Repeated cycles of 5-fluorouracil chemotherapy impaired anti-tumor functions of cytotoxic T cells in a CT26 tumor-bearing mouse model

Clinically, 5-FU based chemotherapy for treatment of colorectal cancer contains approximately eight cycles. To mimic the clinical schedule and investigate the therapeutic effect in our colon cancer model, CT26 cells were inoculated and tumor-bearing mice were treated with 5-FU for one to four cycles (Fig. 1a). Tumor volumes and OS were monitored. The maximum dosage for the first cycle is three injections of 5-FU (40 mg/kg). Four injections per cycle for the followed cycles were generally well tolerated. One cycle of 5-FU (C1) treatment could inhibit tumor growth compared with the untreated control group, and more than one cycle of treatment, namely two cycles (C2), three cycles (C3) and four cycles (C4), inhibited tumor growth to a greater degree than C1 (Fig. 1b; Additional file 1: Raw data Fig.1b). The OS of any of the treated groups was longer than the control group but not significantly different between treatment groups (Fig. 1c; Additional file 1: Raw data Fig 1c). Tumor growth was inhibited during chemotherapy but progressed after treatment suspension. Repeated cycles of 5-FU did not improve OS compared with one-cycle treatment. Non-durable immune responses against tumors after chemotherapy might be one of the reasons for unimproved OS, and therefore, we analyzed the immune status after different cycles of 5-FU in our model.

The blood cell count is an indicator for whether to continue chemotherapy in the clinic. We analyzed the absolute number of different immune cells in the spleens and their percentages in the tumors using flow cytometry. The spleen was usually enlarged with tumor growth (data not shown), but the total number of gradient-separated lymphocytes was not significantly different between the treated groups and their controls after a 7-day rest since last 5-FU injection (Fig. 2a; Additional file 2: Raw data Fig 2). CD4 T-cells (Fig. 2b) and B cells (Fig. 2e) were decreased after three cycles of 5-FU. The amount of CD8 T-cells (Fig. 2c) and NK cells (Fig. 2d) in the spleen did not decline significantly after treatment. Although 5-FU was reported to kill MDSCs, resulting in enhanced T cell-dependent antitumor immunity [22], the number of immunosuppressive cells, including MDSCs and Tregs, was not decreased significantly on day 7 after last 5-FU per cycle in our treatment groups (Fig. 2f, g). The percentages of immune subpopulations between the treatment and control groups were not significantly different except for B cells after C3 treatment (Additional file 3: Figure S1 and Additional file 4: Figure S2). Different immune subpopulations have different developmental pathways and recovery rates [23, 24]. B cells may not recover at the detected time point after three-cycle chemotherapy compared with other immune cells in our model. Alternatively, the recovery of B cells might have been impaired after 5-FU C3 treatment. The number of CD4 T-cells and B cells decreased, but cytokines primarily produced by CD4 T-cells (e.g., IL-10, IL-6 and TGF-β) or IgA secreted by B cells [25] were not decreased after C3 treatment (Additional file 5: Figure S4 and Additional file 6: Figure S5). In general, this observation might indicate that the acute myeloid suppression of 5-FU was probably relieved after the 7-day rest. In addition, immune cells that infiltrated into the tumor bed also should be considered. Because the infiltration of immune cells usually reached a peak 2 days after chemotherapy [26], we analyzed their distribution in the tumors on day 3 after last 5-FU injection. Compared with the control groups, the amount of CD45⁺ immune cells, including NK cells, was not changed significantly in the chemotherapy groups (Fig. 3a, c; Additional file 2: Raw data Fig 3). The percentage of tumor infiltrating CD8 T-cells was increased after 5-FU C1 treatment but not in the C3 group (Fig. 3b). Furthermore, the expression of PD-L1 (also called B7-H1) in tumor cells was analyzed, which could lead the anergy of activated T-cells [27]. In our treatment model, a greater number of non-immune cells (CD45⁻ cells, primarily including tumor cells and fibroblast cells) expressed PD-L1 after chemotherapy (Fig. 3d). This observation might imply that anti-tumor immune functions were repressed, and the tumor could relapse after suspension of 5-FU treatment. Elevated PD-L1 after 5-FU C1 treatment also indicated that early intervention of unanticipated aspects of chemotherapy by immune strategies might be needed.

To identify specific anti-tumor immune functions after C1 and C3, the proliferation and cytotoxicity of spleen cells against CT26 were assayed. First, we investigated the proliferation of lymphocytes from different treatment groups using CFSE assays and analyzed the absolute number of proliferated lymphocytes against CT26 (the definition of proliferated cells is illustrated in Additional file 7: Figure S3). The total proliferated CD8 T-cells increased after C1 treatment but decreased after C3 compared with untreated groups, respectively (Fig. 4b; Additional file 2: Raw data Fig 4). The proliferated CD8 T-cells were tumor-specific clones because the proliferation was stimulated by CT26. Other cells, including CD4 T-cells (Fig. 4a), NK cells (Fig. 4c) and B cells (Fig. 4d), were not significantly different between the treated and control groups.

Second, we determined the cytotoxicity of spleen cells against CT26. Lymphocytes from the 5-FU C1 and control group showed 19.5(±0.5) % and 12.4(±1.1) % cytotoxicity against CT26, respectively. In contrast, lymphocytes from 5-FU C3 showed only 8.7(±1.5) % cytotoxicity against CT26 cells, whereas the control group remained at 13.3(±1.3) % (Fig. 5a; A dditional file 8: Raw data Fig 5). The cytotoxicity of lymphocytes from C3 was the lowest. The changes of cytotoxicity was CT26 cell-specific because no significant difference of cytotoxicity was observed between the C3 treated and control groups against 4 T1 cells (Fig. 5b), which is from the synergetic Balb/C mice. This cytotoxicity was executed primarily by CD8 T-cells because the cytotoxicity against YAC-1 (NK sensitive target cells) was approximately 3 % in both groups (Fig. 5c). The decreased cytotoxicity was due to impaired anti-tumor responses after 5-FU C3 treatment because the ratios of CD8 T-cells were not significantly different between the treated and control groups (Additional file 4: Figure S2B).

Third, the culture supernatants of MLTC from different groups were collected to examine the released cytokines using ELISA assays. IFN-γis a well-known anti-tumor cytokine, and its content was higher in 5-FU C1 but decreased in C3 compared with their control groups. Tumor-promoted cytokines such as IL-10, IL-6, TGF-βand IgA in the medium of MLTC were not reduced after 5-FU C3 treatment (Additional file 5: Figure S4A-D; Additional file 8: Raw data Fig S4). Cytokines in serum were also determined by ELISA assays. TGF-β, an immunosuppressive and tumor-promoted cytokine, was increased after 5-FU treatment (Fig. 5e), similar to radiation- and doxorubicin-treated tumor-bearing MMTV/PyVmT mice [28]. The concentration of IL-6 was improved whereas IL-10 and IgA were not changed after 5-FU C3 (Additional file 6: Figure S5A–C; Additional file 8: Raw data Fig S5). Increased IL-6 after 5-FU C3 might help to promote chronic inflammation and residual tumor survival, which was a negative factor in anti-tumor responses. IFN-γ in the serum was too low to be detected (data not shown).

A single round of 5-FU promoted the proliferation and cytotoxicity of CD8T-cells, whereas repeated cycles of chemotherapy impaired anti-tumor immune functions. The impaired tumor-specific responses might be a critical reason underlying non-durable anti-tumor activity. Therefore, the immediate combination of immunotherapy rather than repeated chemotherapy may help to improve and prolong the anti-tumor effect in cancer treatment. Because PD-L1 was more highly expressed after chemotherapy (Fig. 3d), PD-L1 monoclonal antibodies (αPD-L1) and killer cells (CIKs) were administered after 5-FU C1 (Fig. 6a). As shown in Fig. 6b (Additional file 9 : Raw data Fig 6 ), the combined therapy displayed a greater inhibition of tumor growth compared with chemotherapy or immunotherapy alone. The OS of the combination treatment group was also improved (Fig. 6c). Indeed, half of the mice in the combined group were tumor-free for more than 6 weeks after tumor regression (data not shown).

Male 7-week-old Balb/C mice purchased from Vital River Laboratory Animal Technology Co. Ltd. (Beijing, China) were maintained under specific pathogen-free conditions in the animal facility of Cancer Institute, Chinese Academy of Medical Sciences (CAMS). All procedures for animal experiments were approved by the Institutional Animal Use and Care Committee of CAMS. CT26 cells (a colon cancer cell line from Balb/C mice) were obtained from the Cell Resource Center, Peking Union Medical College (which is the headquarters of National Infrastructure of Cell Line Resource, NSTI) on Jan 10, 2015. The cell line was determined to be free of mycoplasma contamination by PCR and culture. Its species origin was confirmed with PCR. The identity of the cell line was authenticated with STR profiling (FBI, CODIS). All results can be viewed on the website (http://​www.​cellresource.​cn/​). 4 T1 cells (a breast cancer cell line from Balb/C mice) and YAC-1 (a mouse lymphoma cell line) cells were purchased from American Type Culture Collection (ATCC), (Manassas, VA, USA) and maintained in our laboratory. CT26 and YAC-1 cells were cultured in RPMI 1640 medium, and 4 T1 cells were cultured in DMEM/F12 medium (Hyclone, Thermo Fisher Scientific, USA). Cells were maintained in basic medium supplemented with 10 % FBS and penicillin/streptomycin at 37 °C in a humidified atmosphere of 5 % CO₂.

CT26 cells (3 × 10⁵) were inoculated subcutaneously (s.c.) in the right flanks of Balb/C mice. Tumor growth was monitored every 3–4 days by palpation. Tumor diameters were measured twice weekly and used to calculate tumor volumes, as described previously [19]. Mouse survival was monitored every other day.

Approximately 2 weeks after tumor inoculation, 5-FU (40 mg/kg) was injected intraperitoneally (i.p.) every other day for a total of three injections for the first cycle treatment (C1). A week after the last 5-FU injection, 5-FU 40 mg/kg was injected every day for a total of four injections for the second cycle (C2) treatment. The third cycle (C3) and fourth cycle (C4) were as same as C2.

For immunotherapy, 1 × 10⁷ CIK cells (200 μl) and anti-PD-L1 mAb (200 μg/mouse) were injected into tumor and peritoneally respectively after C1 at a 3-day interval for six injections.

Splenic cells were isolated by gradient centrifugation using lymphocyte separation medium (Dakewe Biotech, Shenzhen, China). Tumor tissues were minced and digested in RPMI containing 2 % FBS, 1 mg/ml type IV collagenase (Sigma Aldrich) and 300 U/ml DNase I (Sigma Aldrich) for 2 h at 37 °C, and passed through a cell strainer to achieve cell suspension.

For surface staining, cells were suspended in staining buffer and incubated for 30 min at 4 °C in the dark with fluorophore-conjugated anti-mouse mAbs: APC-anti-CD3, PerCP-Cy5.5-anti-CD4, PE-anti-CD8, PE-anti-CD19, PE-anti-CD49b, PE-Cy5-anti-CD11b, PE-anti-Gr-1, and APC-anti-PD-L1. For intracellular Foxp3 staining, cells were fixed and permeabilized according to the manufacturer’s protocol and incubated with PE-conjugated anti-mouse Foxp3 antibodies for 30 min at 4 °C in the dark. All antibodies and Foxp3 fixation/permeabilization kit were purchased from Biolegend. Cells were acquired by a flow cytometer (BD LSRII) and analyzed using Flowjo software.

To analyze the proliferation of different subsets of lymphocytes, separated splenic cells were labeled with CFSE (Life Technologies) according to the manufacturer’s protocol and incubated with mytomycin C (MMC)-treated CT26 cells at a responder:stimulator (R:S) ratio of 10:1. Three days later, cells were collected and stained with the mixture of fluorophore-conjugated anti-mouse mAbs as indicated and analyzed by flow cytometry.

Seven days after the last 5-FU injection of C1 and C3, splenic cells from treated and non-treated tumor-bearing mice were prepared as effector cells. CT26, 4 T1 and YAC-1 were used as the targeted cells, respectively. As described previously [19], target cells were labeled with CFSE and mixed with effector cells at an effector:target (E:T) ratio of 25:1. The mixed cells were spun down and incubated at 37 °C for 4 h. PI (Sigma Aldrich) was added for DNA labeling of dead cells at the end of the incubation period. Samples were analyzed by flow cytometer within 30 min.

As described above [19], splenic cells from differently treated and non-treated tumor-bearing mice were incubated with MMC inactivated CT26 (alive but not proliferative) at a R:S ratio of 10:1. The supernatant of the mixed lymphocyte and tumor cell culture (MLTC) was collected on day 3, and the concentrations of IFN-γ were determined using the mouse IFN-γ ELISA kit (Dakewe Biotech, Shenzhen, China). Serum from treated and control mice were collected and detected via the TGF-β ELISA kit (NeoBioscience Ltd., Beijing, China).

CIKs were generated as previously described [40]. Briefly, lymphocytes from CT26-bearing mice were stimulated with recombinant mouse IFN-γ (1000 U/ml; Peprotech) for 24 h, then transferred to anti-CD3 (145-2C11; Biolegend) pre-coated tissue-culture flasks, and stimulated with 300 IU/ml recombinant human IL-2 every 3 days until cells were harvested on day 10. Anti-PD-L1 (clone 10B5) hybridoma was kindly provided by Shengdian Wang Lab (Institute of Biophysics, Chinese Academy of Sciences, Beijing, China). Anti-mouse PD-L1 monoclonal antibodies (αPD-L1) were purified from ascites of nude mice [41].

The statistical significance between groups was determined by Student’s t-test, and tumor volumes were analyzed using two-way ANOVA. The Kaplan-Meier survival plot was assessed for significance using the log-rank test. All statistical analyses were performed using GraphPad Prism software version 5 (GraphPad Software, Inc.), and P < 0.05 was considered significant.