IFN-γ down-regulates the PD-1 expression and assist nivolumab in PD-1-blockade effect on CD8+ T-lymphocytes in pancreatic cancer

We collected 88 peripheral T lymphocytes samples, of which 48 were obtained from pancreatic ductal adenocarcinoma (PDAC) patients (23 female and 25 male patients with a median age of 62 years; age range, 41–76 years), and 40 were obtained from healthy donors (20 female and 20 male patients with a median age of 60 years; age range, 40–79 years) at the Department of General Surgery, Sir Run Run Shaw Hospital between December 2016 and May 2017. The patients with PDAC enrolled in this study only if newly diagnosed, confirmed by pathological diagnosis and underwent radical surgery. None of them had acute or chronic infections, inflammatory processes, a history of autoimmune disease, or received radiotherapy, chemotherapy, or immunotherapy before surgery. Four of the 48 PDAC patients donated 50 ml peripheral blood for subsequent T-lymphocytes culture and biological experiments. All healthy donors and patients or their guardians provided written informed consent for scientific research statement. All experiments were approved by the Research Ethics Committee of Sir Run Run Shaw Hospital, School of Medicine, Zhejiang University. All of the research protocols were carried out in accordance with approved guidelines of the Sir Run Run Shaw Hospital, School of Medicine, Zhejiang University.

For peripheral T-lymphocytes: EDTA-anticoagulated peripheral blood was stained with with APC-conjugated mouse anti-human CD3 antibody (BD Pharmingen), FITC-conjugated anti-human CD8 antibody (BD Pharmingen), PE-conjugated mouse anti-human CD279 antibody (BD Pharmingen) or isotype control antibodies. After incubation for 30 min at 4 °C in dark. For cultured T-lymphocytes: 2 × 10⁵ cells were stained with 5 μl fluorochrome-conjugated antibodies or isotype control antibodies mentioned above for 30 min at 4 °C in dark. Next, the cells were washed twice with cold PBS. Washed cells were assayed on an BD LSRFortessa flow cytometer (BD Biosciences). Data were analyzed using FlowJo 10.0.7 software.

Peripheral blood mononuclear cells (PBMCs) were obtained from the venous blood by using a lymphocyte-separating medium (Ficoll-Paque, MP Biomedicals, Carlsbad, USA). All the PBMCs were incubated in a humidified incubator at 37 °C in 5% CO₂ for 2 h. Then, suspended cells were separated and collected to culture. The density of cells was adjusted to 1 × 10⁶/mL with 5 ml RPMI 1640 medium (Gibco) containing 5% heat-inactivated autoserum and rhIL-2 (500 U/ml, PeproTech) and incubated in flask, in which was pre-coated with OKT3 (100 ng/ml, Miltenyi Biotec), at 37 °C in 5% CO₂ for 24 h. Fresh IL-2 and medium were added to each group every 2 days. Human pancreatic cancer cell line PANC-1, BxPC-3 and MIAPaCa-2 were obtained from Chinese Academy of Sciences (Shanghai, China). PANC-1 cells were cultured in DMEM medium (Gibco, USA) supplemented with 4.5 g/L glucose and 10% fetal bovine serum (Gibco, USA) at 37 °C in presence of 5% CO₂. BxPC-3 cells were cultured in RPMI 1640 medium (Gibco, USA) supplemented with 10% fetal bovine serum (Gibco, USA) at 37 °C in presence of 5% CO₂. MIAPaCa-2 cells were cultured in DMEM medium (Gibco, USA) supplemented with and 10% fetal bovine serum (Gibco, USA) at 37 °C in presence of 5% CO₂.

The proliferation and viability of cultured T cells were detected using the trypan blue exclusion method and the automated cell counter system (TC20 automated cell counter, Bio-Rad) according the manuscript.

T-lymphocytes from each group were cultured on 6-well plates at concentration of 1 × 10⁶/well for 24 h. The cytokine concentrations of IFN-γ, TNF-α and IL-2 in the supernatants were detected and quantified using Cytometric Bead Array Human Th1/Th2 Cytokine Kit II (BD Biosciences, USA) with a flow cytometry system according to the manufacturers instruction as previous research [26].

PANC-1, BxPC-3 and MIAPaCa-2 cells were used as target cells, stimulated group or control group cells were used as effector cells mixed in the proportion 5:1, 10:1 and 20:1. Target cells were cultured on 96-well plates at concentration of 1 × 10⁵/ml in 0.1 ml of RPMI 1640 medium containing 10% heat-inactivated fetal bovine serum for 24 h. Effector cells were inoculated onto the culture plate according to effector/target ratio in 0.1 ml of respective medium that were described in the cell culture. Some wells containing only effector cells or target cells were set as controls. There were three parallel wells for each cell count. Culture plates were placed in a humidified incubator at 37 °C in 5% CO2 for 48 h. Next, the in vitro cytotoxicity of the cells against the pancreatic cancer cells was determined using a Cell Couning Kit 8 (CCK-8, Dojindo). Optical density (OD) of each well was read at wave length of 450 nm after 4 h incubated, and cytotoxic activity was calculated as follows: Cytotoxic activity, % = [1 - (OD_{effect and target cells} - OD_{effector cells})/OD_{target cells}] × 100.

All mice were obtained from the animal unit of Zhejiang University (Zhejiang, China). BALB/C nude mice (4–6 weeks old) were used in all experiments. All animal experiments were carried out in the animal unit of Zhejiang University (Zhejiang, China) according to procedures authorized and specifically approved by the animal ethics committee of Zhejiang University (Reference Number: ZJU20170126). Mice were sacrificed by CO₂ inhalation or cervical dislocation at desired time points. Subcutaneous PDAC mouse models were established by subcutaneous injection of BxPC-3 cells (3 × 10⁶) into the left axilla of BALB/C nude mice. Then, subcutaneous tumors with a longitudinal diameter of 1 cm were peeled from subcutaneous mouse models after sacrificed. Four mice were used in each group for ectopic studies. Tumor tissues were washed in D-Hanks’ buffer. Necrotic tissues were removed from tumors, and tumor tissues were cut into about 1mm³ pieces. One tumor piece was implanted in the right axilla of recipient BALB/C nude mice under anesthesia. Three days after subcutaneous implantation, the mice were subjected to adoptive transfer therapy of T cells, and the tumor growth was recorded. Tumor volumes for each mouse were monitored with a caliper, every 4 days or 7 days, by measuring in two directions (length and width). The volume was calculated as length × (width)² / 2. After the experiment, after anesthesia with 100% carbon dioxide, the mice were sacrificed by CO2 inhalation or cervical dislocation, and then the tumor was removed. Immunofluorescence staining was used in these tumors to detect the infiltration of CD8+ T cells.

All the transplanted tumor samples were fixed by 4% PFA at 4 °C overnight and embedded into paraffin. Paraffin-embedded tissues were sectioned into 5 μm sections. The sections were deparaffinized and rehydrated by dimethylbenzene, gradient ethanol series and double-distilled water. After washed by PBS three times for 5 min, antigen retrieval was performed by boiling the sections in citric acid buffer (PH6.0) for 15 min. Cooled sections were washed by PBS, blocked by 5% normal goat serum for 45 min, and incubated with rabbit anti-human CD8 antibody at dilution of 1:100 overnight at 4 °C. Next day, the sections were stained by Alexa Fluor 594-conjugated goat anti-Rabbit IgG antibody (ab150084, Abcam), followed by DAPI staining. Slides were observed under Zeiss Observer A1 microscope. The mean number of CD8+ T cells in four microscopic fields of 40x objective was scored independently by two authors in a blinded manner.

We used flow cytometry to compare the levels of PD-1 expression on peripheral CD8+ T lymphocytes in 48 patients newly diagnosed with pancreatic ductal adenocarcinoma (PDAC) and 40 healthy donors (Fig. 1a). PD-1 was expressed at significantly higher levels on CD8+ T lymphocytes from patients with PDAC than from healthy donors (PDAC vs. healthy donors, 52.39 ± 2.20 vs. 39.43 ± 2.45, respectively; p < 0.001) (Fig. 1b). These results suggest that increased expression of PD-1/PD-L1 may be associated with the evolution and progression of PDAC.

To determine whether a PD-1 checkpoint-blockading agent enhances the immunological function of primary T lymphocytes from patients with PDAC, we isolated these patients’ peripheral T lymphocytes that expressed high levels of PD-1 and cultured them with different concentrations of nivolumab on the first day of induction. The T lymphocytes were treated with human immunoglobulin G4 as a negative control or with pembrolizumab as a positive control. The T lymphocytes were also treated with a CD3-activating antibody followed by the addition of recombinant human IL-2 to the culture medium. Cultured T lymphocytes were extracted on day 7, and their immune phenotypes, particularly PD-1 expression, were measured using flow cytometry. We found that PD-1 expression on CD8+ T lymphocytes decreased in a concentration-dependent manner and that the best blocking effect was achieved with 10 μg/ml of nivolumab (Fig. 2a).

As an immune adjuvant, IFN-γ enhances the immunogenicity of tumor vaccines and promotes the immune response of antigen-specific T cells. To detect the effect of IFN-γ on PD-1 expression by T lymphocytes in patients with pancreatic cancer, we added IFN-γ at a concentration of 1000 U/ml to the culture medium on day 1 during T-lymphocyte induction. T lymphocytes were isolated from the same four patients and treated with the CD3-activating antibody and recombinant human IL-2 as described above. Cultured T lymphocytes were extracted on day 7, and PD-1 expression was determined by flow cytometry (Fig. 2b). The immune phenotype of PD-1 expression of T lymphocytes was lower than that of cells not treated with IFN-γ (18.63 ± 0.82 vs. 47.38 ± 1.69, respectively; p < 0.0001) (Fig. 2c). However, the inhibitory effect of IFN-γ on PD-1 expression was less than that achieved with 0.1 μg/ml (18.63 ± 0.82 vs. 13.65 ± 1.22, respectively; p < 0.05) (Fig. 3b) or 10 μg/ml of nivolumab (18.63 ± 0.82 vs. 1.94 ± 0.31, respectively; p < 0.0001) (Fig. 2c).

To maximally inhibit the PD-1 checkpoint, we combined IFN-γ with 10 μg/ml of nivolumab to activate primary T lymphocytes (10-μg/ml mAb + IFN-γ group). On day 7 of culture, we used flow cytometry to detect PD-1 expression on CD8+ T lymphocytes and found that the 10-μg/ml mAb + IFN-γ group showed lower levels of PD-1 expression than the IFN-γ group (1.47 ± 0.47 vs. 18.63 ± 0.82, respectively; p < 0.0001) (Fig. 2c). Based on this result, we further tested the inhibitory effect of 0.1 μg/ml of nivolumab, a lower concentration, combined with IFN-γ (0.1-μg/ml mAb + IFN-γ group). Unexpectedly, we found no difference in the PD-1-blockade effect between the 0.1-μg/ml mAb + IFN-γ group and 10-μg/ml mAb + IFN-γ group (1.73 ± 0.78 vs. 1.47 ± 0.47, respectively; p = 0.78) (Fig. 2c). These results suggest that combination with IFN-γ allows a reduction of the dose of nivolumab and might minimize the adverse effects of nivolumab monotherapy. Moreover, the PD-1 expression was lower in the 0.1-μg/ml mAb + IFN-γ group than in the IFN-γ group (1.73 ± 0.78 vs. 18.63 ± 0.82, respectively; p < 0.0001) (Fig. 2c) and 0.1-μg/ml mAb group (1.73 ± 0.78 vs. 13.65 ± 1.22, respectively; p < 0.001) (Fig. 2c). This result further demonstrates that combining nivolumab and IFN-γ inhibits PD-1 expression greater than does nivolumab or IFN-γ as monotherapy.

To test the effects of IFN-γ on proliferation, cell viability, and cell density, we used the trypan blue exclusion method and an automatic counter to analyze the cells on days 1, 7, and 14. Peripheral T lymphocytes were adjusted to 1 × 10⁶/ml and 5 ml/group on day 1. During the extended culture time, the proliferation rates differed (Fig. 3a). Combined treatment with 0.1 μg/ml nivolumab and IFN-γ yielded the largest fold-increase in the number of viable cells compared with those of the control (Ctrl) group, IFN-γ group, and 0.1-μg/ml mAb group. However, there was no significant difference between the 0.1-μg/ml mAb group and Ctrl group (day 7: 4.85 ± 0.24 vs. 4.92 ± 0.10 × 10⁶, respectively, p = 0.79; day 14: 17.27 ± 1.30 vs. 16.82 ± 0.64 × 10⁶, respectively; p = 0.77) (Fig. 3a). To measure the levels of cytokines secreted from cultured cells, induced T lymphocytes were separated from each group on day 7, transferred to six-well plates, and cultured in RPMI 1640 medium without cytokines or serum; the culture supernatants were collected at 24 h after transfer. The levels of the cytokines IFN-γ, tumor necrosis factor-α (TNF-α), and IL-2 in the supernatants of all groups were quantified. The concentrations of IFN-γ, TNF-α, and IL-2 differed significantly among the groups. The highest concentrations of TNF-α, INF-γ, and IL-2 were detected in the 0.1-μg/ml mAb + IFN-γ group (Fig. 3b), suggesting that application of the anti-PD-1 antibody combined with IFN-γ at the primary stage of T-lymphocyte induction significantly up-regulated cytokine secretion.

Next, we assessed the cytotoxic activities of induced T lymphocytes in vitro. After 7 days of culture, induced T lymphocytes were collected from each group and incubated with BxPC-3, PANC-1, or MIA PaCa-2 cell lines at an effector/target ratio of 5:1, 10:1, or 20:1 (Fig. 3c). The cytotoxic activity of each group varied as a function of the effector/target ratio, and the highest lytic activity was detected using a 20:1 ratio (Fig. 3c). Further, the 0.1-μg/ml mAb + IFN-γ group exhibited the highest cytotoxic activity against pancreatic cancer cells, compared with the Ctrl group, IFN-γ group, and 0.1-μg/ml mAb group (Fig. 3c). These results suggest that nivolumab combined with IFN-γ significantly enhances the immunological functions of the T lymphocytes of patients with PDAC.

The tumor microenvironment plays an important role during tumor progression and treatment. The microenvironment of pancreatic cancer contributes particularly strongly to immune tolerance. To mimic this status, we tested the cultured T lymphocytes in nude mice bearing subcutaneous pancreatic cancer cell lines. The cultured T lymphocytes from each group were intravenously injected into the tail 3 days after subcutaneous implantation with BxPC-3 cells and five times at 4-day intervals thereafter (Fig. 4a). The tumor size was measured at each injection, and the last measurement was performed before sacrifice (Fig. 4b). Significant inhibition of tumor growth was observed in 0.1-μg/ml mAb-IFN-γ-treated mice compared with the IFN-γ, mAb, or Ctrl groups 31 days after the injection of BxPC-3 cells (Fig. 4c). The mean tumor sizes of the 0.1-μg/ml mAb-IFN-γ, IFN-γ, 0.1-μg/ml mAb, and Ctrl groups were 5.54 ± 0.62, 7.63 ± 0.32, 8.69 ± 0.85, and 11.06 ± 0.61 mm, respectively (Additional file 1: Figure S1). Mice treated with T lymphocytes in the 0.1-μg/ml mAb-IFN-γ group showed the highest tumor suppressive activity on day 31 after inoculation (Fig. 4d). Furthermore, immunofluorescence staining of CD8+ T cells in tumor tissues showed that the number of CD8+ T lymphocytes infiltrated in tumor tissues from the 0.1-μg/ml mAb-IFN-γ group was higher than that from the Ctrl, IFN-γ, or mAb groups (Fig. 4e). These results suggest that adoptive transfer treatment of T lymphocytes, which were stimulated by nivolumab and IFN-γ, could break through the inhibitory immune microenvironment and may therefore provide a novel approach to suppression of pancreatic cancer.