Pathology Grand Rounds: May 11, 2023 - Jennifer C. Jones, MD, PhD

May 11, 2023

Information

Extracellular Frontiers in Health & Disease by Jennifer C. Jones, MD., PhD

ID9920

To CiteDCA Citation Guide

00:00Hi, everyone. Thank you so much
00:02for joining us today. I'm Tiffany,
00:05I'm a fourth year PhD student and
00:07the Pathology Grand Round Student
00:08Committee is very excited to have Dr.
00:11Jennifer Jones here with us today.
00:14It's not. Maybe I'll just use the.
00:18Hello.
00:23Hello. Great. Okay, great.
00:27Thank you again for joining us today.
00:30The Pathology Grand Rand Student Committee
00:32is really excited to host Doctor Jones.
00:34Doctor Jones received her
00:37Bachelor's at Princeton University
00:39and her MD&amp;PHD at Stanford.
00:41She also spends a few years at Harvard
00:43and is currently at the National Center,
00:46the NCI and the CCR at NIH.
00:50And she is also the head of the
00:54Translational Nanobiology section.
00:57Currently, her lab does wonderful
00:59work on extracellular vesicles and
01:01identifying the different types
01:03of vesicles secreted by distinct
01:05tumor types and analyzing how they
01:08affect downstream immune pathways.
01:10She is also working on the
01:12development of the characterization
01:13and analysis of these extracellular
01:15vesicles and we are very excited to
01:17hear more about her research.
01:19So please join me in welcoming Doctor Jones.
01:26So thank you very much.
01:29In one of the chats that I
01:32had with one of your faculty
01:33members earlier this morning,
01:34the comment was made.
01:36This all sounds kind of like a mess,
01:39and actually it is.
01:40And what I'm going to show you is
01:43what we're going to try to make
01:45sense of the mess that is all of
01:48the extracellular structures and
01:49complexes and what they can tell us.
01:52So there is a book called The Commotion
01:57in the Blood that is a bit of a back
02:02story of of how I got started with this.
02:05First, these are the the folks
02:07who have done the the legwork,
02:09the hard work,
02:10the experiments behind these things
02:12that I'm going to show you and.
02:14So they really deserve all the
02:16credit for for getting this done.
02:19And I hope I'll show you
02:21collectively where this ends up.
02:24I'm a radiation oncologist,
02:26so I just wanted to loop you in
02:29on my perspective and interest on
02:34these problems.
02:35There's a a phenomenon called abscopal
02:38immune responses where you radiate one
02:41side of a tumor and the immune system.
02:44Causes other sites to regress.
02:47We have had some clinical trials at at
02:50NIH looking at this and as you know with
02:53immunotherapy it doesn't always work.
02:55In fact, it often doesn't work
02:56as well as you might want it to
02:58and the abscope will response.
03:00The combination of radiation
03:01and immunotherapy works even
03:03less often than that.
03:06So I want to find tools that
03:09help me unpack and understand.
03:11What's happening when those
03:13immune responses are going right,
03:15what we need to do to drive
03:16them in the right direction.
03:18This is the structure of the talk.
03:19I'm going to talk about not just the
03:21motivations that they just went through,
03:23but also the basics about extracellular
03:25vesicles and other extracellular things.
03:27The technology development we've
03:28been doing to crack the nut.
03:30Some basics discoveries that we've
03:32made as we're beginning to actually
03:33start to leverage these tools now
03:35and some conclusions that really
03:37are not really what we expected,
03:39but where we're going.
03:40So for those of you who haven't seen it,
03:42you may or may not have a book
03:44or two in your past that sort of
03:46stoked your interest in the field
03:48that you went into in science.
03:49This was a book in the 90s about
03:52tumor immunology in the early days and
03:55Cole's Toxin and all of these things.
03:57And that really led me to want to
04:00follow a path of studying these things.
04:01Overall,
04:04it's all about cells in the immune
04:06system of the cells in the immune system
04:08and the way that that book frames it.
04:10So there's more in our blood
04:12than just the cells.
04:13There is an assortment of vesicles,
04:16things with lipid bilayers and
04:17various cargo released in different
04:19ways from different cells.
04:20They're also obviously lipoproteins.
04:22They're also ribonuclear proteins complexes.
04:25They're also classes of extracellular
04:28things called exomeres and
04:31supermeres isolated in different
04:33ways from these broadly speaking,
04:37I'm going to focus on.
04:39The vesicles,
04:40with a caveat that I can't guarantee to you
04:44with the technology that anybody is using,
04:48that every single vesicle is actually
04:50vesicle and not just a particle that
04:52happens to have the same density,
04:54size, or other properties, but.
04:57This is this is where we are with the field.
05:00So Ev's are heterogeneous exosomes.
05:03Perche tends to be less than 150 nanometers.
05:07They have protein surface markers.
05:09They have nucleic acid cargo
05:11inside and sometimes outside,
05:13stuck to the surface,
05:15and there's widespread interest in them
05:17from tumors and other types of cells.
05:20People are interested in them because
05:22everybody who's interested in some tissue,
05:25some disease,
05:25some something wants to look at the
05:28vesicles coming from that tissue,
05:29that disease.
05:30And so it's really a great framework
05:33for potentially doing systematic
05:35systems biology if we could be
05:37organized and structured about this.
05:42So I want you to remember this
05:44figure because it's important.
05:45Strictly speaking exosomes come
05:48from multi vesicular bodies.
05:50In the cell and released into
05:53the extracellular space,
05:55other vessels are set from the surface.
05:57Those might be called micro
05:58vessels or microparticles.
05:59But exosomes imply a certain Biogenesis
06:05in the liquid biopsy community.
06:07So most of my clinical colleagues,
06:08when they do exosome studies or
06:10liquid biopsies for biomarkers,
06:12they'll take a biofluid,
06:14they'll isolate the exosomes often with
06:16some kit and then they'll do some other.
06:18Cargo assay RNA or DNA,
06:20some sequencing to identify some biomarkers.
06:25The reason there's so much huge
06:28excitement about this is because
06:31robust consistency in the protocols
06:35and a useful payload in the readouts
06:40has had some big successes.
06:42So Exosome diagnostics has
06:45prostate cancer assays.
06:47Which can help predict the the aggressiveness
06:53and discriminate between high grade
06:55and low grade prostate cancer and
06:58indicate to a patient based on their PS:.
07:01A and this test whether or not
07:04they need to get a biopsy.
07:06This was the result of one of the
07:09randomized control trials where they
07:11looked at this and showed the benefit.
07:15Of combining this test with standard of care.
07:19There have been since 3 randomized
07:21control trials and so I'm not saying
07:23that this kind of assay is not useful.
07:26But if you talk to Johan Skog who
07:29developed these assays you'll be
07:30the first to tell you yes 3 Two of
07:33the three Rna's that they isolated
07:35in the first versions of this test
07:38were not actually in vesicles at all.
07:41They were Co isolated with those.
07:44And so
07:47remember I showed you the figure
07:48where it really means something very
07:50specific in terms of Biogenesis.
07:51We have another part of the field that's
07:54really approaching this strictly as
07:56a I don't know about the Biogenesis,
07:57I don't really care.
07:59I'm taking a blood sample and and
08:02doing a procedural based thing.
08:04So there is an ontology initiative
08:07that we've started because if you
08:09want to create atlases you have
08:11to be speaking the same language.
08:13And I'll get to that at the very
08:15end because I know ontology talks
08:16make everybody fall asleep.
08:17So I just got a couple of slides to to
08:20share with you and thank those of you
08:22who feel about the the survey for us.
08:26But back to the original idea,
08:27So how do we do this systematically
08:31and correctly?
08:32What we want to do is to sort
08:36subsets of the vesicles and then
08:37look at the the cargo and the
08:39messages that those sets contain.
08:43So the state of the field is taking
08:46the whole mess of vesicles from all
08:49the different cells and we want to
08:51parse them out into vesicles from
08:53particular sources and and study those.
08:58So we need to know how do we reliably
09:01identify the specific subsets.
09:04So with immune cell substate sets
09:06you might use CD3 for a T cell.
09:11CD14 for a type of monocyte,
09:14which are the right markers to use for
09:16different types of vesicles and then
09:18how to reliably assay them Once you
09:20know which ones you want to assay,
09:22how do you do the assay in a reliable way.
09:25So we're in an in between space.
09:28There's a lot of cellular biology and
09:30this is pretty mature at this point and
09:32there's a lot of molecular diagnostics.
09:34It's individual molecule assessed in mass,
09:37these are packets and so.
09:40We're looking at packets of informations
09:41and sets of packets of those informations.
09:43So it's a fundamentally different
09:46type of bioinformatics.
09:48And Joshua Welch,
09:49in my group Staff Scientist,
09:52has led the development of three software
09:56tools which help improve the rigor and
09:59reproducibility of three important
10:00tools that we use for characterizing.
10:03Individual E v's and repertoires of E
10:06V's and I'm going to walk through these.
10:08One that's targeted at EV flow
10:11cytometry for single E V's,
10:13resisted pulse sensing that
10:14measures size and concentration,
10:16and MPA pass and Multiplex analysis system,
10:20which is useful for assessing
10:22repertoires broadly.
10:26So these are the specific
10:30tonologies I'm going to talk about.
10:34So for single EV studies we've produced
10:40these advanced protocols for labeling,
10:43sorting, a framework which I'll show
10:46you or how to do and organize and
10:49report the the studies and then also
10:52do your assays in a calibrated way.
10:54So rather than an arbitrary
10:55unit sharing calibrated units,
10:56which if in everybody probably
10:59does philositometry,
11:00you know your skills are actually.
11:01Arbitrary, they're not
11:03calibrated for essence units.
11:07And then I'm going to show you how
11:09we're stepping towards integrating
11:10this into a comprehensive Atlas.
11:12So our labeling protocols and
11:16sorting we're done on the Astrios,
11:20basically a next generation Moflow XDP
11:23den air system with this we showed here.
11:29Remember this picture?
11:29I'm going to come back to it.
11:31These are vesicles that we
11:33can see from DC 2.4 cells.
11:35And this bottom thing is not
11:37a population of vesicles,
11:38it's the noise of the instrument.
11:41So you can see how we're really
11:43hugging the bottom limits of detection.
11:46We can't separate the vesicles
11:48from the noise any better because
11:50of the limits of detection and.
11:52Polystyrene beads can't be used as
11:55calibrators because they refract.
11:56They have a different refractive index,
11:58so they they fundamentally can't
12:01be actual calibrators.
12:03This is what happens when your particles
12:06are lower than the limited fraction.
12:09Fortunately,
12:09HIV and most of our extracellular
12:12vesicles is about the same size,
12:14so we took advantage of that and
12:16took two different HIV variants,
12:20one that was.
12:23CCR5 trophic and one that's CX
12:26CR4 trophic and labeled 1 red,
12:28labeled 1 green,
12:29which showed that we could sort
12:31them and that they remain retain
12:33their trophism for their specific
12:38for their specific type and
12:40the recipient cell line.
12:42So this shows fidelity.
12:44It's not something you're ever
12:46going to do if you're producing.
12:48Therapeutic vessels and
12:49you want to produce a lot.
12:51It's feasible to do this
12:52if you're studying viruses,
12:53but I think our group,
12:55Vanderbilt's group and a group and
12:57the Netherlands may be the only
12:59groups who've ever really done this
13:01and it takes like 48 hours nonstop.
13:04We set up COTS in the lab and all of that.
13:08So it's it's not a
13:12scalable approach.
13:17So when we did this we we had
13:19so many people look at our data
13:22and say how did you do that?
13:23I just, I just don't believe those results.
13:27And so we we took that challenge
13:29on and we said okay,
13:31we're going to prove it and we
13:34need to help each other be able
13:36to look at each other's data
13:38and know what we can believe,
13:39what we what we can trust in terms
13:41of the integrity of the data.
13:43So this led to this formation of
13:45a Tri society ISAF ISAC ISTH flow
13:48cytometry group where we worked on
13:51how can we improve the rigor in
13:53the field so that we can speak the
13:55same language for EV flow cytometry
13:57and improve our data approach.
14:01So this was the product of a
14:06surprisingly long time working with
14:08groups who do things differently.
14:11And we basically set out these
14:15guidelines to help basically tell you
14:20when you're designing your experiment,
14:22you're setting things up.
14:22What do you need to do to
14:24help people reproduce it?
14:25How do you prove that what
14:27you're looking at is Ev's?
14:28How do you validate it across
14:29instruments and settings?
14:30And how do you make your data shareable,
14:34transparent, and ideally interoperable?
14:40So this is where Josh's coding
14:42and technology development skills
14:44have really come into play.
14:45He tackled both the single EV
14:48analysis problem and the EV
14:50repertoire problem in flow cytometry.
14:52Single EV analysis Low is highly
14:55quantitative, but it has terrible
14:58sensitivity for single EV's.
15:00If you do it on a bead based
15:02way like a Multiplex way,
15:03it's high throughput.
15:04It's multi, multi, parametric but.
15:06It's only semi quantitative and
15:08you can't really assess the full
15:11range of complexity.
15:12So for single EV flow cytometry
15:16he developed FCM passive
15:17software that basically
15:21derives the collection angle of the
15:23actual optics of the actual machine
15:25at the time that you're doing.
15:27So if the engineer came in and
15:29fiddled with the alignment,
15:30it would you'd have to,
15:32it wouldn't be the same collection angle.
15:34But once you've derived the collection angle,
15:36if you collect the proper calibrators,
15:38then you can convert your data from those
15:41flow cytometry arbitrary units using ME
15:45theory to calibrated SI standard units
15:49of nanometers and for your fluorescence.
15:51You can also calibrate with molecular
15:53equivalence of soluble fluorescents and
15:56generate calibrated fluorescence as well.
16:02This led to. A bunch of papers,
16:04a bunch of our reports.
16:06This is still something that we're trying
16:09to get out and use more commonly so
16:11that we can more actively engage with
16:13and sort of share data with each other.
16:16For the EV repertoire analysis this involves.
16:21We prototyped a lot of this using
16:23the Miltonie Multiplex Exosome Kit,
16:25which is a bead set of almost 40 beads.
16:30Several micronic piece which capture
16:33based on one antibody type 1 epitope type
16:36captures the vesicles and then you go in
16:39and you detect with a different antibody.
16:42And so with this I'm going to walk you
16:45through some of the results that we see.
16:47But he's written the software to really
16:50facilitate the complexity of this and
16:52all of that data that has to get analyzed
16:54all at once in a calibrated way compared
16:56between experiments etcetera, so.
17:00This is a heat map showing several
17:03experiments we did with different
17:06antibody capture B combinations,
17:08different biofluid types and it's all
17:12calibrated again and with fluorescence
17:17and and the appropriate controls.
17:19So what you can see is CSF is unique,
17:23it's it's sort of standing off on its own,
17:25it's very different from plasm and serum.
17:28Plasma and Serum are relatively similar
17:30to each other and in the way that
17:32PCA and RT Sneeze are parsing them.
17:35So that's what we've done for that
17:37and now we have worked,
17:39we're working on stitching those together
17:42into a more comprehensive Atlas type
17:44approach where we can integrate single EV
17:47data with Multiplex EV repertoire data.
17:52There's also. Resistive pulse sensing.
17:54So if any of you are doing small
17:57particle work you may use a Nano
17:59site nanoparticle tracking analyzer.
18:00Resistive pulse sensing like
18:01a Spectradine or an eyes on.
18:03SO this is specifically resistive pulse
18:06sensing that works with the output and
18:08interface of the spectradine instruments.
18:10Those use little chips and what we
18:12found was if we took the same sample
18:15and reran it on a set of chips
18:18we get a different result every
18:20time just with the standard beads.
18:22And that's not good.
18:24So we developed a way to use this
18:27bike in and then reanalyze the
18:31data to normalize the data.
18:33So essentially it appropriately scales
18:39so that it is calibrated and it
18:41makes a difference in your data.
18:42So the plot on your left is of data
18:48that was not processed with RPS pass,
18:50and you can see there's a huge.
18:52Variation in those when we look at
18:56that with where the the spike has been
18:59used to appropriately scale the data,
19:01you can see that we can more
19:03clearly discriminate the the, the,
19:07the qualities of the size of and
19:10the concentration of those Ev's.
19:13So all three of these we're working with.
19:17Baylor and other collaborators
19:18to to work on integrating these
19:21into tools that people can access
19:23comprehensively and shared data.
19:30So this has been relatively quick.
19:33You know, this is something
19:34we've really been working hard
19:36on for the last 5-6 years,
19:39but that has made the difference
19:41as the field has gone from being
19:43able to go from Western lots.
19:44To flow cytometry where we don't
19:46know our limits of detection,
19:48now we know our limits of detection,
19:50we can articulate them and really
19:52reproducibly state what the results are.
19:57So then okay, we've done it.
19:59Can the whole field do it.
20:02In November we had a an EV
20:07reference material study.
20:08This was really spearheaded by
20:10Joshua Walsh who who recognized
20:12that as much as we try to.
20:15Teach everybody what they need to know.
20:19And we want the data to
20:21actually be consistent,
20:22to be calibrated, to be well,
20:25to 1st be capable,
20:26but then to be calibrated,
20:28to have the whole data set
20:31reproducible, etcetera.
20:36It's too complex when it's a really
20:38complex sample, so we use this.
20:41Fluorescent recombinant EV reference
20:44material from Anne Hendricks,
20:45which is now available from Sigma Millipore.
20:50It's called recombinant exosomes.
20:52From them, they rebranded it just because
20:55they thought that would sell better.
20:57Sigma did not, not.
21:00And and to really get at the heart of
21:03where we need to make inroads in the field,
21:06we went to the manufacturers.
21:08So Josh basically offered all of
21:11the instrument manufacturers the
21:12opportunity to take a sample.
21:14We shipped it off to them and send us back
21:18the results with a a set of sort of criteria.
21:21We want you to report back this,
21:23this, this and this.
21:25So everything was fair.
21:26It was transparent up front.
21:27This is what we wanted because
21:30when you buy instruments,
21:31you need the manufacturer to be able
21:33to tell you how to properly use it.
21:36In this study,
21:37this is only the beginnings,
21:40but this shows I won't go into
21:41all of the results.
21:43But basically in that wheel of all
21:45the criteria we'd like to have met,
21:49some are very good and others
21:51are in the process of learning.
21:54And so if we redid this today,
21:57some of those on the bottom row.
21:59Would be either nearly completely
22:02filled in or filled in.
22:04So this was really a good
22:05opportunity to work with industry
22:07to start trying to pioneer this.
22:09So OK, so if we can do it,
22:13how do we share this with the field?
22:15So this needs to be centralized,
22:17accessible.
22:18So we've been working with Baylor as
22:21part of the ERCC common fund effort
22:25to develop the Nanoflow repository.
22:29And so that's the beginnings of a
22:34shared way to deposit the data.
22:38This Baylor Group is also as
22:42in parallel the XRNA Alice.
22:44So you can see what I was talking
22:46about before where we want to
22:47tie the surface phenotyping data,
22:48the individual EV phenotyping data,
22:50the repertoire phenotyping data
22:52then along with the RNA cargo.
22:55Data.
22:57They do have the the skeleton and
23:01the background and the infrastructure
23:03of the xrna Atlas there.
23:08So this is moving us closer to
23:11doing what we want to do,
23:12which is to be able to look at subsets,
23:14identify markers to pull out
23:16subsets to look at the RNA.
23:18Then we hit a roadblock which
23:21was one I expected, but.
23:24You know,
23:26most RNA seq methods require
23:29nanogram levels of RNA.
23:32When you get to the subsets,
23:34you're probably in less than 100
23:36picogram kind of range of of RNA.
23:39So we tried to we decided to
23:40test whether or not we could use
23:43single cell sequencing methods,
23:44not in the single cell mode but in.
23:48Bulk using that as the library preparation
23:51method for looking at EVRN A's.
23:57And remember I showed you the picture
23:59of the DC 2.4 E V's on the flow
24:01cytometer and I said remember these?
24:03That's the cell line we chose.
24:06It grows like weeds.
24:07It's a little mouse dendritic cell
24:09line that Ken Rock made back in the
24:111990s and it feeds itself GMCSF.
24:13So these are the happiest
24:14cells you could ever.
24:16Want they grow like weeds,
24:20and they've also had a lot of
24:22manipulation in their background.
24:23So I outlined here all of the background
24:27that I kind of ignored until one of
24:30our reviewers pushed us to instead of
24:34TEM get cryoem to really hammer out
24:38the exact size of these and what we
24:41couldn't see in the TEM on the left.
24:44You can see really clearly on the
24:46right we have retroviral capsules or
24:48something that looks awfully a lot like them.
24:50And I got a call from the lab
24:53who was helping us with this.
24:55Not a call,
24:56it was worse than that,
24:58an e-mail that was carbon copied to
25:01the then scientific director of all of
25:04NIH saying what do you not understand
25:07about B SL1 samples for a B SL1 lab.
25:13This cell line is sold by Sigma and
25:17mercury pour has a B SL1 cell line.
25:21And I say look, I'm really sorry,
25:24I don't know what that is.
25:25It could be a mishmash of
25:27rearrangements of any of those things.
25:29In this background it's also a
25:31mouse cell line and they have
25:33lots of endogenous rector viruses,
25:35so I have no idea what that is, but.
25:38I'll I'll get to the bottom of it.
25:41And this is now.
25:42I've lost count of how many years later
25:46we decided we wanted to apply this
25:47RN A/C approach to those because I
25:49wanted to figure out what's what is it.
25:51I don't want to just do a PCR for this,
25:53that and the other thing,
25:54I want to know what's in it.
25:57So we've done Proteomics and we've
25:59done RNAC and it turns out we find
26:01a dominant species and it turns
26:03out it's Mouse Maloney virus which
26:05was part of its background.
26:10So that is a xenotropic virus,
26:14meaning it doesn't go from mouse cells to us,
26:16it just stays between mouse cells and it
26:18doesn't go from mouse cell to mouse cell
26:20unless the cells are actively dividing,
26:22which, well, those do. And so
26:29another reason I wanted
26:31to go down this crazy Rd.
26:33is because of the hers where
26:35we know that those modulate.
26:37Responses to immunotherapy or
26:39their indications that they do.
26:41And so we wanted to have a pipeline,
26:43a method that would allow us to elucidate
26:46the presence or absence or the types
26:49of herbs in our human EV samples.
26:52So we're collaborating with
26:54Kendall Jensen at Tijan and
26:56Yasmine Belkade's group at NIAID.
26:58She's just accepted the position
27:00to run the Pasteur Institute.
27:02So unfortunately we're going
27:03to lose her soon.
27:04But we're working very hard to get
27:06this all tied together before she
27:09goes to have a comprehensive pipeline
27:12that would include conventional RN,
27:14A's and the Hearse.
27:15So I want to show you some results
27:18of all these tools that we've been.
27:20Working on and here a couple of the examples.
27:23I'll show you a little bit of kidney cancer,
27:27prostate cancer, colon cancer, CNS diseases.
27:32But first,
27:34if you could live a day in my shoes,
27:37you get a question just about every day.
27:42I want to start a study and
27:43I want to look at exosomes.
27:45That's what people say to me and I want
27:48to know what kind of blood tube I need.
27:50And that's a really hard what
27:52do you want to do with it?
27:53What do you, what do you want to look at.
27:56So to help us figure out what is
27:59our right blood collection tube,
28:01we decided to compare for the SST tubes,
28:05EDTA tubes and the strect
28:08DNA&amp;RNA complete tubes.
28:09This is the comparisons that we
28:11did to suss out the impacts of
28:14platelets and not platelets and.
28:17Ways that you do the spins,
28:18we counted the particles that were remaining
28:21after we did the depletions etcetera.
28:23And what you see is that we had
28:28a surprise which is that CD62 P,
28:33CD242A,
28:34some platelet markers were not
28:36only elevated in samples where
28:38you froze the sample and then
28:40you spin out the platelets,
28:41which is a terrible idea,
28:42but a lot of people do it.
28:45It was also elevated in the struck DNA tubes.
28:48So maybe something with the
28:49fixation of the struck DNA tube
28:51that's causing shedding of these
28:55these vesicles.
28:55And so we've we've looked at this further.
28:58But you know it's this kind of
29:00quantitative analysis that helps us
29:02assess the not only the integrity
29:04but also the repertoire and the
29:06relative abundance of these different
29:08types of vesicles in the solution.
29:11So for us, for our lab,
29:12for our protocols,
29:13we're doing SST tubes and complete RNA tubes.
29:18I I actually think that plasma
29:20DTA tubes are also great.
29:24I spoke to somebody earlier
29:26today about oncosomes.
29:27These are large vesicles shed
29:29by tumor cells which are like
29:31larger than 800 nanometers,
29:33sometimes larger than a Micron.
29:35So every platelet depleting protocol
29:37that you do to spin out the platelets
29:39is going to remove the oncosomes.
29:43I I don't have a good solution.
29:46If you want to study those,
29:48I think you have to go directly to
29:51processing the onpisomes separately.
29:53So our approach is showing us good
29:58fidelity and differences in tumor types.
30:01So Long story short we compared a bunch
30:04of different EV's from different tumors.
30:06This is something that
30:08we've already published and.
30:10You can see Epcam is more commonly spread
30:12or sort of more highly expressed in these
30:15samples from the epithelial tumors and
30:17from the Seglio bus, I mean, it's good.
30:20You wouldn't expect Epicam so much in those.
30:23The tetraspanins and CD44 are
30:26up in the in in both.
30:30So for kidney cancers,
30:33there's not a great molecular handle.
30:37For pulling out kidney cancers.
30:39So Marsha Lenahan and Maria Marino
30:43have this amazing set of cohorted
30:46patients and data and studies and
30:48information they've learned about
30:50one hippo window and and other
30:53forms of hereditary kidney cancers.
30:57So we worked with them to look at some
31:00of the different tumor types that we
31:03could prototype with and then begin to
31:05look at those samples and patients.
31:07And so we tried a battery of different
31:11markers and we found some that
31:15really hadn't been expected and they
31:17have some of the same features.
31:19Some of them are also Tetra spannons
31:21and they're also Stemmus markers.
31:23So this is consistent with what we saw
31:25was really elevated in the other two types.
31:27So maybe we're finding that there's kind
31:29of a a malignant signature as opposed
31:33to a specific type of tumor signature.
31:37In the types of markers they express.
31:41Then we worked with collaborators to
31:43look at the EV profiles in malignant CSF
31:47samples and other CSF samples including
31:52autoimmune diseases and viral diseases.
31:55And you can see again here
31:57we see that same pattern.
32:01And so
32:04once we have the markers,
32:05we do the pull down.
32:06Do we actually see differences in the RNA?
32:09So this was one of my early proof of
32:13principles examples where we just took
32:17a thoracentesis sample from
32:19a patient of mine who had
32:23very metastatic prostate cancer,
32:25which is PSMA positive the therapeutic
32:29tap and in the therapeutic tap in
32:31the biospecimen protocol we're able
32:33to pull down the PSMA positive EV's.
32:36Compared to the PSMA negative
32:39EV's compared to the bulk sample.
32:41And you can see there are several
32:44RNA's which are highly associated
32:46with the PSMA positive ones,
32:48which you would have missed if
32:50you were looking at the soup of
32:52everything because there's so many
32:54other kinds of vesicles that compete
32:56in that type of identification.
32:58So I thought PSMA was going
33:00to be a great marker,
33:01but really what marker should we be using?
33:07I insinuated and I really feel like
33:09the markers that we choose are not
33:11going to be the same markers that we
33:14use in the context of intact tissue.
33:16It may relate more to their phenotype.
33:18So we did a large screen of 170
33:22different EV surface markers across
33:25some of those kidney cancer patients.
33:28Other CSF sample was just a massive cohort.
33:31So if you're squinting at this
33:32from the back of the room,
33:33you can see there's sort of a
33:35tartan Plaid kind of pattern.
33:38There's a a sample down here
33:40where it's all blown out.
33:42It turned out,
33:43turned out that person had a radioisotopic
33:46treatment for metastatic prostate
33:48cancer couple weeks before and was
33:52having ramped up marrow production.
33:53I don't have any other samples like that,
33:56but clearly this is. Not,
33:58we're not going to understand much from that.
34:01But then there are sections where
34:03you see more of some workers,
34:05less of other markers and in sets.
34:07And if you break down those sets and
34:10you look and you say CSF versus serum,
34:12they're really different.
34:15Looking at PC A's,
34:16if you look at tumors versus immune,
34:19this tumor, that tumor,
34:21they're also very separable.
34:23Well,
34:23some of them are separable more than others.
34:27And then looking at the CSF samples from
34:31patients with or without brain tumors,
34:33you can also see that there
34:36are differences that we see.
34:41So we also worked with Steve Jacobson
34:45in NININDS and he studies both Ms.
34:50and he MTSP, the HTLV associated
34:54tropical Myelotis myelo.
35:00******* peripheresis.
35:03So these are CSF samples from those
35:07patients and patients also with who carry
35:10the HTLV virus but are asymptomatic.
35:12That's what the AC's are or other
35:15viral diseases and you can see
35:18that turns out that hand patients,
35:21the ones with active disease associated
35:24with HTLV in the nervous system.
35:26Have higher CDA than CD2E V counts.
35:31We've followed that up with
35:34another set of samples,
35:36again that size and as as well
35:38as other markers and we still
35:40see that robust difference.
35:41It's it's it's very,
35:43it's not a massive magnitude,
35:45but it's very consistent.
35:47So which of these EV markers
35:49relate to the biological state,
35:51meaning the biological state
35:52of the cell that made them?
35:55And this is work that was done with a
35:59colleague who had a really interesting
36:02biological phenotype they were studying.
36:05They made some knockout cell lines.
36:07And what you see here is our
36:09B plus antibody control,
36:10the knockout line, the control line,
36:12the knockout line, the control line.
36:14And what you can see is that
36:17there are some genes that are just
36:20missing from the knockouts and there
36:23are some genes that are missing.
36:25From the controls.
36:26So we're really getting a sense of
36:29changes in these related to that.
36:32So we're really just starting
36:35to apply these and learn more.
36:37We have also found a pattern
36:40in metastatic potential,
36:42so match sets of cell lines that have
36:44different metastatic potential on
36:46their markers and so now we want to
36:51move forward further with that so.
36:54You know,
36:54I started talking about the
36:56commotion in the blood,
36:58or as one of the earlier professor said,
37:02the the mess that is the
37:04extracellular space these days.
37:08I think the reason why we've wrangled and
37:11learned so much from the immune system is
37:14being able to be so systematic about it.
37:17And so I've tried to begin to wrangle the
37:20extracellular space into the same way,
37:23to establish some foundations
37:24to make a consistent Atlas,
37:26to then begin to study the specific
37:29markers related to tumors,
37:31relating them to phenotypes and
37:32all of these other things.
37:34So the survey that I sent you guys
37:37and thank you for those who who went
37:41slog through it to to to humor me.
37:44The the bottom line is,
37:45do we need an extracellular ontology?
37:48We have a cellular ontology.
37:50But when you take a liquid biopsy,
37:53you have no idea.
37:54There's nothing that's a single
37:56marker that can tell you that a
37:58vesicle came from an exosomal pathway.
38:01In fact,
38:02the biologists are really kind
38:03of working out all the specifics
38:05of the exosomal pathway anyway.
38:07So then you try to frame the
38:09ontology of the extracellular
38:11space in the related ontologies.
38:13So I just mentioned,
38:15do we need an extra cellular one?
38:17There's a cellular one,
38:21I don't know, you guys can tell me,
38:22but I have asked my liquid biopsy colleagues,
38:25is it probably pretty true that
38:28you categorize the things that
38:31you're using for biomarkers,
38:34classify them really by what it is
38:37that you isolated or how you isolated?
38:39I say yeah, okay so.
38:43And the nano material
38:45field is super detailed.
38:47They have a Nano Nano Particle Ontology,
38:51the NPO, that's all about formulation,
38:54this is the shell, this is the surface,
38:56this is the, it's extensive.
38:59So how do we just approach
39:01the mess that's in between?
39:03So hence the survey and I didn't ask
39:08the question I wanted to ask because
39:10it was so strongly objected to.
39:13My first question was going to be
39:15what do you think an exozone is?
39:17A BCD? But people decided
39:20that was too controversial,
39:22so instead we asked more obliquely,
39:28maybe obtusely.
39:31This is a selection of ways to
39:33classify extracellular vesicles.
39:35Which one do you think is most central
39:38to harmonizing with later system?
39:414 vesicles,
39:42the largest proportion that the
39:45highest answer is based on biological
39:49considerations like Biogenesis.
39:51And so I think that message of the
39:54EV community of what distinguishes
39:57A vesicle from a non vesicle and
40:01an exosome and microparticles
40:03or other things it's getting
40:06through in response to the non
40:09vesicular extracellular particles.
40:12Even the EV people, the ISAF people,
40:16say we don't know what the Biogenesis is.
40:19For the most part, the top answer is based
40:21on biochemical considerations, composition.
40:24Is it a lipid biolayer? What's in it?
40:29Informally? And I don't know if this
40:31is ever going to get published or not,
40:32but we did a we did a beta test.
40:35I used my friends and colleagues
40:37at NIH as Guinea pigs.
40:39We have a couple of listservs for
40:42the liquid biopsy group and the EV
40:44interest group and we sent it to them
40:46and it was even more extreme when we
40:50focused on the liquid biopsy groups.
40:53It's about composition, what is it,
40:54what we're looking at and the
40:57EV folks about everything,
40:59not just vesicles and when asked about
41:03everything without dividing into vesicles
41:05or non vehicular extracellular particles.
41:08The EV group still focused on Biogenesis,
41:12so I'm working on the analysis of who
41:15answered what and it should be interesting.
41:18But I've been at meetings where
41:22people stand up and they ask me
41:23why do you care what it's called,
41:26if it's a good biomarker?
41:27And honestly,
41:28if the biomarkers is a good biomarker,
41:30that's great.
41:31It's just if you want to stitch
41:33the data together and understand
41:36how our data relates to each other.
41:39Everybody who does omics and assays
41:42and atlases knows that there has to
41:44be a common framework it's set on.
41:47So I just want to.
41:49In addition,
41:49I really have to thank you all
41:52for inviting me to come speak.
41:53It's really an honor for me as a
41:55young scientist to speak to you guys
41:58learn from you, get your feedback.
42:01I also need to thank the laboratory
42:04pathology kind of that be my mentors.
42:08Past, present and current.
42:10As you know I was thinking last night
42:14I couldn't say this takes a village.
42:16This actually takes like lots of villages.
42:19So these are some of the villages
42:24who have and they're continuing to
42:26help me and I'll take questions.
42:28But as a sneak peek I had bought
42:32on behalf of our residency program
42:34director some slides about the.
42:37Residency program at at NIH If there
42:41are folks who are interested in it
42:43at lunch and I'll just e-mail it to
42:47anybody who's interested, thank you.
42:59Should I open the chat and see if
43:01there are questions in the chat? Okay
43:11act stating for CME credit.
43:14Texting for CME credit, so I don't
43:17think those are questions. Yeah,
43:21refer to analyze the EV in
43:25the context of that area.
43:29Yeah, there's a whole group of ISA which
43:33is interested in not only the Ev's,
43:38the host Ev's, but also the Ev's.
43:41Of you know, across the microbiome
43:45or infections, that's become a very
43:48interesting part of COVID work.
43:50Kendall's done some work on that at Tgen.
43:53Several people have have done
43:55a lot of work on that. Yeah.
43:58In that context how do you
44:01differentiate the post PR?
44:05Yeah, so it depends on your assay, right.
44:08So if you. Have species specific
44:12antibody clones that can begin to
44:14differentiate some of it and that's
44:16been applied in some model systems.
44:20I don't know if it's been
44:23applied in clinical settings.
44:27And then in terms of the informatics
44:31for you know like RNA analysis it
44:32would it would be based on the genomes.
44:37There's certainly overlap where you can't
44:40discriminate I would imagine my final
44:43question on the basis of the buy markers,
44:49the efforts pull down subset,
44:52yeah that's what we're doing and
44:56that's why we had we've had such an
44:58extensive focus on which markers to use.
45:01And then once we do the pull downs,
45:03how do you make that work robustly for
45:05the very small amount that you pull down?
45:08So one thing that struck me and
45:11I think anybody who's interested
45:14in doing liquid biopsies of EV's
45:17should probably understand this in
45:19a milliliter of blood, you know,
45:21you might have 3 circulating tumor cells,
45:2410 circulating tumor cells.
45:26There's something on the order
45:28of about a billion EV's.
45:30And there's something on the order of
45:3510 to the 16th versus 10 to the 18th,
45:38like a billion billion lipoprotein particles.
45:42So, So those since they're so close and
45:45overlapping in size with the vesicles,
45:47those become the main complicator.
45:52And it's what I like about the affinity
45:55pull down part is that you can.
46:00Directly interrogate A membrane receptor,
46:02another membrane receptor,
46:03and know that you're dealing with
46:05something that is likely something that
46:07has a little bit by later because it
46:09has a Tetra span and then thanks, yeah,
46:17the best questions.
46:47Yeah. So, So, yes, yes and yes.
46:50So the the question for folks
46:52online who maybe didn't hear it was
46:56are there. Impacts of the cellular,
46:59the state of the cell in terms of its
47:02metabolism or other stressors that
47:04affect the type of vesicles produced
47:07And are there impacts also on the ways
47:10that cells receive vesicles. So you know
47:172004 Arnie Levine showed that P53 was
47:21central regulator of producing exosome.
47:23So there's. From way back there's there's
47:26been an understanding that genotoxic stress
47:29hence my interest as a radiation oncologist.
47:31Radiation kicks off a surge of these and
47:38no so you can give a sublethal dose and
47:43it it stimulates the exozone pathway.
47:47So so this is there's clearly a.
47:51Very wide heterogeneous range
47:53of types of vesicles.
47:56There are these exosomes,
47:57they're small ones made in the vesicles or
47:58other types that are shut off the surface.
48:00There are the,
48:01I guess you could call them apoptosomes,
48:03the ones that are shut in
48:05the context of apoptosis.
48:08I think we are only scratching the
48:11surface of those different types the
48:15in the 80s or 90s they originally described.
48:19These little vesicles and microscopy
48:21is platelet dust where they just
48:23kind of kick out the garbage.
48:25So there was first an idea that these
48:27are garbage bags and there was this idea
48:30that they're sophisticated endocrine
48:32systems of communicating between cells.
48:34I think it's both and and a
48:36lot of stuff in between.
48:38So for me, I'm going to be
48:41looking for different types of
48:43vesicles with different types of.
48:46Aberrant DNA damage,
48:50Addux and other things.
48:53So yes genotoxic stress increases
48:58exosome production per se.
49:01Also probably stress and loving
49:04There also is starving cells to this
49:10is really kind of related to some
49:12work that Raghu Glory has talked
49:14a lot about which is that the.
49:16Pancreatic cells,
49:17which are essentially just
49:20ravenous for resources,
49:22take up these therapeutic
49:24vesicles that he produces.
49:26And he thinks that that's because
49:28of their metabolic state and
49:30receptor affinity for vesicles
49:31compared to surrounding tissue,
49:33which doesn't seem to pick up
49:35those therapeutic vesicles as well.
49:37But any type of vesicle that you look at,
49:42you can find.
49:45Yin and Yang and a lot of these things.
49:47So there are the E V's or exosomes that
49:53cause essentially vaccinating effects,
49:55tumor stimulation.
49:57There are other vesicles which are
49:59clearly inhibitory that promote a more
50:03mildly suppressor type phenotype,
50:05which are which which do what?
50:07Until we systematically start
50:09breaking the groups apart,
50:11there are a lot of mysteries
50:12that are hard to unravel.
50:23Yeah.
50:25So that's a really good question.
50:28As with all of it, part of
50:29the answer is it depends.
50:34So if you make synthetic
50:37ones and you inject them,
50:39it depends on how you made them.
50:42They could just.
50:43Go first pass and get largely
50:46taken up in the spleen or the
50:48liver and they may make only
50:49kind of one round through,
50:50so it might really matter
50:52which way you inject them.
50:56In other cases where you've made
50:58them under other conditions,
51:00they circle around quite
51:01a bit longer beforehand.
51:03I I think the common understanding
51:07is that probably the turnover
51:09overall is something like 6 hours.
51:12But it's relatively rapid.
51:16But for me as a radiation oncologist,
51:19I don't feel like I need to run in
51:22and get a sample in the first hour.
51:25There are a lot of things that happen
51:272448 hours later that take that long
51:30to start to manifest and be able to be
51:34discernible even if you did seamless
51:36offstaining in the affected tissue.
51:43It seems to be quite rapid,
52:01so I think there's a myth that
52:04every vesicle a cell relieves,
52:07shoots out and heads straight
52:09for the bloodstream and.
52:11Circulates and then whatever
52:15the kidneys, clear some.
52:16There lots of urine studies which look at
52:20vascular populations deliver clear some.
52:22I suspect that the clearance is dependent
52:25on the surface markers like selectins
52:28and organ specific distributions,
52:32but
52:34I wish I knew who first said this,
52:36but I I've heard it said that.
52:41The blood is sort of our ocean within.
52:45So in the context of organisms evolving
52:50through mammals and vertebrates to have a
52:54circulating system that those circulating
52:58systems strikingly reflect the oceans and
53:02those salinity conditions. Other things
53:07there's a researcher at MIT.
53:11Who? Sally Chisholm,
53:14who discovered when she was a
53:17postdoc Prochlorococcus bacteria,
53:19which are responsible for some
53:22ridiculous amount of the world's
53:24CO2 metabolism in the oceans.
53:26Like when you fly over in some areas are kind
53:28of green and some areas are kind of blue.
53:30It's different. Prochloroccus.
53:32They shed vesicles and there's a
53:37thought that's part of how they.
53:40Communicate and cross regulate.
53:42But I think in our compact systems,
53:46there's probably a great deal of
53:48vesicle release that impacts the
53:51local tumor microenvironment and
53:53is not necessarily part of what
53:56processes out and which which
53:59stay and which get processed out.
54:01We talked a little bit about it at dinner.
54:04We just we have to find better ways of
54:06studying the extracellular spaces I think.
54:24So I was hoping I could find
54:26unique sorts of things,
54:27but that's not what I'm finding.
54:29I'm finding patterns among classes of cells
54:33as opposed to unique this versus that.
54:37And I I think you know, if you think
54:39about anatomic pathology and how you
54:41take a chunk of tissue and you look at it,
54:43so PSMA, that's pretty indicative of a
54:47prostate cancer cell in a certain state.
54:50If you took a chunk of prostate tissue out,
54:52if you took a piece of my perotid,
54:55you'd also see high levels of PSMA.
54:58So PSMA is not really a good
55:03prostate cancer necessarily marker.
55:07So I I actually think the the
55:13best classifying markers will will
55:15probably not be exactly the same
55:17as those which have been defined
55:20so far in intact tissue contexts.
55:24All
55:29right, I put everybody to sleep.
55:32Thank you, everybody.