IJO International Journal of Oncology 1019-6439 1791-2423 D.A. Spandidos 10.3892/ijo.2022.5387 IJO-61-02-05387 Articles Immunomodulatory antitumor effect of interferon-beta combined with gemcitabine in pancreatic cancer BlaauboerAmber 1 2 Van KoetsveldPeter M. 2 MustafaDana A.M. 3 DumasJasper 3 DoganFadime 2 Van ZwienenSuzanne 2 Van EijckCasper H.J. 1 HoflandLeo J. 2 Department of Surgery, Erasmus Medical Center, 3015 CN Rotterdam, The Netherlands Department of Internal Medicine, Division of Endocrinology, Erasmus Medical Center, 3015 CN Rotterdam, The Netherlands Department of Pathology, The Tumor Immuno-Pathology Laboratory, Erasmus Medical Center, 3015 CN Rotterdam, The Netherlands Correspondence to: Professor Leo J. Hofland, Department of Internal Medicine, Division of Endocrinology, Erasmus Medical Center, 80 Wytemaweg Street, 3015 CN Rotterdam, The Netherlands, E-mail: l.hofland@erasmusmc.nl 08 2022 01 07 2022 61 2 97 12022022 06042022 Copyright © 2022, Spandidos Publications 2022

Resistance to gemcitabine is common and critically limits its therapeutic efficacy in patients with pancreatic cancer. Interferon-beta (IFN-β) induces numerous antitumor effects and synergizes with gemcitabine treatment. The immunomodulatory effects of this treatment regimen have not yet been described. In the present study, the antitumor effect of IFN-β combined with gemcitabine was investigated in immune competent mice. Mouse KPC3 cells were used in all experiments. Treatment effects were determined with cell proliferation assay. Reverse transcription-quantitative PCR was used to measure gene expression. For in vivo experiments, cells were subcutaneously injected in immune competent mice. For immune profiling, NanoString analysis was performed on tumor samples of treated and untreated mice. Baseline expression of Ifnar-1 and Ifnar-2c in KPC3 cells was 1.42±0.16 and 1.50±0.17, respectively. IC50 value of IFN-β on cell growth was high (>1,000 IU/ml). IFN-β pre-treatment increased the in vitro response to gemcitabine (1.3-fold decrease in EC50; P<0.001). In vivo, tumor size was not statistically significant smaller in mice treated with IFN-β plus gemcitabine (707±92 mm3 vs. 1,239±338 mm3 in vehicle-treated mice; P=0.16). IFN-β alone upregulated expression of numerous immune-related genes. This effect was less pronounced when combined with gemcitabine. For the first time, to the best of our knowledge, the immunomodulatory effects of IFN-β, alone and combined with gemcitabine, in pancreatic cancer were reported. Prognostic markers for predicting effective responses to IFN-β therapy are urgently needed.

 pancreatic cancer interferon-beta gemcitabine immune system Funding: No funding was received. Introduction

Pancreatic cancer is one of the most aggressive malignancies and highly resistant to currently applied cancer therapies (1). Even after curative-intent surgery, cure is exceedingly rare, as demonstrated by a 5-year overall survival (OS) of less than 20% (2). Effective adjuvant systemic therapies are necessary to improve survival outcomes.

Previously, the PRODIGE 24/CCTG PA.6 trial reported an impressive survival benefit with adjuvant modified–FOLFIRINOX (mFOLFIRINOX) compared with adjuvant gemcitabine (median OS 54 vs. 35 months; P<0.001) (3). However, due to the high toxicity rate of this treatment regime, gemcitabine is still being recommended, in particular for patients with poor or declining performance status.

Resistance to gemcitabine is a major impediment of successful treatment and is primarily due to molecular mechanisms limiting the intracellular uptake and metabolic activation of gemcitabine, and thus, its overall efficacy (4).

Type I IFNs (IFN-α and -β) have been proposed as potential adjuvant to gemcitabine treatment in order to improve survival outcomes in patients with pancreatic cancer (5,6). Type I IFNs are pleiotropic cytokines that were originally identified as viral replication suppressor. Further characterization of their biological effect revealed a wide range of antitumor effect; for instance, direct inhibitory effects on tumor cells, anti-angiogenesis, enhanced immunogenicity of tumors as well as other immune stimulatory effects (7,8). Thereby, type I IFNs induce synergistic effects on pancreatic cancer cells when co-treated with gemcitabine in vitro (9,10). Recently, the chemo-sensitizing effect of IFN-β in pancreatic cancer, formed in immune deficient mice, was confirmed and upregulation of gemcitabine transporter-coding genes was identified (9).

Both IFN-α and -β act via the type I IFN receptor (IFNAR) complex, of which IFNAR-1 and IFNAR-2c are the most important subunits (11). Binding to this receptor activates the JAK-STAT signaling pathway, which subsequently initiates the transcription of numerous interferon-stimulated genes (ISGs) that are responsible for mediating the biological activities of type I IFNs.

To accomplish an effect of type I IFNs, the presence of the IFNAR complex is necessary (11,12). Previously, expression of IFNAR-1 and IFNAR-2c in 91.5 and 68.1% of pancreatic cancer tissues, respectively, was reported (13).

Although IFN-α and -β bind to the same receptor and induce signals through similar mechanisms, they have different binding affinities. Previous studies reported a 50-fold higher receptor-binding affinity for IFN-β than IFN-α, resulting in more potent, elicited at much lower concentrations, antitumor effects (10,14).

Despite strong evidence of the more potent and safer antitumor effects of IFN-β compared with IFN-α, only a few studies focused on adjuvant IFN-β therapy in pancreatic cancer. In addition, the immunomodulatory effects of IFN-α and -β are less described and primarily investigated in IFN-α. Hence, in the present study, the potential immunomodulatory effect of IFN-β towards pancreatic cancer cells was revealed for the first time. The unique KPC3 cell line was used, derived from the clinically relevant KPC mouse model, which mimics the immune phenotypic features, aggressiveness, and gemcitabine-resistant character of human pancreatic cancer (15,16). Thereby, the present study focused on expression of gemcitabine transporter- and activating-coding genes, as potential target to increase gemcitabine efficacies.

 Materials and methods <sec> <title>Cells and culture conditions

The mouse KPC3 pancreatic cell line is derived from a primary pancreatic ductal adenocarcinoma tumor of a female KrasG12D/+;Trp53R172H/+;Pdx-1-Cre (KPC) mouse and was kindly provided by Dr van Montfoort (Leiden University Medical Center, Leiden, The Netherlands) (15). Origin of cells was confirmed using short tandem repeat profiling (Powerplex Kit; Promega Corporation). Cells were cultured in RPMI-1640 medium supplemented with 5%fetal calf serum (FCS), penicillin (1×105 U/l), and L-glutamine (2 mmol/l). FCS was purchased from Sigma-Aldrich; Merck KGaA. Media and other supplements were obtained from Gibco; Thermo Fisher Scientific, Inc. FCS was purchased from Sigma-Aldrich; Merck KGaA. Cells were cultured in a 5% CO2 atmosphere at 37°C and routinely validated as Mycoplasma-free. Culture conditions were as previously described in detail (17).

After trypsinization, KPC3 cells were plated in 24-well plates at the appropriate density in order to obtain 80% confluency at the end of the experiment. The next day, incubations were started in quadruplicate and control cells were vehicle-treated. Medium and compounds were refreshed after 3 days. All cell culture experiments were carried out at least twice in quadruplicate. Mouse recombinant IFN-β-1a (Bio-Connect) and gemcitabine (Sigma-Aldrich; Merck KGaA) stock dilutions were diluted in distilled water.

 Cell proliferation assay

Effects of IFN-β and/or gemcitabine on cell growth were assessed by measuring the total DNA amount per well, as a measure of cell number. After treatment, media were removed and plates were stored at −20°C until DNA measurement. Measurement of total DNA was performed with the bisbenzimide fluorescent dye (Hoechst 33258; Sigma-Aldrich; Merck KGaA) as previously described in detail (17).

 Reverse transcription-quantitative (RT-q) PCR

RNA isolation, cDNA synthesis and RT-qPCR were performed as previously described (9), but using different primers (Table SI). Two housekeeping genes were used to normalize mRNA levels using the Vandesompele method: hypoxanthine-guanine phosphoribosyl transferase 1 (Hprt1) and glucuronidase beta (Gusb) (Gibco; ThermoFisher Scientific, Inc.) (18).

 Mice

A total of 28 male C57BL/6 mice (8–10 weeks old) were purchased from Charles River Laboratories. All mice were housed in groups of seven and maintained at room temperature on a daily 12-h light/12-h dark cycle in ventilated cages with autoclaved bedding. Water and autoclaved laboratory rodent diet were provided ad libitum. All mouse experiments were controlled by the animal welfare committee of the Erasmus University Medical Center (Rotterdam, The Netherlands) and approved (approval no. AVD101002017867) by the national central committee of animal experiments, in accordance with the Dutch Act on Animal Experimentation and European Union (EU) Directive 2010/63/EU.

 In vivo experiments

Mice were randomized in four groups (n=7 each) and subcutaneously injected in the flank with 100,000 low passage (passage number 3) KPC3 cells in 100 µl PBS/0.1% BSA (Gibco; Thermo Fisher Scientific, Inc.). Cultured KPC3 cells were harvested at 80% confluency and only single-cell suspensions of greater than 90% viability were used for injection. Tumor size and body weight were measured twice weekly. Tumor volume was calculated as (width^2 × length)/2 using a caliper. Treatment was started when tumor volumes reached ~50 mm3. Mice in the control group and in the IFN-β monotherapy group received daily an intraperitoneal (i.p.) injection of 100 µl of distilled water or 10,000 units IFN-β. Mice randomized to the gemcitabine monotherapy group received two times a week (day 2 and 5) an i.p. injection of 50 mg/kg gemcitabine. Mice in the combination group received upon start of the treatment, daily an injection of 10,000 units IFN-β i.p. and, on day 2 and 5, an i.p. injection of 50 mg/kg gemcitabine. The effect of gemcitabine monotherapy in this model has been previously described (19).

 Necropsy procedures

Mice were sacrificed by cervical dislocation under 5% isoflurane anesthesia when tumor volume reached 1,000 mm3 or when the wellbeing of the mice could no longer be maintained. Tumors were resected during necropsy and tumor volumes were measured. Tumors were divided into two parts and subsequently snap-frozen in liquid nitrogen and fixed overnight at 4°C in freshly prepared 4% formaldehyde solution, and prepared for paraffin sectioning (5-µm thick).

 NanoString analysis

RNA was extracted according to the manufacturer's instructions using the RNeasy Plus Micro kit (Qiagen). RNA samples were diluted in RNA free water and stored at −80°C. The 2100 Bioanalyzer (Agilent Technologies, Inc.) was used to measure RNA Quality Control. Total RNA concentrations were corrected to include fragments seized between 300 and 4,000 nucleotides. A total of 200 ng RNA was hybridized to the PanCancer IO 360 Panel (NanoString Technologies, Inc.) at 67°C for 17 h. The advanced analysis module (version 2.0) of nSolver™ software (version 4.0; NanoString Technologies, Inc.) was used for data analysis. A total of 8 out of 11 housekeeping genes were selected for normalization with the geNorm algorithm embedded in the advanced analysis module (Table SII). Expression threshold was calculated as twice of the average expression of the negative controls. Gene expression below the threshold in more than 80% of the samples were excluded from further analysis. Normalized data were log2 transformed. Differentially expressed genes were identified with simplified negative binomial models, mixture negative binomial models, or log-linear models based on the convergence of each gene. Adjusted P-values were calculated with the Benjamini Hoghberg method. Genes were considered differentially expressed when the adjusted P<0.05.

 Statistical analysis

Statistical analysis was performed using GraphPad Prism version 3.0 (GraphPad Software, Inc.). The half maximal effective concentration (EC50) on cell growth was calculated using non-linear regression curve fitting program. Effect of IFN-β was set on 100% and used as control to analyze the combined effect of IFN-β pre-treatment and gemcitabine in in vitro experiments. One-way ANOVA followed by Tukey's multiple comparisons test was used for comparisons between treatment groups. In all analyses, P<0.05 was considered to indicate a statistically significant difference. Data are indicated as the mean ± SEM.

 Results <sec> <title>IFN-β sensitivity in KPC3 cells in vitro

Relative mRNA expression of the IFN receptor-coding genes, Ifnar-1 and Ifnar-2c, were comparable in KPC3 cells (1.42±0.16 and 1.50±0.17, respectively) (Fig. 1A). Baseline expression of ISGs in untreated KPC3 cells was relatively low (Fig. 1B). The growth-inhibitory effect of IFN-β was not very potent, as shown by an EC50 >1,000 IU/ml and maximal inhibition of 51% after 7 days (P<0.001 vs. control) (Fig. 1C).

 Effect of IFN-β pre-treatment on the response to gemcitabine in vitro

Next, it was analyzed whether IFN-β pre-treatment could sensitize KPC3 cells, reflected by a decrease in the EC50 value of gemcitabine. Although 72 h after 1,000 IU/ml IFN-β pre-treatment alone had no statistically significant effect on the cell amount, gemcitabine sensitivity was slightly increased, as shown by a 1.3-fold decrease in EC50 value compared with untreated control cells (EC50 1.5 ng/ml vs. EC50 1.1 ng/ml respectively; P<0.001) (Fig. 2A and C). Expression of Oas1a and Ifit1 was strongly upregulated following 72 h after 1,000 IU/ml IFN-β treatment (51- and 13-fold increase respectively; both P<0.001 vs. untreated cells), whereas expression of Mx1 was not affected by IFN-β treatment (Fig. 2B).

Baseline expression of transporter- and metabolizing-coding genes in KPC3 cells was low and even undetectable for Slc28a1 and Slc28a3 (Fig. S1). Expression was not upregulated after IFN-β treatment (Fig. S1).

 In vivo validation of IFN-β combined with gemcitabine in immune-competent mice

To study the immunomodulatory antitumor effect of IFN-β, low-passage KPC3 cells were subcutaneously injected in C57BL/6 mice. Mice were randomized into four treatment arms: daily H2O (control), daily 10,000 units IFN-β, twice weekly 50 mg/kg gemcitabine, or the combination of daily 10,000 units IFN-β plus twice weekly 50 mg/kg gemcitabine (Fig. 3A). All mice were sacrificed at day 21, after which tumors were collected for analysis. None of the treatment arms resulted in a statistically significant tumor growth inhibition over time, although lowest tumor volumes were observed in the combined treatment group, suggesting an additive effect rather than a synergistic effect (Fig. 3B-D). After 21 days of treatment, tumor volume was 707±92 mm3 in the combination treated-mice compared with 1,239±338 mm3 in the vehicle treated-mice (P=0.16) (Fig. 3C). Tumor volumes in gemcitabine and IFN-β mono-treated mice were 1,162±232 and 962±271 mm3, respectively (both P>0.05 vs. untreated mice).

No significant weight loss was observed in any of the treatment groups, indicating that all drugs were well tolerated (Fig. 3E). Expression of transporter- and metabolizing-coding genes of gemcitabine was low in untreated KPC3 tumors and were not affected by any treatment (data not shown).

 Effect of IFN-β on expression of immune-related genes

To specifically address the immunomodulatory capacity of IFN-β, a targeted gene expression array was performed on tumor samples of treated and untreated mice. Differentially expressed genes upon treatment compared with untreated mice are revealed in Fig. 4A. After adjusting for multiple testing, only a few genes were significantly altered by the different treatment groups. Specifically, gemcitabine downregulated expression of Itpk1, Igf2r, Bmp2, Dusp1, and Cdkn1a, but increased Tgfbr1. IFN-β only upregulated Gdp2, and in combination-treated tumors, Itpk1 and Lilra5 were respectively down- and upregulated (Fig. 4A, Table SIII).

Without adjusting P-values, an extensive number of genes was found to be differentially expressed (Fig. 4B and Table SIII). In total, 35 genes were upregulated by IFN-β, while 38 genes were downregulated. In gemcitabine-treated tumors, 30 and 69 genes were respectively up- and downregulated. A further 32 genes were upregulated by combination therapy, whereas 56 genes were downregulated (Fig. 4B). Top ten up- and downregulated genes are summarized in Table I. Gemcitabine and combination therapy commonly affected several genes (Snca, Lilra5, Fgfl3, Siglecf, Prkaa2, Selp, Hk2, Erol1, Ndufa412 and Vegfa), while IFN-β alone induced a different gene expression profile (Table I).

IFN-β monotherapy markedly upregulated expression of several immune-related genes. As a consequence, 7 out of 9 immune-related pathways were upregulated in IFN-β-treated tumors compared with untreated mice (range pathway score 1.67-0.54) (Fig. 4C and D). By contrast, gemcitabine induced a suppressive effect on these pathways (range pathway score-1.11-0.39), but when co-treated with IFN-β, 6 out of 9 immune-related pathways were increased (range pathway score 0.96-0.43).

Expression of IFN signaling pathway regulators (Ifi35, Irf1, Irf7 and Stat2) and four ISGs (Gbp2, Gbp3, H2-d1 and Uba7) were induced by IFN-β alone. Moreover, expression of several ISG chemokines (Ccl19, Cxcl10 and Cxcl11) were increased in these tumors (Table II). In combination-treated tumors, the regulatory factors Ifi35 and Trim21 were upregulated as well as two well-known pro-apoptotic ISGs (Mx1 and Oas1a). By contrast, gemcitabine downregulated IFN-related genes (H2-K1, H2-T23, Ifi203, Igf2r and Irf2). Numerous genes involved in antigen presentation were stimulated by IFN-β; for instance, Tap1, Tap2, Psmb9 and Psmb10. This effect was less pronounced in combination-treated tumors and also affected a different set of genes (Trim21, Ctsw and Lag3). Remarkably, the myeloid compartment pathway was downregulated by IFN-β, but stimulated by gemcitabine and combination therapy through upregulation of Clec7a, Lilra5, and P2ry13.

Analysis reported no evident differences in pro-tumor pathways, such as angiogenesis and metastasis, and were equally inhibited in all treatment groups compared with untreated tumors (Fig. 4C).

 Discussion

Pancreatic cancer is a highly aggressive malignancy with limited treatment outcomes. Despite promising recent advances in systemic therapies, for instance mFOLFIRINOX, gemcitabine is still being recommended for patients with a poor or declining performance status (3).

Type I IFN-based adjuvant therapies have been widely studied in pancreatic cancer, primarily due to their potential synergistic effects when co-treated with gemcitabine (5,6,9,10). Thereby, type I IFNs induce several direct antitumor effects, including apoptosis and cell growth arrest as well as critical immune stimulatory effects on various immune cells (7,8).

To date, studies have primarily focused on the direct antitumor effects of the less effective IFN-α. The immunomodulatory effects of IFN-α and -β are less described in pancreatic cancer and primarily investigated in IFN-α. To the best of our knowledge, this is the first study evaluating the antitumor effects of the more promising IFN subtype IFN-β, alone and combined with gemcitabine, on pancreatic cancer cells in immune competent mice. The unique KPC3 pancreatic cancer cell line was used, generated from the clinically relevant and non-immunogenic KPC mouse model (15).

A trend towards smaller tumor volumes in combination-treated mice was observed, suggesting an additive effect rather than synergistic effect. Moreover, tumors displayed a differential gene expression profile upon treatment. In particular, IFN-β alone induced expression of numerous immune-related genes, such as chemokines (Ccl19, Cxcl10 and Cxcl11) and antigen processors (Tap1 and Tap2). Thereby, expression of Irf7 was upregulated, which primarily regulates the immunomodulatory capacities of IFN-β, as well as other IFN signaling pathway regulators (e.g., Ifi35, Irf1 and Stat2) (20,21). Notably, Irf1, which promotes expression of ISGs, and Irf7 are key factors in the positive feedback regulation of IFN-β production and potentiate ISG expression as they directly target ISGs, even in the absence of IFN signaling (22–24). It should be emphasized, however, that only mRNA expression was evaluated. Further studies should demonstrate that IFN-β treatment indeed results in effective antitumor immune responses.

Controversially, gemcitabine induced an immune suppressive effect, whereas addition of IFN-β to gemcitabine solely induced a subset of immune-related genes, which was less pronounced compared with IFN-β mono-treated tumors. Since gemcitabine promotes the infiltration of particularly M2 macrophages in the tumor microenvironment, it may have diminished the immune stimulating effect of IFN-β when administered together (25).

Efficacy of several anticancer therapies, including chemotherapies, partially depend on an intact type I IFN signaling for the promotion of both direct (tumor cell inhibition) and indirect effects (antitumor immune responses) (26). Thus, impaired IFN signaling may as a consequence contribute to therapy resistance in patients with cancer. KPC tumors respond poorly to gemcitabine therapy, particularly orthotopic tumors, which is consistent with clinic outcomes, as only 5–10% of gemcitabine-treated patients show an objective radiographic response at the primary tumor site (27). Regarding in vivo research in KPC mice, the most frequently used gemcitabine concentrations are 50 and 100 mg/kg (16,28,29). To avoid any potential toxicity, mice were treated with 50 mg/kg gemcitabine. As expected, no significant tumor inhibition was observed with gemcitabine alone. Moreover, type I IFN signaling was diminished in gemcitabine-treated tumors as several IFN-related genes were downregulated when compared with untreated tumors (for instance, H2-K1, H2-T23, Ifi203, Igf2r and Irf2). Meanwhile, combination therapy increased expression of IFN regulator factors (for instance, Ifi35 and Trim21) as well as two well-known pro-apoptotic ISGs (Mx1 and Oas1a), but was still insufficient to significantly inhibit tumor growth.

Over the years, numerous studies highlighted the significant role of immune host-mediated mechanisms in the response to type I IFNs, even in IFN-resistant tumor cells (30–33). However, most studies used high-dose intratumoral IFN-β concentrations (33–35). Although intratumoral IFN-β concentrations were not evaluated in the present study, the relatively low concentration IFN-β (i.p. administered) as well as the resistant character of KPC-3 cell to IFN-β therapy may have resulted in insufficient intratumoral levels to induce significant tumor growth inhibition. It is plausible that IFN-β will exert stronger antitumor responses in IFN-β sensitive tumors. In fact, it has been previously shown that IFN-β combined with gemcitabine synergistically reduced tumor volumes in immune deficient mice, bearing an IFN-β sensitive tumor, when compared with untreated mice (9). Moreover, the relatively short half-life time of IFN-β may limit sufficient circulating IFN-β concentrations. Strategies to increase the half-life time of IFN-β, such as IFN-based conjugates or PEGylated form of IFN-β, have demonstrated promising results to achieve higher serum concentration, requiring lower and less frequent doses compared with the conventional IFNs (36,37).

While recombinant IFN therapies are generally given as exogenous pharmaceuticals, it is suggested that the autocrine and paracrine actions of endogenous type I IFNs on tumor growth control (both the direct and indirect effects) are much stronger. New IFN-related cancer treatment strategies, such as STING and RIG-I agonists, have emerged as promising and effective strategies to produce significant amounts of endogenous type I IFNs and are currently being examined in (pre-) clinical studies in various types of cancer, including pancreatic cancer (8). IFN-β gene therapy induced by viral vectors provides another promising strategy to achieve high intratumoral IFN-β levels and has demonstrated potent antitumor efficacy in several pre-clinical cancer models, including pancreatic cancer, with low toxicities (38–40). Discrepancies between treatment outcomes largely depend on differences in IFN-β sensitivity, highlighting the need for accurate biomarkers to predict IFN-β treatment response. Expression of the active IFN-receptor subunits (IFNAR-1 and IFNAR-2c) is required to form a high-binding affinity site and to initiate signal transduction leading to the induction of ISGs (11,41). However, despite expression of both receptor subunits, KPC3 cells responded poorly to IFN-β, suggesting other alterations involving the IFN downstream signaling pathway. In fact, IFN dysregulation can occur through several mechanisms such as loss or silencing of key signaling effector proteins and components (JAKs, STATs and IRFs) or through upregulation of negative regulators (SOCS1/3) (42,43). A careful evaluation of type I IFN signaling may provide important insights into sensitivity to IFN-β therapies and may possibly contribute to a more personalized medicine approach in the future.

Previously emerged evidence has suggested that type I IFNs may have dual roles in antitumor immunity, which may be even harmful and cause further adaptive resistance to therapies, including chemotherapy, radiotherapy and immune checkpoint blockade (44–47). Two hypotheses have been proposed for this controversial effect. First, continuous exposure of type I IFNs may upregulate PD-L1 expression in tumor cells, which subsequently promotes immune resistance trough interaction with PD-1+ immune effector cells (48). Thereby, prolonged type I IFN stimulation may induce an IFN-related DNA damage resistance signature (IRDS) that indicates an unfavorable response to DNA-damaging interventions such as chemotherapy and radiotherapy (44). Previously, IRDS scoring strategies have been used to identify patients with breast and lung cancer and showed higher expression of specific ISGs in poor responders to chemotherapy (49). These paradoxical findings are not yet fully understood and may also differ among types of cancer and depend on the subtype of type I IFNs (50).

In conclusion, for the first time the immunomodulatory potential of exogenous IFN-β combined with gemcitabine was revealed in the immune competent KPC3 mouse pancreatic model. The interplay between tumor cells and type I IFNs is complex and still not fully understood. The dynamic role of type I IFNs should be carefully considered to fully exploit its therapeutic value as anticancer drug. Further studies should focus on timing and duration of type I IFN administration as well as prognostic markers for predicting effective antitumor responses.

 Supplementary Material Supporting Data Supporting Data Supporting Data Acknowledgements

The authors would like to thank the animal welfare takers Kim van der Leest and Lisette Dinnessen (Erasmus Medical Center, Rotterdam, The Netherlands) for their help during the mouse experiments.

 Availability of data and materials

All data are available from the corresponding author on reasonable request. All raw gene expression array data are available on https://figshare.com/s/ec1cfa7d1d66f8528ec3.

 Authors' contributions

AB, PMVK, CHJVE and LJH conceptualized the study. AB, PMVK, DAMM and LJH performed formal analysis. CHJVE acquired funding. AB, PMVK, JD and SVZ conducted investigation. AB, PMVK, DAMM, CHJVE and LJH provided the methodology. PMVK, CHJVE and LJH supervised the study. FD validated the data. AB, CHJVE and LJH performed visualization. AB wrote the original draft. PMVK, DAMM, FD, CHJVE and LJH wrote, reviewed and edited the manuscript. AB, PMVK, CHJVE and LJH confirm the authenticity of all the raw data. All authors read and approved the final manuscript.

 Ethics approval and consent to participate

The present study was controlled by the animal welfare committee of the Erasmus University Medical Center (Rotterdam, The Netherlands) and approved (approval no. AVD101002017867) by the national central committee of animal experiments, in accordance with the Dutch Act on Animal Experimentation and European Union (EU) Directive 2010/63/EU.

 Patient consent for publication

Not applicable.

 Competing interests

The authors declare that they have no competing interests.

 References RahibLSmithBDAizenbergRRosenzweigABFleshmanJMMatrisianLMProjecting cancer incidence and deaths to 2030: The unexpected burden of thyroid, liver, and pancreas cancers in the United StatesCancer Res7429132921201410.1158/0008-5472.CAN-14-015524840647 BengtssonAAnderssonRAnsariDThe actual 5-year survivors of pancreatic ductal adenocarcinoma based on real-world dataSci Rep1016425202010.1038/s41598-020-73525-y33009477 ConroyTHammelPHebbarMBen AbdelghaniMWeiACRaoulJLChonéLFrancoisEArtruPBiagiJJFOLFIRINOX or gemcitabine as adjuvant therapy for pancreatic cancerN Engl J Med37923952406201810.1056/NEJMoa180977530575490 AmrutkarMGladhaugIPPancreatic cancer chemoresistance to gemcitabineCancers (Basel)9157201710.3390/cancers911015729144412 LinehanDCTanMCStrasbergSMDrebinJAHawkinsWGPicusJMyersonRJMalyapaRSHullMTrinkausKTanBRJrAdjuvant interferon-based chemoradiation followed by gemcitabine for resected pancreatic adenocarcinoma: A single-institution phase II studyAnn Surg248145151200810.1097/SLA.0b013e318181e4e918650621 OhmanKALiuJLinehanDCTanMCTanBRFieldsRCStrasbergSMHawkinsWGInterferon-based chemoradiation followed by gemcitabine for resected pancreatic adenocarcinoma: Long-term follow-upHPB (Oxford)19449457201710.1016/j.hpb.2017.01.01228162923 De MaeyerEDe Maeyer-GuignardJType I interferonsInt Rev Immunol175373199810.3109/088301898090844879914943 BlaauboerASiderasKvan EijckCHJHoflandLJType I interferons in pancreatic cancer and development of new therapeutic approachesCrit Rev Oncol Hematol159103204202110.1016/j.critrevonc.2020.10320433387625 BlaauboerABooySvan KoetsveldPMKarelsBDoganFvan ZwienenSvan EijckCHJHoflandLJInterferon-beta enhances sensitivity to gemcitabine in pancreatic cancerBMC Cancer20913202010.1186/s12885-020-07420-032967656 TomimaruYEguchiHWadaHTomokuniAKobayashiSMarubashiSTakedaYTanemuraMUmeshitaKMoriMSynergistic antitumor effect of interferon-β with gemcitabine in interferon-α-non-responsive pancreatic cancer cellsInt J Oncol3812371243201121347515 DomanskiPColamoniciORThe type-I interferon receptor. The long and short of itCytokine Growth Factor Rev7143151199610.1016/1359-6101(96)00017-28899292 WagnerTCVelichkoSChesneySKBirocSHardeDVogelDCrozeEInterferon receptor expression regulates the antiproliferative effects of interferons on cancer cells and solid tumorsInt J Cancer1113242200410.1002/ijc.2023615185340 BooySHoflandLJWaaijersAMCrozeEvan KoetsveldPMde VogelLBiermannKvan EijckCHType I interferon receptor expression in human pancreatic and periampullary cancer tissuePancreas4499105201510.1097/MPA.000000000000020225072284 BooySvan EijckCHDoganFvan KoetsveldPMHoflandLJInfluence of type-I Interferon receptor expression level on the response to type-I Interferons in human pancreatic cancer cellsJ Cell Mol Med18492502201410.1111/jcmm.1220024460759 LeeJWKomarCABengschFGrahamKBeattyGLGenetically engineered mouse models of pancreatic cancer: The KPC model [LSL-Kras(G12D/+);LSL-Trp53(R172H/+);Pdx-1-Cre], its variants, and their application in immuno-oncology drug discoveryCurr Protoc Pharmacol7314.39.114.39.20201627248578 OliveKPJacobetzMADavidsonCJGopinathanAMcIntyreDHonessDMadhuBGoldgrabenMACaldwellMEAllardDInhibition of Hedgehog signaling enhances delivery of chemotherapy in a mouse model of pancreatic cancerScience32414571461200910.1126/science.117136219460966 Herrera-MartínezADFeeldersRAde HerderWWCastañoJPGálvez MorenoMDoganFvan DungenRvan KoetsveldPHoflandLJEffects of Ketoconazole on ACTH-producing and Non-ACTH-producing neuroendocrine tumor cellsHorm Cancer10107119201910.1007/s12672-019-00361-631102172 VandesompeleJDe PreterKPattynFPoppeBVan RoyNDe PaepeASpelemanFAccurate normalization of real-time quantitative RT-PCR data by geometric averaging of multiple internal control genesGenome Biol3RESEARCH0034200210.1186/gb-2002-3-7-research003412184808 BlaauboerAvan KoetsveldPMMustafaDAMDumasJDoganFvan ZwienenSvan EijckCHJHoflandLJThe Class I HDAC inhibitor valproic acid strongly potentiates gemcitabine efficacy in pancreatic cancer by immune system activationBiomedicines10517202210.3390/biomedicines1003051735327319 IwanaszkoMKimmelMNF-κB and IRF pathways: Cross-regulation on target genes promoter levelBMC Genomics16307201510.1186/s12864-015-1511-725888367 SolisMGoubauDRomieu-MourezRGeninPCivasAHiscottJDistinct functions of IRF-3 and IRF-7 in IFN-alpha gene regulation and control of anti-tumor activity in primary macrophagesBiochem Pharmacol7214691476200610.1016/j.bcp.2006.06.00216846591 PlatanitisEDeckerTRegulatory Networks involving STATs, IRFs, and NFκB in inflammationFront Immunol92542201810.3389/fimmu.2018.0254230483250 MariéIDurbinJELevyDEDifferential viral induction of distinct interferon-alpha genes by positive feedback through interferon regulatory factor-7EMBO J1766606669199810.1093/emboj/17.22.66609822609 SatoMHataNAsagiriMNakayaTTaniguchiTTanakaNPositive feedback regulation of type I IFN genes by the IFN-inducible transcription factor IRF-7FEBS Lett441106110199810.1016/S0014-5793(98)01514-29877175 DeshmukhSKTyagiNKhanMASrivastavaSKAl-GhadhbanADuggerKCarterJESinghSSinghAPGemcitabine treatment promotes immunosuppressive microenvironment in pancreatic tumors by supporting the infiltration, growth, and polarization of macrophagesSci Rep812000201810.1038/s41598-018-30437-230097594 ZitvogelLGalluzziLKeppOSmythMJKroemerGType I interferons in anticancer immunityNat Rev Immunol15405414201510.1038/nri384526027717 TemperoMPlunkettWRuiz Van HaperenVHainsworthJHochsterHLenziRAbbruzzeseJRandomized phase II comparison of dose-intense gemcitabine: Thirty-minute infusion and fixed dose rate infusion in patients with pancreatic adenocarcinomaJ Clin Oncol2134023408200310.1200/JCO.2003.09.14012885837 MillerBWMortonJPPineseMSaturnoGJamiesonNBMcGheeETimpsonPLeachJMcGarryLShanksETargeting the LOX/hypoxia axis reverses many of the features that make pancreatic cancer deadly: Inhibition of LOX abrogates metastasis and enhances drug efficacyEMBO Mol Med710631076201510.15252/emmm.20140482726077591 CookNFreseKKBapiroTEJacobetzMAGopinathanAMillerJLRaoSSDemuthTHowatWJJodrellDITuvesonDAGamma secretase inhibition promotes hypoxic necrosis in mouse pancreatic ductal adenocarcinomaJ Exp Med209437444201210.1084/jem.2011192322351932 BelardelliFGresserIThe neglected role of type I interferon in the T-cell response: Implications for its clinical useImmunol Today17369372199610.1016/0167-5699(96)10027-X8783497 BelardelliFFerrantiniMCytokines as a link between innate and adaptive antitumor immunityTrends Immunol23201208200210.1016/S1471-4906(02)02195-611923115 RizzaPCaponeIMorettiFProiettiEBelardelliFIFN-α as a vaccine adjuvant: Recent insights into the mechanisms and perspectives for its clinical useExpert Rev Vaccines10487498201110.1586/erv.11.921506646 SpaapenRMLeungMYFuertesMBKlineJPZhangLZhengYFuYXLuoXCohenKSGajewskiTFTherapeutic activity of high-dose intratumoral IFN-β requires direct effect on the tumor vasculatureJ Immunol19342544260201410.4049/jimmunol.140110925217157 SchiavoniGSistiguAValentiniMMatteiFSestiliPSpadaroFSanchezMLorenziSD'UrsoMTBelardelliFCyclophosphamide synergizes with type I interferons through systemic dendritic cell reactivation and induction of immunogenic tumor apoptosisCancer Res71768778201110.1158/0008-5472.CAN-10-278821156650 YangXZhangXFuMLWeichselbaumRRGajewskiTFGuoYFuYXTargeting the tumor microenvironment with interferon-β bridges innate and adaptive immune responsesCancer Cell253748201410.1016/j.ccr.2013.12.00424434209 BordenECInterferons α and β in cancer: Therapeutic opportunities from new insightsNat Rev Drug Discov18219234201910.1038/s41573-018-0011-230679806 PepinskyRBLePageDJGillAChakrabortyAVaidyanathanSGreenMBakerDPWhalleyEHochmanPSMartinPImproved pharmacokinetic properties of a polyethylene glycol-modified form of interferon-beta-1a with preserved in vitro bioactivityJ Pharmacol Exp Ther29710591066200111356929 RavetELulkaHGrossFCasteillaLBuscailLCordelierPUsing lentiviral vectors for efficient pancreatic cancer gene therapyCancer Gene Ther17315324201010.1038/cgt.2009.7919911032 EndouMMizunoMNagataTTsukadaKNakaharaNTsunoTOsawaHKunoTFujitaMHatanoMYoshidaJGrowth inhibition of human pancreatic cancer cells by human interferon-beta gene combined with gemcitabineInt J Mol Med15277283200515647844 OhashiMYoshidaKKushidaMMiuraYOhnamiSIkarashiYKitadeYYoshidaTAokiKAdenovirus-mediated interferon alpha gene transfer induces regional direct cytotoxicity and possible systemic immunity against pancreatic cancerBr J Cancer93441449200510.1038/sj.bjc.660271316106250 DeonarainRChanDCPlataniasLCFishENInterferon-alpha/beta-receptor interactions: A complex story unfoldingCurr Pharm Des821312137200210.2174/138161202339328812369858 BudhwaniMMazzieriRDolcettiRPlasticity of Type I Interferon-mediated responses in cancer therapy: From Anti-tumor immunity to resistanceFront Oncol8322201810.3389/fonc.2018.0032230186768 PatelSJSanjanaNEKishtonRJEidizadehAVodnalaSKCamMGartnerJJJiaLSteinbergSMYamamotoTNIdentification of essential genes for cancer immunotherapyNature548537542201710.1038/nature2347728783722 WeichselbaumRRIshwaranHYoonTNuytenDSBakerSWKhodarevNSuAWShaikhAYRoachPKreikeBAn interferon-related gene signature for DNA damage resistance is a predictive marker for chemotherapy and radiation for breast cancerProc Natl Acad Sci USA1051849018495200810.1073/pnas.080924210519001271 ErdalEHaiderSRehwinkelJHarrisALMcHughPJA prosurvival DNA damage-induced cytoplasmic interferon response is mediated by end resection factors and is limited by Trex1Genes Dev31353369201710.1101/gad.289769.11628279982 PostAEMSmidMNagelkerkeAMartensJWMBussinkJSweepFSpanPNInterferon-stimulated genes are involved in Cross-resistance to radiotherapy in tamoxifen-resistant breast cancerClin Cancer Res2433973408201810.1158/1078-0432.CCR-17-255129661777 BenciJLXuBQiuYWuTJDadaHTwyman-Saint VictorCCucoloLLeeDSMPaukenKEHuangACTumor interferon signaling regulates a multigenic resistance program to immune checkpoint blockadeCell16715401554.e12201610.1016/j.cell.2016.11.02227912061 TerawakiSChikumaSShibayamaSHayashiTYoshidaTOkazakiTHonjoTIFN-α directly promotes programmed cell death-1 transcription and limits the duration of T cell-mediated immunityJ Immunol18627722779201110.4049/jimmunol.100320821263073 PitrodaSPStackMELiuGFSongSSChenLLiangHParekhADHuangXRoachPPosnerMCJAK2 inhibitor SAR302503 abrogates PD-L1 expression and targets therapy-resistant non-small cell lung cancersMol Cancer Ther17732739201810.1158/1535-7163.MCT-17-066729467274 DohertyMRCheonHJunkDJVinayakSVaradanVTelliMLFordJMStarkGRJacksonMWInterferon-beta represses cancer stem cell properties in triple-negative breast cancerProc Natl Acad Sci USA1141379213797201710.1073/pnas.171372811429229854

IFN-β sensitivity in KPC3 cells in vitro. (A) Baseline mRNA expression of the IFN receptor subunits Ifnar-1 and Ifnar-2c and (B) baseline mRNA expression of the interferon-stimulated genes Oas1a, Ifit1 and Mx1. (C) Dose response curves of IFN-β on cell amount (as measured by total DNA content per well), after 3 and 7 days of treatment. Values represent the mean ± SEM of at least two independent experiments in quadruplicate and are shown as the percentage of control. IFN-β, interferon-β; Ifnar-1, interferon alpha and beta receptor subunit 1; Ifnar-2c, interferon alpha and beta receptor subunit 2; Oas1a, 2′-5′ oligoadenylate synthetase 1a; Ifit1, interferon induced protein with tetratricopeptide repeats 1; Mx1, MX dynamin like GTPase 1.

Effect of IFN-β pre-treatment on gemcitabine response in KPC3 cells. Cells were pre-treated for 72 h with 1,000 IU/ml IFN-β, followed by 72 h gemcitabine monotherapy. (A) Effect of 72 h 1,000 IU/ml IFN-β monotherapy. (B) Fold change in mRNA expression of Oas1a, Ifit1, and Mx1 between no treatment and after 72 h 1,000 IU/ml IFN-β. (C) Overall gemcitabine response in untreated control cells (solid line) vs. 72 h IFN-β pre-treated cells (dotted lines). Controls represent vehicle control for non-IFN-β pre-treated and treatment with IFN-β for IFN-β pre-treated cells. Values represent the mean ± SEM of at least two independent experiments in quadruplicate and are shown as a percentage of control. ***P<0.01 vs. control and ###P<0.01 vs. other gene. IFN-β, interferon-β; Oas1a, 2′-5′ oligoadenylate synthetase 1a; Ifit1, interferon induced protein with tetratricopeptide repeats 1; Mx1, MX dynamin like GTPase 1.

In vivo antitumor effects of IFN-β alone and in combination with gemcitabine, in immune competent mice. (A) Experimental design for in vivo experiments. KPC3 mouse pancreatic cancer cells (1×105/100 µl PBS/0.1% BSA) were subcutaneously injected in C57BL/6 mice. Treatment was started when tumor volumes reached 50 mm3. Groups of mice received daily an i.p. injection of IFN-β (10,000 units), two times a week (at day 2 and 5) an i.p. injection of gemcitabine (50 mg/kg), or the combination of IFN-β plus gemcitabine. Mice in the control group received daily an i.p. injection of 100 µl distilled water. (B) Time course of change in tumor volume. (C) After 21 days of treatment, mice were sacrificed and tumor volumes were measured. (D) Tumor images after 21 days of treatment. (E) Body weight difference from start until end of treatment. Values represent the mean ± SEM. IFN-β, interferon-β; i.p, intraperitoneal.

Nanostring analysis PanCancer IO 360 panel. (A) Volcano plot displays-log10(p-value) and log2 fold change of each gene with the selected covariate. Highly statistically significant genes fall at the top of the plot above the horizontal lines, and highly differentially expressed genes fall to either side. Horizontal lines indicate various P-value thresholds. The 40 most statistically significant genes are labelled in the plot. (B) Venn diagram shows the number of up- and downregulated genes in the three treatment arms compared with control. Overlapping areas represent the number of genes that were altered by both treatment arms. (C) Heatmap displays directed global significance scores of each sample. Directed global significance statistics measure the extent to which gene sets of a gene are up- or downregulated with the variable. Red denotes gene sets whose genes exhibit extensive overexpression with the covariate, blue denotes gene sets with extensive under expression. (D) Pathway scores of immune-related pathways. Increasing pathway scores corresponds to increasing expression. IFN-β, interferon-β.

Top ten up- and down-fold regulation.

Gemcitabine vs. control IFN-β vs. control Combination vs. control
Gene name Log2 fold change P-value B.H. P-value Gene name Log2 fold change P-value B.H. P-value Gene name Log2 fold change P-value B.H. P-value
Snca 2.87 0.00479 0.0887 Gbp2 3.6 4.27×10−5 0.0261 Snca 2.1 0.0307 0.266
Fgf13 2.17 0.00234 0.0731 Cd36 2.78 0.0034 0.116 Lilra5 1.98 0.000121 0.0369
Fcer1a 1.72 0.0288 0.22 Ccl19 2.63 0.0164 0.251 Fgf13 1.69 0.0135 0.197
Gzmb 1.72 0.0166 0.161 Gbp3 2.59 0.000323 0.0638 Siglecf 1.67 0.0122 0.193
Siglecf 1.71 0.0106 0.12 Cxcl10 2.36 0.00302 0.115 Prkaa2 1.38 0.0161 0.206
S100a8 1.64 0.00117 0.0731 Cxcl11 2.13 0.00139 0.0943 Sfrp1 1.29 0.0302 0.266
S100a9 1.61 0.00293 0.0745 Cd1d1 1.89 0.00804 0.196 Col11a1 1.28 0.0191 0.212
Prkaa2 1.57 0.00711 0.101 Fcer1a 1.76 0.0412 0.377 Clec7a 1.06 0.000881 0.095
Lilra5 1.43 0.0025 0.0731 Cpa3 1.72 0.0147 0.242 P2ry13 1.01 0.000932 0.095
Cpa3 1.38 0.0265 0.214 Nlrc5 1.63 0.0175 0.258 Eomes 0.978 0.00744 0.162
A2m −2.09 0.0276 0.216 Il11 −1.68 0.014 0.241 Nkg7 −1.69 0.0286 0.266
Sfrp4 −1.98 0.0226 0.203 Ero1l −1.58 0.00223 0.113 Selp −1.44 0.0104 0.192
Nos2 −1.86 0.00582 0.094 Arg1 −1.47 0.0394 0.375 Tnfrsf11b −1.44 0.0207 0.222
Ero1l −1.47 0.00156 0.0731 Adm −1.41 0.0212 0.267 Hk2 −1.42 0.00219 0.112
Hk2 −1.41 0.00227 0.0731 Fstl3 −1.12 0.000866 0.0882 Ero1l −1.27 0.00483 0.139
Vegfa −1.15 0.00281 0.0745 Bnip3 −1.09 0.00845 0.197 Ndufa4l2 −1.27 0.0059 0.152
Selp −1.14 0.0355 0.244 Vegfa −1.07 0.0102 0.202 Adm −1.22 0.025 0.239
Ndufa4l2 −1.22 0.00788 0.105 Cdkn1a −1.07 0.0013 0.0943 Vegfa −1.1 0.00387 0.124
Itpk1 −0.989 3.5×10−5 0.0211 Dusp1 −1.06 0.00021 0.0638 P4ha2 −1.02 0.0015 0.102
Bnip3 −0.984 0.00789 0.105 Slc2a1 −1.02 0.00446 0.134 Col5a1 −1.03 0.0401 0.295

B.P., Benjamini Hoghberg.

Up- and downregulation of immune-related genes.

Gemcitabine vs. control IFN-β vs. control Combination vs. control
Antigen presentation ↑ Ctsw, Cd276, Gzmb, H2afx, Hells, Sfxn1, Snca ↓ Atf3, Bnip3, Ccl4, Cd274, H2-K1, H2-M3, H2-T23, Parp4, Rad50, Srebf1, Tnks ↑ Cd1d1, Cxcl10, Cxcl11, Cd36, Cybb, H2-D1, H2-T23, Irf1, Lck, Ldhb, Parp9, Parp12, Psmb9, Psmb10, Tap1, Tap2, Tapbp, Uba7 ↑ Trim21, Ctsw, Lag3, Snca ↓ Bnip3, Hdac4, Mdm2, Mlh1, Nkg7, Rad51c, Srebf1
↓ Bnip3, Gzme, Ill1, Itgav, Ldha, Pms2
Costimulatory signalling ↑ Cd274, Mtor, Rictor ↑ Il18bp, Lck, Psmb10, Psmb9 ↑ Havcr2, Hck, Il18bp, Lag3
↓ Ptgs2
Cytokine and chemokine signalling ↑ Ccl4, Jak3, Rock1 ↑ Ccl19, Cxcl10, Cxcl11, Il18bp, Stat2 ↑ Cxcr2, Il18bp, Stat2
↓ Rock1
↓ Il11
Immune cell adhesion and migration ↑ Cd276, Clec5a, Clec7a ↑ H2-D1, H2-T23 ↑ Clec7a, Cybb
↓ Cd274, H2-K1, H2-M3, H2-T23, Pvr, Rock1, Selp ↓ Itgav, Pvr ↓ Cdh1, Pvr, Rock1, Selp
Interferon signalling ↓ H2-K1, H2-T23, Ifi203, Igf2r, Irf2 ↑ Gbp2, Gbp3, H2-D1, Ifi35, Irf1, Irf7, Stat2, Uba7 ↑ Ifi35, Mx1, Oas1a, Trim21
JAK-STAT signalling ↑ Cdkn1a, Jak3, Lif, Mtor ↑ Stat2 ↑ Stat2
↓ Cdkn1a, Il11, Lif ↓ Cdkn1a
Lymphoid compartment ↑ Eomes, Gzmb, Lck, Tbx21 ↑ cd1d1, cxcl10, cxcl11, eomes, ikzf3, stat2, tbx21 ↑ Eomes, Lag3, Mx1, Stat2
↓ Cd274, Igf2r, Pvr ↓ F2rl1, gzme, pvr ↓ Igf2r, Pvr
Myeloid compartment ↑ Clec5a, Clec7a, Fcer1a, Lilra5, P2ry13, S100a8, S100a9 ↑ Fcer1a, P2ry13 ↑ Clec7a, Col11a1, Cybb, Hck, Lilra5, P2ry13
↓ Areg, Arg1, Cdkn1a, Ier3, Lif, Ptgs2
↓ Areg, Cdkn1a, Dll4, Ier3, Lamb3
↓ Areg, Ccl4, Cdkn1a, Ier3, Lmab3, Lif, Nos2
NF-kappaB signalling ↑ Ltb ↑ Psmb10, Psmb9 ↓ Tnfrsf11b