JACC: Heart Failure Vol. 1, No. 6, 2013� 2013 by the American College of Cardiology Foundation ISSN 2213-1779/$36.00Published by Elsevier Inc. http://dx.doi.org/10.1016/j.jchf.2013.09.003

Gene Expression Profiles of Peripheral BloodMononuclear Cells Reveal TranscriptionalSignatures as Novel Biomarkers of CardiacRemodeling in Rats With Aldosteronism andHypertensive Heart Disease

Ivan C. Gerling, PHD,* Robert A. Ahokas, PHD,y German Kamalov, MD, PHD,zWenyuan Zhao, MD, PHD,z Syamal K. Bhattacharya, PHD,z Yao Sun, MD, PHD,zKarl T. Weber, MDzMemphis, Tennessee

From the *

Center, Mem

of Tennesse

Cardiovascu

Tennessee. T

HL73043,

contents are

official views

they have no

Manuscri

accepted Sep

Objectives In

Division of Endocrinolo

phis, Tennessee; yDepar

e Health Science Center

lar Diseases, University o

his work was supported,

R01-HL90867 (Dr. We

solely the responsibility of

of the National Institutes

relationships relevant to

pt received July 22, 2013;

tember 5, 2013.

searching for a noninvasive surrogate tissue mimicking the pro-oxidant/proinflammatory hypertensive heartdisease (HHD) phenotype, we turned to peripheral blood mononuclear cells (PBMCs). We tested whether iterationsin [Ca2þ]i, [Zn

2þ]i, and oxidative stress in cardiomyocytes and PBMCs would complement each other, eliciting similarshifts in gene expression profiles in these tissues demonstrable during the preclinical (week 1) and pathological(week 4) stages of aldosterone/salt treatment (ALDOST).

Background In

appropriate neurohormonal activation contributes to pathological remodeling of myocardium in HHD associatedwith aldosteronism. In rats receiving long-term ALDOST, evidence of reparative fibrosis replacing necroticcardiomyocytes and coronary vasculopathy appears at week 4 associated with the induction of oxidative stress bymitochondria that overwhelms endogenous, largely Zn2þ-based, antioxidant defenses. Biomarker-guided predictionof risk before the appearance of cardiac pathology would prove invaluable.

Methods In

PBMCs and cardiomyocytes, quantitation of cytoplasmic free Ca2þ and Zn2þ, H2O2, and 8-iosprostane levels andisolation of ribonucleic acid (RNA) and gene expression together with statistical and clustering analyses andconfirmation of genes by in situ hybridization and reverse-transcription polymerase chain reaction were performed.

Results C

ompared with controls, at weeks 1 and 4 of ALDOST, we found comparable increments in [Ca2þ]i, [Zn2þ]i, and

8-isoprotane coupled with increased H2O2 production in cardiac mitochondria and PBMCs, together with thecommon networks of expression profiles dominated by genes involved in oxidative stress, inflammation, and repair.These included 3 central Ingenuity pathway-linked genes: p38 mitogen-activated protein kinase, a stress-responsiveprotein; nuclear factor-kB, a redox-sensitive transcription factor and a proinflammatory cascade that it regulates;and transforming growth factor-b1, a fibrogenic cytokine involved in tissue repair.

Conclusions S

ignificant overlapping demonstrated in the molecular mimicry of PBMCs and cardiomyocytes during preclinical andpathological stages of ALDOST implies that transcriptional signatures of PBMCs may serve as early noninvasive andnovel sentinels predictive of impending pathological remodeling in HHD. (J Am Coll Cardiol HF 2013;1:469–76)ª 2013 by the American College of Cardiology Foundation

gy, University of Tennessee Health Science

tment of Obstetrics and Gynecology, University

, Memphis, Tennessee; and the zDivision of

f Tennessee Health Science Center, Memphis,

in part, by National Institutes of Health R01-

ber), and 2R01-DK62103 (Dr. Gerling). Its

the authors and do not necessarily represent the

of Health. All other authors have reported that

the contents of this paper to disclose.

revised manuscript received August 28, 2013,

The pathophysiological origins of the heart failure associatedwith hypertensive heart disease (HHD) are rooted inneurohormonal activation that includes circulating effectorhormones of the renin-angiotensin-aldosterone system (1,2).

See page 477

It is the endocrine actions of the renin-angiotensin-aldosterone system circulating hormones that account for

Abbreviationsand Acronyms

ALDOST = aldosterone/salt

treatment

ANOVA = analysis of

variance

HHD = hypertensive heart

disease

NFkB = nuclear factor-kB

NO = nitric oxide

PBMC = peripheral blood

mononuclear cell

PTH = parathyroid hormone

RT-PCR = reverse-

transcription polymerase

chain reaction

Gerling et al. JACC: Heart Failure Vol. 1, No. 6, 2013PBMCs as Novel Biomarkers December 2013:469–76

470

a systemic response the majorfeatures of which include a pro-oxidant phenotype involvingdiverse tissues, in which the rateof reactive oxygen species gener-ation overwhelms endogenousantioxidant defenses consistingof crucial Zn2þ-dependent en-zymes (3,4); an immunostimu-latory state, in which activatedcirculating lymphocytes and monocytes (peripheral blood mono-nuclear cells [PBMCs]) elaborateproinflammatory cytokines (5–9),and a catabolic state with pro-teolytic degradation of skeletalmuscle and elevations in circu-

lating parathyroid hormone (PTH) that account for boneresorption and predispose to fracture (10–15). The failingheart likewise undergoes a progressive hormone-mediatedstructural and biochemical remodeling with ongoing loss ofcardiomyocytes.

Biomarkers that, during a preclinical stage, would revealrisk before as well as the presence of cardiac pathology wouldadvance the overall evaluation and management of patientshaving heart failure with hypertension (16,17). We postu-lated that such biomarkers are likely to originate in patho-physiological events leading to cardiac remodeling. Insearching for a surrogate tissue, tracking the pro-oxidant/proinflammatory pathway to cardiomyocyte necrosis, weturned to PBMCs (lymphocytes and monocytes) as integralto cellular/molecular mechanisms of tissue repair and cardiacpathology (18). Our rationale in targeting PBMCs as aputative surrogate was based on the following: 1) the geneexpression profile of multiple tissues is shared with PBMCs(19); 2) PBMCs are integral to the neuroendocrine-immunesystem interface invoked during stressor states with themicroarray profiling of their transcriptome used to identifymolecular signaling pathways; 3) PBMC activation revealsmolecular signatures in chronic heart failure and predicts itsseverity (5–9,20,21); 4) our previous studies identified theactivation of the PBMC transcriptome at preclinical (week 1)and pathological (week 4) stages of aldosterone/salt treat-ment (ALDOST) in rats (22,23); and 5) the availability ofPBMCs for clinical interrogation as a noninvasive targetthat could be monitored serially during the course ofneurohormonal activation to predict risk and pathologybefore and at its appearance and response to intervention.

Activated PBMCs during ALDOST express genesassociated with oxidative stress, antioxidant defenses, andproinflammatory cytokines implicating, an immunosti-mulatory state (22,23). Cardioprotection, in turn, wasassessed by monitoring temporal iterations in cationcomposition, redox status, and an altered PBMC tran-scriptome (22–25). Hence, these cells lend themselves aspotential biomarkers of noninvasive and serial sampling

during preclinical and pathological stages of heart failurewith hypertension (HHD) and to monitoring car-dioprotective strategies (3–9,25). We hypothesized that thepathophysiological changes seen with ALDOST arepotentiated by integrated interactions between hormonesand divalent cations (i.e., Ca2þ and Zn2þ as pro-oxidantand antioxidant), and that these iterations will causeunique and yet similar shifts in gene expression in hearttissue and PBMCs, thus providing characteristic molecularand transcriptional signatures.

Methods

Animal model. Eight-week-old male Sprague-Dawley rats(Harlan Laboratories, Indianapolis, Indiana) were used aspreviously reported (22,23). They received aldosterone(0.75 mg/h) via an implanted minipump (Alzet, Cupertino,California) together with 1% NaCl/0.4% KCl in drinkingwater (ALDOST), and standard laboratory chow (Teklad2215 Rodent Diet, Harlan Laboratories, Madison, Wis-consin) for 1 and 4 weeks (6 each group). Six unoperated on,untreated age-/sex-matched rats served as controls. Thisstudy was approved by the institution’s Animal Care andUse Committee. The investigation conforms to theGuidelines for the Care and Use of Laboratory Animalspublished by the National Institutes of Health.PBMCs and cardiomyocyte Ca2D, Zn2D, H2O2, and 8-isoprostane. Quantitation of divalent cations H2O2 and8-iosprostane was performed, as previously described (23,25).Isolation of ribonucleic acid (RNA) and gene expressionanalysis. We isolated RNA using Trizol Reagent (LifeTechnologies, Carlsbad, California), as previously described(26). The RNA underwent gene expression analysis usingAffymetrix software (Affymetrix Inc., Santa Clara, Cal-ifornia), as previously described (22). This analysis wasconducted by the Molecular Resource Center Core Lab atour institution using the GCOS 1.4 software from Affy-metrix Inc. We previously reported on the efficacy of ourcore facility to replicate the relevant expression array data byquantitative reverse-transcription polymerase chain reaction(RT-PCR).Statistical and clustering analysis of gene expressiondata. Expression data files were uploaded into the Gene-Spring DX software (Agilent Technologies, Santa Clara,California) for statistical and cluster analysis, as previouslydescribed (27). Light filtering was initially used to removethose probe sets that were not present in at least 1 of thesamples, followed by a 1-way analysis of variance (ANOVA)(p < 0.05) with Benjamini-Hochberg correction for mul-tiple hypothesis testing. Hierarchical clustering of the genesfrom our ANOVA was carried out to further sort those intosubgroups with specific expression patterns, depending ontheir up- or down-regulation in response to ALDOST.Data mining of gene probe sets identified by the dataanalysis. The main data mining tool used was IngenuityPathways Analysis (Ingenuity Systems, Inc., Redwood City,

Table 1Cytosolic Free [Ca2þ]i and [Zn2þ]i in PBMCs andCardiomyocytes

Groups PBMCs Cardiomyocytes

[Ca2þ]i, nM

Control 41.8 � 3.7 29 � 4

ALDOST week 1 70.7 � 6.2* 70 � 5*

ALDOST week 4 90.8 � 8.1y 80 � 5*

[Zn2þ]i, nM

Control 0.080 � 0.001 0.76 � 0.12

ALDOST week 1 0.175 � 0.022z 1.77 � 0.09*

ALDOST week 4 0.171 � 0.026z 1.64 � 0.08*

Values are mean � SEM; *p < 0.05. yp < 0.001 ALDOST vs. controls; 6 in each group. zp < 0.01.ALDOST ¼ aldosterone/salt treatment; PBMCs ¼ peripheral blood mononuclear cells.

Table 2H2O2 Production in PBMCs and Cardiac Mitochondriaand 8-Isoprostane Levels in PBMCs andCardiomyocytes

H2O2

GroupsPBMCs

(Mean Channel Brightness)Mitochondria,

pmoles/mg protein/min

Control 61.0 � 6.3 91.0 � 13.4

ALDOST wk 1 95.0 � 9.1* 118.3 � 8.4yALDOST wk 4 173.8 � 18.5* 157.6 � 13.2y

8-Isoprostane

PBMCs, pg/mg protein Cardiomyocytes, pg/mg protein

Control 162.3 � 25.4 32.8 � 5.8

ALDOST wk 1 259.9 � 37.8y 46.4 � 9.5yALDOST wk 4 214.6 � 22.3y 341.8 � 50.1y

Values are mean � SD. *p < 0.001 ALDOST vs. controls; 6 in each group. yp < 0.05.Abbreviations as in Table 1.

JACC: Heart Failure Vol. 1, No. 6, 2013 Gerling et al.December 2013:469–76 PBMCs as Novel Biomarkers

471

California), a subscription-based, Web-delivered (www.ingenuity.com) application that allows users to discover,visualize, and explore networks relevant to their experimentalresults. Our use of this pathway analysis approach waspreviously described in detail (27). The pathway analysis notonly serves as a tool to gain insights but also to eliminatefalse-positive genes because random genes (such as falsepositives) are much less likely to connect to pathways thangenes that are changed due to shifts in biological processesand biochemical pathways.Confirmation of genes by in situ hybridization and RT-PCR. We used RT-PCR and in situ hybridization asalternative approaches to evaluate and confirm geneexpression in heart tissue. RT-PCR was conducted on totalRNA extracted from cardiac tissue using Trizol Reagent(Invitrogen, Carlsbad, California) by methods describedpreviously (28). Gene-specific probes and primer sets werededuced using Universal ProbeLibrary Assay Design soft-ware (Roche Applied Science, Indianapolis, Indiana). In situhybridization of sections (16 mm) fixed in 4% formaldehydewas conducted, as previously described (29).

Results

Intracellular Ca2D and Zn2D. Cytosolic free [Ca2þ]i and[Zn2þ]i concentrations were determined in PBMCs and car-diomyocytes. An approximate doubling of these divalentcationswas found at 1week anddidnot change appreciably at 4weeks of ALDOST. The magnitude of these changes wassimilar in both PBMCs and cardiomyocytes, albeit the abso-lute amount of cytosolic free Ca2þ and Zn2þ differs betweenthese respective noncontractile and contractile cells (Table 1).Oxidative stress and free radical–induced damage. Anincrease in intracellular [Ca2þ]i and intramitochondrial[Ca2þ]m stimulates the generation of H2O2 by theseorganelles and consequent free radical–induced damage tolipids determined by increased levels of 8-isoprostane. BothPBMCs and mitochondria demonstrated a significantincrease in H2O2 generation at 1 week and with a furtherincrease at 4 weeks of ALDOST and an increase in 8-isoprostane levels in both PBMCs and cardiomyocytes at1 week and 4 weeks (Table 2).

Gene expression profiles of the heart. We examined geneexpression in heart tissue collected after 1 and 4 weeks ofALDOST and compared them with controls, using theAffymetrix Rat 230 2.0 expression array. Of the 31,099probe sets on this array, 22,538 showed expression (presentor marginal signal) in at least 1 of the 15 samples. Weconducted ANOVA on the 22,538 probe sets to definegenes that were affected by ALDOST (Benjamini-Hochberg multiple test correction p � 0.01) and identified83 probe sets (Online Table 1). Hierarchical clusteringindicated that each of these probe sets belonged to 1 of 2expression patterns that either quenched or amplified inresponse to ALDOST (Fig. 1A). However, it must beemphasized that the effect of ALDOST was already evidentat preclinical 1 week.

The list of probe sets clearly affected by ALDOST wassystematically mined for pathway connectivity using theIngenuity database. The 83 probe sets represented 52different genes for which the Ingenuity knowledge data-base had information to connect these genes to other genesin the database. Six of these genes, however, could not beconnected to any other genes from the list of 52. Theremaining 46 genes were connected in 4 networks with 13,13, 12, and 8 of these genes belonging to each cluster. The3 networks with >10 ALDOST-associated genes in eachwere fused into a single network using the merge networkfunction with the resulting network shown in Figure 1B.These 3 major networks were linked to pro-oxidant/proinflammatory-related genes NFkB, p38 MAPK, andTGF-b1.Gene expression profiles in PBMCs. We analyzed geneexpression in PBMCs collected from ALDOST ina manner similar to that described for heart tissue. We usedthe Affymetrix RAE 230A expression array and found that10,352 of the 15,923 probe sets on this array showedexpression (present or marginal signal) in at least 1 of the 15samples. We conducted ANOVA on the 10,352 probe setsto define genes that were affected by ALDOST using

Figure 1 Heat Map of Gene Expression Profiles in Heart Tissue From Control and Aldosterone-Salt Treated Rats

(A) Heat map of gene expression for control, week 1 and 4 aldosterone/salt treatment (ALDOST). Results from hierarchical clustering of the 83 probe sets in heart tissue

affected by ALDOST. Red indicates relatively high expression, whereas blue indicates relatively low expression. Each group had 6 rats. (B) Network created by the Ingenuity

server from the genes in heart tissue shown in A. The gray nodes indicate genes that are significantly affected by ALDOST, whereas the white icons represent genes, gene

families, and molecules that are needed to connect those ALDOST-affected genes to each other in a network.

Gerling et al. JACC: Heart Failure Vol. 1, No. 6, 2013PBMCs as Novel Biomarkers December 2013:469–76

472

a more stringent p value of �0.001, but did not applya multiple test correction to this dataset. This analysisrevealed that 55 probe sets were affected by the treatment

(Online Table 2). We conducted a hierarchical clusteringanalysis of these 55 probe sets with the resulting gene treeshown in Figure 2A. The genes that belong to subclusters

Figure 2 Heat Map of Gene Expression Profiles in PBMC From Control and Aldosterone-Salt Treated Rats

(A) Heat map of the gene expression for the 3 groups. Results from hierarchical clustering of the 55 probe sets in peripheral blood mononuclear cells (PBMCs) affected by

aldosterone/salt treatment (ALDOST). The solid bars on the right indicate the 42 probe sets that follow a clear pattern of increase or decrease in expression after 1 week of

ALDOST. Red indicates relatively high expression, whereas blue indicates relatively low expression. (B) Network created by the Ingenuity server from the 42 genes in PBMCs

shown in A. The gray nodes indicate genes that are significantly affected by ALDOST, whereas the white icons represent genes, gene families, and molecules that are needed to

connect those ALDOST-affected genes to each other in a network.

JACC: Heart Failure Vol. 1, No. 6, 2013 Gerling et al.December 2013:469–76 PBMCs as Novel Biomarkers

473

Figure 3In Situ Hybridization and RT-PCR Assessmentof TGF-b1

In situ hybridization with a transforming growth factor b1 (TGF-b1)–specific probe on

coronal sections of hearts obtained from rats serving as controls (A) and at

4 weeks aldosterone/salt treatment (ALDOST) (B). High-density TGF-b1 expressionin heart tissue is seen as the yellow-red areas corresponding to sites of fibrosis.

At 4 weeks ALDOST, tissue TGF-b1 was also evaluated using reverse-transcription

polymerase chain reaction (C). *p < 0.05 by the Student t test.

Gerling et al. JACC: Heart Failure Vol. 1, No. 6, 2013PBMCs as Novel Biomarkers December 2013:469–76

474

with a clear pattern of decrease or increase in response toALDOST are noted with a solid bar to the right of the genetree. It should be noted that for all of these 42 genes whoseexpression was affected by ALDOST, the responsesappeared greater at 1 week, long before cardiac pathologywas found.

The list of 42 probe sets, consisting of 27 different genes,clearly affected by ALDOST was mined for pathwayconnectivity using the Ingenuity database. Four of thesegenes could not be connected to any other genes from thelist of 27. The remaining 23 genes were connected in 2networks (with 12 and 11 genes in each) that were fusedinto a single network shown in Figure 2B. As in heart tissue,the network in PBMCs was linked to pro-oxidant/proinflammatory-related genes NFkB, p38 MAPK, andTGF-b1.Gene expression by in situ hybridization and RT-PCR.To confirm the central role of TGF-b1 in the genesis ofremodeling observed at sites of fibrosis, we conducted insitu hybridization on coronal sections and RT-PCR analysison RNA extracted from heart tissue. The results are shownin Figure 3. In situ hybridization demonstrated a moreintense signal for this cytokine in hearts of rats receivingALDOST (Fig. 3B) compared with tissue harvested fromcontrols (Fig. 3A). Further confirmation of this responsewas demonstrated by RT-PCR with a significant up-regulation of TGF-b1 in rats receiving ALDOST (Fig. 3C).

Discussion

We hypothesized that marked cation dyshomeostasis andoxidative stress induced by ALDOST will affect both themyocardium and PBMCs and reveal similar biochemicalchanges and gene expression profiles at preclinical andpathological stages of cardiac remodeling in HHD. Ourfindings imply that the PBMC transcriptome can serve asa surrogate marker of pathological events in the heart.

Increased concentrations of cytosolic free [Ca2þ]i and[Zn2þ]i were demonstrated in both PBMCs and car-diomyocytes during weeks 1 and 4 of ALDOST. Activationand induction of oxidative stress by these immune cells areknown to be Ca2þ dependent, whereas cytokine up-regulation is Zn2þ dependent. Increased intracellular Ca2þ

and Zn2þ levels are driven by PTH and up-regulation ofmetallothionein, a Zn2þ-binding protein (30). Ca2þ andZn2þ are integral to regulating oxidative stress and antiox-idant defenses in cardiomyocytes and PBMCs, respectively.PTH-mediated excessive [Ca2þ]i accumulation in PBMCswith induction of oxidative stress during mineralocorticoid/salt treatment can be prevented by parathyroidectomy (31),Ca2þ/Mg2þ/vitamin D3 supplementation (32), cotreatmentwith a calcium channel blocker (24), or cotreatment withspironolactone, which prevents secondary hyperpara-thyroidism (33). Increments in [Zn2þ]i-based antioxidantdefenses and corresponding cardioprotection are conferredby ZnSO4 supplementation (34–36), a Zn2þ ionophore

(35), or a b3 receptor agonist, which leads to endothelialnitric oxide (NOS) synthase–based generation of NO andNO-induced release of inactive Zn2þ bound to metal-lothionein (37,38). We found unequivocal evidence of

JACC: Heart Failure Vol. 1, No. 6, 2013 Gerling et al.December 2013:469–76 PBMCs as Novel Biomarkers

475

increased H2O2 production and consequent lipid perox-idation in both cardiomyocytes and their mitochondria, aswell in PBMCs during the preclinical stage.

To gain more specific and target tissue–based biomarkers,Affymetrix expression arrays were used to evaluate thecomprehensive gene expression patterns in heart tissue andPBMCs. Hierarchical clustering of genes eliciting a statisti-cally significant differential expression revealed that almostall of the effects of ALDOST were already present at 1week, and at 4 weeks, their expression remained, by andlarge, similar to that at 1 week.

To better illustrate the underlying processes of geneexpression abnormalities induced by ALDOST, we sub-jected the lists of differentially expressed genes to networkanalysis. The network created by genes whose expression inheart tissue is significantly affected by ALDOST wasdominated by genes involved in oxidative stress, inflamma-tion, and cell growth, as well as a large number of connec-tions to other genes related to growth factors, such asinsulin, vascular endothelial growth factor, tumor necrosisfactor a, and platelet-derived growth factor.

Using cardiac tissue and PBMC transcriptomes, weunraveled a similar signal-transducer-effector pathway to beinvolved that includes the 3 major signaling cascades: p38mitogen-activated protein kinase, a stress responsive protein;nuclear factor kB (NFkB), a downstream redox-sensitivetranscription factor and its proinflammatory target genes(e.g., ICAM-1, MCP-1, and TNF-a); and tissue repair viatransforming growth factor b1, a fibrogenic cytokine.However, we would not speculate on complete concordanceof gene expression profiles between noncontractile immunecells and contractile cardiomyocytes during stressor states.We recently reported on similarities and differences in geneprofiling between 2 contractile tissues, heart and muscle, inresponse to 4 weeks of ALDOST (39).

As demonstrated by in situ hybridization to detect andlocalize messenger ribonucleic acid expression and immu-nohistochemical labeling of protein in our previous studies(18), the network created by the PBMC transcriptome wassignificantly affected by ALDOST and dominated by genesinvolved in oxidative stress, inflammation, and cell growthand death. Central to these networks, together with a largenumber of connections to other genes, we found genesintegral to the process of inflammation, such as IL-6, NFkB,and p38 MAPK, and growth factors, such as hepatocytegrowth factor and transforming growth factor b1 to beinvolved in the preclinical stage. Furthermore, the networkcontains a number of caspases and apoptosis-inhibitinggenes, such as STAT3. Using the PBMC transcriptome asbiomarkers in patients with decompensated heart failure,transcripts related to inflammation and oxidative stress wereexpressed in abundance, including NFkB, MCP-1, tumornecrosis factor a and transforming growth factor b1 (3–9). Itwas recently suggested that PBMC gene expression profilingmay identify the preclinical stage of systolic heart failure(21), and the up-regulated PBMC gene network may

provide a molecular signature distinguishing patients withheart failure from healthy controls (8).

Our investigation of gene expression profiles in hearttissue and PBMCs indicated that although there are simi-larities in the molecular pathways activated by ALDOST,differences also exist. These differential expression patternsare perhaps due to the fact that the 2 processes are played outsomewhat differently in these 2 tissues. PBMCs are theeffectors of inflammation where the coronary vasculature isone of their targets. Within the networks created from the 2different tissues, we identified 3 central genes that arecommon: p38 MAPK, NFkB, and TGF-b1. We previouslyshowed that NFkB plays a central role in the proin-flammatory cardiac phenotype seen at week 4 ALDOST,together with the proinflammatory genes it regulates (18). Inthose studies, NFkB and 3-nitrotyrosine, an index ofoxidative damage, were shown to be amplified in the heartby immunohistochemistry. Both NFkB and p38 mitogen-activated protein kinase are redox sensitive and activatedby reactive oxygen species. Cotreatment with eitherN-acetylcysteine, an antioxidant, or a Zn2þ ionophoreattenuated these responses, implicating the relevance ofantioxidant defenses (18).Study limitations. We did not obtain serial blood samplesfrom the same animal at weeks 1 and 4 of ALDOST, nordid we investigate all transcriptome findings in heart tissuewith RT-PCR methodology. We did, however, confirm thegene chip data and central Ingenuity pathway analysis withour earlier interrogation of mRNA expression and up-regulated proteins using in situ hybridization and immu-nohistochemistry, respectively (18). Furthermore, the use ofpathway analysis can eliminate false-positive genes becausethose genes are much less likely to connect in pathways thantrue positives affected by changes in biological processes.Finally, we did not conduct a prediction analysis, as sug-gested by Heidecker et al. (40).

Conclusions

[Ca2þ]i and [Ca2þ]m overloading of cardiomyocytes andPBMC [Ca2þ]i occur, together with the induction ofoxidative stress, in rats during ALDOST. There is a simi-larity in the transcriptome of PBMCs and heart tissue that isfirst evident during the preclinical stage and persists into thepathological stage of chronic aldosteronism with HHD.These findings raise the prospect that PBMCs are capable ofrevealing the early diagnostic signs as a noninvasivebiomarker expression profile predictive of cardiac remodel-ing and thereby has the potential to serve as a novel sentineland useful molecular signature in monitoring car-dioprotection and management strategies in human HHD.

Reprint requests and correspondence: Dr. Karl T. Weber,Division of Cardiovascular Diseases, University of TennesseeHealth Science Center, 956 Court Avenue, Suite A312, Memphis,Tennessee 38163. E-mail: KTWeber@uthsc.edu.

Gerling et al. JACC: Heart Failure Vol. 1, No. 6, 2013PBMCs as Novel Biomarkers December 2013:469–76

476

REFERENCES

1. Delafontaine P, Akao M. Angiotensin II as candidate of cardiaccachexia. Curr Opin Clin Nutr Metab Care 2006;9:220–4.

2. Molavi B, Mehta JL. Oxidative stress in cardiovascular disease:molecular basis of its deleterious effects, its detection, and therapeuticconsiderations. Curr Opin Cardiol 2004;19:488–93.

3. Radovanovic S, Savic-Radojevic A, Pljesa-Ercegovac M, et al. Markersof oxidative damage and antioxidant enzyme activities as predictors ofmorbidity and mortality in patients with chronic heart failure. J CardFail 2012;18:493–501.

4. Caruso R, Verde A, Campolo J, et al. Severity of oxidative stress andinflammatory activation in end-stage heart failure patients are unalteredafter 1 month of left ventricular mechanical assistance. Cytokine 2012;59:138–44.

5. Yndestad A, Damås JK, Eiken HG, et al. Increased gene expression oftumor necrosis factor superfamily ligands inperipheral bloodmononuclearcells during chronic heart failure. Cardiovasc Res 2002;54:175–82.

6. Zhao SP, Xu TD. Elevated tumor necrosis factor alpha of bloodmononuclear cells in patients with congestive heart failure. Int J Cardiol1999;71:257–61.

7. Satoh M, Iwasaka J, Nakamura M, Akatsu T, Shimoda Y, Hiramori K.Increased expression of tumor necrosis factor-a converting enzyme andtumor necrosis factor-a in peripheral bloodmononuclear cells in patientswith advanced congestive heart failure. Eur J Heart Fail 2004;6:869–75.

8. Cappuzzello C, Napolitano M, Arcelli D, et al. Gene expressionprofiles in peripheral blood mononuclear cells of chronic heart failurepatients. Physiol Genomics 2009;38:233–40.

9. Vo TK, de Saint-Hubert M, Morrhaye G, et al. Transcriptomicbiomarkers of the response of hospitalized geriatric patients admittedwith heart failure. Comparison to hospitalized geriatric patients withinfectious diseases or hip fracture. Mech Ageing Dev 2011;132:131–9.

10. Shane E, Mancini D, Aaronson K, et al. Bone mass, vitamin D defi-ciency, and hyperparathyroidism in congestive heart failure. Am J Med1997;103:197–207.

11. Alsafwah S, LaGuardia SP, Nelson MD, et al. Hypovitaminosis D inAfrican Americans residing in Memphis, Tennessee with and withoutheart failure. Am J Med Sci 2008;335:292–7.

12. Sugimoto T, Tanigawa T, Onishi K, et al. Serum intact parathyroidhormone levels predict hospitalisation for heart failure. Heart 2009;95:395–8.

13. Carbone LD, Cross JD, Raza SH, et al. Fracture risk in men withcongestive heart failure. Risk reduction with spironolactone. J Am CollCardiol 2008;52:135–8.

14. van Diepen S, Majumdar SR, Bakal JA, McAlister FA, Ezekowitz JA.Heart failure is a risk factor for orthopedic fracture: a population-basedanalysis of 16,294 patients. Circulation 2008;118:1946–52.

15. Kenny AM, Boxer R, Walsh S, Hager WD, Raisz LG. Femoral bonemineral density in patients with heart failure. Osteoporos Int 2006;17:1420–7.

16. López B, González A, Varo N, Laviades C, Querejeta R, Díez J.Biochemical assessment of myocardial fibrosis in hypertensive heartdisease. Hypertension 2001;38:1222–6.

17. Zannad F, Rossignol P, Iraqi W. Extracellular matrix fibrotic markersin heart failure. Heart Fail Rev 2010;15:319–29.

18. Sun Y, Zhang J, Lu L, Chen SS, Quinn MT, Weber KT. Aldosterone-induced inflammation in the rat heart. Role of oxidative stress. Am JPathol 2002;161:1773–81.

19. Liew CC, Ma J, Tang HC, Zheng R, Dempsey AA. The peripheralblood transcriptome dynamically reflects system wide biology: a poten-tial diagnostic tool. J Lab Clin Med 2006;147:126–32.

20. Seiler PU, Stypmann J, Breithardt G, Schulze-Bahr E. Real-time RT-PCR for gene expression profiling in blood of heart failure patientsdapilot study: gene expression in blood of heart failure patients. Basic ResCardiol 2004;99:230–8.

21. Smih F, Desmoulin F, Berry M, et al. Blood signature of pre-heartfailure: a microarrays study. PLoS One 2011;6:e20414.

22. Gerling IC, Sun Y, Ahokas RA, et al. Aldosteronism: an immunos-timulatory state precedes the proinflammatory/fibrogenic cardiacphenotype. Am J Physiol Heart Circ Physiol 2003;285:H813–21.

23. Ahokas RA, Warrington KJ, Gerling IC, et al. Aldosteronism andperipheral blood mononuclear cell activation. A neuroendocrine-immune interface. Circ Res 2003;93:e124–35.

24. Ahokas RA, Sun Y, Bhattacharya SK, Gerling IC, Weber KT.Aldosteronism and a proinflammatory vascular phenotype. Role ofMg2þ, Ca2þ and H2O2 in peripheral blood mononuclear cells.Circulation 2005;111:51–7.

25. Kamalov G, Ahokas RA, Zhao W, et al. Temporal responses tointrinsically coupled calcium and zinc dyshomeostasis in cardiac myo-cytes and mitochondria during aldosteronism. Am J Physiol Heart CircPhysiol 2010;298:H385–94.

26. Wu J, Kakoola DN, Lenchik NI, Desiderio DM, Marshall DR,Gerling IC. Molecular phenotyping of immune cells from young NODmice reveals abnormal metabolic pathways in the early induction phaseof autoimmune diabetes. PLoS One 2012;7:e46941.

27. Gerling IC, Singh S, Lenchik NI, Marshall DR, Wu J. New dataanalysis and mining approaches identify unique proteome and tran-scriptome markers of susceptibility to autoimmune diabetes. Mol CellProteomics 2006;5:293–305.

28. Zhao W, Zhao T, Chen Y, Qu Y, Gerling IC, Sun Y. Modification ofoxidative stress on gene expression profiling in the rat infarcted heart.Mol Cell Biochem 2013;379:243–53.

29. Zhao W, Zhao T, Chen Y, Ahokas RA, Sun Y. Oxidative stressmediates cardiac fibrosis by enhancing transforming growth factor-beta1 in hypertensive rats. Mol Cell Biochem 2008;317:43–50.

30. Thomas M, Vidal A, Bhattacharya SK, et al. Zinc dyshomeostasis inrats with aldosteronism. Response to spironolactone. Am J PhysiolHeart Circ Physiol 2007;293:H2361–6.

31. Vidal A, Sun Y, Bhattacharya SK, Ahokas RA, Gerling IC, Weber KT.Calcium paradox of aldosteronism and the role of the parathyroidglands. Am J Physiol Heart Circ Physiol 2006;290:H286–94.

32. Goodwin KD, Ahokas RA, Bhattacharya SK, Sun Y, Gerling IC,Weber KT. Preventing oxidative stress in rats with aldosteronism bycalcitriol and dietary calcium and magnesium supplements. Am J MedSci 2006;332:73–8.

33. Chhokar VS, Sun Y, Bhattacharya SK, et al. Hyperparathyroidismand the calcium paradox of aldosteronism. Circulation 2005;111:871–8.

34. Gandhi MS, Deshmukh PA, Kamalov G, et al. Causes and conse-quences of zinc dyshomeostasis in rats with chronic aldosteronism.J Cardiovasc Pharmacol 2008;52:245–52.

35. Kamalov G, Ahokas RA, Zhao W, et al. Uncoupling the coupledcalcium and zinc dyshomeostasis in cardiac myocytes and mitochondriaseen in aldosteronism. J Cardiovasc Pharmacol 2010;55:248–54.

36. Wang J, Song Y, Elsherif L, et al. Cardiac metallothionein inductionplays the major role in the prevention of diabetic cardiomyopathy byzinc supplementation. Circulation 2006;113:544–54.

37. Cheema Y, Sherrod JN, Zhao W, et al. Mitochondriocentricpathway to cardiomyocyte necrosis in aldosteronism: cardioprotectiveresponses to carvedilol and nebivolol. J Cardiovasc Pharmacol 2011;58:80–6.

38. Khan MU, Zhao W, Zhao T, et al. Nebivolol: a multifaceted antiox-idant and cardioprotectant in hypertensive heart disease. J CardiovascPharmacol 2013 Aug 4 [E-pub ahead of print].

39. Cheema Y, Zhao W, Zhao T, et al. Reverse remodeling and recoveryfrom cachexia in rats with aldosteronism. Am J Physiol Heart CircPhysiol 2012;303:H486–95.

40. Heidecker B, Kittleson MM, Kasper EK, et al. Transcriptomicbiomarkers for the accurate diagnosis of myocarditis. Circulation 2011;123:1174–84.

Key Words: aldosteronism - calcium - cardiomyocytes - gene chiparray - oxidative stress - peripheral blood mononuclear cells - zinc.

APPENDIX

For supplemental tables, please see the online version of this article.