Introduction

Genetic modification of T cells to express a chimeric antigen receptor (CAR) has emerged as an effective therapeutic treatment for patients with B cell hematological malignancies over the last years. CAR T cells are generated from peripheral T cells isolated from blood of patients. Based on the differential expression of CD62L, CCR7, CD45RA and CD45RO, these peripheral T cells can be divided into five subsets: naive T (TN) cells, which are antigen-unexperienced, effector T (TEFF) cells, which migrate to sites of inflammation and promote pathogen clearance, and memory T cells, which persist long-term to allow protection against subsequent infections. Memory T cells include stem cell memory (TSCM), central memory (TCM) and effector memory (TEM) cells . In humans, T cell differentiation follows a linear progression where less differentiated cells give rise to more differentiated progeny: TN > TSCM > TCM > TEM > TEFF. During differentiation of TNtowards TEFF cells, the proliferative potential and memory functions are declining, while effector functions increase. Notably, the two markers, CD62L and CCR7 are only expressed on TN and early differentiated (TSCM and TCM) cells. During T cell isolation and subsequent cultivation, cells are usually activated using cytokines and stimulating antibodies to induce T cell proliferation and survival. In the past, IL-2 was most frequently used for cytokine support, thereby driving T cell cultures towards terminally differentiated T cells. More recently, IL-7 and IL‑15 are applied for T cell cultures in an effort to maintain a more naïve- or memory-like T cell phenotype.
Despite its promising results, CAR T cell therapy still needs to overcome various hurdles to become standard therapy for all patients in need. Automated processes have been developed to address the complicated manufacturing process. However, the most suitable T cell phenotype for CAR-mediated tumor therapy is a matter of debate. In general, naive and early memory T cells, are favored for cellular immunotherapy products due to their higher plasticity, longer persistence and greater capability to proliferate and differentiate into highly cytolytic effector cells. Along this line, a beneficial antitumoral function and cell persistence was associated with a high amount of less differentiated CAR T cells not only in patients with B-cell malignancies but also in patients with neuroblastoma.
For the generation of CAR T cell products, lentiviral vectors (LVs) pseudotyped with the glycoprotein of the vesicular stomatitis virus (VSV-G), harboring a broad tropism, are commonly used. Optimizing gene delivery through engineering of vector particles offers the potential to improve and simplify genetic modification of T cells. In this regard, receptor-targeted LVs (RT-LVs) specifically transducing CD3, CD4 or CD8 T cells have been described. All three vector types were recently shown to mediate the generation of CAR T cells directly in vivo in humanized mouse models. RT-LVs use a cell surface protein of choice as entry receptor, which can be achieved through pseudotyping with engineered glycoproteins from paramyxoviruses displaying a receptor-specific targeting domain, such as a single-chain antibody fragment (scFv) or designed ankyrin repeat molecule (DARPin). However, the T cell specific LVs available so far do not discriminate between the differentiation phenotype and exhaustion status of T cells.
Here we describe the generation of a RT-LV that is specific for a T cell marker expressed on less differentiated T cells: CD62L. The specificity of this vector was mediated by displaying a CD62L-specific scFv on measles virus (MV)-based RT-LVs. The resulting CD62L-LV mediated efficient gene delivery and preserved a higher degree of less differentiated CAR T cells upon long-term culture. CAR T cells generated through short-term incubation with CD62L-LV controlled tumor burden in an in vivo setting.