Subscribe
Share
Share
Embed
A review of a recent publication by Loroit and Italiano et al in Annals of Oncology, identifying a novel predictive biomarker of immune-checkpoint blockade resistance.
Proteomics in Proximity discusses 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 unlock biological mechanisms. Co-hosted by Olink's Dale Yuzuki, Cindy Lawley and Sarantis Chlamydas.
Welcome to the
proteomics in proximity podcast, where
your co hosts, Dale Yuzuki,
Cindy Lawley, and Sarantis Chlamydis 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 UN,
lock biological
mechanisms. Here we have your hosts,
Dale, Cindy, and
Cerantis. Here we are with another
episode of Proteomics in
Proximity. I'm your host, Dale Yuzuki, with
my co-host Cindy
Lawley. And Sarantis, tell me this
and Sarantis. And we are here
today to talk about cancer,
and specifically cancer
immunotherapy, and looking at new
biomarkers for
predicting immunotherapy response.
We are looking at an article, uh, published
in 2021, one called
"plasma proteomics identifies
leukemia leukemia inhibiting
factor LIF as a
novel predictive biomarker of
immune checkpoint blockade
resistance." Now, this particular
topic that is a, uh, mouthful. It
is a mouthful. Okay. The
first law author is Loriot. Like I
mentioned, it was in, The Annals
of Oncology, and it was published in
interesting, right, is we're talking
about immune checkpoint blockade
resistance. So these are PD-1
and PD-L1,
uh, therapies that have really
revolutionized cancer
treatment. Specifically, I mean,
backing up a bit immune checkpoint,
uh, blockade, uh,
inhibitors, right?
These particular
ICB's,
so-called, is credited with saving the
life of Jimmy Carter, right? He
came down with very serious
melanoma. He's still
alive. I think he's, what,
there, but his life has been
saved. And this particular paper
is explaining new
predictors of ICB
resistance because it's not
equally effective in other types of cancers.
Go ahead. Uh,
sorry, I just wanted to add a little bit in
there. So just stepping back. One more step
back is tumors tend to
evade the immune system, right?
So they tend to grow,
ahh, often unchecked.
And, uh, one of the
mechanisms for that hiding or
that evasion of the immune system
is this PD-1,
and it's ligand,
PD-L1, uh, that
connection. And so these
checkpoint inhibitors cut that
connection and allow
them the tumor in a
subset of patients, not majority of patients
that are treated. It's actually
only, what is it? 35%
effective. Something around there
in a subset of patients, jimmy Carter
included. Uh, this was actually
an effective way to open
up the immune system, to see the
tumor and then infiltrate the
tumor. And so some of the ways of
evaluating how these ICB therapies are
working is by looking at infiltration of
the immune system into a tumor.
Thank you. Yeah. And for
melanoma, it's extremely
effective. It's for the other
cancers where it's not very effective.
Solid tumor, solid tumors.
Very good. Sarantis, you
were just going to ask. No,
just to mention that
it's a therapy based on antibodies, right,
that targets, uh, either PD-1,
PD-1 in conjunction
with other chemotherapies, where, um, the
success rate is getting increased. But
still, as you mentioned Cindy, 35% of the
patient respond to the treatment.
And, uh, I think it's really important
that new biomarkers or, uh,
new, uh, molecules that are
druggable and they can help increase, uh,
the rate and the outcome of
the eventual outcome of the patient
goes to this
paper as well. Right. And as far as
predicting, uh,
ICB, uh, responsiveness, uh,
I
was involved in ctDNA analysis
for a couple of years, right, from 2015 to
all the rage was looking at tumor mutational
burden. That is looking at right.
TMB, TMB, yeah, TMB, TMB
TMB. Looking at whole exomes,
looking at tissue
biopsies, looking at
circulating tumor DNA is what ctDNA
is and trying to get this
idea of, right, okay, if
there's a high mutational load of the
particular cancer, this person is
more likely to respond to
immune checkpoint inhibitor treatment.
And they tried to normalize
different enrichment methods, they tried to
get all this work around
it, and it really
wasn't all that predictive. I mean, the R
values were about 0.7.
And then a, uh,
multiplex immunohistochemistry was all
the rage. I saw that
wave. Another what we call
companion diagnostics. So these
are tools that
physicians are trying to use in order
to guide an understanding of
who might be better responsive
to such therapies. Yeah, please. Go ahead,
Dale. And so companies like Akoya
and other spatial right
imaging, I think Millipore also had
one hundred-plex kind of thing.
Anyway, there are all these different
companies doing multiple, like, uh,
particular antibodies and
immunostating. Uh, that's what
immunohistochemistry is. You're staining
with antibody and detecting it with
fluorescence. And they're trying to get that
and they're able to up the
predictive power just
incrementally. So that is why the
hunt for biomarkers is
still on. And how do I
predict?
I'm sorry, another point. Also
here, there's a lot of immunohistochemisry is
based on biopsies, right, and quite
invasive. And then I think
the most important part here is in this
paper, they see the value of having the
plasma protein that's less invasive and
can, they can support all the findings
of things, uh, that they can see within
the cell, for example. And the question is,
right, even with the biopsy samples,
even with the sort of invasive methods, it's
still not predictive enough, right.
Where they call it imperfect
predictors. Right. They're still
seeking better
predictors because still, even with those
guided tools,
majority are not. So if we're going to give
the right treatment to the right patient at
the right time, we have to have better tools
for that. And, and can I just say that this
paper is so elegantly
written, I highly recommend it. I just
think these authors, I don't know if you've
seen them speak, both Loriot, as well as,
um, Antoine, uh,
ah, Italiano. Amazing
speakers, both on YouTube, highly recommend.
Yeah, that's Italiano.
And Loriot they're at the
Gustav Roussey is that in Paris?
or it's multiple instances, it's their
cancer center. Grand
Paris I think is where the clinical trials
for these cohorts, the validation and
discovery cohorts are based. They're still
recruiting. These are phenomenally
cutting edge. So looking
at tumor mutational burden, right? From a
genomics perspective, you're looking at IHC
from a tissue histology-based
perspective. There's even effort
around microstaellite instability
to go back to the genomics tool, another
genetic tool, right? Looking at it again,
still not good enough.
And Sarantis, you mentioned, right, the
power of circulating biomarkers,
you might say. Well, haven't been people
been looking at this before?
I think. Yes, there
was all the right tools, actually, right? I
mean, I think here is something to say
that's impressive about Olink technology, right?
Because uh, gives this high throughput thing
and before this was Mass Spec, it was a
little bit more difficult. It's not
so much out throughput. I think
it's more easy nowadays with having this NGS
based technology and things like that,
right?
And uh, it's a cytokine, right? So cytokines
are these low abundant,
(absolutely) proteins, right? So very low
abundance and not easily seen with mass spec
without a large amount of sample or
subtraction methods. And so having the Olink
PEA that hooks this proximity
extension assay, that hooks these low
abundant proteins out of solution and
helps us count their abundance,
uh, is I think,
just a great use
case. And again, talking
about how you
close the loops, right, and you go a little
bit more to the molecular you have the LIF
protein, but then brings uh, together
drug targets like STAT3, for
example, right? I mean you open up
completely a pathway of uh,
target proteins that could be druggable
but also could be potential novel
biomarkers. It's not like one protein,
there's a completely (new) pathway. And I
think tools in the future like Olink
Insight that we introduce it will help
a lot of people understand pathways and
how other actually target molecules
right. Can help and enhance drug
development processes. And biomarker
identification and ah, Sarantis,
you mentioned Olink Insight. For those
who haven't heard the podcast with Hilda
Anderson talking about Olink Insight, I
refer to that. I think it's
episode number five.
At any rate,
Sarantis, there's a question that I have,
right, from, again looking
at soluble
biomarkers in plasma.
Uh, the question I have is,
is this signal sufficient? I mean,
somebody may have advanced tumor
status. I mean this particular paper
used advanced
NSCLC, uh, as
their sort of test
case, uh, is there enough
signal there? I mean, somebody has a tumor
in a particular area of their body,
but we're looking at the entire sort of
plasma. I mean, the blood volume of
an individual is on the order of what, a
couple of liters, right? Maybe not
true. Yeah, that's true. It's always difficult
to force
here. They have much samples, right? We try
biopsies much samples we have
plasma samples and they try to see
correlations. And then they did, as Cindy
mentioned, a really elegant way to check
with the assay to see, for example, B
cells or, uh, differentiate
adult T cells, for example. I think
it's, uh, looking in the omic's
perspective gives more power to this
respect. That'S
my and the imaging as well.
Right? I think,
Sarantis, you said the
Insight tool allows
you to see multiple pathways and
interactions among pathways. And we expect
most diseases we've seen are polygenic. And
there's going to be multiple pathways that
are involved in probably a signature of
maybe 20, 20 or so proteins that
we'll need to understand to really
use it for, say, ah, a companion
diagnostic in the future if we move through
clinical utility in this case, It's
rather remarkable that it's one
cytokine that has that
signal where low versus
high were predictive
of, um, progression, uh,
free survival. Did I say that right?
PFS, yeah. And overall
survival. And uh, that's kind of
remarkable. But I also want to point out
the tertiary lymphoid structure
correlation. So we've got
this imaging data from these biopsies. As
you say, it's very invasive. Right. And
that you can actually see and
count these lymphoid structures, which are
just sort of localized,
um, immune
system, adaptive, uh,
immune, uh, clusters that
are responding to the tumor.
And that also was a correlate
to that same cytokine, which
I find that to be pretty
remarkable. Absolutely. That's a great
point Cindy. So, Sarantis,
I just want you to take us from the top in
terms of talking about the experimental
design. Yeah, I
mean, they try a cohort of
patients actually treated with
immune system checkpoint point blockade, uh, anti
PDL-1. And
uh, they use the Olink
Explorer 1536,
like more than 1500
proteins. And uh,
among this, was there like, our,
uh, cohort, their
discovery cohort, and then there was a
validation cohort with Flamation panel
with more than 290, uh,
patient 96
proteins, where they see
actually really nice correlation with data
that they can see with the Explore, uh,
panel. That's also really nice
because our Target and Explore
panel approach, they go hand by
hand, and that's really great and really
nice to see how we reproduce our assays,
even though from high throughput to low
throughput. Right. And that's
really great and nice
finding. And uh,
uh, one of the first biomarkers that
popping up to their data
set was LIF that, uh,
high levels of LIF in
this plasma was correlated
with the poor clinical outcome of the
patient with the immune blockade.
That was the first and most
remarkable, uh,
actual data. Yeah, it's a
really good point, Sarantis. Like, they call
it discovery cohort and validation cohort.
But not only is it a discovery
cohort where they discovered the signal and
then validated it, but they also used our
discovery tool. Right. The Explore
at that time, this was one of the first
Explore publications and then they
moved it to a lower plex where they could
get the same markers and of
course, that's going to be a lower cost. So
they were able to do more samples in that
validation cohort, which I like, I hadn't
thought about the fact that it
parallels the discovery and
validation also parallels the intent,
to some extent of how
groups are using our products. Sometimes
moving to a Target 96, but
sometimes even moving to a
Flex panel or a lower Plex
panel. Um, yeah. Anyway,
I like that characterization that
you made. And if I understand correctly,
the discovery and validation
cohorts from two independent
clinical trials, is that correct?
That's right. Exactly.
Both
still recruiting. Right. So still, I would
expect more to. Come, uh, here and different
type of cancer, right? Different type
of solid cancer for
bladder to non-small cell lung
cancer. They have different
type of prostate, I think. Also they have
different cancer
types that's like, more
universal. Looked like more universal
biomarkers, actually. Yeah,
I missed that. That's an interesting
aspect, too. Yeah.
M good, um, catch.
Yeah.
Both out of Gustav Roussey,
as Dale mentioned earlier,
this
was the, uh, MATCH-R
cohort. So I have it written down. I'll
actually read it off, uh, prospective
trial to study the evolution of clonal
architecture of tumors from patients
treated with molecular targeted agents.
So that's the MATCH-R, that was the
discovery cohort where they used the
Explore. And, um, I think you mentioned
Cerantis 90 some odd
samples, uh, or
patients in that were used in that.
Um, and then the
premise, the validation cohort
where they used the Target or qPCR
readout it was the predictive markers of
immune related adverse events
in patients treated with immune
stimulatory drugs. So, again, that's the
PREMISE trial. So yeah,
a lot of great stuff coming
out of this group. So then as far
as, uh, Cindy, you mentioned
this idea of a
tumor lymphoid or
tertiary lymphoid structure,
right? That's right. And
this is then a feature of
good response, is that correct?
Yeah. Uh,
I believe the relationship was low
LIF with tertiary
lymphoid structures.
Meaning, my understanding,
uh, and Sarantis, please check me on
this. My understanding is when you
have, um, these sort of
ectopic lymphoid organs that
develop in these non lymphoid tissues, that
they're a sign of
immune infiltration from adaptive
immunity. No. Okay.
So, uh, that's
telling us that there's some element of
the tumor that's visible to the immune
system. And that's like the
Holy Grail, right? The gold.
Yeah. You can manage to not turn
on autoimmunity in some issue.
Right. You don't want the immune system to
go haywire and get inflammation out
of control. That's what kills a lot of
COVID patients. Right. You
want to keep that inflammation in check, but
you want to open up the
tumor to the immune system. So you have
a very targeted kill,
as opposed to sort of traditional
chemotherapy agents that would
take us to death's door
because it killed, um, non
specifically would kill off all fast
growing cells. Right. Which is why people
lost hair, lost nails. These are fast
growing cells in the body, and you just
carry it. Right. I'm oversimplifying,
obviously. But this is the
promise of the future of,
uh, cancer treatments. Yes.
And it's like these
tertiary lymphoid structures are
a sort of in vivo
validation. Right. The
marker
is measuring what we want to
measure, which is this person
responding or a person not going to respond.
Right. Yeah,
exactly. Uh, without having to go in and
do a biopsy to see these tertiary
lymphoid structures. Right.
As you said earlier, Sarantis, if you
can evaluate it in a blood
draw, that's
exciting. Yeah. I mean, you think about how
sick these patients are. Right. These were
advanced NSCLCs was a
majority of the samples. I think bladder
cancer was a minority of the
original 95 samples. And I think
there had other tumor types in there
to represent the discovery
phase. Uh, Sarantis,
can you mention oh, go ahead, Cindy, go
ahead. Yeah, you just
reminded me, um, that in the results,
they actually characterized the difference
in progression free survival and overall
survival. And we're talking months. Right.
These are people that are maybe living
two months versus 21
months. It's a great point.
Ah, and I think even by the time they did
analysis and of course, you
have to qualify for clinical trials, so you
have to have already had,
um, a treatment fail. Right.
Uh, there's several checkpoints I think they
have through their cancer center that
Antoine talks about. But we're
talking about,
um, people that are very
sick and at very high
risk of
dying. Uh, and
this gives them can sometimes
maybe, um, give them
five to ten times more time,
um, than they would have had
otherwise, and hopefully with a less
invasive, um, therapy.
Although I don't know as much about the side
effects of this. Cindy, you bring
up a good point in that these were
advanced, by advanced, meaning they
were metastatic. Right. They already
had stage four,
or stage three, really high
stage lung, uh,
cancer, and had already spread to other
organs. And I think I
was thinking about this this morning, where
at the 50% overall
survival curve, it was maybe
five months. And then,
uh, with the LIF, uh, low,
that's five times, that
over about 20 months.
And so I was asking my boys, taking them to
school, how much is 15 months
worth? Right. Meaning, ah,
from five month point to 15
months more. That's
the difference. And you think when every
month counts, I mean, these people
have right. Advanced,
uh, uh, basically fatal lung
cancer. But with this
treatment, you're given what,
a year, over a year,
how much is a year of your life worth?
Yeah. And it's such a personal
choice, right. Because it also depends
upon what the quality of life
is in that year. Right. Which is
something that's really
yes. As we get older, we
think more and more about these things.
Right. I think it's a nice way to monitor
right. In the way that you can see the
response and you can have something
to monitor. And I think that's
a useful tool if you would like to combine
other therapies in the future. Right.
Because, you know, when you monitor how
people, they respond, having something
that is less invasive, more easy to
handle, in a way more easy to understand,
and, uh, I think it's a great
tool, and it's a great finding. It opens
new ways of seeing science in the future.
Right. Yeah. Interesting. Good point. Yeah.
Because if it's not working, you get off it.
Right. And you just get that quality
of life back. Perhaps. But
yeah, it just depends upon what the side
effects of any given therapy are. And
certainly the ones today are much
better, uh, overall
than the ones 20 years ago.
Now, since a high
level of this LIF, (leukemia)
inhibitory factor right.
Leads, uh, to poor outcome
did I read this correctly?
It's actually now considered a drug
target?
Yeah. Go ahead, Cindy. Please. Go
ahead. Yeah, I actually don't know.
Please, Sarantis. I think they
have some thinking of I don't know
if it's, um, somewhere in the discussion the
measure that they're thinking, like, for
antibody against exactly.
Against monoclonal antibodies
against Liver. It would be a target, I'm
guessing. Yes. Because I think it's like,
uh, cytokine that is like, uh,
uh, differentiation regulates
differentiation.
It depends on if
it's causal or if it's a
thing, if it's the
results of this, right. So if
it's a useful thing to monitor to
know how
someone's responding, then we wouldn't want
to drug it. But I think that's interesting.
Right? Yeah. Uh, you would imagine the
knockout mice are in,
right? Yeah,
that's right. They're in generation II.
They're the F-1. Got the F-1 cross going
on right now.
Jackson Labs is probably yeah, there you
go. Over this. Well, there
you go.
Uh uh, did they
give also any clue in terms of what their
next steps would be? I mean, since
this is a single marker right.
Pretty straightforward to implement,
I
think. Looks like, since the
numbers they're still small, I think.
But Cindy, correct me if I'm wrong there. I
think it needs more bigger numbers
right. And bigger cohorts and different
clinical trials. I don't know. Cindy,
correct me if wrong. You are you are the
expert of the of the big samples and
big cohorts. Yeah. I mean, what was it,
And 292 in
the sort of validation,
uh, in cancer,
naturally, the numbers tend to be
relatively small.
Right. In terms of they're not in the
thousands. They're in the hundreds. Cindy,
do you think that
these numbers need to be
larger? Oh, for sure.
Absolutely. Yeah. No doubt.
No doubt at all.
Uh,
but I think
the burden of
this and this is a question the
burden of use of this
as a companion diagnostic, which
I think is the immediate
proposal, I think further down the
line, establishing causality rather than
consequential that it's
consequential to the state
that, um, there
must be a lower burden for
establishing a biomarker as opposed to
establishing a monoclonal antibody
therapy. And
I don't have that
set out in data, but I think that's
the media or the near term,
uh, exciting promise
here is that it could
be used as a companion diagnostic. And they
actually broke down the use of LIF
high versus low. And
the TPS scores right. Which were the ones
that Dale mentioned. So,
um, the tumor mutation
burden, along with basically trying to
characterize this
ligand expression, uh,
the tumor mutation burden, the
microsatellite instability, and the
immunohisto-chemistry, using those
which are established scores in
conjunction with LIF, showed,
uh, utility. Right. So I thought that
was a logical progression in
their analysis that I thought was
interesting, but also requires larger
sample sizes, because now you're breaking
your numbers into quadrants, and you need
sufficient sample numbers in each of those
quadrants. Right.
Yeah. Interesting.
Well, any other
final comments before we wrap this one up?
Cindy, go
ahead. I'm so sorry. I just am so
excited about this paper. Sarantis, I think
I've stepped on you three times.
Um,
I just highly
recommend that people keep an eye on these
authors, because, like I said, they are
elegant writers.
They speak with
such clarity, and
they have been driving the use of
genomic tools in
demonstrating, um, utility and
cancer for years. Right. Cutting
edge. Right on that
cusp. So sorry. Sarantis, your
turn. Oh, I
think the method, for me, the most important
thing go from invasive to non invasive
methods. Uh,
that's the clever thing on this story. And,
of course, I mean, you can back it up with
multiomics like imaging, get an
assay, but, uh, I think the method is
brilliant. That's pretty much how I
see it. All right, well,
thank you for joining us today.
Till next time. Take care.
Thank you very much. Thank you.
Bye bye. Bye bye.
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
l info@olink.com.