Integral Molecular hosted the webinar “Comprehensive Specificity Profiling of Antibody-Based Therapeutics Using the Membrane Proteome Array” on October 7, 2021. The following is the video transcript of a portion of the webinar presented by Dr. Timothy MacLachlan, Executive Director, Department of Preclinical Safety, Novartis Institutes for Biomedical Research.
In the webinar, Dr. MacLachlan discussed results of a recent survey of industry developers on their chosen methods used for profiling specificity, comparing traditional immunohistochemistry Tissue Cross-Reactivity (TCR) studies with emerging cell array technologies such as the MPA. The full survey results were published in MacLachlan et al., 2021, Regul Toxicol Pharmacol.
Introduction (17:29)
Thank you, Rachel. I'm Tim MacLachlan. I'm from Novartis working in preclinical safety. I'm going to talk to you a little bit today about a survey that we did with a consortia of other biopharmaceutical companies to understand what technologies that they've been using to understand specificity of their antibody therapeutics.
Original FDA Requirements for Specificity Testing of Monoclonal Antibodies (17:50)
Now the history of the technologies that we use for this come from guidance documents from the FDA and from ICH originating in 1997 that advised sponsors to take antibodies and pan them across tissue sections, 32 human tissues from three donors, and evaluate where your antibodies bind by immunohistochemistry techniques. This will tell you where your antibody binds, but it won't tell you what your antibody binds.
Interestingly, in both of these documents, they open the door for the opportunity for sponsors to utilize other techniques than immunohistochemistry against tissues as they become available and validated.
Alternatives to Tissue Cross-reactivity (TCR) Studies (18:39)
So, the question really is what are these alternatives? And over the last several years these have evolved to be in array format. Originally starting with slides that have random peptide libraries spotted on them and then panning the antibody against this with a fluorophore to see what it might bind to. More recently this has allowed for the spotting of human GST tagged proteins that are produced in yeast or bacteria to be spotted onto these slides. And the coverage is fairly decent from a human membrane proteome or even intercellular proteins. This allows evaluation of the binding of the antibody against a fully folded protein. Most recently, this has allowed an understanding of the binding of the antibody in the cellular context. One approach might be to spot plasmids that have a library of human membrane proteins spotted across a slide and then growing cells over those plasmids and then panning the antibody, again, over across the slide. An alternative way of doing this is by transfecting cells in a suspended format. And then panning, again, the antibody across those cells and then detecting binding by a flow cytometry method. This is a really valuable way to understand specificity and that it can tell you exactly what your protein binds to. Now it will not tell you in what tissue and cellular context it would bind to.
No Specificity Profiling Platform is Perfect (20:21)
So there certainly are some positives and negatives to either approach. The tissue cross-reactivity approach utilizes immunohistochemistry so it's dependent on the antibody being a good IHC reagent, which is actually quite rare for therapeutic antibodies. The IHC methods may not actually fully expose all proteins, so the coverage of the full human membrane proteome might not actually be true in the case of tissue cross-reactivity. There have been some cases of false positive staining and if you do get staining, it's unclear if that's actually your target protein in an unexpected location or if it's a true cross-reactant.
Now with the arrays, we know for a fact that it's not the full complete human proteome, at least not at this stage yet. Perhaps someday it would be. And we know that the protein is outside of its normal environments, we're utilizing workhorse cell lines to do this. Some sponsors have done quite a bit of this work, although some have not utilized this technology at this point in time. And as such, because of that lack of familiarity and experience, some consider this as not validated. So really needs to generate a little bit more widespread experience with it.
Conducting the Biosafe TCR 2.0 Survey (21:49)
So based on this, a group of us who focus on preclinical safety within the biotechnology innovation organization, wanted to survey the industry to see what kind of technologies that they have been using. We used as a launching pad, a paper that we had published on the tissue cross-reactivity study about 10 years ago. We wanted to ask what sponsors’ experience with that assay has been since then. And since these alternatives, these array methodologies, have been evolving we wanted to know if sponsors have been using these, if they'd been using them for early screening, and what kind of platforms that they've been using them with, and what their experiences have been.
TCR Studies Have Rarely Led to Changes in Clinical Practice (22:34)
So briefly, here is an overview of the results from this survey that has now been published. We asked about sponsors’ experience with the tissue cross-reactivity study. There were 26 respondents and most of the respondents had done up to 30 to 40 different tissue cross-reactivity studies over the last 10 years.
They don't do them frequently in nonclinical species, this is not required, only human tissue cross-reactivity is required. Although a small fraction of these sponsors do them to try to validate some of their nonclinical species. They use tool antibodies frequently for target localization early in development programs. They have filed some INDs without these TCR studies or they're considering in the future, and it hasn't always been with oncology INDs. Most are running the tissue cross-reactivity study alongside the GLP tox with GLP tox material.
In the majority of sponsors’ experiences, the tissue cross-reactivity data has not predicted a safety finding in the preclinical studies, and the results of this have not led to a halt in development of a candidate. It hasn't affected the overall preclinical program, and hasn't led to a change in clinical trial design, and hasn't led to a clinical hold by regulators. That being said, there are individual examples and anecdotes where the assay was valuable, suggesting that there's still value in performing this assessment to catch that rare, unexpected event.
Increasing Adoption of Cell Array Technology (24:29)
Now, have sponsors started to use alternatives to the tissue cross, namely arrays? Approximately half of the 26 respondents have begun to use arrays for cross-reactivity screening. The majority of these have not found that they (arrays) are generating overtly uninterpretable datasets. They report mostly clean or rare, false positive datasets that they deem to be quite valuable in understanding the specificity of their antibody.
A question was posed in discussion with some of these sponsors if there is an opportunity to replace the tissue cross-reactivity study with these new technologies. It was noted that clearly, as I pointed out earlier, that the datasets provide information different from what is generated from the TCR, which is what with the arrays versus the where with the tissue cross. Certainly, some have suggested, "Well, we could just combine this with a tissue cross assessment," which some do. Certainly, we at Novartis for non-oncology indications are currently combining this with a tissue cross, but looking long-term, this does not necessarily seem to be financially feasible in the long-term. And so, we are looking for opportunities to potentially replace with more experience with this array technology.
Experience Submitting Cell Array Data to Regulatory Agencies (26:03)
So, the bottom line is that there is growing experience with these alternative methods, but we have discovered through discussing the results of this survey with our health authority colleagues at the FDA that the regulatory agencies need to see more of this data, even if it's just exploratory, because they need to understand what the variability is, and what the value of the data that they're getting out of it is. Only half of the respondents that have been performing these note that the data has been sent to health authorities, and we understand that they haven't received any concerning health authority feedback on any of these. We do know that there are some sponsors that are just resistant to replace. They've had over a decade, more than a decade, of experience with tissue cross-reactivity as the means to understand specificity and they haven't had any hiccups with this, that it's just fine. The alternative methods have their own problems, we don't have experience with it, so they want to stick with it. So, I think to each their own, really, what we are advocating is that sponsors have a choice in what technology that they use and we long-term would hope that health authorities would accept whatever technology a sponsor would choose to generate this kind of data.
Conclusions (27:35)
So, combining all of the information we got out of the survey, we pretty much are where we were 10 years ago when we published our original paper with the tissue cross-reactivity in 2010. It's not often generating critical safety information. It's not identifying a critical safety issue. However, there's that rare, rare circumstance, perhaps one in 10 years, where it did find something that needed to be followed up on. And that there's growing experience with these alternative methods and that we're not quite ready to completely replace it for non-oncology indications, but we should monitor over time and see what sponsors’ experiences are.
We're really not interested in just lumping this on to existing data because, while we do believe them to be complementary, we actually believe one to be sufficient for understanding the information that we want to gain from this. The regulatory agencies need to see more of this data and there are some, as I mentioned, some resistance to replace. So, engagement with health authorities for the use of these methods is critical.
We have hypothesized that there are a variety of different potential scenarios that sponsors could consider in submitting cross-reactivity data to health authorities. We could have an array screen coupled with a standard full TCR assay. As I mentioned, this could be quite expensive and we're not exactly sure how often that would actually yield more valuable data than just one standard TCR or one array screen alone.
We think that it's quite possible that a combination of an array screen coupled with data from a target localization study using a tool antibody tissue IHC would be the best-case scenario and is often done in the early evaluation of a target by antibody developers to understand specifically where their target is going to be and where the tissues that they're going to want to monitor are. This is much easier done in a research, non-GLP, setting with an antibody that is optimized for immunochemistry, not necessarily the one that is optimized for therapeutic use.