myocardial expression of cc- and cxc-chemokines and their receptors in human end-stage heart failure

Cardiovascular Research 47 (2000) 778–787www.elsevier.com/ locate /cardiores

www.elsevier.nl / locate /cardiores

Myocardial expression of CC- and CXC-chemokines and their receptors inhuman end-stage heart failure

a,c , c,f f c c,f c,d*˚Jan K. Damas , Hans G. Eiken , Erik Øie , Vigdis Bjerkeli , Arne Yndestad , Thor Ueland ,g b a a gTheis Tønnessen , Odd R. Geiran , Halfdan Aass , Svein Simonsen , Geir Christensen ,

c,e f a c,e˚ ˚Stig S. Frøland , Havard Attramadal , Lars Gullestad , Pal AukrustaDepartment of Cardiology, Division of Heart and Lung Diseases, The National Hospital, Oslo, Norway

bDepartment of Cardiothoracic Surgery, Division of Heart and Lung Diseases, The National Hospital, Oslo, NorwaycResearch Institute for Internal Medicine, The National Hospital, University of Oslo, Oslo, Norway

dSection of Clinical Immunology and Infectious Diseases, Medical Department, The National Hospital, Oslo, NorwayeSection of Endocrinology, Medical Department, The National Hospital, Oslo, Norway

fMSD-Cardiovascular Research Center, The National Hospital, Oslo, Norwayg ˚Institute for Experimental Medical Research, Ulleval Hospital, University of Oslo, Oslo, Norway

Received 10 February 2000; accepted 15 May 2000

Abstract

Objectives: Chemokines regulate several biological processes, such as chemotaxis, collagen turnover, angiogenesis and apoptosis.Based on the persistent immune activation with elevated circulating levels of chemokines in patients with congestive heart failure (CHF),we have hypothesised a pathogenic role for chemokines in the development of CHF. The objective of this study was to examine mRNAlevels and cellular localisation of chemokines and chemokine receptors in human CHF. Methods: We examined explanted hearts from tenpatients with end-stage heart failure (all chambers) and in ten organ donors using an RNase protection assays and immunohistochemicaltechniques. Results: Our main findings were: (i) expression of eight chemokine and nine chemokine receptor genes in both failing andnonfailing myocardium, (ii) particularly high mRNA levels of monocyte chemoattractant protein (MCP)-1 and CXC-chemokine receptor4 (CXCR4), in both chronic failing and nonfailing myocardium, (iii) decreased mRNA levels of MCP-1 and interleukin (IL)-8 in thefailing left ventricles compared to failing left atria, (iv) decreased chemokine (e.g., MCP-1 and IL-8) and increased chemokine receptor(e.g., CCR2, CXCR1) mRNA levels in failing left ventricles and failing left atria compared to corresponding chambers in the nonfailinghearts and (v) immunolocalisation of MCP-1, IL-8 and CXCR4 to cardiomyocytes. Conclusion: The present study demonstrates for thefirst time chemokine and chemokine receptor gene expression and protein localisation in the human myocardium, introducing a newfamily of mediators with potentially important effects on the myocardium. The observation of chemokine dysregulation in humanend-stage heart failure may represent a previously unknown mechanism involved in progression of chronic heart failure. 2000Elsevier Science B.V. All rights reserved.

Keywords: Cytokines; Gene expression; Heart failure; Infection / inflammation; Monoclonal antibodies

1. Introduction nisms such as promotion of cardiomyocyte hypertrophy [3]and apoptosis [4], as well as alternation in extracellular

Inflammatory processes seem to be involved in the matrix in the myocardium [5].pathogenesis and progression of congestive heart failure Chemokines represent a family of inflammatory cyto-(CHF). Previous studies have shown that inflammatory kines that control chemotaxis of leukocyte subsets intocytokines, e.g., tumor necrosis factor (TNF)-a, interleukin inflamed tissue [6]. While CC-chemokines are potent(IL)-1 and IL-6-related cytokines, may induce myocardial chemoattractants and activators for monocytes and lym-dysfunction [1,2] and cardiac remodelling through mecha- phocytes, most CXC-chemokines attract neutrophils [7].

However, in addition to chemotaxis, several functional

*Corresponding author. Tel.: 147-230-736-28; fax: 147-230-736-30.˚E-mail address: [email protected] (J.K. Damas). Time for primary review 28 days.

0008-6363/00/$ – see front matter 2000 Elsevier Science B.V. All rights reserved.PI I : S0008-6363( 00 )00142-5

by guest on February 6, 2016

Dow

nloaded from

˚J.K. Damas et al. / Cardiovascular Research 47 (2000) 778 –787 779

responses, including adherence to endothelium, enzyme (left and right ventricles and atria) from ten patients withsecretion and induction of respiratory burst, are observed end-stage CHF [coronary artery disease (CAD), n57;in vitro after chemokine stimulation of these leukocyte idiopathic dilated cardiomyopathy (IDCM), n52; valvularsubsets [8,9]. heart disease, n51] undergoing heart transplantation were

So far, some studies have suggested a role for used in the study (seven men and three women; aged 21 tochemokines in the pathogenesis of various heart diseases, 57 years; mean, 45.663.5 years). All patients were insuch as atherosclerosis [10], myocarditis [11] and reperfu- NYHA functional classes III and IV, and all had a leftsion damage of ischemic myocardium [12]. Recently, we ventricular ejection fraction (EF) ,35%. All were treatedhave demonstrated elevated circulating levels and en- with angiotensin converting enzyme inhibitors, 60% withhanced release from monocytes and platelets of both CC- digitoxin, 60% with b-blockers, but none were receivingand CXC-chemokines in CHF [13,14], suggesting a pos- intravenous ionotropic support before transplantation. Non-sible pathogenic role of these mediators also in the failing human heart tissue was obtained from suitable heartprogression of heart failure. Indeed, chronic low-grade donors (left atria, n56) and from subjects whose heartsinflammation, as reflected in activated vascular endo- were rejected for cardiac donation because of functionalthelium and the presence of infiltrating inflammatory cells, criteria or no recipient available (left ventricles, n54). Thehas been found in the failing myocardium [15]. cause of death of the donors was acute cerebrovascularChemokines may be important mediators in this process by accident or trauma, and none of the organ donors had apromoting attraction and invasion of activated leukocytes. history of heart failure. The nonfailing hearts were ob-Furthermore, recent reports indicate that, besides being tained from six men and four women, aged 19 to 45 yearspotent inducers of chemotaxis and leukocyte activation, (mean, 37.963.6 years). In addition, total RNA extractedchemokines may have several other biological properties from human heart tissue was obtained from Clontech (Palowith relevance to the pathogenesis of heart failure, such as Alto, CA, USA). Clinical, demographic and hemodynamicinduction of cell proliferation and involvement during data of the CHF study population are presented in Table 1.organogenesis [16] and cardiogenesis [17]. Based on the Tissue aliquots from the failing myocardium of cardiacpossible pathogenic role of chemokines in the development recipients were removed from the still-contracting heartsand progression of heart failure, we have examined immediately after explantation, and immersed in liquidmyocardial gene expression and cellular localisation of nitrogen before storage at 2808C until use. Care was takenchemokines and their corresponding receptors in human not to sample scarred, fibrotic, or adipose tissue, endo-end-stage heart failure. cardium, epicardium, or great vessels. Routine histology

was performed in all explanted hearts and there were nosigns of acute or chronic myocarditis in any of the hearts

2. Methods included in the study. Hearts from actual donors and fromsubjects rejected as heart donors were kept on iced water

2.1. Patients and myocardial tissue samples for 42 to 206 min (mean ischemic time, 122632 min)before tissue was processed as described above.

Studies were performed in accordance with institutionalhuman studies guidelines and conform with the principles 2.2. RNA preparationoutlined in the Declaration of Helsinki. Explanted hearts

Total RNA was extracted from frozen atrial and ven-

Table 1aClinical, demographic and hemodynamic data for CHF patients

Patient EF CI PCW Drug treatment2% l /min?m mmHg

No. Sex Age (years) Diagnosis ACE-I Digitoxin b-blockers

CHF patients1 M 42 CAD 18 1.3 22 1 2 1

2 F 57 CAD 20 1.9 36 1 1 1

3 F 49 CAD 10 1.6 14 1 1 1

4 M 21 VD 25 2.3 20 1 2 2

5 M 40 CAD 20 1.8 28 1 2 1

6 M 54 IDCM 15 1.5 21 1 2 2

7 M 51 CAD 30 2.4 9 1 1 2

8 F 57 CAD 17 2.3 25 1 1 2

9 M 37 IDCM 20 1.8 24 1 1 1

10 M 48 CAD 26 1.7 33 1 1 1

Mean6S.D. 45.669.8 20.165.8 1.960.4 23.268.1a CAD, coronary artery disease; IDCM, idiopathic dilated cardiomyopathy; VD, valvular disease; EF, left ventricular ejection fraction; CI, cardiac index;

PCW, pulmonary capillary wedge pressure; ACE-I, angiotensin-converting enzyme inhibitors.


Dow

nloaded from

˚780 J.K. Damas et al. / Cardiovascular Research 47 (2000) 778 –787

tricular tissue using a modification of the acid guanidinium 2.4. Immunohistochemistrythiocyanate (GTC)–phenol–chloroform method [18].

Heart tissue from failing hearts (left ventricles and atria)Briefly, the frozen tissue was minced in liquid nitrogen,and from nonfailing hearts (left atria) were fixed with 4%placed in GTC-solution and homogenised with an Ultra-paraformaldehyde and embedded in paraffin wax. TheTurrax homogeniser for 30 s. Subsequently, 2 mol / lparaffin-embedded myocardial tissue was cut into 7-mmsodium acetate, pH 4, and chloroform–phenol, 5:1 (v /v),sections, dewaxed in xylene, and subsequently rehydratedpH 4, were added. The mixture was shaken vigorously.in descending concentrations of ethanol. The sections wereAfter centrifugation (12,000 g for 20 min), the aqueousblocked with normal horse serum and subsequently incu-phase was collected, and the RNA was precipitated by thebated with mouse monoclonal anti-human IL-8, anti-addition of an equal volume of isopropanol. A doublehuman MCP-1 or anti-human CXCR4 antibodies (R&Dextraction and DNase I treatment (RQI DNase, Promega,Systems, Minneapolis, MD, USA). To block the presenceMadison, WI, USA) was routinely used to eliminate smallof endogenous peroxidase activity, the slides were incu-amounts of DNA contamination. The RNA pellet wasbated with 0.3% hydrogen peroxide in methanol for 30 minwashed with 75% ethanol and resuspended in RNA storageat room temperature and were washed in phosphate-buf-solution (Ambion, Austin, TX, USA) and stored at 2808Cfered saline (PBS) for 5 min. Anti-IL-8, anti-MCP-1 anduntil used. The integrity of the extracted total RNA wasanti-CXCR4 immunoreactivity were amplified by theassessed by agarose gel electrophoresis and ethidiumavidin–biotin–peroxidase system (Vectastain Elite kit,bromide staining. RNA concentration and purity wereVector laboratories, Burlingame, CA, USA) according toevaluated in duplicate by measuring the absorbance atthe manufacturer’s instructions. Briefly, the sections were260/280 nm using a spectrophotometer (GeneQuant; Phar-incubated with biotinylated horse anti-mouse IgG for 30macia, Uppsala, Sweden).min at room temperature, washed in PBS, and incubatedwith the avidin–biotin–peroxidase complex for 30 min.After a final wash in PBS, the slides were incubated with2.3. RNase protection assay (RPA)diaminobenzidine as the chromogen in a commercialmetal-enhanced system (Pierce Chemical), and the sectionsRPA was used for the detection and quantification ofwere counterstained with hematoxylin. Omission of pri-mRNA species. Multi-probes hCK5 (eight chemokinemary antibody and replacing it with nonimmune mouseprobes), hCR5 (eight CC-chemokine receptor probes) andIgG served as the negative control. In addition, sectionshCR6 (seven CXC-chemokine receptor probes) were avail-were stained with van Gieson to identify extracellularable with reagents for in vitro transcription and RPAcollagen in the myocardium.(RiboQuant; Pharmingen, San Diego, CA, USA). The

following anti-sense RNA probes, which were able to2.5. Statistical analysishybridise with target human mRNA, were synthesised:

lymphotactin (Ltn), regulated on activation normally T-For comparison between different individuals or cardiac

cell expressed and secreted (RANTES), interferon g-in-chambers, one-way ANOVA and the Bonferonni multiple

ducible protein (IP)-10, macrophage inflammatory proteincomparison procedure were used. Data were log-trans-

(MIP)-1a and -1b, MCP-1, IL-8 and inducible (I)-309, theformed when not normally distributed. Probability values

CC-chemokine receptors (CCR)1–5, TER1 (CCR8), theare considered significant when P,0.05. Data are ex-

CXC-chemokine receptors (CXCR)1–4, Burkitt’s lym-pressed as mean6S.E.M. if not otherwise stated. Analysis

phoma receptor (BLR)-1, -2, V28 (CX CR) and ribosomal3 utilised the computer program PRISM (GraphPad, Sanprotein (rp) L32, and glyceraldehyde-3-phosphate dehydro-

Diego, CA, USA).genase (GAPDH).

For all hybridisation assays, |2 mg total RNA from each6patient or donor sample were mixed with 0.5–2310 cpm 3. Results32of probe (a- P-UTP, 3000 Ci /mmol). Protected fragments

were separated in a denaturing 6% polyacrylamide gel for 3.1. Gene expression of chemokines and chemokine90 min. The dried gel was exposed to a phosphorimaging receptors in the human myocardiumscreen (Cyclone system; Packard, Meriden, CT, USA) for20 h followed by densitometric analysis using 1D Quan- RPA was used to detect and quantify chemokine andtifier (Phoretix, Newcastle, UK). A quantitative estimate of chemokine receptor mRNA abundance in total RNAthe mRNA levels for chemokines and chemokine receptors extracted from explanted failing (n510; four chambers, i.e.was attempted by comparing the results with the mRNA 40 samples) and nonfailing hearts (n510; six left atria,levels of rpL32 and GAPDH genes. The intra-experimental four left ventricles). Out of 23 different chemokine andcoefficients of variation (C.V.) were 13.4% (rpL32, n520) chemokine receptor genes tested, 17 were detected in bothand 15.2% (GAPDH, n520). failing and nonfailing hearts using RPA (Fig. 1A–C).


Dow

nloaded from


These genes were expressed at different levels in all Notably, MCP-1 mRNA levels were up to fivefold higherchambers of the failing hearts. The relative levels of gene than for any other chemokines, and CXCR4 was clearlyexpression in both failing and nonfailing hearts were the highest expressed chemokine receptor gene (Table 2).highest for MCP-1, IL-8, RANTES and MIP-1a, repre-senting both CC- and CXC-chemokines (Fig. 1 and Table 3.2. Quantitative assessment of chemokine and2). Furthermore, the CC-chemokine receptors CCR1, chemokine receptor gene expression in failing andCCR2, CCR4 and CCR5, and the CXC-chemokine re- nonfailing myocardiumceptors CXCR1, CXCR2, V28 and CXCR4 were expressedin both failing and nonfailing hearts (Fig. 1 and Table 2). Several significant differences between chemokine and

Fig. 1. RNase protection assay (RPA) of chemokine (A), CC-chemokine receptor (B) and CXC-chemokine receptor (C) gene expression in all chambers ofa failing heart, i.e., left ventricle (LV), left atrium (LA), right ventricle (RV), right atrium (RA), and from left ventricle and left atrium of nonfailing donorhearts. The autoradiographic images from this analysis shows the in vitro transcribed RNA from multi-probes on the left panel (labeled Probes) and thecorresponding RNase-protected fragments after hybridisation with human mRNA on the six right panels. By visual inspection, the autoradiograms showmarked differences in gene expression between the cardiac chambers, and the high gene expression of MCP-1 in left atrium compared with left ventricle(see also Fig. 2 and Table 2). Note, each probe band migrates slower than its protected bands due to flanking sequences that are not protected by mRNA.


Dow

nloaded from


Table 2aMyocardial gene expression of chemokines and corresponding receptors in human end-stage heart failure

Chemokines Atria Ventricles Receptors Atria Ventricles

Right Left Right Left Right Left Right Left

MCP-1 10.162.4 14.562.5 6.760.3 3.860.6*** CCR2 0.360.1 0.560.1 0.360.1 0.560.2IL-8 1.060.2 1.260.1 0.560.3 0.360.1* CXCR1, 0.160.1 0.960.2 0.260.1 0.660.2†

CXCR2 0.260.1 0.760.2 0.360.2 0.760.2†RANTES 2.060.2 1.460.2 1.660.2 0.960.1 CCR1 1.060.6 0.860.1 1.060.6 1.260.2MIP-1a 0.360.1 0.460.1 0.260.1 0.360.1 CCR5 0.260.1 0.360.1 0.260.1 0.360.2SDF-1 2 2 2 2 CXCR4 1.960.7 2.460.5 1.160.2 2.060.6

a Quantification of chemokine and chemokine receptor mRNA levels in ten failing hearts (atrium and ventricle, right and left). Data are normalisedagainst the housekeeping gene rpL32 and are given as a percentage of rpL32 expression (mean6S.E.M.). Stromal cell-derived factor (SDF)-1 was notexamined.

***P,0.001 vs. left atrium.*P,0.05 vs. left atrium.†P,0.01 vs. right ventricle.

chemokine receptor gene expression in failing and nonfail- expression were significantly raised in left atria compareding hearts (left ventricles and atria) were observed. First, to left ventricles (see below), no differences in chemokineMCP-1, IL-8, RANTES and MIP-1a mRNA levels were mRNA expression between these chambers were found insignificantly lower in failing left ventricles and atria the nonfailing hearts (Fig. 2). Finally, in contrast tocompared to corresponding chambers in the nonfailing chemokine mRNA levels, chemokine receptor mRNAhearts (Fig. 2). Second, while MCP-1 and IL-8 mRNA expression, represented by CCR1, CCR2 and CXCR1, was

Fig. 2. Relative mRNA levels of MCP-1 (A), IL-8 (B), RANTES (C) and MIP-1a (D) in left ventricles (FLV) and left atria (FLA) from ten patients withend-stage CHF and in nonfailing left ventricles (NFLV, n54) and nonfailing left atria (NFLA, n56). mRNA levels were quantified using an RNaseprotection assay (RPA) and phosphorimaging, and were normalised against the housekeeping gene product, rpL32. Data are presented as a percentage ofrpL32 expression.


Dow

nloaded from


significantly increased in both left ventricles and atria fromfailing myocardium compared to samples from nonfailinghearts (Fig. 3). Similar patterns of gene expression werealso found for CCR4, CCR5, V28, CXCR2 (data notshown) and CXCR4 (Fig. 3), comparing failing andnonfailing hearts, although the differences did not reachstatistical significance. In contrast, the expression of rpL32relative to GAPDH was stable between failing and nonfail-ing myocardium and very similar results were found onanalysing chemokine and chemokine receptor genes rela-tive to the second control gene GAPDH (Fig. 4, see alsoMethods).

3.3. Quantitative assessment of chemokine and Fig. 4. Relation between mRNA levels of rpL32 and GAPDH in leftchemokine receptor gene expression in different ventricles (FLV) and left atria (FLA) from ten patients with end-stage

CHF and in nonfailing left ventricles (NFLV, n54) and nonfailing leftchambers of the failing heartatria (NFLA, n56).

Several differences were detected in both chemokineand chemokine receptor mRNA levels between the cardiac found for RANTES and MIP-1a, although the differenceschambers in the failing myocardium. MCP-1 (Figs. 2A and did not reach statistical significance (Table 2).Variations in5A) and IL-8 (Fig. 2B) mRNA levels were significantly gene expression between cardiac chambers were alsoelevated in the left atria compared to left ventricles in the found for chemokine receptor mRNA levels, although afailing hearts. Similar patterns of gene expression were systematic difference between left atria and left ventricles

Fig. 3. Relative mRNA levels of CCR1 (A), CCR2 (B), CXCR1 (C) and CXCR4 (D) in left ventricles (FLV) and left atria (FLA) from ten patients withend-stage CHF and in nonfailing left ventricles (NFLV, n54) and nonfailing left atria (NFLA, n56). mRNA levels were quantified using an RNaseprotection assay (RPA) and phosphorimaging, and were normalised against the housekeeping gene product, rpL32. Data are presented as a percentage ofrpL32 expression.


Dow

nloaded from


MCP-1-ir and CXCR4-ir in left atria and left ventricles offailing hearts and in left atria of nonfailing hearts (Fig. 6).Cardiomyocytes and the vascular smooth muscle cellsdisplayed IL-8-ir, MCP-1-ir and CXCR4-ir. Histochemicalanalysis using van Gieson revealed myocardial depositionof collagen predominantly in perivascular tissue and theseareas were not immunoreactive to either IL-8, MCP-1 orCXCR4. The degree of anti-CXCR4 and anti-MCP-1immunostaining was consistent with the mRNA levelsshowing increased CXCR4-ir and decreased MCP-1-ir infailing compared to nonfailing myocardium. Furthermore,MCP-1-ir was higher in atria compared to ventricles in thefailing heart. As for IL-8-ir, there were only minordifferences between failing and nonfailing myocardium.Omitting primary antibody and replacing it with non-immune IgG abolished the signals, demonstrating spe-cificity of the IL-8, MCP-1 and CXCR4 antibodies.

4. Discussion

The present study is the first to demonstrate that thehuman myocardium, both under failing and nonfailingconditions, expresses significant levels of several CC- andCXC-chemokines and their corresponding receptors, withparticularly high levels of MCP-1 and CXCR4. This wasconfirmed at both gene and corresponding protein levels.

Fig. 5. Relative chemokine (A) and chemokine receptor (B) mRNA Finally, by immunohistochemistry, we demonstrated thatlevels (represented by MCP-1 and CXCR4, respectively) in different

the chemokines MCP-1 and IL-8 and the chemokinecardiac chambers of ten end-stage failing hearts. mRNA levels werereceptor CXCR4 are located to the cardiomyocytes and notquantified using an RNase protection assay (RPA) and phosphorimaging,only to ‘contaminating’ leukocytes, endothelial cells orand were normalised against the housekeeping gene product, rpL32. Data

are presented as a percentage of rpL32 expression. Note, all failing left fibroblasts. Chemokines have recently attracted consider-ventricles have decreased levels of MCP-1 mRNA compared to failing able attention as important mediators of inflammation andleft atria.

host defence. Our findings in the present study suggest thatthe ‘chemokine system’ also may represent a previously

was absent (Fig. 5B and Table 2). However, the mRNA unrecognised factor in the pathogenesis of heart failure.expression of the receptors CXCR1 and CXCR2 was Although chemokines are normally considered beneficialsignificantly higher (P,0.01) and tended to be higher for in wound healing, hematopoiesis, organogenesis and in theCCR2 and CXCR4 in the left compared to right ventricles clearance of infectious organisms [16,17], dysregulated(Table 2). In contrast to these variations in chemokine and expression of chemokines has been associated with chronicchemokine receptor mRNA expression according to ana- inflammatory conditions such as atherosclerosis [10],tomical site, the variation in mRNA levels of the house- arthritis [19], bronchial asthma [20], and inflammatorykeeping genes for rpL32 and GAPDH were almost similar bowel syndrome [21]. Several reports indicate a role forto the intra-experimental C.V., indicating stable expression immunologic and inflammatory processes in the patho-in heart failure and equal cellular content in the myocardial genesis of CHF [1–5,13,14]. We would like to suggest thatsamples. chemokines may be important mediators in this process by

promoting attraction and invasion of activated leukocytes3.4. Immunohistochemical analysis into the failing myocardium. Indeed, interstitial monocyte

infiltration in the myocardium with development of aCellular localisation was investigated for the highest number of pathological changes characterising CHF, in-

expressed CC-, CXC-chemokine and chemokine receptor cluding cardiac hypertrophy, ventricular dilatation andgene (MCP-1, IL-8 and CXCR4, respectively) in left depressed contractile function, is found in transgenic miceventricles and atria from failing and in left atria from with myocardial over-expression of MCP-1 [11]. Some ofnonfailing hearts. Immunohistochemical analysis of the these pathological changes may be caused by the ability ofmyocardium revealed IL-8-like immunoreactivity (IL-8-ir), MCP-1, as well as other chemokines, to induce production


Dow

nloaded from


Fig. 6. Representative photomicrographs showing staining and localisation of IL-8-ir (B), MCP-1-ir (C) and CXCR4-ir (D) in myocardial sections of leftventricular tissue of a failing heart. (A) Absence of staining after replacing of primary antibody with mouse non-immune IgG. Original magnification:3400.

of proteolytic enzymes, reactive oxygen species and in- both CC- and CXC-chemokine receptor mRNA expressionflammatory cytokines in recruited myocardial monocytes in human myocardium, with particularly high levels of[22–24]. However, the present study suggests that the CXCR4 and anti-CXCR4 immunostaining of car-cardiomyocytes themselves also may have an active role in diomyocytes, suggest that chemokines may directly modu-inflammatory processes, not only by producing several late cardiomyocyte function. The observation of embryonicchemokines and in particular MCP-1, but also by express- lethality and developmental defects, including cardiacing several chemokine receptors of both the CC and CXC ventricular septum defects in CXCR4 knock-out mice,subtypes. indicates a direct chemokine-mediated effect on the

Although the primary function of chemokines is thought myocardium [29]. Furthermore, enhanced cardiomyocyteto be recruitment of circulating leukocytes into sites of apoptosis appears to be involved in the pathogenesis ofinflammation, recent studies suggest that chemokines may end-stage CHF [30] and, notably, CXCR4 may triggeralso mediate other biological effects, such as cell prolifer- apoptosis in various cell types [28,31,32], underscoring theation [17], collagen turnover [25], modulation of matrix pathogenic potential of our findings of CXCR4 expressionmetalloproteinase activity [26], angiogenesis [27] and in the failing left ventricle. Finally, the myocardial expres-induction of apoptosis [28]. Several of these are clearly of sion of chemokine receptors is also of interest with regardinterest with regard to the pathogenesis of heart failure. to the development of dilated cardiomyopathy in AIDSThus, although there are no reports of direct chemokine- patients. By being co-receptors for human immuno-mediated effects on cardiomyocytes, our novel findings of deficiency virus (HIV) [33], CCR5 and CXCR4 may be of


Dow

nloaded from


importance in permitting HIV entry into the myocardium, 5. Conclusionwhich recently has been demonstrated in AIDS-relatedcardiomyopathy [34]. Although further studies are needed The present study demonstrates for the first timeto clarify the potential role of chemokines and their chemokine and chemokine receptor mRNA expression andreceptors in the pathogenesis of CHF, our discovery of a protein synthesis in the human myocardium, with CC- and‘chemokine network’ in the human heart, located to the CXC-chemokines, and chemokine receptors located to thecardiomyocytes, suggests that locally produced cardiomyocytes. The observation that the end-stage failingchemokines may directly exert an effect upon the myocar- heart expresses decreased levels of chemokines and in-dium, possibly working through an autocrine /paracrine creased levels of chemokine receptors may represent onemechanism. of several different maladaptive mechanisms responsible

Our comparison between failing and nonfailing hearts for progression of advanced heart failure. The ‘chemokinesuggests downregulation of chemokine expression and system’ may represent a previously unrecognised patho-upregulation of chemokine receptor expression in the end- genic factor in the development of CHF and may introducestage failing myocardium. While MCP-1 and other a new family of mediators with potentially importantchemokines may be upregulated in an earlier stage of heart effects on the myocardium.failure [35], our findings suggest that these mediators aredownregulated in the myocardium of patients with end-stage heart failure. In fact, a dynamic regulation of Acknowledgementsinflammatory mediators in the myocardium may well exist,with upregulation in acute heart failure and in the ‘hy- This work was supported by the Norwegian Council ofpertrophic phase’ of CHF, and with downregulation in Cardiovascular Disease, Research Council of Norway anddecompensated CHF, characterised by a dilated left ventri- Medinnova Foundation. We thank Anne Brunsvig forcle [36]. excellent technical assistance.

Caution is needed when comparing chemokine expres-sion in failing and nonfailing myocardium. Individuals‘donating’ control tissue may have been exposed to Referencesstressful stimuli and hypoxia before death. Chemokinegenes are transcribed early upon activation [37], and

[1] Finkel MS, Oddis CV, Jacob TD et al. Negative inotropic effects ofhypoxia has been shown to be a potent inducer of cytokines on the heart mediated by nitric oxide. Sciencechemokine gene expression [38]. Although we found no 1992;257:387–389.

[2] Yokoyama T, Vaca L, Rossen RD et al. Cellular basis for theassociation between chemokine gene expression and is-negative inotropic effects of tumor necrosis factor-a in the mam-chemic time before donor tissue sampling, we cannotmalian heart. J Clin Invest 1993;92:2303–2312.exclude some transcription of these genes before explanta-

[3] Hirota H, Yoshida K, Kishimoto T, Taga T. Continuous activation oftion in the donor group. Furthermore, the interpretation of GP 130, a signal-transacting receptor component for interleukin-6chemokine protein expression is difficult taking into related cytokines, causes myocardial hypertrophy in mice. Proc Natl

Acad Sci USA 1995;92:4862–4866.consideration the complexity of tissue-specific regulation[4] Krown KA, Page MT, Nguyen C et al. Tumor necrosis factorof chemokine and chemokine receptor expression. Differ-

alpha-induced apoptosis in cardiac myocytes. J Clin Investences in the clearance of the chemokines by proteolysis1996;98:2854–2865.

could partly contribute the overall picture of differential [5] Bozkurt B, Kribbs SB, Clubb Jr. FJ et al. Pathophysiologicallytissue expression of chemokines [39]. Accordingly, further relevant concentrations of tumor necrosis factor-a promote progres-

sive left ventricular dysfunction and remodeling in rats. Circulationstudies are needed to examine these aspects of chemokine1998;97:1382–1391.regulation in myocardial tissue.

[6] Taub DD, Oppenheim JJ. Chemokines, inflammation and theIn the failing, but not in the nonfailing, myocardium, weimmune system. Ther Immunol 1994;1:229–246.

found differences in the expression of chemokines between [7] Bagglioni M, Dewald B, Moser B. Interleukin-8 and relateddifferent chambers of the heart. In particular, the MCP-1 chemotactic cytokines: CXC and CC-chemokines. Adv Immunoland IL-8 mRNA levels were higher in all failing left atria 1994;55:97–179.

[8] Jaeschke H, Smith CW. Mechanisms of neutrophil-induced paren-compared to failing left ventricles. We hypothesise thatchymal cell injury. J Leukoc Biol 1997;61:647–653.these findings may reflect differences between left atria

[9] Waltz A, Meloni F, Clark-Lewis I, von Tscharner V, Baggiolini M.and ventricles with regard to hemodynamic burden and 21[Ca ] changes and respiratory burst in human neutrophils andiischemic exposure. Thus, while left ventricle in these monocytes induced by NAP-1/ Interleukin-8, NAP-2 and gro /patients represents end-stage myocardium with downregu- MGSA. J Leukoc Biol 1991;50:279–286.

[10] Boring L, Gosling J, Clearl M, Charo IF. Decreased lesion forma-lated MCP-1 expression, left atria may possibly representtion in CCR22 /2 mice reveals a role for chemokines in thean ‘earlier stage’ of heart failure. Whatever the reasons, ourinitiation of atherosclerosis. Nature 1998;394:894–897.

observations clearly underscore the importance of studying [11] Kolattukudy PE, Quach T, Bergese S et al. Myocarditis induced bydifferent chambers individually when comparing failing targeted expression of the MCP-1 gene in murine cardiac muscle.and nonfailing myocardium. Am J Pathol 1998;152:101–111.


Dow

nloaded from


[12] Kukielka GL, Youker KA, Michael LH et al. Role of early [26] Unemori EN, Marsters JC, Bauer EA, Amento EP. The product ofreperfusion in the induction of adhesion molecules and cytokines in the GRO gene regulates collagen turnover by inducing tissuepreviously ischemic myocardium. Mol Cell Biochem 1995;147:5– inhibitor metalloproteinases. Arthritis Rheum 1991;34:S117.12. [27] Koch AE, Polverini PJ, Kunkel SL et al. Interleukin-8 as a

¨[13] Aukrust P, Ueland T, Muller F et al. Elevated circulating levels of macrophage-derived mediator of angiogenesis. ScienceC-C chemokines in patients with congestive heart failure. Circula- 1992;258:1798–1801.tion 1998;97:1136–1143. [28] Hesselgesser J, Taub D, Baskar P et al. Neuronal apoptosis induced

˚[14] Damas JK, Gullestad L, Ueland T et al. CXC-chemokines, a new by HIV-1 gp120 and the chemokine SDF-1 alpha is mediated by thegroup of cytokines in congestive heart failure—possible role of chemokine receptor CXCR4. Curr Biol 1998;8:595–598.platelets and monocytes. Cardiovasc Res 2000;45:428–436. [29] Tachibana K, Hirota S, Ilizasa H et al. The chemokine receptor

[15] Devaux B, Scholz D, Hirche A, Klovekorn WP, Schaper J. Upregu- CXCR4 is essential for vascularization of the gastrointestinal tract.lation of cell adhesion molecules and the presence of low grade Nature 1998;393:591–594.inflammation in human chronic heart failure. Eur Heart J [30] Aukrust P, Gullestad L, Frøland SS. Is apoptosis an important1997;18:470–479. pathogenic factor in cardiovascular disease? Eur J Clin Invest

[16] Luster AD. Chemokines—chemotactic cytokines that mediate in- 1999;29:369–371.flammation. N Engl J Med 1998;338:436–445. ¨[31] Berndt C, Mopps B, Angermuller S, Gierschik P, Krammer PH.

[17] Nagasawa T, Hirota S, Tachibana K et al. Defect of B-cell 1CXCR4 and CD4 mediate CD95-independent cell death in CD4 Tlymphopoiesis and bone-marrow myelopoiesis in mice lacking the cells. Proc Natl Acad Sci USA 1998;95:12556–12561.CXC chemokine PBSF/SDF-1. Nature 1996;382:635–638.

[32] Herbein G, Mahlknecht U, Batliwalla F et al. Apoptosis of CD81 T[18] Chomezynski P, Sacchi N. Single-step method of RNA isolation by

cells is mediated by macrophages through interaction of HIV gp120acid guanidinium thiocyanate–phenol–chloroform extraction. Anal

with chemokine receptor CXCR4. Nature 1998;395:189–194.Biochem 1987;162:156–159.

[33] Lu Z, Berson JF, Chen Y et al. Evolution of HIV-1 co-receptor usage[19] Barnes DA, Tse J, Kaufhold M et al. Polyclonal antibody directed

through interaction with distinct CCR5 and CXCR4 domains. Procagainst human RANTES ameliorates disease in the Lewis rat

Natl Acad Sci USA 1997;94:6426–6431.adjuvant-induced arthritis model. J Clin Invest 1998;101:2910–

[34] Barbaro G, Di Lorenzo G, Grisorio B, Barbarini G. Incidence of2919.dilated cardiomyopathy and detection of HIV in myocardial cells of[20] Lamkhioued B, Renzi PM, Abi-Younes S. Increased expression ofHIV-positive patients. N Engl J Med 1998;339:1093–1099.eotaxin in bronchoalveolar lavage and airways of asthmatics contri-

[35] Shioi T, Matsumori A, Kihara Y et al. Increased expression ofbutes to the chemotaxis of eosinophils to the site of inflammation. Jinterleukin-1b and monocyte chemotactic and activating factor /Immunol 1997;159:4593–4601.monocyte chemoattractant protein-1 in hypertrophied and failing[21] Z’Graggen K, Walz A, Mazzucchelli L, Strieter RM, Mueller C. Theheart with pressure overload. Circ Res 1997;81:664–671.C-X-C chemokine ENA-78 is preferentially expressed in intestinal

[36] Ono K, Matsumori A, Shioi T, Furukawa Y, Sasayama S. Cytokineepithelium in inflammatory bowel disease. Gastroenterologygene expression after myocardial infarction in rat hearts: possible1997;113:808–816.implication in left ventricular remodeling. Circulation 1998;98:149–[22] Vonhof S, Brost B, Stille-Siegener M et al. Monocyte activation in156.congestive heart failure due to coronary artery disease and idiopathic

[37] Freter RR, Alberta JA, Lam KK, Stiles CD. A new platelet-deriveddilated cardiomyopathy. Int J Cardiol 1998;63:237–244.growth factor-regulated genomic element which binds serine /[23] Rollins BJ, Walz A, Baggiolini M. Recombinant MCP-1/JE inducesthreonine phosphoprotein mediates induction of the slow immedi-chemotaxis, calcium flux, and respiratory burst in human mono-ately-early gene MCP-1. Mol Cell Biol 1995;15:315–325.cytes. Blood 1991;78:1112–1116.

[38] Karakurum M, Shreeniwas R, Chen J et al. Hypoxic induction of[24] Buttke TM, Sandstrom PA. Oxidative stress as a mediator ofinterleukin-8 gene expression in human endothelial cells. J Clinapoptosis. Immunol Today 1994;15:7–10.Invest 1994;93:1564–1570.[25] Unemori EN, Amento EP, Bauer EA, Horuk R. Melanoma growth-

stimulatory activity /GRO decreases collagen expression by human [39] Devalaraja MN, Richmond A. Multiple chemotactic factors: finefibroblasts. J Biol Chem 1993;268:1338–1342. control or redundancy? Trends Pharmacol Sci 1999;20:151–156.


Dow

nloaded from

myocardial expression of cc- and cxc-chemokines and their receptors in human end-stage heart failure

Documents