Preventive Treatment with PD-1 Antibody Increases Tissue-resident Memory T Cells Infiltration and Delays Esophageal Carcinogenesis

Esophageal cancer is one of the most common upper gastrointestinal malignant tumors, and is mainly divided into two subtypes: esophageal adenocarcinoma (EAC) and esophageal squamous cell carcinoma (ESCC). ESCC, one of the leading causes of cancer-related deaths, accounts for more than 90% of all esophageal cancer cases in China according to the statistics of the World Health Organization (1). The pathogenesis of esophageal cancer is still not fully understood. Most studies have shown that geographical location and poor living habits, such as smoking, obesity, hot drinks, and lack of trace elements, are leading risk factors for ESCC (2, 3). Normal esophageal epithelial cells become disordered after exposure to various carcinogenic factors, experiencing from low-grade intraepithelial neoplasia (LGIN) to high-grade intraepithelial neoplasia (HGIN), these cells are eventually contributed to the occurrence of ESCC (4).

At present, the priority in the treatment of ESCC is surgery combined with other non-invasive treatment methods (5). The early detection rate of esophageal cancer has gradually increased with the emergence of various cancer screening tests (6). Endoscopic treatment has become the main treatment for patients with early stage ESCC or HGIN because of the minimal trauma to patients (7). However, even with timely surgical treatment, the mortality rate of esophageal cancer has not significantly decreased. The 5-year survival rate has not exceeded 20% in the past many years (1). The high mortality can be attributed to tumor recurrence and metastasis, which are mainly due to the lack of serosal protection of the esophagus and abundant lymph nodes under the esophageal mucosa (8). Under these special anatomic conditions, tumor metastasis can occur in the early stages of ESCC or HGIN. In addition to the special anatomic structure of the esophagus, the cause of tumor metastasis largely depends on its unique immunosuppressive microenvironment. When immunosuppressive cell function prevails, the normal antitumor immune function is suppressed, which causes tumor cells to escape immune system surveillance and continue to expand or metastasize distantly (9). Therefore, it is extremely important to find effective measures to reverse the occurrence of tumors before the disease completely deteriorates.

With in-depth knowledge of immune cells in the tumor microenvironment, scientists have created many tumor therapies, such as tumor vaccines, chimeric antigen receptor T-cell therapy (CAR-T), and immune checkpoint inhibitors (ICIs; ref. 10). Among these, immune checkpoint inhibitors have provided alternatives for cancer treatment that can restore tumor-infiltrating lymphocyte's (TIL) antitumor function and have been used for the treatment of ESCC worldwide (11). Data from several ongoing clinical trials demonstrate that PD-1 antibody (Ab) monotherapy or combination chemotherapy can effectively improve the overall survival rate of patients with advanced ESCC (12). Neoadjuvant immunotherapy used in the perioperative period is the latest trial in the clinical treatment of ESCC. Studies have found that PD-1 Ab neoadjuvant therapy can enhance the systemic immune activity of antitumor T cells and reduce the scope of tumor lesions providing strong support for subsequent surgical treatment (13). However, in clinical practice, ICI therapy is characterized by a delayed response, in which patients experience tumor remission only after a period of treatment, but the exact cause has not been revealed (14).

Tissue-resident memory T cells (TRM), which form and reside mainly in peripheral nonlymphoid organs, are a unique group of memory T cells that use CD69 and/or CD103 as surface markers and serve as the first line of defense against foreign pathogens (15). In recent years, studies have confirmed the presence of TRM cells in the tumor microenvironment. Serving as a specialized subpopulation of TILs, these cells can secrete cytokines and chemokines to directly kill tumor cells or recruit local intrinsic immune cells (16). Ganesan and colleagues pointed out that the higher the proportion of TRM cells in the tumor microenvironment, the stronger the antitumor function of cytotoxic T cells (17). Furthermore, there is substantial evidence that at high density of TRM present in solid tumors is strongly associated with a good prognosis for patients with tumors (18). TRM cells are also closely associated with the efficacy of tumor immunotherapy. CD8⁺ TRM is the first T-cell subpopulation to express immune checkpoint molecules and early responders to cancer immunotherapy (19). The level of CD8⁺ TRM in the tumor microenvironment may also be used to predict patient responsiveness to ICI treatment (20).

In view of the benefits of ICIs in both advanced patients with ESCC and the perioperative period, we assumed that the application of ICIs in early stage ESCC (HGIN stage) may be able to reverse the progression of neoplasia and effectively control the recurrence or metastasis of cancer. In our study, 4-NQO was used to induce ESCC in mice, and esophageal lesions were observed at different time points (21). We first evaluated the efficacy of treating ESCC in mice with the PD-1 Ab at different time points, as well as the changes in immune cells in the immune microenvironment. Second, preventive treated mice were reexposed to the oncogenic environment to observe the condition of esophageal lesions and immune cell infiltration. We found that PD-1 Ab treatment increased CD8⁺TRM cell infiltration for a long period, and these cells colonized in the tumor microenvironment could still exert antitumor effects when mice drank the carcinogen solution again. These findings provide new strategies for the treatment of patients with early stage ESCC.

All animal studies were conducted in accordance with the guidelines of the Medical Research Center of Beijing Chao-Yang Hospital (Beijing, China) and approved by the Institutional Animal Care and Use Committee of Capital Medical University (Beijing, China). Six-week-old female C57BL/6J mice were purchased from Charles River Laboratories and were fed under specific pathogen-free conditions. To construct animal models for LGIN, HGIN, and ESCC, mice were fed drinking water containing 100 μg/mL 4-nitroquinolone-N-oxide (4-NQO, Sigma Aldrich) for 16 weeks and then sterilized drinking water (21). To identify the time-course experiments characterizing mouse esophageal malignant progression, mice were sacrificed every 4 weeks and their esophagi were processed into H&E slides until 28 weeks (16 weeks of 4-NQO exposure and 12 weeks of sterilized drinking water).

After determining the approximate esophageal lesion conditions in mice at different time points during the induction of ESCC by the chemical carcinogen 4-NQO, C57BL/6J mice were randomly divided into the Cont-IgG, PD-1 Ab preventive, and PD-1 Ab treatment groups. At week 18, mice in the PD-1 Ab preventive treatment group were treated with anti-mouse PD-1 Ab (BE0146, BioXCell, RRID:AB_10949053) 3 times a week for 3 weeks. At week 24, mice in the Cont-IgG and PD-1 Ab treatment groups were treated with rat IgG2a isotype control (BE0089, BioXCell, RRID:AB_1107769) or anti-mouse PD-1 Ab 3 times a week for three weeks separately. When treatment was completed, all mice were sacrificed and their esophagi were removed for subsequent experiments.

The esophagus of all mice was completely removed, and esophageal lesions were collected, fixed in 10% formalin, embedded in paraffin, sectioned into 5 μm sections and then subjected to H&E staining IHC analyses and multiplexed immunofluorescence (mIF). Esophageal lesions were classified as LGIN, HGIN, or ESCC, as defined by the WHO Classification of Tumors and reviewed by two certified pathologists.

Unstained sections and tissue microarrays (product number: HEsoSqu060CD-01) were used for H&E and IHC staining, respectively. A tissue microarray containing 45 specimens including normal esophageal mucosal tissue, HGIN and ESCC, was purchased from Shanghai Outdo Biotech. Ethical approval was granted by the ethics committee of Shanghai Outdo Biotech Company. Antibodies for CD8(ab217344, 1:500, RRID:AB_2890649), CD103 (ab182422, 1:500, RRID:AB_2753196), PD-L1(ab213480, 1:500, RRID:AB_2773715)and Rabbit IgG, monoclonal- Isotype Control (ab172730, 1:500, RRID:AB_2687931) were purchased from Abcam Inc. Ab for PD-1(80170–1-RR, 1:500, RRID:AB_2918871) was purchased from ProteinTech Group.

Unstained sections from mouse esophagus were used for multiplexed Immunofluorescence staining with Opal 4 Multiplex reagents (PerkinElmer) by the manufacturer's instructions. Staining was conducted with DAPI as a nuclear stain, CD8 opal 520, CD103 opal 570, and isotype opal 620. CD8⁺ TRM is defined as both CD103- and CD8-positive cells. Double-positive staining for CD8/CD103 was indicated by blue nuclei and red/green membranes. Three high magnification fields of view (HPF, X400) were randomly collected from each section, and the number of CD8 and CD103 double-positive stained cells were counted. The immunofluorescence staining results were counted and tallied by two pathologists under double-blind conditions.

All data were analyzed using GraphPad Prism 6.0 (GraphPad Software, RRID:SCR_002798). The staining intensities of PD-1/PD-L1 and CD103⁺CD8⁺ T cells were analyzed using rank-sum tests. Differences in mice weight were analyzed by unpaired t-tests. Differences in mice overall survival were analyzed by Kaplan–Meier analysis. P ≤ 0.05 were considered statistically significant defined as P ≤ 0.05; *, P ≤ 0.01**, P ≤ 0.001; ***, P ≤ 0.0001****.

The data generated in this study are available upon request from the corresponding author.

Previous studies have shown that mice exposed to a 4-NQO solution eventually develop ESCC. To clarify the pathologic processes of the mouse esophageal epithelial lesions after exposure to the chemical carcinogen 4-NQO at different time points, we established the ESCC model in C57BL/6J mice following a well-recognized protocol with 4-NQO carcinogen treatment (Fig. 1A). At the beginning of the experiment, we prepared in total 56 mice, which were randomly divided into 8 groups. A group of mice was randomly executed in order to observe the lesions every 4 weeks and their esophagi were processed into hematoxylin and eosin (H&E) slides to identify their lesion conditions. We classified mouse esophagi as normal epithelial, LGIN, HGIN, or ESCC (4).

At week 4, two of the seven mice started to show LGIN. At week 16, six of the seven mice showed HGIN. At week 20, one mouse in the experimental group developed ESCC. At week 28, all mice had ESCC. According to the experimental results above, we selected 18 weeks after exposure to 4-NQO as the PD-1 Ab preventive treatment timepoint when most mice displayed HGIN, and 24 weeks after exposure to 4-NQO as the PD-1 Ab treatment timepoint when most mice displayed ESCC. A bar plot (Fig. 1B) shows the different time points of mouse lesion conditions and a histogram (Fig. 1C) was used to count the proportion of mice esophagi lesions in different periods of 4-NQO exposure. Representative mice esophagi images and H&E staining of normal esophageal epithelium, LGIN, HGIN, and ESCC induced by the chemical carcinogen 4-NQO are shown in Fig. 1D. Furthermore, we counted the changes in body weight of mice during the induction of the chemical carcinogen 4-NQO and we found that the body weight of the control group increased faster than that of 4-NQO–induced mice (Fig. 1E).

To provide a rationale for the use of PD-1 mAb in the early stages of ESCC, we examined the expression of immune checkpoints PD-1 and its ligand PD-L1 at different stages of esophageal lesions by IHC. As shown in Fig. 2A and B, we found that the expression of PD-1 and PD-L1 protein was significantly higher in the HGIN and ESCC stages than in the normal esophageal epithelium. In Fig. 2C and D, the human ESCC tissue microarray, total protein expression of PD-1 and PD-L1 in patients with esophageal carcinoma in situ and ESCC was much higher than in normal esophageal mucosal epithelial tissue. These results provide a basis for subsequent immune checkpoint therapies.

After studying the expression of immune checkpoint proteins in mice at different lesion conditions during 4-NQO induction, we chose to treat mice with PD-1 Ab as a preventive treatment at the HGIN stage (week 18) and treatment at the ESCC stage (week 24) or with rat IgG as isotype-matched controls (Fig. 3A). Images of esophagi-visible lesions also showed that esophagi in the PD-1 Ab preventive treatment group had less visible thickening and redundancy than in the PD-1 Ab treated group or control group (Fig. 3B and C). We also focused on changes in body weight among the three groups. We found an increase in body weight in both the PD-1 Ab preventive treatment group and PD-1 Ab treatment group compared with the IgG-treated group; however, the PD-1 Ab preventive treatment group showed a more significant increase in body weight at the end of the experiment (Fig. 3D). These results suggest that treatment with PD-1 Ab at the HGIN stage (preventive treatment) is a more effective way to impede the development of squamous esophageal carcinoma in mice. In addition, we used mIF to observe the effects of therapy at different times on CD8⁺TRM cell infiltration. We observed that CD103⁺CD8⁺ T cells were significantly elevated in both the preventive treatment and treatment groups compared with the control group injected with IgG (Fig. 3E and F).

After we found that PD-1 Ab treatment can effectively increase CD8⁺TRM cell infiltration, we aimed to explore whether PD-1 Ab preventive treatment can induce CD8⁺TRM cells to colonize the tumor microenvironment for a long period to exert anti-tumor effects. We treated two groups of HGIN-stage mice with PD-1 Ab and the other two groups with rat IgG as isotype-matched controls at week 18, and then sacrificed two groups (one control group and one PD-1 Ab preventive treatment group) immediately after treatment, and the other two groups were sacrificed 6 weeks later (Fig. 4A). We found that there was no difference between two PD-1 Ab preventive treatment group in the number of tumors (Fig. 4B and C). In addition, we found that the total number of CD8⁺TRM cells did not change significantly in these two groups, which reminds us that once these cells are formed and reside in the tumor microenvironment, they can continue to exert anti-tumor effects for a long period of time (Fig. 4D and E).

Since the development of ESCC is closely related to lifestyle and environmental factors, it is possible that patients are reexposed to carcinogenic factors even after surgical/medical treatment. We then reexposed control and PD-1 Ab preventive treatment mice to 4-NQO (Fig. 5A) and compared the changes in survival, body weight, tumor number, and immune cell infiltration in the tumor microenvironment. The results showed that mice that received PD-1 Ab preventive treatment had significantly longer survival and a reduction in tumor number than the control group (Fig. 5B–D). Although we observed a decrease in body weight after reexposure to 4-NQO in the preventive treatment group, there was a remarkable improvement compared with that in the control group (Fig. 5E). We then analyzed the immune cell infiltration in the tumor microenvironment and found that the percentage of CD8⁺ TRM in the tumor microenvironment was significantly higher in mice that received PD-1 Ab preventive treatment (Fig. 5F and G). This suggests that CD8⁺ TRM cells that previously resided in the tumor microenvironment could play a significant anti-tumor role when mice were reexposed to the oncogenic environment.

In China, ESCC is the main pathological type of esophageal cancer. Surgery or endoscopic treatment combined with chemoradiotherapy does not significantly improve patient’ prognosis because of the special anatomical position of the esophagus and its unique immunosuppressive microenvironment (8). Therefore, choosing appropriate treatment methods to delay or reverse tumor progression and prevent recurrence or distant metastasis is of vital importance in ESCC treatment. Immune checkpoint inhibitors have elicited increasing interest in clinical practice and research field (22). Unlike indiscriminate destruction by chemotherapy drugs, immune checkpoint inhibitors (ICIs) can kill tumor cells by restoring the body's antitumor immunity, which has gained great success in the treatment of ESCC (10, 23). Based on the anti-tumor mechanism of ICIs, we speculated whether the use of PD-1 Ab in the early stages of ESCC or HGIN could reverse tumor progression by restoring the body's effective anti-tumor immunity.

First, we used the chemical carcinogen 4-NQO to simulate the carcinogenic process in C57BL/6J mice, which undergo atypical proliferation (LGIN and HGIN) and eventually evolve into ESCC (21). Next, we observed the expression of immune checkpoint PD-1 and its ligand PD-L1 during carcinogenesis. In our study, we found that PD-1 and PD-L1 expression was significantly elevated from the HGIN stage in both mice and human tissues compared with the normal esophageal mucosal epithelium, which usually reflects T cells dysfunction (24, 25). In addition, this finding provides us a theoretical basis for the use of PD-1 Ab therapy during the HGIN period. We chose to start PD-1 Ab treatment at 24 weeks and PD-1 Ab preventive treatment at 18 weeks to investigate whether the use of PD-1 Ab could slow down the progression of the disease or provide a complete cure. We found that after treatment with the PD-1 Ab in the ESCC phase, the mice showed a slight increase in body weight compared with the control group. However, the mice treated with PD-1 Ab during the HGIN phase showed a significant increase in body weight compared with the control and PD-1 Ab treatment group, and a reduction in esophageal lesions was also clearly observed. This may be due to the fact that PD-1 monoclonal antibodies perform a different anti-tumor function compared with chemotherapeutic drugs (23). Although PD-1 Ab therapy can restore the normal anti-tumor function of T cells, these cells cannot exert an anti-tumor function as chemotherapeutic drugs in a short period of time (26). In addition, even if partial T-cell function is restored, these cells cannot fight against the high tumor load and immunosuppressed state of the body. Therefore, PD-1 Ab preventive treatment for patients in the HGIN stage can lead to better outcomes.

To identify the reason why PD-1 Ab treatment effectively delayed disease progression in HGIN-phase mice, we next analyzed the infiltration of immune cells in the tumor microenvironment, especially TRM cells. TRM cells, which can be divided into CD4⁺ TRM cells, CD8⁺ TRM cells, Treg TRM cells, and many other cell subpopulations, colonize peripheral non-lymphoid organs, including tumor tissues, and are an important cell population for adaptive immunity in peripheral tissues (15). In the tumor microenvironment, CD8⁺ TRM can secrete cytotoxic factors, such as IFNγ, granzyme A and B or recruit other immune cells including NK cells and B cells into the tumor microenvironment to kill tumor cells (27), while CD4⁺ TRM plays a role in antigen presentation and assists CD8⁺ TRM formation (28). Numerous studies have shown that the abundance of TRM cells in the tumor microenvironment is significantly and positively correlated with the prognosis of solid tumors, and that a sufficient number of TRM cell infiltrations is a prerequisite for T cell-based tumor immunotherapy (mainly including immune checkpoint inhibitors for ICIs, CAR-T, and ACT therapy; refs 29–31).

We found that treatment of mice with PD-1 Ab in both HGIN and ESCC phases could increase the infiltration of CD8⁺TRM cells, which is also similar to the results of Han and Edward (19, 20). Subsequently, to investigate whether CD8⁺ TRM formulated in tumor microenvironment could resident long and exert anti-tumor effect persistently, for groups of mice in the HGIN phase were treated with PD-1 Ab or rat IgG (as control), and then two groups of mice were sacrificed immediately at the end of the treatment and the other two groups were sacrificed 6 weeks later. We found that treatment with the PD-1 Ab increased the colonization of CD8⁺TRM cells in the tumor microenvironment for a long period in both groups. However, at the end of treatment, the number of CD8⁺ TRM gradually decreased with the decrease in tumor burden.

Because the development of ESCC is closely related to living habits and environmental factors, it is still possible for patients to be reexposed to a cancer-causing environment even after surgical/pharmaceutical treatment. Next, untreated mice and PD-1 Ab preventive treatment mice were reexposed to 4-NQO to observe whether CD8⁺TRM cells colonized in the tumor microenvironment could help resist the oncogenic effects of 4-NQO. We found that mice that received PD-1 Ab preventive treatment had significantly longer survival when reexposed to the 4-NQO environment than untreated mice. Analysis of their tumor microenvironment revealed a significant increase in the proportion of CD8⁺ TRM compared with the untreated groups. This phenomenon also shows us the immune memory function of CD8⁺TRM cells; when the tumor antigen is reexposed, CD8⁺TRM residing in the tumor microenvironment regains a strong tumor-killing function (32). In addition, a study reported that TRM cells may exert resistance to oxidative DNA damage in cells through the nucleotide pool oxidative repair enzyme NUDT1 (33), which is similar to the oncogenic effects of many ESCC risk factors including 4-NQO, and may also help us to explain that PD-1 Ab preventive treatment can significantly prolong survival time in mice reexposed to carcinogenic factors.

In conclusion, for patients with early-stage ESCC, it is especially important to choose the appropriate treatment to slow down disease progression and prevent recurrence and metastasis. Immunotherapy has achieved great success in the current treatment of ESCC, with CD8⁺TRM cells playing an important role. Our study revealed that the use of the immune checkpoint inhibitor PD-1 Ab in the early stage of ESCC was effective in delaying disease progression and increasing CD8⁺TRM cell infiltration of the immune microenvironment for a long period. In addition, we found that CD8⁺TRM cells exerted significant antitumor effects when mice that had received PD-1 Ab preventive treatment were reexposed to carcinogenic factors. Our study provides new ideas for the clinical treatment of patients with early stage ESCC.

No disclosures were reported.

Z. Xiao: Conceptualization, data curation, writing–original draft, project administration, writing–review and editing. R. Yan: Supervision. H. Liu: Software. X. Huang: Validation. Z. Liang: Visualization. G. An: Conceptualization, resources, project administration. Y. Ge: Conceptualization, resources, project administration.

We would like to thank every author who made an indispensable contribution to this research. This study was supported by the Beijing Chaoyang Hospital Scientific Innovation Project (22kejjzd-9) and Y. Ge had been awarded this grant.

The publication costs of this article were defrayed in part by the payment of publication fees. Therefore, and solely to indicate this fact, this article is hereby marked “advertisement” in accordance with 18 USC section 1734.