In vivo effects of a combined 5-HT 1B receptor/sert antagonist in experimental pulmonary hypertension

Transcription

1 Cardiovascular Research (2010) 85, doi: /cvr/cvp306 In vivo effects of a combined 5-HT 1B receptor/sert antagonist in experimental pulmonary hypertension Ian Morecroft 1, Louisa Pang 1, Marta Baranowska 2, Margaret Nilsen 1, Lynn Loughlin 1, Yvonne Dempsie 1, Caroline Millet 1, and Margaret R. MacLean 1 * 1 FBLS, University of Glasgow, Glasgow G12 8QQ, UK; and 2 Medical University of Bialystok, Mickiewicz Str. 2A, Bialystok, Poland Received 28 May 2009; revised 26 August 2009; accepted 1 September 2009; online publish-ahead-of-print 7 September 2009 Time for primary review: 41 days Aims A mechanism for co-operation between the serotonin (5-hydroxytryptamine, 5-HT) transporter and 5-HT 1B receptor in mediating pulmonary artery vasoconstriction and proliferation of pulmonary artery smooth muscle cells has been demonstrated in vitro. Here we determine, for the first time, the in vivo effects of a combined 5-HT 1B receptor/serotonin transporter antagonist (LY393558) with respect to the development of pulmonary arterial hypertension (PAH) and its in vitro effects in human pulmonary artery smooth muscle cells (hpasmcs) derived from idiopathic PAH (IPAH) patients.... Methods We determined the effects of LY as well as a selective serotonin transporter inhibitor, citalopram, on right and results ventricular pressure, right ventricular hypertrophy, and pulmonary vascular remodelling in wildtype mice and mice over-expressing serotonin transporter (SERTþ mice) before and after hypoxic exposure. We also compared their effectiveness at reversing PAH in SERTþ mice and hypoxic mice. Further, we examined the proliferative response to serotonin in IPAH hpasmcs. We also clarified the pharmacology of serotonin-induced vasoconstriction and 5-HT 1B receptor/serotonin transporter interactions in mouse isolated pulmonary artery. Citalopram had a moderate effect at preventing and reversing experimental PAH in vivo whereas LY was more effective. LY was more effective than citalopram at reversing serotonin-induced proliferation in IPAH hpasmcs. There is synergy between 5-HT 1B receptor and serotonin transporter inhibitors against serotonin-induced vasoconstriction in mouse pulmonary arteries.... Conclusion 5-HT 1B receptor and serotonin transporter inhibition are effective at preventing and reversing experimental PAH and serotonin-induced proliferation of PASMCs derived from IPAH patients. Targeting both the serotonin transporter and 5-HT 1B receptor may be a novel therapeutic approach to PAH Keywords Pulmonary hypertension Serotonin Hypoxia Serotonin transporter 5HT 1B 1. Introduction Pulmonary arterial hypertension (PAH) is characterized by an increase in PA pressure, increased pulmonary vascular resistance, vascular remodelling, and right ventricle failure. 1 Approximately 70% of patients with the familial form of PAH have a mutation in the gene encoding bone morphogenetic protein (BMP) receptor type II (BMPR-II). 2 However, there is variable phenotypic expression of PAH among carriers of mutated BMPR-II genes, likely to be related to a second hit caused by environmental and/or genetic modifiers. Studies have implicated serotonin, the serotonin transporter (SERT), and the 5-HT 1B receptor (5-HT 1B R) in the pathobiology of PAH. 3 For example, in mice, chronic hypoxia-induced PAH and dexfenfluramine-induced PAH are dependent on peripheral serotonin syntheisis. 4,5 Human pulmonary arterial smooth muscle cells (hpasmcs) from idiopathic * Corresponding author. Tel: þ , Fax: þ , m.maclean@bio.gla.ac.uk Published on behalf of the European Society of Cardiology. All rights reserved. & The Author For permissions please journals.permissions@oxfordjournals.org.

2 594 I. Morecroft et al. PAH (IPAH) patients show increased expression of the SERT related to increased serotonin-induced proliferation. 6 The 5-HT 1B R mediates constriction in human pulmonary arteries (PAs) 7 and proliferation of hpasmcs. 8 There is co-operation between the 5-HT 1B R and SERT in hpasmcs. For example, 5-HT 1B R-mediated phosphorylation of ERK1/2 and perk1/2 nuclear translocation depends on SERT activity, monoamine oxidase activity, and reactive oxygen species. 8 In bovine PASMCs, there is also co-operation between the SERT and 5-HT 1B Rs via rho kinase activation and perk1/2 nuclear translocation. 9 Mice over-expressing the gene for human SERT (SERTþ mice) have elevated pulmonary pressures and are more susceptible to hypoxia-induced PAH. 10 In vivo such SERT over-expression increases rho kinase-dependent pulmonary remodelling and in vitro rho kinase potentiates 5-HT 1B receptor-stimulated ERK activation and proliferation. 11 Hence, co-operation between SERT and 5-HT 1B R has been demonstrated and explained mechanistically in vitro, but to date there has been no examination of the significance of this interaction in vivo. Recently, combined 5-HT 1B R and SERT antagonists have been described that may become clinically available. 8,9 One of these, LY393558, is very potent in vitro against serotonin-induced vasoconstriction in rat isolated PAs. 12 It was therefore of great interest to investigate whether LY had additional effects over and above those of the selective SERT inhibitor citalopram against experimental murine PAH models. We examined the effects of citalopram and LY in two models of PAH, the SERTþ mouse and the hypoxic mouse. We determined their effectiveness at preventing PAH as well as reversing established PAH. Further, we clarified the pharmacology of the contractile response to serotonin in mouse PAs and examined how this was altered after in vivo treatment with LY or citalopram. Finally, we examined the effects of the antagonists on serotonin-induced proliferation of PASMCs derived from IPAH patients. 2. Methods 2.1 In vivo The generation and characterization of the SERTþ mice have been described previously. 10 Animal procedures were conducted in accordance with the United Kingdom Animals (Scientific Procedures) Act 1986 (Home Office licence PPL60/3773) and with the Guide for the Care and Use of Laboratory Animals published by the US National Institutes of Health (NIH publication No , revised 1996). 2.2 Hypoxia Prevention study Female 5 6 month wildtype (WT) mice (C57BL/CBA) or SERTþ mice (n ¼ 8 10) were maintained in normoxic or hyobaric/hypoxic conditions for 14 days as previously described. 10,13 Mice were treated orally by gavage for the duration of the hypoxic exposure with either LY (30 mg kg 21 day 21, Eli Lilly) citalopram (20 mg kg 21 day 21 ; Tocris Cookson) or vehicle [1%carboxymethylcellulose (Sigma)]. Doses were selected on the basis of published K i values for the transporter and 5-HT 1B R Reversal study To examine the effects of the inhibitors on established hypoxia-induced PAH, mice were maintained in hypobaric/hypoxic conditions for 2 weeks. Drug treatment was then started (as described above) and the mice exposed to a further 2 weeks of hypoxia. 2.3 Characterization of PAH Pressure measurements Under isoflurane (1.5% in O 2 ) anaesthesia, right ventricular systolic pressure (RVP) was measured via a 25-gauge needle advanced into the right ventricle trans-diaphragmatically. 10,13 Heart rate was derived from the RVP and systemic arterial pressure was recorded via a cannula placed in the carotid artery Right ventricular hypertrophy Right ventricular hypertrophy (RVH) was determined as ratio of RV weight to left ventricular weightþseptum [RV/(LVþS)] Remodelling Remodelling was assessed as described previously. 13 Briefly, lung sections (four to six mice per group) were stained with Elastica-Van Gieson stain and microscopically assessed in a blinded fashion. PAs (25 80 mm external diameter) were considered muscularized if they possessed a distinct double-elastic lamina visible for at least half the diameter in the vessel cross-section. The percentage of remodelled vessels was calculated as number of muscularized vessels/total number of vessels counted 100. Group size¼100 vessels assessed/ animal, average vessel size 70 mm external diameter. 2.4 Myography PAs of 350 mm internal diameter (i.d.) were studied as described previously. 12,13 Briefly, PAs from normoxic mice were set up at tensions equivalent to their mean in vivo RVP (12 15 mmhg), whereas PAs from hypoxic mice were set up at tensions equivalent to the elevated in vivo mean pressures observed after exposure to hypoxia (25 30 mmhg). Antagonists were allowed a 45 min equilibrium period. LY was dissolved in DMSO and the relevant serotonin cumulative response curves (CRCs) were also carried out in the presence of DMSO Effects of chronic dosing with LY and citalopram Responses to serotonin were examined in PAs removed from the vehicle-, LY , and citalopram-dosed mice after assessment of RVP Acute effects of serotonin agonists and antagonists To ensure that the effects observed in the vessels removed from the dosing study animals were not altered by the dosing regime or prolonged exposure to anaesthesia, we also examined PAs removed from naïve mice, for example, PAs removed from mice exposed neither to drugs nor anaesthesia. In these arteries, we used further agonist and antagonists to further examine the pharmacology of the serotonin response. Antagonists were allowed a 45 min equilibrium period. The antagonists studied were: 5-HT 1B antagonist GR55562 (1 mmol/l; Tocris, UK), the 5-HT 2A receptor antagonist ketanserin (30 nmol/l; Tocris, UK), the SERT inhibitor citalopram (1 mmol/l; Tocris, UK), and the combined 5-HT 1B /SERT inhibitor LY (100 nmol/l; Eli-Lilly, USA). The responses to the 5-HT 2A agonist a-methyl5-ht and the 5-HT 1 agonist 5-carboxamidotryptamine (5-CT) were also determined. In a separate series of experiments,

3 SERT and 5HT 1B receptors in pulmonary hypertension 595 Figure 1 Effects of LY (þly, 30 mg kg 21 day 21 ) and citalopram (þcit, 20 mg kg 21 day 21 ) on development of right ventricular systolic pressure (RVP) (A, D; n ¼ 5 9), pulmonary vascular remodelling (B, E; n ¼ 4 5) and right ventricular hypertrophy (RVH) (C, F; n ¼ 5 9) in WT (A C) and SERTþ (D F) mice. Mice were subjected to 2 weeks of hypoxia. Data are mean+sem. ***Value significantly greater than corresponding value in vehicle treated normoxic mice (P, 0.001). Value significantly less than corresponding value in vehicle treated mice ( P, 0.05, P, 0.01, P, 0.001). Value significantly greater than corresponding value in LY treated mice ( P, 0.05, P, 0.001). Value significantly greater than corresponding value in WT vehicle treated mice ( P, 0.05, P, 0.01, P, 0.001). WTH, wildtype hypoxic; SERTþH, SERTþ hypoxic; v, vehicle treated; n, number of mice. we investigated if there was in vitro synergism between GR55562 and citalopram in PAs removed from wildtype mice. To this effect, we selected concentrations of GR55562 and citalopram that alone had no effect on the contractile response to serotonin and then examined the response in the presence of both antagonists applied together. 2.5 Proliferation of hpasmcs In order to examine the relevance of our results to a relevant human model, we studied distal hpasmcs from patients with non-familial IPAH and hpasmcs from macroscopically normal lung (from patients undergoing lung resection for suspected malignancy). The investigation conforms with the principles outlined in the Declaration of Helsinki. Cells were prepared by Professor N. Morrell, Department of Medicine, University of Cambridge School of Clinical Medicine, UK as described previously. 2 Cells were plated out in 24-well plates at a density of cells/well and grown in full medium for 24 h before serum starvation for 24 h. Cells were pre-incubated with citalopram (1 mmol/l) or LY (1 mmol/l) for 20 min before adding 5-HT (1 mmol/l) and a final concentration of 0.2% FCS. After 18 h, 0.5 mci [ 3 H]-thymidine was added and proliferation stopped after a further 6 h. Cells were washed with cold PBS twice, washed

4 596 I. Morecroft et al. a third time with cold 5% trichloroacetic acid, and lysed with 0.5 M NaOH. The radioactivity was measured in a liquid scintillation counter. Stimulation indices were calculated as the fold-increase from the basal rate of proliferation as determined by [ 3 H]-thymidine incorporation. Cell counts were performed using a haemocytometer at day 5. Assays were carried out in duplicate. In order to normalize data, cell count mean of the FCS wells of each assay, carried out in duplicate, was calculated and each well of the same assay normalized to this value. 2.6 Analysis In vitro studies, responses were normalized by expressing as per cent of a 50 mm KCl response. Where a maximum response to serotonin or agonist was achieved, the log[mol/l] of serotonin giving a half maximum response (pec 50 ) in the presence/absence of inhibitors was determined. Comparisons were made using ANOVA on all data sets presented (unless otherwise indicated), followed by Neuman Keuls post hoc test. All results are expressed as mean+sem and P, 0.05 considered significant. 3. Results 3.1 Prevention study Wildtype Neither LY nor citalopram had any effect on PAH parameters under normoxic conditions. Under hypoxia, mice developed PAH and LY had a much greater effect than citalopram. Hypoxia-induced increases in RVP were completely inhibited by LY393558, whereas citalopram only partially reduced RVP (Figure 1A). LY markedly attenuated pulmonary vascular remodelling while citalopram had only a modest effect (Figure 1B) and LY attenuated the rise in RV/(LVþS) while citalopram had no effect (Figure 1C). Remodelling in small PAs of the mice studied is illustrated in Figure SERT1 Even under normoxic conditions, vehicle-treated SERTþ mice demonstrated elevations in RVP and RVH and these elevations were ablated by LY393558, but not citalopram (Figure 1D and F). With hypoxia, SERTþ mice developed exaggerated elevations in RVP, pulmonary vascular remodelling, and RVH compared with hypoxic WT mice (Figure 1D F). LY had a profound effect on the elevation in RVP, reducing it to a much greater extent than that observed with citalopram (Figure 1D). Likewise, LY ablated the increase in pulmonary vascular remodelling to a greater extent than citalopram (Figure 1E) and normalized the hypoxia-induced increase in RV/(LVþS) whereas citalopram only partially protected against RV hypertrophy (Figure 1F). Remodelling in small PAs of the mice studied is illustrated in Figure Reversal study Figure 3 shows that following this additional 2 week exposure to hypoxia, there was a further increase in RVH and pulmonary Figure 2 Photomicrographs showing hypoxia-induced remodelling in small PAs from wildtype (WT) and SERTþ mice. The figure demonstrates that LY reversed the remodelling to a greater extent than citalopram (see Figure 1 for data analysis). Note the readily identifiable double elastic lamina (white arrow) and newly formed smooth muscle layer (white asterisk) in the hypoxic group. Scale bars ¼ 10 mm.

5 SERT and 5HT 1B receptors in pulmonary hypertension 597 vascular remodelling in the WT mice (not SERTþ mice) compared with mice subjected to 2 weeks of hypoxia Wildtype LY significantly reversed the hypoxia-induced elevation in RVP (Figure 3A) and reduced remodelling (Figure 3B) and RVH (Figure 3C). Citalopram had no effect on RVP and reduced remodelling and RVH to a lesser extent than LY SERT1 SERTþ mice demonstrated an increase in hypoxia-induced PAH compared with the WT mice (Figure 3D F). LY reversed the elevation in RVP and reduced pulmonary vascular remodelling and RVH (Figure 3D F). Citalopram reduced the elevation in RVP and remodelling to a lesser extent than LY but reversed RVH to a similar extent (Figure 3D F). Systemic blood pressures and heart rate were not affected by hypoxia and did not change in any of the treatment groups (data not shown). Figure 3 Effects of LY (þly, 30 mg kg 21 day 21 ) and citalopram (þcit, 20 mg kg 21 day 21 ) treatment on the reversal of right ventricular systolic pressure (RVP) (A, D; n ¼ 5 9), pulmonary vascular remodelling (B, E; n ¼ 4 5) and right ventricular hypertrophy (RVH) (C, F; n ¼ 5 9) in WT (A C) and SERT (D F) mice. Data are mean + SEM. ***Value significantly greater than corresponding value in vehicle-treated normoxic mice (P, 0.001). Value significantly less than corresponding value in vehicle treated mice ( P, 0.05, P, 0.01, P, 0.001). Value significantly greater than corresponding value in LY treated mice ( P, 0.05). Value significantly greater than corresponding value in WT vehicle treated mice ( P, 0.05, P, 0.01). # Value significantly greater than WTH 2 weeks ( # P, 0.05, ## P, 0.001, paired t-test). WTH, wildtype hypoxic; WTH 2 wks, data after 2 weeks of hypoxia; WTH 4 wks, data after four weeks of hypoxia. SERTþH, SERTþ hypoxic; v, vehicle treated; n, number of mice.

6 598 I. Morecroft et al. 3.3 Vasoconstrictor responses to serotonin Prevention study In PAs from WT mice, the response to serotonin was markedly attenuated by LY393558, in a non-competitive fashion, but un-affected by citalopram (Figure 4A; Table 1). There was a 50% increase in the response to serotonin in WT mice arteries after exposure to hypoxia (Figure 4B; Table 1) and this was inhibited by LY while in contrast, citalopram augmented the response to serotonin (Figure 4B; Table 1). In PAs from normoxic SERTþ mice, the response to serotonin was ablated by LY but not by citalopram (Figure 4C; Table 1). The sensitivity to serotonin was greatly increased in SERTþ PAs in response to hypoxia and this was markedly attenuated by LY393558, whereas citalopram treatment was without effect (Figure 4D; Table 1). The E max to serotonin was higher (P, 0.05) in PAs from WT hypoxic mice than from SERTþ hypoxic mice Reversal study From Figure 4 and Table 1, it can be seen that the responses to serotonin are similar to those observed in the vessels removed from Figure 4 Serotonin-induced contraction in isolated PAs from mice chronically dosed with LY (þly, 30 mg 21 kg 21 day 21 ) and citalopram (þcit, 20 mg 21 kg 21 day 21 ). PAs removed from mice in prevention study: WT (A, B) and SERTþ (C, D). PAs removed from mice in reversal study: WT (E) and SERTþ (F). Data are mean+sem. v, vehicle treated. See Table 1 for statistical analysis.

7 SERT and 5HT 1B receptors in pulmonary hypertension 599 Table 1 Serotonin-induced contraction in PAs from mice chronically dosed with LY (30 mg 21 kg 21 day 21 ), citalopram (1Cit, 20 mg 21 kg 21 day 21 ) or vehicle (1% carboxymethylcellulose) pec 50 E max... Prevention study Normoxic WT vehicle WTþLY * WTþCitalopram SERT þ vehicle SERT þ þly SERT þ þcitalopram Hypoxic WT vehicle k WTþLY # WTþCitalopram # SERT þ vehicle k SERT þ þly SERT þ þcitalopram ** Reversal Study WT vehicle WTþLY WTþCitalopram SERT þ vehicle SERT þ þly SERT þ þcitalopram pec 50 denotes -logarithm of EC 50 ; E max, maximal contraction. Data are shown as mean + SEM. n ¼ 4 7 animals. *P, 0.01 and P, vs. normoxic WT vehicle; P, 0.05 and P, vs. SERTþvehicle; k P, vs. normoxic data; # P, vs. hypoxic WT data; **P, 0.05 and P, vs. hypoxic SERTþ data; P, 0.01 vs. WT vehicle reversal data; P, vs. SERTþ vehicle reversal data. mice used in the prevention study, with respect to pec 50 and E max. LY did not, however, reduce E max (Figure 4E and F) in the way that it did in the prevention study vessels (Figure 4A D). However, the pec 50 of serotonin in arteries removed from dosed WT and SERTþ mice was significantly inhibited by LY393558, for example, the antagonist was acting in a competitive fashion. Citalopram was without effect (Figure 4E and F; Table 1) Acute effects of antagonists The 5-HT 1B R antagonist GR55562 and the 5-HT 2A R antagonist ketanserin both decreased the potency of serotonin in naive PAs from WT mice, with ketanserin having the greatest effect. LY decreased the potency further while citalopram was without significant effect (Figure 5A). Consistent with a response to both 5-HT 2A - and 5-HT 1 -receptor activation, both the 5-HT 2A R agonist a-methyl-5-ht and the 5-HT 1 R agonist 5-CT induced contraction in the WT PAs (Figure 5). Consistent with a greater response via 5-HT 2A -activation, a-methyl-5-ht was more potent that 5-CT (Figure 5B; Table 2). The sensitivity to serotonin was decreased in PAs from SERTþ mice (Figure 5D and E, Table 2). That this was due to SERT over-expression is shown by the ability of citalopram to abolish this effect, increasing the potency of serotonin as we have reported previously (Figure 5D). 4 In SERTþ mouse PAs, both ketanserin and LY inhibited the response to serotonin, whereas the 5-HT 1B R antagonist GR55562 had no effect (Figure 5D; Table 2). That 5-HT 1B Rs were present is shown by the presence of a GR55562-sensitive response to the 5-HT 1 agonist 5-CT (Figure 5E). a-methyl-5-ht also induced a potent contraction in SERTþ PAs, consistent with a 5-HT 2A -mediated response. We investigated possible synergy between the 5-HT 1B R and the SERT in vitro in WT mouse PAs by selecting concentrations of GR55562 and citalopram that, on their own, had no effect on serotonin-induced contraction. We then examined the response to serotonin in the presence of both antagonists applied together at these concentrations. Figure 5C (and Table 2) demonstrates that the combination did antagonize the response to serotonin in a competitive fashion, demonstrating in vitro synergism. 3.4 Proliferation of hpasmcs Serotonin did not, while the positive control, PDGF did, induce proliferation of cells from control lungs (Figure 6A). Serotonin did, however, induce proliferation of cells derived from IPAH patients as did PDGF. LY induced a 58% reduction of serotonin-induced proliferation in these cells (reducing proliferation below that observed with FCS consistent with this containing serotonin), whereas citalopram induced only a 22% reduction (Figure 6A). To confirm these results in terms of cell count, Figure 6B shows that both serotonin and PDGF induced proliferation as determined by cell count. This was inhibited by citalopram and further ablated by LY Discussion We have previously demonstrated a mechanism for the interaction between SERT and the 5-HT 1B Rinin vitro pulmonary fibroblasts and in hpasmcs 8,11 and here we wished to examine the in vivo effects of a combined SERT/5-HT 1B R inhibitor (LY393558) in the SERTþ model of PAH as well as in the hypoxic mouse model. We also wished to examine if the effect of this antagonist was greater than a SERT antagonist (citalopram) alone. We wished to study the ability of these antagonists to prevent hypoxia induced PAH in WT and SERTþ mice. As clinical PAH is well established by the time it presents, we also wished to examine their ability to reverse established PAH in our models. As previously shown, SERTþ mice exhibited PAH and increased hypoxia-induced PAH. 10 This has also been shown in mice that have increased SERT expression in PASMCs. 14 In mice exposed to 2 weeks of hypoxia and the antagonists simultaneously (prevention study), citalopram reduced the hypoxia-induced RVP and remodelling in WT mice, but did not affect the elevation in RVH. This may suggests that SERT activity contributes more to remodelling of the PAs in WT mice than to pulmonary pressures consistent with evidence that pulmonary vascular remodelling is mediated via SERT activity. In WT and SERTþ mice, LY was more effective that citalopram at preventing hypoxia-induced increases in

8 600 I. Morecroft et al. Figure 5 Effect of selective serotonin antagonists and agonists on serotonin-induced vasoconstriction in PAs from naïve control, non-dosed mice. WT PAs: effects of antagonists (A) and agonists (B). SERTþ PAs: effects of antagonists (D) and agonists (E). Concentrations of antagonists used: citalopram (SERT, 1 mmol/l), LY (SERT plus 5-HT 1B, 200 nmol/l), ketanserin (5-HT 2A, 30 nmol/l) and GR55562 (5-HT 1B,1mmol/ L). Data are mean + SEM. See Table 2 for statistical analysis. (C) the effects of 0.3 mmol/l citalopram and 0.3 mmol/l GR55562 either on their own or in combination, on serotonin-induced vasoconstriction in WT PAs. RVP, remodelling, and RVH. Curiously, citalopram was unable to prevent the elevated RVP and RVH observed in the normoxic SERTþ mice. This suggests that SERT over-expression alone is not responsible for this PAH phenotype and that other influences are activated in SERTþ mice. Our results suggest this may be 5-HT 1B R activation as the dual SERT and 5-HT 1B R antagonist LY was effective at preventing these important indices of PAH; the effects of LY were superior to citalopram suggesting involvement of the 5-HT 1B R. In the mice included in the reversal study, drugs were administered during a further 2 weeks of hypoxic exposure. The additional exposure to hypoxia increased remodelling and RVH further in WT mice but not in SERTþ mice where presumably the effects were already maximal. LY partially reversed all the established indices of PAH in these WT and SERTþ mice. Citalopram was less effective than LY393558, with the exception that it induced an equivalent reversal of RVH in the SERTþ mice. Together, these observations suggest that a combined inhibitor may be more effective than a SERT inhibitor alone in a clinical setting where PAH is always well established upon presentation. The K i against SERT is very similar for both LY and citalopram ( nm). The increased ability of LY to prevent and reverse PAH suggests that SERT over-expression may have facilitated the activity of the 5-HT 1B R receptors. We have previously shown that the 5-HT 1B R is exquisitely prone to pharmacological synergy induced by elevations in pulmonary vascular tone or Gq-receptor activation. 15,16 Indeed, if we examine the contractile response to the 5-HT 1 agonist 5-CT in Figure 5, we can see that the response to 1 mmol/l was 20% (of response to 50 mm KCl) in the SERTþ mice vessels, but negligible in the WT mice vessels. The response to 10 mmol/l 5-CT was 60% in the SERTþ mice vessels, but only 30% in the WT mice vessels. As responses were inhibited by the 5-HT 1B R antagonist GR55562, this suggests that there is increased activity of the 5-HT 1B R in PAs removed from the SERTþ mice. LY also has affinity at the 5-HT 2A (K i : 100 nm) 17 and as our in vitro studies (discussed below) demonstrate that serotonin-induced vasoconstriction in isolated PAs can also be mediated by the 5-HT 2A R, this may have contributed to this effect. Treatment with LY or citalopram did not affect systemic blood pressures or heart rate consistent with a selective effect on

9 SERT and 5HT 1B receptors in pulmonary hypertension 601 Table 2 Effect of selective serotonin antagonists and agonists on serotonin-induced vasoconstriction in PAs removed from naïve control, non-dosed mice pec 50 E max... Wildtype Antagonists 5-HT þly N.M. N.M. þ Citalopram[1] þketanserin * þgr55562[1] þgr55562[2] þcitalopram[2] þcitalopram[2] þ GR55562[2] ## Agonists 5-HT CT N.M. N.M. a-methyl-5-ht * SERT þ Antagonists 5-HT þly þ Citalopram k þgr CTþGR ** Agonists 5-HT CT # a-methyl-5-ht The effects of acute in vitro administration of antagonists are shown: ketanserin (30 nmol/l, 5-HT 2A ), GR55562 ([1]: 1 mmol/l; [2]: 30 nmol/l 5-HT 1B ), LY (200 nmol/l, SERT and 5-HT 1B ) and citalopram ([1]: 1 mmol/l; [2]: 30 nmol/l SERT). pec 50 : -logarithm of EC 50 ; E max, maximal contractile effect; 5-HT serotonin, 5-CT 5-carboxamidotryptamine, N.M., no maximum response achieved, and n, number of mice. Data are mean + SEM. n ¼ *P, 0.05, P, 0.01 and P, vs. WT 5-HT; # P, 0.05, k P, 0.01 and P, vs. SERTþ5-HT; ** P, 0.01 vs. SERTþ5-CT, ## P, 0.05 vs. WT 5-HT. the pulmonary circulation. This is consistent with evidence that, uniquely to the pulmonary circulation, the SERT and the 5-HT 1B R co-operate in mediating both vasoconstriction and proliferation. 8,12 The limited efficacy of citalopram in the prevention of PAH in this study is in contrast with a study by Guignabert et al. 18 demonstrating that the SERT inhibitor fluoxetine completely attenuates the development of PAH in the monocrotaline rat model. Fluoxetine, however, is not a selective SERT inhibitor in that it is also a 5-HT 2A receptor antagonist, 19 can alter calcium uptake and sensitivity 20 and up- or down-regulate the 5-HT 2B receptor. 21 These effects may have contributed to its effects in the monocrotaline model which also exhibits a different pathobiology to either SERT over-expression or hypoxia. Hypoxia increased the contractile response to serotonin in PAs from the dosed mice, an effect we have previously demonstrated in the hypoxic rat and attributed to an increase in 5-HT 1 -receptor mediated constriction. 16,22 Consistent with this, LY inhibited the responses to serotonin in PAs from both normoxic and hypoxic WT mice, having a greater effect on the maximum response to serotonin in the PAs from the hypoxic mice. Citalopram, however, potentiated the responses to serotonin in the hypoxic WT PAs and had no effect on responses to serotonin in the normoxic WT PAs. This is consistent with evidence that the 5-HT 1B R and the SERT co-operate in mediating serotonin-induced vasoconstriction 12 and may explain why LY was more effective than citalopram in reversing the effects of hypoxia on RVP. We compared the effects of chronic dosing in vivo with an acute administration of LY and citalopram in vitro on PAs derived from naïve non-dosed mice. In addition, as we have not previously determined the pharmacological profile of the serotonin-induced contraction in mouse PAs, we examined the effects of a 5-HT 2A R antagonist (ketanserin), a 5-HT 1B R antagonist (GR55562), a 5-HT 2A agonist a-methyl5ht, and a 5-HT 1 agonist 5-CT. The results were entirely consistent with the observations made in the arteries removed from the chronically dosed mice. In WT and SERT PAs, the response to serotonin was potentiated by citalopram, but profoundly inhibited by LY In addition, in WT mouse arteries, responses were inhibited by GR55562 and ketanserin and there was a contractile response to the 5-HT 2A agonist a-methyl5ht and the 5-HT 1 agonist 5-CT, consistent with the response to serotonin being mediated by both the 5-HT 1B R and 5-HT 2A R. In the SERTþ mouse vessels, responses were also inhibited by ketanserin and there was a contractile response to a-methyl5ht, consistent with a 5-HT 2A R-mediated response. Citalopram potentiated the contractile response while LY markedly inhibited the response. The 5-HT 1B R antagonist failed to inhibit the contractile response to serotonin. That 5-HT 1B Rs were present and functional is demonstrated by the presence of a GR55562-sensitive response to 5-CT. These results are consistent with our hypothesis, discussed above, that over-expression of the SERT leads to interactions with the function of the 5-HT 1B R such that there is co-operation between the 5-HT 1B R and SERT in mediating the contractile responses to serotonin. The results show (Figure 4) that LY acted in a noncompetitive fashion in arteries from mice exposed to 2 weeks of hypoxia (prevention study) whereas in the arteries removed from the mice that were exposed to 4 weeks of hypoxia (from the reversal study), LY acted in a competitive fashion. According to receptor theory, this phenomenon can be observed when there is an increase in receptor reserve. Thus, when receptor reserve is low, an antagonist may reduce the maximum response to an agonist whereas when reserve is increased, the same concentration may not affect the maximum response but shift the CCRC in a competitive fashion. Serotonin activates multiple receptors, any of which may be increased by exposure to prolonged hypoxia and this is the likely explanation for this observation. We investigated further if there was a synergistic interaction between the SERT and the 5-HT 1B R in vitro. We chose reduced concentrations of GR55562 and citalopram that on their own

10 602 I. Morecroft et al. Figure 6 Serotonin-induced proliferation in hpasmcs. Control: cells derived from control patients. IPAH: cells derived from IPAH patients. (A) The effect of 1 mmol/l serotonin (24 h stimulation) in the absence and presence of either 1 mmol/l citalopram or 1 mmol/l LY Data are expressed as the fold increase from the basal rate of proliferation as determined by [ 3 H]-thymidine incorporation. Data are combined from six repeat experiments are mean+sem. *P, vs. control; # P, 0.01 vs. serotonin and P, vs. serotonin. Data shown are representative data from n ¼3 IPAH/control patients. The response to platelet derived growth factor (PDGF, 10 mg/ml, 24 h) is included as a positive control. (B) Serotonin-induced proliferation of hpasmcs from IPAH patients assessed by cell counting. Cells were grown in 5% FCS and proliferation assessed in the absence of treatment (control 5% FCS), in presence of serotonin (1 mmol/l), serotonin plus citalopram (1 mmol/l), and serotonin plus LY (1 mmol/l). After 5 days of treatment, cells were counted and results were normalized to control 5% FCS. Data shown are representative data from n ¼ 3 IPAH patients. *P, vs. control; # P, 0.01 vs. serotonin, and P, vs. serotonin. The response to platelet derived growth factor (PDGF, 10 ng/ml) is included as a positive control. had no effect, and then examined the effect of both given in combination at these concentrations. The results showed that the combination inhibited the contraction, inducing a one-fold log shift in the pec 50. Figure 5A and Table 2 show that we were unable to calculate the pec 50 for LY393558, but it is clear that it induced a two-fold log shift in the CRC to serotonin and hence this was more effective than the combination of the two antagonists. There could be many reasons for this. We were using reduced concentrations of GR55562 and citalopram in order to demonstrate synergy. In addition, GR55562 has a K i of 100 nm against the 5-HT 1B R compared with the K i of LY of 1 nm. Additionally, as there is evidence for the presence of the 5-HT 2A R in this artery, the ability of LY to inhibit 5-HT 2 receptors may have contributed to this effect. LY was significantly more effective than citalopram at reversing pulmonary vascular remodelling in vivo in the mouse models. We wished to investigate whether this was relevant to the human situation and hence examined the effect of these antagonists in PASMCs derived from normal human lung and lung from IPAH patients. Previous studies have demonstrated the presence of the 5-HT 1B R, 5-HT 2A R, and the SERT in control and IPAH patients and observed an increase in SERT expression in IPAH PAs and PASMCs. 6,23 We demonstrate that there is serotonin-induced proliferation in hpasmcs derived from IPAH patients. While cells from control human lung did proliferate in response to PDGF, they did not respond to serotonin. This is consistent with other studies that have demonstrated that serotonin-induced proliferation of PASMCs is via SERT activity and proliferation depends on SERT expression. It is hence difficult to demonstrate serotonin-induced proliferation in primary cultures of normal hpasmcs where SERT expression is low or absent. 23 Indeed, serotonin-induced proliferation in hpasmcs derived from IPAH patients has been attributed to increased SERT expression. 23 We demonstrated that the effect of LY was twice as great compared with citalopram at reversing the proliferative effects of serotonin in PASMCs from IPAH patients, as assessed by [ 3 H]-thymidine incorporation and cell counting. This is consistent with evidence that the 5-HT 1B R and the SERT co-operate in mediating hpasmc proliferation in vitro. 8,9 Our results confirm that in vivo, combined 5-HT 1B R and the SERT blockade is effective at reversing and preventing experimental PAH and that the contractile and mitogenic effects of serotonin play a role in experimental PAH. This is entirely consistent with our recent studies showing that hypoxia- and dexfenfluramine-induced PAH is ablated in mice which are devoid of tryptophan hydroxylase1 and hence peripheral serotonin. 4,5 While the models we have studied are more relevant to secondary PAH associated with hypoxic lung disease, they are also clinically relevant as it has been shown that SERT overexpression may influence the development of PAH in some cohorts of patients A retrospective study showed, however, that SERT inhibitors alone did not increase survival. 26 The results of ongoing clinical trails with citalopram will be of interest as these will either confirm or contradict this observation. In

11 SERT and 5HT 1B receptors in pulmonary hypertension 603 man, it is also the 5-HT 1B R that mediated vasoconstriction in human PAs. 7 In addition, there is co-operation between the SERT and 5-HT 1B R in mediating mts1-induced proliferation of hpasmcs. 8 Consistent with this, we show here co-operation between the SERT and 5-HT 1B R in serotonin-induced proliferation of PASMCs from IPAH patients. Our study therefore suggests that dual blockade of the SERT and 5-HT 1B R may be a novel therapeutic strategy for the treatment of PAH. Acknowledgements We thank Eli Lilly for the donation of LY and Professor Nick Morrell, University of Cambridge for provision of human PASMCs. Conflict of interest: none declared. Funding This work was supported by the British Heart Foundation [PG/08/037/ 24921]; and the Biotechnology and Biological Sciences Research Council [BB/D007623/1]. M.B. was a European Respiratory Society Research Fellow. References 1. Chin KM, Rubin LJ. Pulmonary arterial hypertension. J Am Coll Cardiol 2008;51: Yang XD, Long L, Southwood M, Rudarakanchana N, Upton PD, Jeffery TK et al. Dysfunctional smad signaling contributes to abnormal smooth muscle cell proliferation in familial pulmonary arterial hypertension. Circ Res 2005;96: Dempsie Y, MacLean MR. Pulmonary hypertension: therapeutic targets within the serotonin system. Br J Pharmacol 2008;155: Dempsie Y, Morecroft I, Welsh DJ, MacRitchie NA, Herold N, Loughlin L et al. Converging evidence in support of the serotonin hypothesis of dexfenfluramineinduced pulmonary hypertension with novel transgenic mice. Circulation 2008; 117: Morecroft I, Dempsie Y, Bader M, Walther DJ, Kotnik K, Loughlin L et al. Effect of tryptophan hydroxylase 1 deficiency on the development of hypoxia-induced pulmonary hypertension. Hypertension 2007;49: Marcos E, Fadel E, Sanchez O, Humbert M, Dartevelle P, Simonneau G et al. Serotonin-induced smooth muscle hyperplasia in various forms of human pulmonary hypertension. Circ Res 2004;94: Morecroft I, Heeley RP, Prentice HM, Kirk A, MacLean MR. 5-hydroxytryptamine receptors mediating contraction in human small muscular pulmonary arteries: importance of the 5-HT1B receptor. Br J Pharmacol 1999;128: Lawrie A, Spiekerkoetter E, Martinez EC, Ambartsumian N, Sheward WJ, MacLean MR et al. Interdependent serotonin transporter and receptor pathways regulate S100A4/Mts1, a gene associated with pulmonary vascular disease. Circ Res 2005;97: Liu YL, Suzuki YJ, Day RM, Fanburg BL. Rho kinase-induced nuclear translocation of ERK1/ERK2 in smooth muscle cell mitogenesis caused by serotonin. Circ Res 2004;95: MacLean MR, Deuchar GA, Hicks MN, Morecroft I, Shen SB, Sheward J et al. Overexpression of the 5-hydroxytryptamine transporter gene effect on pulmonary hemodynamics and hypoxia-induced pulmonary hypertension. Circulation 2004;109: Mair KM, MacLean MR, Morecroft I, Dempsie Y, Palmer TM. Novel interactions between the 5-HT transporter, 5-HT1B receptors and Rho kinase in vivo and in pulmonary fibroblasts. Br J Pharmacol 2008;155: Morecroft I, Loughlin L, Nilsen M, Colston J, Dempsie Y, Sheward J et al. Functional interactions between 5-hydroxytryptamine receptors and the serotonin transporter in pulmonary arteries. J Pharmacol Exp Ther 2005;313: Keegan A, Morecroft I, Smillie D, Hicks MN, MacLean MR. Contribution of the 5-HT1B receptor to hypoxia-induced pulmonary hypertension converging evidence using 5-HT1B-receptor knockout mice and the 5-HT1B/1D-receptor antagonist GR Circ Res 2001;89: Guignabert C, Izikki M, Tu LI, Li ZL, Zadigue P, Barlier-Mur AM et al. Transgenic mice overexpressing the 5-hydroxytryptamine transporter gene in smooth muscle develop pulmonary hypertension. Circ Res 2006;98: MacLean MR. Pulmonary hypertension, anorexigens and 5-HT: pharmacological synergism in action? Trends Pharmacol Sci 1999;20: MacLean MR, Morecroft I. Increased contractile response to 5- hydroxytryptamine(1)-receptor stimulation in pulmonary arteries from chronic hypoxic rats: role of pharmacological synergy. Br J Pharmacol 2001;134: Pullar IA, Boot JR, Carney SL, Cohen ML, Colvin EM, Conway RG et al. In vitro activity of LY393558, an inhibitor of the 5-hydroxytryptamine transporter with 5-HT1B/1D/2 receptor antagonist properties. Eur J Pharmacol 2001;432: Guignabert C, Raffestin B, Benferhat R, Raoul W, Zadigue P, Rideau D et al. Serotonin transporter inhibition prevents and reverses monocrotaline-induced pulmonary hypertension in rats. Circulation 2005;111: Damjanoska KJ, Van de Kar LD, Kindel GH, Zhang Y, D Souza DN, Garcia F et al. Chronic fluoxetine differentially affects 5-hydroxytryptamine2A receptor signaling in frontal cortex, oxytocin- and corticotropin-releasing factor-containing neurons in rat paraventricular nucleus. J Pharmacol Exp Ther 2003;306: Pacher P, Ungvari Z, Kecskemeti V, Koller A. Serotonin reuptake inhibitor, fluoxetine, dilates isolated skeletal muscle arterioles. Possible role of altered Ca2þ sensitivity. Br J Pharmacol 1999;127: Kong EK, Peng L, Chen Y, Yu AC, Hertz L. Up-regulation of 5-HT2B receptor density and receptor-mediated glycogenolysis in mouse astrocytes by long-term fluoxetine administration. Neurochem Res 2002;27: MacLean MR, Sweeney G, Baird M, McCulloch KM, Houslay M, Morecroft I. 5-Hydroxytryptamine receptors mediating vasoconstriction in pulmonary arteries from control and pulmonary hypertensive rats. Br J Pharmacol 1996;119: Eddahibi S, Humbert M, Fadel E, Raffestin B, Darmon M, Capron F et al. Serotonin transporter overexpression is responsible for pulmonary artery smooth muscle hyperplasia in primary pulmonary hypertension. J Clin Invest 2001;108: Eddahibi S, Chaouat A, Morrell N, Fadel E, Fuhrman C, Bugnet AS et al. Polymorphism of the serotonin transporter gene and pulmonary hypertension in chronic obstructive pulmonary disease. Circulation 2003;108: Willers ED, Newman JH, Loyd JE, Robbins IM, Wheeler LA, Prince MA et al. Serotonin transporter polymorphisms in familial and idiopathic pulmonary arterial hypertension. Am J Respir Crit Care Med 2006;173: Kawut SM, Horn EM, Berekashvili KK, Lederer DJ, Widlitz AC, Rosenzweig EB et al. Selective serotonin reuptake inhibitor use and outcomes in pulmonary arterial hypertension. Pulm Pharmacol Ther 2006;19: