Transcriptional and physiological responses to chronic ACTH treatment by the mouse kidney.

We investigated the effects on urinary steroid and electrolyte excretion and renal gene expression of chronic infusions of ACTH in the mouse. ACTH caused a sustained increase in corticosteroid excretion; aldosterone excretion was only transiently elevated. There was an increase in the excretion of deoxycorticosterone, a weak mineralocorticoid, to levels of physiological significance. Nevertheless, we observed neither antinatriuresis nor kaliuresis in ACTH-treated mice, and plasma renin activity was not suppressed. We identified no changes in expression of mineralocorticoid target genes. Water turnover was increased in chronic ACTH-treated mice, as were hematocrit and hypertonicity: volume contraction is consistent with high levels of glucocorticoid. ACTH-treated mice exhibited other signs of glucocorticoid excess, such as enhanced weight gain and involution of the thymus. We identified novel ACTH-induced changes in 1) genes involved in vitamin D (Cyp27b1, Cyp24a1, Gc) and calcium (Sgk, Calb1, Trpv5) metabolism associated with calciuria and phosphaturia; 2) genes that would be predicted to desensitize the kidney to glucocorticoid action (Nr3c1, Hsd11b1, Fkbp5); and 3) genes encoding transporters of enzyme systems associated with xenobiotic metabolism and oxidative stress. Although there is evidence that ACTH-induced hypertension is a function of physiological cross talk between glucocorticoids and mineralocorticoids, the present study suggests that the major changes in electrolyte and fluid homeostasis and renal function are attributable to glucocorticoids. The calcium and organic anion metabolism pathways that are affected by ACTH may explain some of the known adverse effects associated with glucocorticoid excess.

there are various pathological, genetic, and environmental circumstances that directly or indirectly lead to activation of the hypothalamo-pituitary-adrenal (HPA) axis and thereby cause an increase in blood pressure. Patients with Cushing syndrome and those exposed to chronic stress exemplify this process: ACTH causes concomitant adrenal gland hypertrophy, adrenocorticosteroid production, and hypertension. Studies of the etiology of the causes of HPA-mediated blood pressure increases have focused generally on the mineralocorticoid and glucocorticoid properties of adrenal secretions, and although very many contributory factors have been implicated (27, 45), no clear-cut mechanism has emerged. Other cardiovascular risk factors including obesity, dyslipidemia, and diabetes are also associated with glucocorticoid excess.

Increased mineralocorticoid activity might be expected for three reasons. First, secretion of the principal mineralocorticoid, aldosterone, in the zona glomerulosa is driven in part by ACTH (46). Second, intermediates in corticosteroid biosynthetic pathways (e.g., deoxycorticosterone) are also stimulated by ACTH, and these also have mineralocorticoid properties (23, 42). Third, endogenous glucocorticoids (e.g., cortisol and corticosterone) are also potential agonists of mineralocorticoid receptors (MR), particularly when circulating levels are high enough to overcome the protective 11-oxidation step that inactivates glucocorticoid hormones in MR target tissues (39).

Arguably, long-term blood pressure is determined primarily by the kidney. Mineralocorticoids are associated with sodium and water reabsorption by the distal tubule, leading to volume expansion at least in the short to medium term. Glucocorticoids also influence renal function independent of MR via glucocorticoid receptors (GR): GR-specific ligands will affect glomerular filtration rate (GFR) and proximal tubular function (3, 37), and GR activation may also influence sodium transport in the distal nephron (16). In the case of HPA activation we have shown (6) that both MR and GR processes can increase epithelial sodium channel (ENaC) activity.

In the present study we have employed chronic ACTH treatment in mice as a model of HPA-mediated hypertension, defining a time course for altered steroidogenesis and electrolyte metabolism. In combination with a microarray analysis of renal gene expression, we have further resolved the responses to ACTH excess into MR- and GR-mediated components. We have identified novel pathways, validated physiologically, that may explain some of the pathophysiological consequences of HPA hypertension.

MATERIALS AND METHODS

ACTH treatment.

Groups of age-matched male C57BL/6 mice (Harlan Olac) weighing ∼25 g were fed a diet containing 0.3% Na (SDS Diets, Witham, UK) with free access to water in a temperature- and light-controlled (12:12-h light-dark cycle) room. On day 0 an osmotic minipump (model 2002, Alzet, Charles River) was implanted subcutaneously under halothane anesthesia. Pumps contained either ACTH (Synacthen, Ciba-Geigy; 2.8 μg/day) or vehicle (0.9% NaCl). Animal studies were undertaken under UK Home Office license, following review by the local ethics committee. All experiments were performed in accordance with the UK Home Office Animals (Scientific Procedures) Act of 1986.

Metabolic cage studies.

For metabolic studies, mice were housed individually in cages (model 3600M021, Techniplast) with free access to food and water as described above. After a 4-day period of acclimatization, daily food and fluid intakes and body weights were recorded and urine samples were collected for later analysis. Measurements were made for three consecutive days to obtain baseline recordings. Food and water intake and excretion of steroids and electrolytes were stable during this period and were therefore averaged to generate a single baseline value (0 in Figs. 1–3). Experimental measurements were then made for 12 further days (days 1–12 in Figs. 1–3) after minipump implant. Mice received either ACTH (n = 10) or saline vehicle (n = 10). In a subset of animals, GFR was measured from the clearance of FITC-inulin, as described previously (6).

Tissue collection.

Further cohorts of ACTH- and vehicle-treated mice were prepared for tissue sampling and blood pressure measurements. Mice were housed in pairs, and blood pressure was measured by tail-cuff plethysmography as previously described (14). Mice were killed by CO2 inhalation after restraint stress, maximizing plasma corticosterone levels. Heparinized blood was collected by cardiac puncture, and plasma was stored for measurement of electrolytes, osmolality (freezing point depression), and hormones (32).

Urinalysis.

Urinary sodium, potassium, and calcium were measured by flame photometry (BWB Technologies UK). Phosphate was measured by a colorimetric method (QuantiChrom DIPI500, BioAssay Systems). Urinary aldosterone, deoxycorticosterone, and corticosterone were measured by specific ELISAs (1, 2). The effects of ACTH on urinary aldosterone levels were used to validate an ELISA for mouse urine and have been published previously (1).

Creatinine in urine was determined with the creatininase/creatinase-specific enzymatic method (8) and a commercial kit (Alpha Laboratories UK) adapted for use on a Cobas Fara centrifugal analyzer (Roche Diagnostics). Within-run precision was coefficient of variation (CV) < 3%, while intrabatch precision was CV < 5%.

Mouse albumin measurements were determined with a commercial microalbumin kit (Olympus Diagnostics) adapted for use on a Cobas Fara centrifugal analyzer. The immunoturbidimetric assay was standardized against purified mouse albumin standards (Sigma Chemical UK), with samples diluted in phosphate-buffered saline as appropriate.

RNA preparation and microarray processing.

Kidney RNA for microarray analysis was prepared by the TRIzol method and then processed through standard Affymetrix protocols, with one round of cDNA amplification. Processed RNAs were hybridized to the Affymetrix Mouse Genome 430 2.0 GeneChip. RNA processing and microarray analyses were carried out by ARK Genomics (Roslin Institute, Edinburgh, UK). Data were extracted through the GCOS software, and CEL files were used for further data processing. CEL files were imported into Bioconductor (18), normalized by RMA in the Affy (17) module, and statistically analyzed with the Limma (41) package and also with the Rank Products (RankProd) package (9). Gene Ontology (4) and KEGG pathway enrichment analysis was done with the Webgestalt tool (49). Microarray data are available in the ArrayExpress database (www.ebi.ac.uk/arrayexpress) under accession number E-MEXP-1761.

Real-time RT-PCR.

Selected genes were quantified by preoptimized RT-PCR assays in kidneys from separate cohorts of vehicle- and ACTH-treated mice. Total RNA was extracted from tissue samples with the Qiagen RNeasy system and reverse transcribed into cDNA with random primers and the QuantiTect DNase/reverse transcription kit (Qiagen, Crawley, UK). cDNA (equivalent to 1 ng total RNA) was incubated in triplicate with gene-specific primers and fluorescent probes (using predesigned assays from Applied Biosystems, Warrington, UK) in 1× Roche LightCyclerR480 Probes mastermix. PCR cycling and detection of fluorescent signal was carried out with a Roche LightCyclerR480. A standard curve was constructed for each primer probe set with a serial dilution of cDNA pooled from all samples. Results were corrected for the mean of expression of β-actin and 18S ribosomal RNA. Neither 18S RNA nor β-actin was affected by ACTH treatment.

Statistical analyses.

Data are presented as means ± SE. After tests for Gaussian distribution, comparisons were made by either unpaired t-test or ANOVA with Bonferroni post hoc test. For cumulative balance data, box plots are presented and medians are compared with the Mann-Whitney test.

RESULTS

Adrenocortical function.

Adrenal mass and plasma corticosterone were increased threefold by ACTH treatment (Table 1). Urinary corticosterone values started to increase after 1 day of ACTH treatment (Fig. 1) and continued to rise, reaching a plateau at day 10, when rates of excretion were ∼65-fold higher than in vehicle-treated control mice (P < 0.001). Urinary aldosterone also increased initially, reaching a peak that was 9.6-fold higher than in control mice at day 1 (Fig. 1). Thereafter, levels declined and had returned to pretreatment values by day 10. Urinary deoxycorticosterone followed a pattern similar to that of corticosterone and at the end of the ACTH treatment period was 80-fold higher than control mice (Fig. 1). It is notable that in a previous study of a mouse model of congenital adrenal hyperplasia, hypertension was observed with similar high levels of urinary deoxycorticosterone when both urinary corticosterone and aldosterone were reduced (32).

Table 1.

Body weight, organ weight, blood pressure, and plasma metabolite data for C57BL/6J mice treated with vehicle or ACTH

	Vehicle	P Value	ACTH
PNa, mmol/l	142.6 ± 1.0	<0.01	147.9 ± 1.5
PK, mmol/l	4.79 ± 0.23	<0.05	3.86 ± 0.27
POsm, mosmol/kgH2O	306 ± 1	<0.01	316 ± 2
Hematocrit, %	42.4 ± 1.1	<0.001	48.2 ± 1.1
Pcort, nmol/l	133 ± 29	<0.01	745 ± 124
PRA, ng·ml−1·h−1	11.6 ± 2.2	NS	8.0 ± 0.88
PRC, ng·ml−1·h−1	1,520 ± 210	NS	1,991 ± 282
AGT, ng·ml−1·h−1	304 ± 23	NS	312 ± 20
ΔBody wt, g	1.5 ± 0.25	<0.05	2.36 ± 0.38
Heart wt, mg/g body wt	6.39 ± 0.25	NS	5.98 ± 0.24
Adrenal wt, mg/g body wt	0.11 ± 0.1	<0.001	0.28 ± 0.03
Thymus wt, mg/g body wt	1.75 ± 0.11	<0.001	0.63 ± 0.11
Systolic BP, mmHg	104.8 ± 2.4	<0.01	127.8 ± 2.9
GFR, ml/min	0.43 ± 0.08	<0.05	0.62 ± 0.01

Data are means ± SE for C57BL/6J mice treated with either vehicle (n = 8) or ACTH (n = 8) for 14 days. PNa, plasma sodium; PK, plasma potassium; POsm, plasma osmolality; Pcort, plasma corticosterone; PRA, plasma renin activity; PRC, plasma renin concentration; AGT, angiotensinogen; ΔBody wt, change in body weight; BP, blood pressure; GFR, glomerular filtration rate. Statistical comparisons were made by t-test, and P values are as given; NS, not significant.

Fig. 1.

Daily urinary excretion of corticosterone (A), aldosterone (B), and deoxycorticosterone (C). Values are means ± SE; n = 5 (ACTH, ○) and 6 (vehicle, ●) except for deoxcorticosterone control values, where n = 3 because of sample insufficiency.

Water and electrolyte balance.

Urine flow rate was increased almost from the start of treatment (Fig. 2) and was matched by increased drinking (data not shown). By the second week of treatment, however, the polydipsia exceeded the polyuria and ACTH-treated mice showed a significantly more positive cumulative water balance (Fig. 2). Antinatriuresis was observed on the first experimental day in both ACTH- and vehicle-treated mice (Fig. 2). This probably relates to the effects of surgery, since food (and therefore sodium) intake during the first 24 h was similarly reduced. By day 2 of treatment, food intake had recovered to preoperative levels in both groups of mice and was not thereafter affected by ACTH. Control mice maintained a neutral sodium balance throughout the experiment. In contrast, ACTH treatment induced a significant natriuresis. Indeed, sodium excretion exceeded intake from day 3 onward, and ACTH-treated mice were in a negative sodium balance during both the first and second weeks (Fig. 2). ACTH did not affect urinary potassium excretion (data not shown), but these mice were nevertheless mildly hypokalemic (Table 1). The urinary Na-to-K concentration ratio was thus significantly higher in ACTH-treated mice than in control mice (Fig. 2). Although ACTH-treated mice were in a positive water balance, there was no evidence for expansion of plasma volume. Indeed, hematocrit was significantly higher in mice receiving ACTH (Table 1): the hypernatremia and increased plasma osmolality may therefore reflect the concentrating effects of a reduction in plasma volume. Systolic blood pressure was significantly increased (Table 1). Chronic ACTH treatment increased GFR (Table 1) and induced the development of albuminuria (Fig. 2), particularly during the second phase of the experimental regimen.

Fig. 2.

A: daily urine flow rate in mice receiving either ACTH (■; n = 5) or saline vehicle (□; n = 6) for 12 days. Baseline excretion (day 0) is the mean of measurements obtained on 3 consecutive days before implant of the osmotic minipump. B: cumulative sodium balance for days 1–6 and 7–12 of treatment. Open bars, vehicle; filled bars, ACTH. C: daily sodium excretion in mice receiving ACTH (■) or saline vehicle (□). D: cumulative sodium balance for days 1–6 and 7–12 of treatment. Open bars, vehicle; filled bars, ACTH. E: urinary sodium-to-potassium concentration ratio. □, Vehicle; ■, ACTH. F: urinary albumin-creatinine ratio. Open bars, vehicle; filled bars, ACTH. For A, C, E, and F, data are means ± SE. Statistical comparisons were made with 2-way ANOVA, which showed in both cases a highly significant effect of treatment (P < 0.001). Post hoc comparisons were made with Bonferroni test. *P < 0.05; **P < 0.01. For B and D, medians and ranges are shown and comparisons were made with the Mann-Whitney test.

Organ weights.

Body and organ weight data are summarized in Table 1. Thymus weights were reduced by ACTH, in keeping with modulatory effects of excess glucocorticoid activity. Heart and kidney were unaffected, and liver was slightly increased. Body weight gain across the experimental cohorts was consistently greater in ACTH-treated mice than in control mice. In mice used for microarray studies, those treated with ACTH showed a significant gain compared with vehicle-treated control mice; in contrast, there were no significant differences in animals placed in metabolic cages.

Microarray analysis.

Gene expression analysis revealed a relatively small degree of differential gene expression between control and ACTH-treated mice, in terms of both numbers of genes and the extent of modulation. Only 186 probe sets were differentially expressed by twofold or more (Supplemental Table S1). Only 12 probe sets were differentially expressed by 5-fold or more, and only 3 exceeded 10-fold. After accounting for redundant probe sets (probe sets mapping to the same gene annotation), this corresponded to 85 genes at least twofold upregulated by ACTH and 64 genes similarly downregulated. Gene function (Gene Ontology and KEGG pathway) enrichment analysis revealed several biological pathways that were statistically significantly affected in the up- and downregulated gene sets. Among these pathways were those involved in glucocorticoid hormone signaling, calcium homeostasis, and xenobiotic metabolism. To validate the expression data from the microarray studies, quantitative real-time RT-PCR analysis was carried out on selected genes (Table 2). Parametric (Limma) and nonparametric (Rank Products) statistical tests gave results for selection of differential expressed gene lists similar to the fold change criteria.

Table 2.

RT-PCR data for selected genes

Symbol	Gene	Vehicle	P Value	ACTH
Fkbp5	FK 506 binding protein 5	42 ± 7	0.01	196 ± 52
Aqp4	Aquaporin 4	43 ± 8	0.05	104 ± 22
Cyp2b10	Cytochrome P-450 family 2 subfamily b, polypeptide 10	6 ± 6	0.01	280 ± 116
Ren1	Renin	422 ± 158	0.81	775 ± 412
Calb1	Calbindin-28K	34 ± 4	0.05	68 ± 16
Trpv5	Transient receptor potential cation channel, subfamily V	109 ± 11	0.05	67 ± 12
Cyp241a	Cytochrome P-450 family 24 subfamily b, polypeptide 1	284 ± 93	0.5	595 ± 299
Slc8a1	Solute carrier family 8a member 1	49 ± 7	0.4	65 ± 11

P values are for the t-test with biological replicates.

Mineralocorticoid-related genes.

The effects of ACTH on genes implicated in mineralocorticoid actions are shown in Table 3. Sgk1, which mediates mineralocorticoid effects on ion channel activity, and Scnn1a (amiloride-sensitive ENac, α-subunit) were upregulated (4.4- and 1.7-fold, respectively), although the effects were not statistically significant after microarray multiple testing correction. The expression of the other ENaC subunits was not altered. Similarly, Kras, Nedd4, and Fxyd4, which are also implicated in mineralocorticoid signaling, were not changed. Mineralocorticoid (but not glucocorticoid)-induced hypertension is usually associated with renin (Ren) suppression. Neither renin gene expression levels nor plasma renin levels were affected, although angiotensin-converting enzyme (Ace) was markedly suppressed. The lack of effect of ACTH on Ren expression was confirmed by RT-PCR (Table 2). MR (Nr3c2) gene expression was not affected. The apelin receptor (Agtrl1) was significantly downregulated by ACTH treatment. This may be relevant to blood pressure control since apelin is a vasoactive peptide. In the kidney, the apelin receptor is expressed in all nephron segments, with highest expression in the glomerulus (21). Apelin has complex effects on the pre- and postglomerular microvasculature regulating renal hemodynamics (21), and the physiological impact of decreased receptor expression during ACTH excess is unclear.

Table 3.

Microarray gene expression data for selected genes

Symbol	Gene	Vehicle	P Value	ACTH
Mineralocorticoid-related genes
Ren1	Renin	235 ± 29	0.97	295 ± 127
Ace	Angiotensin-converting enzyme	360 ± 21	0.11	160 ± 29
Agtrl1	Angiotensin receptor-like 1 (Apelin receptor)	122 ± 17	<0.01	15 ± 3
Sgk1	Serum/glucocorticoid-regulated kinase	309 ± 78	0.1	1,292 ± 229
Nr3c2	Mineralocorticoid receptor	29.7 + 0.9	0.94	28.7 ± 0.7
Kras	Kirsten-Ras	291 ± 1	0.94	307 ± 6
Sccn1	Epithelial sodium channel α	218 ± 19	0.077	330 ± 29
Nedd4	Neural precursor cell expressed, developmentally downregulated 4	984 ± 40	0.99	992 ± 26
Fxyd4	Fxyd containing ion transport regulator 4	674 ± 50	0.66	556 ± 22
Glucocorticoid-related genes
Nr3c1	Glucocorticoid receptor	442 ± 22	0.27	332 ± 24
Hsd11b1	11β-Hydroxysteroid dehydrogenase type 1	1,759 ± 248	0.04	654 ± 59
Hsd11b2	11β-Hydroxysteroid dehydrogenase type 2	570 ± 28	0.85	616 ± 39
Fkbp5	FK 506 binding protein 5	382 + 33	0.02	1,024 ± 70
Pck1	Phosphoenolpyruvate carboxykinase 1	2,950 ± 335	0.45	4,444 ± 685
Agt	Angiotensinogen	257 ± 13	0.21	431 ± 67
Atpa1	Na-K-ATPase	4,188 ± 117	0.43	5,274 ± 430
Slc12a1	Sodium potassium chloride cotransporter	2,937 ± 530	0.91	3,116 ± 151
Clcnka	Chloride channel Ka	271 ± 6	0.05	460 ± 33
Kcnj1	ROMK channel	1,024 ± 94	0.55	1,240 ± 50
Calcium regulation-related genes
Calb1	Calbindin-28K	3,238 ± 285	0.04	6,332 ± 332
Cyp27b1	Cytochrome P-450 family 27 subfamily b, polypeptide 1	18 ± 5	0.14	136 ± 65
Cyp24a1	Cytochrome P-450 family 24 subfamily b, polypeptide 1	472 ± 163	0.43	1,076 ± 182
Gc	Group-specific component	63 ± 6	0.47	148 ± 60
Sgk	Serum/glucocorticoid-regulated kinase	309 ± 78	0.1	1,292 ± 229
Organic ion transporters and metabolizing enzyme genes
Cyp2b10	Cytochrome P-450 family 2 subfamily b, polypeptide 10	23 ± 4	0.22	317 ± 104
Cyp4a14	Cytochrome P-450 family 4 subfamily a, polypeptide 14	350 ± 233	0.43	990 ± 374
Aldh1a1	Aldehyde dehydrogenase family 1, subfamily a1	45 ± 9	0.19	149 ± 40
Aldh1a7	Aldehyde dehydrogenase family 1, subfamily a7	137 + 17	0.16	378 ± 91
Gpx6	Glutathione peroxidase 6	202 ± 27	0.12	409 ± 52
Ttpat	Tocopherol (α) transfer protein	151 ± 11	0.09	395 ± 83
Sult1a1	Sulfotransferase family 1a phenol preferring, member 1	163 ± 15	0.02	488 ± 66
Slc22a7	Solute carrier family 22a member 7	1,702 ± 29	0.22	828 ± 246
Slc22a8	Solute carrier family 22a member 8	2,510 ± 266	0.01	900 ± 39
Slco1a1	Solute carrier organic anion transporter, member 1a1	2,381 ± 260	0.22	634 ± 261

P values are corrected for multiple testing.

Glucocorticoid-related genes.

Glucocorticoid response elements are found in the promoters of a large number of genes; those in Table 3 were selected on the basis of associations with glucocorticoid signaling, with well-established metabolic effects of glucocorticoids, or with previously identified renal target genes. GR (Nr3c1) and hydroxysteroid 11β-dehydrogenase 1 (Hsd11b1) expression were downregulated, possibly to mitigate the effects of excess hormone. Similarly, Fkbp5, an immunophilin protein that affects GR trafficking, lowers agonist binding affinity, and causes glucocorticoid resistance, was upregulated more than twofold. Agt, Pck1, and Tat (not shown) are classical GR-regulated genes but were not affected by ACTH. Of the glucocorticoid-inducible genes that have been implicated in the regulation of renal function, only the chloride channel Ka (Clcnka) was significantly affected by ACTH.

Calcium homeostasis.

Of the 85 genes upregulated by greater than twofold, 5 were annotated in the Gene Ontology (GO) as being involved in calcium homeostasis and the regulation of vitamin D activity (GO: 0006766 and GO: 0019842; also see Table 3). Sgk1 has been shown to control expression of calbindin-28K (Calb1), which binds and transports Ca2+ from apical to basolateral surfaces of epithelial cells. Calb1 is thought to act in concert with an apical membrane calcium channel (Trpv5) and a basolateral Na/Ca exchanger (Slc8a1) to promote calcium reabsorption in epithelial cells of the distal tubule. Accordingly, expression of these three genes was quantified by RT-PCR: increased expression of Calb1 was confirmed, Slc8a1 was unaffected (as it was in the microarray analysis), and Trpv5 (not represented on the microarray) was down- rather than upregulated.

Vitamin D also transcriptionally regulates Trpv5 (20), and, in the microarray, three genes controlling vitamin D activity were affected: cytochrome P-450, family 27, subfamily b, polypeptide 1 (Cyp27b1) encoding the 1α-hydroxylase enzyme responsible for the final step in the synthesis of the hormonally active 1,25-dihydroxyvitamin D; cytochrome P-450, family 24, subfamily a, polypeptide 1 (Cyp24a1) encoding the deactivating 24-hydroxylase enzyme; and group-specific component (Gc), otherwise known as vitamin D binding protein. Real-time PCR confirmed the magnitude of Cyp24a1 changes, but the effect was not statistically significant, as with the microarray data.

The mixed response of genes regulating renal calcium excretion precludes a coherent predictive model of renal calcium handling. We therefore measured urinary calcium and phosphate excretion directly, finding both to be increased strikingly by ACTH (Fig. 3). Notably, genes involved in renal phosphate reabsorption (e.g., Npt2) were not differentially expressed on the microarray.

Fig. 3.

Twenty-four-hour urinary excretion of calcium (A) and phosphate (B) in mice receiving either ACTH (■; n = 5) or saline vehicle (□; n = 6) for 12 days. Baseline excretion (day 0) is the mean of measurements obtained on 3 consecutive days before implant of the osmotic minipump. Data are means ± SE. Statistical comparisons were made with 2-way ANOVA, which showed in all cases a highly significant effect of treatment (P < 0.001). Post hoc comparisons were made with Bonferroni test. *P < 0.05; **P < 0.01.

Transporters and exchangers.

Slco1b1, Slc22a7, and Slc22a8 are expressed on basolateral membranes of proximal tubular cells and are thought to have a role in organic ion excretion. The expression of all three was decreased by ACTH (Table 3). A compensatory downregulation of excretion might be stimulated by the need to retain endogenous molecules either in the face of the catabolic effects of glucocorticoids or to offset the increased glomerular filtration induced by ACTH (6). It is notable that the renal expression of ABC transporters (generally involved in reabsorption of chemicals at the basolateral membrane) was unaffected. The expression of Ttpa [tocopherol (α) transfer protein] was increased twofold by ACTH. This protein mobilizes the antioxidant vitamin E into the circulation and may therefore be induced to limit the oxidative stress effects of glucocorticoid excess. Likewise, glutathione peroxidase activity also protects against oxidative damage by reducing lipid and hydrogen peroxides to alcohols and water. However, a role for the peroxidase isozyme encoded by Gpx6, the gene regulated by ACTH, has yet to be established.

The remaining genes in Table 3 encode enzymes involved in xenobiotic metabolism that prototypically are induced by dexamethasone. It is likely, therefore, that they are responding to excess glucocorticoids in the present experiment. Most are involved in lipid metabolism. Cyp4a14 changes are of particular interest since the enzyme catalyzes the conversion of arachidonic acid to a prohypertensive compound, 20-hydroxyarachidonic acid, and may also promote steroid hormone clearance. Sult1a1 is a glucocorticoid-inducible sulfotransferase enzyme that potentially stimulates steroid clearance by conversion to water-soluble sulfate metabolites.

Other interesting genes.

A complete list of genes regulated by ACTH is available in Supplemental Table S1. Among the upregulated genes, those that are potentially relevant to cardiovascular risk are ApoCIII and Angptl. These are inhibitors of lipoprotein lipase (Lpl), which is itself downregulated. The transporters Slc6a15 (a chloride-dependent amino acid neurotransmitter transporter) and Slc9a8 (NHE8) were downregulated.

DISCUSSION

The metabolic and renovascular effects of murine ACTH excess are of particular interest because of an expanding list of genetically modified mouse strains in which metabolic disease can be linked to activation of the HPA axis (7, 31, 32). In the present study plasma aldosterone and corticosterone were initially both stimulated by ACTH, but the effect on aldosterone was transient (34) and only the effect on corticosterone was sustained. Persistent glucocorticoid excess implicates GR-mediated pathways in the ACTH phenotype, but MR-mediated responses may also be physiologically important for two reasons. First, corticosterone will bind MR if concentrations are high enough to overcome the protective barrier of Hsd11b2 (25). Second, deoxycorticosterone, a weak MR agonist, was increased to a level of cardiovascular significance (32).

We anticipated that microarray analysis would define the effect of chronic ACTH on renal gene expression, and help to delineate between MR- and GR-mediated pathways. These studies have two major limitations. First, MR- and GR-mediated changes in gene expression often converge on the same pathways, and we are not able to unequivocally attribute changes in gene expression to a given receptor. This is exemplified by the more than fourfold increase in Sgk1, which is under the transcriptional control of both MR and GR (26). Second, relatively few cells within the kidney express MR, and it is possible that whole kidney analysis is insufficiently sensitive to detect physiologically significant changes in the expression of target genes. We were, for example, unable to detect significant changes in the expression of key sodium transport proteins such as α-ENaC (Scnn1a; 1.7-fold), Na-K-Cl cotransporter (NKCC)2 (Slc12a1; 1.09-fold), or Na-Cl cotransporter (NCC) (Slc12a3; 1.24-fold) despite the fact that they are regulated by corticosteroids and show altered physiological activity in ACTH-treated mice (6). Nevertheless, the correlation between transcriptional and physiological changes in the present model has been powerful and has identified important responses to ACTH in several different areas.

Sodium and water homeostasis.

We previously found (6) that ACTH excess stimulates renal tubular sodium reabsorption and increases ENaC activity via MR and GR pathways. In the present study, however, ACTH excess prompted a rapid and sustained natriuresis, as has been shown in rats (24). There was no evidence for sodium retention, and glomerular hyperfiltration may be sufficient to offset enhanced ENaC activity. Moreover, inhibition of sodium reabsorption in the proximal tubule is suggested by phosphaturia as well as the decreased expression of Na/H exchanger (NHE)8 and several organic anion transporters. Calciuria is consistent with impaired sodium reabsorption in the thick limb of Henle, and, combined, these events may explain the negative sodium balance observed in ACTH-treated mice. The suggestion that corticosteroid-induced effects in one nephron region may be compensated by secondary effects elsewhere is illustrated by a recent study in which human GR was overexpressed in the collecting duct of mice (33): GR-stimulated gene expression was offset by opposing changes in upstream nephron segments that did not express human GR, and blood pressure homeostasis was preserved. Countervailing effects such as these underpin the “escape” phenomenon and mitigate the pathophysiological consequences of long-term corticosteroid excess. Escape from the sodium-retaining effects of aldosterone is well documented, and the present study suggests that the kidney also adapts to excess glucocorticoid. At the molecular level, expression changes to three genes may combine to limit renal glucocorticoid signaling. First, GR (Nr3c1) was reduced to a level that, in the hippocampus for example, is sufficient to affect HPA activity (38). Second, an increase in Fkbp5, which binds GR, would reduce receptor affinity for glucocorticoids (47). In squirrel monkeys, for example, Fkbp5 is partly responsible for glucocorticoid resistance (13). Third, a reduction of Hsd11b1, which catalyzes the conversion of inactive 11-dehydro derivatives of glucocorticoid to the active parent compound, would reduce glucocorticoid activity at a local level, particularly in the renal medulla (10). However, it has been reported that HSD11B1 functions as a dehydrogenase in human proximal tubular cells (19), indicating that glucocorticoid activity in different sections of the tubule may be influenced by local redox potentials controlling the directionality of conversions between corticosterone and 11-dehydrocorticosterone as well as by differences in enzyme expression levels.

In contrast to sodium, ACTH-treated mice had a positive water balance. This may be consistent with overactivation of MR, but the effects of GR seem to predominate in terms of plasma volume. Increased hematocrit is strongly suggestive of a GR influence on vascular permeability (22), and we recently confirmed (6) contraction of plasma volume during ACTH excess. The expression of genes involved in renal water conservation (Aqp2, Avpr2, and Slc14a2, for example) was largely unchanged by chronic ACTH treatment. Aquaporin 4, which is constitutively expressed in the basolateral membrane of the collecting duct (43), was significantly upregulated and may play a permissive role for enhanced renal water reabsorption.

The complexity of the pathophysiological response to ACTH excess is illustrated by the separation of sodium and water homeostasis, and yet the pivotal role of sodium in the hypertensive response is clear: ACTH did not increase blood pressure in mice maintained on a diet that was essentially sodium free (6). We now propose that the development of hypernatremia is an important hypertensive mechanism: the redistribution of water outside the vascular compartment causes contraction-hypernatremia, enhancing the release of several vasoactive agents (6, 29) and exerting a significant pressor effect.

Potassium homeostasis.

Volume contraction may mean that potassium depletion is actually more severe than indicated by plasma concentration alone. Potassium depletion is considered a hallmark of mineralocorticoid excess, attributable to activation of MR in the distal nephron, but our data do not support this hypothesis: ACTH-treated mice indeed maintained an inappropriately robust potassium excretion, but there was no kaliuresis and mice were in neutral potassium balance throughout the experiment. We hypothesize that hypokalemia reflects a redistribution of potassium out of the vascular space coupled to a renal conservation defect. The mechanisms underpinning the redistribution remain unknown, but the renal defect may reflect both impaired reabsorption through the paracellular cation shunt in the thick limb of Henle (5) and persistent secretion in the distal nephron (6). Renal potassium channels were not identified as differentially expressed by microarray analysis. The expression level was, however, generally low. Renal outer medullary potassium channel (ROMK) (Kcnj1; 1.22-fold) was abundantly expressed, but its function is determined by splice variants and cannot be determined with whole kidney preparations.

Microarray analysis revealed novel effects of ACTH on at least six genes related to calcium metabolism. Glucocorticoids are well known to inhibit bone calcium reabsorption, and in the present study the urinary excretion of calcium and phosphate were both markedly increased by ACTH. Normally, filtered calcium is reabsorbed along the length of the tubule (30). In the proximal tubule and thick limb of Henle, which account for the bulk of reabsorption, calcium is reclaimed by both transcellular and paracellular mechanisms (15). In the distal convoluted tubule, reabsorption is exclusively transcellular, involving coordinated regulation of an apical channel, Trpv5, cytoplasmic calbindin 28K (Calb1), and basolateral Na/Ca exchanger (Slc8a1). In the distal tubule, calcium reabsorption is regulated by Sgk1: null mice have reduced levels of Calb1 and Trpv5 (40). In the present study, ACTH increased expression of both Sgk1 and Calb1, which would be predicted to reduce calcium excretion. However, Trpv5 was downregulated, and Slc8al was not significantly affected by ACTH treatment, which may contribute to calciuria. Regulation by Sgk1 is dependent on the sodium/hydrogen exchanger regulator Slc9a3r2, which was not stimulated by ACTH treatment, suggesting misregulation of calcium reabsorption in the distal nephron. Indeed, renal calcium reabsorption is controlled by calcium-sensing receptors within the kidney and by hormones such as parathyroid hormone and vitamin D and calcitonin. The microarray data did not reveal any effect on parathyroid hormone signaling but found increases in three genes associated with vitamin D (Cyp27b1, Cyp24a1, and Gc). Cyp27b1 encodes the 1α-hydroxylase enzyme that converts 25-hydroxyvitamin D (the main circulating form of vitamin D) to the biologically active 1,25-dihydroxyvitamin D. Cyp24a1 encodes an inactivating enzyme that hydroxylates vitamin D at the 24 position. Gc encodes a high-affinity serum vitamin D binding protein (DBP). The functional relevance of ACTH-induced changes in these three genes is difficult to explain. One possibility is that calciuria is secondary to the glucocorticoid-mediated inhibition of bone reabsorption. Increased Cyp27b1 could therefore be a compensatory mechanism, increasing active vitamin D and preserving calcium. Although DBP reduces the free active component of vitamin D in the circulation, it also facilitates the renal preservation of vitamin D stores and hence will promote the conservation of calcium (44). DBP also facilitates the uptake of active vitamin D in target tissues (12). An increase in Cyp24a1, in contrast, does not fit the hypothesis of compensatory vitamin D-dependent calcium uptake since 24-hydroxylation is an effective mechanism of negative-feedback control of 1,25-dihydroxyvitamin D synthesis. However, it has been suggested that Cyp24a1 expression in vitamin D target tissues may have a role in regulating hormone activity (28).

Xenobiotic metabolism.

We anticipated that ACTH would regulate genes involved in either the production of or the protection from damaging chemicals: glucocorticoid treatment is known to increase the synthesis of reactive oxygen species (21, 36) as well as offering protection in the kidney from inflammatory cell infiltration following treatment designed to cause renal tubular damage (35). In the present study, changes associated with damage limitation predominate, with a group of 10 or more genes encoding transporters and enzymes involved in xenobiotic metabolism. Interestingly, similar patterns of gene changes have been reported for the genetically obese and diabetic db/db mouse (11, 35, 48). This suggests coordinate control possibly mediated by one or more of a group of established transcription factors (Ahr, Car, Nr1i2, Pparα, Nrf2) for xenobiotic transport genes. However, the expression of none of these factors was affected in the microarray. The precise functions of the ACTH-affected xenobiotic transport genes are not well defined. It seems probable that the transporters are involved in the proximal transport of organic anions or cations, the clearance of which would be compromised by glucocorticoid-induced changes in renal function. These enzymes may be required to detoxify oxidized lipids and other active compounds generated as a consequence of glucocorticoid effects on metabolism.

Summary.

In conclusion, chronic ACTH treatment causes a sustained increase in circulating glucocorticoids. The pattern of electrolyte excretion and microarray analysis of gene changes reveal little that could be ascribed directly to MR. Although GR-dependent changes on renal electrolyte metabolism and gene transcription were observed, the link between these changes and blood pressure control will require further investigation. Many of the gene changes could be considered to be part of a glucocorticoid escape mechanism. GR signaling appeared to be downregulated, and at least 10 genes involved in the metabolism of xenobiotics were affected. The downregulation of Hsd11b1 suggests that local glucocorticoid signaling may well be altered, and the link between enzyme activity and altered redox state warrants further investigation. ACTH induced a marked calciuria and phosphaturia. Although genes controlling vitamin D activity and also renal epithelial calcium transport were affected, further studies to localize sites of expression of these genes are needed to determine their role in causing or counteracting the calciuretic effect of glucocorticoids.

GRANTS

D. R. Dunbar was funded by the Wellcome Trust Cardiovascular Research Initiative and Functional Genomics Initiative. J. J. Mullins is a recipient of a Wellcome Trust Principal Research Fellowship. L. C. Evans was funded by a British Heart Foundation Ph.D. studentship. E. A. S. Al-Dujaili, L. J. Mullins, and C. J. Kenyon were funded by the UK Medical Research Council. M. A. Bailey was a Wellcome Trust Intermediate Fellow. This work was partially funded by a British Heart Foundation CoRE Award.

DISCLOSURES

No conflicts of interest are declared by the author(s).