Abstract
Background/Aim: Recent studies suggest that PD-L1 expression in immune cells, rather than tumor cells, plays a key role in tumor immunity. Trefoil factor family 1 (TFF1) is a secreted protein expressed mainly by the gastrointestinal epithelium and is related to the development of malignant disease. This study investigated the effects of TFF1 on tumor immunity in a xenograft mouse model of colorectal cancer (CRC). Materials and Methods: MC38 cells were implanted in wild-type (WT) and TFF1KO mice, and the tumor micro-environment was investigated using immunohistochemistry. The circulating immune cells were analyzed using flow cytometry. Results: Tumor growth was suppressed in TFF1KO mice. In the tumor microenvironment, CD8- and CD4-positive T cells and CD11c-positive dendritic cells (DCs) were frequently found in TFF1KO mice. When an immune checkpoint inhibitor was administered to these mice, almost half of the tumors in TFF1KO mice showed a complete response. The number of circulating PD-L1/DCs was markedly associated with tumor volume, with TFF1 deletion accelerating this effect and its injection decreasing it. These findings indicate that loss of TFF1 activates tumor immunity via frequent T-cell priming by DCs, and eventually suppresses tumor growth in CRC. In addition, the number of circulating PD-L1/DCs was identified as a predictive marker of checkpoint-inhibiting therapy efficacy. Conclusion: Loss of TFF1 resulted in accelerated immune response to colorectal cancer. Further studies are needed to investigate the precise mechanisms of TFF1 in immunotolerance and develop a novel TFF1-inhibiting immunotherapeutic strategy for CRC.
Colorectal cancer (CRC) is the third most common cancer and remains one of the leading causes of cancer-related deaths worldwide (1). Although multimodality therapy, including surgery, chemotherapy, targeted therapy, and radiation therapy, has significantly improved the prognosis of patients with CRC (2, 3), the emergence of chemoresistance often hinders the effect of treatment, and alternative strategies, including immunotherapy, have been considered (4). Since the development of PD-1/PD-L1 inhibitors for the treatment of malignant diseases, immunotherapy has become a novel and important option for various types of cancers. Although immunotherapy exhibits tremendous effects in some patients with malignant melanoma and lung cancer (5), only a small portion of patients with CRC benefit from these therapies (6). Thus, a novel strategy is necessary to predict or strengthen the effect of immunotherapies (7-10).
In recent years, the PD-L1/PD-1 axis has become a central target of immunotherapeutic approaches for treating malignant disease (11). The cell-surface receptor PD-1 is expressed in T cells and binds to the PD-L1 ligand, which is abundantly expressed in tumor cells. This binding leads to attenuated T-cell activity and immunotolerance. However, in the tumor microenvironment, PD-L1 is highly expressed in immune cells as well as tumor cells, such as DCs, macrophages, myeloid-derived suppressor cells (MDSCs), and Tregs (12). While the effects of PD-L1 on immune cells of the tumor microenvironment remain unclear, recent studies have shown that PD-L1 expressed in immune cells plays an important role and may be even more predictive of the response to immunotherapy than expression of PD-L1 in tumor cells (13-16). Several studies suggest that PD-L1 expression on T cells and DCs also attenuates T-cell activation (13, 17, 18), activated T cells express PD-L1 (19) and PD-L1 on DCs is up-regulated by the uptake of tumor antigens (13, 18). Thus, it is of great importance to understand the functional contributions of PD-L1 expression on T cells and DCs as this information will contribute to the continued progress of cancer immunotherapy.
Trefoil factor family 1 (TFF1) is one of three members of the trefoil factor family. TFF1 is most highly expressed in the gastrointestinal epithelium, including the stomach, colon, and pancreas (20). Recent studies have revealed that TFF1 also functions as a tumor suppressor to inhibit gastric (21), pancreatic (22), Barrett’s epithelial (23) and hepatocellular carcinogenesis (24), whereas other studies have indicated that TFF1 has oncogenic activity (25, 26). The role of TFF1 in the progression of CRC and its prognostic value have not yet been extensively evaluated (27), and there have been no investigations into the association between TFF1 and tumor immunity. In this study, we utilized TFF1-deficient mice to investigate the effects of TFF1 on tumor immunity in a mouse model of CRC.
Materials and Methods
Mice. All animal experiments were conducted in compliance with the guidelines of the Institute for Laboratory Animal Research, Nagoya University Graduate School of Medicine (Approval Number: M230096-002). TFF1-knockout (TFF1KO; knockout-first allele:Tff1tm1a(EUCOMM)Wtsi) mice were purchased from the International Mouse Phenotyping Consortium. C57BL/6J male mice were purchased from SLC Japan (Nagoya, Japan) as control wild-type (WT) mice.
Cell line. The C57BL/6J colon adenocarcinoma cell line MC38 was obtained from Kerafast (Boston, MI, USA). MC38 cells were cultured in DMEM (high glucose, FUJIFILM Wako, Osaka, Japan) containing 10% fetal bovine serum (FBS; Thermo Fisher Scientific, Waltham, MA, USA), 0.1 mM nonessential amino acids (Thermo Fisher Scientific), 1 mM sodium pyruvate (Thermo Fisher Scientific,), 10 mM HEPES (Thermo Fisher Scientific), and 100 U/ml penicillin–streptomycin at 37°C in a humidified atmosphere with 5% CO2.
Xenograft mouse model of CRC. WT C57BL/6J and TFF1 KO mice were injected subcutaneously (s.c.) with 1.0×106 MC38 cells into the dorsum. The tumor volume (mm3) was measured twice a week and was calculated as follows: tumor volume=(length × width2)/2. In the anti-PD-1 antibody treatment model, mice were injected intraperitoneally (i.p.) with anti-PD-1 monoclonal antibody [10 mg/kg, Ultra-LEAF™ Purified anti-mouse CD279 (PD-1) - Biolegend (San Diego, CA, USA) Cat No. 114122] on days 3, 7, 10, 14 and 17 after MC-38 inoculation. In the TFF1-treatment model, recombinant TFF1 was injected subcutaneously into the mice [1 μg/mouse, Recombinant Murine TFF1; Peprotech (Cranbury, NJ, USA) Cat No. 315-31] on days 3, 7, 10, and 14 after MC-38 inoculation. All mice were sacrificed on day 21. The tumors were harvested, and EDTA-anticoagulated peripheral blood was collected from the heart.
Flow cytometry. Flow cytometry was performed using fluorochrome-conjugated antibodies (Table I). Peripheral blood samples were processed by VersaLyse Lysing Solution (Beckman Coulter, Brea, CA, USA) to lyse erythrocytes and then incubated with Fc Block (BD Biosciences, Franklin Lakes, NJ, USA) for 10 min on ice. For cell-surface markers, samples were incubated with antibodies for 15 min. For analysis of intracellular markers, samples were fixed and permeabilized with Foxp3/Transcription Factor Staining Buffer Set (Thermo Fisher Scientific), according to the manufacturer’s instructions, then incubated with antibodies for 15 min. After washing, samples were analyzed using LSRFortessa X-20 (BD Biosciences) and FlowJo software (BD Biosciences).
Antibodies for flow cytometry.
Histology and immunohistochemistry (IHC). Specimens were fixed for 24 h in 10% formalin/phosphate-buffered saline. Then, the specimens were transferred to 70% ethanol and embedded in paraffin. Histologic sections (4 μm) were stained with hematoxylin and eosin (HE) or processed for IHC analysis. IHC was performed using the antibodies shown in Table II. After blocking, the slides were incubated with primary antibodies overnight and then with secondary HRP-conjugated antibodies for 30 min. Visualization was performed by using the EnVision Detection System (Dako, Tokushima, Japan). Counterstaining was performed with hematoxylin.
Antibodies for immunohistochemistry.
Statistical analysis. Categorical variables were compared using Fisher’s exact test. Continuous variables were analyzed with Kruskal–Wallis test or the Mann–Whitney U-test. Graphs are shown as box-and-whisker plots (centerline, median; box limits, upper and lower quartiles; whiskers, 1.5× interquartile range; points, outliers). Differences for which a p-value was less than 0.05 were considered statistically significant. The statistical analysis was performed using the SPSS software program version 28.0 (IBM Japan Ltd., Tokyo, Japan).
Results
Tumor growth was suppressed in TFF1 KO mice. We first investigated a xenograft CRC mouse model in which MC38 cells were implanted into WT and TFF1KO mice. In this xenograft mouse model, the tumors constituted MC38 cells and were uniformly characterized across all mice, while the microenvironment of the tumors differed depending on the mice (WT and TFF1KO). In addition, recombinant TFF1-treated WT mice (subcutaneous injection) were analyzed, resulting in three groups of mice with a gradient amount of TFF1 highest in WT + TFF1 and lowest in TFF1KO (Figure 1A).
Tumor growth is suppressed in TFF1KO mice. (A) Schema of the mouse model. (B) Representative images of tumors in WT+TFF1, WT, and TFF1KO mice. Scale bars: 10 mm. (C) Tumor volume after mouse sacrifice. (D, E) Cleaved-caspase 3 expression in the tumor assessed using IHC. Scale bars: 100 μm. (F, G) Ki67 expression in the tumor assessed using IHC. Scale bars: 100 μm. *p<0.05, **p<0.01.
We first monitored tumor growth in WT + TFF1 (n=5), WT (n=15), and TFF1KO mice (n=17) for up to 21 days (Figure 1A). The greater tumor growth suppression observed in the TFF1KO mice than in the WT mice suggests a suppressive tumor microenvironment in the TFF1KO mice (Figure 1B and C). No significant difference existed between WT and WT + TFF1, suggesting that subcutaneously administered TFF1 lacked sufficient ability to regulate tumor size. To further investigate the cause of the difference in tumor growth, tumors were assessed using IHC. The number of cleaved caspase-3-positive cells, which represents apoptosis of tumor cells, was significantly higher in TFF1KO mice (Figure 1D and E). However, the labeling index of Ki67 did not show a significant difference between WT and TFF1KO tumor cells (Figure 1F and G). These data suggest that the difference in tumor growth was explained by cell apoptosis rather than proliferative ability.
Activated T cells accumulated in the tumor microenvironment in TFF1KO mice. Given that the tumor microenvironment resulted in cell apoptosis in TFF1KO mice, we hypothesized that tumor immunity might be activated by the loss of TFF1. To test for differences in tumor immunity in our mouse model, the tumor microenvironment and circulating immune cells in these mice were assessed using IHC and flow cytometry, respectively.
IHC analysis of the numbers of CD8-positive cytotoxic T cells and CD4-positive helper T cells in the tumor microenvironment, as expected, revealed markedly greater counts in the TFF1KO mice (Figure 2A and B), suggesting T-cell activation in the TFF1KO mice. Interestingly, CD8-positive cells were less in WT + TFF1 mice, indicating that TFF1 treatment can affect immune cells but lacks sufficient capability to regulate tumor size. Next, IHC analysis of granzyme B (a T-cell activation marker) and Tim3 (a T-cell exhaustion marker) showed that granzyme B-positive cells were most prevalent in TFF1KO mice, indicating TFF1’s immune-suppressive role. However, Tim3 expression did not differ between the groups (Figure 2C and D).
Activated T cells accumulate in the tumor microenvironment of TFF1KO mice. (A, B) Representative images and quantification of IHC staining for CD8 and CD4 in the tumor. Scale bars: 100 μm. (C, D) Representative images and quantification of IHC staining for granzyme B and Tim3 in the tumor. Scale bars: 100 μm. *p<0.05, **p<0.01.
Next, to explore the role of PD-L1 expression in immune cells, the proportion of PD-L1-positive cells among CD8-positive cells was assessed using IHC. While PD-L1 was extensively expressed by the tumor cells, there was a certain number of CD8-positive T cells that coexpressed PD-L1. When compared between the groups, PD-L1/CD8 cells were more frequently found in TFF1KO mice (Figure 3), indicating that a high number of PD-L1/CD8 cells represented an active immune status in the tumor microenvironment.
PD-L1 expression in CD8-positive T cells. (A) Representative images of double immunohistochemistry (IHC) for CD8 and PD-L1 in the tumor microenvironment. Arrows indicate double-positive cells. (B) Quantification of IHC images. Scale bars: 100 μm. **p<0.01.
Circulating T cells do not represent the tumor micro-environment. The results of our mouse model described above indicate not only that tumor immunity is activated in TFF1KO mice but also that CD4/CD8/PD-L1 expression in the tumor microenvironment is associated with tumor immunity. Next, we analyzed blood samples to determine whether the immune cells in peripheral blood showed immune activity. The blood samples were analyzed using flow cytometry (gating strategy is shown in Figure 4A), and no significant differences in CD8 and CD4 cells between WT and TFF1KO mice were found (Figure 4B). Then, Tim3 (a marker for T-cell exhaustion) and granzyme B (a marker for T-cell activation) were analyzed, revealing no significant differences between WT and TFF1KO mice (Figure 4C). The number of PD-L1/CD8-positive cells did not show a significant difference between the groups (Figure 4D), suggesting that circulating active T cells did not represent the tumor microenvironment. In contrast, the number of Foxp3-positive cells (Treg) decreased, depending on the amount of TFF1, and was lowest in TFF1KO mice (Figure 4E), supporting the idea that immune suppression was up-regulated by TFF1.
Circulating T cells do not represent the tumor microenvironment. (A) Gating strategy for circulating lymphocytes. (B) Ratio of CD8 and CD4 cells in CD3-positive lymphocytes. (C) Ratio of Tim3 and granzyme B cells in CD8-positive cells. (D) Ratio of PD-L1 cells in CD8-positive T cells. (E) Ratio of Foxp3 cells in CD4-positive T cells. *p<0.05, **p<0.01.
Checkpoint inhibitors worked favorably in TFF1KO mice. Given that tumor immunity is activated in TFF1KO mice, we hypothesized that checkpoint-inhibitor therapy might be more effective in TFF1-deficient mice. To examine this hypothesis, WT (n=14) and TFF1 KO (n=16) mice with MC38 tumors were treated with an anti-PD-1 antibody (Figure 5A). Tumor growth was suppressed in both groups, and disappeared (commonly described as achieving a complete response) more frequently in TFF1 KO mice (43.8%; n=7/16) than in WT mice (7.1%; n=1/14) (Figure 5B and C). Because the tumor microenvironment was not available due to the disappearance of the tumor in TFF1KO mice, the immune activity in these mice was investigated by assaying peripheral immune cells. As expected, CD4 and CD8 cells did not differ between WT and TFF1KO mice (data not shown), whereas the number of Foxp3-positive Tregs was significantly lower in TFF1KO mice (Figure 5D), which is consistent with Figure 4E. Interestingly, TFF1KO mice exhibited a significantly increased number of PD-L1/CD8-positive cells (Figure 5E). These results indicate that the number of circulating PD-L1/CD8 cells represents immune activity in this setting and could be a marker of the efficacy of checkpoint inhibitors.
Checkpoint inhibitor works favorably in TFF1KO mice. (A) Schema of mouse models. (B) Representative images of the tumors in WT and TFF1KO mice. Scale bars: 10 mm. (C) Percentage of complete response to anti-PD-1 antibody. (D) Ratio of Foxp3 cells in CD4-positive cells. (E) Ratio of PD-L1 cells in CD8-positive cells. *p<0.05, **p<0.01.
DCs were frequently found in the tumor microenvironment in TFF1 KO mice. To further investigate the mechanisms by which PD-L1/CD8 cells are induced, DCs in the tumor microenvironment and peripheral blood were analyzed. CD11c expression in the tumors was assessed using IHC to analyze the number of DCs and the expression of PD-L1. As anticipated, the number of CD11c-positive cells were highest in TFF1KO mice (Figure 6A and B), and double IHC revealed that the number of PD-L1/CD11c cells increased gradually, depending on the loss of the amount of TFF1 (Figure 6C and D). These data suggest that antigen presentation occurred more frequently in TFF1KO mice than in WT mice, resulting in frequent T-cell priming and activation of tumor immunity in TFF1KO mice.
Dendritic cells (DCs) are frequently found in the tumor microenvironment of TFF1KO mice. (A) Representative images of IHC for CD11c in the tumor. Scale bars: 100 μm. (B) Quantification of staining in the images of (A). (C) Representative images of double IHC for CD11c and PD-L1 in the tumor. Arrows indicate double-positive cells. Scale bars: 100 μm. (D) Quantification of images of (C). *p<0.05, **p<0.01.
Circulating DCs represent immunoactivity. To evaluate whether the status of DCs found in the tumor microenvironment can be recaptured in peripheral blood, flow cytometry was performed on WT + TFF1 (n=5), WT (n=10), and TFF1KO (n=12) mice. The gating strategy is shown in Figure 7A, in which CD11c- and MHC-II-positive cells were recognized as DCs. The TFF1KO mice showed the highest number of DCs in the peripheral blood (Figure 7B), suggesting that circulating DCs represent the tumor microenvironment. In addition, the number of PD-L1/DCs showed a progressive increase, depending on the loss of the amount of TFF1 (Figure 7C), and when the anti-PD-1 antibody-treated mice were evaluated, PD-L1/DCs were more frequently found in TFF1KO mice than in WT mice (Figure 7D). These results indicate that circulating PD-L1/DCs cells represent immune activity against CRC tumors, while circulating PD-L1/CD8 cells do not.
Circulating dendritic cells (DCs) represent immunoactivity. (A) Gating strategy of DCs. (B) Ratio of DCs in CD45-positive WBCs. (C) Ratio of PD-L1 in CD11c-positive DCs. (D) Ratio of PD-L1 in CD11c-positive DCs found in anti-PD-1 antibody-treated mice. *p<0.05, **p<0.01.
Number of PD-L1/CD11c depends on tumor volume and host TFF1. We then hypothesized that the number of PD-L1/CD11c cells might depend on the tumor volume. Thus, the tumor volume and PD-L1/CD11c cells were plotted for all mice groups, revealing a strong association between them (r=0.631 and p<0.01 in WT mice; r=0.778 and p<0.01 in TFF1KO mice, r=0.953 and p<0.05 in WT+TFF1 mice, respectively, Figure 8A). As tumor volume increased, the number of PD-L1/CD11c cells also rose, with this trend being more pronounced in TFF1KO mice and less pronounced in WT + TFF1 mice. We then compared the ratio of tumor volume and PD-L1/CD11c cells (calculated by tumor volume divided by PD-L1/CD11c cell number) and found a progressive decrease depending on the loss of the amount of TFF1 (Figure 8B). These data suggest that the opportunity for antigen presentation increases not only with larger tumor volumes but also with reduced TFF1 burden, ultimately leading to T-cell activation and suppressed tumor growth.
Number of PD-L1/CD11c depends on tumor volume and TFF1 burden. (A) Scatter plot of tumor volume and PD-L1/DCs in WT+TFF1, WT, and TFF1KO mice. (B) Tumor volume/PD-L1 ratio. *p<0.05, **p<0.01.
Discussion
Our results revealed that loss of the amount of TFF1 resulted in an active immune response, suppressed tumor growth, and a favorable effect of checkpoint inhibitors in a CRC mouse model. While the activity of immune cells in TFF1KO mice was clearly up-regulated in the tumor microenvironment, T-cell activation cannot be identified in the peripheral blood, probably because these cells remain in the tumor microenvironment. However, circulating PD-L1/DCs represent the tumor microenvironment, suggesting the possible usefulness of PD-L1 expression on DCs in predicting the efficacy of checkpoint-inhibitor treatment for patients with CRC.
Immunotherapy with PD-1/PD-L1 inhibitors has been successful in a subset of patients suffering from several cancer types, and recent studies have revealed the importance of PD-L1 expression in immune cells rather than tumor cells. Some studies have revealed that PD-L1 expression on lymphocytes also contributes to the exhaustion of T cells (17), while other prospective clinical studies have revealed that patients with high levels of PD-L1 on T cells respond well to checkpoint inhibitor therapy (16, 28). Retrospective clinical studies also suggested that patients with high PD-L1 expression on tumor infiltrating lymphocytes (TILs) had a better prognosis than those with low expression (29-31). Thus, the role of PD-L1 expression on immune cells has been controversial. Of note, recent studies using PD-L1−/− mice revealed that host expression of PD-L1, especially in antigen-presenting cells (APCs), is essential for the efficacy of checkpoint blockade therapy (32, 33). Our study also revealed that a high number of PD-L1-expressing DCs (found in TFF1KO mice and large tumors) was associated with the efficacy of anti-PD-1 therapy. Taking these observations together, one can assume that DCs express PD-L1 upon the uptake of tumor antigen, but the activation of DCs might be suppressed by PD-L1 expression. Thus, the administration of checkpoint blockade therapy can reestablish the immune activity of these PD-L1-expressing DCs. Our study indicated that PD-L1-expressing DCs can be used as a marker to predict the efficacy of checkpoint inhibitor treatment; however, it remains elusive whether PD-L1 on DCs contributes to the activation or suppression of tumor immunity in our mouse model. Further experiments are needed to clarify the mechanisms by which PD-L1-positive DCs contribute to immune activity.
While the role of TFF1 in the development of malignant disease has been controversial, recent investigations using TFF1-deficient mice support the tumor-suppressive role of TFF1. The main findings of these studies are that 1) TFF1 is expressed in gastrointestinal epithelial cells, and 2) loss of TFF1 in these epithelial cells results in carcinogenesis. This evidence is convincing but does not necessarily mean that the activity of TFF1 is limited to epithelial cells. TFF1 is a secreted protein that exists not only in epithelial cells but also in the bloodstream (34). We hypothesized that host TFF1 might be responsible for the systemic regulation of organs and/or diseases, including the immune response to malignant diseases. As shown in this study, host TFF1 was associated with tumor immunity, supporting our hypothesis. Surprisingly, however, the effect of host TFF1 activity contradicts the tumor-suppressive role of TFF1 in epithelial tissue, wherein host TFF1 is observed to suppresses the immune response, thus supporting the growth of CRC. From an oncological perspective, the actions of TFF1 appear contradictory; however, one can assume that host and epithelial TFF1 share a similar function, that is, to suppress certain kinds of cellular activity. Epithelial TFF1 inhibits the malignant activity of epithelial cells, such as proliferation and invasion, whereas host TFF1 down-regulates immune cell activity. Moreover, TFF1 might influence immune cell migration, as was found in epithelial cells. Further experiments are needed to reveal the functional mechanisms by which TFF1 regulates the cellular activity of various cells, organs, and biological systems.
Conclusion
We found that PD-L1 expression on T cells and DCs was mediated by TFF1 and represented immune activity in a CRC mouse model. Further studies are needed to evaluate TFF1 inhibitors as candidates to support immunotherapy for CRC patients.
Footnotes
Authors’ Contributions
Conceptualization, J.Y., A.O.; Methodology, Resources, Writing, Funding Acquisition, J.Y.; Investigation, Visualization, T.J.; Data Curation, A.O., M.S., T.B., Y.M.; Supervision, Y.Y. and T.K.; Project Administration, E.T.
Conflicts of Interest
The Authors have no conflicts of interest to declare in relation to this study.
Funding
This work was supported by the Japan Society for the Promotion of Science (JSPS) and Grants-in-Aid for Scientific Research (KAKENHI) (grant numbers 20H03751).
- Received July 5, 2024.
- Revision received July 22, 2024.
- Accepted July 23, 2024.
- Copyright © 2024 The Author(s). Published by the International Institute of Anticancer Research.
This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY-NC-ND) 4.0 international license (https://creativecommons.org/licenses/by-nc-nd/4.0).
References
In this issue
Jump to section
Related Articles
Cited By...
- No citing articles found.