←
Caleb Perez29 posts
Caleb Perez avatar

Caleb Perez

@calebrperez

PhD Student | Birnbaum Lab | CAR T Cell Librarian @MITdeptofBE @KochInstitute

◷ Joined September 2014▧ 8 Photos and videos
Caleb Perez avatar
Caleb Perez@calebrperez·3 May 2024

Excited to share some of my PhD work was pre-printed! Using single-cell sequencing, we took a deep look at a library of CAR signaling architectures to try to identify signal-intrinsic drivers of CAR-T function🧵👇 @MEBirnbaum @kochinstitute @MITdeptofBE biorxiv.org/content/10.1101/…

MIT Dept of BE avatar
⟳ Caleb Perez retweeted
MIT Dept of BE@MITdeptofBE·20 May 2024

.@MEbirnbaum has been chosen to receive this year’s Bose Award for Excellence in Teaching. This award celebrates a faculty member who exemplifies exceptional dedication, care, and creativity in undergraduate education.

Post attachment
Blake Smith avatar
⟳ Caleb Perez retweeted
Blake Smith@_blake_smith·15 May 2024

An incredibly impressive body of work from the newly-minted PhD @calebrperez comparing clinical candidate CAR-T architectures head-to-head at single cell resolution, shedding light on mechanisms of differential CAR-T efficacy in vivo with insights into better construct design 🚗

Excited to share some of my PhD work was pre-printed! Using single-cell sequencing, we took a deep look at a library of CAR signaling architectures to try to identify signal-intrinsic drivers of CAR-T function🧵👇 @MEBirnbaum @kochinstitute @MITdeptofBE biorxiv.org/content/10.1101/…

Caleb Perez avatar
Caleb Perez@calebrperez·3 May 2024

The lab will have to ask these questions without me though, as I'm defending in two weeks! DM me for details if you want to hear more about my projects, or better yet, if you’ve got a job opening 🙏🙏🙏

Caleb Perez avatar
Caleb Perez@calebrperez·3 May 2024

I’m excited to see how @MEBirnbaum continues to use caRNA-seq, and other technologies like it, to answer such important questions. We think that more datasets like those presented here, combined with computational advances in ML, predictive tools for ICD design are within reach!

Caleb Perez avatar
Caleb Perez@calebrperez·3 May 2024

That’s one of many questions stemming from our work. To name a few: How do these transcriptional profiles change in different antigen binders, different cell types? What other phenotypes are accessible to T cells, beyond the small slice of diversity we searched here?

Caleb Perez avatar
Caleb Perez@calebrperez·3 May 2024

We found that the solid tumor microenvironment significantly changed transcriptional responses. This suggests that optimal signaling will be context-dependent... so why do we use the same ICD architectures in solid and liquid tumors?

Post attachment
Caleb Perez avatar
Caleb Perez@calebrperez·3 May 2024

Does this functional discordance reflect fundamental shifts in RNA response stemming from the solid tumor context? We performed caRNA-seq on our library, directly from an in vivo solid tumor. This is a first-of-its-kind dataset, mapping a CAR library to in vivo gene expression

Post attachment
Caleb Perez avatar
Caleb Perez@calebrperez·3 May 2024

So far, everything has been in liquid tumors, but do these observations extend to solid tumors? We tested the same candidate CARs against melanoma, and found huge functional differences. Our k7 candidate fails to control solid tumor growth, in contrast to what we saw in leukemia

Post attachment
Caleb Perez avatar
Caleb Perez@calebrperez·3 May 2024

This was a surprise, as tonic signaling is typically associated with👆exhaustion and👇persistence. But these observations are all in the context of conventional ICDs (BBz/28z). We posit that tonic activation of alternative signaling programs might be beneficial for persistence

Caleb Perez avatar
Caleb Perez@calebrperez·3 May 2024

Given such strong k7-induced cures, this begged the question, what sets k7 ICD architectures apart from others? Turned out that increased tonic signaling—turned on in the absence of antigen—were exclusive to k7 CARs, which seemed to help the k7 candidate CAR persist for months

Post attachment
Caleb Perez avatar
Caleb Perez@calebrperez·3 May 2024

But we didn’t stop there! To evaluate the functional consequences of each transcriptional phenotype, we picked individual CARs from each class for extensive follow-up. We found that the k7 candidate CAR induced long-term persistence and durable cures in mouse models of leukemia

Post attachment
Caleb Perez avatar
Caleb Perez@calebrperez·3 May 2024

To highlight one key finding from this: all approved CARs use either 28z or BBz ICDs, which are thought to be quite different. But they lie in the same class (k6) in the context of our whole library, suggesting that we are vastly undersampling potential phenotypes in the clinic!

Post attachment