Defining an Optimal Dual-Targeted CAR T-cell Therapy Approach Simultaneously Targeting BCMA and GPRC5D to Prevent BCMA Escape-Driven Relapse in Multiple Myeloma

Abstract

CAR T-cell therapy for multiple myeloma (MM) targeting B-cell maturation antigen (TNFRSF17; BCMA) induces high overall response rates; however, relapse occurs commonly. Implicated in relapse is a reservoir of MM if cells lacking sufficient BCMA surface expression (antigen escape). We demonstrate that simultaneous targeting of an additional antigen-here, G protein-coupled receptor class-C group-5 member-D (GPRC5D)-can prevent BCMA escape-mediated relapse in a model of MM. To identify an optimal approach, we compare subtherapeutic doses of different forms of dual-targeted cellular therapy. These include (1) parallel-produced and pooled mono-targeted CAR T-cells, (2) bicistronic constructs expressing distinct CARs from a single vector, and (3) a dual-scFv "single-stalk" CAR design. When targeting BCMA-negative disease, bicistronic and pooled approaches had the highest efficacy, whereas for dual-antigen-expressing disease, the bicistronic approach was more efficacious than the pooled approach. Mechanistically, expressing two CARs on a single cell enhanced the strength of CAR T-cell/target cell interactions.

Keywords: adoptive cellular therapy; antigen escape; chimeric antigen receptor; immunotherapy; multiple myeloma.

Figure 1.

Dual-targeted CAR T cells express both scFv's efficiently and specifically lyse target antigen–positive cells. A, BCMA/GPRC5D dual-targeted CAR strategies evaluated. (i, ii) simultaneous 1:1 infusion of independent CAR T cells manufactured in parallel; (iii–iv) bicistronic dual-CAR expression on T cells via a construct with a “self-cleaving” 2A peptide; (v) tandem-scFv, “single stalk” CAR design. SP, signal peptide; S, spacer; TM transmembrane domain. B, Expression of individual BCMA-targeted and GPRC5D-targeted scFvs on the surface of primary donor T cells by flow cytometry with reagents specific to either the BCMA-targeted scFv or the GPRC5D-targeted scFv. C, Total CAR expression and transduction efficiency as measured by flow cytometry with antibody to the spacer domain shared among all CARs. FMO, fluorescence-minus-one. D, Cytotoxicity measured by ATP-dependent bioluminescence after 24-hour coculture of CAR T cells with a monolayer of 3T3-aAPC ffLuc⁺ cells expressing the indicated antigen(s); normalized to aAPCs alone. Mean ± SD from representative biological triplicate shown.

Figure 2.

Upfront treatment with dual-targeted CAR T cells prevents BCMA escape–mediated relapse in a rechallenge model. A, Experimental scheme, 2 × 10⁶ cells of the human bone marrow tropic myeloma cell line OPM2-WT were injected via tail vein into NSG mice and allowed to engraft and expand for 14 days. Mice were randomized to treatment with donor T cells (3 × 10⁶) gene-modified as indicated. Mice in the treatment arms that showed signs of xenogeneic GvHD were euthanized, and on day 105 the remaining long-term surviving mice were challenged with OPM2-BCMA KO cells (2 × 10⁶). B, Kaplan–Meier curves of overall survival after OPM2-WT injection. P values are indicative of each arm compared with BCMA-Δ control. C, Bioluminescent imaging of tumors over time. D, Kaplan–Meier curves after OPM2-BCMA KO injection. P values are indicative of each arm compared with BCMA-41BBz arm.

Figure 3.

Pooled and bicistronic CAR approaches demonstrate enhanced efficacy over the single-stalk approach at subtherapeutic doses in an established BCMA-escape model. A, Experimental scheme, NSG mice were injected intravenously with 2 × 10⁶ OPM2 tumor cells, including a subpopulation (5%–10%) of BCMA-KO cells. Each OPM2 population had been modified to express a distinct luciferase with a nonoverlapping substrate (OPM2-BCMA KO, firefly luciferase; OPM2-WT, membrane-tethered Cypridina luciferase) for in vivo imaging from the same animals over time. Day 14, mice were randomized for treatment with 2.5 × 10⁵ gene-modified T cells. B, Kaplan–Meier curves for mice treated with the indicated CAR T cells. C, Tumor burden by bioluminescent imaging over time of OPM2-WT membrane-tethered Cypridina luciferase⁺ (vargulin substrate) and OPM2-BCMA KO ffLuc⁺ (d-luciferin substrate).

Figure 4.

Bicistronic CAR approach demonstrates enhanced efficacy over the pooled approach at subtherapeutic doses in a model when target cells exclusively express both antigens. A, Experimental scheme: NSG mice were injected intravenously with a pure population of 2 × 10⁶ OPM2-WT cells (endogenously expressing BCMA and GPRC5D). After a 14-day engraftment/expansion period, mice were randomized to the 4 treatment groups shown in B, each receiving a single injection of 2.5 × 10⁵ gene-modified T cells. B, Tumor burden imaged at the time of treatment and day 14 posttreatment with gene-modified CAR T cells. C, Kaplan–Meier curves of overall survival. D, Summary of results. In the presence of BCMA-escape (GPRC5D-only expressing) target cells, the single-stalk CAR approach was less efficacious than pooled or bicistronic strategies, which were similarly efficacious to each other. In contrast, when targeting BCMA/GPRC5D dual-expressing target cells, the bicistronic strategy showed enhanced efficacy compared with the pooled approach.