May 31, 2023 • 44 mins
Welcome to Olink Proteomics in Proximity Podcast!
Below are some useful resources from this episode:
Highlighted pre-print article: Malarstig A, Grassman F, Dahl L, et al. Evaluation of Circulating Plasma Proteins in Breast Cancer: A Mendelian Randomization Analysis. ResearchSquare 2023.04.04. DOI: https://doi.org/10.21203/rs.3.rs-2749047/v1
Highlighted platform that was used to measure proteins in this study with a next-generation sequencing (NGS) readout (Olink® Explore 3072): https://olink.com/products-services/explore/
Learn more about the consortiums or cohorts mentioned in the podcast:
UK Biobank Pharma Proteomics Project (UKB-PPP) is currently performing one of the world’s largest scientific studies of blood protein biomarkers conducted to date: https://www.ukbiobank.ac.uk/learn-more-about-uk-biobank/news/uk-biobank-launches-one-of-the-largest-scientific-studies
Human Proteome Atlas (HPA) aims to map all human proteins using various omics technologies: https://www.proteinatlas.org/
SCALLOP consortium is a collaborative framework for discovery and follow-up of genetic associations with proteins on the Olink Proteomics platform: https://olink.com/our-community/scallop/
KARMA cohort is a prospective screening cohort for breast cancer in Sweden: https://karmastudy.org/ongoing-research/the-karma-cohort/
Swedish Twin Registry (referred as “Twin Gene cohort” in the podcast) is the largest of its kind, containing genetic information about ~87,000 twin pairs: https://ki.se/en/research/the-swedish-twin-registry
Additional published articles and books mentioned during the podcast:
Hood, Leroy and Price, Nathan. The Age of Scientific Wellness: Why the Future of Medicine Is Personalized, Predictive, Data-Rich, and in Your Hands, Cambridge, MA and London, England: Harvard University Press, 2023. https://doi.org/10.4159/9780674293465
Suhre K, McCarthy MI, Schwenk JM. Genetics meets proteomics: perspectives for large population-based studies. Nat Rev Genet. 2021 Jan;22(1):19-37. doi: 10.1038/s41576-020-0268-2. Epub 2020 Aug 28. PMID: 32860016. https://pubmed.ncbi.nlm.nih.gov/32860016/
Would you like to subscribe to the podcast on your favorite player or app? You can do so here:
Apple Podcasts: https://apple.co/3T0YbSm
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
(00:04):
Welcome to the
Proteomics in Proximity podcast, where
your co-hosts Dale Yazuki, Cindy
Lawley, and Sarantis Chlamydus from Olink
Proteomics talk about the intersection of
Proteomics with genomics for drug target
discovery, the application of proteomics to
reveal disease biomarkers, and current
trends in using proteomics to
(00:25):
unlock biological mechanisms. Here we have
your hosts, Dale, Cindy, and
Sarantis. Hello, everybody. I'm
Sarantis. I'm together today with Dale
and Cindy for another episode
of our
great podcast, Proteomics in
Proximity. We are all very happy
to have like a guest, Professor
Johann Schwenk
(00:45):
who holds a
position at the University of
KTH University [Royal Institute of Technology in Sweden].
And today he's
a protein expert and
a professor in
Translational Proteomics. And today we'll
discuss a little bit about his
new research, about his
research interest and how proteins can
enable multiomics approaches. Actually,
(01:07):
Jochen,
thank you very much for joining
today. And I would like to
start the discussion asking you (01:13):
What
does translational proteomics mean to you?
Yeah,
I think when we started thinking
about the title for a
professorship, translation was really a hot
topic at the time (01:28):
to bring
something you've been doing in the lab
into a clinical context. But I think
it turns out to be much more than this. This is
actually to explain also what we're
doing in the lab to others, so that
the community can engage into
our research and we can even find a
broader utility. So it's
still the idea of
(01:48):
connecting the lab environment
with clinical and population
health. So I think
hopefully one day we'll contribute to that
activity. That's great.
And I saw also that you study biochemistry
in Tübingen, University of Tübingen. That's quite
famous from the biochemistry industry worldwide. Do
you have any story that you would like to
record for your first paper? For example,
(02:09):
being in Tübingen, and that would be great
to hear.
Yeah,
Tübingen of course gave me a
fantastic time. We were very
small number of students per semester. I had
very close connection to the
professors. I got a chance to go to
a Lindau Nobel Laureate
meeting there. And it was really,
(02:30):
I think, an inspirational time that
sort of created
a lot of curiosity about science. And
then after that I moved a little bit more
into technology. So in the early two
2000s, when I did my Masters and PhD, I
worked with Luminex-based
assays, which at that time
was really new.
Then sort of that took me then
(02:50):
to join the Protein Atlas in
2005 as a postdoc. And somehow I
got stuck with this fantastic project.
I'm
still around and
learn every day something new about
proteins. I mean, just being
there and to work with Mathias Uhlén
and all the colleagues has been truly
inspirational. So you've been at
SciLife Labs since 2005, and that was when
(03:13):
it began at KTH, is that correct?
Yes. SciLife lab was
inaugurated in 2010. So actually,
it was my and three other groups
that moved into the building under
construction, I think it was in
October 2010.
So I consider myself very much of an
oldie when I think about my time at
SciLife Lab. I've seen it
(03:33):
change, grow, and now, I think,
become a very prominent research institute
in Europe. So it's, I
think, fantastic and very
much also gave me opportunities to learn
about other technologies and to
learn how information about
proteins can be useful. And I
have many stories to tell, but
one of which is, for instance, I have a
(03:54):
little bit of a side activity project around
GPCRs, and that, for instance, I think
wouldn't have been possible if I would just
be sitting somewhere in a lab and
not be exposed to all these different
activities. Sure. And for those not
familiar, GPCRs
or G-
coupled protein receptors.
GCPRs, is that correct?
(04:15):
G-protein coupled
receptors. G-protein coupled receptors. I
got to get my acronym straight.
It's a really important drug
target, right? Membrane
proteins. Yeah, membrane proteins that are
important drug targets. Correct?
Exactly. Yeah.
And they're at the SciLife
Laboratories. Well, you said that you were
involved in the Human Protein Atlas way back
(04:37):
in 2005, so therefore
the genome had just been
finished in 2002.
2003.
Those must have been pretty exciting
times because there was a big pivot
and interest and focus on the
proteome, is that correct?
Yeah, of course. And then at
that time, Mathias
and us were producing all these
(04:58):
antibodies was fairly unique, and people
were not really sure whether that would add
any value to the field
that's dominated by mass spectrometry. But I
think now we've shown, and
the way that we've brought in new
data, trying to understand the data that we
generate, and then sort of give feedback
to other data types with
localization of
(05:18):
subcellular compartments
of proteins that
I think are really
super valuable and help us to disentangle
the complex biology that we
live in. And
my real interest is
proteins in the circulation. So that's even
more complicated because
you're under sort of the
(05:39):
constant exchange of molecules
in all parts of the body.
So it's not as organized as looking at
subcellular localizations, but it's still
fascinating. And I guess
that's something I really
sort of fell in love with, and I
really enjoy doing.
Great.
(05:59):
How would you see Olink because
you are a biochemist? I
mean, you are a mass spec
expert. How will you see
Olink fitting on this pipeline of
mass spectrometry? How do you see mass spec and Olink
working together from your experience?
Because you have a big experience in this
field.
I'm very fortunate to get to know Ulf Landegren
many, many years back, and we've been
(06:20):
sort of seeing each other on a regular
basis, partly because Uppsala [Olink] and Sweden [SciLife Labs] are
very close and sort of been
doing things in parallel. And of course,
fantastic to see the journey that with the
proximity extension, proximity ligation,
and all these different
versions of this concept have
now, I think,
been the driver for using antibodies as
(06:42):
molecular tools. I mean, there was
just a paper in Nature Methods, I think,
just the other day. Again antibodies
conjugated with oligonucleotides. I think
that's giving people much more
bigger field to play and use these
reagents. So
of course I've seen how it
started and we've been
very late to the game. My lab,
(07:03):
or the unit that I'm heading
at SciLife Lab, started to
introduce Olink in 2017.
And since then, we've been super
happy to have the system in-house
and do this
for others, users that
come to SciLife Lab to just want to
have data, but also for our
research. So I think, again, any
(07:24):
data type adds a value to what we do. And I
think Olink has truly
enabled us to do many things we weren't
able to before. So,
fantastic. I'm dying
to ask
in front of you,
based on what you see, the
opportunity in front of you
with this technology, whatever
(07:45):
technologies, right. What is it that you're
most excited about for the
future? Are there aspects
of
work that you've been doing or a
direction that you're going in, that
you would like to share, that you're
comfortable sharing:
I just want to know
what's the part that makes
you go into
(08:07):
flow? What do
you want to do next?
Yeah, I started
doing a lot of assay development myself
when I worked with Luminix 20 years
back. That sounds a little bit silly when I
say this, but it's the truth.
Of course, that sort of
has always been something to try, something
(08:28):
new to maybe try, something that's
difficult, maybe not immediately
rewarding, but in the
long term, something that could
be very fruitful or something that
makes you proud as the researchers that you
say [to yourself] (08:40):
"Okay, this is something I believed in,
and I see it's happening." So the
next sort of moment for me,
when I had this type of thinking was
when COVID started and when
lots of people went into serology
testing or protein testing in
the classical way, when me
and my colleagues at KTH, we said,
"Let's try something different and use dried
(09:00):
blood spots. Let's not ask
people to come to the clinic, let's send
the devices back to them,
to their home, so they can collect
bloods in their kitchen
and sofa, wherever, and send
them the samples back to us, to the lab,
where we can do the research."
So that, I think, really
inspired a lot of new ways of
(09:21):
doing this. When you think about cutting
costs, simplifying workflows, freeing
the time of people in the clinic,
but also to think about
doing health monitoring. I
think people
always often ask me, what do I think is
proteomics best used
for? And I think it's monitoring.
It's like looking at who you are and looking
(09:42):
how you change. And I think this
combination, I think, now is sort
of shaping towards something I really
am passionate about. And dried blood spots
is a fantastic tool.
It's more challenging than doing it
the classical way. There's so much more to
learn, and then maybe even go
further. When
Sarantis and I talked a couple of
(10:04):
weeks back, also to look at even smaller
sample volumes, looking at other body
fluids, such as interstitial
fluid, that could even tell us something
extra that blood is not able to tell us.
We'd need a
baseline to understand the reference of
dried blood spots to be making that
comparison, right? Yeah. And
(10:24):
so enabling in areas where we
can't get a blood draw,
a phlebotomist out to do a
blood draw.
I think that's going to be really
important. Dale?
So if you can give us some background
on dried blood spots, and I'd appreciate it,
because my only familiarity with it was when
I had my first child, and they did a
(10:45):
heel prick, and then they went ahead
and used the blood from
that little lancet onto a
particular card. Is there something
special about the material they
use for a dried blood spot?
And what are the challenges as far
as working with proteins in that context?
I think
let's say if you think
(11:06):
from an analytical
perspective and a precision
perspective, the dry blood spots your
kids
donate, they are just put in a
filter paper to do a plus/minus test.
So it's really a binary answer you're
after. But if you really want to look
at subtle changes in the human
phenotype, I think then you need
to ensure that the precision of the
(11:28):
material you use in your system is
there.
Normally when you have a dry blood spot, you
get sort of a donut distribution of the red
blood cells, so it really matters where you
do the punch. So you want to avoid these
type of things, especially if you want to do
it at scale, if you want to do it
consecutively. So I
started to work with a local company that
(11:49):
was founded by one of my colleagues at
KTH. And they use a
microfuidic system to exactly collect
ten microliters. So just knowing
that what you put into your system is ten
microliters, of course,
then there are different levels of
hematocrits, there are different other
things that you need to consider, but you
at least eliminate some of the concerns that
(12:10):
you have. That,
I think, is really the key. And of course,
it's a simplicity of this
procedure that you can assure it's
easy for people to do. And
they manage, even though they may not
be trained. I failed also
when I did it the first times.
But if you
get used to it, the quality is
(12:31):
really excellent. And I think there are also
studies showing that more and more are using
other devices
that are out there, I
think now you have a new material
which is sort of similar to plasma,
but it has some bonus. And the question is,
how do you manage that bonus? Is
it something that is a challenge, it's a
burden that makes it difficult for
(12:52):
you
to be analytically
precise? Or does it open up
opportunities that were not possible
when you looked at the regular blood,
plasma samples? Because of,
let's say, the
hematopoietic cells that are still there,
they may leak out something that could be
really exciting. So it's
this balance between things that, I think ...
(13:13):
When you have a discussion a little bit
with other proteomics expert about
dry blood spots, there's always a question
about how you control and
normalize. Because, I don't know,
from what I have heard,
actually, it's not easy to have always
the same type of dry blood spots. I'm
guessing that there's a lot of varieties,
a lot of variation can be there. Would you
(13:34):
have any idea how one can normalize this
data in order to have like, longitudinal
studies or studies for different cohorts? Do
you have any idea on that? That would be
great to hear, actually.
Yeah. I mean, there are analytical
concepts that you can think about to do
precision. You, probably similar
to other studies, try to
find some housekeeping
(13:54):
markers. And we found some, for
instance, that are related to skin.
The skin, when you do
the landset,
or punching through your upper
layers of the skin, these proteins will
probably always be there.
So trying to figure out
what are the markers that are constant. And
then again, what we talked about
(14:14):
before, if you have a phenotype that
is changing, then you can sort of do this
resampling. You can learn from the
resampling what are the
constant constituents and what are those
that are variable, what are those that are
unreliable.
And again, the more data you have, the
easier it is to make that exercise,
because you can rank things
(14:34):
in a much more refined way.
That's very knowledgeable. You mean
using housekeeping kind of
housekeeping proteins in a way, right, to
normalize? That's pretty much the idea
around? Yeah, exactly.
And then, of course, it's just also a matter
of using different
statistical models to do
normalization and things like this. I
(14:56):
mean, whenever you have
a variable sample source, I
guess we have that also in
plasma, different
degrees of hemolysis, different fat
content, different
hydration states, they can
influence so many things. So I think just
keeping being on your toes when you look at
the data and not get carried
(15:17):
away too quickly,
I think
something that's very helpful.
Coming back again, and I'm sorry I
monopolized the questions, coming back
again to Cindy's
question (15:30):
Do you think there will be a
new breakthrough? You'll be like going
through new matrices, like
interstitial fluid, for
example, do you think that new materials
will open new ways, new research areas,
and learn quite a lot? What is your
feeling about that? What is your vision?
I
think we should accept
(15:52):
the concept that not all material
will be informative for all
studies we do. So I think if we
find the niche, that they are
informative. So again, this
study we did on interstitial fluid, we
also detected, for instance, antibodies
against SARS-CoV-2 in interstitial
fluid. And then of course, that
is a proof-of-concept we did
(16:13):
because we were curious and we had
material, or we had plus/minus as the
phenotype. But imagine you're
treating someone with melanoma, with a
biologics. How can you assure
that the biologic actually
reaches the area where it should act?
I think these things could,
of course, be much more informative than
looking at a blood sample where you
(16:34):
say, yeah, it's in your system,
but we don't know if it actually reached the
point where it should be
doing the job. So again,
these things, I think, open up new ways
and, and trying these
these new methods of sample
collection. And then, of course,
having the perfect tool that analyzes these
(16:54):
samples. And again, the
fantastic low
volume requirement of Olink
has, for us, been this
perfect match.
So we're super happy that we have a
tool that we can test these ideas and we
can demonstrate it's actually
feasible.
To return to what you're excited about in
(17:15):
terms of these longitudinal studies,
have you had much interaction with the UK
Biobank in terms of samples at
scale? I guess you don't have to worry about
sort of the dried blood spot collection.
I mean, that's really promising, but
here it is. We have a huge
data set. Have you been involved much with
the UK Biobank?
Indirectly, yes. I mean, I've been
(17:36):
talking to Chris Whelan and others. And of
course, when
Karsten Suhre, Mark McCarthy, and I
started to write this review in
Nature Genetics a couple of years back,
we thought of UK Biobank as the
audience,
especially Mark
McCarthy who I consider my mentor.
(17:56):
He was in Stockholm and I talked to him
and said, "Mark, you're doing this fantastic
work. And I think proteomics like Karsten
Suhre has shown, is a perfect match with
genetics. Can't we write up something as
bringing different perspectives together
into one piece of
information?" And that's sort of how this
whole idea started. We actually called
up Karsten and said, "Karsten, we have this idea,
(18:17):
do you want to join?" And this is sort of
where we joined forces. I learned so much
about genetics, and others learned about
proteomics. So I think that
sort of was, of course,
the dream
coming true as writing
something that adds value, but learning
something at the same time. And then, of
course, UK Biobank
(18:38):
being, as has been shown, a
fantastic study
now being powered by all these
new data that is coming out.
But, yeah, again,
it's often a one timepoint
picture, but we want to create a movie of
our lives, right? And the movie tells the
story much better. And we should probably
just explain Mark McCarthy, although I don't
(18:59):
think he needs an introduction. He's such
a well-known figure in
our world, certainly, but he's at
Genentech, of course, but he's one of these
geneticists that has
crossed over into industry. And
just anything he
focuses on I like to
keep an eye on, because it
moves and shakes. He was at
(19:19):
the International Congress of Human
Genetics, and so involved in the
leadership, talking about how to
increase diversity in genetics. And
I love
that Nature Genetics paper. So I just wanted
to say, "Karsten,
you, and Mark M, it's just
such a pleasure to have
you even talking about our technology. It's
(19:41):
very exciting."
And I think,
of course, we wanted to be
as agnostic and fair as
possible, because I think
every technology has its pros
and cons, and I think it's up
to everyone to make a decision what is the
best fit for the situation.
Absolutely. But I guess
(20:02):
coming back to your question about
longitudinal studies,
which we've been also doing
locally, led by
Mathias Uhlén, and we've been working with
Jochen Schwenk from the SCALLOP
cohort. You know, of course, that
that is when it all sort of comes to life,
right? When you see a signature, you can
understand stability, you can understand
(20:24):
that a person has had an infection,
things go up, things go down, but someone
loses weight, things change. So that's
when the information actually becomes
clearer. And that's a
fascinating thing (20:36):
to be able to look
at this real time biology.
I appreciate you talking about this
review paper for the audience.
The paper I believe you're talking about is
"Genetics meets Proteomics:
Perspectives for Large Population-based
Studies." It was in Nature Review Genetics in
January 2021.
(20:56):
I'm trying to remember a
different Karsten Suhre review. I
think you're talking about maybe one from
2017
or 2019. At any rate,
the ability to
monitor real time health as
people transition from a state of
health to one of disease.
I finished a book recently, "The Age of
(21:17):
Scientific Wellness," from Leroy Hood
and Nathan Price,
and it talked about these disease
transitions, where if
somebody's healthy, they don't have
symptoms, but something's happening
in the body, something's happening with
their metabolism, something's happening with
their metagenomics, something's happening
with their proteomics and the
circulation. And that is just
(21:39):
this fascinating thing because you're
talking about wellness, right? We need to be
sampling "well" people. And I think
the UK Biobank gives this unique
perspective. I'd like to hear your
perspective on that.
Yeah, of course, I mean,
UK Biobank offers, as far as I
understand, really a range
of phenotypes. I assume
(22:01):
some involvement was in selecting particular
sort of
disease groups and enriching them for those
that are maybe more prevalent than others.
But just to have that
breadth is really amazing
because often you're limited
to
certain sample collections.
And maybe I
(22:21):
take another sort of, open another bracket
and take a little detour here. But again,
when you do this dried blood spot random
sampling that we did, you include everyone.
You don't include only the ones that are
sick, and they only come when they're sick.
So, you know, okay, CRP [C-reactive protein] and all the other
friends, they're all already up, right?
But we want - so how do you get
that cross sectional, that
true sort of population-based
(22:43):
variance? And I think that's only possible
in a coordinated way, like UK Biobank did.
And there are other biobanks that
all of us in the U.S. and
others are trying to do similar
things. That when you learn this
is the human variability with all the
genetics, with lifestyle, with
social economic factors
influencing
(23:04):
who you are on a molecular level.
So yeah, fantastic. And then having
proteomics in that play is of course
something I get particularly excited about.
And it's this
combination - I'm sorry, go
ahead. I'm sorry, Dale,
go first.
Oh no, Sarantis, this is your show. Go right
(23:24):
ahead, please.
Thank you. Now I'm talking
about how you mentioned about
proteomics, genomics and
disease, and there's
a preprint with Anders Malarstig now
that recently came out, and they're going to
be soon published - finger crossed - with Olink.
Would you like to share a little bit
information about the research there
(23:45):
and the cohort that you use and what's your
main findings? Because it's exciting
to use the circulating proteome
to identify prognosis [biomarkers] for
breast cancer, early prognosis for breast
cancer. I'm happy to hear a little bit more
from you, actually.
Yeah, so
we're talking about the KARMA Cohort, which
is a Swedish breast cancer population
(24:05):
cohort that invites all women
in Sweden undergoing mammography
screening to participate. This is
spearheaded by Per Hall and
Kamila Czene at Karolinska
Institute. And with both I've been
collaborating already for quite some
time and we recently got
some funding to continue our collaborations
and then brought in also
Olink data that we generated
(24:27):
in the lab
to look at breast
cancer risk. So in addition to this
paper that you mentioned, Sarantis, there's
also another one that's been
circling around now
where we wanted to primarily
identify - Can we use proteins
to predict short term risk of
breast cancer? Genetics can do that on
a more longer period
(24:49):
of time, but can proteins
add something to it? And then, of
course, with Anderson and
the leader and the SCALLOP
consortium, and we want to also to bring
also genetics into this. And
I'm not a geneticist. So for
me, again, it's always fascinating to see
proteomics data in action.
(25:09):
Which is what
it makes me most proud
because I think
what it is exciting to
learn when other datas inform you about
your own data and when others take the data
that you generate or that you
know more about and they tell you new
stories. And then, of course,
(25:29):
the KARMA cohort is really a population-based
code and it's really unique in a sense that
we're not only looking at
breast cancer cases,
or the study could
continuously collect sample
patient information or personal information,
and eventually some of these persons will
become patients. And luckily, then
(25:50):
we would have, let's say, a blood
sample from the last time when the patient
was still a person, so to speak, when you
think about these two categories. So we can
go back in time and see are there any
things in the
prior history that could
lead towards okay, you
are actually on a much different
trajectory than the remaining
(26:10):
individuals, so that's all that
people have different lives. We know drugs
played a role, pre- post-menopausal
plays a role,
hormonal replacement therapy plays a role.
So lots of things happen. But then genetics
can tell you an unbiased story about all
these phenotypes and that,
again, gives a new angle to
(26:30):
this whole problem. And in this study,
we used Mendelian
Randomization and found five
interesting proteins that
presumably have a causal role in breast
cancer. And of course, this is the study.
Now, it's only nowadays
600 individuals, but still, it's a
really very fine selection of
(26:51):
samples that could
lead the way. And then again, taking the
road that genetics has
taken, we can use data that exists in
other biobanks and we can sort
of look, do we see the
same associations in these? And
that's, of course,
when multiomics doesn't
become a picture, it becomes a movie, where
(27:12):
we take different [aspects] of these
relationships.
What a great illustration
and analogy.
That's a
great analogy, right? Not a picture. We've
got a movie.
Yeah. And I think that's what
we need, right?
You take
(27:33):
a look at a picture and you interpret so
many things into this, whether you
know something about the painter or
the time when the painting was made.
But if you have a movie, it tells you
much more. It tells you a dynamic that
you cannot really see in a
picture. But anyway, so
again, we had this opportunity, and then,
(27:54):
Asa [Hedman] and Anders [Malarstig] have been really leading
this together with colleagues at Olink
and others, to sort of
find out whether these things we
identified in the Swedish KARMA study also
we can see in the UK Biobank or
in Finngen and it seems so to be the case.
And of course, that gives much more
certainty about that. These are interesting
(28:14):
findings to follow up. And
again, I think what we talked, I guess
before this podcast started,
that then you can
start to develop drugs, and you can see what
actually happens when you give someone a
drug that addresses one of these proteins.
Then you, again, start a new movie,
right? But,
on a different direction.
And again, then use
(28:36):
proteomics to follow and see what
happens. So,
that's fascinating, I think. And so
who does that follow up?
Are you involved in that
kind of
obviously,
proteomics
is your field,
and I just wonder if there's
(28:58):
another
function that takes that to the
clinical trial or to the
test bed to
try out
these drugs that
affect these pathways.
Yeah.
Of course, I guess it would require that
we have the right partners who would have
(29:20):
the libraries to do drug screening on these,
and sort of it's an army of new
things to engage.
But of course,
primarily to see that
what we do in these
studies has a value and then
again, translate it back to functional
studies, which, again, is something
I think will also happen in the next couple
(29:42):
of years, is taking all these big
biobank screenings back into some sort of
functional studies to see, okay, is it
really the molecule? Is it really the
phenotype? Is it
really the drug or the
lifestyle effect? And
that's going to be sort of looping back
where it started, from cellular
studies into
(30:03):
systemic studies, and then back into the -
Wouldn't it be amazing to
find a lifestyle effect that
we never thought might have
an impact?
Yeah. Having the tools to
be able to start parsing these things. It's
fascinating.
(30:26):
To talk a little bit about the
study itself, right? This Nature
Preprint looked at 300
individuals from this mammography study who
had breast cancer
diagnosed over those two years, that
they took a look at them, and then you
matched it with 300 normal individuals from
that same study. You had genotype
information, right? And the
(30:47):
genotypes combined with the
Explore 3000 [platform] in
terms of the 2900
proteins. So you had 600
individuals, 2900
proteins, and then you discovered
800 pQTLs [protein Quantitative Trait Loci]
and controlling
737 proteins. And I
thought that was fascinating, right? That we
(31:07):
have genetic control of
737 proteins that can
identify the variants
pQTLs, and then you can drill down
and get five likely
causative proteins of breast cancer. Do I
understand that correctly? That these
five proteins you identified
were previously not
investigated or investigated with
(31:28):
certain sort of weak associations
of breast cancer? But the take home
message is that these five proteins were
new discoveries.
Yeah, that's our
understanding. I mean, of course,
maybe someone else has already figured this
out, but not told the public about
it. But I
think
an important aspect is also to say
(31:50):
that these 300 cases,
they were not cases at the time of
sampling, they were future cases. So when
they were actually sampled, they were still
considered persons, not patients.
So that's, again, also
to understand,
then we have this list
of proteins that all tell
different stories. And I think it's
(32:12):
fascinating to be sort of in your
mind, thinking about what actually the
role is. But what we need now is, of
course, the hard data that tells us this
is true, or this is
actually the opposite.
And I guess to back up
to the original KARMA study, which was out
of KTH. There were some
70,000 women from Karolinska who
(32:33):
volunteered. From Karolinska.
Okay. 70,000 women,
though, over a couple of years. Is
that correct? I think about
the effort
involved.
Yeah. This is the nice thing about
doing science. In Sweden -
I originally come from
Germany - it's a different
(32:54):
system. But in Sweden,
maybe because of the Nobel Prize, maybe
because of the public interest
in science, there's much easier
engagement. And in
women, I think, in other countries, have
this regular sort of health checkup. So
there's this mammography screening program,
and then you get basically asked, "Do
(33:14):
you want to participate?" And then Per Hall
and his colleagues do the
magic and keep people
engaged, and people
follow, which
is super. And
then you mentioned briefly
the power of replicating these
results, because that,
(33:35):
I think, is an important dimension of this
paper, in that it wasn't just a
single finding in a particular
population that you were able
to find. We're actually able then to go
back to was it? Finngen and the UK
biobank? And then look
at the genotypes, look at the
protein levels, and then being able to
actually show, yes, this connection holds
up. I think that's pretty
(33:56):
significant.
Could you comment on that?
Again,
this work was spearheaded by Asa Hedman,
but again, what I see is
you have, let's say you create this
currency, let's say the
pQTLs. This is a currency you can
go and you can pay in other
(34:17):
countries or in other
biobanks. You can use that currency to
exchange information. And
this is, I guess, what genetics has
really enabled us to do. And now
proteomics is learning how it can do
it. We have different
technologies, they may have different
outcomes, different information. But again,
you can anchor it on the genetics. You can
use the pQTLS, you can
(34:39):
use them as instruments in mendelian
randomization to exchange this information.
And that's
amazing. Yeah. Here it is. You're
talking about empowering proteomics with
genomics, right?
Turning it around instead of coming at, I
mean, I come from a genomics background. So
I think of it in terms of proteomics adding
to the genomics. Here it is. You come from
the proteomics background, and it's the
(35:00):
genetics that is really enriching
the findings. And I think that's
great. Jochen, I have a basic
question, and it's very basic. We
mentioned that one of the factors
could be lifestyle, like
environment, but also could be the hormones.
Do you have, let's say, relatives,
like mother, sister, or twins
(35:20):
that you can control in this cohort? I
imagine there will be also some twins that
you can, let's say, somehow
discriminate and identify the genetic
background versus the environment. This
is the first part of my
question. The second part is for sure,
you would check, like, post- and
pre- menopausal. Then have you seen
differences? What's
(35:41):
your feedback on that? What's your
experience around this type of
observations?
Yeah, I mean, we have had
previous studies that we published using
other own technologies that we
used 510 years ago, where we
specifically looked at hormonal replacement
therapy as one of the factors which,
to our surprise, had really a long
(36:02):
lasting effect on the women's
proteome, which
again, is really
quite significant.
And then, of course, that pre- and
post-menopausal breast cancer.
Again, this is something I've been
learning from my colleagues that I work
with, is very different.
Then, of course, you need to
disentangle. So,
(36:23):
I'm sorry, I want to click
back to what you just said about
hormone replacement therapy having a lasting
effect on the proteome.
What do you mean by that?
Meaning that it shifts
the proteome? But what about the
risk, the cancer risk, right.
Because certainly the women's health
study here in the U.S.
(36:44):
had led to some concerns around
that. I'm just curious
if that's part
of the impact on the proteome,
do you think? Or maybe what we
found in this other study is that we
had a subset of women that really
we could sort of see that previous
use of hormones had
(37:04):
a significant change
in their proteome and also increased their
future risk that they were developing breast
cancer. So, of course, this really
sort of showed up. But it's a
small subset of all the women
that we tested.
It's a great thing to follow up.
(37:25):
Of course, we still need to understand is
what is the effect of taking hormones?
Do you actually have remodeling of
some
reproductive pathways that
constantly do something?
And if they get sort of, let's say, pushed
off track, they will stay on that off
track path for a longer period of time. And
(37:45):
then there will be feedback loops with,
let's say, the liver and other organs to
just try to adapt
with the sort of external trigger.
So that was sort of part
of our sort of understanding of
the use of drugs. But again, it
showed that taking medication has
a quite substantial effect on
your proteome. And we
(38:06):
found it fascinating that it actually seemed
to be consistent
over many years, a
picture of real time biology, the
proteome, right.
And, what we
know about effects of certain
treatments, what we know about effects of
certain drugs, we're just scratching the
surface. Right. A number of our pharma
(38:27):
partners and customers of Olink
are finding out so much with just
a limited set of proteins.
They're not looking at the proteome, they
might be looking at a panel
of 50 or 90 [proteins], or what
have you. But there is just so much to
learn about the biology.
Sarantis, you started to ask, I think you
were down a path of a couple of questions. I
(38:48):
was curious.
I want
to ask just a more
philosophical question about
if you have some, let's
say, mother,
sisters, some relatives, or if
you have some twins that you can follow.
And you can see the change comes from
genetic background, or comes from the
proteomics background, or combination, or
(39:10):
neither. Do you have any experience on
that? Have you seen some patterns
around [that]?
I don't think we
necessarily looked into this.
But I've been working with
another twin cohort from Sweden called Twin
Gene, which the name says
has a quite clear
focus on these aspects.
(39:31):
And
no,
not that I think in particular, but of
course, it's, again, what we
pass on to our children is something
that will be, in the future,
helpful for them to know.
And maybe
they will change their lives when they know.
(39:51):
Okay. I'm at a higher risk of a certain
disease because both my parents
passed away. I guess you see a lot of these
breast cancer studies and effects in
Iceland, I think, right.
But
not in these studies. I cannot
recall that we actually specifically looked
into this. Yeah. What was I think really
interesting about this particular paper
(40:12):
on breast cancer is that you looked
into so many different kinds
of connections
in terms of inherited risk
as well because the title
is paper, "Evaluation of
Circulating Plasma Proteins in Breast
Cancer and Mendelian Randomization
Analysis," you're actually
looking at, then, the entire genetic
(40:32):
backgrounds of unrelated
individuals and just saying what
is elevating that particular
risk. And understand, these five
proteins that were differentially
regulated were basically
lifetime
exposures. That a person
was exposed to a high level of
protein throughout their whole life. And I
think that's what makes this really
(40:54):
fascinating, right? The
proteomics being informed by the
genetics controlling the levels of
protein, and then saying these
five proteins actually become drug
targets, which I thought was
just a fascinating realm.
Before we wrap up, Jochen, would you
like to make any final comments?
Either, I don't know, about
(41:14):
where we are,
where we're going, working with
Olink? Oh, I understand, right? We
didn't even talk about a very
famous
postdoc, famous at Olink,
Philippa [Pettingill] came out of your lab. I don't know
if you want to talk about what it was like
working with her. She has helped
me, Cindy, with her title. She's
Director [of Application Sciences]. She's a superstar.
(41:36):
She runs the field application
scientist team within
the European
region, and she
is absolutely magnificent.
She's also helped lead
our discussions around
statistical analyses in the
UK Biobank Project. She's
just such a magical
human being to
(41:58):
have at Olink. We're so lucky to
have her. And she is a product
launched
out of your lab. At
some point,
you had an impact on
her trajectory. So, yes,
please, anything you have to say about her
would be greatly appreciated. We had
(42:18):
hoped we'd be able to have her on, but
we weren't able to get her
into the timing that we had
going. No, I
mean, all the success that
she has now is because
of her
engagement, her knowledge, and her
curiosity. But, yeah, it was fantastic
to work with her. She was with me about
(42:39):
one and a half, two years. It
was
inspirational and fun
from the first to the last day.
And I think
to see someone leaving the lab
and making such a
wonderful career
is fantastic. I guess if my
contribution is that I showed her all these
(43:00):
different tools that we had in lab,
including Olink and others, and we talked a
lot about the different assays, the
different concepts. So if that has
helped her in achieving these
fantastic things that she's doing with
you, it makes me proud and
happy. I think she deserves it.
And I wish, of course, her
all the success, and
(43:20):
anytime we see her, we see
each other on the media calls,
it's like old friends.
I think she came out
absolutely a leader, and I think
she has such great things to say
about the time that she spent in your
lab. And I think that's
(43:40):
pretty sweet. Thank you.
That's great to hear. All right,
well, thank you very much for joining us
today, Jochen. We've really enjoyed the
conversation. Thank you for having me.
It was fantastic. And
continue with this great podcast.
It's really a treasure. Thanks a lot
for setting this up and running
it. You're so
kind. Thank you.
(44:02):
Okay, well, I think
that's it. Thank you.
Thank you for
listening to the Proteomics in Proximity
podcast brought to you by Olink
Proteomics. To contact the hosts
or for further information, simply
email info@olink.com.
Welcome to the
Proteomics in Proximity podcast, where
your co-hosts Dale Yazuki, Cindy
Lawley, and Sarantis Chlamydus from Olink
Proteomics talk about the intersection of
Proteomics with genomics for drug target
discovery, the application of proteomics to
reveal disease biomarkers, and current
trends in using proteomics to
(00:25):
unlock biological mechanisms. Here we have
your hosts, Dale, Cindy, and
Sarantis. Hello, everybody. I'm
Sarantis. I'm together today with Dale
and Cindy for another episode
of our
great podcast, Proteomics in
Proximity. We are all very happy
to have like a guest, Professor
Johann Schwenk
(00:45):
who holds a
position at the University of
KTH University [Royal Institute of Technology in Sweden].
And today he's
a protein expert and
a professor in
Translational Proteomics. And today we'll
discuss a little bit about his
new research, about his
research interest and how proteins can
enable multiomics approaches. Actually,
(01:07):
Jochen,
thank you very much for joining
today. And I would like to
start the discussion asking you (01:13):
What
does translational proteomics mean to you?
Yeah,
I think when we started thinking
about the title for a
professorship, translation was really a hot
topic at the time (01:28):
to bring
something you've been doing in the lab
into a clinical context. But I think
it turns out to be much more than this. This is
actually to explain also what we're
doing in the lab to others, so that
the community can engage into
our research and we can even find a
broader utility. So it's
still the idea of
(01:48):
connecting the lab environment
with clinical and population
health. So I think
hopefully one day we'll contribute to that
activity. That's great.
And I saw also that you study biochemistry
in Tübingen, University of Tübingen. That's quite
famous from the biochemistry industry worldwide. Do
you have any story that you would like to
record for your first paper? For example,
(02:09):
being in Tübingen, and that would be great
to hear.
Yeah,
Tübingen of course gave me a
fantastic time. We were very
small number of students per semester. I had
very close connection to the
professors. I got a chance to go to
a Lindau Nobel Laureate
meeting there. And it was really,
(02:30):
I think, an inspirational time that
sort of created
a lot of curiosity about science. And
then after that I moved a little bit more
into technology. So in the early two
2000s, when I did my Masters and PhD, I
worked with Luminex-based
assays, which at that time
was really new.
Then sort of that took me then
(02:50):
to join the Protein Atlas in
2005 as a postdoc. And somehow I
got stuck with this fantastic project.
I'm
still around and
learn every day something new about
proteins. I mean, just being
there and to work with Mathias Uhlén
and all the colleagues has been truly
inspirational. So you've been at
SciLife Labs since 2005, and that was when
(03:13):
it began at KTH, is that correct?
Yes. SciLife lab was
inaugurated in 2010. So actually,
it was my and three other groups
that moved into the building under
construction, I think it was in
October 2010.
So I consider myself very much of an
oldie when I think about my time at
SciLife Lab. I've seen it
(03:33):
change, grow, and now, I think,
become a very prominent research institute
in Europe. So it's, I
think, fantastic and very
much also gave me opportunities to learn
about other technologies and to
learn how information about
proteins can be useful. And I
have many stories to tell, but
one of which is, for instance, I have a
(03:54):
little bit of a side activity project around
GPCRs, and that, for instance, I think
wouldn't have been possible if I would just
be sitting somewhere in a lab and
not be exposed to all these different
activities. Sure. And for those not
familiar, GPCRs
or G-
coupled protein receptors.
GCPRs, is that correct?
(04:15):
G-protein coupled
receptors. G-protein coupled receptors. I
got to get my acronym straight.
It's a really important drug
target, right? Membrane
proteins. Yeah, membrane proteins that are
important drug targets. Correct?
Exactly. Yeah.
And they're at the SciLife
Laboratories. Well, you said that you were
involved in the Human Protein Atlas way back
(04:37):
in 2005, so therefore
the genome had just been
finished in 2002.
2003.
Those must have been pretty exciting
times because there was a big pivot
and interest and focus on the
proteome, is that correct?
Yeah, of course. And then at
that time, Mathias
and us were producing all these
(04:58):
antibodies was fairly unique, and people
were not really sure whether that would add
any value to the field
that's dominated by mass spectrometry. But I
think now we've shown, and
the way that we've brought in new
data, trying to understand the data that we
generate, and then sort of give feedback
to other data types with
localization of
(05:18):
subcellular compartments
of proteins that
I think are really
super valuable and help us to disentangle
the complex biology that we
live in. And
my real interest is
proteins in the circulation. So that's even
more complicated because
you're under sort of the
(05:39):
constant exchange of molecules
in all parts of the body.
So it's not as organized as looking at
subcellular localizations, but it's still
fascinating. And I guess
that's something I really
sort of fell in love with, and I
really enjoy doing.
Great.
(05:59):
How would you see Olink because
you are a biochemist? I
mean, you are a mass spec
expert. How will you see
Olink fitting on this pipeline of
mass spectrometry? How do you see mass spec and Olink
working together from your experience?
Because you have a big experience in this
field.
I'm very fortunate to get to know Ulf Landegren
many, many years back, and we've been
(06:20):
sort of seeing each other on a regular
basis, partly because Uppsala [Olink] and Sweden [SciLife Labs] are
very close and sort of been
doing things in parallel. And of course,
fantastic to see the journey that with the
proximity extension, proximity ligation,
and all these different
versions of this concept have
now, I think,
been the driver for using antibodies as
(06:42):
molecular tools. I mean, there was
just a paper in Nature Methods, I think,
just the other day. Again antibodies
conjugated with oligonucleotides. I think
that's giving people much more
bigger field to play and use these
reagents. So
of course I've seen how it
started and we've been
very late to the game. My lab,
(07:03):
or the unit that I'm heading
at SciLife Lab, started to
introduce Olink in 2017.
And since then, we've been super
happy to have the system in-house
and do this
for others, users that
come to SciLife Lab to just want to
have data, but also for our
research. So I think, again, any
(07:24):
data type adds a value to what we do. And I
think Olink has truly
enabled us to do many things we weren't
able to before. So,
fantastic. I'm dying
to ask
in front of you,
based on what you see, the
opportunity in front of you
with this technology, whatever
(07:45):
technologies, right. What is it that you're
most excited about for the
future? Are there aspects
of
work that you've been doing or a
direction that you're going in, that
you would like to share, that you're
comfortable sharing:
I just want to know
what's the part that makes
you go into
(08:07):
flow? What do
you want to do next?
Yeah, I started
doing a lot of assay development myself
when I worked with Luminix 20 years
back. That sounds a little bit silly when I
say this, but it's the truth.
Of course, that sort of
has always been something to try, something
(08:28):
new to maybe try, something that's
difficult, maybe not immediately
rewarding, but in the
long term, something that could
be very fruitful or something that
makes you proud as the researchers that you
say [to yourself] (08:40):
"Okay, this is something I believed in,
and I see it's happening." So the
next sort of moment for me,
when I had this type of thinking was
when COVID started and when
lots of people went into serology
testing or protein testing in
the classical way, when me
and my colleagues at KTH, we said,
"Let's try something different and use dried
(09:00):
blood spots. Let's not ask
people to come to the clinic, let's send
the devices back to them,
to their home, so they can collect
bloods in their kitchen
and sofa, wherever, and send
them the samples back to us, to the lab,
where we can do the research."
So that, I think, really
inspired a lot of new ways of
(09:21):
doing this. When you think about cutting
costs, simplifying workflows, freeing
the time of people in the clinic,
but also to think about
doing health monitoring. I
think people
always often ask me, what do I think is
proteomics best used
for? And I think it's monitoring.
It's like looking at who you are and looking
(09:42):
how you change. And I think this
combination, I think, now is sort
of shaping towards something I really
am passionate about. And dried blood spots
is a fantastic tool.
It's more challenging than doing it
the classical way. There's so much more to
learn, and then maybe even go
further. When
Sarantis and I talked a couple of
(10:04):
weeks back, also to look at even smaller
sample volumes, looking at other body
fluids, such as interstitial
fluid, that could even tell us something
extra that blood is not able to tell us.
We'd need a
baseline to understand the reference of
dried blood spots to be making that
comparison, right? Yeah. And
(10:24):
so enabling in areas where we
can't get a blood draw,
a phlebotomist out to do a
blood draw.
I think that's going to be really
important. Dale?
So if you can give us some background
on dried blood spots, and I'd appreciate it,
because my only familiarity with it was when
I had my first child, and they did a
(10:45):
heel prick, and then they went ahead
and used the blood from
that little lancet onto a
particular card. Is there something
special about the material they
use for a dried blood spot?
And what are the challenges as far
as working with proteins in that context?
I think
let's say if you think
(11:06):
from an analytical
perspective and a precision
perspective, the dry blood spots your
kids
donate, they are just put in a
filter paper to do a plus/minus test.
So it's really a binary answer you're
after. But if you really want to look
at subtle changes in the human
phenotype, I think then you need
to ensure that the precision of the
(11:28):
material you use in your system is
there.
Normally when you have a dry blood spot, you
get sort of a donut distribution of the red
blood cells, so it really matters where you
do the punch. So you want to avoid these
type of things, especially if you want to do
it at scale, if you want to do it
consecutively. So I
started to work with a local company that
(11:49):
was founded by one of my colleagues at
KTH. And they use a
microfuidic system to exactly collect
ten microliters. So just knowing
that what you put into your system is ten
microliters, of course,
then there are different levels of
hematocrits, there are different other
things that you need to consider, but you
at least eliminate some of the concerns that
(12:10):
you have. That,
I think, is really the key. And of course,
it's a simplicity of this
procedure that you can assure it's
easy for people to do. And
they manage, even though they may not
be trained. I failed also
when I did it the first times.
But if you
get used to it, the quality is
(12:31):
really excellent. And I think there are also
studies showing that more and more are using
other devices
that are out there, I
think now you have a new material
which is sort of similar to plasma,
but it has some bonus. And the question is,
how do you manage that bonus? Is
it something that is a challenge, it's a
burden that makes it difficult for
(12:52):
you
to be analytically
precise? Or does it open up
opportunities that were not possible
when you looked at the regular blood,
plasma samples? Because of,
let's say, the
hematopoietic cells that are still there,
they may leak out something that could be
really exciting. So it's
this balance between things that, I think ...
(13:13):
When you have a discussion a little bit
with other proteomics expert about
dry blood spots, there's always a question
about how you control and
normalize. Because, I don't know,
from what I have heard,
actually, it's not easy to have always
the same type of dry blood spots. I'm
guessing that there's a lot of varieties,
a lot of variation can be there. Would you
(13:34):
have any idea how one can normalize this
data in order to have like, longitudinal
studies or studies for different cohorts? Do
you have any idea on that? That would be
great to hear, actually.
Yeah. I mean, there are analytical
concepts that you can think about to do
precision. You, probably similar
to other studies, try to
find some housekeeping
(13:54):
markers. And we found some, for
instance, that are related to skin.
The skin, when you do
the landset,
or punching through your upper
layers of the skin, these proteins will
probably always be there.
So trying to figure out
what are the markers that are constant. And
then again, what we talked about
(14:14):
before, if you have a phenotype that
is changing, then you can sort of do this
resampling. You can learn from the
resampling what are the
constant constituents and what are those
that are variable, what are those that are
unreliable.
And again, the more data you have, the
easier it is to make that exercise,
because you can rank things
(14:34):
in a much more refined way.
That's very knowledgeable. You mean
using housekeeping kind of
housekeeping proteins in a way, right, to
normalize? That's pretty much the idea
around? Yeah, exactly.
And then, of course, it's just also a matter
of using different
statistical models to do
normalization and things like this. I
(14:56):
mean, whenever you have
a variable sample source, I
guess we have that also in
plasma, different
degrees of hemolysis, different fat
content, different
hydration states, they can
influence so many things. So I think just
keeping being on your toes when you look at
the data and not get carried
(15:17):
away too quickly,
I think
something that's very helpful.
Coming back again, and I'm sorry I
monopolized the questions, coming back
again to Cindy's
question (15:30):
Do you think there will be a
new breakthrough? You'll be like going
through new matrices, like
interstitial fluid, for
example, do you think that new materials
will open new ways, new research areas,
and learn quite a lot? What is your
feeling about that? What is your vision?
I
think we should accept
(15:52):
the concept that not all material
will be informative for all
studies we do. So I think if we
find the niche, that they are
informative. So again, this
study we did on interstitial fluid, we
also detected, for instance, antibodies
against SARS-CoV-2 in interstitial
fluid. And then of course, that
is a proof-of-concept we did
(16:13):
because we were curious and we had
material, or we had plus/minus as the
phenotype. But imagine you're
treating someone with melanoma, with a
biologics. How can you assure
that the biologic actually
reaches the area where it should act?
I think these things could,
of course, be much more informative than
looking at a blood sample where you
(16:34):
say, yeah, it's in your system,
but we don't know if it actually reached the
point where it should be
doing the job. So again,
these things, I think, open up new ways
and, and trying these
these new methods of sample
collection. And then, of course,
having the perfect tool that analyzes these
(16:54):
samples. And again, the
fantastic low
volume requirement of Olink
has, for us, been this
perfect match.
So we're super happy that we have a
tool that we can test these ideas and we
can demonstrate it's actually
feasible.
To return to what you're excited about in
(17:15):
terms of these longitudinal studies,
have you had much interaction with the UK
Biobank in terms of samples at
scale? I guess you don't have to worry about
sort of the dried blood spot collection.
I mean, that's really promising, but
here it is. We have a huge
data set. Have you been involved much with
the UK Biobank?
Indirectly, yes. I mean, I've been
(17:36):
talking to Chris Whelan and others. And of
course, when
Karsten Suhre, Mark McCarthy, and I
started to write this review in
Nature Genetics a couple of years back,
we thought of UK Biobank as the
audience,
especially Mark
McCarthy who I consider my mentor.
(17:56):
He was in Stockholm and I talked to him
and said, "Mark, you're doing this fantastic
work. And I think proteomics like Karsten
Suhre has shown, is a perfect match with
genetics. Can't we write up something as
bringing different perspectives together
into one piece of
information?" And that's sort of how this
whole idea started. We actually called
up Karsten and said, "Karsten, we have this idea,
(18:17):
do you want to join?" And this is sort of
where we joined forces. I learned so much
about genetics, and others learned about
proteomics. So I think that
sort of was, of course,
the dream
coming true as writing
something that adds value, but learning
something at the same time. And then, of
course, UK Biobank
(18:38):
being, as has been shown, a
fantastic study
now being powered by all these
new data that is coming out.
But, yeah, again,
it's often a one timepoint
picture, but we want to create a movie of
our lives, right? And the movie tells the
story much better. And we should probably
just explain Mark McCarthy, although I don't
(18:59):
think he needs an introduction. He's such
a well-known figure in
our world, certainly, but he's at
Genentech, of course, but he's one of these
geneticists that has
crossed over into industry. And
just anything he
focuses on I like to
keep an eye on, because it
moves and shakes. He was at
(19:19):
the International Congress of Human
Genetics, and so involved in the
leadership, talking about how to
increase diversity in genetics. And
I love
that Nature Genetics paper. So I just wanted
to say, "Karsten,
you, and Mark M, it's just
such a pleasure to have
you even talking about our technology. It's
(19:41):
very exciting."
And I think,
of course, we wanted to be
as agnostic and fair as
possible, because I think
every technology has its pros
and cons, and I think it's up
to everyone to make a decision what is the
best fit for the situation.
Absolutely. But I guess
(20:02):
coming back to your question about
longitudinal studies,
which we've been also doing
locally, led by
Mathias Uhlén, and we've been working with
Jochen Schwenk from the SCALLOP
cohort. You know, of course, that
that is when it all sort of comes to life,
right? When you see a signature, you can
understand stability, you can understand
(20:24):
that a person has had an infection,
things go up, things go down, but someone
loses weight, things change. So that's
when the information actually becomes
clearer. And that's a
fascinating thing (20:36):
to be able to look
at this real time biology.
I appreciate you talking about this
review paper for the audience.
The paper I believe you're talking about is
"Genetics meets Proteomics:
Perspectives for Large Population-based
Studies." It was in Nature Review Genetics in
January 2021.
(20:56):
I'm trying to remember a
different Karsten Suhre review. I
think you're talking about maybe one from
2017
or 2019. At any rate,
the ability to
monitor real time health as
people transition from a state of
health to one of disease.
I finished a book recently, "The Age of
(21:17):
Scientific Wellness," from Leroy Hood
and Nathan Price,
and it talked about these disease
transitions, where if
somebody's healthy, they don't have
symptoms, but something's happening
in the body, something's happening with
their metabolism, something's happening with
their metagenomics, something's happening
with their proteomics and the
circulation. And that is just
(21:39):
this fascinating thing because you're
talking about wellness, right? We need to be
sampling "well" people. And I think
the UK Biobank gives this unique
perspective. I'd like to hear your
perspective on that.
Yeah, of course, I mean,
UK Biobank offers, as far as I
understand, really a range
of phenotypes. I assume
(22:01):
some involvement was in selecting particular
sort of
disease groups and enriching them for those
that are maybe more prevalent than others.
But just to have that
breadth is really amazing
because often you're limited
to
certain sample collections.
And maybe I
(22:21):
take another sort of, open another bracket
and take a little detour here. But again,
when you do this dried blood spot random
sampling that we did, you include everyone.
You don't include only the ones that are
sick, and they only come when they're sick.
So, you know, okay, CRP [C-reactive protein] and all the other
friends, they're all already up, right?
But we want - so how do you get
that cross sectional, that
true sort of population-based
(22:43):
variance? And I think that's only possible
in a coordinated way, like UK Biobank did.
And there are other biobanks that
all of us in the U.S. and
others are trying to do similar
things. That when you learn this
is the human variability with all the
genetics, with lifestyle, with
social economic factors
influencing
(23:04):
who you are on a molecular level.
So yeah, fantastic. And then having
proteomics in that play is of course
something I get particularly excited about.
And it's this
combination - I'm sorry, go
ahead. I'm sorry, Dale,
go first.
Oh no, Sarantis, this is your show. Go right
(23:24):
ahead, please.
Thank you. Now I'm talking
about how you mentioned about
proteomics, genomics and
disease, and there's
a preprint with Anders Malarstig now
that recently came out, and they're going to
be soon published - finger crossed - with Olink.
Would you like to share a little bit
information about the research there
(23:45):
and the cohort that you use and what's your
main findings? Because it's exciting
to use the circulating proteome
to identify prognosis [biomarkers] for
breast cancer, early prognosis for breast
cancer. I'm happy to hear a little bit more
from you, actually.
Yeah, so
we're talking about the KARMA Cohort, which
is a Swedish breast cancer population
(24:05):
cohort that invites all women
in Sweden undergoing mammography
screening to participate. This is
spearheaded by Per Hall and
Kamila Czene at Karolinska
Institute. And with both I've been
collaborating already for quite some
time and we recently got
some funding to continue our collaborations
and then brought in also
Olink data that we generated
(24:27):
in the lab
to look at breast
cancer risk. So in addition to this
paper that you mentioned, Sarantis, there's
also another one that's been
circling around now
where we wanted to primarily
identify - Can we use proteins
to predict short term risk of
breast cancer? Genetics can do that on
a more longer period
(24:49):
of time, but can proteins
add something to it? And then, of
course, with Anderson and
the leader and the SCALLOP
consortium, and we want to also to bring
also genetics into this. And
I'm not a geneticist. So for
me, again, it's always fascinating to see
proteomics data in action.
(25:09):
Which is what
it makes me most proud
because I think
what it is exciting to
learn when other datas inform you about
your own data and when others take the data
that you generate or that you
know more about and they tell you new
stories. And then, of course,
(25:29):
the KARMA cohort is really a population-based
code and it's really unique in a sense that
we're not only looking at
breast cancer cases,
or the study could
continuously collect sample
patient information or personal information,
and eventually some of these persons will
become patients. And luckily, then
(25:50):
we would have, let's say, a blood
sample from the last time when the patient
was still a person, so to speak, when you
think about these two categories. So we can
go back in time and see are there any
things in the
prior history that could
lead towards okay, you
are actually on a much different
trajectory than the remaining
(26:10):
individuals, so that's all that
people have different lives. We know drugs
played a role, pre- post-menopausal
plays a role,
hormonal replacement therapy plays a role.
So lots of things happen. But then genetics
can tell you an unbiased story about all
these phenotypes and that,
again, gives a new angle to
(26:30):
this whole problem. And in this study,
we used Mendelian
Randomization and found five
interesting proteins that
presumably have a causal role in breast
cancer. And of course, this is the study.
Now, it's only nowadays
600 individuals, but still, it's a
really very fine selection of
(26:51):
samples that could
lead the way. And then again, taking the
road that genetics has
taken, we can use data that exists in
other biobanks and we can sort
of look, do we see the
same associations in these? And
that's, of course,
when multiomics doesn't
become a picture, it becomes a movie, where
(27:12):
we take different [aspects] of these
relationships.
What a great illustration
and analogy.
That's a
great analogy, right? Not a picture. We've
got a movie.
Yeah. And I think that's what
we need, right?
You take
(27:33):
a look at a picture and you interpret so
many things into this, whether you
know something about the painter or
the time when the painting was made.
But if you have a movie, it tells you
much more. It tells you a dynamic that
you cannot really see in a
picture. But anyway, so
again, we had this opportunity, and then,
(27:54):
Asa [Hedman] and Anders [Malarstig] have been really leading
this together with colleagues at Olink
and others, to sort of
find out whether these things we
identified in the Swedish KARMA study also
we can see in the UK Biobank or
in Finngen and it seems so to be the case.
And of course, that gives much more
certainty about that. These are interesting
(28:14):
findings to follow up. And
again, I think what we talked, I guess
before this podcast started,
that then you can
start to develop drugs, and you can see what
actually happens when you give someone a
drug that addresses one of these proteins.
Then you, again, start a new movie,
right? But,
on a different direction.
And again, then use
(28:36):
proteomics to follow and see what
happens. So,
that's fascinating, I think. And so
who does that follow up?
Are you involved in that
kind of
obviously,
proteomics
is your field,
and I just wonder if there's
(28:58):
another
function that takes that to the
clinical trial or to the
test bed to
try out
these drugs that
affect these pathways.
Yeah.
Of course, I guess it would require that
we have the right partners who would have
(29:20):
the libraries to do drug screening on these,
and sort of it's an army of new
things to engage.
But of course,
primarily to see that
what we do in these
studies has a value and then
again, translate it back to functional
studies, which, again, is something
I think will also happen in the next couple
(29:42):
of years, is taking all these big
biobank screenings back into some sort of
functional studies to see, okay, is it
really the molecule? Is it really the
phenotype? Is it
really the drug or the
lifestyle effect? And
that's going to be sort of looping back
where it started, from cellular
studies into
(30:03):
systemic studies, and then back into the -
Wouldn't it be amazing to
find a lifestyle effect that
we never thought might have
an impact?
Yeah. Having the tools to
be able to start parsing these things. It's
fascinating.
(30:26):
To talk a little bit about the
study itself, right? This Nature
Preprint looked at 300
individuals from this mammography study who
had breast cancer
diagnosed over those two years, that
they took a look at them, and then you
matched it with 300 normal individuals from
that same study. You had genotype
information, right? And the
(30:47):
genotypes combined with the
Explore 3000 [platform] in
terms of the 2900
proteins. So you had 600
individuals, 2900
proteins, and then you discovered
800 pQTLs [protein Quantitative Trait Loci]
and controlling
737 proteins. And I
thought that was fascinating, right? That we
(31:07):
have genetic control of
737 proteins that can
identify the variants
pQTLs, and then you can drill down
and get five likely
causative proteins of breast cancer. Do I
understand that correctly? That these
five proteins you identified
were previously not
investigated or investigated with
(31:28):
certain sort of weak associations
of breast cancer? But the take home
message is that these five proteins were
new discoveries.
Yeah, that's our
understanding. I mean, of course,
maybe someone else has already figured this
out, but not told the public about
it. But I
think
an important aspect is also to say
(31:50):
that these 300 cases,
they were not cases at the time of
sampling, they were future cases. So when
they were actually sampled, they were still
considered persons, not patients.
So that's, again, also
to understand,
then we have this list
of proteins that all tell
different stories. And I think it's
(32:12):
fascinating to be sort of in your
mind, thinking about what actually the
role is. But what we need now is, of
course, the hard data that tells us this
is true, or this is
actually the opposite.
And I guess to back up
to the original KARMA study, which was out
of KTH. There were some
70,000 women from Karolinska who
(32:33):
volunteered. From Karolinska.
Okay. 70,000 women,
though, over a couple of years. Is
that correct? I think about
the effort
involved.
Yeah. This is the nice thing about
doing science. In Sweden -
I originally come from
Germany - it's a different
(32:54):
system. But in Sweden,
maybe because of the Nobel Prize, maybe
because of the public interest
in science, there's much easier
engagement. And in
women, I think, in other countries, have
this regular sort of health checkup. So
there's this mammography screening program,
and then you get basically asked, "Do
(33:14):
you want to participate?" And then Per Hall
and his colleagues do the
magic and keep people
engaged, and people
follow, which
is super. And
then you mentioned briefly
the power of replicating these
results, because that,
(33:35):
I think, is an important dimension of this
paper, in that it wasn't just a
single finding in a particular
population that you were able
to find. We're actually able then to go
back to was it? Finngen and the UK
biobank? And then look
at the genotypes, look at the
protein levels, and then being able to
actually show, yes, this connection holds
up. I think that's pretty
(33:56):
significant.
Could you comment on that?
Again,
this work was spearheaded by Asa Hedman,
but again, what I see is
you have, let's say you create this
currency, let's say the
pQTLs. This is a currency you can
go and you can pay in other
(34:17):
countries or in other
biobanks. You can use that currency to
exchange information. And
this is, I guess, what genetics has
really enabled us to do. And now
proteomics is learning how it can do
it. We have different
technologies, they may have different
outcomes, different information. But again,
you can anchor it on the genetics. You can
use the pQTLS, you can
(34:39):
use them as instruments in mendelian
randomization to exchange this information.
And that's
amazing. Yeah. Here it is. You're
talking about empowering proteomics with
genomics, right?
Turning it around instead of coming at, I
mean, I come from a genomics background. So
I think of it in terms of proteomics adding
to the genomics. Here it is. You come from
the proteomics background, and it's the
(35:00):
genetics that is really enriching
the findings. And I think that's
great. Jochen, I have a basic
question, and it's very basic. We
mentioned that one of the factors
could be lifestyle, like
environment, but also could be the hormones.
Do you have, let's say, relatives,
like mother, sister, or twins
(35:20):
that you can control in this cohort? I
imagine there will be also some twins that
you can, let's say, somehow
discriminate and identify the genetic
background versus the environment. This
is the first part of my
question. The second part is for sure,
you would check, like, post- and
pre- menopausal. Then have you seen
differences? What's
(35:41):
your feedback on that? What's your
experience around this type of
observations?
Yeah, I mean, we have had
previous studies that we published using
other own technologies that we
used 510 years ago, where we
specifically looked at hormonal replacement
therapy as one of the factors which,
to our surprise, had really a long
(36:02):
lasting effect on the women's
proteome, which
again, is really
quite significant.
And then, of course, that pre- and
post-menopausal breast cancer.
Again, this is something I've been
learning from my colleagues that I work
with, is very different.
Then, of course, you need to
disentangle. So,
(36:23):
I'm sorry, I want to click
back to what you just said about
hormone replacement therapy having a lasting
effect on the proteome.
What do you mean by that?
Meaning that it shifts
the proteome? But what about the
risk, the cancer risk, right.
Because certainly the women's health
study here in the U.S.
(36:44):
had led to some concerns around
that. I'm just curious
if that's part
of the impact on the proteome,
do you think? Or maybe what we
found in this other study is that we
had a subset of women that really
we could sort of see that previous
use of hormones had
(37:04):
a significant change
in their proteome and also increased their
future risk that they were developing breast
cancer. So, of course, this really
sort of showed up. But it's a
small subset of all the women
that we tested.
It's a great thing to follow up.
(37:25):
Of course, we still need to understand is
what is the effect of taking hormones?
Do you actually have remodeling of
some
reproductive pathways that
constantly do something?
And if they get sort of, let's say, pushed
off track, they will stay on that off
track path for a longer period of time. And
(37:45):
then there will be feedback loops with,
let's say, the liver and other organs to
just try to adapt
with the sort of external trigger.
So that was sort of part
of our sort of understanding of
the use of drugs. But again, it
showed that taking medication has
a quite substantial effect on
your proteome. And we
(38:06):
found it fascinating that it actually seemed
to be consistent
over many years, a
picture of real time biology, the
proteome, right.
And, what we
know about effects of certain
treatments, what we know about effects of
certain drugs, we're just scratching the
surface. Right. A number of our pharma
(38:27):
partners and customers of Olink
are finding out so much with just
a limited set of proteins.
They're not looking at the proteome, they
might be looking at a panel
of 50 or 90 [proteins], or what
have you. But there is just so much to
learn about the biology.
Sarantis, you started to ask, I think you
were down a path of a couple of questions. I
(38:48):
was curious.
I want
to ask just a more
philosophical question about
if you have some, let's
say, mother,
sisters, some relatives, or if
you have some twins that you can follow.
And you can see the change comes from
genetic background, or comes from the
proteomics background, or combination, or
(39:10):
neither. Do you have any experience on
that? Have you seen some patterns
around [that]?
I don't think we
necessarily looked into this.
But I've been working with
another twin cohort from Sweden called Twin
Gene, which the name says
has a quite clear
focus on these aspects.
(39:31):
And
no,
not that I think in particular, but of
course, it's, again, what we
pass on to our children is something
that will be, in the future,
helpful for them to know.
And maybe
they will change their lives when they know.
(39:51):
Okay. I'm at a higher risk of a certain
disease because both my parents
passed away. I guess you see a lot of these
breast cancer studies and effects in
Iceland, I think, right.
But
not in these studies. I cannot
recall that we actually specifically looked
into this. Yeah. What was I think really
interesting about this particular paper
(40:12):
on breast cancer is that you looked
into so many different kinds
of connections
in terms of inherited risk
as well because the title
is paper, "Evaluation of
Circulating Plasma Proteins in Breast
Cancer and Mendelian Randomization
Analysis," you're actually
looking at, then, the entire genetic
(40:32):
backgrounds of unrelated
individuals and just saying what
is elevating that particular
risk. And understand, these five
proteins that were differentially
regulated were basically
lifetime
exposures. That a person
was exposed to a high level of
protein throughout their whole life. And I
think that's what makes this really
(40:54):
fascinating, right? The
proteomics being informed by the
genetics controlling the levels of
protein, and then saying these
five proteins actually become drug
targets, which I thought was
just a fascinating realm.
Before we wrap up, Jochen, would you
like to make any final comments?
Either, I don't know, about
(41:14):
where we are,
where we're going, working with
Olink? Oh, I understand, right? We
didn't even talk about a very
famous
postdoc, famous at Olink,
Philippa [Pettingill] came out of your lab. I don't know
if you want to talk about what it was like
working with her. She has helped
me, Cindy, with her title. She's
Director [of Application Sciences]. She's a superstar.
(41:36):
She runs the field application
scientist team within
the European
region, and she
is absolutely magnificent.
She's also helped lead
our discussions around
statistical analyses in the
UK Biobank Project. She's
just such a magical
human being to
(41:58):
have at Olink. We're so lucky to
have her. And she is a product
launched
out of your lab. At
some point,
you had an impact on
her trajectory. So, yes,
please, anything you have to say about her
would be greatly appreciated. We had
(42:18):
hoped we'd be able to have her on, but
we weren't able to get her
into the timing that we had
going. No, I
mean, all the success that
she has now is because
of her
engagement, her knowledge, and her
curiosity. But, yeah, it was fantastic
to work with her. She was with me about
(42:39):
one and a half, two years. It
was
inspirational and fun
from the first to the last day.
And I think
to see someone leaving the lab
and making such a
wonderful career
is fantastic. I guess if my
contribution is that I showed her all these
(43:00):
different tools that we had in lab,
including Olink and others, and we talked a
lot about the different assays, the
different concepts. So if that has
helped her in achieving these
fantastic things that she's doing with
you, it makes me proud and
happy. I think she deserves it.
And I wish, of course, her
all the success, and
(43:20):
anytime we see her, we see
each other on the media calls,
it's like old friends.
I think she came out
absolutely a leader, and I think
she has such great things to say
about the time that she spent in your
lab. And I think that's
(43:40):
pretty sweet. Thank you.
That's great to hear. All right,
well, thank you very much for joining us
today, Jochen. We've really enjoyed the
conversation. Thank you for having me.
It was fantastic. And
continue with this great podcast.
It's really a treasure. Thanks a lot
for setting this up and running
it. You're so
kind. Thank you.
(44:02):
Okay, well, I think
that's it. Thank you.
Thank you for
listening to the Proteomics in Proximity
podcast brought to you by Olink
Proteomics. To contact the hosts
or for further information, simply
email info@olink.com.
Popular Podcasts
24/7 News: The Latest
The latest news in 4 minutes updated every hour, every day.
Therapy Gecko
An unlicensed lizard psychologist travels the universe talking to strangers about absolutely nothing. TO CALL THE GECKO: follow me on https://www.twitch.tv/lyleforever to get a notification for when I am taking calls. I am usually live Mondays, Wednesdays, and Fridays but lately a lot of other times too. I am a gecko.
The Joe Rogan Experience
The official podcast of comedian Joe Rogan.