Paraffin-embedded tumor tissues were obtained from The Third Affiliated Hospital of Zhengzhou University. The specimens were collected under an Institutional Review Board (IRB)-approved protocol, with informed consent obtained. Antigen retrieval was performed using the heat-induced epitope retrieval (HIER) method with a microwave oven. The slides were heated in sodium citrate buffer for 15 min at 95°C using a microwave oven, followed by a 30-min cooling period at room temperature. Manual immunohistochemistry (IHC) was optimized and performed using a 1:200 dilution of the antibody against CD155 (no. D8A5G) and Ki67 (no. 9449; both from CST, Danvers, MA, USA). The tissue sections underwent deparaffinization in xylene, dehydration and rehydration in ethanol, and steamer antigen retrieval. Subsequently, the sections were blocked with 10% serum and incubated overnight with primary antibodies at 4°C in a light-protected environment. Afterward, the sections were incubated with the corresponding secondary antibody for 1 h at room temperature. Following chromogenic staining and nuclear counterstaining, the sections were observed under a microscope and photographed (27).

Tissues from the indicated mouse organs were dissected, fixed in 4% paraformaldehyde for 24 h, embedded in paraffin, and sectioned (4 μm). Paraffin sections were deparaffinized and rehydrated, followed by standard hematoxylin and eosin (H&E) staining (28).

The cervical cancer cell lines (HCA1 and CAC1C), the breast cancer cell lines (MDA-MB231 and MCF-7), the immortalized cervical epithelial cell line H8, and the human embryonic kidney cell line 293T were purchased from the Cell Bank of Shanghai Academy of Chinese Sciences (Shanghai, China). All of these cell lines, except for the 293T cell line, were cultured in RPMI 1640 medium (no. 11875093; Gibco, Grand Island, NY, USA) with 5% fetal bovine serum (no. 10099; Gibco) and 100 U/ml penicillin/streptomycin (no. 15140148; Invitrogen, Carlsbad, CA, USA). The 293T cell line was cultured in high-glucose Dulbecco’s modified Eagle’s medium (DMEM) (no. 12491015; Gibco) with 10% fetal bovine serum (no. 10099; Gibco) and 100 U/ml penicillin/streptomycin (no. 15140148; Invitrogen).

The CD155 short hairpin RNA (shRNA) lentivirus (shCD155) and the control shRNA lentivirus (NC) were constructed by Gene Pharma (Shanghai, China). The shRNA sequence for shCD155 was 5′-GGGCCAAGTGCACATCATT-3′, while that for NC was 5′-TTCTCCGAACGTGTCACGT-3′. For stable knockdown, the shRNA or the NC was cloned into the pCDH–CMV–shRNA–GFP vector. HCA1 and MDA-MB231 cells were transfected with the NC or the shCD155 lentivirus. The plasmid for overexpressing CD155 was purchased from GenePharma (Shanghai, China). The lentiviral vector pCDH–CMV–CD155–GFP was used for transfection. MCF-7 and CAC1C cells were transfected with the NC or the CD155 lentivirus. To produce lentiviral vectors, the 293T cells were transfected with the shRNA transfer plasmids and packaging plasmids (psPAX2 and PMD2.G) using calcium phosphate precipitation. The viral supernatant was collected at 72 h post-transfection, concentrated via ultracentrifugation, and resuspended for storage at −80 °C. The viral titer [in transducing units (TU) per milliliter] was determined by transducing Jurkat cells with serial dilutions of the concentrate and quantifying the green fluorescent protein (GFP)-positive cells by flow cytometry after 72 h.

Colony formation assays were performed in a six-well format. A titration of cell number of each of the cell lines was performed to identify the best cell seeding condition. By day 14, the medium was removed and the cells washed with phosphate-buffered saline (PBS) and stained using 500 μl of 0.005% (w/v) crystal violet solution in 25% (v/v) methanol for at least 1 h at room temperature (29). Cell viability was examined using a CCK-8 assay (Sevenbio, Nanjing, China). The cells were seeded in 96-well plates at a density of 4 × 10³ cells per well in 200 μl medium for 24, 48, and 72 h. The absorbance was detected at 450 nm after the cells were treated with 10% CCK-8 at 37°C for 2 h. Cell viability was calculated as the ratio of the optical density (OD) values of the drug-treated samples to those of the controls (30).

The CD155–BBz CAR construct was generated by fusing the extracellular domain of TIGIT derived from the TIGIT cDNA with the transmembrane and cytoplasmic domains of 4-1BB and CD3z. The control CAR (MOCK-CAR) was generated by deleting the BBz.CD3z signaling domain from the CD155-BBz CAR. The CAR sequence targeting CD19, an antigen-irrelevant species-specific negative control, was created using a scFv domain (derived from the FMC63 antibody) and the transmembrane and cytoplasmic domains of 4-1BB and CD3z. All sequences were synthesized by Shanghai Sangon Biotech (Shanghai, China). Human T cells were isolated from the PBMCs obtained from healthy human donors and cervical and breast cancer patients using a human T-cell isolation kit (Miltenyi, Bergisch Gladbach, Germany). Subsequently, the isolated T cells were stimulated with anti-human CD3/28 magnetic beads (no. 11161D; Gibco) for 24 h. The human T cells were cultured in RPMI 1640 medium (no. 11875093; Gibco) supplemented with 5% fetal bovine serum (no. 10099; Gibco) and 100 U/ml penicillin/streptomycin (no. 15140148; Invitrogen). Lentiviruses were produced by transfecting 293T cells using Lipofectamine 3000 (no. L3000001; Invitrogen) according to the manufacturer’s instructions. On day 3, the activated T cells were transduced with the CD155-BBz CAR and CD19-CAR lentiviruses. The transduction efficiency was assessed by measuring the GFP expression using flow cytometry at 7 days post-transduction (31).

For the flow cytometry analysis, the cells were first gated on the forward scatter (FSC) vs. the side scatter (SSC) plot to identify the live cell population. Doublets were excluded by gating on FSC-H vs. FSC-A. The target positive population was then defined based on the fluorescence intensity using the unstained or isotype control sample as a reference. The cells were stained with ~2 µl antibodies for 30 min at 4°C, washed with PBS (no. P2272; Sigma, St. Louis, MO, USA), and acquired on a flow cytometer (cytoFLEX; Beckman, Brea, CA, USA). PE-anti-human CD155 (no. 337609; BioLegend, San Diego, CA, USA), PE-anti-Ki67 (no. 571599; BD, Franklin Lakes, NJ, USA), PerCP-anti-human CD4 (no. 300529; BioLegend), APC-anti-human CD8 (no. 344746; BioLegend). Analysis was conducted using FlowJo software (FlowJo, Ashland, OR, USA). For intracellular staining, the cells were fixed and permeabilized with Intracellular Fixation and Permeabilization Buffer Set prior to the addition of the intracellular staining antibody sets.

The cytotoxicity of the CAR T cells was assessed using the lactic dehydrogenase (LDH) release assay (no. ab65393; Abcam, Cambridge, UK) according to the manufacturer’s instructions. The cells were plated in 96-well plates at various effector-to-target (E/T) ratios (i.e., 1:1, 5:1, and 20:1) for a duration of 8 h. The levels of LDH release in the experimental and maximum conditions were measured, and the percentage of cytotoxicity was determined by dividing the experimental LDH release by the maximum LDH release (32). The culture supernatants were collected after 24 h of co-culture to measure the release of interleukin 2 (IL-2) and interferon gamma (IFN-γ) with ELISA following the manufacturer’s instructions (no. S2050 and no. SIF50C; R&D, Minneapolis, MN, USA). After the reactions, the absorbance at 450 nm was read using a spectrophotometer (NanoDrop Technologies, Wilmington, DE, USA). The concentrations of IL-2 and IFN-γ were calculated based on a standard curve (33).

NOD-SCID mice (females, 6–8 weeks old) were obtained from Vital River (Beijing, China) and housed in a specific pathogen-free (SPF) environment. Tumor cells were subcutaneously injected into the hind limb of the mice on day 0. After inoculation, the mice were randomly divided into two groups. CAR T cells, resuspended in 100 µl of PBS, were administered via tail vein injection at the indicated time points. Tumor growth was measured with digital calipers every 2 days. Mice were euthanized when the tumor volume reached 1,500 mm³. The tumor volume was calculated using the formula: V = (length × width × width) × 0.5. Blood, spleen, and liver samples were collected at the time of sacrifice to determine CAR T-cell infiltration and to assess organ damage severity (34).

For general toxicity due to the CAR T-cell therapy in mice, three groups of non-tumor-bearing mice were included. CD155.CAR, CD19.CAR, or Mock-T cells, resuspended in 100 µl of PBS, were administered via tail vein injection at the indicated time points. The body weight of all mice was measured once every 2 days for 1 month. The serum was collected at the end of the experiment. The activity of serum enzymes such as aldolase, alanine aminotransferase (ALT; Sigma), and creatine kinase (CK) and the serum creatinine values (Abcam, Waltham, MA, USA) were determined following the manufacturer’s protocol (35).

Statistical analyses were performed using GraphPad Prism 8.0 software. Data are displayed as the mean ± SD. Statistical tests included the two-tailed Student’s t-test and one-way and two-way ANOVA, with Sidak–Bonferroni correction applied to correct for multiple comparisons, when applicable. Data distribution was assumed to be normal, but which was not formally tested. For in vivo experiments, overall survival was depicted by a Kaplan–Meier curve, and a log-rank test was used to compare differences in the survival between groups. ns, *p < 0.05, **p < 0.01, and ***p < 0.001 were considered for statistical significance.

Tumor targeting by CAR T cells requires the expression of tumor-associated antigens (TAAs) on the surface of tumor cells. To investigate the expression of CD155 in primary breast and cervical cancer, we examined the variations in the CD155 expression among breast cancer, cervical cancer, and normal tissues. IHC was conducted using 30 samples of human cervical cancer and 10 samples of peri-cancerous tissues (Supplementary Table S1). In addition, 30 samples of human breast cancer paired with adjacent normal mammary tissues were included in the analysis (Supplementary Table S2). The results demonstrated that the expression of CD155 in cervical and breast cancer tissues was significantly higher than that in para-carcinomic tissues (Figures 1A, B, D, E). To further examine the prognostic potential of CD155 in cervical cancer, we then analyzed the impact of the expression of CD155 on the prognosis of patients with cervical and breast cancer in the Human Protein Atlas database (https://www.proteinatlas.org/). The results showed that the expression of CD155 was negatively correlated with patient survival in cervical and breast cancer (Figures 1C, F). Based on the analysis results of IHC and the databases, normal human tissue microarrays were collected and IHC staining performed to examine the expression of CD155. In contrast to malignant samples, the marginal and healthy samples exhibited low levels of CD155 in the limited cells (Figure 1G). In general, all of these results provide further evidence of the tumor-specific expression of CD155 at the protein level, suggesting that CD155 could be a potential target for the treatment of cervical and breast cancer.

The available results suggest that CD155 is associated with tumor cell proliferation in the progression of cervical and breast cancer (Supplementary Figure S1). To investigate whether CD155 plays a role in the pathogenesis of cervical and breast cancer, we performed gain- and loss-of-function of CD155 in the cervical and breast cancer cell lines. A higher expression of CD155 has been confirmed in two cervical cancer cell lines (HCA1 and CAC1C) and in one breast cancer cell line (MDA-MB231). Conversely, a lower expression of CD155 has been observed in another breast cancer cell line (MCF-7). In addition, the immortalized cervical epithelial cell line (H8) showed no expression of CD155. These findings were confirmed using flow cytometry (Supplementary Figure S2A). The overexpression and silencing of CD155 in tumor cell lines were confirmed by flow cytometry (Supplementary Figure S2B). The colony formation assays and the CCK-8 assays all showed that CD155 overexpression remarkably promoted, whereas the silencing of CD155 inhibited, the growth of cervical and breast cancer cells (Figures 2A, B). We then verified this phenomenon in human tissue specimens. The correlation between CD155 and Ki67 in tumor tissues was examined using IHC, and the results showed that the level of CD155 was positively related to the expression of Ki67 in the cervical and breast cancer tissues (Figure 2C). Collectively, our data strongly demonstrated that CD155 serves as a tumor proliferation-promoting factor and may serve to encourage further studies to identify innovative therapeutic approaches for patients with cervical and breast cancer.

In general, CAR T-cell therapy is antigen-specific, and its efficacy is dependent on the level of expression of the targeted antigen on the tumor cell (36). To evaluate the CD155 antigen specificity, we transduced the activated T cells either with a second-generation CD155-specific CAR (CD155.CAR) or a CD19 control CAR (CD19.CAR) (Figure 3A). The CD155-specific CAR expression on human T cells transduced with lentiviral particles was analyzed using flow cytometry through detection of GFP fluorescence (Figure 3B). We then evaluated the function of the CAR T cells against the CD155⁺ tumor lines by assessing the cytolytic potential, the proliferative capacity, and the cytokine secretion. To evaluate the cytolytic potential of the CD155.CAR T cells, we performed a 12-h co-incubation experiment against HCA1, MDA-MB231, MCF-7, and H8. All of the CD155⁺ cervical and breast cancer cell lines tested exhibited cell killing, even at low E/T ratios. However, the CD155 low or negative cell lines did not show sensitivity to killing by CD155.CAR T cells (Figure 3C). Consistently, the CD155.CAR T cells upregulated the secretion of IFN-γ and IL-2 compared with the Ctrl-T cells (Figures 3D, E). Taken together, these data suggest that CD155.CAR T cells exhibit specificity and cytotoxicity against CD155⁺ cervical and breast cancer cells through greater secretion of IL-2 and IFN-γ.

To evaluate the in vivo tumor suppression activity of the CD155.CAR T cells, a cervical cancer xenograft mouse model was established by the subcutaneous transplantation of HCA1 cells in NOD-SCID mice. After 7 days of tumor cell implantation, the mice received CAR T treatment via tail vein injection (Figure 4A). Compared with the control group, the CD155.CAR T cells were effective in controlling the growth of CD155⁺ tumors and showed safety profiles as indicated by the body weight loss (Figures 4B, C). As a result, the mice in the experimental groups received CD155.CAR T cells, which significantly extended the overall survival (Figure 4D). Similar efficacy results were observed in the breast cancer xenograft model established by the subcutaneous transplantation of MDA-MB231 cells in NOD-SCID mice (Figures 4E–H). At the end of the treatment study, the remaining tumor xenograft samples were collected to examine the proliferation of tumor cells. As shown in Figure 4I, in the CD155.CAR T-cell treatment group, the proliferation of tumor cells was significantly decreased as determined by Ki67 staining. Collectively, these in vivo experiments indicated that CD155.CAR T cells improve the control of established cervical and breast cancer xenografts of tumor-expressing CD155 target antigens.

Although the low or negative expression of CD155 on normal cells was confirmed, as shown in Figure 1, it was also necessary to evaluate the safety after CD155.CAR T-cell therapy. The general toxicity in non-tumor-bearing mice due to the CAR T-cell administration was evaluated in non-tumor-bearing NOD-SCID mice as mice were not expected to express high levels of CD155 in any organ. To determine the safety profile of the CD155.CAR T cells and to further discuss their translation into the clinic, the non-tumor-bearing mice in each group received CD155.CAR-T, CD19.CAR-T, or Mock-T cells infused via the tail vein. The mice were monitored for 30 days and were collected to determine the weights of the liver, spleen, lung, and kidneys. There were no significant differences in the weights of these organs, demonstrating the good safety profile of the CD155.CAR T cells (Figure 5A). Further histological analysis of the liver, the spleen, the lung, and the kidneys by H&E staining revealed no obvious abnormalities in the CD155.CAR T-cell-treated mice (Figure 5B). CAR T-cell administration showed a remarkable safety profile in terms of changes in the serum enzymes such as aldolase, ALT, CK, and serum creatinine (Figure 5C). There was no significant decrease in the body weights of the mice in the three groups (Supplementary Figure S3). Furthermore, we investigated whether the injected T cells might colonize other organs to cause tissue injury. There was only minimal infiltration of GFP-positive CAR T cells in all of the examined tissues across all treatment groups (Figure 5D). This finding provides strong evidence that, in the absence of substantial antigenic stimulation, CD155.CAR T cells do not preferentially expand, persist, or accumulate in these vital organs. Together, CD155.CAR T cells were well tolerated and did not result in obvious adverse events in mice, suggesting that CD155 is a safe and potent antitumor target for CAR T cells and is worthy of further development for clinical translation.

Patients with cervical and breast cancer have reported abnormalities in their lymphocyte function. Thus, the antitumor activity of the CD155.CAR T cells derived from cervical and breast cancer patients’ T cells needed to be verified. The PBMC collected from one patient with cervical cancer [squamous cell carcinoma, human papillomavirus (HPV)-positive, TNM stage II] was used to isolate the T cells, and the T cells were then used for transduction with the CD155-BBz CAR lentiviruses (Supplementary Figure S4A). In agreement with previous reports, the patient’s T cells (P-T cells) proliferated more slowly than the T cells of a healthy donor (HD-T cells) (Supplementary Figure S4B). However, the patient’s T cells had similar CD4/CD8 ratios to those of the T cells derived from the healthy donor (Supplementary Figure S4C). To determine the cytotoxicity of the CD155.CAR T cells from different sources against the CD155+ tumor cells, a co-incubation experiment that included both T cells and tumor cells was performed. The results showed that the patient’s CD155.CAR T cells had a comparable antitumor activity to the CD155-based CAR T cells derived from the healthy donor (Figures 6A, B). Correspondingly, both CD155.CAR T cells showed effective antitumor activity against the xenografts formed by the HCA1 and MDA-MB231 cells in NOD-SCID mice (Figures 6C–H).