How MNR Can Assist Us in Understanding Disease

February 28, 2022

Information

Feb. 24, 2022

Yale Pathology Grand Rounds

Stephen C. Meredith, MD, PhD

ID7487

To CiteDCA Citation Guide

00:00I'm gonna give Steve plenty of
00:03time to share with us his talk,
00:06but let me say at the outset that it's
00:08a little daunting to introduce Doctor
00:11Meredith to you and do him justice,
00:14but I'll give it a shot.
00:15So Steve was educated at Brandeis University
00:18in both biology and English literature,
00:21and then he next went to
00:24Washington University in in medical
00:27school after medical school,
00:28he came up to University of
00:30Chicago in the mid 70s.
00:31In a joint residency in Pathology
00:34and PhD program in Biochemistry,
00:38where he studied emoji with the
00:41noted biochemist Francois Kezdy,
00:43who was sort of a central European
00:47figure who came to Chicago and and did
00:50some Seminole work in carboxypeptidase,
00:53I think and was on faculty.
00:55So Steve after following his education
00:58in both pathology and biochemistry,
01:01he was hired on to the Department
01:03of Pathology.
01:04As a junior faculty member in the
01:06early 80s and he has developed an
01:10international career in studying
01:12disease from both a philosophical
01:14and scientific perspective,
01:16and so let me flesh that out for you a bit.
01:19Doctor Meredith holds appointments
01:21in the Divinity school, the College,
01:24the Graduate School in Biochemistry and
01:27Biophysics, and the medical school,
01:28and he teaches in all of these
01:31departments actively in the
01:33Divinity School and college.
01:34He has focused his career on
01:37looking at disease philosophically.
01:39Specifically the problem of pain and evil,
01:43and he has taught for 20 years or more.
01:48Courses in the Division school
01:49and college on James Joyce,
01:51Ulysses Thomas Mons, Magic Mountain, St.
01:56Thomas Aquinas,
01:57problem of evil and other brothers
02:01K Dusty Zewski and he teaches these
02:04courses regularly and they are a
02:07well attended major courses in in in
02:09these schools and he has a sort of
02:13an international reputation looking
02:14at disease from this perspective.
02:17And now today we're going to hear
02:20about his other half of him,
02:22which is his scientific pursuits
02:25in in the in the medical school
02:27and and the biophysics department.
02:30He has taught for 20 years the
02:32foundational curriculum in medical school,
02:34including the courses on biochemistry
02:37and the courses on cellular
02:40pathology and immunology.
02:43He runs an active lab that has been
02:46independently funded for 40 years
02:48and he investigates the pathogenesis
02:50of disease using biophysical
02:52and biochemical techniques,
02:54specifically focused on structure,
02:57determination, and disease pathogenesis.
03:00He's used arcane,
03:02you know,
03:02difficult techniques such as solid
03:04state NMR and has made Seminole
03:07contributions in our understanding
03:09of the structure of beta amyloid.
03:11Which was really undetermined until
03:14Steve and his group of collaborators
03:17did it in the early 2000s.
03:19He has also actively studied tumor
03:23immunology with Hans Schreiber and,
03:25specifically the development of
03:28purification of tumor antigens for
03:31cancer vaccines and his work in
03:34peptide design and peptide structure
03:36has been Seminole to my career.
03:39In in in my journey as a.
03:42Translational.
03:42Biologists developing enzyme
03:45therapies for diseases.
03:47Steve has had a remarkable career
03:50and it's he's a kind and gentle
03:53mentor and an insightful scholar.
03:56And it's my pleasure to introduce you
03:59to him and to hear his talk today,
04:02which is on NMR and disease.
04:05So Steve take it away.
04:07Thank you Demetrius, that was a kind,
04:10even embarrassing introduction.
04:12Thank you also for the invitation
04:16and it's great to be here.
04:19Catch up with some old friends and share
04:25this this scientific data with you.
04:29Too bad it couldn't be in person,
04:31but as they say,
04:32half a loaf is better than none.
04:34And it's great to be able to talk to you.
04:38I'm gonna share my screen.
04:41And hopefully this will work again.
04:43Yes, there it is OK.
04:46So I started out my professional
04:50life as an anatomic pathologist.
04:52I still consider myself one even
04:55though the clinical work as you age,
04:58you have to give up some things
05:01and that has gone by the wayside.
05:04But when I was doing clinical work,
05:07a lot of it was autopsies,
05:08and I really enjoyed it and thought
05:10it added a lot to medical practice.
05:14In fact, I lament that for the most,
05:19mostly for some very bad reasons,
05:22medical autopsies are now rather uncommon.
05:25I still think that even with
05:28all the incredible radiological
05:29technology that's available,
05:31all the rest of the biological world,
05:35autopsies still have a lot to teach us.
05:38Now, this statement should not be surprising.
05:43Considering how complex the human body is,
05:47clinicians.
05:50Still, amazingly, do get some things wrong.
05:54What's remarkable is how often they
05:57now get it right, but even that said,
06:01autopsies are illuminating,
06:03and I will quote here the great and
06:08pioneering pathologist and anatomist Bashat,
06:11who wrote open up a few corpses.
06:14You will dissipate it once the darkness
06:17that observation alone could not dissipate.
06:20In other words, he was talking about
06:23observation at the bedside.
06:25After death, open up the bodies,
06:27see what is inside.
06:29Now a friend and colleague of mine,
06:33Tobin Sosnik.
06:35In the biochemistry department
06:37likes to say this,
06:38biology works on all length scales.
06:43When I started my residency
06:45in anatomic pathology,
06:46this was sometime before the Polynesian War.
06:51I had this opportunity as Demetrius was
06:54saying to study biochemistry and I did
06:57my work on collagen's collagen structure.
06:59And at the same time I was doing
07:03bone biopsies and I mean the whole
07:06thing going into the operating room,
07:08getting the core, embedding it,
07:11cutting the slices,
07:12staining and then reading it and
07:15doing quantitative histomorphometry.
07:18Now I mentioned that because.
07:21I had on the one hand pathology and on
07:25the other hand protein science and.
07:29When I was working on college and this
07:31was again before the dawn of time,
07:33it was at that time big news that there
07:37was more than one type of collagen.
07:39At last count there were 29.
07:41Now I mentioned this because I think it is.
07:46I will lament about one more thing
07:49and that is that the tension that
07:52sometimes exists within pathology
07:54departments and I'm happy to say this
07:57has been minimal in my own department.
08:00But that tension is describable.
08:05On this slide.
08:08Channeling the song from the
08:10musical Oklahoma,
08:11the the farmer and the cowman shouldn't
08:14be friends and I would say that of
08:17the anatomic pathologists and the
08:20biochemist as someone who throughout
08:23my career has on occasion in both the
08:28pathology and biochemistry departments
08:30sometimes been a pathologist to
08:33the biochemist and sometimes been
08:35a biochemist to the pathologist.
08:37But in the spirit of interdisciplinarity.
08:40What I'm going to talk about for
08:43the range of the talk is 2 areas
08:45of structural biology research in
08:48which I am currently involved,
08:50and I believe that structural
08:52biology does at the atomic level,
08:54what anatomic pathology typically does
08:57at the cellular and macroscopic level,
09:01but it's all looking at structure
09:04to help help us understand disease.
09:07So the topics of the of the of
09:10this talk will be first a little.
09:12A few introductory remarks about
09:15structural biology and the use
09:17of NMR for that.
09:19Then I'll talk mostly about bein
09:22amyloid and Alzheimer's disease.
09:24I will finish with a if there's time.
09:27With a few remarks about a project
09:30I'm involved in with Jerry Turner at
09:33the Brigham on the pollutant and the
09:37infant and inflammatory bowel disease.
09:41OK, so. As I think you all know,
09:44crystallography is still the
09:47mainstay of structural biology.
09:49It was determined it was developed 1st and
09:52it is still an incredibly useful tool for
09:56studying structural biology of proteins,
09:58but somewhere around the 1980s and
10:021990s techniques were developed by NMR.
10:06Come to determine protein structures
10:10and for a while in the early 1990s
10:13it looked as if these would be
10:15two more or less coequal methods
10:18for determining protein structure,
10:21but that is not in fact the
10:23case because if you have your.
10:25As I like to say, garden variety,
10:28small globular protein that you want
10:31some structural information about,
10:33you will almost certainly get there faster
10:37and easier by doing X ray crystallography.
10:41And the reason for that is that there
10:45are limitations in NMR spectroscopy.
10:48Now, of course,
10:49if there were only limitations,
10:51one wouldn't do it at all,
10:53but the limitations.
10:54There are two problems.
10:57The first problem is that
10:59NMR is structural data.
11:00These are some of the techniques
11:02on the top of the slide for
11:05determining protein structure by NMR,
11:07but the two problems are first that NMR
11:11structural data are usually under determined.
11:15In other words,
11:16what you get from NMR is
11:17two kinds of information.
11:19You get interatomic distances
11:22and you get bond torsional angles
11:25and you put them together.
11:27Using molecular dynamics simulation
11:29constrained by the NMR data and
11:32you get a structure, but it's not.
11:35It's not fully determined
11:36or not determined enough.
11:38In many cases there simply aren't
11:40enough interatomic distances and
11:43torsional angles as you would like.
11:46The second big problem is
11:49the size limitation problem.
11:51The sharpness of the peaks depends on
11:54rapid tumbling of the molecule in solution.
11:57Now,
11:57as proteins get bigger and bigger and bigger,
11:59they tumble more slowly and
12:02through faster relaxation.
12:04What that translates into in the
12:06readout is broader and broader peaks,
12:09and eventually you can't
12:10distinguish one peak from another,
12:12so there's a size limitation.
12:15The world record for the molecular
12:17weight of the biggest protein
12:20determined structure determined
12:21by NMR is 900 kilodaltons by
12:24Kurt View Trick for Agro Yell,
12:26but that's a kind of a special case
12:30because that is a hep tumor with a
12:327 fold axis of rotational symmetry,
12:36so you only have one seventh of the
12:38peaks that you'd expect for a protein
12:40of that size if it were a monomer, so.
12:45So as I said.
12:48If there were only limitations,
12:50you'd consign NMR to the dust heap
12:52of history and forget about it,
12:55but obviously that is not the case.
12:58So the question is,
12:59what can NMR do that?
13:01Other structural techniques cannot do,
13:03and I say this even with the
13:07amazing revolution in cryo EM.
13:09The first thing is that
13:11not every protein can be
13:13crystallized now. Even some of the
13:16what I called before globular.
13:18A garden variety, small globular proteins.
13:21Even some of those cannot
13:23be crystallized even today,
13:25but in addition to that I'll
13:27be talking about amyloids,
13:28which are not crystalline and not soluble.
13:32And the second major use of NMR is
13:35to study protein dynamics because
13:38with NMR you can see things in
13:41the protein that you can't see,
13:44oftentimes in crystallography or cryo em.
13:50Now I like to put a human face on disease.
13:54There is real suffering in disease and
13:57on the left you have the patient or
14:00gusta D 51 year old patient that was
14:04examined by Doctor Ellis Alzheimer
14:06in Frankfurt in 1901 and on the
14:10right is Doctor Alzheimer with his
14:12wife and three children and I always
14:15think that when I look at this.
14:18Woman she has this confused and
14:22forlorn expression on her face.
14:25Now as we know what Alzheimer's did,
14:28what Alzheimer did was not only studied
14:32the patient while she was alive,
14:34but studied her brain after she died
14:38of Alzheimer's disease and he used very
14:42innovative techniques of histopathology.
14:46And this is from his drawings
14:49of neurofibrillary tangles.
14:51This is I'm sure familiar to most of you,
14:54but on the top is a normal elderly brain,
14:58not quite normal,
14:59because if it were really normal
15:01it would be inside the skull,
15:02not out of it,
15:05but on the bottom Alzheimer's
15:07disease and you can see the severe
15:09cerebral atrophy when you slice
15:11it open along that plane.
15:14Here you see the atrophy better,
15:17which gives large ventral ventricles.
15:21And here are the two lesions.
15:25That Alzheimer described.
15:29Which are. Trichoplax with a
15:32dense core of beta amyloid and
15:35on the left and on the right.
15:37Neurofibrillary tangles,
15:39made mostly of Tau.
15:41Hyperphosphorylated Tau.
15:45Now beta amyloid is a
15:47series of set of proteins,
15:49most of them having 40 or 42 amino acids.
15:53Most of the proteins having
15:5540 or 42 amino acids.
15:57It is derived from sequential
15:59proteolytic processing of a
16:02single pass transmembrane protein
16:05called amyloid precursor protein.
16:08Non especially imaginative name.
16:11But I want to point out.
16:15Ben amyloid,
16:16with the failure of the some
16:18of the monoclonal antibodies
16:19and other anti amyloid drugs.
16:21People are a little bit sour
16:24about amyloid these days,
16:25but in case you go too far with
16:27that and think that amyloid has
16:29nothing bit amyloid has nothing
16:31to do with Alzheimer's disease.
16:33I include in this slide some
16:36and only some of the many point
16:39mutations within the sequence
16:41of a beta that gives rise to.
16:44Early onset familial Alzheimer's
16:47disease and related diseases.
16:51Now, once secreted,
16:52beta amyloid undergoes a
16:55process of fibril formation,
16:58and it's a process called
17:00nucleation polymerization,
17:01where it first forms a set of
17:04soluble oligomers and most people
17:06believe that these are the most
17:08toxic species and eventually you
17:11get a critical point of nucleation
17:14after which polymerization into
17:16the fibril occurs rapidly and you
17:18can follow this through a set of.
17:21A number of different simple assays.
17:24This one is thioflavin T
17:27fluorescence thioflavin.
17:28T binds to the fibril,
17:30but not much to earlier forms of the peptide.
17:34So you can see here that there's
17:36a lag period during which
17:38nothing seems to be happening.
17:40But in fact, nucleation is happening,
17:42and once that happens,
17:44thioflavin T fluorescence takes
17:47off and then the process finishes.
17:50OK,
17:50why use solid state NMR
17:53as Demetrios mentioned,
17:55this is not a simple technique.
17:56This is not high throughput.
17:58So why bother with it?
18:00Well first of all,
18:02amyloid fibrils are not crystalline,
18:04so you can't do crystallography.
18:06They are not soluble.
18:08You can't use solution anymore and.
18:12I will simply say that cryo EM,
18:14while very useful,
18:16has limitations as well.
18:18Other methods for studying amyloid have
18:21low resolution and even alpha folders.
18:24Fantastic technique as it is.
18:27It's absolutely brilliant but but the
18:30problem here is that it gives fairly
18:33moderate resolute levels of resolution.
18:37And where it falls short is on unstructured
18:42domains or poorly structured domains.
18:47So there is an important role that solid
18:50state NMR can fulfill because it is
18:54both high resolution and highly precise,
18:57and they're in contrast to solution NMR.
19:00There is no size limit of the sample,
19:03infinite size macroscopic size.
19:09Now the main kind of information
19:11that you get out of solid standard,
19:14or is interatomic distances.
19:15I'm not going to go into it,
19:18but the techniques are generally
19:21called dipolar recoupling,
19:23and the strength of the dipolar
19:27coupling between two spin systems,
19:30and we're talking about spin 1/2 nuclei.
19:33C13 and N15,
19:35either homonuclear or heteronuclear,
19:38the strength of the coupling.
19:40Is a 1 / R ^3 relationship
19:44therefore exquisitely sensitive
19:46to interatomic distances?
19:50So to summarize,
19:51some early solid state NMR
19:53results that I was involved in.
19:56We found that a beta 10 to 35A
20:01beta 1 to 40 form fibrils composed
20:04of imperatore in register.
20:06Parallel beta strands,
20:08which was not expected and for
20:11awhile not believed, but now it is.
20:16There were two beta strands segments,
20:18roughly residues 16 to 22,
20:21and 30 to 40 residues,
20:231 to 10 are disorganized residues,
20:2725 to 28 are structured,
20:29but not beta strands.
20:32Now continuing work in the field and
20:35I highlight this famous picture from
20:37a paper by any pet covin Rob T Co.
20:41Showing the structural model of a beta
20:44fibrils in which there are there are
20:47parallel in register beta strands.
20:492 segments.
20:50Connected by a structured but
20:52not non beta strand segment and
20:55then you don't see residues one
20:58through 10 in the Spectra.
21:00So everyone said who RA the structure of
21:04beta amyloid fibrils is now understood.
21:07End of story.
21:08But no,
21:09it is not the end of the story.
21:11Why not?
21:15Well, that's because of fibril polymorphism.
21:19OK, and I'm going to talk now
21:22about fibril polymorphism and
21:23the structure of brain amyloid.
21:26Now when you think of protein folding
21:28in terms of a rugged folding landscape,
21:31a polypeptide starts out its
21:33life as an unstructured and
21:37completely polymorphic peptide,
21:39meaning that all Phi and PSI
21:42torsional angles are possible,
21:44and what folding consists of is
21:47decreasing free energy to that own
21:50lowest level by searching for the
21:53perfect set of fayence I angles.
21:56Now, the reason this is a rough landscape
21:59is that it that it there are certain
22:02local minima kinetic traps if you will.
22:05The polypeptide can get trapped in it.
22:08Now I like to talk about amyloids
22:11as incompetent proteins 'cause they
22:12can never get to the promised land.
22:15Amyloids have a Pisgah view of Palestine,
22:19that is I had to bring bring
22:21James Joyce into it.
22:22They can get close to like Moses
22:24but never reach the promised lands.
22:26And for that reason,
22:29amyloids retain the sum of the
22:32polymorphism of the ensemble.
22:34And this is again from any
22:37Petkova and Rob Teeko,
22:39and this is the Seminole insight
22:41that I'm going to talk about
22:43for thinking about getting at
22:45the structure of brain amyloid.
22:48This polymorphism manifests itself
22:50by the fact that there are structural
22:53differences depending on subtle
22:56differences and fibril isation conditions.
22:59The experiment here was to take
23:02two identical samples of a beta.
23:05Let one sit on lab bench,
23:07let the other one swirl slowly on
23:09a Circulator 1 cycle per second and
23:12you can see in these transmission
23:15pictures that the ones that were
23:18quiescent have these striations.
23:20These fibrils,
23:21whereas the agitated ones look
23:24like twisted ribbons.
23:26Furthermore if you Sonic 8,
23:29each of these fibrils.
23:32And.
23:33Then use those to seed fresh
23:36solutions of a beta.
23:39What you find if you'll
23:41pardon this expression,
23:42is that the seed Trump's the
23:46fiber alization condition.
23:48You get progeny fibrils that resemble
23:52the seeds, not the conditions.
23:54Now, what does that tell you?
23:56That tells you that most of this
24:00heterogeneity occurs during the process
24:02of nucleation because if you provide seeds,
24:05you bypass nucleation.
24:08Now you can also see.
24:09Differences in NMR and I'm
24:11not going to go into this,
24:13but these have very different structures.
24:15These quiescent and agitated fibrils,
24:18as determined by solid state NMR.
24:22So this was the clue that we used to
24:25try to get at the structure of brain
24:29amyloid because you can't grow a mouse,
24:32let alone a human.
24:33By feeding C. 13 and N 15.
24:36So what do you do?
24:39Well,
24:39what we did is we got brain.
24:42Started with brain and then from a
24:45patient with Alzheimer's disease and
24:48then isolated the amyloid biochemically.
24:52And then we use that as the seed
24:55which we put into synthetic C.
24:5813 and 15 beta amyloid.
25:01And then we can do solid state on
25:03the ex vivo or as I prefer to say,
25:06exmore 2O fibrils of a beta.
25:10And so we can ask the question.
25:12Whether the fibrils of Alzheimer's
25:15disease brains are quiescent or
25:19agitated or maybe neither one.
25:22Let me give you a couple of quick
25:25clinical vignettes here. Patient one.
25:28It's a 72 year old woman with a tentative
25:32clinical diagnosis of Lewy body dementia
25:35and primary progressive aphasia,
25:37but at autopsy, that isn't what she had.
25:40She had mild atrophy of the
25:42frontal and parietal lobes,
25:44and neuritic abeyta plaques.
25:47She had diffuse amyloid and
25:50neurofibrillary tangles indicative.
25:52Of Alzheimer's disease.
25:54After a very extensive search,
25:57we were able to identify a single Lewy body.
26:00It says is here rare,
26:02but it's actually one.
26:06Now now I have immunostains, belkowski,
26:10stains and so forth showed very
26:13typical neuritic plaques in the cortex,
26:17and even more in the hippocampus.
26:19There were also neurofibrillary tangles,
26:22so this was diagnosed as Alzheimer's disease.
26:26Now, the first hint that something
26:29strange was going on here.
26:31Aside from the clinical history.
26:34Is that? When you use her
26:37brain derived amyloid to seed,
26:41you get neither of the previous
26:43morphologic patterns in transmission,
26:45and these are sort of straight,
26:48not striated,
26:49not twisted looking fibrils.
26:54Now these are the solid state,
26:56a couple of the solid state INNOMAR
26:59Spectra that you get and I'll point
27:01out a couple of things about them.
27:04First, these are incredibly sharp
27:06lines for a solid state NMR.
27:09This makes a spectroscopist want
27:11to sell of eight or other things.
27:16But now let me go on to patient two.
27:19This is an 85 year old woman with
27:21a clinical diagnosis of probable
27:24Alzheimer's disease, and at autopsy.
27:27Severe Alzheimer's disease,
27:29classic Alzheimer's disease, brain atrophy,
27:32loss of neurons, gliosis granular,
27:35vacuolar degeneration or annual bodies,
27:38juridic plaques, neurofibrillary tangles,
27:41the whole 9 yards.
27:44But now when you see you get yet
27:48another pattern in transmission.
27:51These are kind of irregular fibrils.
27:52They're not striated. They're not twisted.
27:55They're not straight.
27:56They're kind of irregular.
27:58Now.
28:00When you do solid state in a bar of
28:03her brain, seated a beta fibrils,
28:05that's the blue and the red is what
28:08I showed you before from patient one.
28:10What you get is again a single set of sharp
28:14chemical shift and that says that there
28:17is a single predominant fibril structure.
28:21You get the same thing whether you see it
28:24with samples from the occipital lobe or
28:26the frontal lobe, and it's not on here,
28:29but also the temporal lobe.
28:33But the faces in both this
28:35is a carbon carbon spectrum.
28:38Same thing in a carbon nitrogen spectrum.
28:42Patient one and patient two
28:45have different chemical shifts.
28:47Different fibril structures.
28:49This is prima facie a very solid excuse.
28:53The pun evidence for a very different
28:57fibril structure and the the.
29:00This paper also included a structural
29:03model of the fibrils from patient one,
29:07and I'm not going to go into it.
29:08But bottom line here is.
29:10This is very different from either.
29:12Kind of.
29:13All synthetic fibril and I can tell
29:16you now that it's also different
29:19from the fibrils and patient too.
29:21So polymorphism and inherent property of
29:25all amyloid fibrils, not just a beta,
29:29it results from the coexistence of
29:32structurally diverse molecular nuclei.
29:35And it leads to different different patients,
29:38different fibrils,
29:39and not my work,
29:42but other people's work have
29:44identified a structure.
29:46Dysfunction relationship among
29:49patients with Alzheimer's disease.
29:52The surprise that I want to
29:55spend a moment pondering.
29:57Is that despite this tendency
30:01towards polymorphism,
30:03the a beta fibrils seated by
30:06brains from different regions of
30:08the brain are not polymorphic.
30:10You get a single structure.
30:12So why is that?
30:15Here are three possible explanation.
30:18The first one is that you have
30:21that the brain tissue environment
30:24permits only one nucleation process,
30:28but this seems kind of unlikely
30:30because you have different fibril
30:32structures in different patients.
30:35Why would that be different?
30:36Well,
30:36maybe there are different brain
30:39environments in different patients.
30:42Probably there are second,
30:46maybe fibril structures or nucleated,
30:49but then many of them or most of them
30:52are cleared by unknown or partially
30:55known amyloid clearance mechanisms.
30:58And the third possibility
30:59is one that I am fond of,
31:02though that's not this quite
31:04the same as evidence,
31:05and that is that the majority of fibrils
31:08present in the brains at the time of death.
31:11Rise from nucleation of a structure at
31:14a single site and then that structure.
31:18Spreads by fragmentation and transport
31:20of the fibrils from one site to
31:24another site in the brain where
31:26they can serve as seeds for growth
31:29of structurally identical fibrils.
31:31OK Sir,
31:32I want to talk a little more now about
31:35the relationship between disease
31:37phenotype and peptide confirmation.
31:42About 20 years ago we did some
31:47work on some of the point mutant
31:50forms of beta amyloid and this
31:53one is the D23 and mutation.
31:55It's also called the Iowa
31:57mutation and beta amyloid.
31:59With this mutation forms fibrils
32:01about 1000 fold faster than wild type
32:04a data so it's a market difference
32:07but the other interesting difference
32:09or the main interesting difference.
32:12Is that people with this mutation in a
32:16beta have a different clinical phenotype?
32:21People with wild type a day
32:24to get neuritic plaques and.
32:26Cerebral amyloid angiopathy is
32:29a partially overlapping set.
32:32Most people with Alzheimer's disease
32:35and neuritic plaques have some level of
32:39cerebral cerebral amyloid angiopathy,
32:41but now, in the case of D.
32:4323 and a beta, it's the other way around.
32:49The vacante cerebral amyloid angiopathy,
32:53and they mostly die of hemorrhagic stroke.
32:56And yes, they have parenchymal
32:58deposition of a beta,
33:00but a lot of it is in the
33:02form of diffuse deposition.
33:04There are compact plaques, too,
33:06but a lot of it is diffuse.
33:08Now we did some solid state in tomorrow
33:10work that I'm not going to go into,
33:12but again to give you the bottom line,
33:14whereas wild type a beta gives you parallel.
33:19Enregistrer beta sheets.
33:21Hey Beta D 23 N forms,
33:24antiparallel beta sheets Now this is,
33:27it turns out is a metastable intermediate.
33:30If you do repeated citing the there is
33:35polymorphism and eventually the more stable
33:38parallel and register beta sheets win out.
33:42But for a long time you get
33:44antiparallel beta sheets so there is a
33:47structural difference there for sure.
33:49So to investigate this further,
33:52we have done some studies in which we
33:56start again with brain of people who
33:59have died of of Alzheimer's disease and
34:02or see a cerebral amyloid angiopathy.
34:06And we now we isolate either
34:09parenchyma with neuritic plaques.
34:11Or blood vessels from leptomeninges.
34:15Now it is true that of course brain
34:18parenchyma has a small amount of
34:20blood vessels, but as it turns out,
34:23you don't really see the the patterns
34:26in NMR and other techniques that you
34:29get from the blood vessel source.
34:32But anyway,
34:33you do the same kind of procedure.
34:36We harvest the amyloid biochemically
34:39we added to synthetic peptides with
34:42the appropriate isotopic labels.
34:44We get replicate fibrils of of that.
34:50We can now study by solid state NMR.
34:54Now we did a series of patients.
34:56I'm not going to go through each one,
34:57but I'm going to show you for
34:59the point of illustration.
35:01Patient one who is a woman who
35:05had both Alzheimer's disease
35:07and cerebral amyloid angiopathy,
35:10and these are the carbon carbon
35:13Spectra on the left in blue,
35:15and I'll use this color scheme
35:18throughout is the vascular amyloid.
35:20On the right is parenchymal amyloid
35:22and what the next slide shows.
35:24Is some of the chemical shift differences.
35:28There are significant chemical
35:30shift statistically and otherwise
35:33significant differences at many sites
35:35between these two kinds of amyloid.
35:40And if you look at the particular
35:42sites these are in the beta sheets.
35:48And we also did X ray diffraction and
35:51just to summarize this this work,
35:54you can see that in some patients
35:57the blue line, the blue line and the
35:59red line are more or less the same,
36:01but in some the blue is much higher
36:04and that would indicate a higher
36:07degree of structural order in the
36:10vascular than the parenchymal fibrils.
36:13OK, so that's another story we have
36:18vascular versus neuritic plaque,
36:21amyloid and different fibril structures.
36:24Now I want to show another little vignette
36:27about polymorphism and amyloid and start
36:30with the with the question why R&PDAPP mice?
36:35Demented or very demented,
36:38why don't they have a lot
36:40of neurological deficit?
36:41Now there's a lot of reasons
36:43why that might be so.
36:44And of course there are many,
36:47many differences between mice and humans,
36:50and sometimes people forget that,
36:52but they shouldn't.
36:53But I'm going to give you
36:56one more possibility.
36:58This mouse is one of the earliest models of
37:00Paul's mouse models of Alzheimer's disease.
37:02They overexpress one particular
37:04movement form of the eight PP,
37:08which is the Indiana mutant form of it.
37:11By about four months,
37:13they start to get cognitive defects,
37:15but they're really not all that severe.
37:18In spite of that, they have basically
37:20amyloid coming out of their ears.
37:22Enormous extracellular,
37:24a beta deposition and some other lesions.
37:28They also have tell the effects.
37:31Oh Hyperfest 4 later.
37:36Tell defects. So the question we're asking
37:42them is what kind of fibrils is there a
37:46difference in the fibril structure that
37:48leads to differences in pathogenesis?
37:51This is from the paper of Riley ET al,
37:54just to show you just how much
37:56Emma Lloyd there is in their mouse
37:58hippocampus and other areas as well.
38:01OK, so if you compare.
38:05Human and mouse a beta.
38:08There are three structural amino acid
38:13point differences in the sequence.
38:18And this is a third species I was
38:20attracted to it for the obvious reason,
38:23but it stands for naked mole rat,
38:26not nuclear magnetic resonance and.
38:31There's a picture of it. Uhm?
38:36You can see that when you fiber
38:38allies these three peptides,
38:39we made it, we made fibers.
38:40We did transmission again and right off
38:43the bat you can see that there are under
38:46a given set of fibril isation conditions.
38:48You get fibrils with rather
38:51different morphology.
38:52Here you see these fluoride striations
38:55in human a beta 40 twists mostly
38:58in the nick and Mallrat a beta
39:0140 and this kind of mixed with a
39:04lot of very thin fibrils.
39:06In a beta mouse, a beta 40.
39:10The reason for including the naked mole
39:13rat is that it is the longest lived rodent.
39:19It is insensitive to pain and it is
39:23supposedly resistant to the effects
39:25of a beta and Alzheimer's disease.
39:29I say this is a story that
39:32is still developing.
39:33OK,
39:34in any case,
39:35if you now take the mouse and human
39:38a beta and you compare them and this
39:42is data that's hot off the presses,
39:45so I don't have a structure for this,
39:48it is coming,
39:48but I think it's pretty clear
39:50that there are going to be major
39:53structural differences between
39:54the Hughes human and mouse a beta,
39:56despite having been fibril eisd
40:01under identical conditions.
40:04And this slide to show up a closeup
40:06of 1 region of the spectrum.
40:08If you look at the upper right
40:11of this spectrum.
40:12This is an alanine peak for what it for,
40:15what it matters, but there's a.
40:17There's a big difference in this region
40:20as well with leucine and glutamine peaks,
40:23so a big difference as well.
40:25So we're in the process of
40:27doing the assignment.
40:28I'm not absolutely sure
40:29about the assignments yet,
40:30but this is very exciting research.
40:33That is work in progress.
40:36OK,
40:36in the last few minutes of the talk
40:38I'm going to switch and talk about
40:40the other story I mentioned earlier,
40:43which is how NMR can be used to study that.
40:46And I'm the dynamics of proteins
40:49in particular of occluding and
40:51it's alpha helical bundle.
40:54The reason for interest in this
40:56is that this is a tight junction
40:59protein in which one particular
41:02residue residue sering 408,
41:04seems to be very important.
41:06In regulating tight junction function.
41:10Now this is from a paper by
41:13David Raleigh and Jerry Turner
41:15collaborator in this work,
41:18and I'm not going to go
41:20through their whole model.
41:21They have tons and tons of data.
41:23Beautiful data on this,
41:25except I'll point out one part of it.
41:27And that is that sehring 508 of occluding.
41:33Can be phosphorylated or not.
41:35It gets phosphorylated by CK2 and
41:38if it is not phosphorylated it
41:41binds zielone and and that makes the
41:49the tight junction more permeable
41:51to water and ions so you get
41:54leakage across the tight junction.
41:59Now we wanted to investigate the
42:01structural basis for this difference,
42:03and as I'll show you,
42:05there's a big problem here.
42:07This slide won't go into it in detail,
42:10but let me go back here.
42:14This slide shows that occludin has a
42:18membrane spanning region for membrane
42:20spanning helices and then it has a
42:23tail that is cytosolic or cytoplasmic.
42:27Now this slide shows the evolutionary
42:31conservation of occludin in its
42:35cytoplasmic domain and it is 100%
42:39identical throughout mammalian evolution,
42:41and he and greater than 90%
42:44conserved in birds,
42:45amphibians and fish.
42:47And that's absolutely true at this
42:50critical sehring 408.
42:54Now here's the problem. Here is a.
42:57Here is a crystal structure done
43:00by Arnon levy of the helical of the
43:05cytoplasmic portion of occludin.
43:08And as you can see,
43:08it consistently 3A helical bundle.
43:12But what I will point out is the in terminus,
43:15in which many of the residues are not seen.
43:18In fact, there's about 30 residues that
43:21are not seen in the crystal structure.
43:24So here's the problem.
43:26You have sehring 408 phosphorylation.
43:29It's crucial for physiological
43:32function of the tight junction.
43:35This region is extremely well
43:38conserved evolutionarily,
43:40but especially the unstructured domain with
43:42serine 408 and the surrounding residues,
43:46and yet sehring 408 is not seen
43:48in the crystal structure. Why?
43:51Because it is in an unstructured domain.
43:54So how do we make sense of this conundrum?
43:57Well, we made a group of peptides
44:00and I'm only going to focus on
44:02the first three of those,
44:04and I'm going to use a little
44:06bit of shorthand here.
44:08Occluding A has a searing 408
44:11substituted by an alanine,
44:13so it's like the D phospho amino acid
44:18serine and occludin with a D at position 408.
44:23I'll call it occlusion D or occlusion 408 D.
44:27And so this is now the same length
44:29as the protein that was used
44:31for the crystal structure.
44:33But with this point mutation
44:35now this third construct.
44:38Lops off the part that isn't seen in the
44:43crystal structure and is unstructured.
44:46Now,
44:47if you express this in acoli,
44:50you get alpha helical structure and
44:54it's the same in both A&D and if you
44:59look at the one with just the helical
45:01bundle the unstructured domain lopped
45:04off slightly higher helical percentage
45:06because you've lopped off this,
45:08the unstructured domain.
45:09But other than that,
45:11these proteins look to have at this low
45:15resolution level very similar structures.
45:18Now this is a.
45:21This is like in HSQC,
45:23except because the protein is very elongated.
45:27We used a technique called.
45:30And trozzi, but it's basically gives
45:32you the same kind of information.
45:36But what this show is first of all,
45:38on the left you see red and blue and
45:40most of the peaks are overlapped.
45:42And this is a clue.
45:44Danae and occludin D.
45:45And with this tells you that is that
45:48even on the very fine structural level,
45:51these proteins have very similar structures,
45:54though not quite identical.
45:56Now, if you compare, on the other hand,
45:59occludin.
45:59D.
46:00You can also do occluding a same
46:03answer if you compare occlusion
46:06D with the Helix only occluding.
46:09Now you see a lot of places where red
46:12and green do not overlap very well,
46:15so this bespeaks large differences
46:18of some kind in the structure.
46:22And you can put some numbers on.
46:23It won't go through the details of it,
46:26but what I want to point out is the
46:29difference in the Y axis scale.
46:31If you compare A&D on top,
46:33these are tiny numbers,
46:35whereas if you compare in this case
46:39D with the Helix only version,
46:43you can see that these numbers are
46:44sometimes quite large, almost one PPM.
46:49Now to get at this further, we used a
46:52technique that was developed by Marius
46:56Floor and Nicholas Fousey which is called.
47:01Oh, which is called paramagnetic relaxation
47:05enhancement and basically what you do
47:08here is you put an electron spin label
47:11on a cysteine by a disulfide bridge.
47:14So you're making mutation at a particular
47:17site in amino acid for assisting,
47:19and then that electron spin causes
47:22rapid relaxation of a nuclear spin.
47:25Now the readout for that is that if the.
47:30Electron Spin label is close to the
47:34nuclear spin that you're looking at.
47:37You see, broadening of the peak,
47:38it gets less intense due to
47:41the paramagnetic label,
47:43so you see decreased peak
47:45volume and peak height.
47:49OK, now I was going to tell you two types
47:52of PRE experiments we did on occludin.
47:56The first kind and the first kind.
47:58We took the the the two,
48:02the two versions of the protein
48:05with the unstructured domain and
48:08now we put a label on either
48:10cysteine 409 which occurs naturally
48:13in the peptide very conveniently,
48:16or we mutated the.
48:19End terminal alanine to assisting
48:21now this is I'm going to call this
48:24insist because the unstructured
48:25domain is attached to the helical
48:28bundle and this arrow indicates
48:30roughly where the little way the
48:33crystallographic structure ends.
48:35Now the second time that experiments,
48:38we take a free peptide and we lay,
48:42and it's either the peptide containing
48:46S 4088 or S408D and then we attach the
48:51paramagnetic spin label to cysteine
48:54409 and naturally occurring, assisting,
48:57and now we add it to the occluding,
49:01in which.
49:02**** be Helix only version of occlusion.
49:08No, I'm just wanna give you the
49:12gross outlines of what we see and if
49:15you look at the overall pattern on
49:18the left of the Paramedic Clinic,
49:21label is at 16509 on the right.
49:25At 8383C you can see that the two
49:29valleys whoops that the two valleys
49:31here are at very different places.
49:34Then then in on the right side picture.
49:37So through mapping like this,
49:40we think that C 409 is roughly
49:44where it's shown in this cartoon
49:46on the bottom, whereas
49:50A383C is roughly at where you see
49:52it in this picture on the bottom.
49:54So basically we're mapping the location
49:57of residues in the unstructured domain.
50:02And this is the experiment intrans
50:05where you get a small difference D
50:08slightly more than a four for the
50:13binding of the peptide to the Helix.
50:18Only version of the peptide of the protein.
50:21Now I have to say that it really
50:24wasn't known before this that the
50:27peptide would bind to the Helix only.
50:30Protein, but we confirmed that
50:33by doing affinity measurements
50:36by microscale thermophoresis,
50:39there is in fact about a six fold
50:42difference of affinity phospho S 408
50:46higher than D phospho 408, and this,
50:49by the way, is the phosphorylated,
50:52not the phospho mimetic.
50:54Now the other thing that is
50:57physiologically very important.
51:01Is that this does lead to a
51:04difference of affinity for the
51:06molecular partner of occluding,
51:09which is 01, and as you can see here,
51:12as predicted by the physiological data,
51:16the S 4088 binds with about a two
51:20fold higher affinity than AS408D.
51:24No, the conclusions.
51:27First, NMR can help us understand disease,
51:30and I've given you some diverse
51:32vignettes to demonstrate that first
51:35there is a structure dysfunction
51:38relationship between amyloid structure
51:40and the Alzheimer's disease phenotype.
51:43This is shown by patient one versus patient.
51:47Two Riddick plaques versus vascular
51:50amyloid mouse versus human amyloid second.
51:54This is important because.
51:57Amyloid a genic peptides can Co fiber
52:00allies as well as Co precipitate.
52:04Alright,
52:04a beta and a synuclein I didn't mention,
52:08but we also did stains for Alpha's
52:11nucleus and this was present in some
52:14places in patient one and there are
52:16probably many more examples of this,
52:19so we all know that there are more
52:21things in heaven and earth than
52:23are dreamt of in our philosophy.
52:25Alzheimer's disease is not just
52:28amyloid period yet there is
52:30amyloid involved in it in a way.
52:35The next lesson from this is that
52:37when you say something is not
52:40seen in the crystal structure,
52:42that does not mean it is unimportant.
52:45It may still yet influence
52:48Physiology and biochemistry,
52:49so the reason for that is that we're not
52:53talking about protein structure along alone.
52:56We're also talking about protein dynamics.
53:00OK, now as I like to say.
53:04When I say I, I mean we and when I say we,
53:07I mean they and I wanna give a shout out to
53:12people who really did the work and these are.
53:17Past and present members of the lab,
53:19from left to right,
53:21J Pittman had tools Rivas Diva,
53:24Jonathan Servic,
53:26Bharat Venkata and Patmore.
53:29And I want to particularly give a
53:32shout out to the tools for Srivastava.
53:35Very talented postdoc who is a superb
53:41biochemist and and spectroscopist
53:45and this is Joe Sock Laban.
53:47Who as the saying goes, knows more about,
53:51has forgotten more about NMR
53:53than I will ever know?
53:56And now other collaborators I
53:58particularly want to thank Rob Tiko,
54:01with whom the brain seated amyloid
54:03work has been done and as well,
54:06a lot of some of the the work on the D.
54:0923 N Jerry Turner,
54:12who has been working on has been working
54:15on tight junctions for a long time.
54:17Though we're doing this work
54:19on occluding and he has,
54:21as I say, a whole system going.
54:23It's wonderful data.
54:26Joseph Orgell doing the fiber diffraction,
54:30the the vascular versus parenchymal
54:33amyloid work was done with Yoshitaka
54:37ISHI and Katherine Chappell.
54:40And I want to also mention gingelly
54:44you at NIH formerly working on
54:47the brain seating and Peter Patel.
54:51Great neuro pathologist in our
54:54department who has who has.
54:56Helped me enormously with his
55:00histopathology diagnosis.
55:02I like to say that science has learned
55:04more from the steam engine than the
55:06steam engine ever learned from science.
55:09Science has also learned more
55:11from disease than disease.
55:13Has learned from science.
55:17The aardvark my my my prediction
55:22for 2002 and I realized
55:24speaking to people in New Haven.
55:25I have to specify that the
55:28White Sox are the good socks
55:31and at that point I will stop.
55:35Right? It's good that
55:37they welcomed your lab.
55:38There must've been a nice lab outing.
55:43Any questions?
55:43I'm sure there are some questions
55:46for Doctor Meredith.
55:47Road open here.
55:48Sure, I'll start. And then we'll go forward.
55:51Stephen, thank you for a wonderful,
55:53very clear presentation. You alluded to
55:56it at the end, although it wasn't
55:58clear when you were doing your data. If
56:01when you harvest these
56:03nucleating fibrils from real
56:06brains, whether it's mouse, human, or humans
56:08with different patients, they're going to
56:10be contaminated at a trace level,
56:12you won't see any NMR with other stuff.
56:14Almost inevitably. Do you think that
56:17is Co nucleating with an accounting
56:20for the different structures?
56:23The different nucleation universe,
56:24which minimum the nucleus
56:26filament finds in this.
56:28In their words, just the
56:29proteins associated with
56:30the A beta. That's really we should
56:32be looking at absolutely absolutely.
56:34I mean, we know what happens
56:38when you have pure peptide.
56:40Right now it's not the same conditions
56:43you can say, and that's true, but it is.
56:46I think it is all about what else is
56:49mixed up in that junk in the lab we refer
56:51to this affectionately as brain goo.
56:54Now what I will say is that
56:56as we do the biochemical.
56:58Purification of the amyloid.
57:01We follow the the NUCLEATING activity,
57:04where the seeding activity So what we
57:07throw away does not have seeding activity.
57:11That said, what does have seating
57:13activity is not pure, not even close.
57:17No, not even close.
57:19So you know if I live another hundred
57:23years or maybe someone else will do this,
57:26I think it would be very important to do
57:28a proteomic analysis of what else is in,
57:31you know,
57:31a fairly systematic proteomic analysis
57:33of what else is in the brain.
57:35Do we know for sure in patient one,
57:38for example,
57:39that she had this diagnosis
57:41of Lewy body disease?
57:42We know that there was at
57:44least one Lewy body,
57:45and there's alpha synuclein by immunostain.
57:48Not a lot,
57:49but some.
57:50Yeah, so can I just follow up?
57:52So why is this any different? Maybe because
57:56you can't eat it any different
57:57than a prion in nucleates formation
58:01of a pathologic. Confirmation
58:04as another friend and colleague
58:06Jim Mastriani likes to say.
58:12Point out that Alzheimer's disease
58:15has been transmitted as a prion.
58:17Now it probably takes a set of special
58:20circumstances, so don't worry too much,
58:24but but I think I think it can be,
58:27and the difference between an amyloid
58:31and prion is basically infectivity.
58:35Data fibrils have very low infectivity.
58:38They don't go through the GI tract very well.
58:42Attempts to transmit it that way.
58:44I think I've not been particularly
58:46successful, and the other thing that happens.
58:48You know that how we got
58:50these monoclonal antibodies?
58:51Is someone tried to transmit this
58:54as prion and got an immune response?
58:56So I think in the case of a beta,
59:00maybe the immune response partially
59:03opposes transmission as a prion,
59:06but that's speculation.
59:07I don't really know.
59:08I think it can be transmitted.
59:11Not to say on this right, of course,
59:13and I I think the question John was
59:15really great because I think that most
59:18people don't know that what's injected
59:20as prion fibrils are actually not pure
59:22and they have many things in them.
59:24So the second thing,
59:25just I'll make my comment about my strone.
59:27I actually do those kinds of experiments
59:30and I'm allowed data is not infection that
59:33says that any kind of what you would call
59:36apprion or infectious agent of TSS is.
59:38That is, it doesn't seriously transmit.
59:41If you put a lot of it into a brain.
59:42The animal. That's very sick.
59:44It may seed,
59:45but it doesn't transmit to a new animal.
59:48OK, so that is that is an important point,
59:51but what I would like to ask you because
59:53I thought your your presentation
59:55was great and very illuminating.
59:57Very instructive for me about
59:58fibroids and how you look at them is
01:00:01if you wanted to think about what
01:00:03makes a a beta amyloid of let's
01:00:06say a prion protein infectious.
01:00:09What would the structure be and what
01:00:11would happen if you found that something
01:00:13that had a very different structure
01:00:15from let's say an infected lymph node?
01:00:18Actually did gave the exact same types of
01:00:22transmission and strain characteristics,
01:00:24which are different which are only
01:00:27for that particular agent and not
01:00:29for something that makes something
01:00:31very similar with a different strain.
01:00:33So I really want to know your inside
01:00:35is what would make an amyloid
01:00:37infectious structurally.
01:00:40Well, I can give you a short
01:00:42answer in a long answer.
01:00:44The short answer is. No idea.
01:00:48The longest somewhat longer answer is it it.
01:00:52First of all, it needs to be.
01:00:54I mean, I think you're talking about
01:00:57the route of entry into the brain.
01:00:59That's the first thing.
01:01:01Why is it that a prion can?
01:01:04I mean a real prion?
01:01:06And I agree with you that being amyloid
01:01:08is it's under some circumstances it can be
01:01:11transmitted from one animal to another,
01:01:14but you have to sort of squirt
01:01:15it directly into the brain,
01:01:17which is not the case for a real prion.
01:01:20But I think you know somehow the
01:01:22prion has to be picked up by the by.
01:01:27For example,
01:01:28the dendritic cells in the
01:01:31nose or the oropharynx.
01:01:33And then hitch a ride on the
01:01:35trigeminal nerve into the brain.
01:01:36Or maybe be able to go through
01:01:38other routes in the GI track.
01:01:39I don't really know how it occurs,
01:01:42but
01:01:42goes by blood. It goes by white
01:01:44blood cells as shown in the 1970s.
01:01:48I think you know this is not
01:01:50necessarily a structural thing.
01:01:52This could be a receptor ligand interaction.
01:01:56I suspect that the different prion strains
01:01:59and of course there are strains of prions
01:02:03that depends on the the differences
01:02:07among them have a lot to do with.
01:02:11How the beta sheets are arranged?
01:02:14In other words, if you do limit digests of.
01:02:18The, let's say the GSS, prion versus.
01:02:23Yakub kreutzfeldt prions you
01:02:25get different limit digest.
01:02:26It suggests that different that they're
01:02:31at different locations of the beta sheets
01:02:33and therefore different relationships.
01:02:353 dimensional relationships
01:02:37among the beta sheets,
01:02:39and that would be really nice to know.
01:02:41I mean,
01:02:42if you get the different prion strains,
01:02:45and I know vittle sure of
01:02:47it is doing work like this.
01:02:49For example in cervid prions
01:02:51and other kinds of prions.
01:02:54So I think that would be very
01:02:55important to know and I'll bet
01:02:57that the different strains have
01:02:58different beta sheet arrangements.
01:03:01Actually, the digests are the same
01:03:03in the organ, but they're very
01:03:06different in the peripheral tissues,
01:03:08so it depends on the organ that they're
01:03:10in and that does not reflect the strain.
01:03:14Well, in that case I really don't know.
01:03:16I just I. I sometimes think that there's
01:03:18a lot of popular notions that are said,
01:03:21and I really think your stuff is interesting,
01:03:23and I think the structural stuff is
01:03:25really interesting and I'm very happy
01:03:26that you said that there's other stuff
01:03:28when you isolate stuff from brain,
01:03:30because most people don't know that.
01:03:33I don't know why they don't
01:03:35know that all you have to do
01:03:37is look at it and it's brown.
01:03:39It's not white and fluffy
01:03:41like a synthetic peptide.
01:03:44Great, well listen, I think we've
01:03:46gone a little bit over time,
01:03:47but I want to thank Steve and for for
01:03:50sharing his work with us and I want
01:03:52to thank you for all your questions
01:03:55and your attention to his talk.
01:03:56Very grateful to you all.
01:04:00Take care everyone and we'll see
01:04:01you next week for grand rounds.
01:04:03Steve stay warm out there in Chicago.
01:04:07I'll see you soon.
01:04:09See you bye bye Many thanks,
01:04:11thank you, thank you. Thank you.