Pathology Grand Rounds: January 12, 2023

January 13, 2023

Information

The Pathophysiology of Multiple Sclerosis, by David Hafler, MD, FANA

ID9372

To CiteDCA Citation Guide

00:00At all to this nice, auspicious start
00:03of the new Grand Round series for 2023.
00:07And I'm happy to have Doctor David
00:10Hafler here is our speaker today.
00:13So since 2009, Doctor Heffler has
00:16been the William S and Lower Styles
00:18actually professor and chairman
00:19of the Department of Neurology,
00:21Professor of Immunology, Immunobiology
00:23here at Yale and his neurologist
00:25and chief of the hospital, David.
00:28He's a he's a clinical research scientist
00:30with an interest in understanding the
00:32path of pathogenesis of inflammatory
00:34CNS diseases by studying
00:36both basic properties of.
00:38Due to regulatory pathways in humans
00:40and they run this function in patients.
00:42But as went on, I won't show you all
00:44the ruminations over the 20 odd years,
00:46but with 47,000 patients,
00:4968,000 controls by identified,
00:51233 genetic variants,
00:53all these have been replicated.
00:55All the original ones are
00:57similarly replicated,
00:57which counts for about half the
01:00estimated heritability for Ms.
01:02So that's great.
01:04You can show these wonderful little figures,
01:06but how did the variants cause the disease?
01:09And I think this remains one of
01:11the major challenges for you,
01:12not modern medicine as it's relatively
01:15easy with these technologies.
01:17It wasn't so easy in 2000 where we are
01:20now with different aluminum athlete
01:22technologies to identify genetic
01:24variants with big patient cohorts.
01:27I think the challenge is how to
01:29go from variance to disease.
01:31So this is an effort.
01:32Collaborate effort with Alex Morrison,
01:34Kyle Farr and Brad Bernstein.
01:36We together did genetic and
01:38epigenetic fine mapping of autoimmune
01:40disease variants and these data up.
01:42They'll publish about 5-6 years ago.
01:45I think it's still hand up, pulled up.
01:48So what we did we turns off for a second.
01:50So you all know DNA goes to RNA
01:53goes to protein, you'll learn that.
01:56And so to go from DNA to RNA,
01:58DNA has done wine very well and
02:00there has to be ways so that poll
02:03two another enzymes can get to the
02:05DNA so you can have transcription.
02:08Well you can then use things like
02:10K27 installation,
02:11K4 methylation maps to identify where
02:14this open chromatid on different cell types.
02:17One can then take those data and
02:19overlay them with genetic variants,
02:21arguing that if a genetic variant,
02:24it's a region where there is
02:26no open chromatin,
02:27that variance is not going to be
02:28playing a role that's help with
02:30every place there's a genetic
02:31variant for chromance,
02:32it's open,
02:33then it's likely to be an influencing
02:35that cell type.
02:36That's what we did as part of the
02:38ENCODE project with Brad Bernstein.
02:40Our lab participated in generating the K27K4
02:44methylation maps of human immune cells and.
02:47We have different diseases.
02:49We have neurologic diseases over here.
02:51Here is urate levels,
02:53renal function, kidney disease,
02:55cholesterol and here the autoimmune diseases.
02:58We'll concentrate on Ms.
02:59If you look at Ms.
03:01We found the genetic variance,
03:04the P values of less than 10 to the minus 30.
03:07We're hitting immune cells.
03:09That wasn't surprising T cells.
03:11Macrophages T regs.
03:13But also what was a bit surprising
03:15is there were hitting B cells,
03:17and more so in Ms.
03:19most any other diseases.
03:20If you look at other autoimmune diseases,
03:23it wasn't the case,
03:25say for lupus and primary biliary serositis,
03:29suggesting that B cells play a
03:32critical role in the disease.
03:33Somewhat unfortunately,
03:34around the same time my dear friend
03:37and colleague Steve Hauser made the
03:39observation paper published New England.
03:41Channel,
03:42if you perform B cell depletion,
03:45this is the two different
03:47studies offer one opera 2
03:48compared to the standard treatment
03:50that time barred interferon.
03:52The 9897% decrease in new lesions dramatic
03:55effect and I would just say clinically
03:57will we see a patient we start them on
04:00B cell depletion when we have a patient
04:03who doesn't respond usually isn't Ms.
04:05that's how good that drug is right now.
04:08So these data fit in very
04:10nicely with our observation.
04:12All of these cells in the disease,
04:14but those are those in your
04:16biologist and neuropathologist.
04:17I apologize,
04:18we did not get hits in the brain.
04:22Now I will say, and I won't share Tom
04:24and show the data today with the paper
04:27that is going to be coming out in
04:29nature from our from our consortium,
04:31identifying 2 haplotypes associated
04:33not with the risk of developing Ms.
04:37but with progression.
04:38And these are helpful types which
04:40are found in neuronal cells.
04:42So it's a separate question
04:43of what causes the disease,
04:45what leads to disease progression.
04:46If you have the risk capital type,
04:48your likelihood of progressing
04:51is significantly increased.
04:53So how do you go from Snips
04:56to to functionality?
04:57So we found hits the enough Capital Region.
05:01There are part of the far paper we
05:04found that steps on the NF Kappa B
05:06binding sites across the genome.
05:08And there's snips happen in the
05:10haplotype social NF Kappa B.
05:12Now it's about 10 to the minus 12,
05:14so they're genetic variance here.
05:16However, the odds ratio is about 1.1,
05:19so not a big effect.
05:21So I want to show you.
05:23Just even though the odds ratio is low.
05:26It has a major biologic fact.
05:28In fact, about 1819% of eugenic kits Ms.
05:32are in the NF Capital B and the
05:35TNF NF Capital B signaling region.
05:38So this is published now 7-8 years
05:41ago by will housing our laboratory.
05:43But basically about 20% of these
05:46healthy subjects.
05:47About 20% of you here are GG's.
05:51Do you know enough to know who you are?
05:53Is your homozygote for
05:55this particular variant?
05:57About 20% of you were A and the
05:59rest of you are had to reside in.
06:01If you are G,
06:03you have about a 20 fold increase
06:06in P50 NF Kappa B.
06:08Activity where if you're a it's
06:10significantly less and we every time
06:12we've looked at a genetic variant,
06:14this is what we find that the biology
06:16of these are quite striking and even
06:19though maybe 1.11 point one to the 233
06:23power becomes a very major effect.
06:29So in summary of the genetics big picture,
06:32the genetics of autoimmune
06:34disease dictates lower activation
06:36threshold of different cell types,
06:37including TH 17 cell B cells and T regs
06:41in the dozen or so genes we've looked at.
06:45I also say we've just started a
06:47collaboration with Steve Robbins just
06:49recruiting here from the Broad Institute.
06:51We had these wonderful techniques
06:53of looking at genetic variants
06:54and different cell types look
06:56into effect on motor function.
06:57So there are tools emerging which allows
07:00us to take a whole genome approaches.
07:02So getting back to the question
07:03I started with,
07:04you know the cause of multiple sclerosis,
07:07here's our working model.
07:10That's the genetically,
07:11as I said earlier, it's a genetically
07:13mediated autoimmune disease.
07:15It's initiating the periphery by T
07:17cells and macrophages that traffic
07:19into the central nervous system.
07:22So the the idea is that there are
07:24microbial antigens likely cross reactive
07:27with myelin that something happens to
07:29activate these antigen presenting cells.
07:32Probably be sales, probably EB though
07:35there's no biology behind that yet.
07:37There's expression of
07:39costimulatory molecules.
07:40We have activation of viral
07:42reactive T cells now we all have
07:44autoreactive T cells in our blood.
07:46I could clone mold reactive T
07:48cells from menu in this room.
07:50And a number of years ago,
07:51back in in the late 80s,
07:53we developed technologies for looking
07:55at autoreactive T cells were able to
07:57show for the first time that there are
08:00in fact autoreactive T cells in humans.
08:02Highly robust response and we identified
08:05a dominant epitope Amal and basic
08:08protein are recognized as 84102 region
08:10which went on to to find how it
08:13bounded MHC in a form post doc in the lab,
08:16kyouka fennick with Don Wiley went
08:19down to crystallize this these clones
08:22recognizing this epitope with the T
08:25cell receptor and MHC but was more.
08:28Interesting to me was the fact that we
08:31also found reactivity in healthy individuals.
08:34Significant reactivity
08:34led to decades of work.
08:37Why do we have autoreactive
08:38T cells in their circulation?
08:40This is work done by will count here,
08:42published in STM about seven years ago.
08:46This is principal component analysis
08:48looking at T cell reactivity using a T
08:51cell library approach against published.
08:53Just point out that the paper no peptide
08:55control anger can the myelin peptides.
08:58You can see that the red is Ms.
09:00patients that they tend to go off to GMCSF
09:04gamma in 17 whereas healthy individuals.
09:08The main reactive T cells need
09:10aisle 10 suppressive soda comma,
09:12which makes terrific sense.
09:14And if you do single cell cloning
09:16rather than the library approach,
09:18you can see that and help the individuals.
09:21They tend to make aisle 10 with single
09:24cell they tend they make all ten,
09:27that is Ms.
09:28patients making 17 GMCSF less gamma.
09:31So suggest that these aisle 10 secreting
09:33cells and all of us may play a role.
09:36For example one has damage.
09:38The brain stroke other factors that
09:41these cells may circulate into the
09:43nervous system involving scar formation.
09:46So.
09:48What we find is,
09:49so we have this situation,
09:51we have aisle 10 secreting cells
09:53in healthy individuals,
09:54but you also have regulatory T cells.
09:56Look to both TR1 and Fox V3 cells which
09:59are preventing this from happening.
10:02But multiple sclerosis is a loss of
10:04these Fox P3 regulatory T cells.
10:06I'll show you some recent unpublished
10:09data related to PRDM one.
10:12So the hypothesis is that
10:14in healthy individuals,
10:15these T regs prevent activation of
10:18autoreactive T cells where's Ms.
10:20patients are defective.
10:21This is work done by Dizzy Begleiten,
10:23Claire Batch Allen,
10:25published now almost 20 years ago,
10:28which was the first demonstration
10:29of T Reg dysfunction in the human.
10:31Autoimmune disease,
10:32these are all new ones that untreated Ms.
10:35patients or healthy donors and
10:37this is the presence of oppression
10:39perforation different ratios of T
10:42regs and you can see this market
10:45demolition diminish chip of Reg
10:47function in vitro in patients with Ms.
10:50and the same thing been found in type one
10:53diabetes everyone to arthritis went on
10:55to show that the T regs and patient Ms.
10:57work done by Margaret Dominguez
10:59Pierre is that these T regs and MSN.
11:02Making games that Fearon.
11:03Here we took T Reg,
11:05stimulated for four hours of PNA on a
11:07mycin and measured gamma secretion,
11:09purified populations and using
11:12sample control,
11:13aisle 17 versus gaming can see this gamma.
11:16They all express.
11:17Foxp 3 is a summary of these data
11:20that these T regs were making.
11:22Gamma,
11:22and I'll just point out I'll show
11:24a little bit in a few minutes
11:26that dysfunctional T Rex.
11:28What's happening is they go from
11:30suppressor cells to effector cells.
11:32And start making game interferon.
11:34So an Ms.
11:35is not just bad genes, not bad environment,
11:37but the bad interaction between
11:39genes and the environment.
11:40It's very interested again
11:43in environmental influences.
11:44The instance of Ms. stops at 2000,
11:47but it's continued to increase.
11:49Ms.
11:49Crohn's disease, type one diabetes,
11:51continues to increase.
11:52And of course that can't be genetics.
11:55So the pathophysiology of Ms.
11:58will involve genetic environmental
12:00factors which lead to the immune response.
12:05So well just give background on this about.
12:09We started looking at microbiome
12:11versus TH 17 cells in the blood and
12:15we started looking at dietary history
12:17and we found that if you ate at a
12:20fast food restaurant more than twice
12:22a week you had increased dial 17 cells.
12:25Statistically significant.
12:27We said really and said,
12:29well wasn't the golden
12:31arches provide that and salt.
12:33So we did an incredibly simple experiment.
12:36This work done by Marcus Kletzel,
12:38we added salt to the culture.
12:40And the same time,
12:42my dear friend and colleague Vijay Kutru
12:44was looking at TH 17 cell induction,
12:47identified SGK one as critical
12:49inducing TH 17 cells.
12:51We are very we do lab things together.
12:54We're very closely he told that SGK one,
12:57I told him that salt and we
12:59did the papers in parallel.
13:00I'll just do a little just on terms
13:02of we couldn't have started competing
13:04who could get this out first?
13:06But we're much more clever than that.
13:08We worked in parallel, sent the papers.
13:10Back-to-back to a weekly journal.
13:14And when the editor said, well,
13:16did you know about each other's work,
13:17we said no, not really,
13:20but no one would have believed
13:22that salt did this,
13:23but having two laboratories
13:24showing the same thing at a much
13:27more credence to the work.
13:28So it was and we've gone on to
13:31look at the effect of of salt and
13:33other factors in terms of fat
13:35in terms of T Reg function.
13:37So I'll just show you some of these data.
13:39They're really quite remarkable.
13:40If you add 40 milliequivalents
13:43of salt to a culture,
13:44you have logarithmic increases,
13:46dial 17 and M RNA and and secretion.
13:50They may be saying yourself,
13:51this is artificial right concentration.
13:54Salt and blood is about 150,
13:56which is what sea water is, turns out
13:59the concentration of salt in tissues.
14:02Is higher, it's about 18190 and
14:05blood is a suppressive condition.
14:08You don't want T cells being
14:10activated in the peripheral blood.
14:11So when T cells traffic into the tissue,
14:14they're in the condition of leading to more
14:17activation which is what we were seeing.
14:19I also say that a recent paper by
14:22another group in Germany showed
14:24increased salt concentration using MRI
14:26magnets in tissue skin tissue of Ms.
14:29patients compared to the controls.
14:31So we also this work done.
14:33By the talented graduate student
14:34of the NADPH Mandatory Hernandez.
14:36And she showed that if you add
14:38salt to T regs.
14:39So here we have T Reg effector cells,
14:41we load them with a green dot at the very
14:43green that stimulates them not dead.
14:44With the dye at T regs to go
14:46from here to here,
14:47they suppress entering the cell,
14:49cycle through the same thing
14:50with sodium chloride,
14:51they lose functionality.
14:55And the mechanism, if you look at SGK
14:58one would solve it goes up in the T regs.
15:01You can knock down the SGK one with the
15:04short hairpin RNA and then if you look
15:07at function here's effective function.
15:10If you knockout SGK one you go from
15:12control to here and restore function.
15:15So it was happening so also inducing
15:17SGK one with gamma difference secretion
15:20and T regs leading lots of function.
15:22So I'm presenting to you the importance
15:25of SGK one as a central factor.
15:28And loss of T Rex function.
15:31So come back to that in a moment,
15:32one of the great surprises in my life.
15:35So then the question what's the
15:37transcriptional circuit driving
15:38dysfunctional foxies through positive
15:40regulatory T cells in autoimmunity?
15:42That's can we identify a master
15:44regulator of T cell dysfunction.
15:47Let me just say this is work
15:49done by Thomas Sabita,
15:50who started as a postdoc is now in the system
15:52professor and there's really represents his,
15:55his original work.
15:56So what we basically did was performed a
15:59comprehensive transcriptomic and epigenomic.
16:01Profiling doing bulk and
16:03signals of RNA seek attack.
16:06Seek for epigenetic regulation genome
16:08wide for chromatid accessibility.
16:11He did transcription factor footprint
16:13analysis and accessible chromatin regions and
16:16look to the E QTL effects of Automeris Lucci.
16:18I'll just just show data on the 1st 2:00.
16:21This is a whole one hour talk in itself.
16:24And then we did a CRISPER activation
16:26based validation of this.
16:28It's regulatory elements getting
16:29at the molecular mechanism.
16:31So it's amazing what we can
16:32do in human biology now.
16:34It's unimaginable years ago.
16:35So first of all let me show
16:37you what we found.
16:39Found increases in PRDM 1.
16:42Now for those of you who are mouse
16:44people who say this doesn't make sense.
16:47It's it's well known that PRD one
16:49increases T Reg function announce
16:51T cells and I'll show you what
16:54what it was and that it we found
16:56that it drives a dysfunctional
16:58sheets something like program.
17:00But what's happened is not the
17:02long form that's increased it's
17:03a short form which inhibits the
17:05long form and here's the kicker,
17:07it drives SGK one of all the kinases and
17:11proteins that could have been induced by
17:14the shore former PhD one it was the SGK 1.
17:17So let me now I showed you the results,
17:19let me show you the data.
17:20See here we looked at T Reg,
17:22this is doing vulgar in DC looking at
17:26overlapping differential expressed genes
17:28between memory T regs and seating for itself.
17:31So you can see this market
17:34increase in PRDM one,
17:35BCL 3 pin three will also regulated.
17:38These are all induced by PRDM one and genes
17:41are downregulated by PRDM one like ID 3,
17:44LBH were down regulated.
17:45So we had this increase in.
17:47PRDM one did a replication of the set of
17:51patients and showed here that the PRDM
17:54one is upregulated in patient with Ms.
17:57but again confusing because
17:59of the mouse data.
18:01It's all about mice of course,
18:03but then we learned that there are two
18:06isoforms in humans that don't exist in mice.
18:09Dry nosed mammals do not express a
18:12short form of PRDM one and so the
18:16short form the dominant negative.
18:18When we looked in normal cell types,
18:20we found that the short form is what's
18:23expression memory cells and T regs,
18:24but not in B cells.
18:27So then we looked at the short
18:29former period you want by PCR,
18:30and indeed it's this short
18:32form that's increased NS,
18:33not the long form.
18:36And we rapidly perform Western
18:38blotting to show that indeed the
18:41short form is what's induced.
18:43We then looked at the data sets in
18:46particular set by oted all that's
18:48published in cell and we found that
18:51the PRD one isoform PRD one is also
18:54increased in most autoimmune diseases.
18:56So appears to be a common regulator of in
18:59T regs among different autoimmune diseases.
19:02And of course PRD one drives the
19:04blimp one and we looked at blimp
19:06one expression and MST regs it was
19:08in and it was it was increased.
19:10So it wasn't memory T Rex and memory T Rex.
19:13By flow. By both PCR and flow.
19:18So then here's the experiment
19:20where we transfected PRD one into
19:22T regs and integer cat cells.
19:24We have the induction of SGK one and here's
19:27measuring SGK one and primary union T Rex.
19:30See this? When we overexpressed
19:31the short form of T regs,
19:34you had an increase in SGK one expression.
19:39And then if you go in and do all the
19:40other experiments, not get SGK one,
19:42you lose the loss of T rate function but the
19:45short form in T Rex become dysfunctional.
19:48The whole story came together.
19:50So this suggests to us that the short form
19:52of PD one may be critical in different
19:55autoimmune diseases and driving dysfunction.
19:58And again we believe it's related to
20:00salt and to genetic factors that will
20:03show the data particular CD toward the
20:05genetic variant in CD28 who drives a P1.
20:10Now switch gears and show some public,
20:13recently published work looking at
20:15T cell traffic between blood and
20:17the CNS and the single cell work.
20:19So.
20:20T cell traffic into the CNS is very tightly
20:23regulated and CXCR 3 positive cells
20:25are the ones that get into the brain,
20:28crossing the correct plexus
20:30near great interest,
20:32and they get into the brain.
20:34And what I'll show you is a T cells
20:36in the central nervous system are
20:38CXCR 3 positive and express Tibet
20:41and make gamma interferon.
20:43We believe that this relates to the
20:45fundamental observation by the late
20:47Ben Barris that astrocytes are driving
20:49homeostatic communication would not.
20:51This is Michael Glee up but also are
20:54driving through secretion of cholesterol
20:56and TGF beta are driving this T bet
20:59induction that isn't the T cells go
21:02into the nervous system then they'll
21:04you in the brain drives this function.
21:07So give them on the knowledge of
21:09particular genre pop or Lotto submitter.
21:11Our Krishnaswamy and David Vandyke and Lee
21:14sang together did this worker with us.
21:17And basically we took spinal
21:20fluid group of patients,
21:21isolated the spinal fluid homonuclear cells,
21:25perform 10X the usual way that
21:28most you're familiar with now.
21:30And then we perform this on 6 healthy donors,
21:34get the bed in the moment, 5 patient with Ms.
21:36And looked at over 100,000 cells
21:39into 50,000 T cell receptor.
21:41Now does show high level summary
21:43of the of this work.
21:46So first in terms of blood
21:48versus spinal fluid,
21:49what we found wasn't surprising that the
21:51majority of cells in the spinal fluid
21:54or T cells what we observed before.
21:56So we started first bikes adding a
21:58spinal fluid from patients with Ms.
22:00right blood spinal fluid.
22:02We found the spinal fluid was very inflamed.
22:05I love this picture and we're sitting
22:07at the immunology repeat this change
22:09that we need to do controls goes that
22:12means we have to do spinal taps on age
22:14match 20 something year old students
22:17to do that. So Full disclosure.
22:20I do not know who would spinal taps, I said.
22:25I do not want to know because I
22:27do not want to be accused
22:29of coercion, collusion.
22:30So only Jenna knows who volunteered.
22:33They're all de identified.
22:35But at the I acknowledge all these wonderful
22:37students who had Spinal Tap stuff.
22:40So these cells acquire an
22:42inflammatory signature.
22:43So just quickly the menu now fate
22:46where the cell progression of
22:48blood and CSF release fate maps.
22:51This is a fading out.
22:52The red is blood, the blue is CSF
22:55via very different characteristics.
22:58They sense potential of heat diffusion,
23:00if any based transition embedding.
23:03I love the words they come up with and
23:05that world, but it's a way of looking at it.
23:08Unsupervised visualization that looks
23:10at geometric distance between data points.
23:14So here's the original tissue
23:15on the fate map.
23:16We can define the CD.
23:188 cells are here,
23:20CD4 cells are here.
23:22And we were able to describe we took
23:24the top 10 differential expressed
23:26genes in each cluster and to find
23:29different populations naive cells,
23:31naive CD4,
23:32CD 8 and really three different
23:34populations CSF cell we called CSF
23:371/2 and three in memory CD8 cells.
23:41So one could do something very interesting,
23:44which is looking at the continuum
23:46between blood and spinal fluid,
23:48getting back to the point of how T
23:51cells change function going to tissue.
23:53So we merge the fate to fusion
23:56operator with the original identity
23:57of itself come up with a tissue score
24:00which is basically the probability
24:02of transitioning from one form to
24:05another in a random walk and you
24:07can see the Tisha score that their
24:09cells are very blood like over
24:11here CD4CD8 cells in transition
24:13and cells is very CSF like.
24:18And so I did have a reality test.
24:21Ensuring that tissue
24:22score captures no biology.
24:23ITG Force is the LA4CD49D and
24:27requires this T cell traffic
24:28into the into nervous system.
24:30It's a treatment for Ms.
24:32blocking T cell traffic and you can see
24:35that it's expressed predominantly in the T
24:38cells and transition CSL against the XR3,
24:41KEMA concord and fatty cell
24:43entry predominant expressed.
24:45It's expressed only really
24:47in PCNSL CR7 Express 9 T.
24:51So she excluded from the brain
24:53and you can see that it's almost
24:55exclusively in the peripheral blood.
24:58So we've found nine clusters
24:59of T cells and spinal fluid.
25:01We use the gene expressions imputed
25:04with something called magic because
25:06basically looked at gene expression
25:09across the gene expression patterns.
25:11You see different patterns
25:13across different tissue scores.
25:14Again,
25:15all the details in the paper have
25:16the 9 clusters and I want to start
25:18with the teach one cluster and
25:20think about single cell data.
25:22There's so many things you
25:23can explore with it.
25:25So to me,
25:26I find a few interesting stories,
25:27validate them biologically,
25:29get the data out in the literature,
25:31and I've been so pleased by how many
25:33others have taken our data and use
25:35it for really important experiment.
25:37That's what we're trying to do.
25:38So even treat bothers TH1 cluster.
25:41So if you look at the TH with the CD
25:44four cells to different populations,
25:47the CSF 3 populations particular
25:49and high amounts of gabito Ferrum
25:52tibets CXCR 3 run 3 stat.
25:544, so an aisle 12 receptor,
25:57another CD4 population of tissue
26:00resident markers like lag 3CD-69
26:03and PRDM 1 interestingly enough
26:06and markets have soda toxicity,
26:08a grandson and grandson K,
26:10which is interesting.
26:12Colleague Michael Brenner at Harvard
26:14recently identified grandson Kay
26:16as being involved in complement
26:18deposition and finally CDA population
26:20markets at tissue residence and,
26:22not surprising cytotoxicity.
26:26So to summarize a lot of data,
26:28what do we find in the blood cells
26:31excluded from entering the CSF and
26:33we found cells that are rich for
26:36for traits necessary for entry
26:38and then finally markers for CSF
26:41entry and tissue dependent changes.
26:44So in CSF we found gamma interferon
26:47signature rest in T cells,
26:49cholesterol homeostasis,
26:50TGIF beta pathway and these
26:53cohabitated receptors will get
26:55to that in just a moment.
26:57So then the question is,
26:58do we actually see gamma difference
27:00creating T cells from spinal fluid?
27:03So first we looked at PPD one.
27:05So this is from another three healthy
27:08subjects looking at blood versus CSF.
27:10This is no stimulation,
27:114 hours of stimulation with PM out of
27:14mice and see this market expression of PD1.
27:17And see itself compared to blood
27:20with stimulation goes even higher.
27:22So there is very high expression of this
27:25Co inhibitory stepter in spinal fluid cells.
27:28We then looked at that together,
27:29interferon response with blood and
27:31you can see this major gamma signature
27:34recapitulating what we found with
27:37the RNC data and compared to blood.
27:40So this is major gamma signature
27:42and T cells from spinal fluid.
27:44I'll show you another experiment
27:46which I found interesting.
27:47This is work we've done.
27:48We did looking at PD1 glioblastoma
27:51and basically we took the PD1 high,
27:54PD1 intermediate,
27:55PD one negative and total T cells
27:58stimulated them and not surprisingly
28:00PD1 high cells do not enter cell cycle.
28:04But they did make gamuts to pharon.
28:073% to 50 to over 50%.
28:10So it suggests to us it's phenocopies,
28:13what we see in the brain,
28:14the cells in the brain,
28:15they're condition in the brain have
28:18high amounts of combinatory molecules
28:20that make gaming jefferon and we wonder
28:23is this what immune privilege is?
28:25Is that what if you privilege
28:27the high expression?
28:28Comunitar molecules can enter the cell cycle,
28:31but they are functional.
28:33So now we know normal spot on the floor.
28:35What about multiple sclerosis?
28:36Did the single cell analysis
28:39say looking at the populations
28:41between healthy blood and Ms.
28:43or no difference?
28:44Not surprising.
28:45We never found any differences before.
28:48But if we look at log fold changes,
28:49I'll just highlight some of them
28:51were just beginning to work out
28:53what these different factors being.
28:54We're intrigued by mallet.
28:56Mallet one which is involved in
28:59gene expression and epigenetic
29:01modulation of gene expression.
29:03We found I 32 it's a pro
29:07inflammatory cytokine induces
29:08TNF alpha associated with Ms.
29:11And we found June a part of the AP1 bonding.
29:16I won't show all these
29:17dates to talk in itself,
29:18but we looked at healthy Ms.
29:20versus non expanded versus expanded itself.
29:24It's tremendous clonal expansion
29:26these cells and again looking at
29:29the different populations within
29:32the aisle 32AP1 and there was
29:34more distinct in the clone expand
29:37itself both in CD4 and CD8 cells.
29:39I think the next decade will
29:41be taking these various.
29:42A fact is we found replicating
29:44them by protein and then seeing
29:46how they involved in Ms.
29:48induction.
29:49But this is really a road map as genetics
29:52work road map for what drives and drives
29:55autoimmunity in the nervous system.
29:57And of course,
29:57what about the brain?
29:58I couldn't resist being a pathology group.
30:02So here we did characterize T cells
30:05in normal human brain prank them up
30:08here different cluster patients.
30:09Some of these involve fresh RNC from
30:13brain that provided by Jack Intel
30:16doing epilepsy surgery and here
30:19are the T cells over here at the
30:21very end and different populations.
30:23While summarizing here we're
30:25looking at the tissue residence and
30:27functional gene expression and T cell
30:30with normal brain prank comma and.
30:32You know,
30:33I mentioned we saw this game
30:35interferon signature.
30:36In the spinal fluid and we see here
30:39in the brain this is RNA seek up
30:42T cells right out of the brain.
30:45Here is with Duke seek his major
30:47gaming different signature.
30:48So these data indicate that the gamma
30:51different signature is there in the brain.
30:53Speculate at the end what that
30:56might be doing also Joe work from
30:58Tomo samita done with Andrew Wang
31:01out loud that's predominant humans,
31:03we do mice, we work with vice people.
31:06And this is data from a few months ago and
31:08she's getting barrage of write her thesis.
31:11But basically we wanted to look at T cells
31:13isolated directly from the house brain.
31:16So we saw this in humans.
31:17The question is to see it in mouse brain.
31:20And so here we're looking at T cells
31:23isolated directly from parenchymal tissue.
31:25We wash out the vessels
31:27and work with HomeGoods.
31:28Lander suggests he T cells
31:29are in the prank him up.
31:31He's gamma to Fearon on the X axis.
31:36Case in the CD three you can see this
31:38prominent cabinet different signature in
31:41CD4CD8 cells as we saw in the brain with
31:4440% that cells are making game interferon.
31:47You can see this here,
31:48but you don't see it,
31:50you don't see it in the
31:51in the peripheral blood,
31:53you only see it in the nervous system.
31:55So it suggests that these Gavin
31:57difference between T cells are
31:59physiologic and won't show that the data.
32:02But if you put if you do this
32:04experiments and germ free.
32:06Animals work done with Noah Palm,
32:08they don't make these cells.
32:10These gamma different secreting T cells
32:12are being driven by gut microbiome
32:15and if you label these T cells.
32:17In the gut with the dye that turns
32:20color with the fluorescent probe,
32:22you can show that all these cells in
32:24the brain are coming from the gut,
32:26similar to what BJ's Country did.
32:29And E model.
32:30But this is normal Physiology,
32:33so we can speculate why is it that T
32:36cells in normal central nervous system?
32:39Our country from the cotton
32:40what are they doing there?
32:41Nature doesn't do this for no reason.
32:44I'm sort of speculate that maybe at
32:46night when you have the lymphatics and
32:48you clean your brain out these T cells
32:51that surfing along secreting gamma
32:53influencing the microglia experiments
32:54were battery to begin it's to look
32:57at Tibet gamma knockouts to see what
33:00happens to synaptic pruning and what it
33:03does to the microglia influencing the
33:06neuronal excellent interactions and by.
33:08In summary,
33:09so is there.
33:11Is there glial tea sub
33:13communications circuits?
33:14This is again from light
33:15Ben Barris sowing TGF.
33:17Beta and cholesterol drove these together.
33:20And after I drive TJF,
33:22painting classes are required for my
33:24survival and these sizeable factors are
33:26produced by astrocytes including cholesterol,
33:28lipoprotein which are found in CSF.
33:31I won't show the data but if you take
33:33spinal fluid or cholesterol and TGF beta,
33:36it drives gamma interferon
33:37almost as much as Isle 12.
33:39So we think the cholesterol in the
33:42brain brains what mainly cholesterol
33:44is driving this gamma driving this
33:47gamma and what it's unknown function.
33:49Gamma interferon in the CNS.
33:52It's known to involve in chemical
33:55production record Plexus.
33:56Gamma has known neuroprotective
33:58function like glutamate clearance,
34:00neuronal survival.
34:01We speculate synaptic pruning
34:04and work done by Yoni Kipness
34:06published a few years ago in nature.
34:09So if you knock out gamma to Ferron in mice,
34:12they developed depression.
34:13How do you measure depression mice?
34:14I don't know, but they clearly had changes
34:17in behavior until they meet them working.
34:20That with a number of individuals
34:22and psychiatry department I can
34:23tell you that if you put animals the
34:25germ free environment and get rid
34:27of these games to creating cells
34:29that market changes in behavior.
34:31So it is psychiatrist one field,
34:33neurologist another.
34:34But what's beginning to happen feels
34:37are colliding of course centered
34:39around their inflammation but to me
34:41the normal Physiology the discovery
34:43of gamma difference between T cells
34:45that are normal Physiology which arose
34:48from studying disease is actually.
34:50More interestingly enough,
34:51observations please to how how things work.
34:56And one other experiment again which
34:58I didn't show the data for is you can
35:01ask the question well what if he's
35:02teased cell receptor as a barcode?
35:04To look at this identical T cell
35:07in spinal fluid versus the blood.
35:10Does it change or is there
35:12a selective migration?
35:13The answer is it changes.
35:15The T cells in the blood that share
35:18the same T cell receptor sequence with
35:21cells in the CSF have very different
35:24characteristics the CSF cells.
35:26Have the gamma signature and other PD.
35:29One signature with the identical
35:31T cell in the blood does not
35:33have markets knife cells,
35:35so this provides strong evidence that
35:37what's happening as the T cells migrate
35:39into the central nervous system,
35:41they're acquiring these phenotypes.
35:44So in summary,
35:45all immune disorders are complex
35:48complex genetic diseases where genetic
35:50variants mapped to the immune system
35:52in MSB cells drive the inflamed
35:55Mon reactive CD4 cells instead.
35:57Let me just comment on EB for a moment.
36:00May have heard that beautiful paper
36:02by Alberto Mascaro clearly showing
36:05that if you are a he looked at
36:081,000,000 recruits in the army.
36:10Identified individuals are EB
36:12negative and follow.
36:14Deerfield their light chains come CPK
36:16the brain shows brain damage and he
36:20showed that the on the serial samples
36:22they collected that when NFL went up
36:25in the serum followed by diagnosis of Ms.
36:28it was 49 and 50 ton
36:31preceded by EB infection.
36:32You have to look at a million
36:34people to find that.
36:35But incredibly provocative
36:37data that's EV trigger Ms.
36:40what's the experiment we need to do.
36:42There's one key experiment.
36:44Which we're working on,
36:46which is to vaccinate patients at risk.
36:49So there's no EV vaccine out there now?
36:53Ohh Danner GSK have one and I'm on a
36:55group of devices trying to convince
36:57GSK to do a subset with the clinical
36:59trial patients at risk that we
37:01can vaccinate and prevent disease
37:03as we prevented SP with measles
37:05vaccination that be the defendant
37:08rather definitive evidence we've
37:10not been able to find any biology
37:12B we've looked in the brain of Ms.
37:15patient just interviewing our graduate
37:17student he said you haven't published
37:19much on EV why haven't you go we've
37:21been looking we haven't found any.
37:23In your own publishing that data right.
37:25But what we have your IRB approval
37:28to do now and start shortly it's
37:31a do tonsil aspirates it,
37:33wants it in this patience.
37:34EB lives in the nasal pharynx.
37:37Then you have any ideas how to
37:39do this but we do singles RDC can
37:41do CV expression so we can fund
37:43the EB signature we once and Ms.
37:45patients but of course there may
37:47be gone by the time we do it.
37:49So once again summary of the
37:52talk autoimmune disorders.
37:53Particular mass or complex genetic diseases,
37:56genetic variance mapped to the
37:58immune system and MSB cells.
38:00We believe Dr.
38:01Inflame months specific CD4 cells in the CNS.
38:05Also mentioned that we're doing
38:07single cell RNA sequencing pre
38:08post treatment THC stated about
38:10two weeks all we've been spending
38:12now year analyzing the data but
38:14the most promising we're seeing
38:16would be cell depletion myeloid,
38:19express myeloid, induction of TNF.
38:22You might say, are you kidding me?
38:24It's working by inducing
38:27inflammatory cytokine.
38:28But go back to the clinical trial,
38:30anti TNF makes Ms.
38:32work works great in IBD and RA.
38:35So the data is now suggesting that
38:37TNF induced by the B cell depletion
38:40is leading to TNF secretion which
38:43then induces increased T Reg
38:45function with TNFR 2 receptor
38:47doing the single cell analysis.
38:49So we'll see where that goes.
38:51Yeah I showed you the T cell
38:53traffic between blood and spinal
38:54fluid and brain tightly regulated
38:56showed the TH one signature and
38:58that the blood and sees that for
38:59functionally different indicating
39:01the CNS shapes homeostatic T cell
39:03states that happens all the tissues
39:05at T cell center in the tissues.
39:08This is phenotypic difference
39:09in healthy and masks and control
39:11particularly more expanded cells and
39:13we're beginning to explore what these
39:15are what they mean and this help us
39:17teach one signature seen healthy
39:19brain and of course with the PRT.
39:21And positive teabags think we may have
39:23identified a major transcriptional
39:25factor driving autoimmunity.
39:28So let me end by thanking I I put
39:31here the members of the lab who made
39:35major contributions work Jenna Pappalardo,
39:37who is now out in West Coast,
39:40an industry.
39:41Tomo, who's assistant professor Tomia,
39:43graduate student,
39:44please thank assistant professor
39:46in our department.
39:47Others in the lab now is the former lab
39:49members who contributed the work I discussed.
39:52Today,
39:52Matt Lincoln and the PRD one projects
39:55computational Margot who did the
39:57work with the TH1 T Rex phone
39:59contact and the Center colleagues
40:01of the Broad Institute.
40:02In particular Brad Bernstein
40:04analyst Kellison,
40:05Chuck Epstein who worked with us,
40:07the PRD one project collaborators
40:09at the MI Year in Yale Genetics and
40:12Marcelo and Yang who we feel are
40:14part of our Neuro inflammation group.
40:17So thank you for your time and just
40:19say here's my e-mail and here's
40:21our last picture of our.
40:22Ms.
40:22Group if we have a lot of meetings
40:25every week, all winter, not really.
40:27But anyway,
40:28thank you for giving me the
40:29opportunity to talk to.
40:30I really appreciate it.
40:43So do we know who's in the chat?
40:46That is OK. Good story here.
40:55OK, question.
40:58Thank you for a great thank you.
41:01What about this are those? Who
41:05are they?
41:11You should ask.
41:13Unable to defrag the hard problem.
41:18Here's how to approach.
41:19Do you have a collaboration?
41:23Change. And what we're doing is we're
41:27taking the T cell receptor, thousands of
41:29T cell receptors, popping them into
41:31reporter cell lines and then using
41:33antigen libraries see what they react to.
41:36We're also have tetramers loaded with
41:39different peptide libraries barcoded.
41:42We do single cell and pull them
41:43out and see what they're reacting
41:45with so far, guess what?
41:46Answer we're finding the spinal fluid
41:48across different patient, anything else?
41:52Beebe. Whether it's primarily I
41:56don't know but we also see that the
41:58activity and what we're doing you
42:00know we can I can tell you that
42:02team cell rachamim based approach
42:04that's and tells me one thing right.
42:06So you'd like non hypothetical need
42:08approaches so we're halfway in the
42:10project and hopefully hear so the
42:12better idea to do the same thing
42:13with cancer antigens or they'll know
42:15and guess what they're recognizing.
42:18Don't just rent apart.
42:24The train.
42:31Yeah, the traffic.
42:36Of the possible.
42:41Requires.
42:48Or.
42:50Stop it.
42:56Sure.
43:07What happens with the message? All Star.
43:19K1 did not qualify. In the South. So.
43:27But since it's increased solid,
43:30expect to see it. But you know.
43:35And then?
43:38And yes, it's laughing.
43:43Disease.
43:45That's where we had she is thought
43:48to subtractive stuff because
43:50these papers the post office.
43:54How much is that?
43:57That there's whereas I think
43:59we have a good working model
44:02for relaxing the EMS not yet to
44:04discover but a good model I have
44:06no idea where cost of progressive.
44:10Before we had trees
44:11that will be 50%.
44:15Now the most important question
44:18here is that we have a plan.
44:23Question.
44:26The station.
44:29Progressive disease.
44:31But that's a major question by suspected
44:34to be progressive disease cells stay
44:36there trapped in the back and forth
44:39that we might see on the sausage.
44:42Is it really correct?
44:43That's terrific questions and
44:45we'll go back and then you have to.
44:48The second version.
44:51This is really interesting people,
44:53normal people have T cells are going.
44:56They're making their gambling responsibly.
45:02Yeah, So what are the time scales involved?
45:07T cell recognizes something
45:09that wasn't changed.
45:10You might not know, obviously, but you.
45:12Visioning that it's terms of making
45:15a change there and then having
45:16some response back afterwards,
45:19well, I I could do it my side.
45:24New brand. I can get
45:26spinal fluid but drain it.
45:29But now that it's very
45:30radically so, it happens. If
45:31we so pre, we don't see them at
45:34the time of reading when you start
45:37acquiring the microphone and
45:39that's when we started seeing it.
45:40That happens very quickly.
45:42So we label them,
45:43I can answer that question.
45:44We might label them in
45:47the gut within 2-3 days.
45:49So, so the experiment I did as
45:51as a postdoc.
45:53The terrific experiment.
45:54There's some new thing called
45:57monoclonal antibodies and we
45:58want to put them into people.
45:59No one has really done
46:00that yet. So we're kind of cowboy.
46:03Please stop the recording now. So I've got 5.
46:09So what we did was
46:10we had invited.
46:15Two and I did it with my 5.
46:21Make these. And we said, Gee,
46:24if we don't find material across state lines,
46:27that would be Massachusetts.
46:29You don't need that TA approval.
46:32That now, but you know IRB approval.
46:35So we're very careful what we did,
46:37we had RV approval and what
46:39we do and so we injected.
46:43And to our patients and we showed
46:46that major biological effects
46:48but they're now sanctified
46:50after one child human anti mouse
46:52antibody would deactivate them. So
46:55we only had to make.
46:58But 11 experiments that anti CD 2
47:02N IG23 coded all the T cell but
47:04did not cross into the central
47:06nervous system showed that.
47:08So I would do this thing.
47:10That's a new call, Microsoft you guys.
47:13There was one solicitation to
47:15Harvard Medical School at the
47:16time and I did the first thing
47:19pretreatment did with respondent.
47:20Few days respond that did the same thing.
47:23We got anti mouse and everything but.
47:25Then we did this that it's standing
47:28after the treatment and majority of
47:30cells lit off the coat anti mass and
47:32said what's going on here? Let's just
47:33screw it up again.
47:37Oh my God they're covered with
47:39antibody with mouse antibody.
47:40So we use this way of labeling all
47:43the peripheral blood T cell and that's
47:46being traffic into the CNS policy.
47:48And because everyone looked
47:50at the blood brain barrier,
47:5180% of the cells traffic within three days.
47:56How did they do it?
47:59Found the entry right before crossing.
48:01Well, because when we took
48:03that we couldn't find the party even though.
48:10No.
48:13Now that doesn't listen all
48:16the circulating T cells.
48:23Invite them. So it's suggested
48:26you can follow the connection.
48:30So it's suggested and that nice people
48:33have gone to replicate that more output.
48:36So the traffic of T cells from the blood
48:39to the nervous system side is very fast.
48:42And I think it's continuing this intro data,
48:43but there are three different
48:45scenes of population in CSF.
48:47One and three are about basically residents.
48:50LCF 2 looks like sales and traffic,
48:54so one opposition doesn't have CD-69
48:57was like a population is garden.
48:59We have a question in chat. Yes, yeah.
49:03What's the concordance of Ms.
49:05and identical twins?
49:06That might be a good group in
49:07which to consider prophylaxis.
49:09And a second related question,
49:10is there any handle on what
49:11causes spontaneous remission?
49:13And by these days goes untreated,
49:15Nope, no one goes and I want to
49:17comes to Yelp goes untreated.
49:18Well occasionally they don't
49:19want to do a patient want.
49:21So the answer Jeff is about
49:233040% and yes that would be a
49:25great group to consider for
49:27prophylaxis but it's a small number,
49:30it's two small numbers.
49:31So what we're doing in the Rs
49:33study is developing the tools
49:35to develop to identify average
49:37patients of intake first.
49:38So the question is what is the incident Ms.
49:41in daughters of patients with Ms.
49:43about one in.
49:4430 It's quite high and we can do
49:47polygenic risk score and increase
49:48it even more so and then we can
49:51use NFL we were fully light chains
49:53to follow them so that's why I'm
49:56proposing GSK first as a subset take
49:595000 and first degree relatives
50:01children at risk before they become
50:03EB positive seriously follow them
50:05with NFL NFL go up MRI them I think
50:09that's attractable study to do.
50:10That's how we're going to try to
50:12do it will cause spontaneous.
50:14So what So what relapse,
50:15it's a really good question.
50:17What happens I think in Ms.
50:19is that it's there's an acute
50:21event there's the attacks occur
50:23very quickly within a day within
50:26forty 2448 hours they come on,
50:27it's T cell trafficking into the CNS.
50:30There is Dima that breakdown,
50:32the barbarian barrier,
50:34gadolinium enhancement.
50:35I think it's the edema that's
50:37causing neurologic symptoms with
50:39time there is retraction,
50:41edema goes away,
50:41the blood brain barrier is closed
50:43and rather than having a big lesion.
50:45With a tiny scar,
50:46like it's just just normal Physiology
50:49of of the lesions resolving as
50:51edema goes away and steroids
50:52makes that happen faster.
50:54I think that's what's happening.
50:58That one. Alright.
51:02Well, thank you very much to David.