GNA15 expression in small intestinal neuroendocrine neoplasia: Functional and signalling pathway analyses

Cellular Signalling 27 (2015) 899–907

Contents lists available at ScienceDirect

Cellular Signalling journal homepage: www.elsevier.com/locate/cellsig

GNA15 expression in small intestinal neuroendocrine neoplasia: Functional and signalling pathway analyses Sara Zanini a,b, Francesco Giovinazzo a,b,⁎, Daniele Alaimo a, Ben Lawrence a, Roswitha Pfragner c, Claudio Bassi b, Irvin Modlin a, Mark Kidd a,⁎ a b c

Gastrointestinal Pathobiology Research Group, Yale University School of Medicine, New Haven, CT 06520-8062, USA Department of Surgery, Laboratory of Translational Surgery, LURM, Hospital of G.B. Rossi, University of Verona, Piazzale L.A. Scuro, IT-37134 Verona, Italy Institute of Pathophysiology and Immunology, Centre for Molecular Medicine, Medical University of Graz, Graz, Austria

a r t i c l e

i n f o

Article history: Received 27 September 2014 Received in revised form 18 January 2015 Accepted 2 February 2015 Available online 17 February 2015 Keywords: GNA15 Gα15 Small intestinal neuroendocrine neoplasia G protein KRJ-I cell line ß1 adrenergic receptor

a b s t r a c t Gastroenteropancreatic neuroendocrine neoplasia (GEP-NEN) comprises a heterogeneous group of tumours that exhibit widely divergent biological behaviour. The identification of new targetable GPCR-pathways involved in regulating cell function could help to identify new therapeutic strategies. We assessed the function of a haematopoietic stem cell heterotrimeric G-protein, Gα15, in gut neuroendocrine cell models and examined the clinical implications of its over expression. Functional assays were undertaken to define the role of GNA15 in the small intestinal NEN cell line KRJ-I and in clinical samples from small intestinal NENs using quantitative polymerase chain reaction, western blot, proliferation and apoptosis assays, immunoprecipitation, immunohistochemistry (IHC) and automated quantitative analysis (AQUA). GNA15 was not expressed in normal neuroendocrine cells but was overexpressed in GEP-NEN cell lines. In KRJ-I cells, decreased expression of GNA15 was associated with inhibition of proliferation, activation of apoptosis and differential effects on pro-proliferative ERK, NFκB and Akt pathway signalling. Moreover, Gα15 was demonstrated to couple to the ß1 adrenergic receptor and modulated proliferative signals through this GPCR. Transcript and protein levels of GNA15 were significantly elevated in primary and metastatic tumours compared to normal mucosa and were particularly increased in low Ki-67 expressing tumours. IHC and AQUA revealed that a higher Gα15 expression was associated with a poorer survival. GNA15 may have a pathobiological role in SI-NENs. Targeting this signalling mediator could provide an opportunity for the development of new therapeutic strategies for this tumour type. © 2015 Elsevier Inc. All rights reserved.

1. Introduction Gastroenteropancreatic neuroendocrine neoplasia (GEP-NENs) includes a heterogeneous group of tumours that typically occur in the gut epithelial lining and presumably originate from neuroendocrine cells or their precursors [1]. GEP-NENs are usually found in the small and large intestines (about 80%) while the remainder are localized in the stomach and pancreas. Recent data demonstrate that the incidence has increased significantly over the last three decades; they now comprise ~2% of all malignancies [2,3]. Although non-transformed neuroendocrine cells can be identified by the expression of specific proteins such as synaptophysin, neuron-specific enolase, and chromogranin A [4], there are at least seventeen different neuroendocrine cell types with a range of biological activity and behaviour [5]. While this ⁎ Corresponding authors at: Dept. of Gastroenterological Surgery, Yale University School of Medicine, 333 Cedar Street, P.O. Box 208062, New Haven, CT 06520-8062, USA. Tel.: +1 203 785 5429; fax: +1 203 737 4067. E-mail address: [email protected] (M. Kidd).

http://dx.doi.org/10.1016/j.cellsig.2015.02.001 0898-6568/© 2015 Elsevier Inc. All rights reserved.

constitutes the largest group of hormone producing cells in the body, their secretory and proliferative regulation remains incompletely understood [4,6,7]. G-protein coupled receptors (GPCRs) represent the largest family of cell-surface molecules involved in signal transmission, and account for more than 2% of the total genes encoded by the human genome [8]. GPCR activation has been linked to many cellular functions including neurotransmission, hormone and enzyme release from endocrine and exocrine glands and the regulation of cell proliferation while aberrant activation has been linked to both tumourigenesis and metastasis [9]. The heterotrimeric G protein, Gα15, and its α subunit belongs to the Gαq subfamily with a molecular weight of 43.5 kDa [10]. Gα15 is expressed in highly specific cell types, e.g. haematopoietic [10,11] and epithelial cells [12] during specific stages of differentiation. In particular, its expression is high in haematopoietic progenitor cells but decreases with cell maturation [13]. One of the features of Gα15 is its ability to couple to a wide range of Gs-,Gi- and Gq coupled receptors [14]. In addition, Gα15 is not effectively attenuated by β-arrestin dependent desensitization [15]. This indicates that an activated Gα15 could support

900

S. Zanini et al. / Cellular Signalling 27 (2015) 899–907

sustained signalling in cells chronically exposed to hormones and neurotransmitters [16]. One example is prolongation of β 1 adrenergic receptor signalling [17,18]. Stimulation of this receptor with nitrosamine 4(methylnitrosamino)-1-(3-pyridyl)-1-butanone (NNK, nicotinederived nitrosamine ketone), a high affinity agonist for ß1 adrenergic receptor (βAR) [19], resulted in both lung and exocrine pancreatic tumours in rats [20]. In this model, Gα15 was noted to regulate (and likely prolong) βAR signalling, with the resultant development of neoplasia. Transcriptome analyses of a small intestinal neuroendocrine neoplasia (SI-NEN) cell line, KRJ-I, has identified expression of both Gα15 as well as βAR [7,21]. Based on this observation, as well as the known role of Gα15 and G-protein mediated signalling in normal and neoplastic cells, we hypothesized that GNA15 was ectopically expressed in neuroendocrine tumours and could play a functional role in regulating cell proliferation. We examined GNA15 expression in these tumours and evaluated the effect of GNA15 silencing on extracellular signalregulated kinase (ERK)1/2, nuclear factor kappa-B (NFκB) and AktP70S6K/P85S6K in the KRJ-I cell line. We next evaluated the role of Gα15 on neuroendocrine tumour proliferation after β1 adrenergic receptor stimulation with NNK. We also examined the clinical relevance of GNA15 expression in tumour tissue. 2. Materials and methods 2.1. Cell lines and enterochromaffin cell isolation Normal small intestinal enterochromaffin (EC) cells were isolated as described; ~1 × 106 cells were obtained per sample [7]. The SI-NEN cell lines KRJ-I, P-STS (both primary tumours), H-STS (liver metastasis), and L-STS (lymph node metastasis) were cultured in media containing M199:Hams F12 (1:1 ratio), 10% foetal bovine serum (FBS) gold, and penicillin and streptomycin (100 IU/mL) [22]. The human pancreatic NEN (P-NEN) cell lines were cultured in media containing RPMI 1640:Hams F12 (1:1 ratio), 10% FBS gold, and penicillin and streptomycin (100 IU/mL) (metastatic carcinoid BON cell line) [23], in media containing RPMI 1640, 10% FBS gold, and penicillin and streptomycin (100 IU/mL) (metastatic insulinoma CM cell line and somatostatinoma QGP1 cell line) [24]. Cells were incubated at 37 °C with 5% CO2. 2.2. KRJ-I cells GNA15 silencing with short hairpin RNA (shRNA) lentiviral particles KRJ-I cells were seeded (2.5 × 105 cells/ml) in 6-well plate (Falcon; BD, Franklin Lakes, NJ) and treated with different concentrations of shRNA lentiviral particles specific for GNA15: sh64: TRCN0000036464_CCGGCCCTATAAAGTGACCACGTTTCTCGA GAAACGTGGTCACTTTATAGGGTTTTTG sh65: TRCN0000036465_CCGGCCTCGCATTGTTTGGGACTATCTCGAG ATAGTCCCAAACAATGCGAGGTTTTTG sh66: TRCN0000036466_CCGGCCATTGTTTCGAGAACGTGATCTCGA GATCACGTTCTCGAAACAATGGTTTTTG sh67: TRCN0000036467_CCGGCCTGCTCGATTCAGCCGTGTACTCGA GTACACGGCTGAATCGAGCAGGTTTTTG sh68: TRCN0000036468_CCGGCAAGAGGTTCATCCTGGACATCTCGA GATGTCCAGGATGAACCTCTTGTTTTTG and non-targeting control (SIGMA Aldrich, St. Louis, MO) for 48 h. We tested 0.5 and 1.5 multiplicity of infection (MOI) of lentiviral particles to obtain at least a 50% reduction in GNA15 expression. 2.3. KRJ-I cell ß1 adrenergic receptor stimulation 2 × 104 cells were seeded in 96-well plate (clear bottom, Costar, Corning, NY) and stimulated with NNK (EC50 = 10−10 M). A subset of

studies were undertaken with a combination of NNK and atenolol (IC50 = 10−9), a highly specific ß1 adrenergic receptor antagonist [25] to evaluate the specificity of the βAR-mediated signalling. 2.4. RNA isolation and reverse transcription RNA was extracted from GEP-NEN cell lines (1 × 106) using TRIzol (Invitrogen, Carlsbad, CA) and cleaned (QIAGEN, RNeasy Kit, Qiagen, Valencia, CA) or from KRJ-I cells after ß1 adrenergic receptor stimulation using the QIAGEN RNA isolation kit (QIAGEN) according to the manufacturers' instructions. After conversion to complementary DNA (High Capacity cDNA Archive Kit; Life Technologies, Carlsbad, CA) [26], quantitative reverse-transcription polymerase chain reaction (qRT-PCR) analyses were performed using Assays-on-Demand and the ABI 7900 Sequence Detection System [27]. All primer sets were obtained from Life Technologies. Primers for GNA15 were: Hs00976150_m1 (Exons 4–5), Hs00976152_m1 (Exons 6–7), Hs00157720_m1 (Exons 1–2), Hs00976149_m1 (Exons 3–4). PCR data were normalized using the ΔΔC T method; the asparagine-linked glycosylation 9 gene ALG9 was used as housekeeping gene [28]. 2.5. Protein extraction Lysates from GEP-NEN cell lines (1 × 106 cells) or GNA15-silenced KRJ-I cells (1 × 10 6 cells) were prepared by adding 200 μL of icecold cell lysis buffer (10 × RIPA lysis buffer [Millipore, Billerica, MA], complete protease inhibitor [Roche, Indianapolis, IN], phosphatase inhibitor sets 1 and 2 [Calbiochem, Gibbstown, NJ], 100 mM phenylmethanesulfonyl fluoride [29], 200 mM Na3VO4 [Acros Organics], 12.5 mg/mL sodium dodecyl sulfate [SDS; American Bioanalytical, Natick, MA]). Tubes were maintained in constant agitation for 30 min at 4 °C and then centrifuged at 12,000 g for 20 min at 4 °C, and supernatant protein was quantified (BCA protein assay kit; Thermo Fisher Scientific, Rockford, IL). 2.6. Western blot (WB) analysis Total protein lysates (20 μg) were denaturated in SDS sample buffer, separated on an SDS-polyacrylamide electrophoresis gel (4, 10%), and transferred to a polyvinylidene fluoride membrane (pore size, 0.45 mm; Bio-Rad, Hercules, CA). After blocking (5% bovine serum albumin (BSA)/phosphate-buffered saline (PBS)) for 60 min at room temperature, membranes were incubated with primary antibodies pAkt (Ser473) and Akt, pERK1/2 (Thr185, Tyr187) and ERK1/2, pNFkB and NFkB, pP70S6K/pP85S6K (Thr389) and P70S6K/P85S6K (all Cell Signaling Technology, Beverly, MA) in 5% BSA/PBS/Tween-20 overnight at 4 °C. Membranes were then incubated with horseradish peroxidaseconjugated secondary antibodies (Cell Signaling Technology) for 60 min at room temperature, and immunodetection performed using Western Lightning Plus-ECL (Perkin Elmer, Waltham, MA). Blots were exposed on X-Omat AR film (Eastman Kodak, Rochester, NY). Cross-detection was avoided by stripping the membranes; protein expression was reported relative to that of β-actin (Sigma-Aldrich, St. Louis, MO). The optical density of the appropriately sized bands was measured using ImageJ software (National Institutes of Health, Bethesda, MD) [26]. 2.7. BrdU proliferation assay KRJ-I cells were seeded (2 × 104) in 96-well plate (clear bottom, Costar, Corning) after GNA15 silencing or ßAR stimulation and proliferation was assed using Bromodeoxyuridine (BrdU) ELISA kit (Roche Diagnostics, Indianapolis, IN) according to the manufacturers' instructions. Briefly, after silencing or stimulation, KRJ-I cells were incubated overnight with BrdU uptake solution. Subsequently, cells were fixed, DNA denatured and anti-BrdU antibody solution added and incubated for 120 min. Thereafter, the final substrate was added and the

S. Zanini et al. / Cellular Signalling 27 (2015) 899–907

chemiluminescent emission was read on a GLOMAX Luminometer (Promega, Madison, WI). Luminescence (relative values) in GNA15silenced cells was compared to control cells infected with a scramble shRNA and represented as fold-decrease.

901

activity, was read on a GLOMAX Luminometer (Promega). Luminescence (relative values) in GNA15 silenced cells was compared to control cells infected with a scramble shRNA and represented as fold-increase. 2.10. Immunoprecipitation

2.8. TUNEL assay KRJ-I cells were seeded (2 × 104) in 96-well plate (clear bottom, Costar, Corning) and apoptosis was assed using TUNEL staining (In Situ Cell Death fluorescent kit, Roche Diagnostics). Briefly, control or sh-transfected cells were fixed with 4% paraformaldehyde for 1 h and permeabilized in a 0.1% Triton X-100 — 0.1% sodium citrate solution for 2 min on ice. Subsequently, 50 μL of TUNEL reaction mixture was added and cells were incubated for 1 h at 37 °C. Cells were counterstained with DAPI before mounting. Positive cells were counted and quantitated versus control (scrambled sh) [30]. 2.9. Caspase 3/7 apoptosis assay KRJ-I cells were seeded (2 × 104) in 96-well plate (clear bottom, Costar, Corning) and apoptosis was assessed using the Caspase-Glo 3/7 apoptosis assay (Promega). Briefly, after silencing KRJ-I cells, they were incubated with 100 μL of Caspase-Glo 3/7 reagent for 1 h at room temperature. The luminescence emission, proportional to caspase 3–7

KRJ-I cells lysate was collected as described (Section 2.6). On ice, in a microcentrifuge tube, 2 μg cell lysates were mixed with 1 μg of GNA15 antibody (ABCAM, Cambridge, MA) or β1 adrenergic receptor antibody (ABCAM). Samples were incubated overnight at 4 °C under gentle agitation. Thereafter, 20 μL of protein G PLUS agarose (Santa Cruz Biotechnology Inc., Santa Cruz, CA) was added to each sample and incubated at 4 °C under gentle agitation for 4 h. Three steps of washing with lysis buffer were then performed at 4 °C for 4 min at 2500 rpm. For each wash the beads were gently mixed with lysis buffer and after centrifugation, the supernatant was discarded. Finally, the complex was eluted from the beads by adding 20 μL of sample loading buffer. The beads were then pelleted and the eluted samples boiled at 98 °C for 5 min. Samples were separated by SDS-page and analysed by WB. 2.11. Immunostaining An established immunohistochemical approach was used to identify target protein and chromogranin A (CgA) in paraffin-

Fig. 1. GNA15 expression in GEP-NEN cell lines. A) RNA expression of GNA15 in normal mucosa (n = 4), normal enterochromaffin (EC) cells (n = 3), normal pancreas (n = 4) and in seven neuroendocrine neoplasia (GEP-NEN) cell lines (4 small intestinal NEN [SI-NEN] cell lines, 3 pancreatic neoplasia [P-NEN] cell lines). GNA15 was expressed at significantly higher levels in NEN cell lines compared with the expression in normal EC cells (#p b 0.01 vs. normal EC cells). B) GNA15 expression was evaluated with all the commercially available primers (spanning exons 1–7). C) Protein expression of GNA15. Expression of protein was confirmed by Western blot in both normal intestinal mucosa and in normal pancreas. The specificity is demonstrated by pre-absorption with the target. D) GNA15 protein expression demonstrating positivity in all 7 cell lines. E) ImageJ quantitation of protein expression demonstrating high expression in 5 cell lines, particularly KRJ-I and H-STS (n = 3). SI = small intestinal; P = pancreas, NM = normal mucosa, NP = normal pancreas. L = lysate, L(PA) = lysate pre-absorbed.

902

S. Zanini et al. / Cellular Signalling 27 (2015) 899–907

embedded tissues [31]. Briefly, deparaffinized sections were incubated overnight at 4 °C with primary antibodies: rabbit anti-GNA15 10 μg/mL (NovusBiologicals, Littleton, CO) and mouse anti-CgA 1:100 (DAKO Corp, Carpinteria, CA). Goat anti-rabbit antibody conjugated to a horseradish peroxidase (DAKO) and goat anti-mouse Alexa fluor 488 conjugated (Invitrogen, Carlsbad, CA) were used as secondary reagents. Sections were rinsed with PBS (American Bioanalytical)/0.1% Tween20, and nuclei were stained with 4′, 6-diamidino-2-phenylindole (DAPI) (1:100 dilution) and the targets were visualized with a fluorescent chromogen (Cyanine-5-tyramide; NEN Life Science Products, Boston, MA). Bound antibodies were observed using immunofluorescent microscopy.

~ 1 cm3) and tumour tissue collected for histology. All studies were undertaken under IRB (IACUC) standards at Yale University. 2.13. Human sample collection Small intestinal (SI) tissues from patients with SI-NENs (n = 33) and normal SI mucosa (n = 13) were obtained according to a standard institutional review board protocol at the Yale University School of Medicine. Protocols included steps to minimize the time from resection to processing and freezing. 2.14. Tissue microarray immunostaining, image acquisition, and data analysis

2.12. Xenograft models To generate xenograft models, we injected 1 × 107 KRJ-I or H-STS cells in 100 μL PBS into the flank of 3 Ns/J-mice (Jackson Laboratory, Bar Harbor, ME) under anaesthesia [22,32]. All mice were euthanized 9 weeks after cell line implantation (palpable tumours evident

Slides from the carcinoid tissue microarray (TMA) YTMA60_2 containing SI-NEN samples (n = 74 in total, primaries n = 57, metastasis n = 17) were stained as previously described [27]. Briefly, for antigen retrieval purposes, sections were immersed in citrate buffer (10 mM sodium citrate, pH 6.0), and subjected to 1 × 10 min high temperature-

Fig. 2. GNA15 silencing with shRNA lentiviral particles. A) 0.5 multiplicity of infection (MOI) and 1.5MOI of short hairpin (shRNA) lentiviral particles were used to infect KRJ1. A scramble sequence, used as control (K-CON) and five different shRNA sequences were used (sh64, sh65, sh66, sh67, sh68). An average reduction of 15.38% ± 15.01% standard deviation (SD) and of 58.86% ± 2.05 SD in the GNA15 expression was observed. B) In GNA15-silenced cells, proliferation assessed by Bromodeoxyuridine (BrdU) uptake was significantly decreased (⁎p b 0.001 vs. scramble sh-silenced cells). C) In GNA15-silenced cells caspase pathway activity, measured by activity of caspase 3/7 was significantly increased (⁎p b 0.05 vs. scramble sh-silenced cells). D) Apoptosis, assessed by staining and quantitation of TUNEL-positive cells was significantly increased in GNA15-silenced cells (⁎p b 0.05).

S. Zanini et al. / Cellular Signalling 27 (2015) 899–907

high pressure treatment followed by treatment with 0.3% H2O2 in methanol for 30 min at 37 °C (inactivate endogenous peroxidases). Slides were then incubated overnight at 4 °C with 10 μg/mL anti-GNA15 rabbit polyclonal antibody (NovusBiologicals). Goat anti-rabbit antibody conjugated to horseradish peroxidase (DAKO) was used as a secondary reagent. For automated analysis, NEN cells or normal mucosal epithelia were identified by the use of an anti-pan-cytokeratin antibody mouse monoclonal antibody 1:100 (Cell Signaling Technology), nuclei and target were visualized respectively by DAPI and with a fluorescent chromogen (Cy-5-tyramide; NEN Life Science Products). Monochromatic, high-resolution (1024 × 1024 pixel; 0.5 μm) images were obtained for each histospot. Areas of tumour or normal epithelia were distinguished from stromal elements by creating a mask from the cytokeratin signal. Coalescence of cytokeratin at the cell surface localized the cell membranes and DAPI was used to identify nuclei. The GNA15 signal from the tumour cells was scored and expressed as signal intensity divided by the cytokeratin mask area. Histospots containing b 10% tumour, as assessed by mask area (automated), were excluded from further analysis. Previous studies have demonstrated that the staining from a single histospot provides a sufficiently representative sample for analysis [33]. 2.15. Statistical evaluation All statistical analyses were performed using Microsoft Excel (Microsoft, Redmond, WA), SPSS (version 17; SPSS Inc., Chicago, IL), and Prism 6 (GraphPad Software, San Diego, CA). The results are expressed as mean ± standard error of mean (SEM). Student t-test was used to compare continuous variables. Univariate analyses for survival were performed with the Kaplan–Meyer method with the log-rank test to verify significance of differences. 3. Results 3.1. GNA15 expression in GEP-NEN cell lines Transcript levels of GNA15 were analysed in isolated normal EC cells, in normal tissue and in the different NEN cell lines using qRT-PCR (primer Hs00157720_m1; covering exons 1–2). In normal human EC cell preparations (n = 3), expression of GNA15 was absent or only marginally evident but expression was detectable in haematopoietic stem cells (positive control) [34] (Fig. 1A). In contrast, GNA15 was significantly expressed in the three SI-NEN cell lines (KRJ-I, H-STS and L-STS; p b 0.01, t-test,

903

unpaired, 2 tailed) and in the two P-NEN cell lines (QGP1 and BON; p b 0.01, t-test, unpaired, 2 tailed). The greatest expression was noted in the KRJ-I cell line (15–743 fold respect to normal EC cells, p b 0.0001 1-way ANOVA test). GNA15 mRNA was marginally evident in P-STS and BON cells (Fig. 1A). To confirm these data, transcript levels (including exons 3–7) of GNA15 were then analysed with three additional commercially available primers. The PCR results identified consistent transcript levels of GNA15 (Fig. 1B) which demonstrated that the complete transcript (NM_002068; includes seven exons [12]) was expressed in these samples. Expression of GNA15 was confirmed by WB (Fig. 1C–E). The antibody was specific (Fig. 1C) and was highly expressed in KRJ-I cells (Fig. 1D). Quantitative assessment demonstrated elevated protein levels in the 5 cell lines that also exhibited elevated mRNA (Fig. 1E). 3.2. Functional evaluation of GNA15 silencing with shRNA lentiviral particles 3.2.1. GNA15 silencing and cell growth The reduction in GNA15 mRNA expression was 15.38 ± 15.0% with 0.5 MOI lentiviral particles and 58.86 ± 2.05% with 1.5 MOI (Fig. 2A). These reductions were confirmed at a protein level (Fig. 2A). 1.5MOI was therefore chosen for all the silencing experiments. Functionally, proliferation in KRJ-I cell line was significantly decreased (p b 0.001) (t-test, unpaired, 2 tailed) by GNA15 silencing (Fig. 2B). In addition, apoptosis, measured by both caspase 3/7 activity and TUNEL assays, was increased (p b 0.05 vs. control) (t-test, unpaired, 2 tailed) (Fig. 2C–D). 3.2.2. Effects on ERK, NFκB and Akt pathway signalling Having determined that GNA15 silencing altered the dynamics of cell proliferation and apoptosis, we next evaluated effects on ERK, NFκB and Akt signal pathways. Viral particles sh66 and sh68 decreased ERK and Akt phosphorylation (and therefore activation) (Fig. 3A). The inhibitory effects were consistent and significant (Fig. 3B). In contrast, phosphorylation of NFκB was increased after silencing with sh66 and sh68 (Fig. 3A–B). The variability between the different sequences for Gα15 silencing may reflect a threshold level of inhibition is required before a measureable effect can be detected in intracellular kinase activation. No significant effect was noted after silencing with sh64, sh65 or sh67 and no effect of any shRNA particle was observed on pP85S6K/P85S6K activation.

Fig. 3. ERK, NFkB and Akt pathway activation after GNA15 silencing. A) Western blot analysis of extracellular signal-regulated kinase (ERK)1/2, AKT, nuclear factor kappa B (Nf-kB) and P70S6K/P85S6K in KRJ-I cells after infection with shRNA lentiviral particles is shown. Beta-actin was used as a control. B) ImageJ quantitation of protein expression in these pathways confirming down-regulation of ERK and AKT signalling by sh66/sh68 and a concommitant upregulation of NFkB signalling (n = 3). ⁎p b 0.05 vs. CON.

904

S. Zanini et al. / Cellular Signalling 27 (2015) 899–907

Fig. 4. Demonstration that Gα15 is directly linked to the β1 adrenergic receptor. A) Co-immunoprecipitation of β1 adrenergic receptor or Gα15 followed by Western blot with antibodies targeting Gα15 or the β1 adrenergic receptor demonstrate that these are linked. Gα15 has a size of 44 kDa and the receptor is 51 kDa. Null = no antibody; + = antibody. B) Western blot analysis of extracellular signal-regulated kinase (ERK)1/2, AKT, and NFkB) in KRJ-I cells after infection with shRNA lentiviral particles (sh66) demonstrates that nitrosamine 4-(methylnitrosamino)-1-(3-pyridyl)-1-butanone (NNK, nicotine-derived nitrosamine ketone)-stimulated signalling through the β1 adrenergic receptor is attenuated by knockdown of Gα15. Conversely, NFkB signalling is amplified. C) ImageJ quantitation of protein expression in these pathways confirming NNK stimulates ERK and AKT signalling and that these effects are significantly inhibited by sh66. NFkB signalling was amplified by sh66 and further increased by NNK (n = 3). ⁎p b 0.05 vs. CON. D) BrdU incorporation assay in KRJ1 infected cells with shRNA lentiviral particles after stimulation with nitrosamine 4-(methylnitrosamino)-1-(3-pyridyl)-1-butanone (NNK, nicotine-derived nitrosamine ketone) and atenolol. ⁎p b 0.05 vs. control, #p b 0.05 vs. NNK stimulation, &p b 0.05 vs. NNK stimulation control (black bars), %p b 0.05 vs. NNK stimulation in GNA15 silenced cells. Dotted line = 1.0. bADR = β1 adrenergic receptor; NNK = nitrosamine 4-(methylnitrosamino)-1-(3-pyridyl)-1-butanone (NNK, nicotine-derived nitrosamine ketone).

3.3. The β1 adrenergic receptor: Gα15 axis and proliferation To demonstrate that Gα15 can couple to the ß1 adrenergic receptor, we used an immunoprecipitation protocol for both Gα15 and ß1 adrenergic receptor. After SDS-polyacrylamide gel electrophoresis and WB analysis with the ß1 adrenergic receptor or Gα15, the presence of the specific bands on the membranes, corresponding to Gα15 (43.5 KDa) and ß1 adrenergic receptor (51 kDa) were identified (Fig. 4A). We next evaluated whether GNA15 silencing altered βAR-mediated cell proliferation in KRJ-I cells. We focused on the proliferative effects of NNK and evaluated whether NNK-mediated signalling was affected by GNA15 silencing. NNK significantly (p b 0.05) activated ERK and AKT phosphorylation (and therefore signalling) in KRJ-I cells (Fig. 4B–C). Silencing GNA15 (sh66) reversed this effect which was also associated

with an elevation in NFκB signalling. Signalling was associated with changes in proliferation. Specifically, NNK significantly increased proliferation in KRJ-I cells (⁎p b 0.05 vs. unstimulated cells) (Fig. 4D). This effect was reversed by atenolol (#p b 0.05 vs. NNK stimulation). In GNA15 silenced cells, NNK-mediated proliferation was significantly lower (&p b 0.05 vs. NNK stimulation in non-silenced cells) (Fig. 4D). A similar reduction in effect was observed after NNK + atenolol stimulation (%p b 0.05 vs. NNK stimulation in GNA15 silenced cells). 3.4. Gα15, β1 adrenergic receptor function and cell-cycle inhibitor regulation We next evaluated expression of genes relevant to proliferation in the NNK-GNA15 model system. We focused on cyclin-dependent kinase inhibitor 2A (CDKN2A; P16INK4A) and cyclin-dependent kinase inhibitor

Fig. 5. Gα15 is involved in β1 adrenergic receptor regulation of cell-cycle inhibitors. P16INK4A, P21WAF1/CIP1, Proliferating Cell Nuclear Antigen (PCNA) and Ki67 RNA expression in control and GNA15 silenced cells after β1 adrenergic receptor stimulation with NNK alone or in combination with atenolol. U = unstimulated, S = NNK-stimulated, S + A = NNK + antagonist; +sH = + shRNA. ⁎p b 0.05 vs. control. %p b 0.05 vs. S (NNK stimulation). &p b 0.05 vs. NNK stimulation alone (black bars). $p b 0.05 vs. unstimulated (with shRNA). Dotted line = 1.0.

S. Zanini et al. / Cellular Signalling 27 (2015) 899–907

905

silenced cells). P21WAF1/CIP1 transcript levels were also increased in NNK- and atenolol-treated cells ($p b 0.05 vs. unstimulated GNA15 silenced cells) (Fig. 5). Transcripts of PCNA, which interacts with P21WAF1/CIP1 to regulate cell cycle progression [36], were decreased in GNA15-silenced cells, both in controls and when stimulated with NNK alone or with atenolol. (⁎p b 0.05 vs. control and $p b 0.05 vs. unstimulated GNA15-silenced cells) (Fig. 5). After NNK stimulation, transcript levels of Ki67 were increased in control cells (⁎p b 0.05 vs. unstimulated cells) (Fig. 5). These were significantly decreased after stimulation with NNK in combination with atenolol. No significant differences were noted in Ki67 in GNA15-silenced cells. 3.5. Gα15 expression in tumour samples

Fig. 6. GNA15 in normal and neoplastic tissues and examination of the correlation between GNA15 expression and other GEP-NEN biomarkers. A) Transcript levels of GNA15 in normal small intestinal mucosa (n = 13), primary small intestinal neoplasia (SI-NEN) (n = 11), and SI-NEN liver metastasis (n = 22). There is a significant difference in mRNA expression between normal mucosa and neoplastic tissues (primaries and metastasis: ⁎p b 0.05 vs. normal mucosa). B) The correlation between GNA15 expression and the Ki67 is shown. Levels of transcript are higher in low proliferating lesions (b1%) (⁎p b 0.05).

1 (CDKN1A; P21WAF1/CIP1) both negatively regulators of the cell cycle [35], and the positive regulators, Proliferating Cell Nuclear Antigen (PCNA) and Ki67. In non-silenced cells, P16INK4A and P21WAF1/CIP1 were inhibited by NNK (⁎p b 0.05) and were activated by atenolol (%p b 0.05 vs. NNK stimulation) (Fig. 5). In GNA15-silenced (sh66) cells, in contrast, transcript levels of both cell cycle inhibitors were increased after NNK stimulation ($p b 0.05 vs. unstimulated GNA15-

An analysis of primary SI-NENs and metastases demonstrated significantly elevated GNA15 mRNA levels (⁎p b 0.05) compared with normal mucosa (Fig. 6A). An analysis of clinical features indicated that GNA15 was expressed differentially between tumours with Ki67 b 1% and Ki67 N 1% (⁎p b 0.05) (Fig. 6B). Specifically, low grade tumours exhibited elevated GNA15 expression. The results were confirmed by immunohistochemical staining of normal mucosa, primary tumours, liver metastasis and xenograft tumours from SI-NEN cell lines (KRJ1-I and H-STS) xenografts (Fig. 7). Co-staining with CgA served as a neuroendocrine tumour marker [37]. 3.6. Quantitative Gα15 immunohistochemistry The immunohistochemical expression of Gα15 in SI-NENs was examined by immunohistochemistry and AQUA quantitation of staining

Fig. 7. Immunohistochemical staining for GNA15 expression. Levels in a normal mucosa, primary SI-NEN, SI-NEN metastasis, KRJ1 and H-STS cell xenografts injected are shown. Nuclei are shown with 40,6-diamidino-2-phenylindole staining (DAPI; blue) (left), GNA15 with cyanine 5 (red) and chromogranin A with Alexa Fluor 488 (green) (middle). On the right is shown the composite view (right). White arrow identifies a single normal EC cell with GNA15 — this was uncommon. In contrast, the primary tumour and xenografts exhibited high expression(dual staining with GNA15/CgA) while expression in the metastases was lower.

906

S. Zanini et al. / Cellular Signalling 27 (2015) 899–907

intensity of YTMA60_2 (Fig. 8A). Levels of Gα15 were significantly elevated in primary tumours compared to tumours with liver or lymph node metastases (⁎p b 0.01) (Fig. 8B). This is consistent with the results demonstrating that transcript levels are higher in SI-NEN primaries compared with metastases (Fig. 6A). A cut-off of 153.9 (AQUA score) corresponding to the median value Gα15 signal intensity in the cytokeratin mask area was used to divide low and high Gα15 expression. The median survival was 13.86 and 6.15 years respectively for the low and high Gα15 expression groups (log-rank p = 0.05; odds ratio 2.25, 95% CI: 1.73–2.78) (Fig. 8C). A similar trend was noted using a cut-off of 156.7 (equivalent to the median value Gα15 signal intensity in the cytokeratin mask area; log-rank p = 0.1, ratio 2.25, 95% CI: 1.77–2.74) (Fig. 8D). 4. Discussion GEP-NENs comprise a diverse group of lesions of different cell types that exhibit distinct biological and clinical behaviours. Despite this, they are largely treated as a uniform neoplastic entity with the resultant limitations in therapeutic efficacies [5]. The identification of differentially expressed factors that could better define the biological behaviour of NENs e.g., GPCR signalling, would provide a basis for the development of alternate therapies. We sought to identify and characterize the expression and functional activity of a novel GPCR signalling mediator to gain insights into the pathobiological mechanisms that regulate the protean biology of these neoplasms. Gα15, a Gαq protein, is a heterotrimeric G-protein linked to GPCRs that has unique functions in haematopoietic cells not performed by other Gq subfamily members [12]. Its expression is specifically decreased when cells lose the capacity to proliferate during granulocytic and erythroid differentiation [38]. It has also been detected in pre-B acute lymphoblastic leukaemia and in CD34-enriched peripheral blood stem cells (PBSC) [39]. Moreover, in contrast to other Gq proteins, Gα15 can be activated by a variety of receptors, which have very

different physiological functions [17]. It is likely that such a G-protein, which can non-selectively link different receptors to the same effector, could produce inappropriate or abnormal signalling. Therefore it is tempting to speculate that its ectopic expression under nonphysiological conditions could be pathological since it has been widely demonstrated that the aberrant signalling through GPCRs is associated with tumourigenesis [40]. The identification of its ectopic expression in SI-NENs may support such a role in neuroendocrine neoplasia. We identified that GNA15 was differentially overexpressed in GEP-NEN cell lines, particularly in tumours of small intestinal origin (N150-fold compared to non-transformed neuroendocrine cells). This suggested an important biological role in tumour function. We hypothesized that expression was linked to proliferation and apoptosis and investigated this in the KRJ-I cell line because these cells expressed the highest GNA15 transcript levels; they also represent a well-characterized small intestinal enterochromaffin cell neuroendocrine tumour model [23]. After silencing with lentiviral particles, the activity of pro-proliferative signalling as well as proliferation per se was significantly decreased while apoptosis was increased. This indicates that the expression of GNA15 may promote cellular proliferation and inhibit cellular apoptosis. ERK and Akt are major signalling pathways involved in directing neoplastic cellular proliferation [41,42] and recently their role has been demonstrated in SI-NEN chemoresistance [26]. Akt signalling is connected at multiple levels to the Ras/mitogen activated protein kinase (MAPK) pathway and both can cross-activate or inhibit each other [43, 44]. The reduction in the activation of Akt and ERK observed in the present study after silencing with lentiviral particles, sh66 and sh68, demonstrate that these pathway could be activated through Gα15. It should be noted that a decreased phosphorylation in p70S6K/p85S6K, two isoforms of the same kinase belonging to the same signalling pathway [45] and downstream of Akt [46], and considered a marker of Akt signalling was not observed. This suggests the principle effector is through ERK, well-known to be upregulated in SI-NENs [47,48].

Fig. 8. Quantitative assessment of GNA15 and relationship to survival. Expression levels of GNA15 determined by immunohistochemistry and AQUA quantitation on the NEN tissue microarray, YTMA60_2. A) Three-colour image of a small intestinal neoplasia demonstrating significant overlap between cytokeratin and cytoplasmic GNA15 staining in the tumour (inset). Blue, nuclei (DAPI); green, tumour mask (cytokeratin–Alexa488); red, GNA15 (Cy5). B) AQUA levels of GNA15 were significantly overexpressed in primary malignant SI-NEN neoplasia compared metastatic neoplasia. (⁎p b 0.01). C–D) Overall survival in patients with low and high Gα15 expressions. Median survival was used as cut-off to distinguish between the two groups. Analysis on all SI-NENs (8C), analysis on primary SI-NENs only (8D).

S. Zanini et al. / Cellular Signalling 27 (2015) 899–907

A number of GPCRs activate NFκB via Gi- or Gq-dependent pathways and an intricate signalling network exists between GNA15-coupled receptors and the IKB kinase (IKK)/NFkB pathway [49]. Furthermore, in HeLa cells transfected with kappaB-driven luciferase reporter the constitutive activation of Gα15 mediated the activation of NFκB signalling [50]. Moreover, NFκB appears to play a critical role in the survival and growth of neoplastic cell lines via cross-talk with other pathways including phosphatidylinositol 3-kinases/Akt [51]. In our experiments, GNA15 silencing increased NFκB activation which was amplified by βAR activation through NNK. We hypothesize that this phenomenon was a consequence of a compensatory GPCR-related mechanism associated with NFκB activation independent of Akt and ERK pathways amplifiable by βAR activity [52–54]. Previous studies demonstrated that KRJ-I cells respond to ß1 adrenergic receptor activation [21] and stimulation of βAR signalling is known to be involved in cancer development [55–57]. Consequently, we investigated the coupling of GNA15 with the ß1 adrenergic receptor and the functional effect on cellular proliferation after a stimulation with NNK, a high affinity ß1 adrenergic receptor agonist that is also cancer promoting [58,59]. In response to NNK stimulation, GNA15-silenced cells did not proliferate and a specific activation of cell cycle inhibitors, P16INK4A and P21WAF1/CIP1, was observed. This was coupled to amelioration of NNK-mediated pro-proliferative signalling. We concluded that the silencing of GNA15 could inhibit cellular proliferation in response to a cancer-promoting factor. Our in vitro findings were confirmed in human tissues where GNA15 was differentially overexpressed in primary and metastatic tissues compared with normal tissues (p b 0.05) or preparations of isolated nontransformed neuroendocrine cells (EC cells). Moreover, the quantitative immunohistochemical analysis of its expression identified that a higher Gα15 expression was predictive of a worse survival in SI-NENs (logrank p = 0.05, ratio 2.25, CI 95%, 1.73–2.78). Interestingly, the high expression of Gα15 was associated with a lower expression of Ki67. However, one limitation of this marker may be related to the potential intratumoural heterogeneity and the differential expression between metastasis and primary tumours [60]. Our data support the hypothesis that GNA15 may be expressed in a subtype of GEP-NENs and its higher expression in primary, particularly slower proliferating tumours, than metastatic tumours could represent an important step in neoplastic progression. 4.1. Conclusion GNA15 is highly expressed at the mRNA and protein levels in GEP-NEN cell lines. On this basis as well as following confirmation that the protein is detectable in clinical samples, we propose that the ectopic expression of this protein in non-haematopoietic system is pathologically significant. Although the precise functional role of GNA15 in haematopoietic lineages is not completely understood, we postulate that ectopic expression in neuroendocrine-derived tumours plays a role in proliferation and resistance to apoptosis. Moreover, its ability to transmit growth regulatory signals in the absence of intrinsic inhibitory modules (e.g. the resistance to β-arrestindependent desensitization [16]) makes this protein a potentially important GPCR signal mediator. This is particularly intriguing in neuroendocrine neoplasms, which originate from neuroendocrine cells whose primary characteristics are to synthesize and release hormones, functions which are typically regulated by GPCRs. As such, further investigation into the mechanism of regulation of GNA15 in

907

tumour cells could provide new insights into tumour pathobiology as well as opens new avenues for therapeutic strategies. Acknowledgements Cooperint Grant 2010, University of Verona SZ. References [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [23] [24] [25] [26] [27] [28] [29] [30] [31] [32] [33] [34] [35] [36] [37] [38] [39] [40] [41] [42] [43] [44] [45] [46] [47] [48] [49] [50] [51] [52] [53] [54] [55] [56] [57] [58] [59] [60]

I.M. Modlin, et al., Lancet Oncol. 9 (1) (2008) 61–72. A. Frilling, et al., Endocr. Relat. Cancer 19 (5) (2012) R163-R85. I.M. Modlin, K.D. Lye, M. Kidd, Cancer 97 (4) (2003) 934–959. G. Kloppel, A. Perren, P.U. Heitz, Ann. N. Y. Acad. Sci. 1014 (2004) 13–27. S. Schimmack, et al., Langenbecks Arch. Surg. 396 (3) (2011) 273–298. I. Drozdov, et al., PLoS One 6 (8) (2011) e22457. I.M. Modlin, et al., J. Clin. Endocrinol. Metab. 91 (6) (2006) 2340–2348. R.T. Dorsam, J.S. Gutkind, Nat. Rev. Cancer 7 (2) (2007) 79–94. J. Wu, et al., Cancer Genomics Proteomics 9 (1) (2012) 37–50. T.T. Amatruda III, et al., Proc. Natl. Acad. Sci. U. S. A. 88 (13) (1991) 5587–5591. T.M. Wilkie, et al., Proc. Natl. Acad. Sci. U. S. A. 88 (22) (1991) 10049–10053. F. Giannone, et al., J. Mol. Endocrinol. 44 (5) (2010) 259–269. Y. Su, M.K. Ho, Y.H. Wong, Neurosignals 17 (1) (2009) 71–81. E. Kostenis, Trends Pharmacol. Sci. 22 (11) (2001) 560–564. F. Giovinazzo, et al., Cell. Signal. 25 (3) (2013) 651–659. G. Innamorati, et al., Cell. Signal. 21 (7) (2009) 1135–1142. S. Offermanns, M.I. Simon, J. Biol. Chem. 270 (25) (1995) 15175–15180. X. Zhu, L. Birnbaumer, Proc. Natl. Acad. Sci. U. S. A. 93 (7) (1996) 2827–2831. H.M. Schuller, et al., Cancer Res. 59 (18) (1999) 4510–4515. A. Rivenson, et al., Cancer Res. 48 (23) (1988) 6912–6917. M. Kidd, et al., J. Mol. Endocrinol. 38 (1–2) (2007) 181–192. R. Pfragner, et al., Anticancer Res. 29 (6) (2009) 1951–1961. Z.L. Siddique, et al., Neuroendocrinology 89 (4) (2009) 458–470. D. Cecconi, et al., Proteomics 7 (10) (2007) 1644–1653. A.A. McLeod, D.G. Shand, Ann. Intern. Med. 96 (2) (1982) 244–245. B. Svejda, et al., Cancer 117 (18) (2011) 4141–4154. M. Kidd, et al., Cancer 103 (2) (2005) 229–236. M. Kidd, et al., Physiol. Genomics 30 (3) (2007) 363–370. A.V. Biankin, et al., Nature 491 (7424) (2012) 399–405. F. Fiory, et al., PLoS One 9 (12) (2014) e113655. http://dx.doi.org/10.1371/journal. pone.0113655.eCollection2014. M. Kidd, et al., Ann. Surg. 243 (2) (2006) 273–280. S. Schimmack, et al., Epigenetics Chromatin 7 (2014) 15. http://dx.doi.org/10.1186/ 1756-8935-7-15.eCollection2014. R.L. Camp, L.A. Charette, D.L. Rimm, Lab. Invest. 80 (12) (2000) 1943–1949. J.L. Gu, et al., J. Cell. Biochem. 88 (6) (2003) 1101–1111. S.C. Wong, et al., J. Pathol. 194 (1) (2001) 35–42. G. Maga, U. Hubscher, J. Cell Sci. 116 (Pt 15) (2003) 3051–3060. B. Lawrence, et al., Endocrinol. Metab. Clin. North Am. 40 (1) (2011) 111–134 (viii). S. Tenailleau, I. Corre, S. Hermouet, Exp. Hematol. 25 (9) (1997) 927–934. M. Pfeilstocker, et al., Leukemia 10 (7) (1996) 1117–1121. R. Lappano, M. Maggiolini, Nat. Rev. Drug Discov. 10 (1) (2011) 47–60. K. Sheppard, et al., Crit. Rev. Oncog. 17 (1) (2012) 69–95. L. Santarpia, S.M. Lippman, A.K. El-Naggar, Expert Opin. Ther. Targets 16 (1) (2012) 103–119. E. Aksamitiene, A. Kiyatkin, B.N. Kholodenko, Biochem. Soc. Trans. 40 (1) (2012) 139–146. A. von Kriegsheim, et al., Nat. Cell Biol. 11 (12) (2009) 1458–1464. S.C. Kozma, G. Thomas, Semin. Cancer Biol. 5 (4) (1994) 255–260. M.Y. Zhao, et al., Clin. Cancer Res. 15 (5) (2009) 1708–1720. A. Tannapfel, et al., Am. J. Clin. Pathol. 123 (2) (2005) 256–260. D. Karhoff, et al., Neuroendocrinology 85 (1) (2007) 45–53. A.M. Liu, Y.H. Wong, J. Biol. Chem. 279 (51) (2004) 53196–53204. M. Yang, et al., J. Immunol. 166 (11) (2001) 6885–6892. A.R. Hussain, et al., PLoS One 7 (6) (2012) e39945. T. Kwon, et al., J. Biol. Chem. 275 (1) (2000) 423–428. A. Gastonguay, et al., Cancer Biol. Ther. 13 (8) (2012) 647–656. T.M. Ahola, et al., Endocrinology 143 (12) (2002) 4620–4626. V.Y. Shin, et al., Toxicol. Sci. 96 (1) (2007) 21–29. P.G. Park, et al., Cancer Res. 55 (16) (1995) 3504–3508. S.W. Cole, A.K. Sood, Clin. Cancer Res. 18 (5) (2012) 1201–1206. W.K. Wu, et al., Cancer Res. 65 (12) (2005) 5272–5277. D.L. Weddle, et al., Carcinogenesis 22 (3) (2001) 473–479. E. Vilar, et al., Endocr. Relat. Cancer 14 (2) (2007) 221–232.