Advances in Treatment of Hemophilia and Targeting dsRNA in Tumors to Overcome Immunotherapy Resistance

November 20, 2020

Information

Yale Cancer Center Grand Rounds | November 10, 2020

Robbert Bona, MD and Jeffrey Ishizuka, MD, DPhil

ID5916

To CiteDCA Citation Guide

00:00The hearts go out to his wife.
00:03Doctor Kellie Martin is two
00:05children tests and Jacob and just
00:08take a moment just to silence
00:10just to recognize Tony's legacy.
00:21Well, thank you, so let's now turn
00:24to our first of two great speakers.
00:28We were very fortunate this year
00:31to recruit Doctor Jeffrey Ishizuka.
00:34Jeff is an assistant professor of medicine.
00:38And Jeff's work.
00:39Previously at Harvard was was
00:42focused on the biology of T cells.
00:44Discovering knew better understanding
00:46of that biology and and and
00:49ultimately leveraging that science
00:50to what is likely to be the next
00:54generation of amino therapies.
00:55And we're really very fortunate to have
00:58Jeff as one of our physician scientists in
01:01the center of molecular Italian Colosseum,
01:04member of the Melanoma program,
01:07and physician scientists in general at.
01:09At Yale and Smile also,
01:11Jeff really excited to hear about your work.
01:14Turn it over to you.
01:17Thank
01:17you so much Charlie, really appreciate
01:20it and let me just project my slides.
01:23How there we go?
01:25Yes, thank you so much and thank you
01:28for the opportunity to speak today.
01:30Today I'm going to be talking to you
01:32about some of the work we've done.
01:34Targeting double stranded RNA in
01:37order to overcome immunotherapy
01:38resistance and also update on
01:40other ongoing projects in the lab.
01:42This is my disclosure slide.
01:45I wanted to begin with the overall survival
01:48curves from the Checkmate 067 trial,
01:51which is likely familiar to this audience.
01:54These curves represent survival in
01:56advanced Melanoma by patients treated
01:58with immune checkpoint blockade.
01:59In this case,
02:00with antibodies targeting PD 1C,
02:02TL A4 or the combination.
02:05I wanted to start here because
02:06Melanoma has been something of
02:08a touchstone for the use of
02:10checkpoint blockade in solid tumors.
02:11First indication approved and remains
02:13one of the indications in which immune
02:15checkpoint blockade is most effective
02:16in these data are outstanding,
02:18particularly when compared with the pre
02:20immunotherapy standard of Care Dakar Busine,
02:22which had an overall survival
02:23of five to 10% at five years.
02:26However,
02:26even in this disease,
02:28large proportion of patients
02:29don't experience durable benefit.
02:31The situation which is which is
02:33actually more challenging in other
02:35diseases where responses are less good.
02:38And this is really the focus of
02:39our work to improve responses
02:41in this disease and in others.
02:46Certainly, however, if you check,
02:48my blockade is rapidly reshaping the
02:49landscape of cancer care across indications.
02:51I was preparing for this talk and I
02:53had to go through and update this slide
02:56because indications have nearly doubled
02:58since its original publication by
02:59Tony Ribas and Jed will Chuck in 2018.
03:01Although many of us have followed
03:03this emerging data very closely,
03:05I have to admit that it gave me pause to
03:08consider the pace of change in this field.
03:11The advancement of PD one access approvals
03:14continues through lymphomas and solid
03:17tumors of desperate tissue origins.
03:19Combination approaches have also
03:20proliferated, including approaches,
03:22approvals in music, leoma,
03:24breast cancer, and others.
03:26Successful combinations include
03:27combinations of checkpoint inhibitors
03:29with other checkpoint inhibitors,
03:31chemotherapies and touristing
03:32kinese inhibitors,
03:33and notably many here at Yale,
03:36have played critical roles in this dance.
03:42Still, for all the advances,
03:43there have been a lot of failures and
03:46there remain a lot of ongoing challenges.
03:48For most, many patients don't respond,
03:50indeed, considered across all indications,
03:52most patients don't respond in a few
03:54of the response rates listed are
03:56really based on earlier trials that
03:58likely overestimated response rates.
04:00Many of them also include
04:02biomarker cutpoints,
04:03PDL 1 positive ITI and this sort of thing.
04:07And in my mind there are really a couple
04:09of big areas in which we can improve.
04:111st for all of the new indications,
04:14few combinations involving novel
04:17targets have been approved.
04:192nd, we have a limited mechanistic
04:22understanding of how these agents work.
04:25Accordingly,
04:25the biomarkers that we used to deploy
04:27them lack sensitivity and specificity,
04:29and there's not a great way to rationally
04:32prioritize combinations with anti PD one.
04:36So it's worth considering for a moment what
04:37we've learned about response and resistance,
04:39not so much in the interest of
04:41an extensive overview for which
04:42we wouldn't have time today,
04:43but in terms of the pathways that have
04:46given the strongest clinical signals to date.
04:48The data shown here are from the study
04:51by Merck of over 300 different patients
04:53across 22 different tumor tissue types.
04:56These figures show responses.
04:59Non response defined as CR
05:01or PR versus no CR PR.
05:04When graphed with tumor mutational burden
05:06on the Y axis and a gene expression profile
05:09representing tumor microenvironment,
05:11inflammation kind of T cell
05:12inflammation on the X axis.
05:14The genes in this profile are listed in the
05:17upper right here and notably include PDL,
05:20one among them as well as several MHC related
05:23genes and kind of T cell related genes.
05:28Tumor mutational burden, as you know,
05:30is often used as a surrogate for
05:32too many antigens and the gene
05:34expression profile really points
05:35to information of the tumor,
05:37micro environment and the authors make
05:39two points that are important here.
05:41First, that these are two of the
05:43strongest predictors they could find.
05:45Reviewing one of the largest
05:46and most comprehensive datasets
05:48that existed at the time.
05:49Really, it's telling us in second
05:51that they appear to predict response
05:52independently of one another.
05:54That is to say that although the
05:56best responses are in that kind of.
05:58Upper right quadrant that you actually get
06:01a good number of responses in a T cell.
06:05Inflamed only micro environment,
06:06or in TMB only TB high only tumors.
06:12For the sake of time today,
06:13I won't spend a lot of time
06:15on TMB or antigen load,
06:17so it's obviously an important consideration.
06:18Instead, I'm just going to talk about
06:20tumor microenvironment information,
06:21which is really the focus of our lab.
06:23Aside from the work by the Merck
06:25Group A number of lines of evidence
06:27have established inadequate tumor
06:28microenvironment information.
06:29As one of the most prominent mechanisms
06:31of resistance to me, no therapy.
06:33Most dramatically,
06:34this occurs in immune desert type tumors,
06:37which entirely lack T cell infiltrate,
06:39as depicted here. However,
06:41it can also occur in a different phenotype.
06:44The so-called immune excluded
06:45tumors which have anti tumor immune
06:47cells at the site of the tumor,
06:49although they are excluded
06:50from the tumor core,
06:51either by physical barriers
06:53or by immune signaling.
06:55Finally,
06:55we believe that there is the T cell
06:57inflamed type of tumor that have
06:59diffuse infiltration of T cells
07:01that tend to be PD L1 positive,
07:03and these are the ones that we
07:06believe respond best to immunotherapy.
07:09To date,
07:09there's been progress in identifying
07:11therapeutic strategies to enhance this
07:13tumor microenvironment information,
07:14many of which involve either real
07:16or simulated infection of the tumor
07:18to trigger anti tumor immunity,
07:20and I think about them in kind
07:22of two big buckets.
07:24The first is the provision of
07:26exogenous sources that mimic
07:28nucleic acid ligands to tumors.
07:30This includes sting agonist,
07:31MDA 5 or rig I agonist,
07:33double stranded RNA sensing
07:35pathways and uncle lytic viruses.
07:37The other is the induction of endogenous
07:40sources of nucleic acid ligands,
07:43primarily endogenous retroviruses,
07:44although others have been published recently,
07:47alualu repeats in humans.
07:51And examples of this include a
07:53deciding in CDK 46 inhibitors.
07:57So my interest in turning these cold
08:00microenvironments hot and kind of
08:02providing these logins to tuners really
08:04developed out of work in the Canings
08:07lab was finishing my postdoctoral work
08:09there and through the type of experiment
08:11that I'm showing here on the left,
08:14you have kind of a transplantable tumor
08:16cell line, something like a B16 Melanoma,
08:19and the way the experiment works is to,
08:22in fact, that cell line with a
08:24library of CRISPR CAS 9 guides
08:26that knockout thousands of.
08:28Immunologically relevant genes.
08:29In the genome and then to kind of
08:33select those guides until you have
08:35a pool of knockout tumor cell lines
08:38that is then implanted into mice
08:40under increasing immune selective
08:42pressure from extremely immunodeficient
08:44mice that lack T cells to mice with
08:46an intact immune cell system.
08:482 mice treated with immunotherapy.
08:50In this case,
08:51the irradiated GM CSF secreting
08:53whole tumor cell vaccine GBX,
08:55plus anti PD one kind of strong
08:58immunotherapy treatment regiment.
08:59Would grow these tumors for about
09:012 weeks and then remove them.
09:04Harvested tumors and sequence the sequence.
09:06The barcodes sequence the guides using
09:09them as barcodes and quantitating.
09:11Enrichment and depletion of each
09:13guy and the way we interpreted
09:15this experiment was to compare high
09:17to lower mean selective pressure.
09:20So immunotherapy treated to
09:22immunodeficient mice, for example,
09:23and to interpret it that guides
09:25that were depleted.
09:27Comparing height alone,
09:28selective pressure represented Jews that,
09:30when deleted, convert sensitivity.
09:32To the mean system,
09:34and therefore potential targets
09:36for combination therapy.
09:38In contrast,
09:39guides that were enriched under
09:41strongly selective pressure suggested
09:42to US jeans that were lost made
09:44tumors resistant to new therapy.
09:48And a lot of the targets that we
09:50found this way actually ended up in
09:52the kind of realm of double stranded
09:54RNA sensing or antiviral triggering,
09:56and this is really the area that
09:58I focused on throughout my time.
10:01And this guy is thinking because a lot of
10:03what we know about viral infection comes
10:06from the study of exonerees viruses.
10:08But of course the genome is comprised
10:11largely of repetitive elements that have
10:14the potential to form double stranded RNA.
10:17These could be small interspersed
10:19nuclear elements and obvious retrovirus.
10:21Endogenous retroviruses are long
10:23interspersed nuclear elements or or others.
10:26And so we considered that that we've
10:29Co evolved with these elements.
10:31With these kind of viral remnants
10:33in many cases and ourselves have
10:35developed systems to regulate double
10:37stranded RNA sensing to distinguish
10:39between double stranded RNA.
10:41That's a result of normal cellular
10:43activity and exogenous viral threats.
10:45And so we thought that by targeting some
10:48of the genes that control this regulation,
10:51we might sensitize tumor cells
10:52to tumor therapy.
10:53Trigger this kind of anti virus state.
10:58And the top hits that we discovered
11:01through this process in the antiviral
11:03sensing arena was this paid.
11:05R18 R is an adenosine deaminase
11:07that acts on double stranded RNA.
11:10It has a long cytoplasmic P-150 isoform.
11:12That's interferon inducible and a short.
11:15Constitu Tively Express P110I support him.
11:19The main known function of edar is to
11:21catalyze the conversion of adenosine
11:22to in a scene and double stranded RNA.
11:25And it's thought that in so doing it
11:27prevents double stranded RNA sensing in
11:30the triggering of antiviral immunity.
11:32Kind of autoimmunity.
11:33Accordingly,
11:33there is an autoimmune syndrome called
11:36Acardi Goutieres syndrome that is
11:38associated with biallelic mutations
11:40of a Darwin on the catalytic domain.
11:43It can be quite severe effects
11:45children and mimics viral infection.
11:47However, Interestingly,
11:47the parents of affected patients
11:49who have monolith mutations in the
11:52catalytic domain have evidence of
11:54increased signatures of interferon
11:55gene expression in the blood,
11:57but have no detectable disease phenotype,
11:59suggesting that there's a gene dose effect.
12:04So to begin to validate our
12:06one as a potential drug target
12:08for combination immunotherapy.
12:09We created dedicated knockout tumor cell
12:12lines again using the B16 Melanoma model.
12:15This transplantable tumor model and
12:16we implanted these into mice under
12:19increasing selective pressure.
12:20It means selective pressure starting
12:23with the extremely immunodeficient
12:24nods give gamma mice that entirely
12:27lack adaptive immunity and have
12:29only impaired innate immunity.
12:31In these mice,
12:32looking at the 8 Arnold tumors,
12:34either P-150 knockouts in Orange,
12:36P-150 P, 110 knockouts in red
12:38compared to controls and Gray,
12:39and looking at tumor volume on the top,
12:42or survival in the bottom,
12:44you can see a sort of minimal
12:46decrease in the growth of the
12:49Darnell tumors compared to controls.
12:51And a minimal increase in survival.
12:55In contrast,
12:55when planted these tumors into wild
12:57type mice with an intact immune system,
12:59you see a significant decrease in
13:01the growth of tumors in a significant
13:03survival advantage for the mice.
13:05Finally,
13:05when we implemented these tumors into
13:08mice and treated with anti PD one,
13:10we saw a near 100% cure rate for
13:12mice treated that were a Darnall and
13:15almost no cures in the control chambers.
13:18So to start to understand
13:20the mechanism of this,
13:21we looked at the tumor micro environment
13:23of untreated a Darnall and control
13:25tumors 14 days after implantation,
13:27and we did this using immuno histo
13:29chemistry and as you can see on the
13:32left in control tumors you have
13:34the immune desert type phenotype.
13:36Almost no CD8T cells infiltrating.
13:38In contrast,
13:39in a Darnall tumors we saw this
13:41T cell inflamed phenotype with
13:42diffuse infiltration of CD8T cells.
13:44Quantitative here on the right.
13:49To understand this more deeply,
13:50we next perform flow cytometry.
13:52Again with tumors 14 days after
13:54implantation in the untreated setting,
13:56and as you might predict,
13:58we saw an increase in CD 45
14:01positive immune cells and a Darnell
14:04tumors compared with controls.
14:06And then looking within the CD 45 compartment
14:08we saw increases in CD 3 positive T cells,
14:12CD 4 positive T cells,
14:13CD 8 positive T cells,
14:15gamma Delta T cells and NK cells.
14:19In contrast, when we looked at
14:22immunosuppressive populations,
14:23including mdse and tumor
14:25associated neutrophils,
14:26we saw significant increases in control
14:30tumors relative to a Darnall tumors.
14:34Finally, to probe the micro
14:35environment yet more deeply,
14:37we perform single cell RNA sequencing.
14:39These are the populations we
14:41recovered with myeloid populations
14:43in the upper right and T cell
14:46populations in the bottom left.
14:48As you can see,
14:50using these density plots that
14:51we adapted for this purpose,
14:53you get a strong signal from suppressive
14:56myeloid populations and to like
14:58macrophages and mdse in control tumors.
15:00But a weaker signal from inflammatory
15:02monocytes and CD8T cells.
15:03In contrast,
15:04in the 8 Arnold tumors you
15:06have hardly any signal from the
15:08suppressive minded populations
15:09and and enrichment of single from
15:11inflammatory monocytes and CD8T cells.
15:16To understand what's driving this
15:18change in the micro environment,
15:20we wanted to study the double
15:22stranded RNA sensing pathways that
15:24we thought could be associated
15:25with the phenotypes we'd observed.
15:28Specifically, we wanted to understand the
15:30role of protein kinase are an MD5 rig,
15:33I and nouns which are both associated
15:35with his internal sensors of nucleic
15:38acids in double stranded RNA,
15:40specifically protein kinase power
15:41is associated with translation
15:43arrest in a pop ptosis.
15:44Upon binding double stranded RNA.
15:47Where is MD5 regarding mass induced type
15:50one interferon in the antiviral state?
15:53To test the role of each of these sensors,
15:56we generated a series of double
15:58and triple knockout tumor cell
16:00lines and probe some of the in
16:02vitro phenotypes that we previously
16:04previously studied in a Darnell tumors.
16:06Specifically,
16:07we looked 1st at growth inhibition.
16:09So when you stimulate control
16:11tumors with interferon in vitro,
16:13there's a slight defect in growth that's
16:17magnified when you knockout eight R1.
16:20Looking at our double knockouts,
16:22we saw no effect of knocking out rig.
16:25I MDA 5 or Mens but saw that knocking
16:28out peak PQR reduced the phenotype to
16:31the levels observed in control tumors,
16:34suggesting a PQR was alone.
16:37Responsible for the in vitro
16:39growth defect that we'd observed.
16:41We next looked at interferon beta production.
16:44And this was again an in vitro Aliza and
16:47tumor cells stimulated with interferon.
16:50As you can see,
16:52control tumors produce no
16:54detectable interferon,
16:54whereas a Darnall tumors
16:57produces significant quantity.
16:58This is maintained from the loss
17:00of Rig I suggesting that guy is
17:02not involved in the phenotype.
17:04However,
17:04following the loss of MDA,
17:06Five Man's or PK are you see a
17:08significant reduction suggesting
17:09that all three of these sensors,
17:11or these two sensors in this adapter
17:13have a role to play in phenotype.
17:17We next wanted to understand which of
17:20these double stranded RNA sensing pathways
17:21was required for the in vivo phenotype of
17:24sensitization to whom checkpoint blockade.
17:26So we took our double and triple knockout
17:28tumor cell lines and implanted them into
17:30mice, treating the mice with PD one.
17:33Antibodies targeting PD one,
17:34and as you can see in our control experiment,
17:37control tumors continue to grow out
17:39as they did previously for us in the
17:41eternal summers respond well to,
17:43you know, therapy.
17:45This phenotype persisted following
17:47loss of PQR, suggesting that PQR is
17:50alone not required for the phenotype.
17:53Similarly.
17:53It persisted following loss of MD5,
17:57suggesting MDA 5 alone does not
18:00explain the phenotype.
18:01However.
18:02Following the deletion of both PK
18:04are in MDA 5 together with eight
18:06or one we no longer observe any
18:08difference between the growth of
18:10eight R1 knowledge control tumors
18:12treated with immunotherapy.
18:14Together,
18:14these results suggested to us that
18:16growth inhibition by PQR or antiviral
18:19sensing by MDA 5 amounts sufficient
18:21mediate sensitivity to no therapy
18:23but that at least one is required.
18:28We next wanted to understand which
18:30double stranded RNA sensing pathway
18:32was required for the enhanced community
18:34filtration for the inflammation in the
18:36tumor microenvironment that we'd observed.
18:39And so we again used our double and
18:41triple knockout tumor cell lines.
18:43In this time return to our habit of
18:45looking at the tumor microenvironment,
18:47dissecting the tumors out,
18:49separating out the cells,
18:50and quantitating them.
18:52To look which sensor was was required.
18:57In our control tumors,
18:58you see a relatively low infiltration
19:01of immune cells that significantly
19:03increased following loss of eight R1.
19:05And Interestingly,
19:06this phenotype is,
19:07if anything exaggerated following
19:09loss of protein kinase are however
19:12it's attenuated following loss of
19:13MBA 5 and oblated following the
19:16loss of the two senses together.
19:18A similar pattern followed when
19:19we looked at the proportion of
19:22the 45 positive immune cells that
19:24was comprised of CD8T cells,
19:25again,
19:26increases in eight are null that
19:28persisted following loss of PQR
19:30was attenuated following loss
19:32of MD5 with loss following the
19:34loss of both sensors together.
19:36When we look at a immunosuppressive mdse,
19:39we saw the opposite pattern
19:41increases in control that persisted
19:43or work were even increased.
19:45Further following loss of PQR and no
19:48loss of the phenotype following loss
19:50of MDA 5 for the two sensors together.
19:57This suggested to us that MBA five
19:58may be playing the predominant role.
20:00And inducing tumor microenvironment
20:02inflammation may darnel tumors.
20:04To confirm this,
20:06we looked at the production of interferon
20:08beta interferon gamma in the tumor
20:11microenvironment of the eternal jiggers.
20:13And we saw a similar pattern again
20:15increases in a terminal tumors that
20:18persisted following loss of PQR but was
20:20lost after law after loss of MD5 or the
20:23two sensors together in the same pattern.
20:25Again looking at tumor
20:28lysate interferon gamma.
20:29So haven't seen having seen
20:31this powerful dual mechanism for
20:33sensitizing tumors to immunotherapy.
20:35We asked whether loss of eight R1 was
20:38sufficient to overcome commonly acquired
20:40mechanisms of resistance to amino therapy,
20:43including genetic aberrations
20:45that have been identified as
20:47enriched when comparing discordant,
20:48responsive,
20:49pretreatment,
20:49and resistant posttreatment lesions.
20:51Matched with the same patient.
20:54Known mechanisms that fit this
20:57description include the loss of MHC one
21:00through mutations of HLA or beta 2M,
21:02loss of targeting,
21:03children expressing through Mino,
21:05editing mutations and interferon sensing
21:07pathways including interferon gamma receptor,
21:09the Jackson,
21:10the stats.
21:13And we focused first on the loss of MHC one,
21:16as mediated by loss of data to microblogging
21:19which has been repeatedly identified as
21:21important in challenging form of resistance.
21:24To create this model we
21:26again use CRISPR CAS 9.
21:28This time deleting beta 2 micro
21:30globulin and eight are together.
21:32Along with creating match
21:34control tumor cell lines.
21:37To validate our model of resistance,
21:39we compared control in beta two of null
21:41tumors in the untreated that is dashed
21:44line state versus the treated state.
21:47That's the solid lines using again,
21:49this strong immunotherapy treatment
21:51regimen of GBX and PD one.
21:54And we did this because the normal
21:56control chambers responded very poorly
21:58to PD one and we wanted to make sure
22:00that we could see a response in control
22:03tumors and then validate that it was
22:05lost in the beta two unknown tumors.
22:08And sure enough,
22:09that's what we did see you can see the
22:11control tumors respond albiate transiently.
22:13Alternately,
22:13do grow out to this strong unit
22:15therapy treatment regiment,
22:16but made it to heaven.
22:18All tumors hardly respond at all.
22:22We next looked at a Darnall tumors.
22:25This is our positive control
22:26experiment using strong again
22:27with therapy treatment regimen.
22:29We got a great response to treatment.
22:31The untreated tumors grow out,
22:33albeit more slowly than controls.
22:36Strikingly, however,
22:37this sensitivity persisted following
22:39loss of beta two microglobulin,
22:41suggesting that loss of a Darwin
22:43in tumors is sufficient to overcome
22:46this mechanism of resistance.
22:48This result was a bit surprising actually.
22:50At first, as it suggests that CD8T
22:52cell recognition with MHC one in
22:54tumors is not in all cases required
22:56for the response to amino therapy.
22:58It also raises the question as to
23:01whether it could be possible to
23:03target tumors that entirely lack
23:04high quality CDH cell antigens.
23:06A lot of ongoing work in the
23:08lab is focused on dissecting the
23:09mechanism of this finding,
23:11and one of the first
23:12things we wanted to know.
23:14Is whether antigenic vaccine GBX,
23:16which was unsuccessful in
23:17translating to human use,
23:19was required for this response.
23:23This is actually pretty new data
23:24or afraid with PD one alone,
23:26and found that indeed you still get
23:28great responses in a Darwin all tumors.
23:32Even without the gmax.
23:34To start to understand this
23:36mechanism further, we again looked
23:38in the tumor microenvironment,
23:39this time focusing on our beta 2M
23:41null compared to control tumors.
23:43And so, as you would expect,
23:46increased immune infiltration
23:47CD 45 positive cells.
23:50But now focused on some of these
23:52MHC one non MHC one restricted
23:55cytotoxic populations and these
23:57include granzyme B positive CD
24:004 positive T cells and NK cells.
24:02With the hypothesis that perhaps
24:04these cells which don't require MHC
24:07one for recognition of tumor cells.
24:09May be involved in the phenotype
24:13we've observed.
24:14We've also begun to dissect the
24:16cytokinin kyma kind drivers,
24:17by which these populations may
24:19be recruited and activated.
24:21These graphs are from side to kinda be
24:24Teresa Beta to null and a Darnall tumors.
24:28The two prominent chemo kinds
24:29were identified so far.
24:31CX CL 10 in CCL 5.
24:34Which are both significantly
24:35increased in our beta to emulate
24:37our one all tumors compared with
24:39beta to a control control tumors.
24:44Notably Ehrenring here at Yale has
24:46described a similar phenotype of
24:47being able to overcome the loss
24:49of MHC one using a modified I'll
24:5118 side kind that he designed.
24:52So this remains another possibility
24:54that we haven't yet explored.
24:56However, we think this type of study
24:59is important 'cause articulating the
25:00general principles by which loss of MHC
25:02one can be overcome could lead to new
25:05treatment approaches to target tumor
25:06specific immune evasion mechanisms.
25:11In summary, I hope I've convinced
25:13you have several points.
25:14First aid are one loss over improves the
25:18response to me to therapy. Specifically,
25:21it can overcome the lack of evidence.
25:23Plain tumor, micro environment and the
25:26loss of antigen presentation by image C1.
25:29Additionally, this phenotype is
25:31driven both by tumor microenvironment,
25:33inflammation mediated by MDA
25:355 and sensitization.
25:36Interferon driven by PK are.
25:40Finally, and I think this may be important.
25:43Tumor cells contain sufficient innate
25:45lightning into drive therapeutic information.
25:47If they are in need.
25:49Nucleic acid sensing
25:50checkpoints are disabled.
25:52And what we think this implies is that
25:54there may be other similar innate
25:56immune checkpoints that limit the
25:57sensing of double stranded RNA or other
26:00nucleic acid ligands that we could
26:01think about as therapeutic targets.
26:05And really, those questions inform the
26:07rest of the work that the lab is doing.
26:10I've mentioned already a focus on
26:12double stranded RNA and eight R1.
26:14We're also applying functional genomics
26:15to try to identify other novel targets.
26:18Really, with the insight that we
26:19have to focus on turning on some of
26:22these pathways of double stranded RNA
26:24sensing or micro violent information.
26:27And then we're involved in human translation,
26:29doing kind of in depth tumor
26:31microenvironment investigation across
26:32several different tumor indications.
26:34We're always looking for new
26:37collaborators there.
26:38And all of this comes under the rubric
26:41of therapeutically targeting the
26:43information in the tumor microenvironment.
26:45In just the last couple of minutes here,
26:47I want to quickly mention some of
26:49the ongoing projects in the lab that
26:51I haven't talked about this far.
26:52First, I mentioned just the project.
26:57Describing how to Riker environment
26:59inflammation can overcome the
27:01loss of MHC one.
27:02This is being led by Jessica Way,
27:05but she's Additionally leading a project.
27:09Looking at human tumors and trying
27:10to turn these pathways on in ex
27:13vivo samples as well as doing deep
27:14dissection of the micro environment.
27:16Where we go is working on novel
27:19strategies to detect double stranded
27:20RNA and to mimic the sensors of double
27:23stranded RNA that we believe will be
27:26compatible with functional genomic
27:28screening in the identification of
27:30novel cancer immunotherapy targets.
27:32And finally,
27:32even Kim who is in the lab focused on the
27:36comparison of discordant response lesions.
27:39So responsive and resistant lesions.
27:42From the same patient trying to
27:44understand novel mechanisms of
27:46resistance to new therapies so
27:47that we can focus on overcoming.
27:50With that I want to thank everybody in
27:52our lab as well as our collaborators
27:55and mentors here at, you know,
27:57have been fantastic.
27:58I also wanted knowledge at Nikki
28:00Ning my form. Enter drumming.
28:01So much of the work that I presented early
28:04derives from from studies with them,
28:06and of course our funding here
28:08at the Cancer Center and the
28:10International Research Alliance.
28:11With that, I will wrap up.
28:13Thank you so much for the chance to present,
28:16and I'm happy to take questions.
28:19Jeff, thank you. That's just
28:21terrific work and really exciting.
28:23And we we have folks can submit questions.
28:27We have one question.
28:28Mike Hurwitz asked.
28:30So given the response in eight R1
28:33knockouts in the absence of MHC class one,
28:36do you think that's function of
28:38CD4T cells or NK cells, or both?
28:42Or some other mechanism? Yeah,
28:44I think that's a great question and we
28:46definitely would love to know that answer.
28:50Best hypothesis Now is that partially
28:52based on some of the work that Ehrenring
28:56is presented in Marcus Bosenberg.
28:58NK cells could be an important player there.
29:01Certainly there increased and we
29:02started to see some cytokines in Kemah
29:04kinds that may activate them further,
29:06but you know, we don't even know for
29:08sure that CD8T cells aren't important.
29:10That's an experiment we're doing now.
29:12We just know they're not recognizing the
29:14tumor, but could they be activated through
29:16cross presentation or another means is
29:18another question that we're investigating.
29:21And then you know, in related work.
29:24Obviously Akiko, Saki,
29:25and Anna Pile of working independently on
29:28Rig Rig I are iguana, which which it is.
29:31But which obviously is not necessarily
29:34related to the function vadar one,
29:36and you know how?
29:38How do you see those two with those
29:41two sort of bodies of work relating?
29:44Yeah, so this is
29:46a great question Charlie and actually
29:48Akiko is one of my mentors here and.
29:52Collaborators and we've talked about this.
29:55We're actually in the process of testing.
29:58Are a guy at. Egotist with the innate
30:03arnolin control tumor cell lines and
30:05you know the colloquial way we we
30:07thought about this is kind of as a
30:09maximum inflammation bomb because what
30:11we've shown is that any interferon
30:13producing stimulus can trigger this
30:158 Arnold amplification of sensing,
30:16and so our hypothesis would be that if
30:18you initiate signaling through a guy,
30:20even if there a guy is not involved
30:23in the pathways we've described here,
30:25you basically create a massive
30:26amplification of interferon,
30:27buy by further knocking out eight
30:29R1 so that remains to be seen,
30:32but that's what I would hypothesize.
30:34Yeah, that's interesting.
30:35It sounds like a great
30:37opportunity to look at that.
30:38Well, I I want to keep us on time,
30:41so Jeff, thank you.
30:42I know there are other questions
30:44coming in and people should certainly
30:46reach out to you directly, Jeff.
30:48But thank you for a superb presentation
30:51and let me now turn to our second speaker,
30:54doctor Robert Bone and Bob Bone is a
30:56professor of medicine in hematology,
30:58and recently the past year joins us as the
31:01director of the Benign Hematology program,
31:03as well as the medical director of
31:06the Hemophilia Treatment Center.
31:08Prior to joining Yale,
31:09Bob was founding faculty member
31:11and leader at the Frank Netter
31:14School of Medicine at Quinnipiac,
31:16as well as a professor of medicine at
31:19the University of Connecticut School of
31:21Medicine and Bob throughout his career,
31:24really has been a leader in in in the
31:27clinical care and sort of advancing
31:30work in hemostasis thrombosis as well
31:32as benign hematologic conditions.
31:35And we're really,
31:36very fortunate Bob to.
31:38That Bob,
31:38now leading this section and sharing
31:40with his work with us.
31:42So Bob thank you.
31:44Thank you, Charlie for that introduction
31:47and for the opportunity to speak today.
31:50Let me just share my screen here.
31:53So good afternoon everybody.
31:56And what I would like to do in the
31:58next 25 minutes or so is discuss with
32:01you some of the advances that have a
32:04curd in the treatment of hemophilia
32:07and what I hope to show you is that
32:09over the past five years there have
32:11really been significant and substantial
32:13advances which came in the background
32:16of really several decades of really
32:18only modest advances in therapy.
32:21So just as a brief review here,
32:24these are excellent disorders,
32:26mostly affecting men,
32:27but can also affect women who might
32:30have low factor levels due to
32:33unequal X chromosome inactivation,
32:35hemophilia A&B or deficiencies
32:37in factor 8 or 9 respectively.
32:39They are clinically identical disorders
32:42and the severity of the disease
32:45is really relies primarily on the
32:48residual factor that is remaining
32:50in the blood with those with severe.
32:53And moderate disease having less
32:55than 5% of factor 8 or factor 9
32:57and those with mild disease having
33:00a higher value and morbidity and
33:02mortality is due to spontaneous
33:04and trauma induced bleeding,
33:06including bleeding into joints which
33:09can cause a hemophilic arthropathy
33:12which we can be quite quite disabling.
33:15And just the history of hemophilia
33:17treatment in the last century
33:19is seen briefly on this slide,
33:21and at the end of World War Two
33:24blood or plasma transfusions were
33:26used to treat patients.
33:28This these were largely ineffective,
33:30is only small amounts of factor 8 or
33:33factor 9 could be transfused in the 1960s.
33:36Cryoprecipitate was discovered
33:37as a source of Factor 8,
33:40and that quickly gave way to the
33:42use of factor concentrates either
33:44factor 8 or factor 9.
33:46Purified from the plasma of
33:4910s of thousands of donors.
33:51And of course,
33:52while this advanced care,
33:54it also exposed individuals to a
33:56number of viral viral particles and
33:59hepatitis C and HIV became a very
34:02significant problem in this population.
34:04And then in the early 90s
34:07recombinant factors 8:00 and 9:00,
34:09or produced and for the developed world,
34:12where economically this was allowable
34:14of the treatment of hemophilia
34:17with recombinant factors 8 and 9.
34:19Became really the standard of
34:23care up until very recently.
34:26There are now about 145 federally
34:29funded hemophilia treatment centers
34:31in this country and of course jeliz is
34:34one of those is one of those centers.
34:37And the therapeutic.
34:38The approach in clinical issues
34:40are outlined here.
34:42Patients with hemophilia can either be
34:44treated in what's known as on-demand
34:46or episodic factor replacement,
34:48which is the treatment with Ivy
34:51Factor 8 or factor 9 to treat a
34:54bleed or prophylactic therapy.
34:56An inhibitor development,
34:57that is an Allo antibody directed
35:00against Factor 8 or less commonly,
35:02factor 9 is a significant problem
35:05for patients and may occur in 30 or
35:0840% of individuals with hemophilia A
35:10and makes treatment very difficult
35:13and the goals of therapy are really
35:15here to prevent
35:16any bleeding. If possible,
35:18prevent joint disease and optimize a
35:21quality of life for these individuals.
35:24And the infusion of factor 8 or
35:25factor 9 by patients is traditionally
35:27given at home intravenously.
35:29Patients from a very young age learn to start
35:32an Ivy and infuse factor 8 or factor 9,
35:35but because of the short
35:37half-life of these drugs,
35:38about 12 hours for factor 8 and
35:4018 to 24 hours for factor 9,
35:42they need to be administered two to
35:45three to sometimes four times per
35:47week to keep the factor levels in a
35:49range that will prevent bleeding.
35:51So this is an onerous thing
35:53for patients to do.
35:54And any advances here would be
35:58greatly appreciated by them.
36:00So here's the obligatory coagulations
36:02slide that I would like to show
36:06to to reinforce and emphasize
36:08the role that Factor 8 and factor
36:119 having blood coagulation.
36:14So what we're seeing here is the
36:17tissue factor initiated pathway and
36:20activation of factor 10 by tissue
36:23factor 7A or activation by factor
36:269 to 9 A by tissue factor 7A.
36:29And 9A is also able to activate 9:50
36:33A O2 pathways to get down to this
36:36all important enzyme factor 10A,
36:39and in this latter reaction factor
36:418 serves as a cofactor for the
36:44enzyme factor 9A.
36:45To act on its substrate factor 10
36:48and increases the rate of reaction
36:51hundreds of 1000 fold when factor 8
36:54is able to align the substrate and
36:56enzyme on a phospholipid surface in
36:59the correct in. In the correct fashion.
37:03One other thing to mention about
37:05Factor 8 before we get into some of
37:08the details of the advances is that
37:11factor 8 travels if you will in the
37:13blood bound to von Willebrand factor.
37:15Von Willebrand factor is seen here in
37:18this linear structure at the bottom,
37:20factor 8 is the yellow diagram above,
37:23and the binding of factor 8 von
37:25Willibrand factor enhances the
37:27half life of factor 8 from about
37:292 hours to about 12 hours.
37:31So this is a very important interaction.
37:34And just to point out here,
37:36'cause this will become important
37:38later is that the binding site
37:40on von Willebrand factor is these
37:42two protein domains,
37:43designated D prime and D3,
37:45and another important point is there
37:47appears to be a large portion of the
37:50factor 8 molecules termed the B domain,
37:52which is not required for factor 8 function,
37:55so you could remove that domain
37:58and in fact factor 8 has a similar
38:01activity than it does with that domain.
38:04So the advances in care of hemophilia
38:07really over the past five to six
38:10years come into three different areas.
38:12One is extended,
38:13half-life factor concentrates,
38:14allowing for patients to infuse
38:16less frequently.
38:17The development of non factor 8
38:19or 9 therapeutics,
38:20and then gene therapy and we'll
38:24go through these individually
38:26in the next 15 minutes or so.
38:28So the extended Half-life products
38:30have been produced by manipulating
38:32the recombinant factor eight
38:34or nine in a number of
38:35different ways, many of which are familiar
38:38to you by either adding polyethylene
38:40glycol or conjugating the factor to the
38:43FC portion of immunoglobulin or albumen,
38:45to improve half-life, or,
38:47in the case of factor 8,
38:49to remove that B domain, which causes
38:52a slight increase in the half life.
38:55And there are now a number of products
38:58that have been approved for use at
39:00our extended Half-life products,
39:02and I'll draw your attention
39:04to the last three on this list.
39:07These are factor 9 products which have
39:09been manipulated by these methods,
39:11seen here and the half life of these
39:14products has been extended from 18 to
39:1724 hours to upwards of 90 or 100 hours.
39:20So this is allowed patients with factor 9
39:23deficiency or hemophilia B to be treated.
39:26Once a week,
39:27once every 10 days and in some circumstances,
39:31even once every two weeks.
39:32So a significant advance for
39:34people needing to give intravenous
39:36therapy themselves at home.
39:38The advances in hemophilia A with factor 8.
39:41However,
39:41a much more modest with this
39:43type of manipulation,
39:45and it turns out that the the
39:47degradation in the catabolism and
39:49clearance from the circulation of
39:51factor 8 is much more linked to the
39:54clearance of von Willebrand factor,
39:56the protein that it's bound to.
39:59So making modifications in the FAQ.
40:00After 8 molecule has really had
40:03minimal effect up until recently
40:06on Factor 8 Half-life.
40:08So an interesting construct has been devised,
40:11and it's shown on the top panel here
40:13and in this construct the D prime and
40:16D3 regions of von Willebrand factor,
40:19the binding region to factor 8,
40:22is linked to an FC portion of an
40:25immunoglobulin and linked to the
40:27B domain less factor 8 molecule,
40:29which also has linked on at
40:32this hydrophilic polypeptide,
40:33which also can extend the half life.
40:36So this product has been called bib 001.
40:39And was treated with.
40:41Was used to treat a handful of
40:43patients in a safety study,
40:45and those results were were
40:47reported in the New England Journal
40:49of Medicine earlier this year,
40:51and patients were either treated
40:53at two different doses of this new
40:56product and the factor a clearance
40:58from the circulation was compared
41:00to the typical factor 8 clearance
41:02seen in the lighter blue bars here
41:05and what you can see I think,
41:07is that the half life of this newer product.
41:11Is now about two days increased,
41:13about five or six fold the half life
41:15of the standard factor 8 product.
41:18So this this product is now in
41:20large scale clinical trials and I
41:22think in the next year or two we
41:24should have some more information,
41:26and this may be an advanced
41:29for for some of our patients.
41:32So let me shift for a minute for
41:34the to the non factor product for
41:36the treatment of hemophilia and I
41:38think their significant advance
41:39has been made here and there are
41:41three drugs that will talk about
41:43will really focus primarily on this
41:45first drug which is called EMAS
41:48ISM AB. A nemesis Omab is a
41:51bispecific monoclonal antibody.
41:53That binds the factor 9 and factor 10,
41:56so it simulates the activity of Factor 8.
42:00Remember that factor 8 is able
42:02to colocalize factor 9 and factor
42:0510 on a phospholipid surface.
42:07This antibody is able to bind factor
42:109A and factor 10 in the circulation an
42:14again simulate the activity of Factor 8.
42:18So this drug is not exactly like Factor 8.
42:21There are.
42:22There are certain differences here.
42:24It binds to factor 8 and nine
42:27in the circulation,
42:28not just on the phospholipid membrane.
42:30It has different infinities
42:32for the substrate and enzyme,
42:34and whether or not that becomes an
42:37issue for this drug will only know
42:40as more experience is is accumulated.
42:43But nonetheless,
42:43this drug is really shown dramatic activity,
42:46so this this is a study that was
42:49published a few years ago in the
42:52New England Journal of Medicine.
42:54Here we had patients who have hemophilia
42:57A with inhibitors to factor 8,
42:59so a challenging group of patients to
43:01treat were treated either with their
43:04typical regimen of recombinant factor
43:067A or factor 8, bypassing activity,
43:08or with Emma system AB given by
43:11subcutaneous injection once a week
43:13and the annual bleeding rate.
43:15Is been been described on this slide
43:17here and you could see if we just look
43:21at these blue histograms for a minute here.
43:23The annualized bleeding rate in the
43:26EMA system app Prophylaxis Group
43:29was about five or six and it was
43:32almost 30 in the standard of care.
43:35Treatment of patients with
43:37hemophilia A and inhibitors.
43:39So a really significant
43:41advantage for these individuals.
43:43And then a second study was published
43:45with looked at patients with hemophilia
43:48A without inhibitors and these.
43:50This was a randomized trial.
43:52Patients were treated with one
43:54of two doses of Emma's is a map
43:56either given weekly or every other
43:58week by subcutaneous injection,
44:01compared with no prophylaxis.
44:02About 100 patients in the trial,
44:04and again the annual annualized
44:06bleeding rate went from about
44:0940 to about one or two.