Background
Due to the high prevalence of colorectal cancer [
1], better insight into regulatory mechanisms involved in cell proliferation in this malignancy is needed, and might ultimately lead to improved treatment. Several receptors can mediate proliferogenic signals. Among these, G protein-coupled receptors (GPCRs) may induce mitogenic signalling and have a role in cancer, including colorectal and pancreatic cancer [
2‐
4]. Moreover, activation of GPCRs and receptor tyrosine kinases (RTKs) may act in concert to enhance cellular proliferation. Thus, an important question is how these signals are integrated in the cells.
GPCRs are heptahelical transmembrane receptors mediating their effects via heterotrimeric G proteins (of either the Gs, Gi, Gq, or G12/13 subtypes) [
5,
6]. While the role of Gs-coupled prostanoid receptors in colon cancer cell proliferation, apoptosis, and migration has been extensively studied [
4], there is less information on the role of Gq-coupled receptors in this malignancy. Stimulation of these receptors leads to activation of phospholipase Cβ (PLCβ) and thereby of protein kinase C (PKC), which may be involved in tumorigenesis [
7]. Elevated expression of PKC βII has been found to be an early promotive event in colon cancer development [
8], and inhibition of PKC β was found to decrease proliferation and induce apoptosis in colon carcinoma cells [
9].
Neurotensin is a peptide that binds to GPCRs. It is mainly formed in the central nervous system and by endocrine cells of the digestive tract, where it acts as a paracrine and endocrine modulator in a variety of gut functions, including vascular smooth muscle activity, gastrointestinal motility, gastric emptying, and intestinal, pancreatic, and biliary secretions [
10]. In addition, neurotensin stimulates growth of the intestinal mucosa under physiological and pathological conditions [
10,
11] and has been found to promote azoxymethane-induced colon carcinogenesis in rats and mice [
12,
13]. Neurotensin has also been implicated in the progression of cancers of the pancreas, breast, lung, and prostate [
10,
11,
14]. Three subtypes of neurotensin receptors have been cloned [
15]. The high affinity NTSR1 receptor and the low affinity NTSR2 receptor both belong to the GPCR family, while the NTSR3/sortilin receptor is a nonspecific receptor with a single transmembrane domain [
15,
16]. The pharmacological and signalling properties of the NTSR2 receptor, which exerts its effects mainly in the central nervous system, are incompletely understood, and appear to be dependent on cell type and species [
16]. The peripheral effects of neurotensin appear to be mediated largely by NTSR1, which activates PLCβ [
14,
16]. Experiments using a specific antagonist or knockdown of the NTSR1 using short interfering RNA suggest that NTSR1 mediates the effects of neurotensin on cancer cells, although NTSR3/sortilin, which is often coexpressed in cancer cells, may modulate NTSR1 signalling [
14,
16]. Splice variants of the NTSR1 were recently detected in prostate cancer cell lines, however, no functional studies of these have been conducted [
17]. Recent data have suggested that the NTSR1 receptor gene may be a downstream target of the extracellular signal-regulated kinase (ERK) and Tcf/β-catenin pathways [
18,
19], and increased expression of NTSR1 during progression of colon tumorigenesis has been reported [
20,
21].
Neurotensin has been found to stimulate proliferation of certain colon carcinoma cell lines [
10,
22]. Reports on intracellular signalling leading to proliferation induced by neurotensin in some other cell types have suggested the involvement of PKC-dependent activation of ERK and protein kinase D (PKD) [
10,
23‐
27], and either dependence or independence of epidermal growth factor receptor (EGFR) transactivation [
23‐
25]. In the pancreatic cancer cell line Panc-1, DNA synthesis induced by neurotensin was independent of EGFR transactivation [
23], whereas in the prostate cancer cell line PC-3, neurotensin stimulated mitogenesis by a PKC-dependent transactivation of EGFR [
24]. In colon carcinoma cell lines neurotensin has been found to activate ERK, as well as PKC, Akt, and nuclear factor κB (NF-κB) pathways [
28‐
31]. Furthermore, neurotensin induced phosphorylation and inactivation of glycogen synthase kinase (GSK), leading to cyclin D1 expression, through mechanisms that were at least partly dependent on PKC [
29]. Neurotensin has also been found to induce a proinflammatory tumour microenvironment and promote cancer cell invasion through pathways that involved NF-κB, PKC, ERK, and the sodium-proton exchanger 1 (NHE1) [
14,
31‐
33].
The aim of the present study was to investigate some of the intracellular signalling pathways involved in mitogenesis induced by neurotensin in human colorectal cancer cells, by examining the HCT116 and HT29 lines and comparing them with Panc-1 cells. The results suggested that while neurotensin acted predominantly through PKC in Panc-1 cells and via EGFR transactivation in HT29 cells, it used both these pathways in HCT116 cells. In the latter cells neurotensin-induced activation of ERK was mediated largely by PKC, while neurotensin-induced activation of Akt was independent of PKC but involved transactivation of the EGFR, apparently by a Ca2+-dependent mechanism. Neurotensin-induced DNA synthesis was mediated mainly by PKC.
Methods
Chemicals
Dulbecco's modified Eagle's medium, N-(2-hydroxyethyl)piperazine-N'-(2-ethanesulfonic acid (Hepes), penicillin and streptomycin were from Gibco (Grand Island, NY). Neurotensin, 12-O-tetradecanoylphorbol-13-acetate (TPA), thapsigargin, epidermal growth factor (EGF), and wortmannin were obtained from Sigma-Aldrich (St. Louis, MO). [2-[1-(3-dimetylaminopropyl)-1H-indol-3-yl]-maleimide] (GF109203X), 4-(3-chloroanilino)-6,7-dimethoxyquinazoline (tyrphostin AG1478), 2'-amino-3'-methoxyflavone (PD98059, and N-[(2R)-2 (hydroxamidocarbonylmethyl)-4-methylpentanoyl]-L-tryptophan methylamide (GM6001/Galardin) were from Calbiochem (San Diego, CA). 7-Methyl-2-(4-morpholinyl)-9-[1-(phenylamino)ethyl]-4H-pyrido[1,2-a] pyrimidin-4-one (TGX-221) was obtained from Cayman Chemical (Ann Arbor, MI). Transforming growth factor α (TGFα) was obtained from Bachem (Bubendorf, Switzerland). 4-Quinazolinamine, N-(3-chloro-4-fluorophenyl)-7-methoxy-6-[3-4-morpholin)propoxy] (gefitinib) was a gift from Astra Zeneca (Cheshire, UK), and cetuximab was kindly provided by Merck KgaA (Darmstadt, Germany). [6-3H]thymidine (20-30 Ci/mmol) and myo-[2-3H]inositol (15.0 Ci/mmol) were from Amersham Biosciences (Buckinghamshire, UK). Antibodies against phosphorylated AktSer473, total Akt, dually phosphorylated ERKThr202/Tyr204, phospho-EGF receptorTyr1173, and phospho-Shc Tyr239/240 were obtained from Cell Signaling Technology (Boston, MA). Anti-ERK and anti-Shc antibodies were obtained from Upstate (Billerica, MA). EGFR antibody (1005) was obtained from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA). Secondary antibodies were purchased from Bio-Rad Laboratories (Hercules, CA) and Licor Biosciences (Lincoln, NE). All other chemicals were of analytical quality. Stock solutions of test compounds were prepared in DMSO (TPA, thapsigargin, wortmannin, PD98059, GM6001, TGX221, gefitinib) or 0.9% NaCl (neurotensin, GF109203X). EGF was dissolved in 4 mM HCl, and TGFα in 4 mM HCl containing 1% bovine serum albumin from Sigma Aldrich (St. Louis, MO). Cetuximab was dissolved in phosphate-buffered saline (PBS). When solutions containing DMSO were used, the final concentration of DMSO was kept as low as possible.
Cell culture
Human colorectal cancer cell lines HCT116 and HT29, and pancreatic adenocarcinoma cell line Panc-1 were obtained from ATCC (Manassas, VA). The cells were maintained in Dulbecco's modified Eagle's medium containing 1 g/l glucose (or 4.5 g/l for Panc-1) supplemented with 10% horse serum (10% fetal bovine serum for Panc-1 cells), penicillin (67 μg/ml), streptomycin (100 μg/ml) and 2 mM glutamine (4 mM for Panc-1). Cells were plated onto Costar plastic culture wells (Corning Life Sciences, Acton, MA) at a density of 50 000 cells/cm2 (25 000 cells/cm2 in the case of Panc-1 cells) in serum-containing medium. The cultures were kept in 95% air/5% CO2 at 37°C. After 24 hours the medium was replaced with serum-free medium and the cells were cultured for 24 hours before stimulation with agonists.
Measurement of DNA synthesis
Neurotensin, TPA, and inhibitors of PKC and EGF receptor were added to serum-starved HCT116 cells as described in the figure legends, and [
3H]thymidine was added 12 hours after stimulation. Serum-starved HT29 and Panc-1 cells were stimulated for 21 hours with neurotensin and EGF before [
3H]thymidine was added. The cells were harvested after three hours pulsing with [
3H]thymidine, and DNA synthesis was measured as the amount of radioactivity incorporated into DNA as previously described [
34]. Briefly, medium was removed, and cells were washed twice with 0.9% NaCl. The cellular material was dissolved with 1.5 ml of 0.5 N NaOH for 3 hours at 37°C, collected, mixed with 1.5 ml H
2O, and precipitated with 0.75 ml 50% trichloroacetic acid (TCA). The acid-precipitable material was transferred to glass fiber filters (GF/C Whatman, GE Healthcare, UK) and washed twice with 5.0 ml 5% TCA, followed by liquid scintillation counting of the filters in a Packard Tri-Carb liquid scintillation counter.
Inositol phosphate accumulation
Cells were labelled with [3H]inositol, 2.5 μCi/ml for 24 hours in serum-free medium. Medium was removed 30 minutes before agonist stimulation and replaced with Krebs-Ringer-Hepes buffer pH 7.4, containing 10 mM glucose and 15 mM LiCI. HCT116 cells were stimulated with neurotensin for 30 minutes, and the reaction was stopped by removing buffer and adding 1 ml ice-cold 0.4 M perchloric acid. Samples were harvested and neutralized with 1.5 M KOH, 60 mM EDTA, 60 mM Hepes, in the presence of Universal indicator. The neutralized supernatants were applied on columns containing 1 ml Dowex AG 1-X8 resin (Bio-Rad Laboratories, Hercules, CA), and inositol phosphates were eluted with 10 ml 1 M ammonium formate/0.1 M formic acid.
Immunoblotting
Aliquots with ~30 000 cells (total cell lysate prepared in Laemmli or RIPA buffer) were electrophoresed on 6-12% (w/v) polyacrylamide gels. This was followed by protein electrotransfer to nitrocellulose membranes and immunoblotting with antibodies against phospho-Akt, total Akt, phospho-ERK1/2, total ERK, phospho-EGFR, total EGFR, phospho-Shc, and total Shc, respectively. Immunoreactive bands were visualized with enhanced chemiluminescence using LumiGLO (KPL Protein Research Products, Gaithersburg, MD), or infrared imaging using Odyssey Infrared Imaging System supplied by Licor Biosciences (Lincoln, NE), respectively.
Statistical analyses
Results are expressed as means ± standard error of the mean (S.E.M). DNA synthesis data were analyzed by one-way ANOVA, and post tests using Bonferroni correction to compare groups, using GraphPad Prism (version 5.01, GraphPad Software, San Diego, California, USA). Results were considered significant when p < 0.05.
Discussion
In the present study, we have found that neurotensin-induced signalling in colon carcinoma cells involves both EGFR-dependent and -independent pathways. In HCT116 cells, stimulation by neurotensin of ERK phosphorylation and DNA synthesis is mediated by PKC, whereas Akt phosphorylation induced by neurotensin is dependent on EGFR-mediated signalling.
In agreement with previous studies in human pancreatic cancer cells (Panc-1) [
23,
25] we found that neurotensin-induced ERK activation and DNA synthesis in the colon cancer cells HCT116 was mainly dependent on PKC and did not involve EGFR transactivation. Thus, the stimulatory effect of neurotensin and TPA on DNA synthesis was of the same magnitude, and stimulation of both DNA synthesis and ERK phosphorylation by neurotensin was inhibited by pretreatment with the PKC blocker GF109203X. Furthermore, while neurotensin stimulated Akt phosphorylation in an EGFR-dependent manner, TPA did not induce phosphorylation of Akt in HCT116 cells. In prostate cancer cells, neurotensin also stimulated ERK phosphorylation in a PKC-dependent manner, but in these cells activation of PKC mediated transactivation of the EGFR [
24].
We did not find that EGF stimulated DNA synthesis significantly in HCT116 cells. A plausible explanation is the autocrine production of TGFα and other ligands, leading to constitutive activation of EGFR in HCT116 cells [
47‐
49]. It was previously reported that while exogenous addition of EGF had no effect on DNA synthesis, due to the production of TGFα [
35], the EGFR was not saturated by the autocrine ligand(s) and could be further activated by exogenous EGF, resulting in integrin α
2 expression, cell adhesion, and micromotion [
37]. It is likely that basal DNA synthesis reflects the effect of this constitutive EGFR activation, consistent with the finding that inhibition of EGFR activity with gefitinib reduced both basal and neurotensin-stimulated DNA synthesis. However, neurotensin still enhanced DNA synthesis compared to its corresponding control.
While neurotensin-induced phosphorylation of ERK and stimulation of DNA synthesis in HCT116 cells were dependent on PKC, we found phosphorylation of Akt induced by neurotensin to be independent of PKC. Moreover, the lack of effect of TPA on phosphorylation of Akt further strengthens the notion that PKC is not involved in activation of Akt in HCT116 cells. Instead, neurotensin-induced phosphorylation of Akt was dependent on EGFR activation, and this effect was mimicked by elevation of intracellular Ca
2+ induced by thapsigargin. Our results thus strongly suggest that neurotensin-induced phosphorylation of ERK and Akt is mediated by different pathways. In contrast, phosphorylation of both ERK and Akt induced by neurotensin was mediated by PKC-dependent EGFR transactivation in prostate cancer cells [
24]. Furthermore, in HT29 cells, both ERK and Akt phosphorylation induced by neurotensin was abolished by pretreatment with gefitinib (Figure
4) or cetuximab (data not shown). These observations are in line with previous studies in HT29 cells, demonstrating that activation of PAR1 and PAR2 receptors led to transactivation of the EGFR through matrix metalloproteinase-dependent release of TGFα [
50,
51]. The different time course of ERK and Akt phosphorylation in HCT116 cells also supports the involvement of different pathways.
Conflicting results have been reported on the effect of neurotensin on EGFR phosphorylation in different cells [
23,
24]. Thus, while neurotensin did not induce transactivation of the EGFR in Panc-1 cells [
23], PKC-dependent transactivation of the EGFR mediated the mitogenic effect of neurotensin on prostate cancer cells [
24]. We found that neurotensin induced phosphorylation of the EGFR and the adaptor protein Shc in HCT116 cells, and that inhibiting the EGFR with cetuximab or gefitinib strongly reduced neurotensin-induced phosphorylation of Akt. These results strongly suggest that the EGFR is transactivated by neurotensin in HCT116 cells and that this transactivation is involved in mediating the Akt phosphorylation stimulated by neurotensin. Since the PI3K/Akt pathway is important in several regulations besides cell proliferation, such as promoting cell survival by enhancing resistance to apoptosis [
52‐
56], the EGFR-mediated activation by neurotensin may have significant roles in the malignant phenotype in these cells.
It is unclear why neurotensin activates different pathways in the different the cell lines. It is known that HCT116 and Panc-1 cells both harbour a
KRAS mutation, while HT29 cells have a mutant
BRAF. Furthermore, HT29 and HCT116 cells harbour mutations in the catalytic α polypeptide of phosphoinositide-3-kinase, (
PIK3CA), and HT29 cells also have mutated p53
http://www.sanger.ac.uk/genetics/CGP/CellLines/, [
57]. While it is known that mutations in
KRAS,
BRAF and
PIK3CA may determine the responsiveness to targeted therapies such as EGFR, MEK or mTOR inhibitors [
53,
57,
58], the consequences of these mutations for neurotensin signalling in the different cell lines are not obvious. Whereas we found that neurotensin treatment stimulated Akt phosphorylation in the three cell lines examined, an earlier report using NTSR1-transfected AV12 cells found that neurotensin inhibited basal and EGF- or insulin-stimulated Akt phosphorylation, which was ascribed to a negative regulation mediated through Gq [
59]. It has been found that classical PKC isoforms mediate feedback inhibition of EGFR transactivation by Gq-coupled receptor agonists [
60]. The degree of EGFR-induced transactivation involvement in signalling by neurotensin may thus depend on the strength of PKC-mediated feedback inhibition in different cells. In this context, it is of interest that HCT116 cells have a higher expression of the classical isoform PKCβII than HT29 cells [
61].
Interestingly, while the results showed that EGFR activation was required for neurotensin-stimulated phosphorylation of Akt, we did not obtain complete inhibition of this effect by pretreatment with neither GM6001, cetuximab or gefitinib. Contrary to this, Akt phosphorylation induced by direct activation of the EGFR by TGFα or EGF was completely suppressed by gefitinib or cetuximab. Also, neurotensin-stimulated Shc phosphorylation was completely suppressed. One possible explanation is that neurotensin also might induce release of ligands that activate ErbB3 or ErbB4 receptors. The HCT116 cells have been found to release several ligands that activate the ErbB receptor family [
62‐
64]. The lack of complete inhibition induced by GM6001 pretreatment could imply that EGFR transactivation could also be induced independently of ligand shedding by an intracellular calcium-mediated mechanism, possibly involving Pyk2 or Src [
43,
44]. Alternatively, neurotensin might induce transactivation of the insulin-like growth factor-1 receptor (IGF-1R), as observed in human colonic epithelial cells [
65]. Another possibility is that neurotensin induces Akt phosphorylation through activation of subtypes of PI3K that are directly activated by GPCRs [
66,
67]. In fact, HCT116 cells have been found to express PI3Kβ [
52], which is activated by GPCRs [
66]. TGX-221, an inhibitor of PI3Kβ [
46], did not affect neurotensin-stimulated Akt phosphorylation when used alone, but it further suppressed neurotensin-stimulated phosphorylation of Akt when combined with gefitinib (data not shown). Thus, it is possible that multiple pathways activated by neurotensin might converge on Akt phosphorylation in a partially redundant manner. In contrast, neurotensin-stimulated phosphorylation of Akt in Panc-1 cells was abolished by pretreatment with TGX-221, indicating involvement of PI3Kβ in this cell line. Although several mechanisms may thus be involved in mediating the effect of neurotensin on phosphorylation of Akt in HCT116 cells, our results suggest that ligand shedding, which may be dependent on Ca
2+ elevation, and the resulting activation of the EGFR is a main pathway.
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
KMM participated in the design of the study, carried out DNA synthesis and immunoblotting experiments, and drafted the manuscript. IHT, MA, and JØ carried out immunoblotting experiments and helped revise the manuscript. MD carried out DNA synthesis and inositol phosphate experiments. OD and TC conceived of the study, participated in the design of the study and helped revise the manuscript. DS conceived of the study, participated in the design of the study, carried out DNA synthesis and inositol phosphate experiments, and helped revise the manuscript. All authors read and approved of the final manuscript.