You’ve Got a Fast CAR
It takes a certain amount of effort to convince myself that Tracy Chapman’s Fast Car is not about lymphocytes bearing chimaeric antigen receptors. Seemingly everyone has a fast CAR and a plan (as the song goes) to get them to the clinic. The CAR modality, so far, has been applied to pro-inflammatory lymphocytes. To outfit Tregs, a classically anti-inflammatory lymphocyte, with a CAR is an inspired choice by the Ferreira Lab at the Medical University of South Carolina. After all, adoptive Treg therapies do represent a panacea of sorts for chronic inflammatory disorders. Having generated and extensively characterized their CAR Tregs, Ferreira’s group uncovered a Janus-like phenotype that is both pro-inflammatory and anti-inflammatory, and, putatively, the ability to drive a germinal centre reaction.
It's a cool paper, and you should read it. As for me, it seemed like the CAR Tregs were behaving like a subset of induced T-follicular helper (Tfh) cells derived from once T-bet-expressing CD4s. The interferon gamma (IFNg) release is consistent with this idea as is the CD40L expression. Furthermore, Tfh cells also make many of the cytokines catalogued in the paper i.e. IL-4, IL-6, IL-10 and IL-17A.
While the repertoire of cytokines secreted is important, so is examining the TCR-antigen contract of activation. Much has been said over the years about how, upon antigen encounter, high affinity TCRs tend to skew Th1, and low or med-affinity TCRs run towards a Tfh-phenotype. While the antigen-binding moiety (the scFv) of the CAR is not a classical TCR, the same rules may apply. And why shouldn’t they? The downstream signaling apparatus is the same as what CAR-less T cells possess.
That the authors see the cytokine balance tip from the pro-inflammatory to the immunomodulatory upon turning down scFv affinity confirms the aforementioned. This is a critical finding because it highlights an important consideration for CAR design. Since antigen affinity correlates with antigenicity thus promulgating a pro- or anti-inflammatory programmes, it behoves one to tune carefully scFv affinity to keep the peace in vivo, so to speak!
Let’s talk IFNg, however. Bien sur, IFNg+ Tregs are real. In fact, there are data showing that IFNg+ Tregs are helpful in preventing GVHD following an organ transplant. Not to be left behind, Tfh cells can and do secrete IFNg as well. IFNg production by Tfh following vaccination shapes the local humoral response in the lung. But we are in uncharted territory with CAR Tregs. It is unclear what is inducing the IFNg release and what that means, functionally speaking. I have thoughts, though!
Now, co-stimulatory molecules help direct CD4 T cell polarisation to distinct Th subsets. I don’t know how much ICAM-1 K562 lines express, but I’m sure they express some given their myelomonocytic provenance. Know that the ICAM-1/LFA-1 axis is a major component of the immunological synapse (the pSMAC, specifically, should you go looking!). Functionally, that axis is critical to Th1, Tfh and Treg differentiation. Again, I don’t know if these polarizing elements exist in the in vitro systems used by Cochrane et al; I’m just spitballing here.
If it is not clear why I have harped on about Tfh cells so much in this little write-up, it’s because of CD40L emergence in activated CAR Tregs. As they say in the paper itself, it’s not clear what kind of B cell help this IFNg-producing CD40L-hi population will provide, if any. Yet, once again: I have thoughts!
CD40 engagement in the presence of IFNg induces T-bet expression in B cells (TLR9 stim is better, but that’s a story for another day). All things in immunology are context-dependent. T-bet+ B cells drive some beautifully specific and robustly protective antibodies in viral infections. But they spell bad news in some autoimmune contexts and obesity (something to consider for CAR Treg treatment for type 1 diabetes). Will CAR Tregs recapitulate this biology? I really, really hope this story has a second season where we find out!
In sum, CAR Tregs possess a seemingly enigmatic biology which functionally sits somewhere in the intersection of Th1, Tfh and Tregs. I do not know if the authors looked at Bcl6 or T-bet expression. Perhaps even ICOS? But I’m certain all of those would have popped in the comprehensive gene expression analyses contained in the paper. And look, Tregs are hard: hard to isolate, hard to expand and, I feel certain, hard to engineer, and keep alive post-engineering. Cochrane and colleagues’ efforts here are ipso facto heroic in my eyes.
CAR Treg therapies herald promise. What’s required are guidelines for CAR design that inform which and co-stimulatory or inhibitory receptors (the authors found interesting outcomes with TIGIT!) must the cells be impelled to express, how exquisitely must TCR affinity be tuned and how precise the CAR expression levels themselves should be. It will also be fun to see how CAR Tregs hold up against Tregs engineered to express a native TCR.
Ultimately, one pair of lines in Tracy Chapman’s song goes You’ve got a fast car/I want a ticket to anywhere. And now with CAR Tregs joining the pantheon of adoptive cell therapies, the field feels just a bit more infinite.
Peace!
