| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Division of Endocrinology (A.T., C.Z.G., G.K.A., V.R., G.H.W.), Brigham and Womens Hospital, and Childrens Hospital Informatics Program (I.S.K.), Boston, Massachusetts 02115
Address all correspondence and requests for reprints to: Alexander Turchin, Division of Endocrinology, Brigham and Womens Hospital, 221 Longwood Avenue, Boston, Massachusetts 02115. E-mail: aturchin{at}partners.org.
| Abstract |
|---|
|
|
|---|
| Introduction |
|---|
|
|
|---|
Animal studies have demonstrated that mineralocorticoid excess in conjunction with a high sodium intake results in myocardial inflammation and fibrosis (7, 8, 9). Damage involves both left and right ventricle as well as both atria and the adventitia of the pulmonary artery, underlining blood pressure-independent nature of the phenomenon (10). These results were confirmed in a transgenic model of intracardiac mineralocorticoid excess overexpressing 11ß-hydroxysteroid dehydrogenase type 2 in cardiomyocytes leading to cardiac fibrosis and heart failure (11).
Adverse effects of aldosterone are abrogated by administration of a mineralocorticoid receptor antagonist in animal models (8, 12, 13, 14), suggesting that induction of cardiac inflammation and fibrosis by aldosterone is mediated by the mineralocorticoid receptor activation. In clinical studies, mineralocorticoid receptor antagonists have been demonstrated to decrease morbidity and mortality in normotensive patients with heart failure (15, 16).
The mechanism by which activation of mineralocorticoid receptor leads to cardiac inflammation and fibrosis is unclear. Mineralocorticoid receptor functions as a ligand-dependent transcription factor. However, immediate consequences of mineralocorticoid receptor activation on gene expression in the heart are poorly understood (17). Most reports focus on either a narrow group of genes or a single time point after aldosterone administration (18, 19). However, a single time point is unlikely to capture the dynamic nature of gene expression (20, 21, 22). In this study we have taken a different approach by determining the temporal profile of the effects of acute in vivo aldosterone administration on gene expression profile in the heart. Importantly, we also included at each time point a response to a placebo injection, therefore separating the effects of the experimental method from those of the aldosterone.
| Materials and Methods |
|---|
|
|
|---|
Male C57BL/6 mice (Charles River, Wilmington, MA) were uninephrectomized at 11 wk of age and allowed 7 d recovery time. Starting at 12 wk of age, all animals were fed 3% NaCl diet (Purina Test Diet Salt Series based on Basal Diet 5755, Pharmaserv Inc., Framingham, MA) for 1 wk. Animals were allowed unlimited access to water.
At the age of 13 wk, mice were administered ip injection of 125 µl mouse serum as control or 250 ng (approximately 10 µg/kg) aldosterone (Acros Organics, Geel, Belgium) diluted in 250 nl ethanol and subsequently in 125 µl mouse serum. Five animals in each of the control and experimental groups were killed at the following intervals after the injection: 0 h (no injection), 0.5, 1, 2, 3, 4, 5, and 12 h.
All mice were killed by deep inhalatory anesthesia with isoflurane. Blood and the heart were collected from every animal.
Biochemical assays
Plasma aldosterone was measured using a RIA kit (Diagnostic Products Corp., Los Angeles, CA). Plasma corticosterone levels were measured using a competitive enzyme immunoassay kit (American Laboratory Products Co., Windham, NH).
Microarray analysis
Mouse hearts were preserved in liquid nitrogen immediately after death of the animal. The total RNA was isolated using Trizol reagent (Invitrogen, Carlsbad, CA) according to the manufacturers protocol. RNA concentration and purity were assessed using a Beckman Coulter spectrophotometer DV 600. All RNA samples used in these experiments had 260:280 ratios between 1.94 and 2.25. Aliquots of the extracted total RNA were assessed for integrity by agarose gel electrophoresis and ethidium bromide staining for visualization of the 28S and 18S rRNA bands. RNA samples from all five animals receiving the same treatment at each time point were pooled.
Microarray analysis was performed using the mouse 430 2.0 GeneChip arrays (Affymetrix, Santa Clara, CA), which contain probes for more than 34,000 known mouse genes. Preparation of the labeled cDNA and microarray hybridization were performed by the Microarray Core Facility at the Childrens Hospital (Boston, MA). The raw data were normalized and analyzed using GeneChip Affymetrix GCOS 1.2 software. The signal values and the present, absent, or marginal calls were computed for all probe sets. Only probe sets with the average signal intensity greater than 1280 that were either unique for the gene or directed against the 3' region of the gene were used in the analysis as described in detail in Results.
The ratios of signal intensities of aldosterone injection to control groups from each of the eight time points were used as an input for Cluster Analysis of Gene Expression Dynamics (CAGED) 1.0 (http://genomethods.org/caged), a program specifically designed for cluster analysis of temporal microarray experiments (23).
The data discussed in this publication have been deposited in the National Center for Biotechnology Informations Gene Expression Omnibus (http://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE3440) (24, 25).
Quantitative real-time RT-PCR
PCRs were performed using total RNA generated for microarray analysis. cDNA was synthesized from 5 µg RNA with the first-strand cDNA synthesis kit (Amersham Biosciences, Buckinghamshire, UK). PCRs were performed by ABI Prism 7000 sequence detection system (Applied Biosystems, Foster City, CA) with the QuantiTect SYBR Green PCR kit (Qiagen Sciences, Valencia, CA) according to the manufacturers protocol. Primers were designed with Primer Express version 2.0 (Applied Biosystems) and were synthesized by Integrated DNA Technologies (Coralville, IA). The sequences of the primers used to detect specific genes were as follows: Plunc, 5'-TGGTCTTGTGCAGAGTCCTG-3' (forward, nt 484), 5'-CCTCCCCTGATTGTCTTTCA-3' (reverse, nt 643); thioredoxin reductase 2, 5'-GCTTCTGGCAAGGAAGACAC-3' (forward, nt 938), 5'-CCCTCAGCAACATCTCCAAT-3' (reverse, nt 1128). The relative level of mRNA expression of a specific gene was calculated based on 
CT method according to the manufacturers instruction and normalized to mRNA level for the housekeeping gene ß-actin.
| Results |
|---|
|
|
|---|
|
It is known that expressed genes are more likely to have higher signal intensity (26). It has been shown that in the majority of human tissues fewer than 25% of all known genes are expressed in each tissue (27). We therefore limited data analysis to the probes with mean signal intensities of all 15 data points in the top 25% by magnitude (above 1280).
mRNA is usually degraded in the cell starting from the 5' end. This is known to lead to more reliable detection of gene expression by probes against 3' regions of the corresponding genes (28). We examined coefficient of variation of probes against 5', middle, and 3' regions of four constitutively expressed mouse genes routinely used as controls in Affymetrix microarrays (Table 1
). Most of the time, coefficient of variation for probe sets against 3' regions was substantially lower than coefficient of variation for either 5' or middle region of the corresponding gene [one exception was the gene for transferrin receptor whose mRNA undergoes a unique degradation pathway starting at the 3' end (29)]. We subsequently verified that this applies to the genes whose mean signal intensities were greater than 1280 with more than one probe set. For these genes the probe sets against 5' regions of the genes had a higher coefficient of variation than probe sets against 3' regions (P < 0.0001; Students t test). Therefore, only data on 5111 probe sets that were either unique for the gene or the closest to the 3' end of the mRNA of all probe sets specific for the gene on the microarray chip were used in further analysis.
|
|
, and protein phosphatase 5, featured prominently in this group; other genes encoded proteins responsible for steroid biosynthesis, inactivation of reactive oxygen species, and structural proteins as well as proteins of unknown function.
|
|
| Discussion |
|---|
|
|
|---|
Elucidation of early response genes for aldosterone has traditionally been difficult because the mineralocorticoid receptor binds the physiological glucocorticoids cortisol and corticosterone with affinity comparable with that of aldosterone (30, 31). There is also evidence that at least in some tissues activation of glucocorticoid receptor may induce effects similar to agonist activation of mineralocorticoid receptor (32, 33). Consequently this experiment was designed to elucidate the effects specific to aldosterone. Whereas both the aldosterone injection and control animals experienced a similar rise in plasma corticosterone levels (presumably due to the stress of the injection), the ratios of expression levels of aldosterone injection to control group for every gene at every time point were used in the data analysis, thereby canceling out the effect of corticosterone on gene expression. This experimental design also controlled for other potential confounders related to the injection which could have influenced gene expression levels.
Animal models of mineralocorticoid excess demonstrate cardiac damage (1, 7, 8), and human clinical trials offer evidence that blockade of mineralocorticoid receptor can decrease morbidity and mortality in patient with heart failure (15, 16). This experiment was designed to focus on the genetic effects of aldosterone that could potentially mediate empirically observed myocardial injury. In rodent models, uninephrectomy and high-salt diet are necessary for demonstration of mineralocorticoid-mediated cardiac damage (7, 12, 34). Whereas the duration of this experiment was too short to observe myocardial injury, experimental conditions (uninephrectomy, high salt diet, and high physiologic levels of aldosterone) were designed to replicate the physiological milieu that over longer periods of time has been shown to lead to cardiac damage.
In this experiment we used a novel software program CAGED to identify genes differentially expressed between aldosterone injection and control groups. CAGED was specifically designed to analyze time series microarray experiments. It employs Bayesian clustering to provide a statistically robust view of the natural dynamic structure of the transcriptome. We also used several noise-reduction strategies, including limiting analysis to microarray probes with high signal intensities (35) and probe sets closest to the 3' end of the transcript (28). Most of the constitutively expressed mouse genes used by Affimetrix as controls had the lowest coefficient of variation in the probes located closest to the 3' end of their mRNAs, supporting this strategy. One exception was transferrin receptor whose mRNA undergoes a unique degradation process starting at the 3' end (29), likely leading to the relative increase in the coefficient of variation of 3' probes observed in this experiment. The noise-reduction approach in conjunction with the use of Bayesian analysis by CAGED permitted us to identify a small number of genes with a unique pattern of expression after an acute aldosterone administration; temporal expression profiles of representative genes were subsequently verified by quantitative real time RT-PCR.
The genes found to be differentially expressed between the aldosterone injection and control groups fell into several functional categories: phosphatases; steroid biosynthesis; inactivation of reactive oxygen species (thioredoxin reductase 2); structural proteins (ankyrin repeat and suppressor of cytokine signaling box-containing protein 15); and others, including genes whose function in the heart is currently unknown (Plunc). All of the genes exhibited a similar expression pattern of suppression (relative to the control group) at 13 h after aldosterone administration followed by a brief rebound increase and a subsequent return to baseline.
Several of the phosphatases whose expression level was affected by aldosterone are known to play a role in steroid signaling: dual specificity phosphatase 7 dephosphorylates ERK (36), and protein phosphatase 5 may be involved in glucocorticoid receptor signaling (37). Several other groups have reported an increase in phosphorylation of ERK1/2 within several minutes after aldosterone administration that is not blocked by inhibitors of transcription and translation (38, 39), thus implying a nongenomic mechanism. Our findings possibly indicate a second, genomic pathway leading to the ERK1/2 phosphorylation (by means of decreasing the expression of the enzyme that dephosphorylates ERK1/2) after a longer lead time, potentially leading to a more sustained effect.
Another important group of genes affected by aldosterone were regulators of tissue damage and inflammation: thioredoxin reductase 2 provides cellular defense against reactive oxygen species (40), and uteroglobin (also known as blastokinin, secretoglobin family 1A, member1, or Clara cell secretory protein) inhibits phospholipase A2 and therefore potentially has an antiinflammatory effect (41). Excess of aldosterone was known to cause inflammatory lesions in the heart as long as more than half a century ago (1). More recently ours and other laboratories have shown that aldosterone and other mineralocorticoids induce inflammatory markers and subsequently lead to development of histologic picture of inflammation in the coronary vessels (12, 34, 42). These effects can be duplicated by administration of 11ß-hydroxysteroid dehydrogenase type 2 inhibitors (42). Our findings of a decrease in the expression level of thioredoxin reductase and uteroglobin after aldosterone administration suggest a possible mechanism for these responses.
Multiple investigators have reported cardiac fibrosis as one of the prominent effects of concomitant exogenous aldosterone administration and a high-salt diet (9, 13, 43). Interestingly, however, no cardiac fibrosis was found in transgenic mice expressing increased levels of aldosterone synthase in the heart leading to elevated intracardiac aldosterone concentration (44). This suggests that development of cardiac fibrosis may be a secondary rather than primary effect of aldosterone. In our experiment we did not find any evidence of increased expression of enzymes involved in collagen synthesis, consistent with this hypothesis.
The experimental design of the study had its limitations. Because corticosterone levels increased in both control and experimental animals, we would not be able to detect a change in the expression levels of genes influenced by both glucocorticoid and mineralocorticoid receptors, most notably serum and glucocorticoid-inducible kinase (45). This may account for the absence of some genes whose expression level was increased by aldosterone. However, serum and glucocorticoid-inducible kinase expression levels in the heart have not been shown to be affected by aldosterone (46). Consequently, we felt that it was important to first identify genetic mechanisms unique to aldosterone, which then could serve as a foundation for further investigation into the complexity of genetic response to mineralocorticoids.
RNA samples in this experiment were pooled (five animals per microarray) to increase the statistical power. The power could be increased further if each individual animals RNA sample was analyzed on its own microarray chip and potential outlier values identified. However, statistical validity of using pooled RNA samples for microarray analysis is well established (47). Importantly, the data at all time points in the experiment were congruent, making it highly unlikely that a single outlier value could have significantly influenced the findings.
In this exploratory study, we deliberately chose an in vivo model of mineralocorticoid excess that closely replicated experimental conditions that were previously reported to lead to cardiac injury. The study was not designed to identify the effect of aldosterone on specific tissues and cells in the heart. This should be the subject of further in vitro studies that will build on the in vivo findings reported here. It will also be important to establish the role of mineralocorticoid receptor in mediating these effects of aldosterone. Investigation of the effects of long-term, rather than short-term, aldosterone exposure on the gene expression in the heart may shed more light on the mechanisms of observed aldosterone-mediated cardiac damage.
In summary, the current study presents the results of an investigation of the timeline of changes in gene expression profile in the heart after acute in vivo aldosterone administration. We used a novel software program, CAGED, to identify a small number of genes whose expression patterns were altered in the aldosterone-injected animals. The known functions of a number of these genes, which have not been previously reported to be affected by aldosterone, are consistent with known effects of aldosterone at both molecular and histologic level.
| Acknowledgments |
|---|
| Footnotes |
|---|
Author Disclosure Summary: A.T., C.Z.G., G.K.A., V.R., I.S.K., and G.H.W. have nothing to declare.
First Published Online April 6, 2006
Abbreviation: CAGED, Cluster Analysis of Gene Expression Dynamics.
Received January 3, 2006.
Accepted for publication March 29, 2006.
| References |
|---|
|
|
|---|
This article has been cited by other articles:
![]() |
V. Ricchiuti, C. G Lian, E. M Oestreicher, L. Tran, J. R Stone, T. Yao, E. W Seely, G. H Williams, and G. K Adler Estradiol increases angiotensin II type 1 receptor in hearts of ovariectomized rats J. Endocrinol., January 1, 2009; 200(1): 75 - 84. [Abstract] [Full Text] [PDF] |
||||
![]() |
L. H. Pojoga, T. M. Yao, S. Sinha, R. L. Ross, J. C. Lin, J. D. Raffetto, G. K. Adler, G. H. Williams, and R. A. Khalil Effect of dietary sodium on vasoconstriction and eNOS-mediated vascular relaxation in caveolin-1-deficient mice Am J Physiol Heart Circ Physiol, March 1, 2008; 294(3): H1258 - H1265. [Abstract] [Full Text] [PDF] |
||||
![]() |
G. Fejes-Toth and A. Naray-Fejes-Toth Early Aldosterone-Regulated Genes in Cardiomyocytes: Clues to Cardiac Remodeling? Endocrinology, April 1, 2007; 148(4): 1502 - 1510. [Abstract] [Full Text] [PDF] |
||||
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |