Free Fatty Acids Exert an Inhibitory Effect on Adrenocorticotropin and Cortisol Secretion in Humans

Abstract

Free fatty acid (FFA) administration stimulates the hypothalamic-pituitary-adrenal (HPA) axis in rats, suggesting that the HPA axis and lipolysis may be linked by a positive-feedback loop. To clarify the influence of FFA on the HPA axis in humans, we studied the effect of lipid load on both basal and stimulated ACTH and cortisol secretion in normal subjects. In six young female volunteers [(mean ± sem) age, 24.4 ± 2.1 yr; body mass index, 23.1 ± 1.2 kg/m²), ACTH, cortisol, FFA, glucose, and insulin levels were measured every 30 min for 330 min during the following procedures: 1) iv saline infusion (from 0 to 330 min); 2) iv FFA infusion (Intralipid 10%, from 0 to 210 min) followed by saline infusion (from 210 to 330 min); 3) human CRH (hCRH) administration (2 μg/kg iv at 90 min) during saline infusion (from 0 to 330 min); and 4) hCRH administration during FFA infusion (Intralipid 10%, from 0 to 210 min, followed by saline infusion from 210 to 330 min). During saline infusion, ACTH and cortisol levels progressively declined. Lipid-heparin emulsion (LHE) infusion strikingly increased circulating FFA levels and, simultaneously, amplified the ACTH and cortisol decrease (P < 0.05). After LHE withdrawal, FFA decrease was associated with an increase (P < 0.05) in ACTH and cortisol levels (restored to baseline values within 60 min). The ACTH and cortisol responses to hCRH, however, were unaffected by LHE that, concomitantly, induced an increase (P < 0.05) in glucose but not in insulin levels. This study shows that an LHE-induced increase in FFA levels has an inhibitory effect on spontaneous ACTH and cortisol secretion in humans. Lipid load, however, does not affect the ACTH and cortisol responses to hCRH; this evidence would indicate that the negative influence of FFA on the HPA axis in humans takes place at the suprapituitary level.

FREE FATTY ACIDS (FFA) are known to influence anterior pituitary function. GH secretion, for instance, is regulated by metabolic factors such as glucose (1, 2) and FFA (2–4). Because GH has a direct lipolytic effect (5), its inhibition by FFA has been hypothesized to reflect a negative-feedback mechanism on somatotroph function.

Among several factors modulating lipolysis, glucocorticoids have been shown to play a role; in fact, several studies demonstrated that the hyperactivation of the hypothalamic-pituitary-adrenal (HPA) axis is associated with the elevation of circulating FFA levels (6, 7). It had been hypothesized that the increase of circulating FFA levels could induce a suppression of HPA activity via a feedback mechanism. This hypothesis was tested in animals by Widmaier et al. (8), who, surprisingly, found a dose-related increase in ACTH and corticosterone levels after the infusion of Intralipid emulsion in rats; based on this evidence, the existence of a positive-feedback loop between FFA and the HPA axis was hypothesized. In fact, in vitro studies showed a direct stimulatory effect of FFA, especially long-chain unsaturated ones, on the adrenal gland (9, 10), although very high FFA concentrations have been shown to increase ACTH release from corticotroph cells (9). On the other hand, other authors reported a dual, dose-dependent effect of FFA on cortisol release from adrenal cells; in fact, oleic acid and linoleic acid were able to stimulate glucocorticoid production in the absence of adrenocorticotropic hormone despite high FFA concentrations, which inhibit ACTH action (11).

The influence, if any, of FFA on HPA axis activity in humans is still controversial. There is clear evidence that cortisol levels increase after a mixed meal; however, studies evaluating the effects of meal composition on the HPA axis showed that high-fat meals, as opposed to high-protein meals, do not modify spontaneous cortisol secretion (12, 13). Similarly, oral fat load did not modify the cortisol response to stress in normal subjects (14). On the other hand, the lipolytic effect exerted by glucocorticoids is well known in humans, and several pathophysiological conditions are connoted by both enhanced HPA axis activity and increased circulating FFA levels. These conditions include obesity (15–17), diabetes mellitus (18), fasting (19), and anorexia nervosa (17, 20).

Based on the foregoing, we aimed to clarify whether there is any influence of high FFA levels on the HPA axis in humans, suggesting the existence of a control of corticotroph function by lipids. To reach this goal, we studied the effect of an acute and high increase of circulating FFA levels induced by the infusion of a lipid-heparin emulsion (LHE) on both basal and stimulated ACTH and cortisol secretion in normal subjects.

Subjects and Methods

Six young female volunteers [(mean ± sem) age, 24.4 ± 2.1 yr; BMI, 23.1 ± 1.2 kg/m²) were studied in their early follicular phase in two consecutive menstrual cycles. All of the subjects gave their written informed consent to participate in the study. The study had been approved by the independent Ethical Committee of the University of Turin.

Each subject underwent the following tests: 1) iv saline infusion (500 ml of 0.9% NaCl solution) from 0 to 330 min; 2) iv 10% LHE infusion (350 ml Intralipid, Fresenius Kabi, Verona, Italy; together with 2500 U heparin, corresponding to 375.2 kcal) from 0 to 210 min, followed by saline infusion (200 ml of 0.9% NaCl solution) from 210 to 330 min; 3) iv human CRH (hCRH) administration (2 μg/kg at 90 min) during saline infusion from 0 to 330 min; and 4) iv hCRH administration (2 μg/kg at +90 min) during iv 10% LHE infusion from 0 to 210 min, followed by saline infusion (200 ml of 0.9% NaCl solution) from 210 to 330 min.

Intralipid is a suspension of soybean oil and glycerol; during Intralipid infusion, an increase of circulating FFA is generated from this suspension after the activation of lipoprotein lipase. To increase the lipoprotein lipase activity, heparin has been added to the suspension.

All tests were performed in the morning between 0830 and 0900 h, after an overnight fast and 30 min after an indwelling catheter was inserted in a forearm vein and kept patent by slow infusion of isotonic saline. All tests were performed in random order and at least 5 d apart.

Blood samples were taken every 30 min from 0 to 330 min. ACTH, cortisol, FFA, glucose, and insulin levels were measured at each time point. In test sessions 3 and 4, blood samples for ACTH and cortisol evaluation after hCRH administration were taken every 15 min from 90 to 210 min.

Plasma ACTH levels (pg/ml; 1 pg/ml × 0.22 = 1 pmol/liter) were measured in duplicate by immunoradiometric assay (Allegro HS-ACTH, Nichols Institute Diagnostics, San Juan Capistrano, CA). The sensitivity of the assay was 1 pg/ml. The inter- and intraassay coefficients of variation ranged from 6.9 to 8.9% and from 1.1 to 3.0%, respectively.

Serum cortisol levels (μg/dl; 1 μg/dl × 27.59 = 1 nmol/liter) were measured in duplicate by RIA (CORT-CTK 125 RIA, Sorin Biomedica, Saluggia, Italy). The sensitivity of the assay was 0.4 μg/dl. The inter- and intraassay coefficients of variation ranged from 6.6 to 7.5% and from 3.8 to 6.6%, respectively.

Plasma FFA levels (mEq/liter × 282 = 1 mg/liter) were measured by enzymatic analysis using the NEFA QUICK BMY kit (Roche Molecular Biochemicals, Yamanouchi, Tokyo, Japan).

Plasma glucose levels (mg/dl; 1 mg/dl × 0.05 = 1 mmol/liter) were measured by glucooxidase colorimetric method (GLUCOFIX, Menarini Diagnostici, Florence, Italy).

Serum insulin levels (mU/liter; 1 mU/liter × 7.17 = 1 pmol/liter) were measured in duplicate by immunoradiometric assay (INSIK-5, Sorin Biomedica). The sensitivity of the assay was 2.5 mU/liter. The inter- and intraassay coefficients of variation ranged from 6.2 to 10.8% and from 5.5 to 10.6%, respectively.

All samples from an individual subject were measured in the same assay.

The data are expressed as mean ± sem of absolute values.

The statistical analysis was carried out using nonparametric ANOVA (Friedman test) and the Wilcoxon test, as appropriate.

Results

ACTH, cortisol, FFA, glucose, and insulin levels at baseline were similar in every session (Table 1 and Figs. 1–4¹²³⁴).

During saline infusion, both ACTH and cortisol levels showed a progressive decline (Table 1 and Fig. 2). FFA, glucose, and insulin levels did not undergo significant variations during the saline session (Table 1 and Figs. 1 and 3³).

LHE infusion led to a clear-cut increase in circulating FFA levels, which reached peak values at 210 min (P < 0.005) (Table 1 and Fig. 1). This rise in FFA levels was associated with a significant decrease in ACTH levels (P < 0.05 vs. baseline), which reached nadir values between 90 and 210 min. This decrease was significantly higher than that recorded during saline infusion (P < 0.05 vs. time points 90, 120, 150, 180, and 210 min during saline). The FFA increase also induced a significant decrease in cortisol levels (P < 0.05 vs. baseline), which was more marked than during saline infusion (P < 0.05 vs. time points 90, 120, 150, 180, 210, and 240 min during saline) (Table 1 and Fig. 2).

After withdrawal of LHE infusion, FFA levels showed a dramatic decrease (P < 0.05). This was associated with a progressive increase of ACTH and cortisol levels, which within 60 min reached values similar to those recorded during the saline session (Table 1 and Fig. 2).

LHE infusion induced a significant (P < 0.05), progressive increase in glucose levels without any significant change in insulin secretion. Glucose levels remained stable after stopping the LHE infusion (Fig. 3).

hCRH induced the well-known, marked increase in both ACTH and cortisol levels (P < 0.05), which was not significantly modified by LHE infusion. Notably, ACTH and cortisol responses to hCRH during LHE infusion have been slightly delayed, but this change was not statistically different (Table 1 and Fig. 4).

Side effects

No side effects were observed during LHE infusion. hCRH induced a transient facial flushing and tachycardia in five subjects.

Discussion

The results of this study demonstrate that a lipid load-induced increase of circulating FFA levels has an inhibitory influence on spontaneous ACTH and cortisol secretion in humans. However, the ACTH and cortisol responses to hCRH are not affected by the lipid load.

The existence of a relationship between the HPA axis and lipid metabolism is based on the well-known lipolytic action of glucocorticoids. In fact, the activation of this axis induces a clear increase of circulating FFA levels (6, 7). However, the influence, if any, of FFA on HPA activity is not well established.

Data from animal studies suggest a complex and not completely understood interplay between the HPA axis and FFA. A dose-related increase of circulating FFA concentrations elicited a stimulatory effect on both corticotroph and adrenal secretion in rats (8); these findings suggested a positive-feedback action of FFA on the HPA axis. However, a dual effect of lipids on adrenal activity also has been described (8, 10, 11, 21). In vitro studies have shown both an increase and a decrease of cortisol/corticosterone secretion from the adrenal gland, depending on the administered dose (8, 10, 11, 21); these findings suggested that circulating FFA could exert a dose-dependent, positive- or negative-feedback action on the HPA axis.

The data concerning the interplay between FFA and the HPA axis in humans are even scantier. Indirect evidence points toward a positive influence of FFA on the HPA axis, in agreement with animal data. In fact, several pathophysiological conditions, such as obesity (15–17), diabetes mellitus (18), fasting (19), and anorexia nervosa (17, 20) are connoted by concomitant enhancement of HPA axis activity and increased circulating FFA levels; theoretically, this HPA hyperactivity would, at least partially, reflect chronically elevated FFA levels.

In this context, our findings surprisingly show that an Intralipid-induced increase of circulating FFA concentrations exerts an inhibitory instead of stimulatory effect on the HPA axis in normal subjects; these findings do not support the hypothesis that exposure to high FFA levels stimulates the HPA axis and, on the contrary, suggest the existence of a negative feedback loop between FFA and HPA in humans. We chose an Intralipid dose equivalent to that used in many studies focused on the effect of FFA on GH secretion, which showed a clear inhibitory effect on somatotroph function (3, 4, 22). Indeed, this dose induced very supraphysiological FFA levels in our subjects, and therefore, we cannot exclude the concept that lower FFA levels could have a different impact (or no impact) on the ACTH and cortisol secretory pattern, as has been observed in animals (8). Moreover, a possible influence of heparin per se on HPA changes observed in our study cannot be definitely ruled out. In fact, controversial findings about the influence of heparin on adrenal function have been reported, showing either inhibiting effects or no effects on it (23–25).

Regarding the physiological relevance of the effect of LHE-induced FFA concentrations on the HPA axis in our study, the inhibitory influence exerted by pharmacological FFA levels does not allow us to state that a physiological control of HPA exerted by circulating FFA levels is operative. However, our results probably allow us to definitely rule out the possibility that FFA have a positive influence on HPA in humans. Thus, chronically elevated FFA levels in obesity (15–17), diabetes mellitus (18), fasting (19), and anorexia nervosa (17, 20) are unlikely to explain the HPA hyperactivation that connotes these pathological conditions.

The lack of influence of the lipid load on the ACTH and cortisol response to hCRH allows some speculation about the site where FFA inhibitory effect could take place. In fact, because the corticotroph response to hCRH is unchanged during exposure to high FFA levels, an action at the hypothalamic level can be hypothesized. Direct evidence supporting a FFA action on hypothalamic neurohormones and neurotransmitters mainly involved in the control of corticotroph function is, to our knowledge, lacking. An action of FFA at the pituitary level has been suggested by some authors who, however, found a stimulatory effect of high FFA concentrations only upon ACTH secretion from rat pituitary in vitro (9). A stimulatory role of FFA on corticosterone secretion as a result of a direct action of lipids on the adrenal glands, as well as the influence of FFA on peripheral glucocorticoid metabolism, has also been reported (10, 23–25). These data could reflect peripheral actions of FFA that, however, seem to be overridden by a major inhibitory influence taking place at a suprapituitary level, at least in humans.

A possible FFA action at the hypothalamic or suprahypothalamic level is suggested by the evidence that FFA possess an electrophysiological effect on the central nervous system and that FFA are taken up by central nervous system cells (26, 27). Moreover, this hypothesis fits well with other evidence indicating that metabolic fuels, including FFA, play a major role in the regulation of other neurohormones, e.g. GnRH and somatostatin (2, 4, 28, 29). At present, to our knowledge, there is no evidence definitely supporting CRH and/or arginine vasopressin modulation by FFA. On the other hand, the absence of the effect of an Intralipid-induced FFA increase on HPA response to hCRH is in agreement with a CRH-mediated mechanism, whereas some change in the corticotroph response to exogenous hCRH could be expected, assuming that FFA may have an inhibiting effect on endogenous arginine vasopressin release (30). Regarding the possibility that FFA may modulate neuropeptides and/or metabolites involved in energy balance, it has been shown that ghrelin, a new gastric hormone mainly involved in metabolic processes, also shows a stimulatory effect on the HPA axis. Some studies both in animals and humans demonstrated that the increase in circulating FFA levels induced by a fat load reduced circulating ghrelin levels (31, 32), although these data have not been confirmed by other studies (our unpublished results, and Ref. 33); thus, a possible involvement of ghrelin in the modulation of HPA function exerted by FFA cannot be excluded. Moreover, changes in glucose levels are known to influence the HPA axis. Hypoglycemia is a well-known stimulus of HPA function, whereas evidence indicating an inhibiting influence of glucose load on the HPA axis in humans is controversial (34, 35). A previous study in rats showed that Intralipid administration was followed by an increase in plasma glucose levels, which could theoretically account for the modulation of ACTH and corticosterone secretion observed in that study (8). The slight but significant increase in glucose levels under lipid load observed in our study could, theoretically, partially explain our findings. However, it must be noted that the glucose increase in the present study was clearly lower than that recorded after the administration of glucagon, which is a well-known stimulus of ACTH and cortisol secretion (36).

In conclusion, this study shows that the increase in circulating FFA levels induced by lipid load has an inhibitory effect on spontaneous ACTH and cortisol secretion in humans. Lipid load, however, does not affect the ACTH and cortisol responses to hCRH; this evidence would indicate that the negative influence of FFA on the HPA axis in humans takes place at the suprapituitary level.

Acknowledgments

We thank Professor F. Camanni for his support of this study as well as Dr. A. Bertagna, A. Barberis, and M. Talliano for their skillful technical assistance.

This work was supported by the University of Turin and the Foundation for the Study of Endocrinological and Metabolic Diseases.

Abbreviations:

 
• FFA,

  Free fatty acid(s);

 
• hCRH,

  human CRH;

 
• HPA,

  hypothalamic-pituitary-adrenal;

 
• LHE,

  lipid-heparin emulsion.

References