Targeting fibroblast activation protein in tumor stroma with chimeric antigen receptor T cells can inhibit tumor growth and augment host immunity without severe toxicity

Abstract

The majority of chimeric antigen receptor (CAR) T-cell research has focused on attacking cancer cells. Here, we show that targeting the tumor-promoting, nontransformed stromal cells using CAR T cells may offer several advantages. We developed a retroviral CAR construct specific for the mouse fibroblast activation protein (FAP), comprising a single-chain Fv FAP [monoclonal antibody (mAb) 73.3] with the CD8α hinge and transmembrane regions, and the human CD3ζ and 4-1BB activation domains. The transduced muFAP-CAR mouse T cells secreted IFN-γ and killed FAP-expressing 3T3 target cells specifically. Adoptively transferred 73.3-FAP-CAR mouse T cells selectively reduced FAP(hi) stromal cells and inhibited the growth of multiple types of subcutaneously transplanted tumors in wild-type, but not FAP-null immune-competent syngeneic mice. The antitumor effects could be augmented by multiple injections of the CAR T cells, by using CAR T cells with a deficiency in diacylglycerol kinase, or by combination with a vaccine. A major mechanism of action of the muFAP-CAR T cells was the augmentation of the endogenous CD8(+) T-cell antitumor responses. Off-tumor toxicity in our models was minimal following muFAP-CAR T-cell therapy. In summary, inhibiting tumor growth by targeting tumor stroma with adoptively transferred CAR T cells directed to FAP can be safe and effective, suggesting that further clinical development of anti-human FAP-CAR is warranted.

Conflict of interest statement

All authors declare that there is no conflict of interest with any financial organization regarding the material discussed in the manuscript.

©2013 AACR.

Figures

Figure 1. Structure of FAP-CAR

Total RNA of 73.3 hybridoma was extracted, reverse transcribed, and the cDNA PCR amplified and inserted into a cloning vector to obtain the sequence of the variable domains of IgG heavy (A) and light (B) chains. The anti-muFAP CAR consists of the anti-muFAP scFv, CD8α hinge and transmembrane (TM) domain, plus 4-1BB and CD3ζ intracellular signaling domains (ICDs), and was cloned into the MigR1 retroviral vector for transduction of primary mouse T cells (C). A fully mouse FAP-CAR construct was also synthesized, which consists of the anti-muFAP scFv, CD8α hinge and CD28 transmembrane domain, plus CD28 and CD3ζ intracellular signaling domains of mouse origin, and was cloned into the MSGV retroviral vector to transduce primary mouse T cells (D). Two first-generation FAP-CAR constructs with a human CD3ζ (C) or a mouse CD3ζ (D) signaling domain were also generated.

Figure 2. Ex vivo assessment of mouse CAR T cells redirected against FAP and signaling in FAP-CAR T cells

(A) Retroviral transduced mouse T cells expressed GFP (MigR1) or GFP and anti-mFAP-CAR. (B) Up-regulation of CD69 on CAR (GFP)-positive CD8 and CD4 T cells was determined following stimulation with BSA- or FAP-coated beads for 18 hours. T cells were stimulated with anti-CD3/anti-CD28 beads as positive control. (C) FAP-CAR T cells were exposed to either BSA-or FAP-coated beads for 10 min. Cell lysates were then prepared and immunoblotted for phospho-ERK, phospho-AKT, and phospho-IKKα/β. Anti-CD3ε antibody was used as a positive control for T cell activation, and β-actin was immunoblotted for equal loading. To determine target-specific cytolytic activity (D) and IFNγ production (E) of FAP-CAR T cells, various Effector:Target ratios of MigR1 and FAP-CAR T cells were reacted with parental 3T3 or 3T3.FAP fibroblasts for 18 hours. * Denotes statistical significance between FAP-CAR-treated 3T3.FAP group versus the other 3 groups, p value < 0.05.

Figure 3. Anti-tumor activities of FAP-CAR T cells in mice bearing flank tumors

Syngeneic mice bearing (A) TC1, (B) LKR, (C) AE17.ova, (E) CT26, and (F) 4T1 tumors were injected intravenously with 10 million FAP-CAR or MigR1 T cells when the tumors reached ~100–150 mm3. Tumor measurements followed. (D) To test the target-specificity of FAP-CAR T cells, AE17.ova tumor cells were also injected into FAP-null C57BL/6 mice. FAP-CAR T cells were given 7 days later. * Denotes statistical significance between untreated, MigR1 and FAP-CAR-treated samples, p value < 0.05.

Figure 4. Enhanced therapeutic responses of FAP-CAR T cells

Mice with AE17.ova flank tumors were injected intravenously with FAP-CAR T cells when tumors were ~100 mm3. A. The overall efficacy of FAP-CAR T cells was enhanced when a second dose of FAP-CAR T cells was given a week later. The grey arrow indicates the injection time of the second dose of FAP-CAR T cells. B. Efficacy could also be enhanced after injection of FAP-CAR T cells lacking the negative intracellular regulator DGKζ. * Denotes statistical significance between untreated and FAP-CAR-treated samples, p value < 0.05. # Denotes statistical significance between single dose FAP-CAR treated group versus double dose group or DGKζ KO FAP-CAR treated group. (C) FAP-CAR T cells enhance efficacy of cancer vaccine. TC1 tumor cells were inoculated into the right flanks of C57BL/6 mice. When tumors reached 200 mm3, one dose of Ad.E7 (109 pfu) was given to the mice contralaterally to their flank tumors (black arrow). FAP-CAR T cells (10 million cells) were given 4 days later (gray arrow). Tumor measurements followed. The values are expressed as the mean ± SEM (n=5). * Denotes significant difference between untreated and the combo groups (p< 0.05).

Figure 5. Adaptive immune response plays a key role in FAP-CAR T cells-induced antitumor response

AE17.ova tumors were injected into both C57BL/6 (A) and NSG mice (B). When tumors reached ~100–125 mm3, a single dose (10 million) of FAP-CAR T cells were adoptively transferred through tail vein into mice. Tumor measurements followed. * Denotes statistical significance between untreated and FAP-CAR-treated samples, p value < 0.05. FAP-CAR induced infiltration of antigen-specific CD8 T cells into tumors. TC1 (C,D) and AE17.ova (E,F) tumors were harvested 8 days after adoptive transfer of FAP-CAR T cells in mice. Tumors were digested and made into single cell suspensions. Cells were stained with fluorochrome-conjugated tetramer loaded with E7- or SIINFEKEL(ova)-peptide, together with anti-CD8 antibody to determine percent tumor-specific CD8 T cells in tumors. * Denotes statistical significance between untreated, MigR1 and FAP-CAR-treated samples, p value < 0.05.

Figure 6. Activation and tumor-infiltration of endogenous T cells following treatment with FAP-CAR T cells

AE17.ova tumor-bearing mice were injected intravenously with 10 million FAP-CAR or MigR1 T cells when tumors reached ~100 mm3. At 3 and 8 days following adoptive transfer, tumors were harvested and digested to check for the frequencies of CD3+ tumor-infiltrating (A) total T cells 3 days post-transfer; (B) TNF-producing cells; (C) IFNγ-producing cells; (D) 4-1BB+ cells; or (E) CD69+ T cells; (F) total T cells 8 days post-transfer. * Denotes statistical significance between untreated, MigR1 and FAP-CAR-treated samples, p value < 0.05.