Abstract
Abstract 951
Adoptive immunotherapy with T cells engineered by gene transfer to express CD19-specific chimeric antigen receptors (CARs) has the potential to induce remissions in patients with advanced B cell malignancies. CARs are synthetic receptors with an extracellular antigen-binding domain (scFv), a spacer domain that provides separation of the scFv from the cell membrane and an intracellular signaling module, most commonly the CD3ζ chain and one or more costimulatory domains such as CD28 or 4-1BB. Several clinical trials with CD19-CAR T cells in small cohorts of patients with B cell tumors have been reported with variable results. Although most studies have used the CD19-specific FMC63 scFv as the tumor-targeting moiety, the extracellular, transmembrane and intracellular CAR domains used in each trial have been distinct, and an emerging paradigm is that including costimulation in the design of the CAR is key to achieving anti-tumor activity in vivo.
In this study, we analyzed the influence of extracellular spacer domain length on the in vitro and in vivo function of CD19-CARs. We constructed a panel of four CD19-CARs comprised of the FMC63 scFv and either a long spacer derived from the IgG4-Fc Hinge-CH2-CH3 domain (229 AA) or a short Hinge domain only spacer (12 AA). Each CAR contained a signaling module of CD3ζ with CD28 (short/CD28; long/CD28) or 4-1BB (short/4-1BB; long/4-1BB). We transduced CD8+ CD45RO+ CD62L+central memory T cells of normal donors with each of the CARs, enriched transduced T cells to >90% purity by immunomagnetic selection using a tEGFR marker encoded in the CAR vector, and expanded CAR transduced T cells using a uniform culture protocol. We compared the in vitro function of T cell lines expressing each of the CD19-CARs and confirmed specific cytolytic activity against CD19+ target cells including K562/CD19, and Raji and JeKo-1 lymphoma cells. Quantitative cytokine analyses showed higher levels of IFN-γ, TNF-α, IL-2 production in T cells expressing CD19-CARs with CD28 costimulatory domain compared to the corresponding constructs with 4-1BB, consistent with prior work. T cells expressing each of the CD19-CARs proliferated in vitro after stimulation with K562/CD19 and Raji tumor cells by CFSE dye dilution, with the strongest proliferation observed in T cells expressing the CD19-CAR ‘long/CD28’, consistent with the highest levels of IL-2 production by T cells expressing this construct.
We then analyzed the in vivo anti-tumor efficacy of each CD19-CAR in immunodeficient NOD/SCID/g−/− (NSG) mice engrafted with firefly luciferase transduced Raji cells. Tumor was inoculated on day 0, and once tumor was established (day 7), the mice received a single dose of 2.5×106̂ T cells expressing each CD19-CAR, a tEGFR control vector, or were left untreated. Surprisingly, only T cells expressing CD19-CARs with a short spacer domain (short/CD28 and short/4-1BB) eradicated the Raji tumors and led to long-term tumor-free survival of all mice. T cells expressing CD19-CARs with a long spacer domain (long/CD28 and long/4-1BB) did not confer a significant anti-tumor effect and all mice expired from systemic lymphoma at a similar time as control and untreated mice. The anti-tumor efficacy in vivo of T cells modified with long spacer CD19-CARs could not be improved by increasing CAR T cell dose 4 fold, or by including additional costimulatory domains into the CD19-CAR (long/CD28:4-1BB). Serial analyses in peripheral blood, bone marrow and spleen showed dramatically lower numbers of transferred T cells in mice treated with long spacer CD19-CARs compared to mice treated with short spacer CD19-CARs or control T cells. Further analysis revealed that despite strong activation in vivo as assessed by upregulation of CD69 and CD25, CD19-CARs with long extracellular spacer domain induced a high rate of activation induced T cell death in vivo. Collectively, these results demonstrate that the extracellular spacer domain that lacks intrinsic signaling function is critical in the design of effective CD19-CARs, and illustrates that tailoring spacer length is likely to be essential for designing effective CARs specific for other tumor antigens.
No relevant conflicts of interest to declare.
Author notes
Asterisk with author names denotes non-ASH members.