Steroids 78 (2013) 347–355

Contents lists available at SciVerse ScienceDirect

Steroids

journal homepage: www.elsevier .com/locate /s teroids

Adrenalectomy amplifies aldosterone induced injury in cardiovascular tissue:An effect attenuated by adrenally derived steroids

Andrew S. Brem a, David J. Morris b, Xiangpo Li a, Yan Ge a, Sunil Shaw c, Rujun Gong a,⇑a Division of Kidney Disease and Hypertension, Rhode Island Hospital, The Warren Alpert School of Medicine at Brown University, Providence, RI, United Statesb Department of Laboratory Medicine, Miriam Hospital, The Warren Alpert School of Medicine at Brown University, Providence, RI, United Statesc Department of Pediatrics, Woman and Infants Hospital of Rhode Island, The Warren Alpert School of Medicine at Brown University, Providence, RI, United States

a r t i c l e i n f o

Article history:Received 17 May 2012Received in revised form 30 November 2012Accepted 12 December 2012Available online 31 December 2012

Keywords:11b-Hydroxysteroid dehydrogenaseVascular endothelial cells11-DehydrocorticosteroneMineralocorticoidsFibrosisInflammation

0039-128X/$ - see front matter � 2012 Elsevier Inc. Ahttp://dx.doi.org/10.1016/j.steroids.2012.12.007

⇑ Corresponding author.E-mail address: Rujun_Gong@Brown.edu (R. Gong

a b s t r a c t

Aldosterone induces fibrotic changes in cardiovascular tissues but its effects have usually been demon-strated in models of pre-existing renal injury and/or hypertension. This study tests the hypothesis thataldosterone can directly induce vascular fibrotic changes in the absence of prior renal injury or hyperten-sion. Experiments were conducted in intact or adrenalectomized (ADX) mice. Mice were divided intogroups and treated for 1 week with vehicle or aldosterone (8 lg/kg/day) ± inhibitor (800 lg/kg/day):CONTROLS, mice treated with aldosterone, ADX-CONTROLS, ADX + corticosterone (CORT 8 lg/kg/day),ADX with aldosterone, ADX with aldosterone plus the mineralocorticoid receptor (MR) antagonist RU-318, ADX with aldosterone + CORT (CORT inhibitor dose), and ADX with aldosterone + 11-dehydro-CORT.Aortic smooth muscle to collagen ratio, aorta intimal thickness (lm), heart weight/body weight ratio(mg/gm), and left ventricular collagen (%) were measured. Prior to sacrifice, blood pressures were normalin all animals. Lower dose CORT alone had no effect on any of the variables examined. Aldosterone expo-sure was associated with extra-cellular matrix accumulation in cardiovascular tissues in intact mice andadrenalectomy exacerbated these effects. RU-318, CORT (inhibitor dose), and 11-deydro-CORT eachattenuated the early fibrotic changes induced by aldosterone. In the heart, aldosterone exposure affectedall the parameters measured and caused intimal hypercellularity with monocytes adhering to endothelialcells lining coronary vessels. Cultured endothelial cells exposed to aldosterone (10 nM) released E-selec-tin, produced collagen, and promoted monocyte adhesion. These effects were inhibited by RU-318 and11-deydro-CORT but not by CORT. Thus, adrenalectomy enhances aldosterone induced early fibroticchanges in heart and aorta. Aldosterone initially targets vascular endothelial cells. MR antagonists and11-dehydro-CORT, an 11b-HSD dehydrogenase end-product, directly attenuate these effects.

� 2012 Elsevier Inc. All rights reserved.

1. Introduction

Since its discovery in the early 1950’s, aldosterone is bestknown for its actions in the kidney regulating sodium, potassium,and acid–base balance. In the last decade or so, numerous investi-gators have shown that aldosterone also has a darker side, promot-ing inflammatory and fibrotic changes in the kidney, heart, andblood vessels [1–6]. Most of the studies documenting the aldoste-rone induced fibrotic changes were conducted in animals withsome form of pre-existing renal injury and/or hypertension [2–5].Moreover, sodium loading, pharmacologic doses of mineralocortic-oids, and extended periods of time were often required for the fi-brotic changes to occur [2,4]. In the kidney and vascular tissue,chronic mineralocorticoid activation can trigger local productionof pro-inflammatory and pro-fibrotic cytokines mediated through

ll rights reserved.

).

the SGK (serum and glucocorticoid kinase) pathway [7,8]. Aldoste-rone induced fibrotic changes in renal collecting duct cells requiremineralocorticoid receptor (MR) activation and protein synthesisto occur [7]; events similar to those associated with aldosterone in-duced electrolyte transport [9,10]. Our laboratory recently showedtwo additional novel findings in renal tissues [11]: first, that cul-tured kidney collecting duct epithelial cells synthesize excess col-lagen after being exposed to aldosterone (10 nM) for only 48 hand second, that kidneys harvested from mice continuously in-fused with aldosterone exhibit early fibrotic changes after only1 week, an effect independent of prior renal injury or blood pres-sure. The degree of fibrotic change was significantly greater in micethat underwent adrenalectomy just prior to the study suggestingthat a factor derived from the adrenal gland might normally pro-tect against aldosterone’s actions in the kidney.

The idea that products emanating from the adrenal gland mightmodify the actions of aldosterone was based on earlier observa-tions involving renal electrolyte transport. Years ago, investigators

348 A.S. Brem et al. / Steroids 78 (2013) 347–355

noted that the usual antinatriuretic effects of physiological dosagesof aldosterone were not readily apparent in adrenally intact ani-mals [12–14]; the expected reproducible increase in trans-epithe-lial sodium transport following exposure to aldosterone was onlyseen in isolated preparations like the toad bladder [15,16] or in ani-mals that underwent prior adrenalectomy [12,14]. Alberti andSharp [17] and later others [18] showed that 11-dehydro-metabo-lites of the endogenous glucocorticoids corticosterone and cortisolsuppress aldosterone stimulated sodium transport and appear toprevent the translocation of aldosterone–MR complex from thecytoplasm to the nucleus of target cells [19]. As mentioned above,11-dehydrocorticosterone also prevented the pro-fibrotic effects ofaldosterone in isolated cultured renal collecting duct cells and innormotensive adrenalectomized mice continuously infused withaldosterone [11]. The observation in isolated renal collecting ductcells is particularly important since those cells do not have theability to enzymatically convert 11-dehydrocorticosterone backto its parent, corticosterone [11].

Animals exposed to aldosterone or DOCA (deoxycorticosteroneacetate) over time do exhibit inflammatory and fibrotic changesin the heart and major arteries [4,20]. While hypertension is usu-ally present, the authors of recent studies have suggested thatthe hypertension may not be the primary causative factor inducingthe inflammation and subsequent fibrosis [21]. Mineralocorticoidreceptor (MR) activation appears necessary however since MRantagonists seem to attenuate the inflammatory and fibroticchanges [5,22]. Our understanding of how aldosterone initiatesthe fibrotic process in vascular tissues and which cells in thevascular bed are initially targeted are areas that need to beexplored.

The present series of experiments were designed to test thehypothesis that aldosterone can directly induce pro-inflammatoryand pro-fibrotic changes in vascular tissue from otherwise normalanimals. Using our novel murine model [11], we now show thataldosterone, at a concentration 8 lg/kg/day, continuously adminis-tered over a week can also induce pro-inflammatory and pro-fibro-tic changes in vascular tissue in the absence of hypertension andprior injury. Vascular endothelial cells appear to be an initial targetfor aldosterone. Moreover, these changes in the animal model canbe blocked by exogenously administered MR antagonists, by corti-costerone, and by 11-dehydro-corticosterone, a metabolic deriva-tive of the glucocorticoid corticosterone, which can be eitherlocally generated by vascular 11b-hydroxysteroid dehydrogenase(11b-HSD) [23] or is available from the circulating plasma [24].In the circulation, endogenous glucocorticoids like corticosteroneand cortisol are almost entirely protein bound but the proteinbinding of the 11-dehydro derivatives is limited [25,26] makingthese agents potentially more bio-available.

2. Methods

2.1. Animal studies

Male C57BL/6 mice that weighed between 20 and 25 g werehoused in the Central Research Facilities of Rhode Island Hospital,which is an AAALAC accredited facility, and studies were per-formed under an approved IACUC protocol. All animals were fedon standard rodent chow containing 0.4% dietary sodium. Micewere anesthetized and received either a sham operation or bilat-eral adrenalectomy. Only adrenalectomized mice were given 0.9%saline as drinking water for the duration of the study period to in-sure their survival. All mice received an Alzet micro pump im-planted subcutaneously. The pump contained either vehicle orindicated steroids. Mice were randomly assigned to one of the fol-lowing groups (n = 4): (1) Sham Control: sham operated mice were

given vehicle (DMSO) as a continuous subcutaneous infusion ofvehicle by micro pump for 7 days; (2) Sham + Aldosterone: shamoperated mice received a continuous subcutaneous infusion ofaldosterone (Aldo) by micro pump at a dose of 8 lg/kg/day for7 days; or (3) ADX Control: Adrenalectomized mice were givenvehicle as a continuous subcutaneous infusion of vehicle by micropump for 7 days; (4) ADX + Aldo: Adrenalectomized mice receiveda continuous subcutaneous infusion of aldosterone by micro pumpat a dose of 8 lg/kg/day for 7 days; (5) ADX + Corticosterone:Adrenalectomized mice received a continuous subcutaneous infu-sion of corticosterone (Cort) by micro pump at a dose of 8 lg/kg/day for 7 days; (6) ADX + Aldo + A: Adrenalectomized mice re-ceived a continuous subcutaneous infusion of aldosterone 8 lg/kg/day and 11-dehydrocorticosterone (A) by micro pump at a doseof 800 lg/kg/day for 7 days; (7) ADX + Aldo + Cort: Adrenalecto-mized mice received a continuous subcutaneous infusion of aldo-sterone 8 lg/kg/day and corticosterone by micro pump at a doseof 800 lg/kg/day for 7 days; (8) ADX + Aldo + RU-318: Adrenalec-tomized mice received a continuous subcutaneous infusion ofaldosterone 8 lg/kg/day and RU-318 by micro pump at a dose of800 lg/kg/day for 7 days. Seven days after surgery, animals weresacrificed. The hearts and aortae were perfused with iced salineand segments were fixed in fixation solutions, frozen for cryostatsectioning, or snap-frozen in liquid nitrogen and stored at �80 �C.

2.2. Blood pressure

Just before sacrifice, mice were anesthetized and prepared for ameasurement of mean arterial pressure. A polyethylene catheterwas inserted into the femoral artery and a computer running WIN-DAS software (DATAQ Instruments, Akron, OH) continuously mon-itored the blood pressure using a pressure transducer.

2.3. Cardiac and aortic histology

Formalin-fixed hearts and thoracic aortae were embedded inparaffin and prepared in 3 lm-thick sections. For general histology,sections were processed for hematoxylin/eosin, periodic acid-Schiff, and Masson-Trichrome staining. A semi quantitative mor-phometric score index was used to evaluate the degree of intersti-tial extra cellular matrix accumulation [3,27]. Presence of extracellular matrix was documented by Trichrome staining of collagen.The magnitude of cardiac and aortic interstitial collagen accumula-tion was graded from 0 to 3 [28]: 0, absent; 1, mild; 2, moderate;and 3, severe. A mean score was calculated using the values ob-tained in 20 random high-power (400�) fields per mouse in fourmice per group. All sections were examined without knowledgeof the treatment protocol.

2.4. Cell culture studies

Human umbilical vein endothelial cells (HUVEC) were pur-chased from VEC Technologies (Rensselaer, NY) and maintainedin MCDB-131 complete medium. HUVEC (2–8 passages) wereseeded on gelatin-coated (1.5%) cultures at approximately 80%confluence. After growth for 24 h in complete medium, cells under-went serum starvation for 6 h in Medium 199. Human monocytes(THP-1) were purchased from American Type Culture Collection(ATCC; Manassas, VA) and were cultured in RPMI supplementedwith 10% FBS. For fluorescence-viable labeling, THP-1 (1 � 107)were incubated in medium that contained 5 ug/ml Calcein-AM(Invitrogen, Carlsbad, CA) at 37 �C for 30 min. Excess dye wasremoved by washing three times with PBS.

A.S. Brem et al. / Steroids 78 (2013) 347–355 349

2.5. Static monocytic adhesion assay

Adhesion studies were performed with the human monocyticcell line THP-1 under static conditions. Fluorescence-labeledTHP-1 cells were re-suspended in Medium 199 (1 � 106 cells/ml).For static adhesion assays, HUVEC monolayers in 12-well plateswere treated with aldosterone 10 nM ± RU-318, corticosterone, or11-dehydro-corticosterone 1 lM for the stated intervals. HUVECcultures were washed twice with PBS before addition of 1 ml(1 � 106) of labeled THP-1 cells per well. The plates were incubatedfor 30 min at 37 �C. After incubation, the monolayers were washedgently three times with PBS and photomicrographs were takenusing a phase-contrast microscope. Adherent monocytes werelysed with RIPA buffer (1% Nonidet P-40, 0.1% SDS, 100 ug/mlPMSF, 0.5% sodium deoxycholate, 1 mM sodium orthovanadate,2 ug/ml aprotin, 2 ug/ml leupeptin, and 5 mM EDTA in PBS), andfluorescence intensity was measured by a Spectramax GEMINIEM fluorescence plate reader (Molecular Devices, Sunnyvale, CA)at an excitation wavelength of 485 nm and emission at 530 nm.HUVEC monolayers that were adhered with non-labeled THP-1cells served as negative controls.

2.6. Sirius Red assay and staining

Sirius Red specifically stains collagens I and III, which are earlymarkers of fibrosis. To assess the quantitative production of extra-cellular matrix substances in cultured HUVEC, the cells were grownto confluence and treated as indicated. Cells were harvested andsonicated for Sirius Red assays. In brief, ammonium sulfate wasadded to the sonicated cells and the contents were incubated un-der a slow constant rocking motion at 4 �C for 24 h and was thencentrifuged. The pellet was re-suspended in acetic acid. An aliquotof the re-suspended sample was mixed with Sirius Red (50 lM)solution (Sigma, St. Louis, MO) made up in 0.5 M acetic acid. After30 min of constant mixing, the pellet was prepared by centrifugeand re-suspended in 1 ml of potassium hydroxide. The optical den-sity of the samples was individually read with a spectrophotometerat a wavelength of 540 nm. In separate but related studies, HUVECcells were grown on slides but under the same culture conditions.At the completion of the experiment, the cells were directly stainedwith Sirius Red and examined under the light microscope.

Fig. 1. Aldosterone induces fibrotic changes both in aorta intimal and medial layers. Panaortae from normal and adrenalectomized (ADX) mice after 1 week. Aldo exposure mCorticosterone (8 lg/kg/day) exposure had minimal effect over the same time period. Pa11-dehydrocorticosterone (a) each attenuated the effects of aldo. The trichrome stained

2.7. Western immunoblot analysis

HUVEC monolayers were lysed in RIPA buffer. Protein concen-trations were determined using a bicinchoninic acid protein assaykit (Sigma, St. Louis, MO, USA). Samples with equal amounts of to-tal protein (60 mg/ml) were fractionated by 7.5–10% SDS–poly-acrylamide gels under reducing condition and analyzed byWestern immunoblot as described previously. The antibodiesagainst E-selection and actin were purchased from Santa Cruz Bio-technology (Santa Cruz, CA).

2.8. Western blot for MR

HUVEC at passage four were lysed in Laemmli buffer (4% SDS,20% glycerol, 10% 2-mercaptoethanol, 0.004% bromphenol blue,and 125 mM tris pH 6.8) resolved on SDS–PAGE and blotted tonitrocellulose membranes. Blots were probed with anti-MR anti-body [29] (1:500 rMR1–18 1D5; Developmental Studies Hybrid-oma Bank, U. of Iowa). Second antibody was used at 1:2500(goat-anti mouse HRP, Thermo Scientific) followed by chemilumi-niscence detection and a Biorad Gelblot to capture images.

2.9. Statistical analyses

One investigator in a blinded manner performed computerizedmorphometric analysis as well as histological scoring. For immu-noblot analysis, bands were scanned and the integrated pixel den-sity was determined using a densitometer and the NIH imageanalysis program. All data are expressed as mean ± SD. Statisticalanalysis of the data from multiple groups was performed by ANO-VA followed by Student-Newman–Keuls tests. Data from twogroups were compared by t-test. P < 0.05 was consideredsignificant.

3. Results

3.1. Aldosterone induces extracellular matrix accumulation andendothelial inflammation in cardiovascular tissues

Two major groups of mice were studied; sham operated adre-nally intact controls and mice adrenalectomized at day 1. All of

el a depicts the effects of aldosterone 8 lg/kg/day (aldo) and corticosterone on thearkedly enhanced the intimal and medial segments especially in the ADX mice.nel b shows the effect of aldo exclusively in ADX mice. RU-318, corticosterone, andphotomicrographs are magnified 200� times.

350 A.S. Brem et al. / Steroids 78 (2013) 347–355

the ADX mice survived the week of study. Blood pressures werenormal in all animals prior to sacrifice on day 7. Previously re-ported plasma aldosterone levels were 671 ± 134 pg/ml in controlsreceiving vehicle, 3410 ± 357 pg/ml in adrenally intact mice in-fused with aldosterone 8 lg/kg/day, 57 ± 27 pg/ml in ADX mice,2868 ± 742 pg/ml in ADX mice infused with aldosterone and46 ± 34 pg/ml in ADX mice infused with corticosterone [11]. Plas-ma corticosterone was only measured in control and ADX mice(242.6 ± 31.6 ng/ml and 17.6 ± 10.1 ng/ml respectively) to docu-ment the effects of adrenalectomy [11].

Vascular tissue from mice treated with aldosterone exhibitedextracellular matrix accumulation, a characteristic early fibroticchange by light microscopy, with a reduced aortic smooth muscleto collagen ratio and aortic intimal thickening (Fig. 1a). Mice trea-ted with aldosterone also demonstrated an increase in the heart tobody weight ratio and showed a greater volume of collagen in the

455

*

00.51

1.52

2.53

3.54

4.5

Aor

ta s

moo

th m

uscl

e to

col

lage

n ra

tio

*&

414.24.34.44.5

Vehicl

eAldo

Vehicl

eCort

Aldo

Aldo+A

Aldo+RU31

8

Aldo+Cort

Vehicl

eAldo

Vehicl

eCort

Aldo

Aldo+A

Aldo+RU31

8

Aldo+Cort

Vehicl

eAldo

Vehicl

eCort

Aldo

Aldo+A

Aldo+RU31

8

Aldo+Cort

*&

3.73.83.94

4.1

Hea

rt to

bod

yw

eigh

t rat

io (m

g/g)

00.51

1.52

2.53

3.5

*

&

0Num

ber o

f adh

eren

tm

acro

phag

e pe

r ves

sel

(a)

(c)

(e)

Fig. 2. Morphologic changes induced by aldosterone in aorta and heart. Panel a demontreated with aldo for 1 week. Aldo exposure is associated with a significant drop in the rathickness in aortae from normal and ADX mice treated with aldo for 1 week. Aldo treatmPanel c depicts the heart to body ratios from normal and ADX mice treated with aldo for 1pronounced in ADX mice. Panel d illustrates the left ventricular collagen volume as a pmarked rise in the ratio most notable in the ADX mice. Panel e shows the effect of aldo eThe effect of aldo was most clearly seen in ADX mice. In all the experiments, RU-318, c⁄p < 0.001 aldo vs either control or ADX control and & p < 0.05 aldo vs either control or

left ventricle (Fig. 2); most of the collagen was peri-vascular in themyocardium (Fig. 3a). More importantly, small coronary vessels inthe hearts of aldosterone treated mice exhibited an increase inwall thickness, an increase in cellularity (Fig. 3a) and vascularendothelial dysfunction with inflammation seen by immunohisto-chemical staining of F4/80 positive monocytes and counting of F4/80 positive monocytes adhering to vascular endothelial cells(Fig. 4).

3.2. Adrenalectomy enhances the aldosterone induced pro-fibrotic andpro-inflammatory changes

In adrenalectomized mice, 8 lg/kg/day of corticosterone alonehad no effect on the aorta or the heart by histologic examination.In contrast, aldosterone treatment for 1 week resulted in greaterextracellular matrix accumulation in both aorta (Fig. 1b) and heart

8090100 * &

010203040506070

A C A

Aor

ta In

tima

med

iath

ickn

ess

(um

)

456

Vehicl

eAldo

Vehicl

eCort

Aldo

Aldo

Aldo+A

Aldo+RU31

8

Aldo+RU31

8

Aldo+Cort

Aldo+Cort

*

&

0123

Vehicl

eAldo

Vehicl

eCort

Aldo+A

Left

vetr

icul

arco

llage

n vo

lum

e (%

)

+AD

ADX

(b)

(d)

strates the smooth muscle to collagen ratio in aortae from normal and ADX micetio, which is more pronounced in ADX mice. Panel b shows the intimal/medial (I/M)

ent correlated with a significant increase in the thickness especially in ADX mice.week. Aldo exposure is associated with a significant rise in the ratio, which is more

ercent of the total in normal and ADX mice. Again, aldo treatment was linked to axposure on the adhesion of macrophages to the walls of micro-vessels in the heart.

orticosterone, and 11-dehydrocorticosterone (a) each blunted the response to aldo.ADX control.

Fig. 3. The effects of aldosterone were most noticeable in the micro-vasculature supplying the heart. Panel a depicts the effects of aldosterone (aldo) and corticosterone onthe hearts from normal and adrenalectomized (ADX) mice after 1 week. Aldo exposure was associated with early peri-vascular thickening and what appears to be an earlyinflammatory response within the micro-vessels. In addition, early mild interstitial fibrotic staining is seen again particularly in the ADX mice treated with aldo. Low dosecorticosterone exposure had minimal effect over the same time period. Panel b shows the effect of aldo exclusively in ADX mice. RU-318, corticosterone, and 11-dehydrocorticosterone (a) suppressed the effects of aldo. The trichrome stained photomicrographs are magnified 200� times.

Fig. 4. Adrenalecotmy amplifies aldosterone induced vascular endothelial dysfunction and inflammation. Representative micrographs demonstrate fluorescent immuno-chemistry staining of monocytes (F4/80, a monocyte/macrophage marker) in green in heart tissue counterstained with Evans blue � 200. In ADX mice exposed to aldosterone,macrophages attach to the endothelial cells lining the micro-vessels within the heart. (For interpretation of the references to color in this figure legend, the reader is referredto the web version of this article.)

A.S. Brem et al. / Steroids 78 (2013) 347–355 351

when compared to similarly treated adrenally intact mice (Fig. 3b).This observation was confirmed by assessing the aortic smoothmuscle to collagen ratio, aortic intimal thickness (lm), heartweight to body ratio (mg/gm), and left ventricular collagen volume(%) (Fig. 2c and d). The endothelial dysfunction induced by aldoste-

rone as seen by monocyte endothelial adhesion was also greatercompared to tissue obtained from adrenally intact mice (Figs. 2eand 4). These findings are consistent with the view that the adrenalgland generates a factor that antagonizes the pro-fibrotic and pro-inflammatory effects of aldosterone.

Fig. 5. Mineralocorticoid receptors are expressed in cultured vascular endothelialcells (HUVEC) as seen by Western blot. THP-1 cells served as negative controls.

352 A.S. Brem et al. / Steroids 78 (2013) 347–355

3.3. MR antagonist RU-318, corticosterone and 11-dehydrocorticosterone can each block the early fibrotic effects ofaldosterone on cardiovascular tissue

The pro-inflammatory and early fibrotic changes induced byaldosterone in ADX mice were almost completely blocked by min-eralocorticoid receptor antagonist RU-318 (800 lg/kg/day)(Figs. 1b and 3b). Previously published studies have also shownthat 11-dehydro derivatives of cortisol and corticosterone effec-tively block the electrolyte transport and the fibrotic actions ofaldosterone in renal tissues [11,17,18]. These 11-dehydro deriva-tives are produced by 11b-HSD type 2 and by the bi-directional iso-form 11b-HSD type 1 functioning in the forward dehydrogenasedirection; both enzymes are expressed in vascular endothelial cells[23,30]. 11-dehydrocorticosterone (800 lg/kg/day) also blocked

Fig. 6. Aldosterone (10 nM) induces extracellular matrix production in HUVEC cells in aHUVEC cells were treated with aldosterone (10 nM) for indicated intervals (0, 3, 12, 24,cells were treated with aldosterone (10 nM) for 48 h in the presence or absence ofdehydrocorticosterone (A: 1 lM, 100 nM, 1 nM), or corticosterone (1 lM) and then harvSirius Red staining of HUVEC cells treated with vehicle or aldosterone (10 nM) in the presin this figure legend, the reader is referred to the web version of this article.)

the aldosterone-induced cardiac and vascular changes in the ADXmice; this agent was as effective as RU-318 (Figs. 1b and 3b). Thealdosterone-ADX mice treated with corticosterone (800 lg/kg/day) also showed an equivalent decrease in early fibrotic changes(Figs. 1b and 3b).

3.4. Aldosterone induces extracellular matrix production in culturedvascular endothelial cells

The observations made in the mice suggested that the aldoste-rone induced changes in the heart and arteries first begins in endo-thelial cells lining blood vessels. Cultured human umbilical veinendothelial cells (HUVEC) were next used to examine how aldoste-rone influences vascular endothelial cell function. We first con-firmed that HUVEC express MR by Western blot (Fig. 5). WhenHUVEC were exposed to aldosterone 10 nM, the HUVEC markedlyincreased their synthesis of collagen as assessed by Sirius Redstaining (Fig. 6) in a time dependent fashion peaking at 48 h. Thisaldosterone induced effect was blocked by both RU-318 (1 lM)and 11-dehydrocorticosterone (1 lM) but not by corticosterone(1 lM).

3.5. Aldosterone triggers endothelial dysfunction and monocyteadhesion to endothelial cells in vitro

We next employed a monocyte static adhesion assay to deter-mine whether aldosterone is able to directly induce endothelialcell dysfunction. Few fluorescent THP-1 cells were found adherentto vehicle (Fig. 7) treated HUVEC monolayers. Aldosterone trig-gered monocyte adhesion peaking after 8–12 h (Fig. 7b); RU-318and 11-dehydrocorticosterone blocked the effects of aldosteronewhile corticosterone had no effect. Aldosterone induced monocyte

time dependent fashion. RU-318 and 11-dehydrocorticosterone block this effect. (a)48 h) and then harvested. The cells were sonicated for Sirius Red assay; (b) HUVECthe MR antagonist RU-318, the glucocorticoid receptor antagonist RU-486, 11-ested and sonicated for Sirius Red assay; (c) Representative micrographs depictingence or absence of RU-318 or A (1 lM). (For interpretation of the references to color

Fig. 7. 11-dehydrocorticosterone (a) and RU-318 functionally block aldosterone elicited monocyte adhesion to HUVEC monolayers. (a) Representative micrographs depictmonocyte adhesion to HUVEC monolayers treated with vehicle, aldosterone in the presence or absence of various substances. Studies were conducted with RU-318 (1 lM), A(1 lM), corticosterone (1 lM), or the glucocorticoid antagonist RU-486 (1 lM). (b) Aliquots of cell lysates harvested at indicated time points were subjected to fluorometricanalysis to quantify the relative amount of adherent monocytes. (c) Western immunoblot analysis of cell lysates for E-selectin expression in HUVEC cells 4 h after indicatedtreatments. Both RU-318 and A functionally blocked the increase in aldosterone induced E-selectin expression.

A.S. Brem et al. / Steroids 78 (2013) 347–355 353

adhesion to HUVEC monolayers, was consistent with prior micro-scopic findings. Monocyte to endothelial adhesion is mediated ini-tially by E-selectin, which is inducible in endothelial cells uponstimulation. Aldosterone strongly stimulated de novo expressionof E-selectin at 4 h without altering HUVEC viability. Westernimmunoblots of whole-cell lysates showed that RU-318 and 11-dehydrocorticosterone (in a dose dependent fashion) each substan-tially suppressed aldosterone induced E-selectin expression in thecultured endothelial monolayers. The glucocorticoid receptor ana-tagonist RU-486 had no effect on the aldosterone induced expres-sion of E-selectin.

4. Discussion

Hans Selye first suggested in the 1950’s that aldosterone has apro-fibrotic effect [31]. Nearly 40 years passed before this conceptwas revisited and more fully studied. In 1992, Brilla and Weberlinked mineralocorticoid excess with the development of myocar-dial fibrosis in a previously uninephrectomized rat model [1]. Intheir experiments, the authors only observed the increase in colla-gen deposition in rats infused with aldosterone (approximately30 lg/kg/day) and fed a high salt diet (10 g added to each liter ofwater) for 8 weeks. Animals fed a low salt diet were protected fromboth the rise in blood pressure and the cardiac fibrosis. Severalyears later, Rocha and colleagues described hypertension withthe presence of cardiac hypertrophy and myocardial necrosis withthe influx of inflammatory cells in rats treated with the nitric oxidesynthesis inhibitor L-NAME, adrenalectomized, and given aldoste-rone (40 lg/kg/day by pump) over a period of 14 days [5]. Necroticfindings were also noted in small arteries and arterioles from these

animals. Subsequently, both groups and later others confirmed andextended the observations of mineralocorticoid induced myocar-dial and peri-vascular inflammation and fibrosis [4,20,21,32]. Innearly all of these animal models, the cardiac and vascular effectsof mineralocorticoids could be blocked or at least attenuated byMR antagonists like spironolactone.

Our in vivo histologic observations of aldosterone induced peri-vascular inflammation and fibrosis occurring in heart and the aortaare very similar to those findings discussed above. However, whatis unique is that the aldosterone induced changes occurred withina week in otherwise normal mice and occurred in the absence ofsalt loading, hypertension, or pre-existing injury or toxicity (i.e.prior nephrectomy, L-NAME exposure, pharmacologic doses ofmineralocorticoids, etc.). Adrenalectomy markedly enhanced thevascular effects of aldosterone in our model suggesting that sub-stances, possibly other steroids derived from the adrenal gland,might play a protective role. The concentration of aldosteroneneeded to induce these vascular changes were somewhat higherthan serum aldosterone values seen in animals after prolongeddiuretic exposure (four to five times above baseline vs two timesbaseline for diuretics [33]) but were considerably lower than 50-fold increase over baseline observed in animals treated with the ni-tric oxide synthase inhibitor, L-NAME [34]. Thus it appears asthough a more modest aldosterone exposure alone over time isenough to produce early fibrotic changes under the right circum-stances. MR antagonists, corticosterone, and 11-dehydrocorticos-terone, an end product of 11b-HSD, functionally are able to blockthese fibrogenic effects. Since the endogenous adrenal steroids cor-tisol in man and corticosterone in rodents, are the principal sub-strates for vascular 11b-HSD dehydrogenase, it is possible thatthese glucocorticoids and their locally synthesized 11-dehydro

354 A.S. Brem et al. / Steroids 78 (2013) 347–355

metabolites exert a significant effect in limiting the harmful vascu-lar effects of aldosterone. In addition, the plasma may also provideample concentrations of unbound 11-dehydro glucocorticoidmetabolites to also blunt the actions of aldosterone. In the circula-tion, endogenous glucocorticoids like cortisol and corticosteroneare highly protein bound but the protein binding for 11-dehydro-corticosterone and the 11-dehydro metabolite of cortisol, corti-sone, is limited, making the metabolites potentially more bio-available at the tissue level [25].

Since 11-dehydro glucocortiocoid metabolites bind poorly toMR as well [35], most investigators have viewed them as inert orinactive. While these compounds do not appear to have any bio-logic activity on their own when tested, they have been consis-tently shown to blunt aldosterone induced sodium transport inrenal related tissues, [17,18] where there is no 11b-HSD reductaseactivity. We have previously shown that 11-dehydrocorticosteronealso blocks aldosterone induced fibrotic changes in the kidney [11].We hypothesized that these same compounds might block theearly fibrotic and inflammatory actions of aldosterone in vasculartissue, and it appears to be so. How these 11-dehydro compoundswork is less clear since there is ample bi-directional 11b-HSD-1and some 11b-HSD-2 expressed in endothelial cells [23,30].Odermatt and colleagues showed that 11-dehydrocorticosteroneprevented the aldosterone–MR complex from translocating to thecell nucleus effectively blocking the hormone’s biologic action[19]. Whether this mechanism operates in vascular tissue remainsunknown at this time and should be the subject of additionalstudy.

Using in vivo experiments, it is difficult if not impossible to dif-ferentiate the anti-mineralocorticoid effects of 11-dehydrocorti-costerone from those of corticosterone in heart and bloodvessels. Our in vivo animal studies clearly show both agents areeffective in blocking the vascular effects of aldosterone. However,the picture becomes clearer in the cultured vascular endothelialexperiments. 11-dehydrocorticosterne attenuates aldosterone’s ef-fects in a dose dependent fashion while corticosterone does not.This is consistent with the view that corticosterone may not beable to directly inhibit the actions of aldosterone and that local11b-HSD-2 and 11b-HSD-1 dehydrogenase cannot sufficientlymetabolize enough corticosterone to 11-dehydrocorticosteroneand protect the vascular endothelial cell against the aldosteroneinduced changes. At least in vascular tissue, factors separate fromor in addition to the local metabolism of endogenous glucocorti-coids influence activation of the mineralocorticoid receptor.

Generation of reactive oxygen species by activation of NAPDHoxidase is one of the inflammatory and pro-fibrotic processes in-duced by aldosterone [36]. This specific pathway is of particularinterest since the enzyme system generates NADP as a bi-productco-factor while synthesizing reactive oxygen radials. NADP servesas the principal co-factor driving the forward dehydrogenase reac-tion of 11b-HSD-1. Under normal circumstances, 11b-HSD-1 is bi-directional and is amply expressed in both vascular endothelialcells and vascular smooth muscle cells [23,30]. The presence of ex-cess NADP would likely shift the reaction equilibrium forwardresulting in the formation of more 11-dehydro metabolites. Thepresence of 11b-HSD-2 in vascular endothelial cells also allowsfor the generation of 11-dehydro metabolites. These metabolitesthen might potentially act as a negative feedback loop attenuatingthe effect of the aldosterone and turning off the additional forma-tion of the oxygen free radials.

While the experimental model is the same as that in our previ-ously published paper describing changes in the kidney [11], thereare some interesting differences in how the renal and vascular sys-tems appear to respond to the aldosterone infusion. In the kidney,there is no obvious cellular inflammation after 1 week but clearevidence of collagen formation and fibrosis [11]. Moreover, in

separate experiments, aldosterone directly induces excess collagensynthesis within 48 h in cultured collecting duct cells in theabsence of inflammatory cells and in the absence of exogenouscytokines. In contrast, the arteries and arterioles taken from aldo-sterone infused mice in this study showed signs of cellular inflam-mation with attachment of circulating macrophages to vascularendothelial cells. When cultured, vascular endothelial cells re-spond to aldosterone in a more nuanced manner by both rapidlyenhancing monocyte adhesion through the release of E-selectinas well as directly inducing collagen synthesis within 24–48 h sim-ilar to the renal collecting duct cells. Thus the kidney’s initial re-sponse to aldosterone seems more pro-fibrotic and the initialvascular response appears to be both inflammatory and pro-fibrotic suggesting tissue specific differences in the pathologicpathways. In both tissues, MR antagonists, 11-dehydro glucocorti-coid metabolites, and endogenous glucocorticoids blunt the actionof aldosterone however possibly suggesting a common pathway.The ability to assess the individual inhibitory effects of 11-dehydrocorticosterone and corticosterone on aldosterone action is easier inrenal tissues since the kidney cannot enzymatically reduce11-dehydro corticosterone back to its parent.

Our observations clearly indicate that adrenal ablation canexacerbate aldosterone induced inflammation and fibrosis in vas-cular tissues. The pathologic changes occur relatively rapidly butrequire persistent exposure to aldosterone and the loss of adrenalsubstances including 11-dehydro-glucocorticoid metabolites formaximum effect. Lastly, vascular endothelial cells appear to bethe initial target for aldosterone. These findings provide supportfor the use MR receptor antagonists and possibly agents, which di-rectly promote the generation of 11-dehydro metabolites, in pa-tients with relevant risk factors.

Conflict of interest

The authors have no real or apparent conflicts of interest toreport.

Acknowledgements

The authors wish to thank Sarah J. Longley and Evelyn Tolbertfor their assistance on this project. This work was supported bythe following Grants: NIH R21 HL093561 (S.S.), P20RR018728(S.S.), RO1DK092485 (R.G.) and University Medical Foundationdevelopmental Grants (R.G.).

References

[1] Brilla CG, Weber KT. Mineralocorticoid excess, dietary sodium, and myocardialfibrosis. J Lab Clin Med 1992;120:893–901.

[2] Young M, Fullerton M, Dilley R, Funder J. Mineralocorticoids, hypertension, andcardiac fibrosis. J Clin Invest 1994;93:2578–83.

[3] Greene EL, Kren S, Hostetter TH. Role of aldosterone in the remnant kidneymodel in the rat. J Clin Invest 1996;98:1063–8.

[4] Sun Y, Ramires FJ, Weber KT. Fibrosis of atria and great vessels in response toangiotensin II or aldosterone infusion. Cardiovasc Res 1997;35:138–47.

[5] Rocha R, Stier Jr CT, Kifor I, et al. Aldosterone: a mediator of myocardialnecrosis and renal arteriopathy. Endocrinology 2000;141:3871–8.

[6] Brown NJ, Nakamura S, Ma L, et al. Aldosterone modulates plasminogenactivator inhibitor-1 and glomerulosclerosis in vivo. Kidney Int2000;58:1219–27.

[7] Leroy V, De Seigneux S, Agassiz V, et al. Aldosterone activates NF-kappaB in thecollecting duct. J Am Soc Nephrol 2009;20:131–44. Epub 2008 Nov 5.

[8] Vallon V, Wyatt AW, Klingel K, et al. SGK1-dependent cardiac CTGF formationand fibrosis following DOCA treatment. J Mol Med 2006;84:396–404. Epub2006 Apr 8.

[9] Muller OG, Parnova RG, Centeno G, et al. Mineralocorticoid effects in thekidney: correlation between alphaENaC, GILZ, and Sgk-1 mRNA expression andurinary excretion of Na+ and K+. J Am Soc Nephrol 2003;14:1107–15.

[10] Naray-Fejes-Toth A, Helms MN, Stokes JB, Fejes-Toth G. Regulation of sodiumtransport in mammalian collecting duct cells by aldosterone-induced kinase,SGK1: structure/function studies. Mol Cell Endocrinol 2004;217:197–202.

A.S. Brem et al. / Steroids 78 (2013) 347–355 355

[11] Brem AS, Morris DJ, Ge Y, et al. Direct fibrogenic effects of aldosterone onnormotensive kidney: an effect modified by 11{beta}-HSD activity. Am JPhysiol Renal Physiol 2010;298:F1178–87.

[12] Barger AC, Berlin RD, Tulenko JF. Infusion of aldosterone, 9-alpha-fluorohydrocortisone, and antidiuretic hormone into the renal artery ofnormal and adrenalectomized unanesthetized dogs: effect on electrolyte andwater excretion. Endocrinology 1958;62:804–15.

[13] Ross EJ. Modification of the effects of aldosterone on electrolyte excretionin man by simultaneous administration of corticosterone and hydrocor-tisone. Relevance to Conn’s syndrome. J Clin Endocrinol Metab 1960;20:229–37.

[14] Morris DJ, Berek JS, Davis RP. The physiological response to aldosterone inadrenalectomized and intact rats and its sex dependence. Endocrinology1973;92:989–93.

[15] Porter GA, Bogoroch R, Edelman IS. On the mechanism of action of aldosteroneon sodium transport: the role of RNA synthesis. Proc Natl Acad Sci USA1964;52:1326–33.

[16] Farman N, Kusch M, Edelman IS. Aldosterone receptor occupancy and sodiumtransport in the urinary bladder of Bufo marinus. Am J Physiol1978;235:C90–6.

[17] Alberti KGMM, Sharp GWG. Identification of four types of steroid by theirinteraction with mineralocorticoid receptors in the toad bladder. J Endocrinol1970;48:563–74.

[18] Brem AS, Matheson KL, Barnes JL, Morris DJ. 11-Dehydrocorticosterone, aglucocorticoid metabolite, inhibits aldosterone action in toad bladder. Am JPhysiol 1991;261:F873–9.

[19] Odermatt A, Arnold P, Frey FJ. The intracellular localization of themineralocorticoid receptor is regulated by 11beta-hydroxysteroiddehydrogenase type 2. J Biol Chem 2001;276:28484–92. Epub 2001 May 11.

[20] Young MJ, Morgan J, Brolin K, et al. Activation of mineralocorticoid receptorsby exogenous glucocorticoids and the development of cardiovascularinflammatory responses in adrenalectomized rats. Endocrinology2010;151:2622–8. Epub 2010 Apr 21.

[21] Young M, Head G, Funder J. Determinants of cardiac fibrosis in experimentalhypermineralocorticoid states. Am J Physiol 1995;269:E657–62.

[22] Rocha R, Martin-Berger CL, Yang P, et al. Selective aldosterone blockadeprevents angiotensin II/salt-induced vascular inflammation in the rat heart.Endocrinology 2002;143:4828–36.

[23] Brem AS, Bina RB, King TC, Morris DJ. Localization of 2 11beta-OH steroiddehydrogenase isoforms in aortic endothelial cells. Hypertension1998;31:459–62.

[24] Sippell WG, Dorr HG, Bidlingmaier F, Knorr D. Plasma levels of aldosterone,corticosterone, 11-deoxycorticosterone, progesterone, 17-hydroxyprogesterone, cortisol, and cortisone during infancy and childhood.Pediatr Res 1980;14:39–46.

[25] Mickelson KE, Forsthoefel J, Westphal U. Steroid–protein interactions. Humancorticosteroid binding globulin: some physicochemical properties and bindingspecificity. Biochemistry 1981;20:6211–8.

[26] Daniel JY, Leboulenger F, Vaudry H, et al. Interrelations between bindingaffinity and metabolic clearance rate for the main corticosteroids in the rabbit.J Steroid Biochem 1982;16:379–84.

[27] Ingelfinger JR. Forestalling fibrosis. N Engl J Med 2003;349:2265–6.[28] Gong R, Rifai A, Tolbert EM, et al. Hepatocyte growth factor modulates matrix

metalloproteinases and plasminogen activator/plasmin proteolytic pathwaysin progressive renal interstitial fibrosis. J Am Soc Nephrol 2003;14:3047–60.

[29] Gomez-Sanchez CE, de Rodriguez AF, Romero DG, et al. Development of apanel of monoclonal antibodies against the mineralocorticoid receptor.Endocrinology 2006;147:1343–8. Epub 2005 Nov 17.

[30] Gong R, Morris DJ, Brem AS. Variable expression of 11beta Hydroxysteroiddehydrogenase (11beta-HSD) isoforms in vascular endothelial cells. Steroids2008;73:1187–96. Epub 2008 Jun 3.

[31] Selye H. Anticortisol action of aldosterone. Science 1955;121:368–9.[32] Rocha R, Rudolph AE, Frierdich GE, et al. Aldosterone induces a vascular

inflammatory phenotype in the rat heart. Am J Physiol Heart Circ Physiol2002;283:H1802–10.

[33] Reungjui S, Hu H, Mu W, et al. Thiazide-induced subtle renal injury notobserved in states of equivalent hypokalemia. Kidney Int 2007;72:1483–92.Epub 2007 Oct 10.

[34] Ikeda H, Tsuruya K, Toyonaga J, et al. Spironolactone suppresses inflammationand prevents L-NAME-induced renal injury in rats. Kidney Int2009;75:147–55. Epub 2008 Oct 15.

[35] Funder JW, Pearce PT, Smith R, Smith AI. Mineralocorticord action: targettissue specificity is enzyme, not receptor, mediated. Science 1988;242:583–5.

[36] Miyata K, Rahman M, Shokoji T, et al. Aldosterone stimulates reactive oxygenspecies production through activation of NADPH oxidase in rat mesangial cells.J Am Soc Nephrol 2005;16:2906–12. Epub 2005 Aug 31.