0:15
Thank you very much for the introduction.
0:17
Yes, I'm here today on behalf of BIOGNOSYS to talk you through one of the platforms that we have.
0:24
So first of all, to give you a bit of a technical overview if you like, and a story as to how we develop this platform, how we've made it more accessible to industry when it comes to smaller inputs, especially on tissue biopsies.
0:40
And then to go a bit more into, you know, how people have used this, so a few of our collaborators and into a bit more data as well.
0:49
But just to give you a bit of an overview for those of us, for those of you who don't know us, I do see some familiar faces here, but a few new ones as well.
0:57
So BIOGNOSYS is a proteomics CRO, so whole proteomics as well as immunopeptidomics, a company based out of Zurich in Switzerland.
1:06
So we spun out of ETH Zurich just over 15 years ago now very much as a mass spec software company, but slowly developed into more of a contract research organisation.
1:18
And today the main part of the business is very much the proteomics CRO business in terms of where else we work and who we work with.
1:28
So we of course have the HQ in Switzerland, but we also have a new lab that's just opened in Newton in Massachusetts, which is looking into processing the high throughput plasma samples.
1:40
Also some cell line work as well.
1:43
But the immunopeptidomic side of things is very much still a focus.
1:49
In terms of our reach, so we do work with both large pharma and biotech, that's the main part of our client base.
1:53
But we also do work through a number of collaborators in academia too.
1:58
We have 3000 publications now using our various proteomic workflows and we have one of the largest mass spec facilities in the world.
2:09
And a bit of news as of recently.
2:10
So last year we joined the Bruker family, which has given us access to newer technologies, a lot of high throughput proteomics as well.
2:21
But moving on from the history lesson and the background there for you.
2:25
So immunopeptidome profiling.
2:27
So this was a venture that BIOGNOSYS looked into roughly 3 years ago now and we were looking into how we can optimise a workflow for HLA peptidomics.
2:36
So both Class 1 and Class 2 and through various sort of market research we did and sort of speaking for academic collaborators, one of the big limitations in immunopeptidomics is the amount of material you need to extract the HLA receptors and the peptides associated with them.
2:52
And when it comes to especially clinical samples, we saw that people were often having to start with a gram of tissue and that's something that just generally isn't feasible.
3:01
So a real sort of push for us was to look at how can we make a very sensitive assay in both discovery immunopeptidomics, also target them as well and how can we reduce the inputs that are needed.
3:15
So we went through several rounds of optimization over several years and there are multiple points that we look to optimise the workflow.
3:24
So first of all, the enrichment for the HLA receptors themselves.
3:27
So there's a number of commercially available HLA antibodies for enrichment, some that are specific certain alleles, A2, A3, A11, A24 being some of the predominant ones.
3:39
But what we found is that we got the biggest, we saw them the highest number of peptides when we used pan HLA antibodies from class one.
3:49
So that was one round of optimisation that we looked into.
3:52
And then of course, being mass spec experts, mass spec the other side of things and by running very long gradients by using an acquisition method known as data independent acquisition or DIA for short.
4:03
For the mass spec people in the audience, we now see in excess of 10,000 peptides when we run discovery studies.
4:10
And on top of that, we also looked at developing a targeted platform more recently, which uses something called parallel reaction monitoring.
4:17
So overall, what we have is a very optimised workflow, one that can work with very minimal inputs as well.
4:25
So you can see in the tissue, the realm of tissue, it's roughly 10 to 15 milligrams can be more depending exactly on the question we want to answer.
4:33
But for general discovery studies, this number works well.
4:36
And then cell wise, we're looking at roughly 25 million.
4:40
And maybe just a final comment just at the back end.
4:43
So as I mentioned earlier, we were sort of born out of mass spec software, what we also use as our own software.
4:49
So Spectronaut and SpectroMine, which again is has increased the amount of the immunopeptidome that we cover in these kinds of studies.
4:59
But just to add a bit more as to where do we get these sample inputs?
5:03
Have we tested other inputs?
5:05
Where do we see some form of plateau?
5:06
Where do we see a good coverage?
5:08
As you can see here for HLA class one and two, we did a number of rampant studies.
5:14
So if you look at the cell line data on the top of that slide there, you can see for Class 1 peptidomics that roughly the sort of 20 - 25 million cell mark.
5:23
We start to see a bit of a plateau where we see roughly 11,000 peptide IDs and we don't see much of a gain when we get to 50 million cells.
5:32
So that's why internally, we work with 20 – 25 million cells and we don't see much gain for class one, if you like.
5:44
And the same can be said for Class 2.
5:46
Again, 25 million seems to be the optimum and a fairly minimal amount compared to the rest of the field.
5:54
Moving on to tissue, so very much limited when it comes to inputs that we get here.
5:59
So we can see when you get to the 10 to 15 milligram mark that we again see a plateau roughly at 11,000 IDs increasing from this up to sort of 45 even up to over 100 milligrams of tissue in DIA.
6:13
So discovery studies we don't see much more of the more peptidome. when it comes to more targeted studies, there is an argument that the lower abundant non conical targets, maybe we should look at higher inputs and increase the concentration of lower abundant targets.
6:27
But for general discovery studies, this sort of 15mg mark seems to work very well.
6:32
Last one and maybe as a final comment, you can see on the bar chart just down here.
6:42
So actually in Class 2, the more tissue we put in, the more we see.
6:47
But as I'm focusing more on the Class 1 side of things today, I will focus on that.
6:52
But if anyone has any questions on more Class 2 analysis, happy to discuss that later on.
6:59
And then a final comment on sort of what we've done with this workflow, the matrices that we've looked at.
7:04
So once we launched the platform, we saw a big increase for a big interest in terms of sort of discovery preclinical studies for immunopeptidomics.
7:14
But something that started to come out the woodwork was having to use this in the clinic when there isn't tissue available, there's like biopsies and something that is often used as a surrogate matrix is PBMCs.
7:26
And what we wanted to look at doing is see how many peptides can we see from clinical samples, specifically PBMCs.
7:34
And as you can see, the coverage is reduced there, but it's a pretty reasonable coverage still of the HLA one immunopeptidome and something that can be used as a readout.
7:46
I would say more when people are looking to modulate the overall immunopeptidome rather than looking at quantifying specific neoepitopes.
7:54
But either way, we see that immunopeptidomics is sort of moving into clinical samples and into the clinic and it's something that we continue to work on at this stage, but moving into sort of a bit of context, if you like, as to what we do.
8:10
So of course, we're a mass spec CRO.
8:12
We're very much focused on the technical.
8:14
So what we rely on is to work with collaborators who can give us a bit more biology and a bit more context to these things.
8:22
So as a first case study, we work very closely with Indivumed therapeutics in Germany and they were actually very helpful in terms of developing this immunopeptidomics workflow.
8:33
And what we look to do with them in this initial study is just a general discovery unbiased profiling of the Class 1 and then the peptidome.
8:42
So as you can see, we took tissue samples we had with cancerous lung tissue samples, worked with 15 mg of fresh tissue.
8:52
And what you can see as outlined in previous slides as well is a really comprehensive coverage for the peptidome.
9:00
So just because I can see I only have 5 minutes left, I won't dwell on this too much, but the general take home message here is you get a very good coverage.
9:17
Finally to move forward into to more of the targeting the immunopeptidomic space. So we have collaborators in Oxford.
9:20
So some of you may know Grey Wolf Therapeutics.
9:23
So Grey Wolf, we're focused on ERAP biology, specifically pivoting ERAP 1 and ERAP 2 enzymes and the outcome of this inhibition leads to a complete modulation of the Class 1 immunopeptidome.
9:37
And specifically when inhibiting ERAP 1, you see an increase in longer Class 1 peptides being presented.
9:51
And what this leads to is some novel responses, but in terms of what we did, so Grey Wolf provided us with a number of epitopes that are presented upon a ERAP 1 inhibition.
10:05
So they selected 30 new antigens that they wanted to quantify and we then looked to develop a targeted assay to work out a rough copy number per cell in these peptide targets.
10:15
And in the initial experiment that we ran, so we took several different cancer cell lines, so from colorectal, ovarian, Melanoma and lung and we ran this panel of 30 peptides to look and see how these copy numbers per cells varied with an ERAP1 inhibitor and without one of them.
10:35
And what you can see quite clearly, so I've only picked out two of the cell lines here.
10:41
But in the ovarian cancer cell line, if we see that upon ERAP1 inhibition, we see a very big increase in peptide 1 and peptide 4 here.
10:50
It's a very nice clear result when you compare to the negative control and the same here for peptide 6 in the Melanoma cell line.
10:58
So using these targeting assays is a very effective way of estimating copy number per cell and further down the line looking into CRT therapies and various different other modalities in the lab.
11:12
But in summary, so we've developed a very optimised HLA one profiling workflow that can be used both in the targeted sense and also an unbiased discovery sense as well.
11:23
We generally search only against the canonical immunopeptidome for now, but we can supplement our searches for the more non canonical sequences as well in those earlier stage stages, which often comes from strand transcriptomic data and quantification of specific peptides can be used to estimate copy numbers per cell, particularly in the TCR-T space.
11:47
Thank you very much for listening and happy to take any questions.
