps we were also inhibiting
29:12and that these might explain the we
29:15one dependency that we were creating.
29:17So it was reported that TPX 2 bound
29:23to the DNA repair protein 53BP1 and
29:28that once Aurora was recruited,
29:31that complex then recruited BRCA 1 and
29:34that this that this complex ultimately
29:38protected stalled replication,
29:39replication forks from MRE 11 degradation.
29:43So we started to look at replication
29:46stress as a potential outcome
29:48of of these inhibitors.
29:49I don't know if people are familiar
29:51with this DNA fiber spreading assay,
29:54but you give a a red
29:56labeled chlorideoxyuridine.
29:57You then chase that with a
29:59green labeled IO deoxyuridine.
30:01If the replication force is progressing,
30:04you'll get a strip of red
30:05followed by a strip of green.
30:07But if it's not, you'll just see the
30:09red because the forks have arrested.
30:11And so as you can see here
30:15with either monotherapy,
30:16we see a little bit of decrease in the
30:19amount of green that gets incorporated.
30:21And then with combination, we see
30:23almost no replication fork progression.
30:25And this is, there's,
30:27there's a software out there
30:28for counting this week.
30:30We've heard that people are
30:31are not very happy with it.
30:32A a lot of this work was done
30:34by people in the lab.
30:36Julian Barantes, a really talented
30:38Yale undergrad named Jackie,
30:40she and Pratima Charasia.
30:41And so we looked at this across
30:44a number of different cell lines
30:46and we see this amplification of
30:49replication stress is measured
30:51by reduced fork progression in
30:53all of our sensitive models and
30:54we don't see it in SCC 61.
30:56And so we've been trying to figure
30:58out what's different about SCC 61.
31:00Just looking at the literature on
31:02this cell line it it is very highly
31:05expressing of Rand 54 L and DSCC
31:07one which may indicate just a a
31:09better background for DNA repair,
31:11but we have a lot more work
31:13to to figure that out.
31:15We've also looked at proximity ligand
31:17assays for our PCNA and RNA Pol
31:212 and accumulation of replication
31:24protein A at sites of replication
31:27for extolling in the we one field.
31:31The lead compound at DAVA Certib
31:33had was found to have a lot
31:35of toxicity when it went into
31:37combination with PARP inhibitors
31:38and ultimately has been withdrawn.
31:41There are a number of second generation
31:43inhibitors that seem to have much cleaner
31:45toxicity profiles moving forward.
31:46We've been working with a a
31:48company named Zentalis that's
31:52developing a Zen assertib and this
31:54has been moving forward predominantly
31:55I think in GYN malignancies.
31:57So far they've reported that
31:59cyclin E expression can be a
32:01favorable biomarker for this.
32:03And this is certainly something that
32:05was also touted for Adavasor Tib.
32:08But what we think we,
32:10we may not need to use that for this
32:12combination because we think that we're
32:14inducing it with the replication stress
32:16that the Aurora inhibitor is inducing.
32:19So here you see cells that were synchronized.
32:23They after they were released
32:26from synchronization,
32:27cyclone goes up in all of them,
32:28then it comes down in all of them.
32:30But if they're treated with the Aurora
32:32inhibitor alone or in combination,
32:34cyclone E is is up by 4 hours.
32:38We've also looked at RNA seek and this will
32:40come back around at the end of the talk.
32:42But we do see a compensatory up
32:45regulation of PLK one in these
32:48cells after Aurora inhibition.
32:49And then if,
32:50if we look sort of in a pathway
32:54process pathway way at what's
32:56happening with the RNA seek,
32:57we see significant focus both
33:00on mitotic cell processes as
33:02well as on replication stress.
33:05So the new model that we have
33:07incorporates the old model.
33:09So the allocer Tib leads to
33:11the weird spindle formation.
33:13We one inhibitor allows this
33:16incompetent cell to enter mitosis
33:18and and lead to mitotic catastrophe.
33:21But we've supplemented this now
33:23with the with the notion that
33:26amplifying replication stress with
33:28Aurora inhibition is necessary to
33:31creating this we one dependency.
33:34This is the the new agent that
33:36that we're working with.
33:37As I mentioned a Zen assertib,
33:39it's been through phase one both
33:41in monotherapy and is tolerable in
33:43combination with a large number of
33:46conventionally cytotoxic agents.
33:48And we've been able to show that we
33:52see very similar synergy with a Zen
33:55assertib as we did with a DAVA certib.
34:01Yes, OK. So that, that that part of
34:05the talk sort of encompasses why
34:07we've gotten so interested in using
34:09this Aurora we won combination.
34:11And I'll, I'll talk to you a
34:13little bit later about how we've
34:14taken this forward to combo match.
34:16Another way in which Aurora has
34:19become very interesting is for
34:20its role in adaptive resistance.
34:22And I think many of you may
34:24know Katie Politi's work on this
34:26in ASA Mertonib resistance.
34:27But so we wanted to ask whether
34:30coordinated inhibition of these
34:32mitotic entry and mitotic checkpoint
34:34kinases would also be effective
34:36in resistance that's in adaptive
34:38resistance that's Aurora driven.
34:40And I'm gonna tell you both one
34:42story where it looks like that's
34:43true and then a countervailing
34:45example where it doesn't look like
34:46it's true and why I think that is
34:50so. And and I should say that I'm
34:55extremely grateful to the Yale Sporin
34:57lung cancer for the pilot funds
34:59that allowed us to do this project,
35:01this part of the project.
35:02So people probably just spent a
35:05couple days at West Campus learning
35:08a lot about lung cancer and you'll
35:10be familiar that we now have
35:12direct inhibitors of K Ras G12C,
35:14which is an important driver
35:15in non small cell lung cancer.
35:17And that although these agents are
35:20very clearly highly active with
35:22response rates in the 40 to 50% range,
35:25they have median progression free
35:27survivals of only six or seven
35:29months related to both
35:34adaptive resistance and mutational
35:36resistance that develops And the
35:39adaptive resistance is Aurora dependent
35:41and goes through map kinase signaling.
35:44And this was very beautifully worked
35:47out by Pure Alito's group in in a
35:50publication a few years ago in nature.
35:52And so we kind of perked up because we're
35:56interested in Aurora and understood that
35:58Aurora is a key effector of KRAS signaling.
36:02It phosphorylates Ralgef and Ral A and
36:06this enhances KRAS driven to moragenesis.
36:10And so we and and in the original
36:13Purelita paper,
36:14they had shown that knockdown of
36:17Aurora A abrogated that effect.
36:20So we looked at the combination of an an
36:24Aurora inhibitor and sodorasib in K Ras
36:27G12C driven non small cell lung cancer cells.
36:31And I'm just going to talk you through the
36:34middle part of I don't know if you're able
36:37to see my not really able to see this.
36:39Are you all right talking about
36:45this? So we created a a quiescent cell
36:49reporter by M Cherry labeling of P27
36:53and exposed cells either to the Aurora
36:56inhibitor sotorasib or the combination.
36:58And if you focus on that little blue
37:02hump in the sotorasib treated cells,
37:06those are cells that are non quiescent
37:08that are escaping from sotorasib and
37:10that's through and you can see the
37:14phospho URC that goes that
37:16goes up as that happens.
37:17And we're able to abrogate that
37:19by adding an Aurora inhibitor.
37:21And so we've looked at synergy.
37:24We do see synergy with Aurora
37:26inhibition and sotorasib in these,
37:28in these models. If we look at
37:36the combination of Aurora and we one
37:39inhibition, we see again the very similar
37:42phenotypes with multipolar spindle
37:43cells when we give an Aurora inhibitor
37:46mitotic catastrophe when we give the
37:49combination again very few normal
37:51mitotic cells after combination therapy.
37:53And so we see this as oh,
37:56and I forgot to tell you that we've
37:59also confirmed this in animal models,
38:00which is at the far right.
38:02You can see the,
38:04the blue line is the combination therapy.
38:06And even after three cycles of therapy,
38:09these cells are or these tumors
38:11are still sensitive.
38:12So we have taken this to combo match.
38:16So for those of you who are not
38:18in the clinical trial field,
38:20when you're doing work with novel
38:23targeted therapies where you may be
38:26working with small drug companies
38:27and they have one of the drugs that
38:29you want to use in your combination
38:31but not the other.
38:32It's historically been very difficult
38:35to conduct trials of combinations.
38:37And this has been a a major barrier
38:40in the field of,
38:41of synthetic lethal therapies.
38:44And so the NCI working together with
38:47the cooperative groups has started
38:49this mechanism called combo match
38:51where they will do single trial Cradas
38:53with the companies for that drug
38:55and provide support for doing a a
38:59phase two trial with the combination.
39:02So we took this concept to combo
39:05match proposing a combination of as
39:08an assertive with BIC 1911 with a
39:11primary endpoint of objective response rate,
39:14secondary endpoints of progression
39:15free survival,
39:16toxicity and duration of response.
39:19All of the combo match trials
39:21require biopsy and you have the
39:23opportunity on baseline material
39:24to do extensive correlatives.
39:26And we propose P53 mutation class comb
39:29mutation markers of replication stress.
39:32And in work that comes from Faye
39:33Johnson's group at MD Anderson and she's
39:36a partner with me on this proposal,
39:38IHC for retinoblastoma.
39:40We've met with the NCI combo
39:44match statistical team and the
39:47agreement is for three cohorts of
39:50P53 mutated solid tumor patients.
39:52Firstly,
39:52recurrent metastatic head neck cancer
39:55that's platinum and PD1 inhibitor,
39:57refractory or ineligible non small
40:00cell lung cancer that's progressed
40:02on two prior lines of therapy
40:04that's P53 mutated including
40:07KRAS G12C that's post inhibitor.
40:10And we wrote sotorasib and adagrasib,
40:12but I'm learning that there are
40:14many new drugs coming in this field
40:15and then other solid tumors that
40:17are P53 mutated and have progressed
40:19on two prior lines of therapy.
40:23The experience with the NCI statistical
40:26folks was interesting and we learned
40:29that it's very necessary to have a
40:31futility analysis for a trial like this.
40:33And so we have generated one of
40:35those and they are also recommending
40:37that we have a rapid phase one,
40:39which there's a mechanism for
40:41doing in the ETCTN.
40:44OK,
40:45so that the the K Ras story was
40:48my example of where I think Aurora
40:50and we won seems to work in the
40:53face of adaptive resistance.
40:55And now I'm going to tell you a story that
40:57where it doesn't seem to work.
40:58And this is work from Seb
41:00Cruz Gomez in the lab.
41:02And this isn't an HPV positive model.
41:05And I think you might remember
41:06from the beginning of the talk,
41:07I said P53 is disrupted in head neck cancer,
41:11but it's because the viral oncoproteins
41:14are degrading a wild type P53 and
41:17in response to replication stress,
41:18you can actually see that
41:20these cells up regulate P53.
41:22And so I think that that maybe is
41:24why we're not seeing the effect here.
41:27But So what Seb has has been working
41:30on is the Aurora mediated adaptive
41:33resistance to pan her inhibitors.
41:36And so I'll,
41:37I'll start by summarizing a lot of
41:39background and say that EGFR inhibitors
41:42in general have not appeared very
41:44active in HPV positive cancer.
41:46And this was initially believed to
41:48be because of PIC 3 CA mutation or
41:51because there wasn't a lot of EGFR around.
41:54I think that that's probably not true,
41:56but a recent publication from Jenny
41:58Grandis and Silvio Gookin showed
42:00that the viral onchoproteins actually
42:02drive expression of her three.
42:04And Seb is now also seeing that
42:06there's her two expressed here.
42:08So he's moved from using a conventional
42:11EGFR inhibitor like erlatinib
42:13to pan her inhibitors.
42:16One of the other projects in our
42:18spore has to do with optimizing pan
42:22Hearn inhibitors for head neck cancer
42:24based on Mark Lemons observation
42:26that the conformational specificity
42:27of these agents is very important
42:30and we've taken a hint from him and
42:32moving from a fat NIB to decamitnib.
42:34And as I think you can see here,
42:36there's strong synergy for decamitnib with
42:40our second generation Aurora inhibitor.
42:43What mediates this here we think
42:46is an EMT like process.
42:48And I think you can appreciate
42:51here the up regulation of Ned 9,
42:53whether we're using a fat nib
42:56or a dacomitinib.
42:57This has been confirmed
42:59in our RNAC experiments.
43:02He's able to show marked synergy for
43:05these agents in in xenograft models.
43:08And the the way this experiment goes
43:11is he treats the the mice for 19
43:13days and then he sacrifices them.
43:15But he kept three alive in each
43:18group and we're now out to about
43:20six weeks and no tumor has regrown
43:22in the combination animals.
43:24So we think that that's pretty,
43:25pretty interesting
43:29in terms of what potentially is driving
43:33resistance to that combination.
43:35We do see pretty dramatic up regulation
43:37of PLK one in the treated cells.
43:40And part of our stand up to cancer
43:42grant has allowed us to collaborate
43:45with the Mandy Paulson group at Fred
43:47Hutch really expert proteomics group.
43:49And so they've been performing
43:52shotgun proteomics on on cell line
43:57experiments that Seb has done.
43:59And we see this very dramatic
44:01up up regulation of PLK 1 as a
44:04potential resistance mechanism.
44:06So he's taking this forward in
44:08combination with PLK 1 inhibitors
44:10with very strong activity.
44:12And this is LED us to go back to it
44:15also in our HPV negative models where
44:18we see pretty strong synergy for PLK 1
44:22inhibition with with we one inhibition.
44:26And so we're planning to take
44:28that forward into animal models.
44:29This work represents so many
44:32contributions from so many people.
44:34I, I first want to call out the
44:36amazing groups in the head.
44:38Next, four, We have three projects.
44:41One, as I mentioned on
44:45specificity for P and her
44:46inhibitors led by Joe Contessa,
44:48Kate Ferguson and Mark Mark Lemon.
44:51Our project that I've just been
44:52telling you all about with synthetic
44:54lethal therapy and the demethylation
44:56project that I alluded to earlier on,
44:59Arnala Stoppe is here.
45:00She's been our very capable
45:02administrator for the team.
45:04We have great cores and
45:07developmental programs.
45:08And then I particularly want
45:09to thank the people in my lab,
45:12Jong Muli, Pratyama, Chaurasia,
45:15Sebastian, Jackie Shee,
45:16Mei Jin and Dixon Adah.
45:19And then I also just have to call out
45:22our collaborators at at Fox Chase,
45:25the Hutch, MD Anderson and Georgetown.
45:27So I left some time for questions.
45:29Thank you,
45:37Roy.
45:40September in Bangkok, that was
45:42when Dixon joined us. You mentioned
45:48that you went to targeted therapy,
45:49but immunotherapy only worked at
45:515%. So the mechanisms of resistance
45:54and this type of paths are any different
45:55than what we see elsewhere. And
45:57would you be, is it,
45:57is it T cell exclusion? Is it
46:00regulatory cells?
46:00Do you have any sense on that?
46:02Yeah. So I, I, I think we've
46:07come back again and again to,
46:11and this is not for true for
46:12I think for the HPV positive
46:14where I'm extremely enthusiastic
46:15about the demethylation story.
46:17But in the HPV negative a lot
46:20of things have failed. So sting,
46:22sting agonists haven't gone anywhere,
46:24OX40 and it looks really like our best
46:30combination partner to date is Platinum.
46:32So we're trying to figure out,
46:35you know what's different about head neck
46:37cancer compared to even lung cancer,
46:38which genomically is so similar,
46:40right And the the tumor microenvironment,
46:45I think in head neck cancer it
46:47has a lot of hallmarks that are
46:49very hostile to immune cells.
46:51So we have very high concentration
46:54of cancer associated fibroblasts,
46:56lots of MDS,
46:57CS macrophages are predominantly M2
47:00and then the tumor is just in a very
47:04fibrotic background that's actually
47:05difficult for T cells to to migrate into.
47:09And so you know that the interest in
47:13platinum is both is it killing MDSCS and
47:15is it disrupting this architecture in
47:18a way that T cells can can migrate in better.
47:21So, so we are working with,
47:24we've had clinical trials here
47:27with Papinimab,
47:28which is a semaphore, a inhibitor.
47:33We have a,
47:34a umbrella trial coming from GSK with a
47:38couple of different combination partners.
47:40But to be honest,
47:42I think it's not so much that,
47:43you know,
47:44this is a very PDL 1 expressing cancer.
47:48I think the PDL 1 inhibitors work.
47:51I think it's that the the environment for the
47:53T cells when they get there is is different.
47:55And I think that that's why the
47:57Co stimulatory approaches have
47:58not been so successful.
47:59Kurt I signature and I was
48:07wondering if you saw any
48:11synergy between, you know,
48:13P1 inhibitor by visa in those.
48:17Yeah. So we've just actually started
48:19sending the samples off for methylation.
48:21We we don't have any data about that yet,
48:22but we've had the same thought Karen.
48:27Fantastic. You know, when
48:29you think of head and neck cancers
48:30that I'm not certainly an expert,
48:31but it seems like local regional
48:33therapies seem to work pretty well for how
48:37do you vision this word say translating
48:40to the clinical space and number one,
48:43the three cancerous lesions.
48:44So like do you think this will
48:46translate into that documentation?
48:48And 2nd, how do you think it translated
48:50around local regional therapy?
48:52Like you envisioned this being
48:53only for the metastatic side or do
48:55you think that these pathways are
48:57activated even if those that get
48:58cured or at least get local regional.
49:01Yeah. So I think in HPV negative disease,
49:06the, the possibility of doing a window
49:10trial to answer that question is,
49:12is quite reasonable.
49:15Just to go back to Roy's question,
49:17A a very interesting observation
49:19in head neck cancer is that
49:22neoadjuvant immunotherapy works
49:23much better than immunotherapy
49:25in the chemo radiation setting or
49:27the recurrent metastatic setting.
49:29And there's also a lot of interest
49:31and I see Mike Chiairazzi sitting
49:32here and and he's working on this in
49:36what what is it that our standard
49:39therapy is doing to tumor draining
49:40lymph nodes that might really be
49:42wiping out part of the immune response.
49:44So I actually am am expecting that we'll
49:47see a lot more use of immunotherapy
49:49in the new adjuvant setting.
49:52Why we haven't been able to incorporate
49:55immunotherapy with chemo radiation I think
49:57is a whole other long talk in and of itself.
50:00But clearly the fact that the studies
50:03were done with no patient selection
50:06and that the immunotherapy was started
50:08at the same time as the radiation,
50:10so all the T cells were migrating right
50:12into the radiation field probably were
50:14two major limitations of those studies.
50:16So I think we'll see a lot more
50:19neoadjuvant immunotherapy maybe
50:20in combination with chemotherapy,
50:22which seems to have a very high
50:23response rate.
50:24So I do think that I'm,
50:25I'm mostly thinking about
50:27this approach for recurrent
50:29metastatic disease and tumors like
50:36a small cell also activity
50:39of Alistair especially fit
50:44out. Do you have some cell lines?
50:45Yeah, we'd love to.
50:50Yeah, we have not done it,
50:51but that would be cool
50:59interest with the new agents.
51:00Yeah, OK, great.
51:04So why doesn't head neck cancer
51:07metastasize distant sites as much as
51:11it's because it it does.
51:14So there there's a, a really cool
51:17paper from the Chicago Consortium,
51:19University of Colorado and Northwestern,
51:22actually written by a former Yale
51:25fellow that shows that as your
51:27local regional therapy improves,
51:29the rate of distant metastases goes up.
51:32And so, you know,
51:34I think it's a matter of the fact that
51:37it's so prone to local regional recurrence,
51:40particularly when it's HPV
51:42negative that has LED that to.
51:44But if you look at the
51:45current immunotherapy trials,
51:46about 1/3 of the patients local regional,
51:48only a third distant Mets,
51:50a third both.
51:52We never have adequate power to
51:55address the distant Mets question
51:57in our chemo radiation studies.
51:59But
52:03right, that's totally related to your talk.
52:06But you did mention that that cancer,
52:07the disparities that exist and I
52:11think you have before with mentions.
52:15Can you say a few words about
52:17anything here? Actually,
52:25yeah. And and I'm sorry, I sort of
52:26let that part of Karen's question go.
52:28But so our surgeons go out in April,
52:31which is head neck cancer awareness
52:33month and they do a lot of free
52:36events where they screen people. The
52:41recruitment of Sarah pay to
52:43to the surgery department.
52:44She has a, a head grant from an IDCR looking
52:48at preclinical models of, of prevention.
52:52Curtis Pickering who joined us from
52:54MD Anderson about two years ago is,
52:57is funded in, in oral pre malignancy.
53:00So I think we, we are developing a program,
53:04I think Sarah's assembling a,
53:05a, a, a tissue bank.
53:08I think that that's something that
53:09that you can expect to see more sure
53:17in the graph.
53:28All of our models are P53 mutated.
53:35OK. Thank you.