AUTHORS
Yingshi Chen, PhD, Aditya Bharti2, Steven Pastor3, Katherine Mueller, PhD4, Bakir Valentic5, Noah Ford6, Jose Arias-Umana7, Jessica Barragan8, Gabrielle Zuern9, Martina Markovska10, Siddharth Chitta11, Caitlin Hopkins12, Alexander Shestov13, Joseph Fraietta, PhD14, Roddy O'Connor, PhD15, Evan Weber, PhD16
AFFILIATIONS
1 The Children's Hospital of Philadelphia, Phialadelphia, United States, University of Pennsylvania, Phialadelphia, United States
2 The Children's Hospital of Philadelphia, Phialadelphia, United States
3 The Children's Hospital of Philadelphia, Phialadelphia, United States
4 The Children's Hospital of Philadelphia, Phialadelphia, United States, University of Pennsylvania, Phialadelphia, United States
5 University of Pennsylvania, Phialadelphia, United States
6 University of Pennsylvania, Phialadelphia, United States
7 The Children's Hospital of Philadelphia, Phialadelphia, United States, University of Pennsylvania, Phialadelphia, United States
8 The Children's Hospital of Philadelphia, Phialadelphia, United States, University of Pennsylvania, Phialadelphia, United States
9 The Children's Hospital of Philadelphia, Phialadelphia, United States
10 University of Pennsylvania, Phialadelphia, United States
11 University of Pennsylvania, Phialadelphia, United States
12 University of Pennsylvania, Phialadelphia, United States
13 University of Pennsylvania, Phialadelphia, United States
14 University of Pennsylvania, Phialadelphia, United States
15 University of Pennsylvania, Phialadelphia, United States
16 The Children's Hospital of Philadelphia, Phialadelphia, United States, University of Pennsylvania, Phialadelphia, United States
ABSTRACT
Background: A major factor limiting the durability of CAR T therapies for hematologic malignancies is poor in vivo CAR T cell persistence. Optimization of CAR signaling, gene editing, or overexpression of cytokines or transcription factors can augment persistence. However, these advances have not yielded a unifying mechanistic framework for durable CAR T cell activity, thereby limiting the rational design of more efficacious CAR T therapeutics.
Methods: Meta-analysis of 16 publicly available RNA-seq datasets from human CAR T cells engineered with various potency enhancements identified NF-κB signaling as the most highly-enriched and conserved transcriptional program. To define NF-κB transcription factor(s) driving this program, we overexpressed or CRISPR-deleted individual canonical (RelA, cRel) or non-canonical (RelB) NF-κB transcription factors in human CD19 CAR T cells. CAR T cell expansion, differentiation, exhaustion, persistence, and antitumor function were assessed using acute and chronic antigen stimulation assays and Nalm6 leukemia xenograft models. Bulk RNA-seq and semi-targeted metabolomics were used to identify pathways and processes governed by NF-κB transcription factors.
Results: Among the three major NF-κB transcription factors, only cRel was necessary and sufficient for CAR T cell persistence and potency. cRel overexpression (cRelOE) generated an effector-like phenotype (P=0.0008), led to sustained ex vivo CAR T cell expansion that lasted >5 weeks (P=0.0002), and augmented cytokine secretion under settings of chronic stimulation (IFNγ, P=0.0142; IL-2, P<0.0001). Conversely, CRISPR deletion of endogenous REL impaired expansion (P=0.0044) and promoted a quiescence-associated phenotype (P=0.0264) with reduced expression of effector markers CD25 and Granzyme B (P<0.001). Nalm6-engrafted mice treated with cRelOE CD19 CAR T cells exhibited a significant survival advantage (60% vs 0% by day 60, P<0.0001) and elevated CAR T cell levels in the blood, spleen, and bone marrow (all P<0.05). RNA-seq and metabolomics revealed that cRel coordinates cholesterol biosynthesis (NES=2.22, P=1.6×10⁻⁸) and glutamine-dependent anabolic metabolism (NES=2.00, P=2.3×10⁻⁸) to support durable CAR T cell persistence and antitumor activity. Importantly, we identified a unique cRel gene signature that was associated with superior in vivo persistence in patients treated with CD19-directed CAR T cell therapy in several independent clinical studies (P=0.0011).
Conclusions: We demonstrate that cRel underpins broadly-conserved NF-κB programs that sustain CAR T cell persistence and antitumor activity. These findings establish canonical NF-κB signaling and cRel as important biological axes that can be leveraged to design CAR T cell therapeutics with enhanced potency and durability.
DISCLOSURES
Yingshi Chen, PhD: No relevant conflicts of interest to declare.