Chimeric Antigen Receptor T Cells Targeting Integrin αvβ6 Expressed on Cholangiocarcinoma Cells

Cholangiocarcinoma (CCA) is a lethal bile duct cancer that responds poorly to current standard treatments. A new therapeutic approach is, therefore, urgently needed. Adoptive T cell transfer using chimeric antigen receptor (CAR) T cells is a new therapeutic modality with demonstrated efficacy in hematologic malignancies. However, its efficacy against solid tumors is modest, and further intensive investigation continues. An important factor that influences the success of CAR T cell therapy is the selection of a target antigen that is highly expressed on cancer cells, but markedly less so in normal cells. Integrin αvβ6 is upregulated in several solid tumors, but is minimally expressed in normal epithelial cells, which suggests integrin αvβ6 as an attractive target antigen for CAR T cell immunotherapy in CCA. We investigated integrin αvβ6 expression in pathological tissue samples from patients with liver fluke-associated CCA. We then created CAR T cells targeting integrin αvβ6 and evaluated their anti-tumor activities against CCA cells. We found overexpression of the integrin αvβ6 protein in 23 of 30 (73.3%) CCA patient tissue samples. Significant association between high integrin αvβ6 expression and short survival time (p = 0.043) was also observed. Lentiviral constructs were engineered to encode CARs containing an integrin αvβ6-binding peptide (A20) derived from foot-and-mouth disease virus fused with a second-generation CD28/CD3ζ signaling domain (A20-2G CAR) or with a fourth-generation CD28/4-1BB/CD27/CD3ζ signaling domain (A20-4G CAR). The A20-2G and A20-4G CARs were highly expressed in primary human T cells transduced with the engineered lentiviruses, and they exhibited high levels of cytotoxicity against integrin αvβ6-positive CCA cells (p < 0.05). Interestingly, the A20-2G and A20-4G CAR T cells displayed anti-tumor function against integrin αvβ6-positive CCA tumor spheroids (p < 0.05). Upon specific antigen recognition, A20-4G CAR T cells produced a slightly lower level of IFN-γ, but exhibited higher proliferation than A20-2G CAR T cells. Thus, the A20-4G CAR T cells with lower level of cytokine production, but with higher proliferation represents a promising potential adoptive T cell therapy for integrin αvβ6-positive CCA.


INTRODUCTION
Cholangiocarcinoma (CCA) arises from the cancerous transformation of epithelial cells lining the bile duct. It is a relatively rare cancer, but its rates of both incidence and mortality are increasing worldwide (1). The highest incidence of CCA was found in the Northeast of Thailand where infection with an oncogenic liver fluke [Opisthorchis viverrini (OV)] is known to be a strong risk factor for CCA (2). Surgical resection is a curative treatment for CCA; however, only 20-40% of tumors are resectable, and the recurrence rate after surgery is high (3). For non-resectable patients, the standard first-line therapy is gemcitabine in combination with cisplatin. However, this therapeutic regimen achieves a 5-year overall survival rate of <5%, and the median overall survival is <1 year (4). Immunotherapy and targeted therapy for this difficult-totreat disease have been reported (5,6); however, the limited efficacy of these therapies highlights the need for an alternative treatment approach.
Generally, CCA and other cancers develop when transformed cells escape immune surveillance. Downregulation of MHC molecules that conceal cancerous cells from T cell recognition is one of many cancer immune escape mechanisms (7,8). To overcome this problem in cancer treatment, adoptive transfer of T cells expressing a chimeric antigen receptor (CAR) has been developed as a promising therapeutic approach. CARs are synthetic receptors that mimic natural T cell receptor function by combining a cancer antigen-binding domain with a T cell activating signaling domain. CAR T cells recognize cancer antigen in a direct, antibody-like fashion, which leads to the activation of intracellular signaling. As a result, CAR T cells kill cancerous cells in an MHC-independent manner. Different generations of CAR T cells have been developed by combining the intracellular part of T cell receptor (CD3ζ) and one or more co-stimulatory domains. Recently, three second-generation CAR (2G-CAR) T cells targeting CD19 for hematologic malignancies were approved by the U.S. Food and Drug Administration (USFDA), namely Kymriah (4-1BB/CD3), Yescarta, and Tecartus (CD28/CD3ζ). However, clinically successful CAR T cell therapies in patients with solid tumors have been limited, and studies to improve the efficacies of these therapies are intensively ongoing. Several research groups have designed other generations of CAR T cells by adding more co-stimulatory domains into the CAR molecule (9,10). Third-generation CAR (3G CAR) T cells consisting of CD28/CD137/CD3ζ (11,12) or CD28/CD27/CD3ζ (13) were created and tested. Fourth-generation CAR (4G CAR) T cells containing CD28/CD137/CD27/CD3ζ have also been produced and proven effective in the treatment of B cell leukemias (10,14,15).
An essential factor that influences the success of CAR T cell immunotherapy is the selection of a target antigen that is highly expressed on the surface of cancerous cells, but that is only minimally expressed on normal cells. Binding between the target antigen on cancerous cells and the extracellular antigen-binding domain of the CAR molecule leads to activation of CAR T cells to kill cancerous cells. An attractive potential target antigen in solid tumors is integrin αvβ6 because it is overexpressed in multiple epithelial malignancies, including pancreatic ductal adenocarcinoma (16), ovarian cancer (17), head and neck squamous cell carcinoma (18,19), breast cancer (20), and CCA (21,22). Integrin αvβ6 is also a target for diagnostic imaging and anti-cancer therapies (19). Several integrin αvβ6-specific therapeutic agents have been studied in clinical trials (23)(24)(25). Notably, the specificity of integrin αvβ6 immunohistochemistry for CCA (100%) surpassed all other tested markers, and the sensitivity was very similar to that of cytokeratin 7 (CK7) (86 vs. 90%) (26). Integrin αvβ6 is, thus, a potential target antigen for development of CAR T cell therapy for CCA.
A single-chain variable fragment (scFv) derived from a monoclonal antibody is most commonly used as the extracellular cancer antigen-binding domain within tumor-specific CAR molecules. However, antigens on cancer cells can also be specifically bound by the peptide with high specificity for the desired target (27). Second-generation (2G) of CAR T cells targeting integrin αvβ6 were recently tested in multiple solid tumors, including pancreatic, breast, and ovarian cancers (28). The antigen-binding domain of this 2G CAR was a peptide containing 20-mers of amino acids (A20) that was derived from the viral capsid protein 1 (VP1) of foot-and-mouth disease virus (FMDV), and it bound explicitly with high affinity (<1 nM) only to integrin αvβ6 (29). However, the effect of CAR T cells targeting integrin αvβ6 in the killing of CCA cells is still unknown.
Expression of integrin αvβ6 has not been previously characterized in CCA tumors or cell lines from patients with liver fluke-associated CCA. To evaluate its potential as a therapeutic target antigen in CCA, we analyzed integrin αvβ6 expression in pathological tissue samples and CCA tumor cell lines. We then compared the anti-tumor activity of A20-2G CAR T cells and A20-4G CAR T cells, containing the CD28/CD3ζ and CD28/4-1BB/CD27/CD3ζ signaling domains, respectively. Both CARs demonstrated integrin αvβ6-dependent anti-tumor activity in models of CCA, but the same anti-tumor activity was not observed in non-transduced (NT) T cells. The results of this study provide proof-of-principle for the use of integrin αvβ6-specific CAR T cells in adoptive T cell immunotherapy of CCA.

Ethical Approval
In this study, paraffin-embedded tissues from patients with CCA were collected at Srinagarind Hospital, Khon Kaen University, Khon Kaen, Thailand. The protocols for collection of the tissue samples and the experimental studies were approved by the Ethics Committee for Human Research, Khon Kaen University (No. HE591063). Written informed consent was obtained from each patient before the study. Peripheral blood mononuclear cells (PBMCs) were obtained from healthy volunteers who had provided written informed consent in accordance with a protocol approved by the Siriraj Institutional Review Board of the Faculty of Medicine Siriraj Hospital, Mahidol University, Bangkok, Thailand (COA No. Si 829/2020).

Immunohistochemistry
Integrin αvβ6 was detected on the formalin-fixed, paraffinembedded tissue sections using immunohistochemistry (IHC). Specifically, 4-µm-thick paraffin sections were deparaffinized in xylene, and then rehydrated in gradient ethanol. These tissue sections were antigen-retrieved in 0.1% trypsin, and endogenous peroxidases were blocked in 3% H 2 O 2 . After blocking of nonspecific binding with 5% skim milk, the tissue sections were incubated overnight with an anti-αvβ6 monoclonal antibody (dilution 1:100, Clone 6.2A1; Biogen Inc., Cambridge, MA, USA) that was generously provided by Dr. Shelia M. Violette. The treated tissue sections were then reacted with EnVision Kit/Horseradish Peroxidase (HRP) TM (Agilent Technologies, Santa Clara, CA, USA) at room temperature (RT) for 1 h. Immunoreactive integrin αvβ6 from those tissue sections were developed using 3, 3 ′ -diaminobenzidine (DAB) solution. The tissue sections were subsequently counterstained with Mayer's Hematoxylin (Sigma-Aldrich Corporation, St. Louis, MO. USA), mounted, and observed under a light microscope (Nikon Eclipse Ti2; Nikon Instruments, Tokyo, Japan). The IHC-stained tissues were evaluated independently by two investigators who had no prior knowledge of patient clinical or survival data. The frequency of integrin αvβ6 was semiquantitatively scored according to the percentage of positive cells, as follows: 0%, negative; 1-25%, +1; 26-50%, +2; and, 50%, +3. The intensity of protein staining was scored, as follows: weak, 1; moderate, 2; and, strong, 3. The expression of integrin αvβ6 was evaluated using H-score by multiplying the frequencies and intensities. The patients were then categorized into two groups according to the median IHC score. Log-rank test was used to analyze the difference between these two groups relative to overall survival.

Immunofluorescence Staining
Cells were cultured on glass coverslips and fixed with 4% paraformaldehyde for 15 min on ice. The cells were then washed with phosphate-buffered saline (PBS) before incubation with 5% bovine serum albumin (BSA) blocking solution for 30 min. The cells were then incubated with a 1:100 dilution of anti-β6 polyclonal antibody (clone ITBG6; Thermo Fisher Scientific, Waltham, MA, USA) at 4 • C overnight. After washing, the cells were incubated with anti-mouse Alexa Fluor R 488-labeled (clone A21206; Thermo Fisher Scientific), and their nuclei were counterstained with Hoechst 33342 dye for 1 h at RT. The coverslips were mounted on a glass slide and immunofluorescence signals were visualized using a Zeiss LSM 800 confocal microscope (Carl Zeiss Microscopy, Jena, Germany).

Construction of Chimeric Antigen Receptor
Second-generation CAR T cells containing A20 peptide for targeting integrin αvβ6 (A20-2G) CAR T cells were created as previously described (28). Briefly, A20 peptide derived from FMDV was placed downstream of a CD124 signal sequence, followed by human c-myc peptide tag (EQKLISEEDL), as shown in Figure 3A. The DNA fragment encoding the required parts was synthesized by Integrated DNA Technologies (Coralville, IA, USA). The A20 codon-optimized gene was sub-cloned into self-inactivating lentivirus vectors (pCDH) containing expression cassettes encoding CD8 short hinge, a CD28 transmembrane domain, and the CD28/CD3ζ (A20-2G) or CD28/4-1BB/CD27/CD3ζ (A20-4G) signaling domains. Transgene expression is driven by the elongation factor-1α (EF-1α) promoter. Plasmid DNA was isolated using a Midiprep Kit (Qiagen, Hilden, Germany) and sequences were verified by DNA sequencing.

Lentivirus Production and T Cell Transduction
Lenti-X 293T cells were transfected with the A20-2G or A20-4G plasmid and two packaging plasmids (psPAX2 and pMD.2G) at a ratio of 5:3:1 using calcium phosphate transfection method. The supernatant was collected at 48 and 72 h post-transfection and filtered through a 0.45 µm filter unit (Merck Millipore) to remove cell debris, followed by concentration at 20,000 × g for 2 h at 4 • C (Sorvall RC-6 Plus Centrifuge; Thermo Fisher Scientific). Virus titer was determined using a qPCR Lentiviral Titration Kit (ABM, Richmond, BC, Canada) according to the manufacturer's instructions.

Immunoblot Analysis
Immunoblotting was used to detect the expression of A20-2G or A20-4G CAR construct following transfection in Lenti-X 293T cells. In brief, the cells were lysed in radioimmunoprecipitation assay (RIPA) lysis buffer. Cell lysate was then resolved by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and subsequently transferred onto a nitrocellulose membrane. The membrane was blocked with 5% skim-milk in Tris-buffered saline (TBS) and 0.1% Tween-20 (TBS-T), and then detected with anti-CD3ζ (clone sc-166435) and anti-GAPDH (clone sc-32233) from Santa Cruz Biotechnology (Dallas, TX, USA). The membrane was incubated with HRP-conjugated secondary antibody (Invitrogen) and the immunoreaction was developed using chemiluminescence reagents (SuperSignal R West Pico Substrate; Thermo Fisher Scientific). The signal from the reaction was captured on X-ray film and quantified using ImageJ program (National Institutes of Health, Bethesda, MD, USA). The expression level of GAPDH was used as loading control.

Cytotoxicity Assays
The monolayer of target cells (1 × 10 4 ) was co-cultured with A20-2G CAR T cells or A20-4G CAR T cells or NT T cells at three different effector to target (E:T) ratios (5:1, 2.5:1, and 1.25:1) for 24-48 h. After removal of the CAR T cells, 100 µl of crystal violet fixing/staining solution was added to each well and incubated for 20 min. The plates were washed and the cell-bound dye was dissolved in methanol. The absorbance was measured at a wavelength of 595 nm using a Sunrise TM Absorbance Microplate Reader (Magellan TM data analysis software version 6.6.0.1; Tecan, Männedorf, Switzerland). Cytotoxicity was calculated using the following formula: [1-(absorbance of monolayer culture with T cells/absorbance of monolayer culture alone)] × 100%.
In addition, cytotoxicity assay was conducted by using threedimensional (3D) spheroid model. In brief, a total number of 2 × 10 3 cancer cells were firstly stained with CellTracker TM Green CMFDA (5-chloromethylfluorescein diacetate) Dye (Thermo Fisher Scientific, Waltham, MA) and then seeded into an ultralow attachment 96-well round-bottomed plate (Corning, NY, USA) containing 2.5% Corning matrigel matrix (Corning, NY, USA). To generate a single spheroid, the plate was centrifuged at 1,000 × g at 4 • C for 10 min and then cultured for 48 h. T cells in culture medium containing 1 µg/ml propidium iodide (PI) were added to the spheroid at E:T ratio of 5:1. After coculturing for 3 days, dead cancer cells, which were stained with PI, were analyzed by a confocal microscope (Nikon Instruments Inc., Melville, NY, USA). Quantification of mean fluorescence intensity (MFI) was conducted by using NIS-Elements software.
Cytotoxicity was calculated by following formula: [(experimental MFI-spontaneous MFI)/(maximum MFI-spontaneous MFI)] × 100. Experimental and spontaneous MFIs were MFI of sample spheroids when co-cultured with or without CAR T cells, respectively. Maximum MFI was MFI of spheroid treated with 0.1% Triton-X 100.

Intracellular Cytokine Staining and T Cell Proliferation Assay
CAR T cells were cultured with target cells in media containing Brefeldin A (BioLegend) at an E:T ratio of 5:1 for 6 h in 5% CO 2 at 37 • C. The CAR T cells were harvested and the cell surface markers were stained with anti-CD3-FITC and anti-CD8-APC antibodies (Immunotools). The cells were then fixed with 4% paraformaldehyde for 15 min. Intracellular cytokine was stained by incubation with anti-IFN-γ-PE (Immunotools) antibody in the presence of 0.5% saponin permeabilization agent on ice for 30 min. The cells were then subjected to flow cytometry to analyze cytokine production levels.
T cell proliferation was assessed by tracking cells labeled with carboxyfluorescein succinimidyl ester (CFSE). Briefly, 1 × 10 5 of T cells were stained with 1 µM CFSE for 10 min at 37 • C. After washing twice with culture media, CFSE-labeled T cells were cocultured with target cell monolayers at an E:T ratio of 5:1 in AIM-V medium supplemented with 5% human serum. On day 3 after co-culturing, CFSE dilution was measured by flow cytometry to estimate proliferation of total CAR T cells. Note that no exogenous cytokines were added during the proliferation assay.

Statistical Analysis
GraphPad Prism 7 software (GraphPad Software, San Diego, CA, USA) was used for statistical analysis. Data are presented as mean ± standard deviation (SD) or standard error of the mean (SEM). For comparison between two groups, a two-tailed t-test was used. For comparisons among three or more groups, one-way analysis of variance (ANOVA) with Bonferroni's post hoc test was used. A p-value <0.05 was considered statistically significant.

Expression of Integrin αvβ6 in Human CCA Tissues and Cell Lines
Immunohistochemistry (IHC) staining was performed to examine the expression of integrin αvβ6 in liver fluke-associated CCA tumors and cell lines, including KKU055, KKU100 and KKU213A. The results revealed the presence of integrin αvβ6 in human CCA tissues, whereas the protein was virtually non-existent all non-tumorous tissues (Figure 1). Low to high intensities of staining signals were observed in CCA tissues (Figures 1B,C). The stained protein was located at both the cell membrane and within the cytoplasm of CCA cells. Twentythree out of 30 (73.3%) human CCA tissue samples showed positive staining, of which 13 tumors (43.3%) showed high level expression. The expression levels of the protein were calculated as H-scores ( Figure 1D). Cumulative survival of patients with low/negative and high expression levels of integrin αvβ6 was compared. All patients with CCA had died by the end of the follow-up period. Analysis of survival times showed a significant difference between the CCA patients who had low/negative and those who had high integrin αvβ6 expression levels. The median survival time was 308.0 days (95% confidence interval [CI]: 178.9-437.1) in the patients with low/negative integrin αvβ6 expression level, and 155.0 days (95% CI: 77.5-232.5) in those with high integrin αvβ6 expression level (log-rank test; p = 0.043) (Figure 1E).

Expression of Chimeric Antigen Receptor Targeting Integrin αvβ6 in Lenti-X 293T and Human Primary T Cells
Second-and fourth-generation CAR targeted against integrin αvβ6 (A20-2G CAR and A20-4G CAR) ( Figure 3A) were expressed using lentiviral vector. The expression of A20-2G CAR and A20-4G CAR in Lenti-X 293T cells was examined by immunoblotting under reducing conditions using anti-CD3ζ antibody. The results showed that A20-2G CAR and A20-4G CAR were expressed at the predicted sizes of 32 and 43 kDa, respectively ( Figure 3B). To generate A20-2G CAR and A20-4G CAR T cells, primary human lymphocytes isolated from a healthy donor were transduced with lentiviruses carrying either A20-2G CAR or A20-4G CAR construct. A representative flow cytometry profile after lentiviral transduction of the lymphocytes is shown in Figure 3C. The median CAR expression of A20-2G CAR was 71.5±17.5%, and A20-4G CAR was 69.6±19.1%, as examined on day five post-transduction ( Figure 3D). The phenotypes of CAR T cells were analyzed by flow cytometry, which showed that within the CD3 + population, there were significantly more cytotoxic CD8 + T cells than helper CD4 + T cells in the groups of NT T cells (61.8±11.28% and 27.7±6.3%, p=0.038), A20-2G CAR T cells (72.7±7.0% and 25.3±8.0%, p=0.004), and A20-4G CAR T cells (73.7±7.0% and 24.6±7.5%, p=0.003) (Figure 3E). Furthermore, the generated CAR T cells were enriched in CD45RA − CD62L + central memory (T CM ) and CD45RA + CD62L + naïve T cells (Figure 3F). The expression of these markers did not significantly differ between the NT T cells and the A20-2G or A20-4G CAR T cells.
Anti-tumor Activities of A20-2G and A20-4G CAR T Cells Against Integrin αvβ6-Expressing Cells The anti-tumor activity of A20-2G and A20-4G CAR T cells against integrin αvβ6-expressing cells was tested by co-culturing these CAR T cells with target cell lines expressing different levels of integrin αvβ6 at the indicated E:T ratio (Figure 4), and using NT cells as control T cells. After co-culture and removal of effector cells, crystal violet solution was added to stain the remaining target cells (Figure 4A). The results revealed that both A20-2G and A20-4G CAR T cells exhibited strong cytotoxic effects in a dose-dependent manner, and both had higher cytotoxic effects than NT T cells. At an E:T ratio of 5:1 and an incubation time of 24 h, the killing activity of A20-2G and A20-4G CAR T cells on integrin αvβ6-negative A375.puro cells was very low (Figure 4B), and the killing activity of A20-2G and A20-4G CAR T cells on integrin αvβ6-positive A375.β6 cells was as high as 65.9±5.06% and 69.4±10.1%, respectively, compared to the killing activity of NT T cells on A375.β6 cells, which was 25.7±6.9% (p<0.05) (Figure 4C). Killing effects were also observed when A20-2G and A20-4G CAR T cells were co-cultured with KKU055 cells (34.6±2.9% and 52.8±7.5%, respectively) compared to the killing activity of NT T cells on KKU055 cells, which was 21.3±0.8% (p<0.05) (Figure 4D). At an E:T ratio of 5:1 and an incubation time of 24 h, no significant cytotoxicity was observed in assays with the KKU100 and KKU213A cells (data not shown). However, after co-culturing for 48 h, the killing activity of A20-2G and A20-4G CAR T cells on KKU100 cells was 62.4±4.7% and 59.4±3.0%, respectively, compared to the killing activity of NT T cells on KKU100 cells, which was 35.1±0.4% (p<0.05) ( Figure 4E). Lastly, the killing activity of A20-2G and A20-4G CAR T cells on KKU213A cells was 50.9±6.1% and 54.5±6.3%, respectively, compared to the killing activity of NT T cells on KKU213 cells, which was 25.6±4.5% (p<0.05) (Figure 4F).

Production of Interferon-γ in and Proliferation of A20-2G and A20-4G CAR T Cells After Co-culturing With Integrin αvβ6-Expressing Target Cells
Production of interferon-γ (IFN-γ) in A20-2G and A20-4G CAR T cells was examined after these CAR T cells were co-cultured with integrin αvβ6-negative A375.puro or integrin αvβ6-positive A375.β6 cells at an E:T ratio of 5:1 and an incubation time of 6 h. As positive control, the CAR T cells were activated using phorbol-12-myristate-13-acetate (PMA) and ionomycin (IONO) in the presence of brefeldin A. In all cases, production of IFN-γ in CD8 + T cell population was examined by intracellular cytokine staining and analyzed by flow cytometry (Figure 6A). When NT T cells, A20-2G CAR T cells, and A20-4G CAR T cells were co-cultured with integrin αvβ6-negative A375.puro cells, the production of IFN-γ was minimal ( Figure 6B). When these cells were co-cultured with integrin αvβ6-positive A375.β6 cells, the production of IFN-γ in A20-2G CAR T cells was clearly increased compared to those co-cultured with NT T cells (4.5±1.9 % vs. 1.1±0.7%; p=0.0156). However, when A20-4G CAR T cells were co-cultured with integrin αvβ6-positive A375.β6 cells, the production of IFN-γ was not significantly increased (1.3±9.5% vs. 1.1±0.7%; p>0.99), (Figure 6B).

DISCUSSION
A more effective treatment for patients with advanced unresectable/metastatic CCA is urgently needed. Adoptive T cell therapy using CAR T cells has provided promising outcomes against hematological malignancies (33), and its potential for treatment of solid cancers is being extensively investigated. The results of the present study provide evidence that A20-2G and A20-4G CAR T cells targeting integrin αvβ6 protein are efficient in killing CCA cell lines, which indicates their potential for treatment of CCA.
A few candidate target antigens are currently being investigated in clinical trials of CCA (34). The integrin αvβ6 protein is overexpressed in several solid tumors, but it is only minimally expressed in normal tissues (29). Here, we show that integrin αvβ6 represents a novel therapeutic target antigen for CAR T cell immunotherapy in patients with CCA. We initially examined the expression of the integrin αvβ6 protein in liver fluke-associated CCA tissues by IHC staining. We found that 73.3% of tumor samples from the Thai patients with this type of CCA had increased expression of integrin αvβ6 (Figure 1). Expression was highly specific to CCA cells compared to adjacent non-malignant biliary epithelia that had an undetectable level of integrin αvβ6. These results are consistent with those observed in CCA tissues from other ethnic groups, including Swiss (26), Japanese (21), and Chinese populations (22). Notably, the survival time of CCA patients who had low/negative integrin αvβ6 protein expression was significantly longer than the survival time of those who had high integrin αvβ6 expression (Figures 1D,E). Integrin αvβ6 has been reported to promote resistance of CCA cells to cisplatin-induced apoptosis (22), which indicates that it should be targeted using other therapeutic approaches, such as immunotherapy. In this study, we also reported the expression of integrin αvβ6 on the surface of patient-derived CCA cell lines, including KKU055, KKU100, and KKU213A cells (Figure 2). We then generated CAR T cells targeting integrin αvβ6 and tested their anti-tumor function in these cell lines. The expression level of the target antigen was reported to affect CAR T cell functionality (35). Thus, we selected a panel of CCA cell lines with different expression levels of integrin αvβ6 to demonstrate the effectiveness of CAR T cells specific to this target antigen.
In previous clinical trials involving patients with B cell malignancy, CAR T cells containing either CD28 or 4-1BB presented different properties, but they showed a similar antitumor response. CD28-based CAR T cells were rapidly The schematic map demonstrates the A20-2G and A20-4G CAR constructs. The sequence of A20 (integrin αvβ6-targeting ligand) was cloned in-frame into lentiviral vectors linked with the sequences of CD8 hinge, CD28 transmembrane domain (TM), costimulatory domain of CD28/CD3ζ (A20-2G), or CD28/4-1BB/CD27/CD3ζ (A20-4G). (B) Expression of A20-2G CAR and A20-4G CAR proteins in Lenti-X 293T cells was detected by immunoblotting analysis using anti-CD3ζ antibody. The data shown represent the mean of three experiments, all of which showed similar results. (C) Representative histogram shows expression of A20-2G and A20-4G CAR on the surface of T cells as examined by anti-cMyc FITC-conjugated antibody and flow cytometric analysis. (D) Expression of A20-2G or A20-4G CAR on transduced T cells was examined and summarized from eight healthy donors as mean ± standard deviation (SD). Non-transduced (NT) T cells were used as control. (E) Representative flow cytometric gating of CD3 + CD56 − T cell population (upper) and summarized data (lower). The phenotypes of CAR T cells were cytotoxic CD8 + T cells and helper CD4 + T cells. Flow cytometric gating was based on cells stained with isotype-match control antibody. (F) The subgroups of A20-2G or A20-4G CAR T cells consisted of CD45RA − CD62L + central memory (T CM ) and CD45RA + CD62L + naïve T cells. These data were derived from four healthy donors (ns, non-significant, *p<0.05, **p<0.01, ***p<0.001).
activated and their cytolytic activities were enhanced; however, their persistence was short (<3 months) (36). In contrast, 4-1BBbased CAR T cells demonstrated slow response and exhaustion, but their survival was longer (>1 year) (37). The findings of previous studies suggested that complete treatment response required persistence of CAR T cells (37,38). Accordingly, a CD27 signaling domain was combined in our 4G CAR design to further support the activation, proliferation, and survival of CAR T cells in vitro and in vivo (39). In our study, we generated the A20-2G CAR construct containing A20/CD28/CD3ζ to be a control, while the A20-4G CAR construct comprising A20/CD28/4-1BB/CD27/CD3ζ (Figure 3A) was designed to combine the favorable properties of CD28, 4-1BB, and CD27. These two constructs were expressed using a self-inactivating lentiviral vector system. The A20-2G and A20-4G CAR proteins of the predicted size could be detected in Lenti-X 293T cells ( Figure 2B). Moreover, A20-2G and A20-4G CAR T cells were successfully generated using T cells isolated from eight healthy donors. The expressions of the A20-2G and A20-4G CARs were 71.5±17.5% and 69.6±19.1%, respectively (Figures 3C,D). The final products contained cytotoxic T cells (CD3 + CD56 − CD8 + ) with a CD45RA − CD62L + phenotype as major populations (Figures 3E,F). This phenotype was reported to support cancer immune surveillance, long-term expansion, and persistence in vivo (40). However, the cell phenotypes did not differ between A20-2G and A20-4G CAR T cells compared to NT T cell control, which suggests that their phenotypes may depend on the manufacturing process that we undertook using PHA-L activation and T cell culture in media containing IL-2, IL-7, and IL-15, which were reported to promote cytotoxic T cell memory phenotype (40,41).
Our data demonstrates that while the A20-2G and A20-4G CAR T cells had minimal cytolytic activity against αvβ6-negative A375.puro cells, the two CAR T cell populations were clearly able to kill αvβ6-positive A375.β6 cells and CCA cells in an E:T ratio-dependent manner (Figure 4), which indicates their specific killing ability. Additionally, both A20-2G and A20-4G CAR T cells killed KKU055 cells within 24 h, which was faster than the time it took for them to kill KKU100 and KKU213A cells within 48 h. It should be noted that high levels of integrin αvβ6 expression on CCA cells did not correlate with high levels of effector activities of the two CAR T cells. The most likely Representative images on the right show bright filed of a spheroid formed by 2 × 10 3 cells for 48 h and a fluorescence image after co-culturing CAR T cells with the spheroid. (B-F) Representative images showing cytolytic activities of A20-2G or A20-4G CAR T cells compared to NT cells at an effector to target (E:T) ratio of 5:1 on the spheroids of A375.puro, A375.β6, and CCA (KKU055, KKU100, KKU213A) at 72 h of co-culturing. Dead of green fluorescence-labeled cells were visualized by PI uptake (Red). (G) Histogram shows cytotoxic activities of NT cells, A20-2G CAR T cells, and A20-4G CAR T cells on tumor cells determined by PI incorporation of dead cells within spheroids. Note that lower level of the percentage of cytotoxicity of A20-4G CAR T cells for KKU213A resulted from more CCA cell lysis. The bar graphs represent mean ± standard error of the mean (SEM) (n = 3; *p<0.05, **p<0.01). explanation is that CCA cells expressing the integrin αvβ6 protein may also express immune checkpoint molecules to suppress CAR T cell function. The upregulation of PD-L1 in CCA cells has been reported (42), and it can induce T cell exhaustion via the engagement of PD-1 on CD8 + T cells (43,44). A combination of CAR T cells and immune checkpoint inhibitor may improve effector functions of CAR T cells for treatment of CCA (45).
The traditional two-dimensional (2D) culture model based on the growth and proliferation of monolayer cells might not represent the condition with the presence of cell-cell and cellextracellular matrix interactions. Thus, three-dimensional (3D) CCA spheroids that appeared like solid tumor were generated to evaluate anti-tumor activities of A20-2G and A20-4G CAR T cells. The results showed that the two CAR T cells could infiltrate into the spheroid and displayed potent anti-tumor activities, as demonstrated by dead cancer cells in the spheroids stained by propidium iodide (PI) (Figure 5). In the CCA patients, T cell infiltration in the CCA tissue is a positive outcome predictor (46). A study using 3D culture system revealed that gene expression in this culture system was much closer to clinical expression profiles than those observed in the 2D culture system (47), indicating the suitability of the 3D culture system for preclinical studies. After co-culturing with αvβ6-positive cells, A20-2G CAR T cells produced greater levels of IFN-γ than A20-4G CAR T cells (4.5±1.9% vs. 1.1±0.7%; p=0.0156) (Figures 6A,B). This may be an advantage of A20-4G CAR T cells since clinical studies of the 4G CAR T cells targeting CD19 found that low levels of IFN-γ might be beneficial to limit CAR T cellmediated cytokine release syndrome (CRS) (10,15), which was often observed in patients who received CAR T cell therapy (48,49). Furthermore, A20-4G CAR T cells showed a higher proliferation rate than A20-2G CAR T cells (74.6±5.9% vs. 56.5±5.5%, p=0.0175) (Figures 6C,D). This higher proliferation rate may result from the incorporation of CD27 into the A20-4G CAR construct because its signaling is known to typically promoting T cell proliferation (38). Our data may also suggest the addition of 4-1BB/CD27 into the A20-4G CAR construct to be more effective than the addition of only CD28 into the A20-2G CAR construct. Thus, A20-4G CAR T cells possibly offer better benefits than A20-2G CAR T cells for CCA treatment because they may cause less severe CRS response and they have a higher proliferation rate.
A previous study reported the use of CAR T cell immunotherapy to treat a patient with advanced unresectable/metastatic CCA that proved resistant to chemotherapy and radiotherapy (34). Sequential infusions of CAR T cell therapies targeted against EGFR and CD133 induced partial response (PR) for 8.5 months and 4.5 months, respectively. That study showed that CAR T cell therapy targeting two or more antigens is feasible for resolving the problem of tumor heterogeneity in CCA (34). Thus, the generation of CAR T cells specific to integrin αvβ6 and other tumor-associated antigens for treatment of CCA warrants further study. Moreover, since CCA is characterized as having desmoplastic stroma and an immune suppressive tumor microenvironment (TME), the combination of CAR T cell therapy with other treatment modalities, such as chemotherapy using gemcitabine/cisplatin drugs (5), immune checkpoint blockade (45), and/or FGFR inhibitor (6), may overcome immune escape mechanisms of CCA.
Here, we report integrin αvβ6, which is upregulated in CCA tissues, to be a promising target antigen for adoptive T cell therapy of CCA. A20-2G and A20-4G CAR T cells targeting integrin αvβ6 were successfully generated, and both were found to effectively kill αvβ6-positive CCA cells in both monolayer cell and spheroid culture systems. The A20-4G CAR T cells were found to be superior to the A20-2G CAR T cells concerning their higher proliferative potential and lower cytokine production. Thus, the A20-4G CAR T cells warrant further study for their therapeutic potential against CCA.

DATA AVAILABILITY STATEMENT
The original contributions presented in the study are included in the article/supplementary material, further inquiries can be directed to the corresponding author/s.

ETHICS STATEMENT
The studies involving human participants were reviewed and approved by Siriraj Institutional Review Board of the Faculty of Medicine Siriraj Hospital, Mahidol University, Bangkok, Thailand. The patients/participants provided their written informed consent to participate in this study.

AUTHOR CONTRIBUTIONS
NP designed and performed experiments, analyzed data, interpreted results, and prepared manuscript. CS and KS partly performed experiments and analyzed data. MJ and PY conceptualized, managed, and supervised the study. NP, CS, KS, TC, JS, SW, JM, MJ, and PY provided materials and reagents, designed experiments, interpreted results, and edited the manuscript. All authors read and approved the submitted version.