Background
Small Cell Lung Cancer (SCLC) is an aggressive tumor characterized by rapid and extensive metastatic dissemination, recurrence after chemotherapy and poor prognosis. Therefore, there is an urgent need of new treatment modalities, and promising results have been achieved in recent phase I-II studies of immunotherapy [
1]. In general, immune checkpoint blockade through monoclonal antibodies targeting PD-1, PD-L1, and/or CTLA-4 have shown unprecedented activity in several metastatic malignancies, including melanoma and Non-Small Cell Lung Cancer (NSCLC) [
2‐
4]. A recent phase I-II trial of the anti-PD-1 antibody nivolumab in patients with recurrent SCLC showed a 10% response rate and a 32% disease control rate [
5]. In addition, different schedules of nivolumab in combination with ipilimumab showed 19–23% response rates [
5]. These results prompted the National Comprehensive Cancer Network to consider the nivolumab-ipilimumab combination in the 2016 guidelines for SCLC treatment.
It is now well established that immune checkpoint blockers re-activate pre-existing, silenced CTL cell responses against tumor neo-antigens, in metastatic melanoma and NSCLC [
6‐
8]. However, SCLC is a tumor lacking HLA class I expression and, accordingly, should be resistant to the activity of HLA-restricted CTLs [
9]. Nonetheless, IFN-γ is capable to restore HLA class I expression [
10] and sensitivity to CTL-mediated recognition of SCLC [
11,
12], as well as of other tumor cells showing down-regulated HLA class I antigen expression [
13]. In view of its immune-modulatory and direct anti-tumor effects, clinical studies of IFN-γ have been performed in different cancers, with some evidence of activity in ovarian and bladder cancer. However, no activity was found in other cancers and adverse effects including toxicity or even tumor progression have been recorded (reviewed in [
14]). These findings may relate, at least in part, to the ability of IFN-γ to activate immune-regulatory loops, for example through the induction of PD-L1 or indoleamine 2,3 dioxigenase (IDO) [
15].
We recently reported that IL-27, a heterodimer cytokine related to both IL-6 and IL-12 cytokine families [
16,
17], has several functional activities in common with IFN-γ, in different cancer cells [
18]. Indeed, IL-27 up-regulates multiple components of the HLA class I antigen presentation machinery in human cancer cells, thus facilitating Cytotoxic T Lymphocyte (CTL) recognition. Moreover, in lymphoid cells, IL-27 induces the expression of the transcription factor T-bet, an inducer of Th1 and CTL responses [
19], which have been involved in the anti-tumor activity of IL-27. IL-27 may also exert direct anti-tumor effects through the inhibition of angiogenesis and neoplastic cell proliferation in different cancers including acute myeloid leukemia [
20], prostate cancer [
21], and melanoma [
22]. In addition, IL-27 inhibits the expression of stem cell and mesenchymal transition genes in NSCLC cells [
23,
24]. In view of its immune-enhancing activities and direct anti-tumor effects IL-27 has been considered as a potential anti-tumor agent [
25]. On the other hand, IL-27 induces the expression of immune-regulatory molecules such as the IL-18 natural inhibitor, IL-18BP [
26], the tryptophan catabolic enzyme IDO and PD-L1 [
27], in neoplastic cells. Therefore, IL-27 may have a dual role in anti-tumor immunity [
28] and shares several immune-regulatory functions with IFN-γ, in relationship to the common usage of the STAT1 intracellular signaling pathway [
18,
19,
23,
29].
At the best of our knowledge, no studies have addressed the effects of IL-27 on SCLC cells, so far. In this study we tested the effects of IL-27 on a panel of SCLC cell lines and found that it is capable to restore HLA class I expression through the up-regulation of peptide transporters and other components of the class I antigen presentation machinery in most SCLC lines tested. We also explored the expression and signaling properties of the IL-27 receptor complex, a heterodimer of IL27RA/WSX1 and GP130 chains [
30]. IL-27 and IL-6 share the usage of the GP130 chain and downstream signaling pathways through STAT1/3 [
30]. In addition, serum IL-6 is elevated in SCLC in relationship with advanced stages, worse prognosis and Neuron Specific Enolase (NSE) levels, suggesting a possible role of IL-6 in SCLC progression [
31], similarly to other cancers [
32]. Therefore, we also tested the possible effects of IL-6 on SCLC cells and dissected the role of the GP130 molecule by the usage of a sIL-6R/IL-6 synthetic ligand, similar to “hyper-IL-6” [
33], in both IL-27-responsive and unresponsive SCLC cells.
Methods
Cells and treatments
The human SCLC cell lines NCI-H69, NCI-H146, NCI-H446, NCI-H82 were purchased from ATCC and NCI-N592 was kindly provided by Dr. J. Minna (NCI, Washington DC). Cells were grown in RPMI 1640, with L-glutamine, 10% FCS and antibiotics (Lonza) and never kept in culture for longer than 4 months, when an aliquot of the original stock was thawed. Treatments with cytokines were performed with slight differences, according to the final use of the stimulated samples. Conditions were set on the bases of preliminary titration experiments. For immunofluorescence and QRT-PCR analyses, cells were seeded in 24-well plates in culture medium at 5 × 104 cells/well and different cytokines were added: IFN-γ (1000 IU/ml, PeproTech, 300–02), IL-27 (100 ng/ml R&D System, 2526-IL-010), IL-6 (50 ng/ml R&D System 206-IL-010) or recombinant human IL-6Rα/IL-6 chimera [sIL-6R/IL-6] (50 ng/ml R&D System 8954-SR-025). Treatments were carried out for 48 h.
For the analysis of tyrosine-phosphorylated STAT proteins, 1 × 105 SCLC cells were incubated in a test tube at 37 °C with or without 50 ng/ ml of IL-27, 20 ng/ml of IL-6, 40 ng/ml of sIL-6R/IL-6 in 0.5 ml of medium for the 10, 30 or 60 min time points. Incubations for the 1, 2, 3 or 4 h time points were carried out in 1 ml culture medium. Treatments for the 6, 18 and 24 h time points were performed in 24-well plates, in 1 ml culture medium. Cells were then rescued by centrifugation and immediately processed.
Immunofluorescence
Immunofluorescence with anti-GP130 PE, anti-IL-27RA/WSX1/TCCR APC (R&D Systems, Clones 28,126 and 191,116), anti-PD-L1 PE or Isotype Control PE (eBioscience Bender, BMS-125983-41 and BMS-124724-41, respectively), anti-human IL-6Rα FITC (R&D Systems, Clone 17,506 FAB227F-025) was performed according to manufacturer’s instructions. Indirect immunofluorescence was performed on 5 to 10 × 104 cells/sample with anti-HLA class I W6/32 mAb (ATCC) and FITC-labeled goat anti-mouse (Jackson Immunoresearch, 115–096-068) according to standard techniques. Cell fluorescence was analyzed by flow cytometry with a FACScan (Becton & Dickinson) using the Cell Quest software or a Gallios (Beckman Coulter). Gating on viable cells was performed using physical parameters and 104 gated events were acquired.
Western blot
Cells were lysed in lysis buffer (20 mM Tris-HCl pH 7.4, 1 mM EDTA, 150 mM NaCl, 1% Brij97) containing 2 mM Na Orthovanadate and protease inhibitors (Roche Diagnostics, Complete Mini 04693124001). Lysates were resolved under reducing conditions by SDS-PAGE (10% or 13% acrylamide) and analyzed by Western blotting using the following antibodies: rabbit anti-phospho-STAT1 (pY701) and anti-STAT1 anti-sera (Cell Signaling Technology, 9167 and 9172, respectively), murine anti-phospho-STAT3 (pY705) and anti-STAT3 mAbs (BD Transduction Laboratories, 612,356 and 610,190, respectively), rabbit anti-SOCS3 (Cell Signaling Technology 2932) and murine α-tubulin or β-actin mAbs (Sigma-Aldrich T6074 and A2228, respectively). Proteins were detected by ECL Prime (GE Healthcare, RPN2232) and visualized by a chemiluminescence gel documentation and analysis system (MINI HD, UVITEC, Cambridge).
RT-PCR analysis
Total RNA was isolated by the NucleoSpin RNA kit (Macherey-Nagel, 740,955.250) and reverse-transcribed using the SuperScript III Reverse Transcriptase (Invitrogen, 18,064–071). Amplification was carried out by the Mastercycler® ep realplex4 instrument (Eppendorf International) using the iQTM SYBR® Green Supermix system (Bio-Rad Laboratories, 170–8882). Quantification of mRNAs relative to housekeeping gene was expressed as 1/ΔCT. Expression levels of mRNAs relative to untreated control were calculated by the ΔΔCT method.
Statistical analysis
Data are expressed as the mean ± standard deviation (SD) of triplicates. The Student’s T test was used for the statistical analysis of experimental replicates. A value of p < 0.05 was considered significant.
Discussion
In this study we show, for the first time, that cultured SCLC cells respond to IL-27 by activating STAT1, and to a lesser extent STAT3, tyrosine phosphorylation and by up-regulating HLA class I surface molecule expression. SCLC is an aggressive cancer, which usually shows down-regulated expression of HLA class I molecules and may therefore escape from immune recognition by CTLs [
9,
10]. SCLC cells express tumor associated antigens, which can be recognized by CTLs including, for example, the cancer/testis antigens MAGE-1 and -3 [
11], the ion channel gBK [
12], recoverin [
34] and the neuron-associated protein Hu [
35]. Several lines of evidence suggest that CTLs may play a protective role in SCLC. Indeed, CTLs recognizing SCLC-associated antigens have been isolated from patients with long-term survival [
36] eventually associated with concomitant autoimmunity due to antigen sharing by SCLC and normal tissues [
34,
35]. In addition, the recent finding that SCLC patients respond to anti-PD-1 antibodies also supports a role for CTLs in SCLC immune control [
5]. IFN-γ is able to restore HLA class I expression and SCLC antigen recognition by CTLs, in vitro, through the induction of different components of the antigen-presentation machinery, among which are TAP1 and TAP2 [
11,
12]. However, IFN-γ is a stronger inducer of PD-L1 expression than IL-27, in SCLC and may therefore limit anti-tumor responses by PD-1 expressing CTLs in vivo. In this respect, IL-27 may offer a better chance to induce HLA class I expression with a limited effect on PD-L1 induction.
The induction of surface HLA class I expression by IL-27 and IFN-γ is related to the activation of the STAT1 signaling pathway, which is an essential mediator of both IFNs and IL-27 biological effects [
19,
23,
26,
27]. Indeed, recent studies by proteomics [
18] or gene expression profiling [
29] in different cell types showed a broad overlap between IFN-γ and IL-27 effects, including the induction of several components of the HLA class I antigen presentation machinery. However, the NCI-H146 cell line was completely unresponsive to IL-27 both in terms of HLA class I induction and STAT1/3 signaling, suggesting heterogeneity in the response to IL-27 in different SCLC. In such cases, IFN-γ may be a more suitable agent to restore CTL responses against SCLC, as also the IL-27-resistant SCLC cells were sensitive to IFN-γ induction. Our present findings also suggested that IL-27 unresponsiveness could be related to specific alterations in the IL-27R complex. Indeed, the IL-27-resistant cell line NCI-H146 showed reduced expression of IL-27RA and GP130 receptor chains, and failed to activate STAT3 signaling in response to sIL-6R/IL-6, a direct agonist of GP130 [
33].
IL-6 is a pro-inflammatory cytokine, which plays an important role also in the progression and immune-regulation in several types of cancers through the activation of the STAT3 pathway [
37,
38]. Particularly, in SCLC patients IL-6 levels were elevated in the circulation and related to worse survival [
31]. In addition, phosphorylated STAT3 was constitutively expressed in the SCLC tumors in vivo [
39] and a very recent report supports a role for IL-6 produced by tumor-associated macrophages in the paracrine activation of STAT3, in SCLC [
40]. Finally, NCI-H446 SCLC cells can produce IL-6 in response to Hypoxia-inducible factor 1α and express IL-6/STAT3-related genes [
41]. Accordingly, in the present report, we could not detect constitutive tyrosine phosphorylation of STAT3 in all SCLC cell lines studied but IL-6 induced STAT3 activation only in the NCI-H446 cell line. The other SCLC cell lines failed to respond to IL-6 but three of them showed STAT3 activation upon sIL-6R/IL-6 stimulation. sIL-6R/IL-6 is similar to hyper-IL-6, an IL-6/IL-6Rα extracellular domain chimera, which mimics natural, soluble IL-6/IL-6Rα complexes and can directly activate the GP130 receptor in cells lacking IL-6Rα [
33]. Indeed, IL-6 signaling through the GP130 chain requires the IL-6Rα, either as cell surface-bound molecule or in a soluble form in complex with IL-6 [
42]. Accordingly, the IL-6-responsive cell line NCI-H446 expressed
IL6RA gene, which was less expressed in the IL-6-unresponsive, sIL-6R/IL-6-sensitive NCI-N592, NCI-H69 and NCI-H146 cells. Our present data may prompt further studies to identify the potential role of natural IL-6/IL-6Rα complexes in the in vivo activation of STAT3 and in the pathogenesis of SCLC.
Previous findings indicated that GP130 signaling, induced by IL-6 ligands, may switch-off the responsiveness to IL-27, through the induction of the suppressor of cytokine signaling, SOCS3, in human hepatocellular carcinoma cells [
29]. SOCS are P-STAT-inducible proteins, which act as potent blockers of STAT protein signaling in a feed back loop [
43]. In view of the potential role of IL-6 ligands in SCLC, we tested whether sIL-6R/IL-6 may alter IL-27 effects in SCLC cells. The interference of sIL-6R/IL-6 was particularly evident in the NCI-H69 cell model, where the concomitant presence of sIL-6R/IL-6 inhibited IL-27-driven surface HLA class I expression by 90%. However, the inhibitory effect of IL-6 signaling seemed not related to SOCS3 induction in NCI-H69 cells because: i) pre-stimulation with sIL-6R/IL-6 followed by removal shortly before IL-27 stimulation failed to interfere with HLA class I induction, ii) SOCS3 is constitutively expressed by NCI-H69 cells and is not further induced by sIL-6R/IL-6 and iii) IL-27 induces STAT1 signaling also in the presence of sIL-6R/IL-6. The role of SOCS3 in cancer is still controversial, because SOCS3 has been reported as a tumor-suppressor or tumor-promoting molecule, in different tumors [
44]. For example, in pancreatic cancer SOCS3 expression correlates with a better prognosis and overexpression of SOCS3 limits tumor growth, while SOCS3 silencing by promoter methylation has opposite effects [
45]. On the other hand, constitutive SOCS3 expression has been reported in human melanomas, where it can inhibit responsiveness to IFNs and may contribute to resistance to IFN-treatment [
46,
47]. Somehow surprisingly, this is not the case of NCI-H69 cells, which respond very well to both IFN-γ and IL-27, in spite of constitutive SOCS3 expression. An alternative explanation for the sIL-6R/IL-6-mediated inhibition of IL-27 effects, is that the concomitant signaling of sIL-6R/IL-6 and IL-27 alters the balance between tyrosine-phosphorylated forms of STAT1 and STAT3 (Fig.
7). Indeed, IL-27 is a weak inducer of STAT3 and a much better inducer of STAT1, whereas sIL-6R/IL-6 is a strong inducer of STAT3. During simultaneous stimulation by the two cytokines, phosphorylated STAT1/STAT3 heterodimer formation may be favored [
48]. Although the functional role of such heterodimers is still poorly understood, their formation may reduce the availability of phospho-STAT1 homodimers, which are mediators of IL-27 intracellular functions. In the case of IFN-γ signaling, the inhibitory effect of sIL-6R/IL-6 is negligible, as IFN-γ is a stronger inducer of STAT1 phosphorylation than IL-27. Whatever the mechanism(s) of IL-6 interference with IL-27 biological activity, it may be overcome in therapeutic settings by the use of the IL-6/IL-6R blocking agents [
49] or STAT3 inhibitors [
50].