Volume 116, Issue 12 p. 2902-2912
Original Article
Free Access

The 5-HT2B receptor plays a key regulatory role in both neuroendocrine tumor cell proliferation and the modulation of the fibroblast component of the neoplastic microenvironment

Bernhard Svejda MD

Bernhard Svejda MD

Institute of Pathophysiology and Immunology, Centre for Molecular Medicine, Graz, Austria

Gastrointestinal Pathobiology Research Group, Yale University School of Medicine, New Haven, Connecticut

Search for more papers by this author
Mark Kidd PhD

Mark Kidd PhD

Gastrointestinal Pathobiology Research Group, Yale University School of Medicine, New Haven, Connecticut

Search for more papers by this author
Francesco Giovinazzo MD

Francesco Giovinazzo MD

Gastrointestinal Pathobiology Research Group, Yale University School of Medicine, New Haven, Connecticut

Search for more papers by this author
Karim Eltawil MD

Karim Eltawil MD

Gastrointestinal Pathobiology Research Group, Yale University School of Medicine, New Haven, Connecticut

Search for more papers by this author
Bjorn I Gustafsson MD, PhD

Bjorn I Gustafsson MD, PhD

Gastrointestinal Pathobiology Research Group, Yale University School of Medicine, New Haven, Connecticut

Department of Internal Medicine, St. Olav's Hospital HF, Trondheim, Norway

Search for more papers by this author
Roswitha Pfragner PhD

Roswitha Pfragner PhD

Institute of Pathophysiology and Immunology, Centre for Molecular Medicine, Graz, Austria

Search for more papers by this author
Irvin M. Modlin MD, PhD, DSc

Corresponding Author

Irvin M. Modlin MD, PhD, DSc

Gastrointestinal Pathobiology Research Group, Yale University School of Medicine, New Haven, Connecticut

Fax: (203) 737-4067

Yale University School of Medicine, 333 Cedar Street, P.O. Box 208062, New Haven, CT 06520-8062===Search for more papers by this author
First published: 11 June 2010
Citations: 52



Fibrosis is a cardinal feature of small intestinal neuroendocrine tumors (SI-NETs) both in local peritumoral tissue and systemic sites (cardiac). 5-HT, a commonly secreted NET amine, is a known inducer of fibrosis, although the mechanistic basis for it and growth factors regulating fibrosis and proliferation in the tumor microenvironment are unclear. We hypothesized that targeting 5-HT2B receptors on tumor cells would inhibit SI-NET 5-HT release and, thereby, fibroblast activation in the tumor microenvironment.


We studied the 5-HT2B receptor antagonist PRX-08066 in NET cell lines (KRJ-I, H720) and in the coculture system (KRJ-I cells: fibroblastic HEK293 cells) using real time polymerase chain reaction, ELISA, Ki67 immunostaining, and flow cytometry-based caspase 3 assays to assess antiproliferative and profibrotic signaling pathways.


In the 5-HT2B expressing SI-NET cell line, KRJ-I, PRX-08066 inhibited proliferation (IC50 4.6x10−9M) and 5-HT secretion (6.9 × 10−9M) and decreased ERK1/2 phosphorylation and profibrotic growth factor synthesis and secretion (transforming growth factor beta 1 [TGFβ1], connective tissue growth factor [CTGF] and fibroblast growth factor [FGF2]). In the KRJ-I:HEK293 coculture system, PRX-08066 significantly decreased 5-HT release (>60%), Ki67 (transcript and immunostaining: 20%-80%), TGFβ1, CTGF, and FGF2 transcription (20%-50%) in the KRJ-I cell line. 5-HT itself stimulated HEK293 proliferation (25%) and synthesis of TGFβ1, CTGF and FGF2. PRX-08066 inhibition of KRJ-I function reversed these effects in the coculture system.


Targeting the 5-HT2B receptor may be an effective antiproliferative and antifibrotic strategy for SI-NETs because it inhibits tumor microenvironment fibroblasts as well as NET cells. Fibrosis and proliferation appear to be biologically interfaced neuroendocrine neoplasia domains. Cancer 2010. © 2010 American Cancer Society.

Small intestinal neuroendocrine tumors (SI-NETs) are cancers originating from serotonin-producing enterochromaffin (EC) cells in the diffuse neuroendocrine system,1, 2 and the carcinoid syndrome reflects excessive serotonin (5-HT) release.2 Carcinoid syndrome symptomatology includes diarrhea, flushing, bronchoconstriction, and fibrosis in the local peritumoral tissue (mesentery) and distantly in the heart or lungs.3-5

5-HT exhibits both mitogenic and fibrogenic effects as noted in fibroblasts, smooth muscle cells, and endothelial cells.6-8 This effect is mediated via G-protein coupled 5-HT receptors, which activate mitogenic pathways through the extracellular signal-regulated kinase (ERK) pathway and JNK activation.9, 10 Previously, we have demonstrated that 5-HT stimulated proliferation (EC50 = 25 nM) of the neoplastic SI-NET (EC cell-derived) cell line, KRJ-I.11 This process was transduced via phosphorylation of ERK, activation of the mitogen-activated protein kinase (MAPK) pathway, and transcription of c-Jun and Ki67, indicating functional 5-HT signaling coupled to proliferation. Other studies have noted that 5-HT modulates valvular subendocardial cell proliferation; human heart valves express messenger ribonucleic acid (mRNA) for 5-HT1B, 1D, 2A, and 2B receptors,12 and 5-HT agonists (fenfluramine, dexfenfluramine, pergolide, cabergoline, ergotamine) are associated with pulmonary fibrosis and valvular heart disease.13, 14 Considerable evidence, therefore, exists that 5-HT2B receptors are involved in cellular pathways that culminate in fibrosis.

Information regarding the structurally complex peritumoral microenvironment and its role in regulating tumor cell function is sparse.15 The tumor microenvironment comprises an admixture of endothelial cells, inflammatory cells, and fibroblasts whose interaction culminates in tumor growth, progression, and the development of fibrosis.16, 17 A series of key events include the production of growth factors, chemokines, and extracellular matrix proteins by a subpopulation of fibroblasts, the cancer-associated fibroblasts (CAFs). In cell cultures, phenotypic features of CAFs can be replicated by TGFβ stimulation, suggesting that this growth factor is critical for the regulation (modulation) of cell function in the tumor microenvironment.18 TGFβ, CTGF, and FGF2 are also synthesized and secreted by NET cells and are considered to modify the behavior of local fibroblasts and endothelial cells in the carcinoid tumor microenvironment and distantly in the endocardium.5, 19

The interaction of tumor cells and fibroblasts has not previously been evaluated in NETs because no appropriate models existed. In other cancers, in vitro models have been used to study cell migration and intercellular cross talk between different cell types,20-23 while in vivo experiments use a suspension of CAFs and tumor cells injected into mice to evaluate enhanced tumor progression.15 We hypothesized that targeting 5-HT2B receptors would inhibit SI-NET 5-HT release and, thereby, fibroblast activation in the tumor microenvironment. Accordingly, we investigated whether targeting the 5-HT2B receptor expressed on KRJ-I cells inhibited cell proliferation and 5-HT production and then the role of 5-HT in the production and secretion of profibrotic factors particularly as a paracrine effector in a NET:fibroblast coculture model system.


Experimental Approach

Studies were performed using normal small intestinal mucosa, small intestinal NETs, normal EC cells, KRJ-I11, 24-26 (SI-NET), NCI-H72027 (atypical bronchopulmonary NET, originally obtained from ATCC), and the human embryonic kidney cell line HEK293 (originally obtained from ATCC).28 The receptor profile of tissues and cell lines were determined using real time polymerase chain reaction (RT-PCR). Effects of the specific 5-HT2B receptor antagonist PRX-08066, originally developed for pulmonary hypertension, were examined on NET cell viability (3-(4,5-Dimethylthiazol-2-Yl)-2,5-diphenyltetrazolium bromide uptake29) and on basal and stimulated (isoproterenol, a β-adrenergic agonist30: 10−8M) 5-HT release using ELISA.11, 30 The antiproliferative effects of PRX-08066 were confirmed by ERK1/2 phosphorylation (ELISA)11 and Ki67 and caspase 3 transcripts quantified using RT-PCR.11, 31, 32 Secretion and transcript levels of profibrotic/angiogenetic factors TGFβ1, CTGF, and FGF2 were determined in response to PRX-08066 and in the coculture model (KRJ-I:HEK293). Activity of caspase 3 and alterations in live/dead cells in the coculture system was analyzed by the PhiPhiLux assay, propidium iodide uptake,26 and flow cytometry. In separate studies, the effects of 5-HT alone on HEK293 proliferation and fibrogenic factor synthesis and secretion, was studied.

Normal EC cell isolation

EC cells were isolated as described, approximately 1 × 106 cells were obtained/sample.30

Cell culture conditions

KRJ-I and NCI-H720 were cultured as described.11, 24-26 HEK293 was maintained in Dulbecco modified Eagle medium (DMEM), 10% fetal bovine serum and antibiotics.28

Coculture of cell lines

HEK293 cells were cultured in a 6-well plate (Falcon, BD, Franklin Lakes, NJ; 4 × 105 cells/mL) for 24 hours. KRJ-I cells (4 × 105 cells/mL) were added using coculture filter well inserts (Falcon, BD) with 1 μm pore size diameter. A 1:1 dilution of Ham's F12/M199 and DMEM medium supplemented with 10% FBS was used for normal growth condition for both cell lines. After 24 hours, the 5-HT2B antagonist was added (10−9 M) to the filter well insert. Cells and medium were harvested after 24 hours for ELISA and RT-PCR.


PRX-08066 (5-HT2B antagonist) was a gift of EPIX Pharmaceutical (Lexington, Mass).

Proliferation Studies

Five × 105 cells/mL, seeded in 96-well plates at 100 μL (4 plates/experimental condition) were stimulated with PRX-08066 (10−7M to 10−11M: n = 6 wells/concentration).25, 26 After 24 hours, mitochondrial activity was measured after adding MTT (3-[4,5-dimethylthiazol-2-ly]-2.5-diphenyltetrazolium bromide: 0.5mg/mL per well) for 3 hours. The optical density was read photospectrometrically at 595 nm using a microplate reader (Bio-Rad 3500, Hercules, Calif).29 Results were normalized to control (unstimulated cells) and the effective half-maximal concentrations calculated.

ERK1/2 signaling pathway analysis

After 2 hours incubation, KRJ-I cells and H720 cells were stimulated with PRX-08066, (10−9M) for 60 minutes and MAPK (ERK) signal activity determined using SuperArray CASE ELISA kits (SABiosciences, Frederick, Md; ERK—FE-002) as described.11

Ribonucleic Acid (RNA) isolation and Reverse Transcription

RNA was extracted from each tissue (30-50 mg), EC cell preparation or cell line (1 × 106, n = 4) using TRIZOL (Invitrogen, Carlsbad, Calif) and cleaned (Qiagen RNeasy kit and DNeasy Tissue kit, Qiagen Inc., Valencia, Calif). After conversion to cDNA (High Capacity cDNA Archive Kit, Applied Biosystems, Foster City, Calif),11, 32 RT-PCR analyses were performed using Assays-on Demand and the ABI 7900 Sequence Detection System.11, 32 All primer/probe sets were obtained from Applied Biosystems. The primers chosen were designed to encompass exon:exon boundaries and, therefore, preferentially amplified cDNA which minimized the possibility of genomic DNA contamination. Non-RT controls were included as negative controls for each of the samples. In preliminary analyses, we ran the PCR mix on gels and confirmed presence of single bands for each primer set used. RT-PCR data was normalized using the ΔΔCT method and the housekeeping gene ALG9 (for KRJ-I, EC cells and SI-NET samples) and β-actin (for HEK cells). The 5-HT2B receptor profile was analyzed in all samples, while profibrotic and cell cycle markers, TGFβ1, CTGF, FGF2, TPH1, caspase 3, and Ki67 transcripts, were measured after 24 hours of incubation with 5-HT2B antagonist (10−9M) in KRJ-I cells cultured for 48 hours.

5HT, TGFβ1, CTGF, and FGF2 secretion

Basal levels of these agents were measured using commercially available ELISA assays (5-HT: BA 10-0900; Rocky Mountain Diagnostics, Colorado Springs, Colo; TGFβ1: DB100B; R&D Systems, Minneapolis, Minn; FGFβ1: DFB50; R&D Systems)11, 30 and in-house assay (CTGF).19 Briefly, cells were seeded into 96-well plates (2 × 105 cells/mL) and agents levels were measured after 1 hour, 2 hours, and 24 hours.

Ki67 immunostaining

KRJ-I cells were immunostained with rabbit anti-Ki67 (Novocastra Laboratories, Bannockburn, Ill) as described using 2 color (nuclear stain, DAPI, blue/green, flourescein isothiocyanate (FITC) anti-Ki67).33 Cells were counted in 10 high-power field (HPF)/section and quantified as percentage of positive cells.34

Caspase 3 pathway activation and Propidium iodide uptake analyzed by FCM

Caspase 3 activity was evaluated in cocultured cells using the PhiPhiLux G1D2 kit (MBL, Nagoya, Japan) as described.35, 36 Briefly, cells were incubated with 75 μL of PhiPhiLux G1D2 for 1 hour at 37°C in 5% CO2. Before the measurements of caspase-3 activity (in the FITC channel), cells were costained with propidium iodide (PI) (5 μg/mL). During FCM, 4 gates were used including PI-positive/FITC negative (dead cells), PI-negative/FITC-negative (live cells, no caspase activity), PI-negative/FITC positive (live cells, undergoing caspase activation) and PI-positive/FITC positive (dead cells with caspase activity). Populations in each gate were quantified using flow cytometry as described.25, 35, 36

Statistical Evaluation

All statistical analyses were performed using Microsoft Excel and Prism 4 (GraphPad Software, San Diego, Calif). Sigmoidal dose responses and nonlinear regression analyses were calculated to identify half-maximal inhibitory (IC50) concentrations. Alterations in signal transduction, transcriptional activation, growth factor release, caspase 3 pathway activation, and dead/alive cells were assessed using 2-tailed, t test for paired data.


5-HT2B Receptor Profile

5-HT2B receptor levels were analyzed in normal SI mucosa, SI-NETs, isolated EC cells, and in KRJ-I, NCI-H720, and HEK293 using RT-PCR. 5-HT2B mRNA was identified in normal mucosa and in all 12 SI-NETs. Although KRJ-I cells, as well as SI-NETs expressed significant (P < .05) levels of the 5-HT2B receptor, no transcripts were identified in normal EC cells, in H720 cells, or in HEK293 cells (Fig. 1). KRJ-I cells, expressing the receptor, were used as a model to investigate effects of PRX-08066 on cell function. H720 cells were used as a tumor control.

Details are in the caption following the image

Transcript expression of the 5-HT2B receptor in SI mucosa (n = 6), SI-NETs (n = 12), isolated normal EC cells (n = 8), KRJ-I (n = 5), NCI-H720 (n = 5), and HEK293 (n = 5) by RT-PCR is depicted. Receptor expression was normalized to ALG9 or beta-Actin (HEK), using the ΔΔCT method. Transcripts were identified in normal mucosa, SI-NETs, and KRJ-I. No transcripts were identified in normal EC cells, NCI-H720 cells, or in HEK293 cells. SI-NETs indicates small intestinal neuroendocrine tumors; RT-PCR, real time polymerase chain reaction; Mean ± standard error of the mean (SEM). *P < .0001 versus EC cells. #P < .05 versus NCI-H720 cells.

Effects in NET Cell Lines

Anti-proliferative effects of 5-HT2B antagonist PRX-08066

Cell proliferation was inhibited in KRJ-I cells (IC50 = 4.6 × 10−10M, P < .05) with a maximum of 20 ± 5% (Fig. 2A). NCI-H720 cell proliferation was not reduced, consistent with the absence of 5-HT2B receptors on this cell line.

Details are in the caption following the image

Effect of PRX-08066 on (A) proliferation, (B) basal, and (C) isoproterenol (10 nM)-stimulated 5-HT secretion in NCI-H720 and KRJ-I measured by MTT uptake and ELISA is depicted. The IC50 for proliferative inhibition of KRJ-I was 4.6 × 10−10M with a maximal effect at 10−7M (A). No effect was observed in NCI-H720 cells. Mean ± SEM; n = 5 separate experiments. *P < .05 versus NCI-H720 cells.

The IC50 for basal secretion in KRJ-I was 6.9 × 10−9M with a maximal inhibitory effect at 10−7M (B). The IC50 for isoproterenol-stimulated secretion in KRJ-I cells was 1.2 × 10−9M with a maximal inhibitory effect at 10−8M (C). No effect was observed in NCI-H720 cells. MTT indicates Mean ± SEM; n = 5. *P < .05 versus NCI-H720 cells.

Effects of PRX-08066 on basal and isoproterenol-stimulated 5-HT secretion

Basal 5-HT secretion was dose-dependently inhibited (IC50 = 6.9 × 10−9M, ICmax: 30 ± 3.4%: 10−7M, P < .05) in KRJ-I cells (Fig. 2B). Isoproterenol-stimulated 5-HT release was inhibited with an IC50 = 1.25 × 10−9M and a maximum inhibition of 60 ± 4.7% (10−8M, P < .05) (Fig. 2C). No effects were observed in NCI-H720 cells.

Effects of PRX-08066 on signaling pathways

ERK phosphorylation was significantly inhibited (P < .05) in KRJ-I cells at a PRX-08066 concentration of 5 × 10−10M, while in NCI-H720, no effect was observed (Fig. 3).

Details are in the caption following the image

Effect of PRX-08066 on ERK phosphorylation in NCI-H720 and KRJ-I cells is depicted. PRX-08066 (5 × 10−10M) significantly (P = .0006) inhibited ERK phosphorylation (∼10%). No effect was observed in NCI-H720 cells. ERK indicates extracellular signal-regulated kinase. Mean ± SEM; n = 5. *P < .05 versus NCI-H720 cells.

Effects of PRX-08066 on transcription and secretion of profibrotic factors

Analysis of transcription and secretion of profibrotic factors in KRJ-I and NCI-H720 cells was examined at 60 minutes and 120 minutes post-PRX-08066 administration (IC50 dose ∼ 1 nM). In KRJ-I cells, TGFβ1 transcription and secretion was significantly (P < .05) decreased at 60 minutes (32 ± 22%, 44 ± 21% decrease) and at 120 minutes (74 ± 22%, 71 ± 25% decrease) (Fig. 4A, Fig. 4D). CTGF transcription and secretion was significantly inhibited at 120 minutes (63 ± 25%, P < .05, 71 ± 26%, P < .01 decrease) (Fig. 4B, Fig. 4E). FGF2 transcription was significantly inhibited at 120 minutes (67 ± 26%, P < .05), while secretion was inhibited at both 60 minutes (60 ± 28%, P < .01) and 120 minutes (95 ± 25%, P < .01) (Fig. 4C, 4F). No effects were noted on NCI-H720 cells.

Details are in the caption following the image

Effect of PRX-08066 on TGFβ1 (A), CTGF (B), and FGF2 (C) transcription at 1 hour and 2 hours in NCI-H720 and KRJ-I is depicted. Inhibition of TGFβ1 transcripts was observed at both 1 and 2 hours, while inhibition of CTGF and FGF2 was noted at 2 hours in KRJ-I. No effect was observed in NCI-H720 cells. Mean ± SEM; n = 5. *P < .05 versus H720 cells, **P < .01 versus NCI-H720 cells. Effect of PRX-08066 on TGFβ1 (D), CTGF (E), and FGF2 (F) secretion at 1 hour and 2 hours in NCI-H720 and KRJ-I. Inhibition of both TGFβ1 and FGF2 secretion was observed at both 1 and 2 hours, while inhibition of CTGF was noted at 2 hours in KRJ-I. No effect was observed in NCI-H720 cells. Mean ± SEM; n = 5. *P < .05 versus H720 cells, **P < .01 versus NCI-H720 cells. H = NCI-H70 cells, K = KRJ-I.

Effects of PRX-08066 on HEK293 cells

MTT uptake and expression of profibrotic and cell cycle factors were investigated in HEK293 cells treated with the 5-HT2B antagonist (10−9M) for 24 hours. No differences in MTT uptake or transcription levels of TGFβ1, CTGF, FGF2, TPH1, caspase 3 or Ki67 were identified (data not shown) consistent with the absence of 5-HT2B on this cell line.

Effects of 5-HT on HEK293 cells

MTT uptake and expression of pro-fibrotic and cell cycle factors were investigated in HEK293 cells treated with the 5-HT (10−6 to 10−11M) for 24 hours. 5-HT dose-dependently and significantly (P < .05) increased MTT uptake with an EC50 = 10−9M (Fig. 5A). Coincubation with the 5-HT2A/C receptor antagonist, ketanserin (10 μM), reversed these effects. Furthermore, 5-HT (10 μM) significantly (P < .05) stimulated transcription levels of Ki67, TGFβ1, CTGF, and FGF2 (Fig. 5B-E) consistent with a stimulatory effect on fibroblast proliferation and production of profibrotic mediators.

Details are in the caption following the image

Effect of 5-HT on HEK293 proliferation (A) and transcription of Ki67 and fibrogenic factors (B-E) is depicted. The EC50 for proliferation of HEK293 cells was 4.6 × 10−10M with a maximal effect at 10−7M (A). The 5-HT2A/C receptor antagonist, ketanserin (10μM), reversed this effect. 5-HT (10−8M) stimulated Ki67, TGFβ1, CTGF, and FGF2 transcription. Mean ± SEM; n = 4. *P < .05 versus control. #P < .05 versus 5-HT. CO = control.

Effects of PRX-08066 on KRJ-I and HEK293 cells in a coculture system

Basal 5-HT release

The KRJ-I and HEK293 cell lines were cultured in a coculture system to examine cross talk between cells (Fig. 6A-B). As expected, a significant inhibitory effect (P < .05) on 5-HT release was noted when KRJ-I cells were incubated with PRX-08066 (10−9M) for 24 hours (Fig. 6C-D). In contrast, isoproterenol (10−8M) stimulated (P < .05) 5-HT release.

Details are in the caption following the image

KRJ-I cells and HEK293 cells established in a coculture system (A) and systematic overview of the coculture model (B) are illustrated. To study intercellular crosstalk, nonadherent KRJ-I cells were maintained in the upper compartment separated from adherent HEK293 by a filter well insert (1 μm pore size). The specific 5-HT2B receptor antagonist was added to the tumor cell-containing compartment. Inhibition of 5-HT release was observed in KRJ-I when PRX-08066 (10−9M) was added (C). No effect on HT-secretion was noted when PRX-08066 was added to the HEK293 cells in the coculture system (D). Isoproterenol (10−8M) stimulated only 5-HT when added to the KRJ-I compartment. CO = control, ISO = isoproterenol. Mean ± SEM; n = 4. *P < .05.

Effects of PRX-08066 on transcript and protein levels of Ki67 and transcription of caspase-3 and profibrotic markers

In KRJ-I cells, transcripts for Ki67 were significantly decreased (84 ± 3.2% P < .01) as was Ki67 protein (36.8 ± 13.1% vs 70.1 ± 18.6% in untreated cells, P < .05). These decreases in Ki67, were associated with an increase in caspase 3 transcript levels (25 ± 19%, P < .05) (Fig. 7A-B). TGFβ1 and FGF2 transcripts were also decreased (51 ± 28%, 43 ± 28%, both P < .05) compared with controls (Fig. 7C-D), and a significant decrease in TPH1 was observed P < .05 (Fig. 7E). In HEK293 cells, Ki67 was significantly decreased (92 ± 28%, P < .05) but no differences in caspase 3 was identified (Fig. 7F-G). Both profibrotic/angiogenic factors TGFβ1 and FGF2 were significantly decreased (78 ± 2%, P < .05, 70 ± 35%, P < .01) compared with untreated cocultured HEK293 cells (Fig. 7H-I) and to PRX-08066-treated HEK293 cells that were cultured in upper chambers in the absence of KRJ-I cells. Both untreated coculture cells as well as HEK293 cells maintained in the absence of KRJ-I cells expressed similar transcript levels of Ki67, TGFβ1, FGF2, and CTGF.

Details are in the caption following the image

Effect of PRX-08066 on Ki67 (A, F), Caspase 3 (B, G), TGFβ1 (C, H), FGF2 (D, I), and TPH1 (E) transcripts after 24 hours in KRJ-I coculture d with HEK293 is depicted. In both cell lines, a significant decrease in Ki67 transcript and profibrotic/angiogenetic factors FGF2 and TGFβ1 was observed. In addition, a significant increase in caspase 3 levels and a decrease in TPH1 expression level was noted in KRJ-I. Mean ± SEM; n = 4. *P < .05, **P < .01. K = KRJ-I, H = HEK293 cells.

Effects of PRX-08066 on caspase 3 mediated apoptosis and cell viability

In KRJ-I cells, PRX-08066 significantly increased the number of dead cells (34 ± 27%, P < .05) compared with untreated controls (Fig. 8C). Although no increase in caspase 3 activity was observed (Fig. 8B), a significant increase in dead/caspase 3 positive cells (28 ± 11%, P < .05) and a reduction in viable cells (98 ± 0.1%, P < .01) were noted (Fig. 8A-D). In HEK293 cells, PRX-08066 caused a significant increase in dead/caspase 3 positive cells (76 ± 20%, P < .01) and caspase 3 activity (52 ± 34%, P < .01) (Fig. 8E-F). In addition, a significant reduction of viable cells (97 ± 1.5%, P < .05) was identified (Fig. 8H).

Details are in the caption following the image

Effects of PRX-08066 on caspase 3 activity and cell death in KRJ-I and HEK293 in a coculture model are depicted. In KRJ-I treated with PRX-08066 (10−9M) for 24 hours, a significant increase in dead cells (C) was noted, associated with a decreasing proportion of viable cells. In coculture d HEK293 cells, a significant increase in the proportion of caspase 3-positive cells (F) and decreased live cells were observed (H). Mean ± SEM; n = 8. *P < .05, **P < .01. K = KRJ-I, H = HEK293 cells.


Normal EC cells do not express the 5-HT2B receptor and its expression in the neoplastic phenotype suggests gain of function and, potentially, a regulatory role in SI-NET interaction with the tumor microenvironment. This phenomenon is tumor-specific; 100% of EC cell-derived SI-NETs in this study expressed this receptor, while the bronchopulmonary NET cell line investigated did not. This selective expression represents not only a unique tumor-related biological event but also that individual SI-NETs may be amenable to selective identification and targeting of the 5-HT2B receptor with PRX-08066. Because 5-HT exhibits an autocrine proliferative role in 5-HT secreting neoplastic NETs,26 targeting these cells with a specific 5-HT2B receptor antagonist provides a novel therapeutic option.

In previous studies, the proliferative activity of 5-HT in KRJ-I and NCI-720 was shown to be dependent on the expression of 5-HT2 receptor subtypes (at least 5-HT2A/C) and was concentration-dependent with EC50 in the nM range.25 Similar proliferation effects have been noted in the 5-HT secreting prostate cancer cell line PC3 (10 nM: 5-HT1A/B and 5-HT2A receptors),37-39 the 5-HT2A expressing breast cancer cell line MCF-740, and in human choriocarcinoma cell lines JEG-3 and BeWO (both 5-HT2A).41 In the current study, we identified a selective mitogenic role for 5-HT2B in KRJ-I cells.

Cell proliferation was significantly inhibited by PRX-08066 in KRJ-I, while NCI-H720 cell viability was not impaired, consistent with absence of the 5-HT2B receptor in the lung NET. In addition, both basal and isoproterenol-stimulated 5-HT release was significantly reduced by the antagonist in KRJ-I cells. This demonstrates that the 5-HT2B receptor is functionally coupled to both proliferative and secretory pathways in this cell line.

To further investigate signal transduction pathways involved in the antiproliferative effect of 5-HT2B receptor antagonist, we investigated phosphorylation of ERK. Recent studies have demonstrated a direct role of 5-HT receptor subtypes in vascular and tracheal smooth muscle cell proliferation. The mechanism is via coupling of 5-HT2A receptors and the ERK pathway, while 5-HT2B receptors activate ERK through the RAS pathway.42, 43, 44 In the current study, ERK1 of 2 phosphorylation was significantly decreased by PRX-08066 in KRJ-I cells. Our findings indicate that the antiproliferative effect of the 5-HT2B antagonist is consistent with inhibition of an ERK-mediated proliferative event. In contrast, no decrease in ERK phosphorylation was noted in the 5-HT2B receptor negative cell line NCl-720. This accentuates the necessity of defining individualized therapeutic strategies based on an individual tumor's receptor profile.25

Escape from TGFβ1 mediated growth control occurs in SI-NETs and is associated with increased ERK phosphorylation and signaling.45 Agents that target the ERK pathway may, therefore, provide an additional strategy for reversing activation of the autocrine TGFβ1 proliferation pathway in SI-NETs. The current study supports this proposal because targeting 5-HT2B receptors, in addition to inhibiting ERK phosphorylation, also resulted in a significant decrease in TGFβ1 transcription and secretion.

For almost a century, it has been recognized that SI-NETs often present with fibrosis in the peritumoral tissue, the adjacent mesentery and peritoneum, as well as in the right side of the heart or lungs.46 We hypothesized that fibroblasts in the tumor microenvironment play an important role in regulating NET-mediated fibrosis and metastasis. Numerous studies attest to the key role of the microenvironment and have established it is an integral part of tumorigenesis through the production of mitogenic, angiogenetic, and proteolytic factors that affect the phenotype of all cells in the vicinity.47 TGβ1 and CTGF are well-characterized profibrotic factors and elevated serum levels have been related to fibrosis in SI-NETs.19 FGF2 is a well-known stimulator of vascular endothelial cells48, 49 and plays a crucial role in tumor angiogenesis.50, 51 5-HT has previously been suggested as an important factor that induces fibrosis.5, 52 In our study, transcript levels and secretion of TGβ1, CTGF, and FGF2 were significantly reduced in KRJ-I cells treated with PRX-08066. 5-HT directly stimulated HEK293 MTT uptake and Ki67 transcription as well as transcription of TGFβ1, CTGF, and FGF2. These effects were reversed by ketanserin, a 5-HT2A/C antagonist, and demonstrates that proliferation and fibrogenic activity of this fibroblast cell line is directly activated by 5-HT. These data suggest an important role for the 5-HT2B receptor and 5-HT produced by tumor cells itself in the regulation of peritumoral fibrosis and angiogenesis. Thus, targeting the 5-HT2B receptor might not only inhibit NET proliferation but also decrease 5-HT release and profibrotic growth factor secretion, hence, abrogating the development of peritumoral fibrosis and possibly angiogenesis.

To further define the role of 5-HT in the pathogenesis of both tumorigenesis and fibrosis, we developed a NET:fibroblast coculture system. This was used to evaluate the effects of selectively targeting 5-HT2B receptors in individual components of this system. HEK293 cells exhibit features of fibroblast-specific intracellular pathways and have been previously used as fibroblast model.53-55 These cells do not express 5-HT2B receptors and cannot be inhibited by PRX-08066, but, as we have noted (above), respond to 5-HT with proliferation and synthesis of fibrogenic factors. Blocking 5-HT2B receptor with PRX-08066 resulted, as expected, in a significant increase of caspase 3 and decrease of Ki67 in KRJ-I cells in coculture. In addition, a significant decrease in TGβ1, CTGF, and FGF2 synthesis was noted, similar to our single cell studies. Interestingly, transcripts of these factors and Ki67 were also decreased in coculture d HEK293 cells, despite the absence of a 5-HT2B receptor on this cell line. Given that levels of these transcripts remained unchanged in HEK293 cells cultured alone and our observations that 5-HT directly stimulate HEK293 activity through the 5-HT2A/C receptor, it is clear that KRJ-I-derived 5-HT in the coculture model is responsible for the cellular cross-talk between tumor cells and fibroblasts. To confirm the biological relevance of the transcript analysis data, we determined caspase 3 activity and cell death in coculture d KRJ-I and HEK293 cells. A significant increase in dead cells was noted in PRX-08066 treated KRJ-I cells and induction of the caspase 3 pathway in HEK293 cells confirming the transcript analyses.

In conclusion, SI-NETs express 5-HT2B receptors as does the EC cell-derived SI-NET cell line, KRJ-I. Targeting this receptor with the specific antagonist, PRX-08066, resulted in a significant decrease in 5-HT secretion and a reduction in KRJ-I cell proliferation. The MEK/ERK pathway played a role in this antiproliferative effect. By using a coculture model of the tumor environment (KRJ-I:HEK293 cells), we identified a decrease in profibrotic/angiogenetic factors in both tumor cells and fibroblasts after inhibition of the 5-HT2B receptor. We postulate that the tumor proliferative activity of SI-NETs (including cell growth and the development of desmoplasia) is associated with the particular microenvironment in the peritoneum and that tumor cells support this necessary milieu through the secretion of profibrotic/angiogenetic factors. Targeting key receptors such as 5-HT2B and blocking critical proliferative pathways in the tumor cells provide a possible therapeutic strategy to prevent tumor progression, fibrosis, and metastasis in this neuroendocrine neoplasia. This may apply to other fibrotic processes associated with neuroendocrine cell dysregulation, eg, Crohn's disease.56


Institutional support was provided by R01-CA115285 (IMM), R01-DK080871-01A2 (MK), Kontaktutvalget St. Olaf's Hospital and Faculty of Medicine-NTNU (BG).