Skip to main content


T Cells (and Bifunctionals) For Everything

Let’s take a one-day break from the coronavirus, but fear not, I will not be forsaking immunology. We’ll take a look at an interesting therapeutic idea involving T-cells and how it’s going.

For some years now, Amgen has been working on a technology called “BiTE“, which stands for “bispecific T-cell engager“. It fits into the general “bifunctional molecule” idea that’s so popular these days. It’s a big deal in protein localization and cellular localization, but there’s a longer history in trying to apply the idea to antibodies. What if each arm of the “Y” structure of a typical antibody bound to a different antigen? That’s not how the body grows them, but we could find uses for such things. There is, in fact, one of these on the market now: emicizumab, which brings together two proteins in the clotting factor cascade in a way that patients with hemophilia A can’t.

You can extend that idea to an even less natural-looking structure by moving to “nanobody” type proteins (single-chain antibodies) and linking them together. The first of these to make it through the clinic was blinatumomab, and its mechanism of action will tell you why Amgen has been sinking so much money into the idea after buying the protein’s original developer (Micromet) in 2012. Bilinatumomab has one small antibody domain that targets a surface protein called CD19, which is found specifically on B cells. There’s a short flexible linker between that one and another antibody domain that targets CD3, a surface protein on T cells. So blinatumomab should, in theory, force T cells over into proximity with B cells, while the binding to CD3 also activates the T cells to start exerting their cytotoxic effects. In a healthy individual this would be a spectacularly bad idea, but if you have B-cell leukemia, it could be just the thing.

If you’re into this sort of thing, you will already have noted the general similarity between such an approach and chimeric antigen receptor T cell therapy. CAR-T involves re-engineering a patient’s own T cells to recognize surface antigens such as CD19 to the same end: causing (in that case) a T-cell attack on the excess B cells in leukemia. Here’s a recent paper comparing the two approaches, but it needs to be noted that it’s an Amgen paper and thus could be more disposed towards the BiTE technology. Both approaches can be effective, with similar side effects (cytokine release syndrome being a major one), and both can show relapses over time. CRS is a real concern with all such approaches – in 2019, Regeneron reported two deaths from cytokine release in a clinical trial of one of their own bispecific antibody candidates (targeting CD3 again for the T cells and CD20 for follicular lymphoma). Interestingly, there’s at least one report of a patient with acute lymphoblastic leukemia who relapsed after CAR-T treatment who then went on blinatumomab and showed a complete remission. As with all immuno-oncology, we’re still finding our way on these things.

One of the ways that Amgen has been trying to improve their BiTE molecules is through improving their half-lives. They’ve been working on one of those (pavurutamab, AMG 701) that brings in T cells as before via CD3 and also targets the B-cell maturation antigen protein (BCMA) to treat multiple myeloma. But in their earnings release the other day, the company disclosed that they’re pausing trial enrollment due to some apparent toxicity events as they bring up the dosing levels in Phase I – that’s the only thing it can mean when you say you’re talking with the FDA about “optimizing safety monitoring and mitigation”. They also disclosed that they’re pausing another longer-half-life BiTE (AMG 673) that targets CD3/CD33 for AML – the company says that they’re going to gather more data in the clinic on the original bifunctional (AMG 330) they have in that combination before proceeding with this one. And yet another BiTE, this one (AMG 596) combining CD3 and an EGFR variant found in glioblastoma, has also been put on hold, and the only only reason given is “while we prioritize our portfolio”. As an aside, don’t you just love press release phrasing?

Now, they still have BiTE programs going – the release specifically mentions AMG 757, targeting delta-like ligand 3 (DLL3) for small cell lung cancer. But the path to extending clinical benefit with new immuno-oncology ideas has not been a smooth one, and it’s not going to be graded and paved any time soon. Amgen had some of the earlier Micromet-derived programs fail on them, such as AMG 110, a CD3/EpCAM (epithelial cell adhesion molecule) candidate, and there will absolutely be more failures to come.

But not all of those will be Amgen’s, nor will all the successes. I haven’t even mentioned another parallel effort from MacroGenics, their “DART” (dual affinity re-targeting) platform and their “TRIDENT” one that brings in a third protein. They’ve gone after some similar targets (there are a number of obvious places to start in immuno-oncology, but the list is not as long as you might like). For example, here’s their CD3/CD19 program in an early iteration. There are applications outside oncology, too, of course, and they just signed a deal last month for one of those with J&J/Janssen (no details!) Another such platform is “CrossMab” from Roche, who have their own technology for making specific antibody light chain/heavy chain combinations. They’re all over the clinic as well, and there are plenty of others (see below!)

This 2019 review will give you an idea of the landscape in this area. It’s anyone’s guess what this will all end up looking like, but we’re going to learn a lot about immunology along the way, and the successes will very much be worth seeing. I’ll leave everyone with the graphic below, from that review paper – this is what an expanding field of research looks like and no mistake:

15 comments on “T Cells (and Bifunctionals) For Everything”

  1. Kevin Law says:

    When one of these Amgen efforts fails in the clinic, we’re all going to be tempted to respond with “another one BiTEs the dust”.

    1. Marko says:

      And repeated such failures have taken a BiTE out of Amgen’s share price, prompting the CEO to proclaim : ” BiTE me ! Why are we wasting our time with this stuff ! “

      1. Larry says:

        Marko beat me to it…

  2. Barry says:

    This is kin to an idea I floated 20yrs ago. Because GrowthFactorReceptors (GFRs) are activated by dimerization, a ligand that binds to two different ones could specifically signal ONLY to cells which display BOTH of them. Since cancer cells are notorious for displaying GFRs that they shouldn’t have, they’re the obvious target for such.

  3. sgcox says:

    In a sense bifunctional antibodies are akin to “molecular glue” or PROTACs which are so popular these days. So why not make small molecules which will glue two types of cells of choice together and manipulate their function ?
    Not sure Fc of BiTe are of any benefit here, can actually call immune system to destroy this anomalous two-cell contraption.

    1. Andrew says:

      Because small molecules are comparatively weak and promiscuous.

  4. George Dorbell says:

    Apologies for going off topic, but retraction request submitted 27 November 2020 regarding the seminal paper on application of Reverse Transcriptase Polymerase Chain Reaction methodology to detection of SARS-CoV-2 in human beings (“The Corman-Drosten paper,” submitted to the journal Eurosurveillance 21 January 2020, published 23 January 2020…)

    …Considered at the annual editorial board meeting of the journal Eurosurveillance on 4 December 2020, then further reviewed anonymously over the next two months by a group of five invited laboratory experts, who, “confirmed that the article was scientifically adequate for its purpose and for the limited data and material available at this early stage in the Covid-19 pandemic…”

    …And that, “Any laboratory deciding to use the primers and protocol suggested in this article would ascertain the assay for its fitness for purpose and compliance with local quality and accreditation requirements; this is what has happened worldwide since the publication of the article. With more data and evolving knowledge, laboratories have since further improved the initial method, as per usual practice…”

    …And so, “In conclusion, after a thorough investigation in which we collected scientific advice from various sources, including several external reviewers, the editorial team—unanimously supported by its associate editors, except for those who were involved as co-authors—has decided that the criteria for a retraction of the article have not been fulfilled.”

    “Response to retraction request and allegations of misconduct and scientific flaws”

    “External peer review of the RTPCR test to detect SARS-CoV-2 reveals 10 major scientific flaws at the molecular and methodological level: consequences for false positive results”

    “Detection of 2019 novel coronavirus (2019-nCoV) by real-time RT-PCR”

    Presumed all signed, sealed and delivered, either in person or through online communication, by the editorial board of the journal Eurosurveillance, on a bright cold day around the end of January 2021, as the clocks were striking thirteen.

  5. ScientistSailor says:

    @Barry, wouldn’t that just turn on the growth signal even further? Or would you arm that bispecific mAb with a toxin?

    1. Barry says:

      cancer cells (in solid tumors, not in lymphomas/leukemias) are routinely treading the line of hypoxia. Flogging one to grow more should tip it into necrosis.

      1. sgcox says:

        This is a first time I heard of promoting cancer cell proliferation as an effective cancer therapy…
        May be bifunctionals promote non-productive heterodimers which block the growth signal from normal homodimers ? Or may be simple Fc mediated recruitment to cancer cells with abundance of receptors not usually present together on normal cells.

    2. Barry says:

      Conventional chemotherapy kills any cell that initiates mitosis. This could selectively light up cancer cells for the chemo.

  6. A Protein Engineer says:

    Do remember that mAbs themselves can be/are bispecific. The Fab arms of an IgG1 binds the antigen on the target and the Fc binds CD16 on NK cells leading to ADCC (e.g., trastuzumab)

  7. Daren Austin says:

    Point of information. IgG4 is often bispecific and I’m not counting the Fc portion, either

    Nature has been doing it for years 🙂

  8. Jb says:

    This may be a great approach if one targeted something like cancer associated glycosylation. There are all sorts of cancer specific glycans, and for years people have tried to develop vaccines against tumor associated sugars but have failed mostly because it is very difficult to mount a robust immune response against glycan targets. However, if there were strategy that’d target the glycan and the underlying protein with a bispecific approach, maybe it’d finally work. The Tn and STn antigens would make for fantastic targets if the underlying peptide could be targeted as well, because the truncated glycan would probably make it easier to target both.

  9. very intersting and useful

Leave a Reply

Your email address will not be published. Required fields are marked *

Time limit is exhausted. Please reload CAPTCHA.

This site uses Akismet to reduce spam. Learn how your comment data is processed.