Mechanism of hematocrit increase induced by the combined administration of morphine and Adriamycin: Role of histamine release

The mechanism by which morphine interacts with Adriamycin to increase hematocrit has been investigated in mice. Treatment with Adriamycin (12.8 mg/kg, iv) or morphine (20 mg/kg, sc) resulted in slight increases in hematocrit 30 min postdose, while animals given sc morphine 30 min prior to iv Adriamycin exhibited significant increases in hematocrit as early as 1 min post-Adriamycin. The hematocrit increase reached maximal levels 12 min post-Adriamycin and returned to basal values 4 hr postdose. Splenectomizing animals prior to morphine and Adriamycin treatment had no effect on the drug-induced hematocrit increase. This indicates that red cell release from splenic contracture is not the mechanism for the hematocrit increase. Measurement of plasma histamine levels following drug treatment demonstrated a marked and rapid rise in plasma histamine levels reaching maximal values 1 min post-Adriamycin. Adriamycin alone triggered this release; however, morphine pretreatment resulted in a higher maximum and more prolonged elevation of plasma histamine levels. Treatment with pyrilamine (3.1-50 mg/kg, ip) prior to morphine and Adriamycin administration partially reversed drug-induced hematocrit increase and protected against resultant lethality. Cimetidine (50-200 mg/kg, ip) treatment was not effective. The temporal relationship between hematocrit and histamine increases suggests a cause/effect relationship between released histamine and hematocrit elevation. Protection by pyrilamine and not cimetidine further supports this cause/effect relationship and indicates the effects is mediated via histamine type 1 receptors.     

Alterations in histamine levels in the rat induced by compound 48/80

Histamine release in the rat was induced in vivo either by a single dose of compound 48/80 injected i.v. or by four repeated, daily doses of the same compound injected i.p. After i.v. injection the levels of blood histamine were determined and after i.p. injections the changes in both tele-methylhistamine and histamine levels in different tissues were investigated. I.v. injection of 48/80 induced a very rapid and marked increase of blood histamine by 7.4 to 11-fold over the control levels within the first two minutes. After repeated i.p. injections of compound 48/80 most tissues showed higher than normal tele-methylhistamine/histamine ratios. The results suggest that agents known to induce release of histamine from mast cells may exert significant changes in blood and tissue histamine levels and that liberated histamine is thereafter extensively catabolized.     

Histamine release from isolated rat mast cells induced by opiates: effect of sterical configuration and calcium

The stereospecificity of the action of opiates on rat mast cells was investigated by means of the l- and d-isomers levorphanol and dextrorphan. The dose-response curves for histamine release induced by the 2 drugs were of a similar shape with a maximum at 2 X 10(-3) M and a pronounced minimum at 5 X 10(-3) M. At concentrations below 5 X 10(-3) M the effect of both drugs resembled that of morphine, i.e. the release occurred rapidly and inhibition was observed with naloxone, codeine, and antimycin A. Levorphanol, dextrorphan, and the antagonist levallorphan at concentrations above 5 X 10(-3) M seemed to be toxic to mast cells. The uptake of 45Ca in connection with histamine release induced by pethidine, levorphanol, and dextrorphan was lower than that of control cells, whereas the uptake induced by morphine did not differ from that of controls. The inhibition of compound 48/80-induced histamine release by morphine was paralleled by a reduced 45Ca uptake. The time course for the inhibitory effect of preincubation with morphine was similar for the histamine released induced by morphine and by compound 48/80. Washing of the cells after preincubation with morphine was without effect on the inhibition of morphine-induced histamine release, whereas the inhibition of compound 48/80 was reduced. The present observation with morphine and compound 48/80 support our previous impression of similarities in their mode of action, but a mechanism implying an interference by morphine with the disposition of calcium could also account for the findings. The observed antagonism between morphine and calcium resembles that of opiate receptors, but histamine release induced by opiates does not involve stereospecific opiate receptors.     

Antiarrhythmic effects of selective beta 1- and beta 2- and nonselective beta-adrenoceptor blockade in normokalaemic and dietary-induced hypokalaemic rats

Hypokalaemia (HK) is a risk factor for development of clinical arrhythmias, and epinephrine (Epi), released after myocardial infarction, may itself induce HK. The effects of selective beta 1- and beta 2-blockade with metoprolol (M) and ICI 118551 (I), respectively, and non-selective blockade with propranolol (P) against early ischaemia-induced arrhythmias were therefore compared in normokalaemic (NK) and dietary-induced HK rats. Plasma Epi, norepinephrine (NE), and K+ levels were also measured. All three blockers [5-10 mg/kg intravenously (i.v.)] were antifibrillatory in NK rats, whereas I and P (2-10 mg/kg) additionally reduced the number of ventricular premature beats. Coronary artery ligation increased plasma Epi levels in both NK an HK rats, but plasma NE increased further in HK rats. HK was also associated with an increased arrhythmia severity and negation of the antifibrillatory action of M. In contrast, I and P retained antiarrhythmic and antifibrillatory effects in HK animals and increased survival. Both I and P increased plasma K+ levels in both NK and HK animals whereas M did not. We concluded that beta 2-receptor-mediated HK after coronary occlusion may play a substantial role in arrhythmogenesis. Protection afforded by beta 2-blocking agents may be due to alleviation of HK in HK rats or to induction of hyperkalaemia in NK rats.     

Indirect adrenergic effect of compound 48/80 in cat cerebral arteries

Compound 48/80 evoked an increase in the spontaneous tritium release from cat cerebral arteries prelabelled with 3H-NA. This increase was abolished after cervical gangliectomy, external calcium removal or in the presence of colchicine. Cocaine brought about an enhancement in the release of radioactivity elicited by compound 48/80 while diphenhydramine did not affect it. These results suggest that compound 48/80 has an indirect adrenergic effect in cat cerebral arteries which is not mediated by the release of histamine from mast cells.     

Increase of plasma renin activity after subcutaneous application of compound 48/80 in the rat

Subcutaneous (s.c.) administration of compound 48/80 (3.0 mg/kg) to conscious rats produced a time-dependent long-lasting increase of plasma renin activity (PRA). A dose-related increase of the hematocrit was also observed after injection of compound 48/80. The onset of the hematocrit increase preceded that of PRA increase. Pretreatment with a dose of more than 20 mg/kg of histamine H1-receptor antagonists such as tripelennamine or diphenhydramine prior to the injection of compound 48/80 (3.0 mg/kg s.c.) attenuated or abolished the effects of compound 48/80 on PRA, hematocrit and plasma extravasation. Pretreatment with cimetidine (histamine H2-receptor antagonist, 40 mg/kg i.p.) had no effect on these plasma variables. The increase of PRA caused by s.c. administration of compound 48/80 was not affected by the pretreatment with propranolol (beta-adrenoceptor antagonist, 10 mg/kg i.p.), which completely inhibited the isoproterenol (0.5 mg/kg s.c.)-induced PRA increase. Administration of compound 48/80 did not induce a significant PRA increase in the nephrectomized rats although the increase of hematocrit following s.c. administration of compound 48/80 persisted despite the absence of kidneys. S.c. administration of compound 48/80 (3.0 mg/kg) led to a significant decrease of histamine content at the site of injection and to a significant increase in plasma histamine concentration without affecting arterial blood pressure. The present data suggest that s.c. administration of compound 48/80 stimulates the release of histamine from cutaneous mast cells, which cause an increase in vascular permeability to plasma protein via the stimulation of histamine H1-receptors, then leads to hypovolemia. The resulting hypovolemia may directly stimulate the juxtraglomerular cells of the kidney to release renin.     

Mast cells in citric acid-induced cough of guinea pigs

It was demonstrated previously that mast cells play an important role in citric acid (CA)-induced airway constriction. To investigate the role of mast cells in CA-induced cough, three experiments were carried out in this study. In the first experiment, 59 guinea pigs were employed and we used compound 48/80 to deplete mast cells, cromolyn sodium to stabilize mast cells, MK-886 to inhibit leukotriene synthesis, pyrilamine to antagonize histamine H(1) receptor, methysergide to antagonize serotonin receptor, and indomethacin to inhibit cyclooxygenase. In the second experiment, 56 compound 48/80-pretreated animals were divided into two parts; the first one was used to test the role of exogenous leukotriene (LT) C(4), while the second one to test the role of exogenous histamine in CA-induced cough. Each animal with one of the above pretreatments was exposed sequentially to saline (baseline) and CA (0.6 M) aerosol, each for 3 min. Then, cough was recorded for 12 min using a barometric body plethysmograph. In the third experiment, the activation of mast cells upon CA inhalation was investigated by determining arterial plasma histamine concentration in 17 animals. Exposure to CA induced a marked increase in cough number. Compound 48/80, cromolyn sodium, MK-886 and pyrilamine, but not indomethacin or methysergide, significantly attenuated CA-induced cough. Injection of LTC(4) or histamine caused a significant increase in CA-induced cough in compound 48/80-pretreated animals. In addition, CA inhalation caused significant increase in plasma histamine concentration, which was blocked by compound 48/80 pretreatment. These results suggest that mast cells play an important role in CA aerosol inhalation-induced cough via perhaps mediators LTs and histamine.     

Effect of nociceptin on histamine and serotonin release in the central nervous system

Role in pain sensation of both nociceptin (NC), the bioactive heptadecapeptide sequence of preproorphaninFQ and of histamine has been widely evidenced in the central nervous system (CNS). In the current series of experiments effect of intracerebroventricularly (i.c.v.) administered NC (5.5 nmol/rat) on histamine and serotonin levels in blood plasma, CSF and brain areas (hypothalamus and hippocampus) was studies and compared to the effect of the mast cell degranulator Compound 48/80(100microg/kg, i.c.v.) and the neuroactive peptide Substance P (50nmol/rat, i.c.v.). It was found that all the three compounds increased the histamine level in the CNS, however their activity concerning the mast cell-, and neuronal histamine release is different. NC could release histamine from both the mast cells and the neurons and it decreased CNS serotonin levels. Substance P was found the most potent in increasing CNS histamine levels. Compound 48/80 treatment resulted in elevated histamine levels both in the CNS and blood plasma. It is concluded that the histamine releasing effects of i.c.v. administered NC and SP are limited to the CNS, but in the effect of Compound 48/80 its blood-brain barrier impairing activity is also involved. Data also demonstrate that NC has significant effect on both the histaminergic and serotonergic neurotransmission in the CNS.     

A novel acute toxicity resulting from the administration of morphine and Adriamycin to mice

The toxic potential of the combination of morphine and Adriamycin (doxorubicin) has been evaluated in mice. Intravenous administration of 20 or 40 mg Adriamycin/kg resulted in a dose-related mortality beginning 4 days postdose. Increasing the dose of Adriamycin to 75 mg/kg caused a biphasic mortality pattern, with 30% of the animals dying within 30 min of drug treatment. Pretreatment with morphine (20 mg/kg) reduced by 80% the dose of Adriamycin required to induce a 30% mortality at 30 min post-Adriamycin administration. Morphine pretreatment also caused a dose-dependent increase in plasma Adriamycin, as measured by total plasma fluorescence. Morphine or Adriamycin administered alone resulted in a slight hematocrit increase. The combination of morphine and Adriamycin caused an increase in hematocrit to maximal levels of approximately 75% from basal levels of about 48%. The increase in hematocrit after morphine plus Adriamycin exceeded the rise caused by higher doses of Adriamycin which, without morphine, resulted in similar plasma Adriamycin levels. The temporal relationship between the elevated hematocrit and early (30 min) mortality suggest a cause/effect relationship between these two events following combined morphine plus Adriamycin treatment.     

What - if any - is the role of adrenergic mechanisms in histamine release from mast cells?

The effects of catecholamines on histamine release from rat peritoneal mast cells, was studied in an in vitro system. It was found that norepinephrine (10(-5)-101(-3) M) exerts a significant, dose related, repressive effect on compound 48/80-induced histamine release. This effect is greatly potentiated by beta-antagonists and is noticeable throughout the concentration range 10(-11)-10(-3) M norepinephrine. Phentolamine diminishes the repressive effect that norepinephrine shows at 10(-5) M. Norepinephrine (10(-5) M) totally inhibits the progressive histamine release induced by both compound 48/80 and strontium (10 muM) in non-Ca2+-depleted cells. The release that is dependent on extracellular calcium is inhibited by norepinephrine. The repressive effect of norepinephrine at 10(-3) is counteracted by 5.6 mM D-glucose, 2-deoxyglucose abolishes this effect. The repression of histamine release by 10(-5) M norepinephrine is not influenced by D-glucose. These results suggest that the effects on histamine release, observed within a low concentration range of norepinephrine (less than 10(-3) M), may be due to alpha-adrenoreceptor mechanisms and an interference in transmembrane calcium transport. Our data further suggest that norepinephrine at 10(-3) M may inhibit oxidative phosphorylation. Isoproterenol and epinephrine (10(-9)-10(-5) M) show little effect on 48/80-induced histamine release in a normal medium. However, when calcium is excluded from the medium, histamine release is potentiated. These results seem to indicate that isoproterenol and epinephrine act by displacing intracellular calcium, making it available for the exocytosis process.     

Prostaglandin E2 prevents diclofenac-induced enhancement of histamine release and inflammation evoked by in vivo challenge with compound 48/80 in the hamster cheek pouch

Based on observations obtained by the use of intravital microscopy, we report that prostaglandins (PGs) can exert inhibitory effects on mast cell-dependent inflammation. Thus, the PG-synthesis inhibitors diclofenac and indomethacin potentiated extravasation of plasma evoked by challenge with the mast cell secretagogue compound 48/80. Although the plasma leakage induced by compound 48/80 was in large mediated by histamine, neither diclofenac nor indomethacin potentiated the plasma leakage caused by exogenous histamine. These findings indicated that endogenous PGs inhibited the mast cell-dependent reaction at the level of mediator release. This mode of action was confirmed, as diclofenac was found to enhance the in vivo release of histamine that ensued challenge with compound 48/80. Moreover, the enhancement of the response to compound 48/80 observed after diclofenac treatment was prevented by local administration of PGE2 (30 nM). This inhibition included both the histamine release and the plasma leakage. In addition, diclofenac enhanced the leukocyte emigration after compound 48/80 challenge, and PGE2 reversed also this effect, suggesting that endogenous PGs (e.g. PGE2) also inhibited the release of chemotactic mediators.     

Plasma membrane fluidity studies by fluorescence polarization in rat mast cells stimulated by compound 48/80

Plasma membrane fluidity measurements were performed on purified living mast cells using a novel non-permeant fluorescence polarization probe TMA-DPH, upon stimulation by compound 48/80. The fluorescence anisotropy increased rapidly after treatment by 48/80 in a dose-dependent way. The effect was found to be specific for mast cells; it was inhibited by the histamine release antagonist FR 7534 in a correlative manner. The role of calcium was examined. The results brought evidence for a plasma membrane fluidity decrease induced by 48/80; a biphasic mechanism was inferred for the histamine release process.     

Circulatory depression and ventricular arrhythmias induced by compound 48/80 in anaesthetized rats.

The effects of graded doses of compound 48/80 on various cardiovascular and respiratory parameters were studied in pentobarbitone-anaesthetized rats. Following intravenous injections, this compound significantly depressed the mean blood pressure (MBP), left ventricular pressure (LVP) and dLVP/dtmax, and caused ventricular tachycardia (VT) or fibrillation (VF) and death. Heart rate (HR) response were variable, and there were no marked changes in airway resistance or blood gases. Pretreatment of the animals with either cimetidine or diphenhydramine significantly prolonged the time of onset of VT/VF but failed to alter the changes in other circulatory variables. A combination of cimetidine and diphenhydramine significantly alleviated the decreases in MBP and LVP and prevented the occurrence of VT/VF. It is suggested that the circulatory depression and the occurrence of ventricular arrhythmias following the administration of compound 48/80 result from activation of H1- and H2-receptors by elevated blood histamine levels due to release of the amine from tissues.     

Mechanism of ionophore A23187 induction of plasma protein leakage and of its inhibition by indomethacin

Calcium ionophore A23187 produced a dose-dependent increase in plasma protein leakage on intradermal injection in rats. Studies with mepyramine and cyproheptadine indicated that histamine and 5-hydroxytryptamine (5-HT) partially contribute to the ionophore action and experiments with compound 48/80 supported these findings. Depletion of plasma kininogen levels with cellulose sulphate administered indomethacin inhibited the ionophore response in a dose-dependent manner. The inhibition was not reversed by intradermally injected prostaglandin E2 (PGE2) in doses up to 50 ng/site, suggesting that PGE2 also may not be an important mediator. It is proposed that the ionophore produces plasma protein leakage by an indirect (through histamine and 5-HT release) and a direct action on vascular endothelial cells and that indomethacin antagonises both actions by inhibiting calcium transport processes.     

Chronic adriamycin treatment impairs myocardial interstitial neuronal release of norepinephrine and epinephrine

Although chronic adriamycin (doxorubicin) treatment is known to induce cardiomyopathic heart failure with sympathetic neurohumoral activation in a dose-dependent manner, its effect on local neuronal catecholamine release at the cardiac sympathetic nerve terminals remains to be clearly determined. Using a cardiac microdialysis technique, we measured dialysate norepinephrine (NE) and epinephrine (Epi) concentrations as indices of myocardial interstitial NE and Epi levels. respectively, in rabbits with chronic adriamycin treatment (ADR) (4 mg/kg/week, 6 weeks, n = 8) and in control rabbits (CNT) (n = 6). Exocytotic release was evoked by the local administration of KCl (100 mM) through the dialysis probe. Basal levels of NE and Epi did not differ between the ADR and CNT groups (NE, 11.6 +/- 6.6 vs. 20.4 +/- 17.2 pg/ml; Epi, 4.0 +/- 0.1 vs. 4.6 +/- 1.7 pg/ml: mean +/- SD). The exocytotic release was suppressed in the ADR compared with the CNT group (NE, 191.4 +/- 144.7 vs. 760.5 +/- 337.8 pg/ml; p < 0.05: Epi, 4.2 +/- 0.4 vs. 20.8 +/- 9.9 pg/ml; p < 0.05). We conclude that chronic adriamycin treatment impairs the neuronal exocytotic release of catecholamine at the cardiac sympathetic nerve terminals.     

Cardiac 3,4-dihydroxyphenylethylene glycol (DHPG) and catecholamine levels in a rat model of left ventricular failure

Cardiac norepinephrine (NE), dopamine (DA), epinephrine (Epi), and dihydroxyphenylethylene glycol (DHPG) (a major neuronal metabolite of NE) content were measured in rats with cardiac failure resulting from left ventricular myocardial infarction (LVMI) induced by coronary ligation. The ratio of DHPG/NE was significantly higher in both the right ventricle and interventricular septum of rats with LVMI compared with controls, reflecting a tendency for cardiac DHPG content to rise and NE content to fall during cardiac failure. Cardiac DA and Epi content did not significantly differ between rats with LVMI and controls. Elevated DHPG/NE ratios apparently reflected the increase in NE turnover that accompanies elevated sympathetic activity in heart failure more precisely than changes in NE levels or DHPG levels alone. Furthermore, DHPG/NE ratios are not influenced by increases in cardiac weight due to cardiac hypertrophy. The DHPG/NE ratios may be a useful index of cardiac sympathetic activity for future studies of the effects of drug treatment of cardiac failure in this animal model.     

A method for determining whether hypotension caused by novel compounds in preclinical development results from histamine release

INTRODUCTION: Many therapeutic agents stimulate histamine release from mast cells, which results in a decrease in blood pressure. The purpose of this study is to establish a method to determine if the mechanism of action, or one of the mechanisms, of hypotensive compounds is related to the release of histamine. The method was developed using a novel hypotensive compound, SC-372. METHODS: In Inactin anesthetized rats, after intravenous administration of SC-372 (0.3-7 mg/kg), the 2 and 7 mg/kg resulted in a dose-dependent decrease in blood pressure. Histamine (0.1 and 1 mg/kg) was injected intravenously to establish whether histamine release was the mechanism of action for the hypotension induced by SC-372. Compound 48/80 (0.1 mg/kg, promotes histamine release) and Cromolyn (1 mg/kg/min, [5 min], prevents histamine release from mast cells) were characterized and used intravenously in combination with/or compared to SC-372. RESULTS: Histamine resulted in a decrease in blood pressure that was unaffected by Cromolyn (1 mg/kg). Administration of Compound 48/80 resulted in a rapid reduction of systemic blood pressure. Intravenous infusion of Cromolyn prior to the injection of Compound 48/80 significantly attentuated the hypotensive response and the increase in histamine levels in the plasma. Intravenous administration of SC-372 resulted in a rapid reduction in blood pressure with a profile similar to that of Compound 48/80. When the rats were treated with Cromolyn prior to the administration of SC-372, both the blood pressure and plasma histamine levels were maintained at their pretreatment control levels. DISCUSSION: These data indicate that Compound 48/80 and Cromolyn can be used in rats to screen for histamine release-dependent drug-induced hypotension and suggest that the rapid decrease in blood pressure caused by SC-372 may result from histamine release from mast cells.     

Effects of psychoactive drugs on plasma catecholamines during stress in rats

In unstressed rats, morphine, pentobarbital, diazepam and alprazolam were without effect on plasma norepinephrine (NE) and epinephrine (E) levels. Amphetamine increased the levels of both catecholamines. In stressed rats, morphine did not affect the stress induced increases in NE and E. Amphetamine enhanced the catecholamine stress-response further. Diazepam reduced stress-induced increases of NE levels. Pentobarbital and alprazolam attenuated stress-induced increases of both NE and E levels; this effect was particularly marked with alprazolam.The rise in plasma norepinephrine (NE) and epinephrine (E) levels during stress has been well documented (DeTurck and Vogel, 1982; McCarty and Kopin, 1979). Considerably less is known about the effects of drugs on these stress-induced increases in biogenic amine levels. In rats, ganglionic blockade by chlorisondamine and blockade of NE release by bretylium has been shown to decrease or reverse stress-induced increases in plasma NE and/or E (McCarty and Kopin, 1979). These effects are not surprising, due to the known action of these drugs on the autonomic nervous system. However, drugs unrelated to the autonomic nervous system can also affect the biogenic amine levels during stress. Fentanyl-oxygen anesthesia reduces plasma NE and E in man during surgery (Stanley, Berman, Green and Robertson, 1980), and diazepam reduces NE levels during third molar extraction without affecting E level (Goldstein, Dionne, Sweet, Gracely, Brewer, Gregg and Keiser, 1982). Recently, we observed that ethanol can reduce stress-induced increases in rat plasma levels of both NE and E (DeTurck and Vogel, 1982). In that case, it has been speculated that this reduction is due to the sedative or anxiolytic properties of alcohol (DeTurck and Vogel, 1982). To test the possibility that these properties can indeed affect stress-induced increases in NE and E, we measured the effects of several psychoactive drugs on plasma catecholamine levels during stress. The drugs were the stimulant amphetamine, the sedative drugs morphine and pentobarbital, and the anxiolytic/sedative drugs diazepam and alprazolam.     

Nadroparine inhibits the hypersensitivity response in the conjunctiva

This study sought to investigate the effects of nadroparine on an in vivo experimental model of type I hypersensitivity response in the rat conjunctiva. Following drug application onto the eye, either before or after challenge with the mast cell degranulator, basic polyamine compound 48/80, the conjunctival histamine content and the nitrite levels in the conjunctival lavage fluid were quantified fluorometrically and spectrophotometrically, respectively. Instillation into the eye of nadroparine inhibited the C48/80-induced decreases in conjunctival histamine and the delayed increases in nitrite levels, without influencing basal mediator levels. Protamine did not induce histamine release and only partially reversed the effects of nadroparine post-challenge, yet it had no effect on the protective action of the drug when administered prior to degranulation. The results showed that nadroparine was equally effective in attenuating the effects of compound 48/80 in the eye when administered topically either before or after challenge.     

Antipruritic and antierythema effects of ascomycete Bulgaria inquinans extract in ICR Mice

The effect of ethanol extract obtained from Bulgaria inquinans on the scratching behavior and vascular permeability changes induced by compound 48/80, histamine and serotonin in ICR mice was studied. The extract dose-dependently inhibited scratching behavior induced by compound 48/80 and serotonin. A significant inhibition was observed at doses of 300 and 600 mg/kg when Bulgaria inquinans extract was administered orally. However, no inhibitory effect was observed on the histamine-induced scratching behavior by the extract, even at a dose of 600 mg/kg. In addition, it also inhibited the increase in the vascular permeability induced by compound 48/80 and serotonin at doses of 300 and 600 mg/kg; however, it failed to inhibit the increased vascular permeability induced by histamine, even at a dose of 600 mg/kg. Bulgaria inquinans extract showed a potent inhibitory effect on histamine release induced by compound 48/80. These results suggest that Bulgaria inquinans extract is effective in cutaneous pruritus and erythema, which were probably mediated by inhibiting the release of histamine from mast cells and antagonizing the effect on serotonin.