Harnessing Phage Display for Peptide and Antibody Discovery and Development

0:00

We start with Claudia Dall’armi.

0:01

Claudia is the head of the Display Technology at IRBM where she leads a multidisciplinary research team in the Translational Biology and Discovery Research division.

0:13

Claudia has more than 20 years of scientific experience across both academia and industry, and she established herself as a leader in peptides and biologics discovery, particularly in the development of mRNA and phage display platform for the identification of therapeutic peptide and antibodies, which I guess is what we are hearing about today.

0:31

So please, Claudia, good morning everyone.

0:38

I'm really pleased to be here and have the possibility to give you an overview on how to harness display technologies in this case specifically phage display for the discovery of peptide and antibodies.

0:53

So as you probably know, phage display has been around for a while.

0:57

So since its discovery and it was in the 80s.

1:02

So since then I've been generated several kind of libraries just transforming the genome of this virus that you can amplify using bacteria.

1:14

So practically you can display a combinatorial peptide library fused with the proteins that are coated on this virus and with that you can generate a high diversity library with the complexity that can easily reach ten to [the power of] nine but you can reach also ten to [the power of] eleven.

1:36

This technology gives the possibility to challenge the target with different strategies to really reach the goal that you have in mind for the binders of your dream.

1:48

And also it is very easy to use.

1:52

So here in IRBM we have over 20 years of experience in this technology, and we have proprietary libraries for peptides and antibodies. Peptides we have library macrocycle peptide by the formation of a disulfide bond between two cysteines, and the single chain library.

2:10

Our libraries are lower [unclear] fused with pIII and the reason why is because this fusion allows the display of small peptide and single chain but most importantly it's a monovalent display of the peptide.

2:25

Meaning that we are really selecting for high affinity binders.

2:32

So just thinking that this is an old guy, this phage display, but it's still shaping the future because it's not the technologies that you can just challenge the recombinant protein in like for example, mRNA display.

2:45

So usually I don't know how you are acquainted, but you challenge your library on a target.

2:52

And then what you do, you collect the phages that are bound to the target, you amplify them, and you reuse them for the second round of selection, and you go on for 2, 3, 4 rounds until you have identified a pool of phages that have really high affinity for the binders.

3:10

So but you cannot use only recombinant protein with this technology.

3:14

You can use cells expressing the target.

3:16

You can go also on cells that you don't know which is the target, but you want to reach for peptides or antibodies that are really able to engage that specific cell type can be cancer cell or whatever cell.

3:28

You can also go in vivo so you can inject the library in the tail of a rodent, and you can collect phages directly from a tissue.

3:38

So you're really able to identify phages that are displaying some peptide or antibody that can allow the penetration of that specific tissue.

3:48

And you also perform a biopanning on slide tissues.

3:57

So you can really check for a distribution and where the phages are binding on a tissue like using for example, spatial imaging.

4:06

All these things can be used for tailoring the biopanning to the specific target.

4:12

But you can use them also for screening.

4:15

So once you have identified the pool of phages that are binding the target, you can screen those as NGS analysis for example.

4:23

But what I really like to combine with NGS is a single clone screening.

4:28

So you are able to isolate thousands of different phage clones, and you can characterise them at the phage stage with the characteristic that you want for the specific binders.

4:40

So just to give an idea, as you know, I'm in peptides now, the modalities that is rising as a modality and how you identify new therapeutic peptides.

4:52

So you can use phage display, mRNA display.

4:55

But in this case, what you do, once you have tailored the biopanning in the phage screening, you can already go for whatever it is, for example, binding the receptor at the cell membrane and internalise it.

5:09

In the case you want to develop, for example, a PDC or a peptide that is conjugated with an oligonucleotide.

5:17

So at the phage stage, you've already selected and identified that hit with that specific characteristic.

5:24

You can do the same for activity modulation for some receptor.

5:28

And you can also identify either way if you are able to display some protein-protein interaction, for example, ligand binding to receptor.

5:37

So we have two nice works punished on this topic.

5:41

But what I really want to focus today is how you can exploit this technology to identify antibodies.

5:52

So I don't know how many of you know that 15% of the antibodies nowadays in the clinic comes from phage display.

6:02

So maybe it is not the first technology that you think about when you want to develop an antibody, but maybe when you want to develop a therapeutic one it can be a good idea.

6:11

Because usually the libraries are made with a human scaffold, so you don't have to lose time afterwards trying to humanise the antibody.

6:20

And they are pretty stable, they're not prone to aggregation, and they have a low immunogenicity.

6:28

So in IRBM, we are fortunate because we can go from the beginning of the screening up to the lead identification until the preclinical candidate nomination, starting from the phage display.

6:41

So of course, once you perform a screening, you want to identify the hit ID, you have to convert it at a small scale, IgG production, then characterise them with many different orthogonal assays until you reach them in the point where you have identified your lead.

6:59

So we have two different antibody libraries, the first one as the diversity inserted in the human CDR3 and it tells them in a framework that is completely human, the complexity is very high to be a phage library because it's almost ten to [the power of] eleven, and we have an insertion of the fragment of the loops that have different length.

7:23

So usually you challenge this library against the target.

7:26

And then once you have identified the sequences, you can convert them to IgG, characterise them.

7:32

And once you have your lead, if you didn't reach the optimal affinity, you can use the second library that is an affinity modulation library where the variability is inserted in the other CDR.

7:44

So radically cloning the CDR3 that you have identified in the first screening in the second library you can just challenge again in a very good screening this time to identify high affinity antibody.

7:58

So you can increase tenfold at least the affinity of the antibody for the target.

8:05

So today I want to show you a case study that we have done and is one of our assets.

8:10

So there is this target CRLF2 or TSLPR that is a cytokine-like receptor that is normally expressed on B cells and is forming heterodimer with L7 receptor, and is implicated in the battery for JAK/STAT and mTOR activation.

8:29

But unfortunately it is rearranged in a big subpopulation of acute lymphoblastic leukaemia.

8:35

So it is over 50% of the patient that this rearrangement in the Ph-like B-ALL, and those patients have poor prognosis, high rate of relapse and no effective therapies beside the standard of care.

8:51

So that is really an unmet need for this kind of patient.

8:54

And we wanted to develop an antibody as a therapeutic for this kind of patient.

9:01

So we started from the phage display.

9:03

So we run a campaign against the recombinant protein and terminating the cells.

9:09

So we are able to find sequences able to recognise both the conditions of the protein we then optimised [unclear] them for binding.

9:17

Of course, like the single phages, but also for the ability to bind and internalise with the receptor, because that can give you already an idea which is the way how you can exploit the antibodies that you're finding as a therapeutic.

9:31

It could be an ADC or a CAR T. So, and then I mean, once we have identified the best hits were like 50 IgGs, they've been produced, converted, produced and characterised at a small scale.

9:46

So we verify the affinity epitope binding, and we verify that actually whatever was happening at the phage stage was happening also at the as an antibody.

9:58

So some of them were able to bind and internalise.

10:02

Some of them they were stuck on the membrane so they could be used for different therapeutic approach.

10:07

So the two best candidates ranked for affinity have been further optimised by the affinity maturation library reaching one digit nanomolar affinity for the candidate that is able to internalise and the picomolar affinity for the two candidates that we are now developing as a CAR T.

10:29

So what I'm going to talk about is the first part.

10:32

So the development of an ADC for this therapy.

10:35

So we took advantage of a payload that is well known and accepted in clinic that is tesirine used for Zynlonta and that's a proof of concept.

10:46

We went ahead with our study and here is just a summary, where you can see that today this is nicely internalising and also has a very nice efficacy in vitro, for a cytotoxicity.

11:00

So we test it in different cell lines and cells expressing different level of the target.

11:07

So then we moved toward an in vivo system.

11:12

So we just sub Q and inoculated HELA cells expressing the target at high level and with just one injection of ADC at 0.7 MPK.

11:23

After two weeks of treatment we have completed tumour regression.

11:27

So the ADC was working very well.

11:31

compared of course to the control and the negative. But we wanted to move further.

11:39

So internally still we have developed an orthotopic xenograft taking advantage of the Mutz-5 that are cells derived from patients.

11:49

So because they didn't have luciferase, we didn't modify them, so we could monitor the engraftment of those cells by micro bleeding.

11:58

So we collected the blood every week until we were able to count the number of Mutz-5s circulating in the blood.

12:07

So that give us an idea how the engraftment was going well.

12:11

So once we are able to detect 1-2% of the cells in the circulating mouse blood, we injected the treatment, and the treatment was given in two different doses 0.4 and 0.7 MPK and of course that was the vehicle.

12:28

So after at the end point of the study, we sacrificed the mice, you can see even microscopically how the treatment worked very well.

12:38

So just looking at the spleen of the treated mice with the both doses, you can see that they come back to the normal sites where they're not treated mice.

12:48

So a very large spleen.

12:50

And if you look to the tissue that they were analysed like bone marrow and spleen, you can see that you have a dose dependence complete [destruction] of the Mutz-5 in bone marrow and spleen and particularly at the higher dose of course.

13:07

But you can see really that there is less tumour, when you treated even in 0.4 MPK.

13:15

So we went further at that, and we decided to see which was the therapeutic opportunity that we have because of course as I mentioned the only therapy available is the VXL that is a standard of care for those patients.

13:34

So we wanted to compare these two treatments.

13:38

So we reinjected the ADC with 0.7 MPK with the control ADC just to be sure that everything was fine.

13:48

And with the we VXL. We treated the mice.

13:51

So once we reached the circulating Mutz level 1%, we injected one single dose of ADC, and we treated for four weeks as is the normally perform the VXL treatment.

14:07

We stopped the treatment for one week and then we analysed what was happening in the mice.

14:12

And as you can see, the ADC was completely promoting the tumour regression where we VXL unfortunately still had some residual disease very visible.

14:23

Even the immunohistochemistry that we perform internally was showing no trace of Mutz-5 in the bone marrow and the overall survival curve is really impressive.

14:33

So really believe that is a good tool for therapeutics.

14:40

So just to give you a quick take away.

14:43

So starting from phage display you can really go straight and find something that can be nominated for, for the clinical point of view. Because you start you can select for the target, you can in this case we identify an antibody that was binding with high affinity internalising with the target and once conjugated with the payload was really efficacious in destroying tumour cells.

15:11

So can be really thought as a monotherapy against the for treatment of CLRF2 rearranged B-ALL.

15:20

One thing that I didn't mention, because the homology and identity between the human and the mouse isoform were really low like 35%.

15:31

We had to run a parallel campaign to develop a surrogate.

15:37

And I don't have time to show the data, but we identified another antibody able to recognise the mouse isoform with the same exact characteristic that is well tolerated.

15:50

So give us also strength about safety using this kind of tool.

15:55

So just to tell you that, I mean, we are fortunate that we can perform all this kind of characterization in house.

16:04

And really I want to show you that from phages display go ahead.

16:08

You can go with the hit ID, lead optimization, up to preclinical development.

16:16

And with that, I would like to thank all the people that have been working to this great project and the contributor.

16:26

And of course thank you all for your attention.

16:30

I will be happy to take any questions.