Lyell Immunopharma : Engineering potent CAR T-cell therapies by controlling T-cell activation signaling parameters using the Stim-R™ technology, a programmable synthetic cell-signaling platform

Engineering Potent CAR T-Cell Therapies by Controlling T-Cell Activation Signaling Parameters Using the Stim-RTM Technology, a Programmable Synthetic Cell-Signaling Platform

Abstract No. 252

Aileen Li, Jessica Briones, Jia Lu, Candace Sims, Quinn Walker, Rowena Martinez, Stefan Siebert, Lora Zhao, Emily Fu-Sum, Sheila Lou, Andrew Jimena, Elizabeth Pedrosa, Hajime Hiraragi, Bijan Boldajipour, Purnima Sundar, Shobha Potluri, Omar Ali and Alexander Cheung

Lyell Immunopharma, 201 Haskins Way, South San Francisco, CA 94080

Background

Results

• Chimeric antigen receptor (CAR) T-cell therapy has produced profound results in certain hematologic malignancies but has been less successful in the treatment of solid tumors
• Studies suggest that T-cell exhaustion plays a role in limiting the ability of CAR T cells to eradicate solid tumors1
• T-cellactivation is a formative event that directs cell fate, function, and durability of mature T cells
• During expansion of T cells to generate a CAR T-cell product, parameters related to activation can affect the phenotypic and functional quality of the resulting cells

Hypothesis

• Controlled delivery of activation molecules during T-cell production can generate CAR T cells with greater potency

Methods

• To control signaling during T-cell activation, we employed our Stim-R epigenetic reprogramming technology, a synthetic cell mimic that mediates precise signal-molecule presentation (Figure 1)

Figure 1: The Stim-R technology is a programmable cell-signaling platform

The Stim-R technology comprises biodegradable lipid-coated silica micro-rods that can present multiple signals in precise densities and stoichiometries. Soluble signals are released in a controlled manner while surface- anchored signals are presented on a synthetic lipid membrane, mimicking physiologic presentation.

• We designed and fabricated Stim-R technology formulations to present T-cell activating
	signals engaging CD3 and CD28 at different densities and stoichiometries
•	Utilizing these formulation variants, we generated arrays of diverse ROR1-targeted CAR
	T-cell products, which we profiled phenotypically and functionally
•	Based on these metrics, we compared Stim-R-generated CAR T cells to CAR T cells

Key Findings

The Stim-R epigenetic reprogramming technology generates potent ROR1-targeted CAR T-cell products with:

• Increased polyfunctionality
• Enhanced cytotoxicity and proliferation in vitro
• Persistence of a unique cell population enriched in both stemness and effector-associated gene signatures following repeated exposure to tumor cell lines in vitro
• Higher peak CAR T-cell number and prolonged CAR T-cell persistence in vivo
• Improved tumor control in vivo

Figure 3: Stim-R CAR T cells showed production characteristics comparable to CAR T cells generated using TransActTM, a conventional T-cell stimulation reagent ("Benchmark")

ROR1-targeted CAR T cells produced using the Stim-R technology or Benchmark showed comparable expansion (A), transduction efficiency (B), and frequency of stem-like cells (C). Data in (A-C) represent the same 4 independent donors. Data in (A-C) were analyzed using a Student's t-test. ns, not significant.

Figure 5: Stim-R CAR T cells showed enhanced cytotoxicity, expansion, and cytokine production in response to repeated ROR1+ target-cell stimulation

(A) Stim-R and Benchmark CAR T cells (red cells) were exposed to repeated antigen stimulation using the ROR1+ H1975 cell line (blue cells). (B) Stim-R CAR T cells showed superior target clearance compared to Benchmark CAR T cells. Target cell density was measured as total intensity of target cell fluorescence by IncuCyte. (C) Stim-R CAR T cells showed higher proliferation in response to target stimulation. (D) Stim-R CAR T cells showed a trend of higher IL-2 and IFNγ production after repeated stimulation. Data in

(B) represent the mean of 2 technical replicates for each donor. Data in (C) represent unique donors (dots) with the mean of the donors superimposed (lines). Data in (D) represent unique donors.

Figure 6: Transcriptomic analysis revealed that Stim-R CAR T cells retained a unique subset of stem-like cells with effector-associated gene signatures and displayed down-regulation of exhaustion-associated gene sets compared to conventional CAR T cells following repeated antigen stimulation

Figure 7: Stim-R CAR T cells exhibited higher peak T-cell numbers in the blood, prolonged persistence, and improved tumor control in vivo

1. In vivo study design: Stim-R and Benchmark CAR T cells were evaluated in the H1975 xenograft solid tumor model in mice. Stim-R CAR T cells showed improved tumor control (B) and overall survival (C) compared to mock- transduced Stim-R T cells (Stim-R mock) and Benchmark CAR T cells. Data in (B) represent individual tumor growth curves (n=10 mice per condition). (D) CAR T cells were measured in the blood over time. Despite a comparable number of CAR T cells detected in the blood 24 hours post-injection,Stim-R CAR T cells showed higher peak CAR T- cell numbers, prolonged persistence, and higher total CAR T cells detected over the course of the study (assessed by calculating AUC) compared to Benchmark CAR T cells. Data represent mean ± SE. (n=10 mice per condition).
   ****P<1×10-4 by Mantel-Cox test (C) or by one-way ANOVA, followed by Tukey's HSD post hoc test for each metric (D; only Benchmark and Stim-R CAR T-cell conditions shown); ns, not significant.

Conclusions

generated using a conventional stimulation reagent to identify lead formulations showing

superior in vitro function (Figure 2)

• We also evaluated the in vivo efficacy of lead Stim-R-generated CAR T cells

Figure 2: An unbiased screening method using the Stim-R technology to optimize TCR and co-stimulation signal strength to produce potent CAR T-cell products

The Stim-R technology was used to independently tune TCR and co-stimulation signal strength during T-cell activation in order to generate an array of CAR T-cell products for profiling. Stim-R CAR T-cell products were ranked based on performance in an in vitro repeated antigen-stimulation assay and the top-performingStim-R product was benchmarked in vitro and in vivo against CAR T cells generated using TransActTM, a commercially available T-cell stimulation reagent currently used for clinical CAR T-cell production ("Benchmark").

Figure 4: Stim-R CAR T cells exhibited increased polyfunctionality in response to acute ROR1+ target-cell stimulation

Stim-R CAR T cells and Benchmark CAR T cells were each co-cultured with a ROR1-expressing cell line for 5 hours. Subsequently, the cells were analyzed for IL-2 and IFNγ expression using flow cytometry (A). Stim-R CAR T cells showed higher population of polyfunctional cells, which express both IL-2 and IFNγ (B), and lower population of non-reactive cells, which express neither IL-2 nor IFNγ (C). Data in (B-C) represent the same 4 independent donors. Data in (B-C) were analyzed using a Student's t-test; *P<0.05.

Single-cellRNA-Seq and bulk RNA-Seq were performed on Stim-R and Benchmark CAR T cells collected on day 10 of a repeated antigen stimulation assay (Figure 5A) from 3 donors. (A) Stim-R CD8+ T cells displayed up-regulation of stemness-associated and effector-associated gene sets compared to Benchmark CD8+ T cells in bulk RNA-Seq gene set enrichment analysis. (B) Stim-R and Benchmark CD8+ T cells are separated on UMAP plot generated in single-cellRNA-Seq analysis. Putative stem-like clusters identified in Stim-R (C1) and Benchmark (C4), respectively, are indicated, with proportions shown in boxes.

(C) C1 in Stim-R (highlighted) exhibited high expression of stemness-associated gene marker TCF7 and protein marker CD27, as well as positivity for effector-associated genes GNLY and CCL5 (center panels). C1 in Stim-R showed significantly higher TCF7 RNA and CD27 protein expression compared to C4 in Benchmark (right panel). (D) Clusters C8 and C14 exhibited enriched module score of TTE gene set2 and high TIGIT protein expression. The proportion of C8+14 decreased in Stim-R CD8+ T cells compared to Benchmark CD8+ T cells. (E) Stim-R CD8+ T cells displayed down-regulation of TTE gene set2 compared to Benchmark CD8+ T cells in bulk RNA-Seq gene set enrichment analysis. ****P<1×10-4 by Wilcoxon test.

• Optimizing signal presentation during T-cell activation using Stim-R epigenetic reprogramming technology enabled the production of potent ROR1-targeted CAR T cells with improved polyfunctionality, persistence, and anti-tumor activity, attributes associated with improved outcomes with other T-cell therapies
• Stim-RCAR T cells retained a stem-like subpopulation with effector-associated gene signatures and showed reduced exhaustion following repeated antigen stimulation
• Our results suggest that enhancement of T-cell products with Stim-R technology may improve therapeutic benefit against solid tumors

Abbreviations

AUC, area under the curve; CAR, chimeric antigen receptor; CCL5, C-C motif chemokine ligand 5; CCR7, C-C motif chemokine receptor 7; CD, cluster of differentiation; CD45RA, CD45 200-245 kDa isoform; E, effector; EGFR, epidermal growth factor receptor; GNLY, granulysin; IFNγ, interferon gamma; IL-2: interleukin 2; MHC, major histocompatibility complex; NES, normalized enrichment score; NSG, NOD scid gamma; PK, pharmacokinetics; RFU, relative fluorescence units; RNA-Seq,RNA-sequencing; ROR1, receptor tyrosine kinase-like orphan receptor 1; SE, standard error of the mean; T, target; TCF1/7, transcription factor 1 protein; TCR, T-cell receptor; TIGIT, T-cell immunoreceptor with Ig and ITIM domains; TTE, T-cell terminal exhaustion.

References

1. Fraietta JA, et al. Nat Med. 2018;24(5):563-571. 2. Oliveira G, et al. Nature. 2021;596(7870):119-125.

Acknowledgements

We acknowledge Brianna Parish, Carson Harms, and Melissa DeFrancesco for their help with Stim-R technology process development. We also acknowledge Blythe Sather, Suman Vodnala, Gary Lee, Stan Riddell, Stephen Hill, and Nick Restifo for their scientific input and strategic guidance. Medical writing and editorial support were funded by Lyell Immunopharma, Inc., and provided by Leslie Mitchell, PhD, and Melanie Styers, PhD, of Verascity Science.

Presented at SITC Annual Meeting 2022; Nov 8-12; Boston, MA, USA