Engineering T cell function using chimeric antigen receptors identified using a DNA library approach

Abstract

Genetic engineering of cellular function holds much promise for the treatment of a variety of diseases including gene deficiencies and cancer. However, engineering the full complement of cellular functions can be a daunting genetic exercise since many molecular triggers need to be activated to achieve complete function. In the case of T cells, genes encoding chimeric antigen receptors (CARs) covalently linking antibodies to cytoplasmic signaling domains can trigger some, but not all, cellular functions against cancer cells. To date, relatively few CAR formats have been investigated using a candidate molecule approach, in which rationally chosen molecules were trialed one by one. Therefore, to expedite this arduous process we developed an innovative screening method to screen many thousands of CAR formats to identify genes able to enhance the anticancer ability of T cells. We used a directional in-frame library of randomly assembled signaling domains in a CAR specific for the tumor associated antigen erbB2. Several new and original CARs were identified, one of which had an enhanced ability to lyse cancer cells and inhibit tumor growth in mice. This study highlights novel technology that could be used to screen a variety of molecules for their capacity to induce diverse functions in cells.

Figure 1. The chimeric cDNA library is diverse in size and composition.

(a) The genetic construct is depicted encoding extracellular single chain anti-erbB2 linked via a c-myc tag and the hinge region from CD8 to the transmembrane domain of CD28. DNA encoding various signaling molecules was randomly ligated into the SfiI site. (b) Plasmids of 20 representative clones were digested with NotI and XhoI and analyzed by agarose gel electrophoresis. (c) DNA from the 20 clones was sequenced and represented as shown (not to scale). Corresponding clone numbers and weights from panel (b) are indicated. Break indicates unsequenced portions.

Figure 2. Novel chimeric receptors can enable Jurkat cells to respond following receptor ligation.

(a) Populations of the human T cell line Jurkat (as listed) were incubated with anti-c-myc antibody and the expression of the activation marker CD69 determined. The boxed region represents cells with significant upregulation of CD69. Cells from the library-transduced population were subjected to fluorescence activated cell sorting and cells within the region cloned. CD69 upregulation in response to a positive control, phorbol myristate acetate (PMA) and phytohemagglutinin (PHA) is also depicted, together with upregulation in cells expressing a known active CAR, αerbB2CD28ζ. (b) Four clones were screened for their ability to secrete IL-2 in response to CAR ligation by erbB2 antigen, compared to the known positive control αerbB2CD28ζ. (c) IL-2 secretion from clones following incubation with anti-c-myc antibody (3 experiments pooled). (d) Expression of CARs in Jurkat clones was determined by flow cytometry following staining with anti-c-myc antibody. Four clones and the positive control Jurkat-erbB2CD28ζ are listed. (Error bars = SEM). *p<0.05, **p<0.005.

Figure 3. Copy number of chimeric receptors in Jurkat cell lines evaluated by PCR.

The intensity of the PCR products from Jurkat cell lines expressing chimeric receptors was compared to standard copy number of parental Jurkat genomic DNA spiked with 1–16 copies of a CAR by (a) PCR and quantified using (b) Metamorph software. Shown is a representative experiment from 3 experiments performed.

Figure 4. Human primary T cells bearing the αerbB2DAP10ζCD27 CAR respond against tumor cells.

(a) Expression of CARs (αerbB2-CD28ζ: 57.99±12.66, αerbB2-DAP10ζCD27: 66.38±22.90. N = 3± SEM, p = 0.54) and (b) expression of CD4/CD8 in transduced human T cells was determined by flow cytometry (representative plot of 3 experiments). (c) The ability of T cells to secrete cytokines in response to CAR ligation was determined by ELISA after overnight incubation with tumor cells (3 experiments pooled). (d) The ability of CARs to mediate T cell proliferation was determined by incorporation of tritiated thymidine following incubation of T cells with anti-c-myc antibody. (Representative of 2 experiments), ***p<0.0005. (e) The antigen-specific cytolytic potential of CAR-expressing T cells was determined in vitro using a ⁵¹Cr release assay using target cells expressing erbB2 or not expressing erbB2. (4 experiments pooled), *p = 0.0147.

Figure 5. Human T cells bearing the αerbB2DAP10ζCD27 CAR inhibit tumor growth in mice.

(a) CAR-expressing human T cells were injected on days 0, 1 and 2 intravenously into NOD-SCID mice injected subcutaneously on day 0 with 24JK-erbB2 sarcomas, and tumor growth monitored (9–10 mice/group). (b,c) Localization of T cells was determined on dissociated tumor and tissues on day 4 of tumor growth using flow cytometry (5 mice/group). (Error bars = SEM). *p<0.05, **p<0.005, NS = not significant.