Chimeric antibody receptors (CARs): driving T-cell specificity to enhance anti-tumor immunity

Preventing Lck Activation in CAR T Cells Confers Treg Resistance but Requires 4-1BB Signaling for Them to Persist and Treat Solid Tumors in Nonlymphodepleted Hosts

PURPOSE

Chimeric antigen receptor (CAR) T cells have shown promise against solid tumors, but their efficacy has been limited, due in part, to immunosuppression by CD4+FoxP3+ regulatory T cells (Tregs). Although lymphodepletion is commonly used to deplete Tregs, these regimens are nonspecific, toxic, and provide only a narrow window before Tregs repopulate hosts. Importantly, CARs have also been shown to inadvertently potentiate Tregs by providing a source of IL2 for Treg consumption. We explored whether disruption of the IL2 axis would confer efficacy against solid tumors without the need for lymphodepletion.

EXPERIMENTAL DESIGN

We developed second- (CD28z) and third- (CD28-4-1BBz) generation CARs targeting EGFRvIII. To eliminate secretion of IL2, 2 amino acid substitutions were introduced in the PYAP Lck-binding motif of the CD28 domain (ΔCD28). We evaluated CARs against B16 melanomas expressing EGFRvIII.

RESULTS

CD28z CARs failed to engraft in vivo. Although 4-1BB addition improved expansion, CD28-4-1BBz CARs required lymphodepletion to treat solid tumors. CARs deficient in Lck signaling, however, significantly retarded tumor growth without a need for lymphodepletion and this was dependent on inclusion of 4-1BB. To evaluate CAR vulnerability to Tregs, we lymphodepleted mice and transferred CARs alone or with purified Tregs. Cotransfer with Tregs abrogated the efficacy of CD28-4-1BBz CARs, whereas the efficacy of ΔCD28-4-1BBz CARs remained unperturbed.

CONCLUSIONS

In the absence of lymphodepletion, CARs targeting solid tumors are hindered by Treg immunosuppression and poor persistence. Here, CARs were modified to circumvent Treg suppression and to simultaneously improve in vivo engraftment. Modified CARs treated solid tumors without a need for lymphodepletion.

Clinical application of Sleeping Beauty and artificial antigen presenting cells to genetically modify T cells from peripheral and umbilical cord blood

The potency of clinical-grade T cells can be improved by combining gene therapy with immunotherapy to engineer a biologic product with the potential for superior (i) recognition of tumor-associated antigens (TAAs), (ii) persistence after infusion, (iii) potential for migration to tumor sites, and (iv) ability to recycle effector functions within the tumor microenvironment. Most approaches to genetic manipulation of T cells engineered for human application have used retrovirus and lentivirus for the stable expression of CAR(1-3). This approach, although compliant with current good manufacturing practice (GMP), can be expensive as it relies on the manufacture and release of clinical-grade recombinant virus from a limited number of production facilities. The electro-transfer of nonviral plasmids is an appealing alternative to transduction since DNA species can be produced to clinical grade at approximately 1/10(th) the cost of recombinant GMP-grade virus. To improve the efficiency of integration we adapted Sleeping Beauty (SB) transposon and transposase for human application(4-8). Our SB system uses two DNA plasmids that consist of a transposon coding for a gene of interest (e.g. 2(nd) generation CD19-specific CAR transgene, designated CD19RCD28) and a transposase (e.g. SB11) which inserts the transgene into TA dinucleotide repeats(9-11). To generate clinically-sufficient numbers of genetically modified T cells we use K562-derived artificial antigen presenting cells (aAPC) (clone #4) modified to express a TAA (e.g. CD19) as well as the T cell costimulatory molecules CD86, CD137L, a membrane-bound version of interleukin (IL)-15 (peptide fused to modified IgG4 Fc region) and CD64 (Fc-γ receptor 1) for the loading of monoclonal antibodies (mAb)(12). In this report, we demonstrate the procedures that can be undertaken in compliance with cGMP to generate CD19-specific CAR(+) T cells suitable for human application. This was achieved by the synchronous electro-transfer of two DNA plasmids, a SB transposon (CD19RCD28) and a SB transposase (SB11) followed by retrieval of stable integrants by the every-7-day additions (stimulation cycle) of γ-irradiated aAPC (clone #4) in the presence of soluble recombinant human IL-2 and IL-21(13). Typically 4 cycles (28 days of continuous culture) are undertaken to generate clinically-appealing numbers of T cells that stably express the CAR. This methodology to manufacturing clinical-grade CD19-specific T cells can be applied to T cells derived from peripheral blood (PB) or umbilical cord blood (UCB). Furthermore, this approach can be harnessed to generate T cells to diverse tumor types by pairing the specificity of the introduced CAR with expression of the TAA, recognized by the CAR, on the aAPC.