Differential response of rapid eye movement sleep to cholinergic blockade by scopolamine in currently depressed, remitted, and normal control subjects

The degree of cholinergic dysregulation of sleep in adult depression was evaluated using scopolamine. On separate sessions, placebo and scopolamine (4.5 micrograms/kg, IM) were administered to 14 patients with unipolar major depression, 16 recovered/remitted patients, and 18 normal controls. Scopolamine increased rapid eye movement (REM) latency (RL), reduced REM activity (RA), REM density (RD), and REM duration, and increased the percentage of stage 4 sleep in all groups. There was a differential effect of scopolamine on RL, RA, and REM duration for the first REM period, and on percentage of stage 4 sleep. Whereas a primary cholinergic hyperactivity could account for the RA and RD responses, the response profile for RL was more compatible with reduced aminergic tone as the proximal cause of the cholinergic hyperactivity. Whether the sleep abnormalities observed in remitted patients reflect an underlying vulnerability for development or recurrence of depression, and/or a scar, remains to be determined.

The pharmacokinetics and pharmacodynamics of adinazolam: multi-ethnic comparisons

The pharmacokinetics and pharmacodynamics of adinazolam and N-demethyladinazolam (NDMAD), its major active metabolite, were compared in 39 healthy male volunteers (13 Asian, 12 Caucasian and 14 African-American). In a four-way, double-blind crossover design, subjects were administered (1) 30 mg oral adinazolam mesylate SR tablets, (2) 10 mg parenteral (i.v.) adinazolam mesylate, (3) 30 mg i.v. NDMAD and (4) placebo. Venous blood samples were collected at specific time intervals after drug administration and assayed for adinazolam and NDMAD concentrations. Sedation was rated at the time of each blood draw according to the Nurse-Rated Sedation Scale, and the digit-symbol substitution test was administered to evaluate psychomotor performance. After i.v. administration of adinazolam, Asians manifested significantly higher Cmax, larger AUC and lower CL of both adinazolam and NDMAD than their Caucasian and African-American counterparts. Likewise, after i.v. NDMAD Asians had significantly higher NDMAD Cmax and AUC than Caucasians and African-Americans. Most of these differences remained statistically significant after controlling for body surface area. With PO adinazolam, Asians also manifested substantially higher Cmax, larger AUC and lower CL for both adinazolam and NDMAD; however, with the exception of Cmax, these differences did not reach statistical significance. These results are in accordance with previous observations for ethnic-related differences in drug pharmacokinetics. In contrast, pharmacodynamic differences were not noted among the three study groups.

CSF testosterone and 5-HIAA correlate with different types of aggressive behaviors

We studied the potential roles of testosterone and serotonin in various forms of aggressive and violent behaviors by measuring each biochemical and behaviour in free-ranging adolescent male nonhuman primates. Our results showed that (1) CSF free testosterone concentrations were positively correlated with overall aggressiveness, but not with measures of impulsivity. (2) CSF 5-HIAA concentrations were negatively correlated with impulsive behavior, and severe, unrestrained aggression, but not with overall rates of aggression. High rates of impulsive behavior were positively correlated with severe, unrestrained aggression, but not overall rates of aggression. (3) Dimensional analyses showed that while subjects with low CSF 5-HIAA exhibited high rates of aggression, high CSF testosterone further augmented rates and intensity of aggression in subjects with low CSF 5-HIAA. We conclude that high CSF free testosterone concentrations are associated with competitive aggression, while low CSF 5-HIAA concentrations are associated with severe aggression which results from impaired impulse control, and perseverance.

In vivo proton magnetic resonance spectroscopy of the normal aging human brain

The effect of age on brain metabolite concentrations was evaluated using localized proton magnetic resonance spectroscopy. This technique allows in vivo measurements of N-acetyl compounds (NA), total creatine (CR), choline-containing compounds (CHO), myo-inositol (MI), glutamate and glutamine (GLX), as well as the percentage of cerebrospinal fluid (CSF) and the brain water content within the brain region studied. Frontal gray matter and frontal white matter brain regions were examined in 36 normal healthy volunteers (19-78 years of age). Using a rigorous absolute quantitation method, with an external reference and atrophy correction, we found relatively stable concentrations of NA, a neuronal marker. In contrast, CR, CHO, MI, and the percentage of CSF increased in the gray matter with age. However, the brain water content decreased significantly with age (r = -0.72; p < 0.0001). No significant age-related changes in metabolite concentrations, CSF or brain water content were observed in the white matter regions. These findings demonstrate that biochemical alterations are associated with aging in the frontal gray matter. There might be an increase in the brain density as indicated by increased metabolite concentrations and decreased brain water content with aging.

Psychobiologic effects of 3,4-methylenedioxymethamphetamine in humans: methodological considerations and preliminary observations

3,4-Methylenedioxymethamphetamine (MDMA) is a phenethylamine with potent effects on serotonergic neurotransmission which has been the object of controversy over its potential as a therapeutic adjunct versus its possible risks for causing neurotoxic injury. This paper discusses the background, methodology and preliminary findings of the first FDA approved Phase I study prospectively evaluating the effects of MDMA administration in humans. Six subjects with prior experience with MDMA were administered two different dosages of MDMA and an inactive placebo utilizing a randomized, double-blind methodologic design. Dosages from 0.25 to 1.0 mg/kg, p.o., were administered. All subjects tolerated the procedures without any overt evidence of physical discomfort or psychological distress. MDMA produced a modest increase in heart rate and blood pressure. The threshold dose for the stimulation of ACTH and prolactin appeared to be between 0.5 and 0.75 mg/kg, with the two higher doses clearly stimulating both ACTH and prolactin. Methodology for assessing MDMA's effects on serotonergic neurotransmission is discussed.

Prenatal stress prevents the desensitization of the corticosterone response to TFMPP by desmethylimipramine, but not by phenelzine, in adult male offspring

Gravid female rats were subjected to one hour of restraint stress twice daily or left undisturbed from days 14-21 of gestation. Adult 105-day old male non-stressed (NS) and stressed (S) offspring were treated once daily with saline, desipramine (DMI) (10 mg/kg, sc) or phenelzine (5.0 mg/kg, sc) for 14 days. Twenty-four hours after the last injection, animals were challenged with saline or 1-(m-trifluoromethylphenyl)piperazine (TFMPP) (5.0 mg/kg, sc), a serotonin1B/2C (5-HT1B/2C) agonist, and plasma prolactin and corticosterone concentrations were measured one hour later. As compared to acute saline administration, TFMPP significantly increased prolactin and corticosterone concentrations in all groups. In NS offspring, both DMI and phenelzine treatment augmented the prolactin response, but blunted the corticosterone response, to TFMPP. In S offspring, the prolactin response to TFMPP also was augmented by phenelzine or DMI treatment, whereas the corticosterone response to TFMPP was blunted during phenelzine treatment. However, DMI treatment was not able to desensitize the corticosterone response to TFMPP in the S rats. The results indicate the adaptive capacity of 5-HT systems to DMI administration was compromised in adult animals exposed to stress in utero.

Ethnicity and psychopharmacology. Bridging the gap

Taken together, the literature reviewed clearly indicates that the disposition and effect of a large number of psychotropic agents are influenced substantially by ethnicity and culture. Recent advances in the realm of pharmacokinetics, pharmacogenetics, and pharmacodynamics have led to a greater understanding of some of the mechanisms responsible for such differences. In comparison, much less currently is known regarding how various psychosocial factors impinge on drug responses in different cultural settings. Progress in research in this area is important for clinical reasons, as psychiatric clinicians will increasingly be confronted with patients with divergent ethnic and cultural backgrounds. In addition, knowledge derived from such research will contribute significantly to a better understanding of how the effects of psychotropic agents are mediated, and also should be valuable for the drug development industry that will have to take into account the increasingly diversifying domestic and international markets.

Relationship of 24-hour urinary free cortisol to 4-hour salivary morning and afternoon cortisol and cortisone as measured by a time-integrated oral diffusion sink

The relationship between salivary corticosteroids integrated over 4-hour periods and urinary free cortisol collected over 24 hours was investigated in normal controls. Twenty-one normal volunteers wore "oral diffusion sink" sampling devices in their mouths for two 4-hour periods (08:00-12:00 hours and 13:00-17:00 hours) and on the same day collected a 24-hour urine specimen. Time-integrated salivary corticosteroid concentrations were determined from the sample devices and urinary free cortisol was measured. Salivary corticosteroids were not consistently higher in the morning than in the afternoon period and did not differ between men and women. Urinary free cortisol levels were higher in women. No salivary corticosteroids measure was significantly correlated with urinary free cortisol. We conclude that time-integrated salivary corticosteroids do not reflect urinary free cortisol levels in normal controls.

Pretreatment with MK-801 inhibits pemoline-induced self-biting behavior in prepubertal rats

The indirect dopamine agonist, pemoline (120-300 mg/kg s.c.), can induce self-biting behavior in the rat. The present study demonstrates that the non-competitive N-methyl-D-aspartate (NMDA) antagonist, dizocilpine (MK-801, 0.2 mg/kg s.c.), significantly attenuates pemoline-induced self-biting behavior, while simultaneously increasing locomotor activity. When animals received a fixed dose of MK-801 with increasing doses of pemoline, a competitive relationship emerged such that high-dose pemoline surmounted the antagonistic effect of MK-801. In contrast to spiperone, delayed administration of MK-801 was ineffective in blocking the subsequent expression of self-biting behavior, suggesting that dizocilpine exerts its protective effect early in the cascade of events which eventually leads to self-biting behavior in this paradigm.

Exposure to threshold doses of nicotine in utero: II. Neuroendocrine response to nicotine in adult male offspring

Groups of gravid female rats were injected subcutaneously with saline (SAL), a low-dose of nicotine (LN) (0.05 mg/kg, bid) or a high-dose of nicotine (HN) (3.0 mg/kg, bid) from day 4 to day 20 of gestation, or were left undisturbed. In adult 120-day-old male offspring, the ACTH and prolactin responses to acute nicotine challenge were evaluated. The experiment was performed on three separate occasions. Based upon dose-response and time-course studies with nicotine in normal animals, the neuroendocrine responses to nicotine (0.75 and 1.0 mg/kg, sc) were measured 7.5 min after nicotine administration, the peak response-time for both hormones. The ACTH response to acute nicotine administration was blunted significantly in the HN rats, but normal in the LN rats, for all three experiments. In two experiments, the prolactin response to acute nicotine administration was blunted significantly in the HN rats, but enhanced significantly in the LN offspring. The results indicate that prenatal nicotine administration can produce long-term neuroendocrine effects involving nicotinic-receptor coupled circuits, with long-term functional sequelae produced by dosages of nicotine considerably smaller than previously shown to be pharmacologically/toxicologically active.

Neuroendocrine responses produced by enantiomeric pairs of drugs that interact with phencyclidine and sigma receptors

The present study characterized the response of the hypothalamo-pituitary-adrenal axis after the acute administration of enantiomeric pairs of drugs that bind to phencyclidine (PCP) and sigma receptors. Rats were injected with the enantiomers of 1-(1-phenylcyclohexyl)-3-methylpiperidine (PCMP), N-allylnormetazocine (SKF 10,047), dioxadrol (dexoxadrol and levoxadrol) or pentazocine, and plasma levels of adrenocorticotropin (ACTH) and corticosterone were determined by radioimmunoassay. The effects of the enantiomers of PCMP and dioxadrol showed stereospecificity as both (+)-PCMP and dexoxadrol increased plasma levels of ACTH and corticosterone but (-)-PCMP and levoxadrol had no effect. Whereas (-)-pentazocine produced greater responses than (+)-pentazocine, the two enantiomers of SKF 10,047 did not show stereoselectivity. Although the potency of the enantiomers of PCMP and dioxadrol parallel their affinity for binding to PCP receptors, the potency of the enantiomers of pentazocine did not. These results suggest that although the stimulation of the hypothalamo-pituitary-adrenal axis by PCP and drugs with PCP-like activity might be due to interactions with PCP receptors, the effects of pentazocine also involve interactions at other sites.

Dissociation between plasma bioactive and immunoactive ACTH concentrations in depressed patients

Previous studies have reported dissociations between plasma cortisol and immunoactive adrenocorticotropic hormone (ACTH) concentrations in both normal controls and in patients with major depression. In order to investigate this issue further, placebo and dexamethasone (DEX) were administered to normal controls and depressed patients at 11 PM, and plasma cortisol and ACTH were measured the following morning at 7 AM. Plasma ACTH concentrations were quantitated by both immunoassay (I-ACTH) and by bioassay (B-ACTH). In 10 normal controls, DEX (0.25, 0.5, and 1.0 mg, PO, elixir) produced a dose-related suppression of cortisol, I-ACTH and B-ACTH, with all three hormones significantly suppressed by DEX (0.5 and 1.0 mg) (p < or = 0.01). In 20 depressed patients, 7 AM plasma ACTH and cortisol concentrations were assessed following a single dose of DEX (0.5 mg). Fifteen patients were classified as suppressors and five as escapers, as reflected by mean (+/- SEM) cortisol concentration of 19.9 +/- 3.0 ng/ml and 81.2 +/- 7.0 ng/ml, respectively. Mean I-ACTH concentrations were comparable in both the escapers (8.6 +/- 1.6 pg/ml) and in the suppressors (7.0 +/- 1.0 pg/ml). In contrast, the mean B-ACTH concentration was more than two-fold higher in the escapers (4.5 +/- 0.5 pg/ml) than in the suppressors (2.2 +/- 0.3 pg/ml) (p < or = 0.001). Eleven of the 20 patients received both placebo and DEX (0.5 mg) on two separate occasions. Although DEX significantly suppressed both cortisol (p < or = 0.0001) and B-ACTH (p < or = 0.01) concentrations, I-ACTH was not significantly reduced.(ABSTRACT TRUNCATED AT 250 WORDS)

Exposure to threshold doses of nicotine in utero: I. Neuroendocrine response to restraint stress in adult male offspring

Gravid female rats were injected subcutaneously with saline or nicotine (3.0 mg/kg and 0.05 mg/kg, bid) from day 4 to day 20 of gestation or were left undisturbed. In adult 120-day old male offspring, the ACTH, corticosterone and prolactin concentrations before, during (15, 30, 45 and 60 minutes) and after (30, 60, 90 and 120 minutes) one hour of restraint stress were studied. Baseline (non-stress) concentrations of each hormone were comparable among the groups. As compared to saline controls, ACTH concentrations were significantly higher during stress at 30 and 60 minutes in the high-dose nicotine (HN) animals, with the average ACTH concentration during stress also being significantly higher in the HN rats. Neither nicotine regimen affected the corticosterone response to stress at any time-point. The prolactin response to stress was significantly reduced in the HN group at 45 and 60 minutes as compared to saline controls, with the average prolactin concentration also reduced during stress. During recovery, average ACTH concentrations were significantly higher in the HN group, and significantly lower in the LN group, with no differences found for either corticosterone or prolactin. The results indicate that exposure to a high-dose of nicotine during gestation, and to a very low-dose as well, produced functional alterations in adult male offspring as manifested by abnormal neuroendocrine responses to restraint stress. However, the differences between the nicotine and saline controls were sometimes as great as between the non-injected controls and the saline controls. Thus, any conclusions drawn about the long-term effects of prenatal nicotine on neuroendocrine responsivity to stress must be tempered by the influence of the repeated injection procedure.

CSF biochemistries, glucose metabolism, and diurnal activity rhythms in alcoholic, violent offenders, fire setters, and healthy volunteers

BACKGROUND

There is an extensive literature describing a central serotonin deficit in alcoholic, impulsive, violent offenders and fire setters. In the present study, we investigated biochemical concomitants of impulsivity and aggressiveness, and the physiological consequences of reduced central serotonin turnover.

METHODS

Forty-three impulsive and 15 nonimpulsive alcoholic offenders and 21 healthy volunteers were studied in the forensic psychiatry ward of a university psychiatric department. The subjects underwent lumbar punctures and oral glucose and aspartame challenges, and their diurnal activity rhythm was measured with physical activity monitors. Discriminant function analyses were used to investigate psychophysiological and biochemical concomitants of aggressive and impulsive behaviors.

RESULTS

Alcoholic, impulsive offenders with antisocial personality disorder had low mean cerebrospinal fluid (CSF) 5-hydroxyindoleacetic acid (5-HIAA) and corticotropin levels and high mean CSF testosterone concentrations. Compared with healthy volunteers, they showed increased physical activity during the daytime. Alcoholic, impulsive offenders with intermittent explosive disorder had a low mean CSF 5-HIAA concentration and blood glucose nadir after an oral glucose challenge, and desynchronized diurnal activity rhythm. Healthy volunteers had mean CSF 5-HIAA concentrations that were intermediate between those of alcoholic, impulsive and nonimpulsive offenders. Alcoholic, nonimpulsive offenders had a significantly higher mean CSF 5-HIAA concentration than all the other groups, including healthy volunteers.

CONCLUSIONS

In the present sample, a low CSF 5-HIAA concentration was primarily associated with impulsivity and high CSF testosterone concentration, with aggressiveness or interpersonal violence.

Personality profiles and state aggressiveness in Finnish alcoholic, violent offenders, fire setters, and healthy volunteers

BACKGROUND

Based on clinical observations in a series of studies on Finnish alcoholic, violent offenders, we asserted that the impulsive offenders represented an extreme group of type 2 alcoholics. We also observed that these subjects were vulnerable to hypoglycemia after the administration of oral glucose load. Furthermore, we believe that while being hypoglycemic, the impulsive offenders are particularly irritable and aggressive. In the present study, we addressed these issues by studying psychological trait and state variables in a new group of violent offenders and fire setters, and age- and sex-matched healthy volunteers.

METHODS

Fifty-eight alcoholic, violent offenders and impulsive fire setters and 21 healthy volunteers were administered the Karolinska scales of personality and the Rosenzweig picture frustration test after an oral aspartame and glucose challenge.

RESULTS

The psychological test results and the criminal histories of the offenders, together with biochemical measurements, suggest that a low 5-hydroxyindoleacetic acid concentration in cerebrospinal fluid in the alcoholic offenders is associated with irritability and impaired impulse control, and a high free testosterone concentration in cerebrospinal fluid is associated with increased aggressiveness, monotony avoidance, sensation seeking, suspiciousness, and reduced socialization.

CONCLUSION

Finnish alcoholic, impulsive offenders have personality profiles characteristic of Scandinavian early-onset male alcoholics with antisocial traits, who have been also referred to as type 2 alcoholics.

Low-dose naltrexone inhibits pemoline-induced self-biting behavior in prepubertal rats

ABSTRACT Pemoline-induced self-biting behavior has been compared to similar self-injurious behavior (SIB) that occurs in the context of some mental retardation syndromes. The opiate antagonist, naltrexone, has been used successfully in the treatment of SIB in individuals with autism or mental retardation. This is the first report of the effect of naltrexone in an animal model of self-biting behavior. Naltrexone (0.01 mg/kg s.c.) significantly reduced the severity of self-biting behavior, but higher doses (0.10-10 mg/kg s.c.) had no such effect. Consistent with these results in the rat, a review of the clinical use of naltrexone in SIB is also suggestive of a relative loss of efficacy at higher doses (apparent therapeutic window). Naltrexone has higher affinity for mu receptors than other opioid receptor subtypes. Thus, the effect of naltrexone in treating SIB appears to be due to (mu-receptor interactions, with the loss of effect at higher doses due to supervening kappa receptor binding. Speculatively, one may ask if dosage reduction would convert naltrexone nonresponders to responders by lowering kappa receptor-mediated effects. Given the co-localization of opioids and dopamine inputs in some neurons, opioid-dopamine system interactions may also be important in pemoline-induced self-biting behavior.

Effects of low-dose dexamethasone on sleep EEG patterns, plasma cortisol, and the TSH response to TRH in major depression

Dexamethasone (DEX) (0.5 mg, P.O.) and placebo were administered at 2300 h in randomized design to 19 patients with major depression and the effects on the sleep electroencephalogram (EEG) were studied. In addition, the thyroid stimulating hormone (TSH) response to thyrotropin releasing hormone (TRH) and basal plasma cortisol concentrations were assessed the following morning. DEX did not affect sleep architecture or continuity variables, including rapid eye movement (REM) latency, REM activity and REM density. Similarly, DEX did not significantly influence the TSH response to TRH (delta max TSH). In contrast, plasma cortisol concentrations were significantly suppressed by DEX. The results indicate that, as opposed to higher dosages of glucocorticoids, 0.5 mg DEX had minimal effects on the sleep EEG or delta max TSH in depressed patients.

Haloperidol plasma levels and clinical response: a therapeutic window relationship

OBJECTIVE

The purpose of the study was to assess the relationship between plasma haloperidol and clinical response.

METHOD

Sixty-nine newly admitted drug-free schizophrenic men were randomly assigned to receive haloperidol, 5, 10, or 20 mg daily for 4 weeks, and clinical response was measured at the end of the fixed-dose period. Haloperidol was assayed by a sensitive and specific radioimmunoassay.

RESULTS

The authors found a curvilinear relationship between clinical response and plasma haloperidol during fixed-dose treatment, with an apparent optimum between 5 and 12 ng/ml. When plasma levels above 12 ng/ml were lowered to the 5-12 ng/ml range, all patients improved to varying degrees and no patient deteriorated. When plasma levels of nonresponders within this therapeutic window were raised above 12 ng/ml (as in routine practice), they, on balance, deteriorated in that they became more dysphoric. With the 20-mg dose, half the patients had plasma levels above 12 ng/ml.

CONCLUSIONS

In this sample of newly admitted schizophrenic men, optimal clinical response occurred with a plasma haloperidol range of 5-12 ng/ml.

Relationship between REM sleep latency and nocturnal cortisol concentrations in depressed patients

Due to conflicting reports on the possible association between shortening of rapid eye movement (REM) latency and increased cortisol secretion in patients with severe depression, this study examined the relationship between REM sleep latency and nocturnal cortisol concentration in 12 outpatients with major depression. The results showed a significant inverse correlation (r = -0.71, P < 0.01) between REM sleep latency and mean (23.00 hours-03.00 hours) plasma cortisol concentration. Age and severity of depression did not contribute to the inverse relationship. REM activity and density during the first REM period showed no significant correlations with the cortisol measures. A review of the literature suggests that this relationship might be unique to subjects with major depression, and again raises the possibility that these biological disruptions may have a common neurochemical basis.

Ethnicity and family involvement in the treatment of schizophrenic patients

In a cross-ethnic study of neuroleptic response, the level of family involvement with the treatment of 26 Asian and 26 Caucasian patients was evaluated. Asian family members were intimately involved, Caucasians much less so. These results quantitatively demonstrated the relatively greater importance of working with family members when treating Asian patients. They also indicate that relatively more effort is needed to increase involvement of Caucasian families in the treatment process.

Dose-dependent effects of scopolamine on nocturnal growth hormone secretion in normal adult men: relation to delta-sleep changes

To explore the sensitivity of nocturnal GH secretion to different degrees of cholinergic blockade, we investigated the effects of two doses of the muscarinic receptor antagonist scopolamine (SCOP; 3.0 and 6.0 micrograms/kg, im) and placebo, administered in a randomized fashion at 2300 h on three nights to eight normal male volunteers. Both doses of SCOP produced significant reductions in mean nocturnal GH concentration compared to the effects of the placebo; the higher dose of SCOP reduced GH to a greater degree than the lower dose, but this difference was not statistically significant (mean, 2.3 micrograms/L after 6 micrograms/kg vs. 3.0 micrograms/L after 3 micrograms/kg). Both SCOP doses significantly shifted GH secretion into later portions of the night, with a significantly greater delay observed after the larger dose. Similarly, a significant delay in the time of the GH rise was produced by SCOP. In contrast, the effects of both doses of SCOP on delta-sleep or sleep onset were small. These data confirm earlier reports demonstrating that cholinergic muscarinic input represents a potentially important source of regulation of nocturnal GH release and suggest that the magnitude of the reduction in GH and the extent of delay in the GH rise time may reflect quantitative differences in the degree of cholinergic blockade. These data are in agreement with recent studies suggesting that the timing of GH release need not be associated with delta-sleep per se.

Corticosterone and prolactin response to TFMPP in rats during repeated antidepressant administration

The corticosterone and prolactin response to acute administration of the 5-HT agonist 1-(m-trifluoromethylphenyl) piperazine (TFMPP) (10 mg kg-1) was assessed in rats treated for 10 days with either saline, amitriptyline (20 mg kg-1 day-1) or nialamide (40 mg kg-1 day-1). For all groups, TFMPP significantly increased both serum corticosterone and prolactin concentrations compared with control animals challenged with saline. However, the corticosterone response to TFMPP was attenuated significantly by nialamide pretreatment, while the prolactin response to TFMPP was enhanced significantly by amitriptyline pretreatment. These results support previous reports that antidepressants differentially affect 5-HT-ergic systems involved in the regulation of corticosterone and prolactin secretion.