1. IntroductionEsophageal carcinoma (EC) is a type of malignant neoplasm of the digestive tract originating from the esophageal epithelium. EC is one of the most common cancers in China, with obvious regional characteristics, and it is predicted that approximately 324,422 people could be newly diagnosed with EC and 301,135 people could die from the disease in China [1]. The principal histological types of EC are esophageal squamous cell carcinoma (ESCC) and esophageal adenocarcinoma (EAC). In China, more than 90% of patients with esophageal cancer have ESCC. Patients with early-stage ESCC lack specific symptoms, and thus are mostly diagnosed at the middle or late stages. Few therapeutic options and a poor prognosis of ESCC account for the low 5-year survival rate of patients, which is ~30% [2].Monoclonal antibodies targeting programmed death-1 (PD-1) or PD-ligand 1 (PD-L1) are representative immune checkpoint inhibitors that have shown good therapeutic effects for EC in clinical trials [3,4,5]. However, these immunotherapeutic approaches produce durable responses in only a subset of patients with ESCC, many of which have primary or acquired resistance [6,7]. The mechanisms of resistance to PD-1/PD-L1 inhibitors include the loss of first and co-stimulation signals, major histocompatibility complex dysfunction, irreversible T cell exhaustion, and an immunosuppressive tumor microenvironment [7,8,9,10]. Thus, there is still a large demand for effective ESCC therapies, and their development is the focus of basic and clinical research.In the past few years, immunotherapy with chimeric antigen receptor (CAR)-expressing T cells has provided significant positive effects for the treatment of hematologic malignancies and has emerged as one of the most promising therapeutic methods in cancer [11,12,13,14]. Therefore, the exploration of CAR-T cell immunotherapy application to ESCC may provide new avenues for the development of effective treatment approaches for this disease.Our ESCC whole genome sequencing (WGS) and RNA sequencing (RNA-Seq) data revealed that epidermal growth factor receptor (EGFR) is one of the most amplification genes and it is highly expressed in ESCC, and it has low or no expression in normal mucosal tissues [15]. Previous studies have also indicated that ~80% [16,17,18,19] of patients with ESCC show high levels of EGFR expression. Therefore, we speculated that EGFR could be an excellent target for CAR-T cell therapy in ESCC. EGFR is overexpressed in various epithelial tumors, and its abnormal activation is closely correlated with cancer occurrence and development [20,21]. According to the literature, CAR-T cells targeting EGFR have been explored in non-small cell lung [22,23] and biliary tract [24], triple negative breast cancers [25,26], and glioblastoma [27,28], and some of these results have already been tested in clinical trials. Thus, Feng et al. used EGFR-targeting CAR-T cells to treat patients with advanced and refractory non-small cell lung cancer with a strong positivity for EGFR expression (more than 50%), and the results indicate that 7 of the 11 enrolled patients were evaluable, including two with significant tumor reduction and five with stable disease [23]. In another phase I clinical trial, EGFR-targeting CAR-T cells were used for patients with recurrent/metastasizing biliary tract carcinoma with a strong EGFR expression, and the findings indicated that among the 17 evaluable patients (of the total 19), 1 had 22 months of complete remission and 10 achieved stable disease [24]. Overall, these results indicate that anti-EGFR CAR-T-cell-based therapy may have a high clinical potential in different cancer types.The aim of this study was to test the feasibility of EGFR-targeting CAR-T cells in the treatment of ESCC and to provide a theoretical basis for preclinical research. Our ESCC whole-genome sequencing (WGS) and RNA-seq data revealed high levels of EGFR amplification and expression in ESCC compared with paracancerous tissues. To target EGFR-expressing ESCC cells, we used the second-generation CAR structure to constructe anti-EGFR CAR in this study because it has been reported that the second-generation CAR structure has a better antitumor effect [29]. We constructed five anti-EGFR CARs with a single chain fragment variable (scFv) derived from EGFR antibodies and used them to transfect T lymphocytes, which were then compared for their in vitro and in vivo anticancer activities. The results showed that CAR1-T, CAR2-T, and CAR4-T cells had a high cytotoxicity against ESCC in vitro; among them, CAR1-T and CAR2-T were also able to clear ESCC in vivo. The successful preparation and functional identification of anti-EGFR CAR-T cells targeting ESCC will lay the foundation for using CAR-T cell-based immunotherapy to treat ESCC. 2. Materials and Methods 2.1. Western Blotting

ESCC cells were lysed on ice for 30 min using a RIPA buffer supplemented with a Protease Inhibitor Cocktail. After measuring the total protein concentration, 50 μg of protein was subjected to SDS-PAGE (5% stacking/8% separating gels) and then transferred to nitrocellulose membranes, which were blocked with 2% BSA for 2 h at 16–24 °C, and then incubated with an anti-EGFR antibody (Cell Signaling Technology, Danvers, MA, USA) overnight at 4 °C. The membranes were then incubated with horseradish peroxidase-labeled secondary antibodies (Jackson ImmunoResearch, West Grove, PA, USA), and proteins of interest were detected using chemiluminescence reagents (ThermoFisher, Waltham, MA, USA). Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) (Proteintech, Wuhan, China) was used as the loading control.

2.2. Cell Lines

HEK293T cells were kindly gifted from Dr. H. Wang and human ESCC cell lines KYSE30, KYSE150, KYSE180, KYSE450, and TE1 were kindly supplied by Dr. Y. Shimada. HEK293T and ESCC cells were cultured in DMEM (GIBCO) and RPMI 1640 (GIBCO) medium, respectively, supplemented with 10% (v/v) fetal bovine serum (FBS) and 100 U/mL penicillin/streptomycin (GIBCO). All of the cells were free of mycoplasma and maintained at 37 °C and 5% CO2.

2.3. Lentivirus Production

HEK293T cells were seeded at 7.0 × 106 cells per 10 cm dish, cultured for 15–18 h, and transfected at ~90% confluence using Lipo3000 (ThermoFisher, Waltham, MA, USA, L3000015) according to the manufacturer’s protocol. Virus-containing supernatants were collected 48 and 72 h post-transfection and filtered. Lentiviral particles were concentrated by centrifugation at 4000 rcf for 1 h at −4 °C in virus concentration tubes (Merck Millipore Ultracel-100 K, UFC910096) and were stored at −80 °C.

2.4. Human CD3+ T Cell Enrichment, Activation, and Multiplication

hPBMCs were obtained from human peripheral blood provided by Shanghai Liquan Hospital and hPBMCs cell isolation service provided by Milestone Biotechnologies. CD3+ T cells were enriched according to the instructions of the T Cell Enrichment Kit (Stemcell Technologies, Vancouver, BC, Canada, 19051), activated with anti-CD3/anti-CD28 Dynabeads (ThermoFisher, Waltham, MA, USA, 11131D) added at a ratio of 2:1, and proliferated. The cells were maintained in X-VIVO15 medium (Lonza, Basel, Switzerland, 04-418Q) supplemented with 5% (v/v) inactivated FBS (GIBCO, Grand Island, NY, USA, 1914970), 100 U/mL penicillin/streptomycin (GIBCO, Grand Island, NY, USA, 15140-122), and 300 IU/mL interleukin (IL)-2 (Sino Biological Inc., Chesterbrook, PA, USA, GMP-CD66).

2.5. Construction of Anti-EGFR CARsScFv1 and scFv2 were developed based on anti-EGFR antibody mAB806 [30,31]; the difference between the two was in that the former was encoded in the VH-VL orientation and the latter in the VL-VH orientation. ScFv3, scFv4, and scFv5 were derived from EGFR antibodies Y022 [32], C10 [33], and cetuximab [34], respectively, and their sequences were encoded in the VH-VL orientation. The structure of anti-EGFR CARs consisted of three parts: (1) the extracellular domain comprising human CD8a signal peptides (nucleotides 1032–1094, GenBank NM 001145873.1), scFv, and CD8a hinges (nucleotides 1443–1577, GenBank NM 001145873.1); (2) the transmembrane domain representing the transmembrane portion of CD28 (nucleotides 515–595, GenBank NM 006139.4); and (3) intracellular domains comprising CD28 (nucleotides 596–718, GenBank NM 006139.4) co-stimulated domains and CD3ζ (nucleotides 363–698, GenBank XM 011510145.2) co-stimulated polypeptides. To facilitate the assessment of the transfection efficiency in anti-EGFR CAR lentivirus-infected T cells, eGFP was linked to CD3Z via the self-cleaving peptide P2A. All anti-EGFR CAR sequences were chemically synthesized and inserted into lentiviral vectors. 2.6. Preparation of Anti-EGFR CAR-T Cells

Freshly enriched primary human CD3+ T cells were activated with Dynabeads (ThermoFisher, Waltham, MA, USA, 11131D) at a 2:1 ratio for 24 h, infected with lentivirus carrying the anti-EGFR CAR, and maintained at the density of 1 × 106 cells/mL.

2.7. Flow Cytometry

The EGFR expression on the surface of the ESCC cell lines, anti-EGFR CAR expression in the T cells, and the proportion of human CD3+ T cells and anti-EGFR CAR-T cells in the mouse peripheral blood were detected by flow cytometry performed using PE-labeled anti-human EGFR (BioLegend, San Diego, CA, USA, 352904) and Brilliant Violet 421TM-labeled CD3 antibodies (BioLegend, San Diego, CA, USA, 300434) in CytoFLEX LX (Beckman Coulter, Brea, CA, USA) and LSRFortessa (BD) instruments.

2.8. Luciferase-Based Cytolysis Assay

Luciferase-expressing ESCC cells were suspended in complete RPMI 1640 medium at the density of 1 × 105 cells/mL, seeded into 96-well plates (Greiner, Kremsmünster, Austria, 655098) (100 μL per well), and cultured at 5% CO2 and 37 °C. After 6–8 h, effector (CAR-T) cells were added (100 μL per well) at different effector to tumor (E/T) cell ratios (0.5:1, 0.25:1, or 0.125:1), and the co-cultures were incubated at the same conditions for 24 or 72 h; then, 10 μL Steady-Glo luciferase substrate (Promega, Madison, WI, USA, E2520) was added and luminescence detected using SynergyH1 (BioTek, Winooski, VT, USA). The percentage of tumor cells lysed by effector cells was calculated based on the luciferase activity: 100%—(RLU of effector and tumor cell co-culture)/(RLU of tumor cells) × 100%, where RLU indicates the relative luminescence units.

2.9. Cytokine ELISA

Effector cells and ESCC cell lines were co-incubated at a ratio of 0.5:1 (1 × 104 tumor cells in each assay) for 3 days, and the supernatants were analyzed for the release of cytokines tumor necrosis factor (TNF)-α, IL-2, and interferon (IFN)-γ using ELISA kits (Dakewei, Shenzhen, China, 1110002, 1110202, and 1117202) according to the manufacturer’s instructions.

2.10. CAR-T Cell Antitumor Function in a Mouse Xenograft Model

Five-week-old female C-NKG mice (Cyagen, Santa Clara, CA, USA) were subcutaneously inoculated with 2 × 106 KYSE150-luci cells. After 14 days, when the tumor volume reached 30–40 mm3, the mice were distributed into PBS, T, CAR1-T, CAR2-T, CAR3-T, and CAR4-T groups (n = 5 per group), randomly, and injected with PBS, 5 × 106 T cells, or 5 × 106 anti-EGFR CAR-T cells (CAR+ 50%) through the tail vein on days 15 and 22. The mice were monitored weekly for body weight and tumor volume; peripheral blood was collected to analyze the ratio of human cells to CAR-T cells.

2.11. Statistical AnalysisAll of the statistical analyses for WGS and RNA-seq were performed using R (Version 4.0.2; https://www.R-project.org, accessed on 25 August 2020) and SPSS software (Version 22.0, https://www.ibm.com/analytics/spss-statistics-software, accessed on 25 August 2020). Student’s t-test was used to compare the expression between the tumor and normal samples. Fisher’s exact test was used to determine the association between risk scores and clinical characteristics. Kaplan–Meier curves were plotted and a log-rank test was performed. All of the subsequent experimental data were statistically analyzed using GraphPad Prism 6 and were presented as the mean ± standard deviation. Differences between groups were analyzed by Student’s t-test. The level of statistical significance was set at p 4. Discussion

ESCC is the predominant histological EC subtype in China; however, the effects of the existing therapeutic approaches are limited. Based on the molecular features of EGFR in ESCC, we speculated that anti-EGFR CAR-T cells might efficiently lyse ESCC cells. To test this hypothesis, we generated five anti-EGFR CAR-T cell lines, two of which showed functional activity against ESCC cells both in culture and in a mouse xenograft model. Our results provide a proof of concept that engineered anti-EGFR CAR-T cells may have a therapeutic potential for ESCC.

Molecularly targeted drugs against EGFR mainly include tyrosine kinase inhibitors (TKIs), such as gefitinib and erlotinib, and monoclonal antibodies such as cetuximab and panitumumab; however, many clinical trials have confirmed that ESCC do not respond to these drugs [37,38,39,40]. The negative clinical outcomes may be due to patients with ESCC having rare TKI-response EGFR driver mutations. Furthermore, the heterogeneity of esophageal cancer and drug resistance developed during treatment may also be responsible for the poor efficacy of molecularly targeted drugs [41].Although ESCC patients carry few EGFR driver mutations, most of them show EGFR gene amplification and overexpression, suggesting the potential of EGFR-targeting CAR-T cells for ESCC therapy. Indeed, among the anti-EGFR CAR-T cells constructed in this study, we successfully identified two cell lines, CAR1-T and CAR2-T, with a cytotoxic activity against ESCC cells in vitro and in vivo. These effective CAR-T cell lines differ from the ineffective ones (CAR3-T, CAR4-T, and CAR5-T) only in that their scFvs are derived from other EGFR antibodies, and it can be hypothesized that the unsatisfactory results in ESCC clearance may be due to conformational variations, which may affect binding to EGFR; however, this speculation requires experimental verification. Severe toxicity, including death, has been reported due to the off-target risk of CAR T cells [42,43]. Thus, scFv selection is crucial, mAb806 binds to the EGFR287–302 epitope, which is exposed when EGFR overexpressed or when EGFR is mutated in cancer cells. Therefore, mAb806 can bind to EGFR overexpressed in cancer cells, but not in normal cells [30,31]. Hence, CAR1-T cells and CAR2-T cells are expected to have an excellent anti-tumor effect in esophageal squamous cell carcinoma under safe conditions. The successful construction and functional identification of anti-EGFR CAR-T cells that are cytotoxic for ESCC cells provide a preclinical basis for the application of CAR-T cell immunotherapy to treat ESCC.Although anti-EGFR CAR-T cells are effective against non-small cell lung cancer and biliary cancer, several clinical trials suggest that CAR-T cells have a lower activity against solid tumors. One of the main reasons for this is that the solid tumor microenvironment leads to T cell exhaustion [44,45]. Therefore, the specific mechanisms underlying CAR-T cell exhaustion in the microenvironment of different solid tumors should be investigated to further improve the anti-neoplastic effect of CAR-T cells. We performed single-cell sequencing of T cells in tumor samples from patients with ESCC. An analysis of the sc-RNA-seq results is expected to assist in understanding the cause of TIL exhaustion, thereby further improving the antitumor effect of CAR-T cells in ESCC. Given the continuous replenishment of multi-omics sequencing data obtained on large ESCC samples, advances in gene editing technology, and the development of new drugs targeting diverse cancer-related immune mechanisms, it is believed that the anti-neoplastic activity of CAR-T cells against ESCC can be further improved through optimization and combination with other immunotherapies.