Interaction between the serotonergic system and HPA and HPT axes in patients with major depression: implications for pathogenesis of suicidal behavior.

Disturbances in the serotonin (5-hydroxytryptamine, 5-HT) system constitute the neurobiological abnormality most consistently associated with suicide. This abnormality could be a marker of vulnerability predisposing individuals to auto-aggressive and impulsive behavior. However, other abnormalities, such as hyperactivity of the hypothalamic-pituitary-adrenal (HPA) axis, have also been described in suicide victims. While inhibitory effects of adrenocorticosteroids on 5-HT(1A) receptor function have been shown in animals, HPA axis hyperactivity does not seem to be responsible for the reduced 5-HT activity found in depressed patients with a history of suicidal behavior. On the other hand, hypothalamic-pituitarythyroid (HPT) axis dysfunction, frequently observed in depression, may represent a compensatory response to reduced central 5-HT neurotransmission. Moreover, in depressed patients with a history of suicidal behavior, the absence of a functional link between HPT and dopamine activity at the hypothalamic level may be implicated in the pathophysiology of suicidal behavior. Future research is needed to determine why compensatory mechanisms are not efficient in patients with suicidal behavior.

Retrospective studies, using prolactin (PRL) response to d-fenfluramine test (d-FEN; a presynaptic serotonin [5-hydroxytryptamine, 5-HT]-releasing and -uptake-inhibiting agent), have suggested that reduced serotonergic functioning may be a marker of increased suicide risk in patients with major depression[1] and schizophrenia.[2] The etiology of this abnormality remains unknown, but it has been suggested[3] that overactivation of the hypothalamic-pituitary-adrenal (HPA) axis by chronic stress, and the associated hypercortisolism, could directly induce changes in 5-HT pathways. Consequently, it has been hypothesized that 5-HT abnormality in patients with a history of suicidal behavior could be secondary to hyperactivity of the HPA axis.

The hypothesis of hyperactivity of the HPA is supported by the postmortem findings of increased cerebrospinal fluid (CSF) corticotropin-releasing hormone (CRH) concentrations and reduced CRH receptor binding sites in the frontal cortex of suicide victims (interpreted as downregulation following CRH hypersecretion) and dexamethasone nonsuppression in some, but not all, suicide attempters.[4] While there is evidence for inhibitory effects of the HPA on 5-HT activity in animals, studies in healthy humans and depressed patients have produced disparate findings. It has been reported, in small samples, that acute treatment with hydrocortisone attenuates PRL response to d-FEN,[5] though subchronic administration of hydrocortisone induces no change in PRL response to d-FEN.[6] It has also been found that pretreatment with dexamethasone decreases Cortisol - but not PRL - response to d-FEN.[7] In depression, some authors have hypothesized[8,9] that raised Cortisol levels could lead to impaired neuroendocrine responses to the d-FEN challenge. This modification could be related more specifically to blunted 5-HT1A receptor function, rather than to reduced 5-HT presynaptic release or 5-HT2 receptor function. Indeed, preclinical studies have shown that chronic hypercorti solcmia inhibits 5-HT1A receptor function, while 5-HT turnover and 5-HT2 receptor function are enhanced.[10]

Lack of effect of HPA hyperactivity on 5-HT responsiveness in depressed patients

In a recent study,[11] we have examined the relationship between HPA axis activity and 5-HT function in major depression, particularly in patients with a history of suicidal behavior. We measured (i) Cortisol levels at baseline and following dexamethasone suppression test (DST; 1 mg orally administered at midnight on day 1); and (ii) PRL, adrenocorticotropic hormone (ACTH), and Cortisol responses to challenge with d-FEN (45 mg orally, at 9 am on day 5) in 71 drug-free Diagnostic and Statistical Manual of Menial Disorders, Fourth Edition (DSM-IV)[ major depressed inpatients (40 with a history of suicide attempt, 31 without) and 34 hospitalized healthy control subjects.

We hypothesized that, if HPA overactivity is at the origin of reduced 5-HT function, high Cortisol levels (basal or post-DST) would be associated with low hormonal responses to d-FEN.

Depressed patients had higher post-DST Cortisol levels (Table I), but similar responses to d-FEN compared with control subjects. Hormonal responses to d-FEN were not correlated with Cortisol levels (basal or post-DST). Among depressed patients, DST suppressors and DST nonsuppressors exhibited no significant difference in endocrine response to d-FEN. Patients were subsequently classified according to their basal Cortisol values (ie, >550 nmol/L). When patients with high basal Cortisol values (n=10) were compared with patients with normal basal Cortisol values (n=61), they showed no significant difference in post-d-FEN values. These results suggest, that high Cortisol levels (at baseline or post-DST) have no significant effect on PRL or ACTH/cortisol responses to d-FEN.

In our sample, DST nonsuppression was associated with psychotic depression (P<0.0008), increased age (P<0.004), and global severity of depression (P<0.04). Although the exact pathophysiology underlying DST suppression remains unclear, it has been suggested that abnormal Cortisol response reflects impaired negative feedback (at the level of the pituitary corticotroph) on endogenous HPA axis hyperactivity (ie, increase in hypothalamic corticotropin-releasing factor and vasopressin drive that overrides the action of dexamethasone).[13]

When patients with a history of suicide attempt were compared with patients without such a history, they showed lower hormonal responses to d-FEN, but comparable basal and post-DST Cortisol levels (Table II). Taken together these results suggest that (i) increased HPA axis activity does not. impair the ability of the brain's 5-HT system to respond normally; and consequently (ii) increased HPA axis activity is not at the origin of reduced 5-HT activity observed in a subgroup of depressed patients with a history of suicide attempts.

However, given the pharmacological properties of d-FEN, the extent of response to this compound, which is dosedependent, depends on a combined effect on the synthesis and release of 5-HT and the stimulation of postsynaptic 5-HT receptors (ie, 5-HT1A or 5-HT2A/2C receptors, or both), without defining which 5-HT receptor subtypes might be dysrcgulated. In this context, it is conceivable that altered 5-HT1A receptor function by glucocorticoids - although not unanimously found in depression[14] - may be compensated by increased 5-HT turnover and 5-HT2 receptor activity, therefore leading to normal responses to d-FEN test.

In fact, even if the hypothesis of hypercortisolism leading to reduced 5-HT function had been confirmed, this hypothesis would not have clarified why about two thirds of patients with a history of suicidal behavior had not exhibited hypercortisolism. One may argue that hypcrcortisolism could be associated with a specific form of suicidal behavior (ie, violent, although studies on this topic are conflicting); but considering hypercortisolism as etiological may be restrictive since HPA axis hyperactivity is not specific to a particular psychiatric diagnosis.

Increased HPT axis activity as a compensatory mechanism for diminished 5-HT function in depressed patients

Hypothalamic-pituitary-thyroid (HPT) axis activity is altered in a substantial proportion of depressed patients. It is generally accepted that approximately one quarter of euthyroid depressed patients have a blunted thyrotropin (thyroid-stimulating hormone, TSH) response to morning administration of protirelin (thryotropin-relcasing hormone, TRH).[15] We have reported that the 11 pm TRH test (200 µg, intravenous [IV]) is more sensitive than the 8 am TRH test, and that the difference in TSH response between 11 pm and 8 am TRH tests (ΔATSH) is an even more sensitive measure: this chronobiological index is reduced in about 70% of inpatients with major depression.[16] It has been suggested that blunted TRH-induced TSH stimulation might reflect a downregulation of the TRH receptors in the pituitary gland secondary to a prolonged increase in hypothalamic TRH stimulation.[15]

On the basis of recent animal studies, the effects of 5-HT on the central regulation of TRH secretion are believed to be predominantly inhibitory. According to this assumption, a reduced 5-HT function could lead to hypersecretion of TRH, and therefore to blunted TSH response to TRH in depression.

However, when depressed patients are classified on the basis of their ΔΔTSH test status, patients with reduced ΔΔTSH values (≤2.5 µU/mL) have hormonal d-FEN responses comparable to those of controls. Conversely, patients with normal ΔΔTSH values (ie, without HPT axis abnormality) show lower PRL and Cortisol responses to d-FEN than controls and patients with abnormal ΔΔTSH values ( [17] ACTH response to d-FEN - which correlated with Cortisol (r=0.66; n=80; P<0.00001) and PRE (r=0.41; n=80; P<0.0003) response- is also lower in patients without HPT axis abnormality when compared with controls (P<0.009) and patients with HPT dysfunction (ie, reduced AATSH values; P<0.0002).

Thus, patients with normal HPT axis activity exhibit reduced PRL and ACTH/cortisol responses compatible with a 5-HT deficit. On the other hand, patients with abnormal HPT axis activity show a level of 5-HT function comparable to that found in healthy subjects. Therefore, according to the TRH hypothesis,[15] one may hypothesize that TRH overactivity, which produces both pituitary TRH receptor downregulation and direct activation of the thyroid gland,[18,19] could also stimulate 5-HT activity. Indeed, it has been found in animal studies that (i) TRH stimulates 5-HT neurotransmission via 5-Hl receptors; and (ii) thyroid hormones enhance 5-HT activity in certain brain areas (such as the cerebral cortex). Furthermore, the reduced central 5-HT activity found in patients with hypothyroidism is reversed by thyroxine replacement, therapy.[20] In the context, of major depression, the effects of increased HPT axis hormones (ie, increased secretion of TRH and elevated circulating concentrations of thyroid hormones well within the physiological range[19]) may be regarded as a compensatory mechanism in order to correct reduced central 5-HT activity.

Schematically, one may define two situations (

The compensatory mechanisms are effective; in this case a signal (such as a decrease in 5-HT function) leads to a series of biological modifications (such as an increase in thyroid axis activity). These modifications may be understood as a repairing process aiming to restore an efficient 5-HT functioning.

The compensatory mechanisms are not effective; in this case the 5-HT dysfunction remains. In depressed patients with a history of suicidal behavior, 5-HT alteration may be understood as a failure of the compensatory mechanisms.

Interactions between the dopaminergic system and the HTP axis in depression

Given the interactions between dopamine (DA) and HPT and 5-HT, one may hypothesize that DA may also be involved in the compensatory mechanisms. It is known that the mesolimbic DA system plays a key role in goaldirected and motivational behavior. In depression, it has been suggested that hypofunction in mesolimbic DA system may be involved in anhedonia and amotivational apathy.[21] DA agonists can facilitate the action of antidepressant drugs in certain animal models of depression and in some depressed patients.

While at the hypothalamic level the interregulations of DA and 5-HT systems are complex and not fully understood, preclinical studies have shown that dopamine D2 receptors stimulate the release of hypothalamic TRH and inhibit. TSH production at the pituitary level. In turn, TRH and thyroid hormones stimulate the DA system, and induce a downregulation of D2 receptors.

To examine the functional relationships between HPT axis activity and DA function in depressed patients, especially in those with a history of suicidal behavior, we measured hormonal responses to 8 am and 11 pm TRH tests and to apomorphine (APO) test in 64 drug-free inpatients with DSM-FV [12] major depression (35 with a history of suicide attempt, 29 without) and 34 hospitalized healthy controls. APO, a direct-acting DA agonist with high affinities for D2 and D3 receptors and a partial agonist at the D receptor, decreases PRL and stimulates growth hormone (GH), ACTH, and Cortisol secretion.[22]

Compared with controls, patients demonstrate lower TRH and TSH responses and lower APO-induced PRL suppression (Table III). PRL response to APO provides an indirect index of central neurotransmission by assessing postsynaptic D2 receptor sensitivity at the pituitary level. A lower PRL response to APO may reflect a decreased D2 receptor function. This abnormality may represent (i) a primary deficit in D2 receptor sensitivity in the pituitary in depressed patients; or (ii) a downregulation of D2 receptors secondary to increased presynaptic DA activity. Cooccurrence of HPT axis and tuberoinfundibular DA dysregulation is compatible with a decreased TRH and D2 receptor function, possibly secondary to increased TRH tone, since TRH stimulates the DA system and induces a downregulation of D2 receptors.

When classifying patients according to their history of suicidal behavior, those with a negative history more frequently have reduced AATSH values ( but comparable hormonal APO responses (ie, PRL, ACTH, and Cortisol), than those with a positive history. In patients without a history of suicide attempt, a negative correlation is found between AATSH values and post- APO ACTH (p=-0.44, P=0.02) and Cortisol (p=-0.50,P<0.008) levels. This correlation is found neither in patients with a history of suicide attempt ( nor in control subjects.

The finding of a negative correlation between AATSH and post-APO ACTH and Cortisol values in patients without a history of suicidal behavior is rather paradoxical. Owing to the regulations between HPT and DA systems, one could have expected a positive correlation and not a negative one (ie, an increase in TRH secretion should have led to a decrease in D2 function). Whether hypofunctionality of D2 receptors exists on both hypothalamic and pituitary levels, the absence of GH, ACTH, and Cortisol response to APO in depression would suggest, an upregulation of other DA receptor subtypes (such as D1) in the hypothalamus. Indeed, GH, ACTH, and Cortisol response to APO reflects primarily stimulation of the hypothalamic releasing hormones (GHreleasing hormone and CRH, respectively) rather than a direct, effect on the pituitary Moreover, Cortisol response to APO, which is correlated to ACTH (p=0.74; n=98; P<0.00001), can be considered as an index of central DA function connected with the regulation of the HPA axis. Thus, the negative correlation between ΔΔSH and post-APO ACTH and Cortisol values in patients without a history of suicidal behavior suggests that the efficacy of compensatory mechanisms requires a new functional balance between HPT and DA systems.

In the depressed group with a history of suicidal behavior, the absence of a functional link between HPT and DA activity in the hypothalamus may play a role in the pathophysiology of suicidal behavior. However, one may note that half of the patients of this group showed HPT and DA functional adjustment (Figure 4; ie, those exhibiting blunted ΔΔTSH values), suggesting that this requirement is not sufficient in the efficacy of compensatory mechanisms. In other words, other processes - so far unknown - are also involved in the efficacy of compensatory mechanisms.

Conclusions

Taken together our findings in depressed inpatients suggest, that:

HPA axis hyperactivity is not responsible for the reduced 5-HT activity found in patients with a history of suicidal behavior.

HPT dysregulation may be regarded as a compensatory mechanism for diminished central 5-HT activity.

Cooccurrence of HPT axis and tuberoinfundibular DA dysregulation is compatible with a decreased TRH and D2 receptor function (possibly secondary to increased TRH tone).

The absence of a functional link between HPT and DA activity in the hypothalamus may be implicated in the pathogenesis of suicidal behavior.

A better knowledge of processes involved in the efficacy of compensatory mechanisms could lead to new therapeutic strategies in patients with recurrent major depressive disorder, especially those with a history of suicidal behavior.

Table I.

Demographic characteristics and biological data for depressed patients and normal control subjects. DST, dexamethasone suppression test; No. of abnormal DSTs, number of subjects with highest post-DST Cortisol level >130 nmol/L; Δ, peak concentration minus baseline value; PRL, prolactin; ACTH, adrenocorticotropic hormone.

	Control subjects (n=34)	Depressed patients (n=71)	P
	Mean (SEM)	Mean (SEM)
Age (years)	39.6 (1.6)	40.2 (1.1)	0.81
Range (years)	20-63	19-60
Gender (M/F)	14/20	28/43	0.86
Weight (kg)	66.5 (1.5)	65.5 (1.4)	0.80
Range (kg)	47-90	42.93
Post-DST cortisol (nmol/L)	38 (4)	98 (12)	0.003
No of abnormal DSTs	1	19	0.008
Feunfluramine test
Baseline PRL (µg/L)	12.4 (0.9)	14.5 (1.1)	0.73
ΔPRL (µg/L)	5.5 (0.8)	4.3 (1.3)	0.09
Baseline ACTH (ng/L)	25.5 (3.1)	24.7 (2.0)	0.98
ΔACTH (ng/L)	11.2 (4.6)	8.1 (1.2)	0.57
Baseline cortisol (nmol/L)	358 (20)	378 (21)	0.90
ΔCortisol (nmol/L)	88 (27)	83 (18)	0.74

Table II.

Demographic characteristics and biological data for depressed patients according to their history suicide attempt. HDRS, Hamilton Depression Rating Scale score (1 7-item); DST, dexamethasone suppression test; No. of abnormal DSTs, number of subjects with highest post-DST Cortisol level >1 30 nmol/L; >Δ, peak concentration minus baseline value; PRL, prolactin; ACTH, adrenocorticotropic hormone.

	History of suicide attempt (n=40)	No history of suicide attempt (n=31)	P
	Mean (SEM)	Mean (SEM)
Age (years)	40.7 (1.4)	39.5 (1.8)	0.50
Range (years)	19-58	19-60
Gender (M/F)	17/23	11/20	0.36
Weight (kg)	66.3 (1.9)	62.8 (1.4)	0.24
Range (kg)	42-93	43-85
Post-DST cortisol (nmol/L)	108 (19)	84 (12)	0.93
No. of abnormal DSTs	12	7	0.33
Feunfluramine test
Baseline PRL (µg/L)	13.4 (1.4)	16.1 (1.7)	0.30
ΔPRL (µg/L)	0.7 (0.9)	8.9 (2.6)	0.0001
Baseline ACTH (ng/L)	27.9 (3.1)	20.8 (2.3)	0.07
ΔACTH (ng/L)	6.4 (1.7)	10.4 (2.0)	0.05
Baseline cortisol (nmol/L)	383 (31)	372 (25)	0.85
ΔCortisol (nmol/L)	48 (21)	128 (28)	0.02

Table III.

Demographic characteristics and biological data for depressed patients and normal control subjects. Δ, peak concentration minus baseline value; TRH, thyrotropin-releasing hormone; TSH, thyroid-stimulating hormone; PRL, prolactin; PRL suppression = (PRLAUC/baseline PRLAUC)x100; GH, growth hormone; ACTH, adrenocorticotropic hormone.

	Control (n=34)	Depressed patients (n=64)	P
	Mean (SEM)	Mean (SEM)
Age (years)	37.7 (1.6)	40.4 (1.4)	0.24
Gender (M/F)	14/20	32/32	0.52
TRH tests
8 AM baseline TSH (µU/mL)	1.20 (0.09)	0.94 (0.06)	0.007
8 AM ΔTSH (µU/mL)	8.43 (0.57)	6.21 (0.41)	2.3 10-4
11 PM baseline TSH (µU/mL)	1.19 (0.11)	0.80 (0.05)	0.002
11 PM Δ(µU/mL)	12.34 (0.65)	8.06 (0.50)	4.7 10-7
Apomorphine test
Baseline PRL (µg/L)	10.6 (0.9)	13.3 (1.0)	0.18
PRL suppression (%)	39.5 (1.5)	28.5 (2.0)	1.2 10-4
Baseline GH (ng/mL)	0.4 (0.3)	0.3 (0.2)	0.88
ΔGH (ng/mL)	12.3 (1.6)	12.5 (1.2)	0.89
Baseline ACTH (ng/L)	26.5 (3.1)	28.1 (2.3)	0.55
ΔACTH (ng/L)	61.5 (19.0)	50.1 (11.6)	0.94
Baseline cortisol (nmol/L)	305 (17)	283 (15)	0.33
ΔCortisol (nmol/L)	175 (31)	117 (17)	0.15