Effect of Arginine on the Hypothalamic–Pituitary–Adrenal Axis in Individuals With and Without Vasopressin Deficiency

Abstract

Context

Arginine stimulates pituitary hormones, like growth hormone and vasopressin, but its effect on the hypothalamic–pituitary–adrenal (HPA) axis is unknown. Arginine may also stimulate the HPA axis, possibly through a mechanism involving vasopressin.

Objective

To investigate the effect of arginine on adrenocorticotropic hormone (ACTH) and cortisol in subjects with and without vasopressin deficiency.

Design

Prospective study, University Hospital Basel.

Participants

38 patients with central diabetes insipidus, 58 patients with primary polydipsia, and 50 healthy controls.

Intervention

Arginine infusion with measurement of ACTH, cortisol and copeptin at baseline and 30, 45, 60, 90, and 120 minutes.

Results

We found different response patterns to arginine: in patients with diabetes insipidus (and low stimulated copeptin levels) median (interquartile range [IQR]) ACTH and cortisol increased from 22.9 (16.8, 38.7) to 36.6 (26.2, 52.1) ng/L and from 385 (266, 463) to 467 (349, 533) nmol/L, respectively. In contrast, median (IQR) ACTH and cortisol levels decreased in patients with primary polydipsia (despite high stimulated copeptin levels): ACTH from 17.3 (12.3, 23) to 14.8 (10.9, 19.8) ng/L and cortisol from 343 (262, 429) to 272 (220.8, 360.3) nmol/L; likewise, in healthy controls: ACTH from 26.5 (17.6, 35.7) to 14.8 (12.1, 22.7) ng/L and cortisol from 471 (393.3, 581.8) to 301.5 (206.5, 377.8) nmol/L.

Conclusion

Diabetes insipidus is associated with increased responsiveness of ACTH/cortisol to arginine. In contrast, arginine does not stimulate the HPA axis in healthy controls or in primary polydipsia.

Arginine is an amino acid known to promote various metabolic effects in humans: insulin is released, plasma concentration of free fatty acids falls and hormones of the pituitary gland such as growth hormone, prolactin, and thyroid stimulating hormone (TSH) are stimulated (1-3). Arginine infusion has therefore been widely used as a simple tool to diagnose growth hormone deficiency in children and adults (4-6).

We have recently shown that arginine infusion is also a potent stimulus of the posterior pituitary and that arginine-stimulated copeptin measurements are a useful tool to diagnose diabetes insipidus with high diagnostic accuracy (7,8).

In view of the broader use of this test in clinical practice, it is of interest to study the mechanism of action of arginine within the hypothalamus and the pituitary gland. In particular, the effect of arginine on the hypothalamic–pituitary–adrenal (HPA) axis has not been studied so far. Korbonits et al assessed the impact of the growth hormone secretagogue hexarelin on the HPA axis and observed a release of adrenocorticotropic hormone (ACTH) and cortisol in response to hexarelin injection (9). The HPA response was further augmented by the addition of corticotropin-releasing hormone (CRH). This led the authors to postulate that the mechanism of hexarelin action did not primarily involve CRH, but possibly another hypothalamic stimulator of ACTH (eg, vasopressin). Since vasopressin is closely related to the HPA axis, it is plausible that it has a modulatory role, such as a synergistic effect of vasopressin with CRH (10-12).

In line with this, we hypothesized that arginine stimulates the HPA axis and that its effect depends on the integrity of the posterior pituitary gland and on vasopressin release. The aim of this study was to investigate the effect of arginine infusion on the secretion of ACTH and cortisol in patients with vasopressin deficiency (ie, central diabetes insipidus) in patients with chronic vasopressin suppression (ie, primary polydipsia) and in healthy controls.

Methods

Study design and setting

This is a secondary analysis of a prospective study conducted at the University Hospital Basel from 2013 to 2017, whose aim was to evaluate copeptin measurements after arginine stimulation as a diagnostic test for diabetes insipidus. The study design has previously been described in detail (8). In short, 3 subject groups were recruited: adult patients with central diabetes insipidus, patients with primary polydipsia, and healthy controls. The final reference diagnoses of central diabetes insipidus and primary polydipsia were based on 3 diagnostic components: results of the indirect water deprivation test according to the protocol of Miller et al, patients’ history, and treatment response (8,13).

Patients with nephrogenic diabetes insipidus, patients suffering from any acute illness, and pregnant patients were excluded. For the group of healthy controls, subjects over 18 years and with normal drinking habits (no history of polyuria defined as urinary output of more than 3000 mL per day) were recruited. Pregnant women and people taking any medication (other than hormonal contraception) were excluded. Written informed consent was obtained from all study participants. The local ethic committee approved the study protocol and the study was preregistered on ClinicalTrials.gov (NCT00757276).

Procedures

The study participants underwent a standardized arginine stimulation test. The test started at 800 am after an overnight fast. Participants were requested to drink freely until 2 h before test start (to avoid hypernatremia in case of diabetes insipidus) and then to stop drinking (to avoid hyponatremia in case of primary polydipsia). During the test, participants were not allowed to drink. If patients were being treated with desmopressin, the drug was discontinued at least 24 h before the test. Baseline data including age, body mass index (BMI), medical history, physical examination, vital signs, and routine laboratory tests were collected. The participants were settled in a supine position and an intravenous catheter was inserted in an antecubital vein. The dosage of arginine (L-arginine-hydrochloride 21%; Braun, B Braun Melsungen AG, Melsungen, Germany) was adapted to the participants’ weight: 0.5 g/kg body weight of arginine were diluted in 500 mL NaCl 0.9% and infused over 30 min. At baseline and at 30, 45, 60, 90, and 120 minutes after arginine infusion, blood was drawn for the measurements of the following hormones: ACTH, cortisol, insulin, glucose, prolactin and copeptin. The blood samples were immediately centrifuged for 10 min at 3000 rpm at 4°C and stored at −70°C for later analysis. ACTH levels were determined with IMMULITE 2000 ACTH, a solid-phase, 2-site sequential chemiluminescent immunometric assay (Siemens Healthcare, Diagnostics Products Ltd., Llanberis, UK). Cortisol levels were determined with Elecsys Cortisol Test 2010 and prolactin levels with Elecsys Prolactin II, both electrochemiluminescence immunoassays (Roche Diagnostics GmbH, Mannheim, Germany). Copeptin levels were measured in one batch with a commercial automated immunofluorescence assay (B.R.A.H.M.S Copeptin-proAVP KYPTOR, Brahms, Thermo Scientific Biomarkers, Henningsdorf, Germany). Routine laboratory measurements were performed using automated biochemical analyses in the University Central Laboratories (University Hospital Basel, Basel, Switzerland).

During the test, blood pressure and pulse were monitored, and a standardized clinical feature questionnaire was used to record possible adverse effects and symptoms. Finally, participants were asked to rate the level of discomfort during the test on a visual analog scale that ranged from 0 to 10 where 0 was no discomfort and 10 was maximal discomfort.

Study objective

The primary objective of this study was to investigate the influence of arginine infusion on the secretion of ACTH and cortisol in patients with central diabetes insipidus, in patients with primary polydipsia, and in healthy controls. A secondary objective was to explore a possible association between ACTH/cortisol and plasma copeptin concentrations following arginine infusion. Further objectives included the effects of arginine infusion on insulin, glucose, and prolactin levels.

Statistical analysis

Baseline demographic and clinical characteristics are presented as frequencies and percentages or as mean and standard deviation (SD). P-values are derived from Wilcoxon signed-rank test for continuous variables and Fisher’s exact test for categorical variables for comparison between patients with diabetes insipidus and patients with primary polydipsia.

We graphically examined the time course of the hormones after arginine infusion in the three groups. For each patient, we derived the maximally stimulated value and the maximal change from baseline (absolute and relative) of the hormones. The maximally stimulated value is defined as the highest hormone level achieved during the test. Summary statistics (median, IQR) are provided for the baseline value, the maximally stimulated value, and the maximal relative change. In a post hoc subgroup analysis, we compared the time course of ACTH and cortisol concentrations after arginine infusion in 2 subgroups of patients with diabetes insipidus: familial diabetes insipidus (magnocellular vasopressin deficiency and parvocellular vasopressin deficiency) and selected patients with acquired diabetes insipidus who presumably only have magnocellular vasopressin deficiency (4 with postoperative, 3 with hypophysitis, and 2 with idiopathic diabetes insipidus).

Maximally stimulated values were tested for a difference between the 3 groups by means of quantile (median) regression (R package “quantreg”), with the respective baseline values as covariates. Diabetes insipidus was defined as the reference group; hence, we estimated the differences between diabetes insipidus and primary polydipsia and between diabetes insipidus and healthy controls. We report estimated effect sizes with 95% confidence intervals (CIs; based on bootstrapping) and P-values.

Additionally, we investigated the association between maximally stimulated ACTH and cortisol levels with copeptin levels and other possible influencing factors. Therefore, separate quantile models were fitted for all possible combinations (baseline value, baseline copeptin, maximally stimulated copeptin, interaction terms). We used Akaike’s information criterion to select the best of these models to explain ACTH and cortisol after arginine stimulation. We performed exploratory subgroup analyses by fitting the best model to each subject group separately. These associations are presented visually by scatterplots.

For analyses of ACTH and cortisol, patients with glucocorticoid treatment were excluded.

All analyses were conducted using the statistical software package R (R Core Team 2018), Version 3.6.0 using 2-sided statistical tests.

Results

Baseline characteristics

The pooled patient data set included 96 patients of whom 38 (39.6%) were diagnosed with central diabetes insipidus and 58 (60.4%) with primary polydipsia. Patients with primary polydipsia were younger than patients with diabetes insipidus and female patients were predominant in all groups. Blood pressure, plasma sodium concentration, and plasma osmolality were in the normal range and slightly higher in patients with diabetes insipidus. The majority of patients suffered from acquired pituitary disease and 6 (15.8%) patients had familial diabetes insipidus; 21 patients with diabetes insipidus (55.3%) suffered from anterior pituitary deficiency, and 17 patients with diabetes insipidus (44.7%) were on glucocorticoid treatment. Psychiatric diseases were more frequent in patients with primary polydipsia (31%).

The control group consisted of 50 healthy controls of whom 23 (46%) were male with a mean age of 30.3 years (SD 9.9). Baseline demographic, clinical, and laboratory characteristics are presented in Table 1.

Time courses of ACTH and cortisol

For analysis of ACTH and cortisol, 18 patients with glucocorticoid treatment were excluded from the analysis.

Baseline levels of ACTH and cortisol were comparable in all 3 subject groups (Table 2). The time course of ACTH and cortisol levels during arginine stimulation revealed different response patterns in the 3 subject groups. In patients with diabetes insipidus, ACTH and cortisol levels increased upon arginine stimulation, whereas in patients with primary polydipsia, ACTH and cortisol levels decreased slightly (Figs. 1 and 2). The enhanced ACTH/cortisol response in diabetes insipidus was consistent, both in patients with familial diabetes insipidus (n = 6) and in those with acquired diabetes insipidus (n = 9) (Fig. 3A and 3B).

The maximal changes of ACTH and cortisol from baseline differed strongly between the 3 subject groups, with a higher maximal increase in ACTH and cortisol in patients with diabetes insipidus compared to patients with primary polydipsia and healthy controls. Absolute and relative values of ACTH and cortisol are given in Table 2.

The baseline-adjusted median (95% CI) difference in the maximally stimulated ACTH values for patients with diabetes insipidus versus patients with primary polydipsia or healthy controls was 22 (8.6, 35.3) ng/L (P-value = 0.0017) and 23.7 (10.8, 36.6) ng/L (P-value = 0.0004), respectively. For cortisol, the baseline-adjusted median (95% CI) difference in maximal values was 126.3 (63, 189.7) nmol/L (P-value = 0.0002) for patients with diabetes insipidus versus patients with primary polydipsia and 176.5 (119.6, 233.3) nmol/L (P-value < 0.0001) for patients with diabetes insipidus versus healthy controls.

Time course of copeptin

Copeptin levels only minimally increased in patients with diabetes insipidus in contrast to a clear increase in patients with primary polydipsia and healthy controls (Table 2) (8).

Association of ACTH and cortisol response with copeptin levels

The best model based on Akaike’s information criterion for maximally stimulated ACTH values included the variables of baseline ACTH, baseline copeptin, and maximally stimulated copeptin. We observed a positive association of maximally stimulated ACTH with baseline ACTH and maximally stimulated copeptin values (Table 3). Subgroup analyses indicate that the positive association of maximally stimulated ACTH levels with maximally stimulated copeptin levels was most pronounced in healthy controls. For patients with primary polydipsia a similar but much weaker effect was observed. For patients with diabetes insipidus our data provided no evidence for such an association (Fig. 4). In contrast, for maximally arginine-stimulated cortisol, our data provide no evidence for an association with maximally stimulated copeptin (Fig. 5).

Insulin, glucose and prolactin levels

Insulin and glucose increased in all 3 groups with a peak after 30 min and declined subsequently (Fig. 6). Glucose levels dropped below the baseline levels at 60 min and increased again thereafter but remained in the range of normal during the test (Fig. 6). Prolactin levels increased in all groups until 45 min after arginine stimulation and declined back to baseline levels until 120 min (Fig. 7).

We did not find any evidence for a difference in maximal changes between subject groups for insulin, glucose and prolactin (data not shown).

Adverse effects and level of discomfort

Vital signs and sodium levels remained stable and in the normal range during the test (except for single patients with diabetes insipidus who experienced mild hypernatremia at the end of the test) (8). Patients reported low levels of discomfort (diabetes insipidus: median visual analog scale score 3.5 [interquartile range (IQR)] 2−4, primary polydipsia: 3 [2–4], healthy controls: 1 [1−3]) (8). The most common adverse effects were nausea, vertigo, headache, or facial paraesthesia and occurred in similar frequency in the 2 patient groups (diabetes insipidus [without glucocorticoid treatment]: 53%, primary polydipsia: 58%). In the group of healthy controls, adverse effects were less common (20%).

Discussion

This study showed different ACTH and cortisol responses after arginine infusion in patients with central diabetes insipidus compared to patients with primary polydipsia and healthy controls. While the HPA axis was stimulated in patients with diabetes insipidus, we observed decreasing levels of ACTH and cortisol in patients with primary polydipsia and healthy controls. The stimulatory effects of arginine infusion on prolactin and insulin release are consistent with previous studies (1-3) and did not differ between the 3 subject groups.

Our results suggest that the HPA axis is influenced and regulated by different central mechanisms in patients with diabetes insipidus versus in those with primary polydipsia or healthy controls. The arginine-stimulated release of ACTH/cortisol was consistent irrespective of diabetes insipidus etiology (ie, acquired vs familial diabetes insipidus). Notably, the observed ACTH and cortisol changes seemed not explainable by acute adverse effects during the test as those were balanced between the patients groups. Other possible HPA axis triggers such as hypoglycemia could be excluded. Hypernatremia was experienced in only 4 patients with diabetes insipidus at the end of the test (range 146−147 mmol/L) and seems therefore also a rather unlikely cause of the observed ACTH release.

Interestingly, a different HPA response pattern from various stimuli in patients with diabetes insipidus versus healthy controls has previously been described (14-17). Pivonello et al observed significantly higher ACTH and cortisol levels upon CRH administration in patients with diabetes insipidus compared to healthy controls (16). This increased HPA responsiveness disappeared after 1-week treatment with desmopressin. Similarly, Itagaki et al observed an activation of the HPA axis after hypertonic saline infusion in patients with diabetes insipidus, but not in healthy controls (17). Possible mechanisms causing the enhanced and distinct ACTH response in diabetes insipidus have been postulated: first, as vasopressin downregulates CRH receptors (18), they might be upregulated in patients with diabetes insipidus resulting in an increased sensitivity of corticotropic cells to CRH stimulus. Second, an explanation may involve the hyperactivation of the renin–angiotensin system and catecholamines due to chronic relative hypovolemia in patients with diabetes insipidus with stimulatory impact on the HPA axis (19,20). Pivonello et al found not only increased responsiveness to CRH stimulation but also increased baseline levels of ACTH and cortisol, pointing to a chronic hyperactivation of the HPA axis and hypercortisolism (16). In our study, we had no indication for a chronic HPA hyperactivation as baseline hormone levels were similar in all subject groups.

Finally, different secretion pathways of vasopressin release might play a role (21-23). The first pathway involves the release of vasopressin from magnocellular neurons into the posterior pituitary gland for regulation of water balance. The second pathway involves the release from parvocellular neurons into the hypophyseal portal system where vasopressin might have neuroregulatory functions in the anterior pituitary gland (10). This second secretion pathway may be preserved or even enhanced in patients with acquired diabetes insipidus (15,24,25) and may therefore stimulate the ACTH and cortisol release (15-17). However, as patients with familial diabetes insipidus and therefore ubiquitous vasopressin deficiency showed a similar increase in arginine-stimulated ACTH/cortisol values, an enhanced parvocellular pathway seems not the most plausible explanation. Rather, our data suggest that not only parvocellular vasopressin secretion but also magnocellular vasopressin secretion may play a role in pituitary corticotroph regulation, which is in line with previous animal data (26).

Arginine has similar effects on the pituitary gland as hexarelin (1-3,8). However, in contrast to the study of Korbonits et al, where an increase in ACTH and cortisol upon hexarelin injection in healthy men was observed, our results did not show a stimulatory effect of arginine on the HPA axis in healthy controls and in patients with primary polydipsia. Hence, our results do not support Korbonits’s hypothesis of a stimulatory effect of vasopressin secretion on the HPA axis—at least in the context of arginine infusion.

The slight decrease in ACTH and cortisol levels in healthy controls and patients with primary polydipsia is most likely explained by the physiological circadian rhythm of these stress hormones (27). Interestingly, we found no difference in ACTH, cortisol, or copeptin response in healthy controls versus patients with primary polydipsia. Given the high prevalence of psychiatric disease in the latter group (28,29)—with possible chronic HPA axis activation (30-32)—one might have expected higher baseline or stimulated ACTH/cortisol levels. Similarly, the chronic suppression of vasopressin through excessive fluid intake in patients with primary polydipsia may affect copeptin and HPA response (8). However, our data did not show a difference in copeptin or HPA response to arginine between patients with primary polydipsia and healthy controls, although basal and stimulated copeptin levels tended to be lower in primary polydipsia. Thus, patients with primary polydipsia may essentially have normally functioning vasopressin-releasing cells apart from phases of high fluid intake and low plasma osmolality.

Our study has strengths and limitations. For analyses comprising the HPA axis, patients on glucocorticoid treatment were excluded (all but 1 were patients with diabetes insipidus), and a small group of patients with diabetes insipidus was, therefore, compared to a quite large group of patients with primary polydipsia and healthy controls. Nevertheless, the remaining number of diabetes insipidus patients (n = 21) was still comparable or higher than that in similar studies (16,17). The subgroup analysis of single patients with familial and acquired diabetes insipidus was done post hoc and the results should be interpreted with caution. The study design did not allow conclusions to be drawn regarding the clinical impact of the hyperresponsiveness of the HPA axis in diabetes insipidus. Furthermore, we did not test whether our findings of arginine-stimulated ACTH and cortisol levels would have been reversed by the application of desmopressin. On the other hand, strengths of this study are its prospective design, the well-defined participant groups and the novelty of the findings about arginine infusion and HPA axis.

In summary, we show that arginine infusion has different effects on the HPA axis in patients with diabetes insipidus compared to patients with primary polydipsia and healthy controls. Arginine infusion stimulated the HPA axis in patients with diabetes insipidus resulting in an increase in ACTH and cortisol. In contrast, no stimulatory effect on the HPA axis occurred in patients with primary polydipsia or in healthy controls. In line with previous studies, our results suggest a hyperresponsiveness of the HPA axis in patients with central diabetes insipidus. The clinical impact of this finding remains unknown. Reassuringly, chronic treatment with desmopressin may normalize HPA axis responsiveness (16).

Acknowledgments

Thermo Fisher Scientific (Henningsdorf, Germany) provided the assays for the measurement of copeptin but was otherwise not involved in any part of the study. We thank all participants for their contribution in our study. We also thank the support staff and study and laboratory personnel at all participating centres, especially Cemile Bathelt, Nina Hutter, Céline Bürgi, Joyce Santos de Jesus, Patrick Simon, and Nicole Salvisberg, as well as Jonathan Thwaite, for English corrections.

Financial Support: This study was supported by a grant awarded to Mirjam Christ-Crain from the Swiss National Science Foundation (SNF-162608), the University Hospital Basel, and a grant of the Goldschmidt-Jacobson Foundation awarded to Katja Bologna.

Clinical Trial Information: Registration number NCT00757276.

Additional Information

Disclosure Summary: The authors have nothing to disclose.

Data Availability: The datasets generated during and/or analyzed during the current study are not publicly available but are available from the corresponding author on reasonable request.

References

Author notes