and Roel A. de WegerJoyce van Kuik, Erica de Koning, Carla C. Baan, Nicolaas de Jonge, Frits H.J. Gmelig-Meyling Jorg van Loosdregt, Matthijs F.M. van Oosterhout, Annette H. Bruggink, Dick F. van Wichen,

ResponseHelper 1−Arteries With Cardiac Allograft Vasculopathy Is Indicative of a Memory T

The Chemokine and Chemokine Receptor Profile of Infiltrating Cells in the Wall of

Print ISSN: 0009-7322. Online ISSN: 1524-4539 Copyright © 2006 American Heart Association, Inc. All rights reserved.

is published by the American Heart Association, 7272 Greenville Avenue, Dallas, TX 75231Circulation doi: 10.1161/CIRCULATIONAHA.105.597526

2006;114:1599-1607; originally published online October 2, 2006;Circulation. 

http://circ.ahajournals.org/content/114/15/1599World Wide Web at:

The online version of this article, along with updated information and services, is located on the

  http://circ.ahajournals.org//subscriptions/

is online at: Circulation Information about subscribing to Subscriptions: 

http://www.lww.com/reprints Information about reprints can be found online at: Reprints:

  document. Permissions and Rights Question and Answer this process is available in the

click Request Permissions in the middle column of the Web page under Services. Further information aboutOffice. Once the online version of the published article for which permission is being requested is located,

can be obtained via RightsLink, a service of the Copyright Clearance Center, not the EditorialCirculationin Requests for permissions to reproduce figures, tables, or portions of articles originally publishedPermissions:

by guest on May 31, 2013http://circ.ahajournals.org/Downloaded from

The Chemokine and Chemokine Receptor Profile of InfiltratingCells in the Wall of Arteries With Cardiac Allograft Vasculopathy

Is Indicative of a Memory T–Helper 1 ResponseJorg van Loosdregt, MSc; Matthijs F.M. van Oosterhout, MD, PhD; Annette H. Bruggink, MSc;

Dick F. van Wichen, BSc; Joyce van Kuik, BSc; Erica de Koning, BSc; Carla C. Baan, PhD;Nicolaas de Jonge, MD, PhD; Frits H.J. Gmelig-Meyling, PhD; Roel A. de Weger, PhD

Background—Despite improvement in short-term patient survival after heart transplantation (HTx), long-term survivalrates have not improved much, mainly because of cardiac allograft vasculopathy (CAV). Cytokines and chemokines areconsidered to play an important role in CAV development.

Methods and Results—We focused on coronary arteries of HTx patients and made an inventory of the infiltrating cells andthe expression of cytokines as well as chemokines and chemokine receptors (C�CR) in the different layers of the vesselwall with CAV. Tissue slides were stained for a variety of cell markers (CD3, CD4, CD8, CD20, CD68, CD79a),chemokines (monokine induced by interferon [MIG], interferon-inducible protein 10 [IP-10], interferon-inducible Tcell-� chemoattractant [ITAC], RANTES [regulated on activation normal T cell expressed and secreted], andfractalkine), and chemokine receptors (CXCR3, CCR5, and CX3CR1). In reference coronary arteries (not transplanted),almost no infiltrating cells were found, and in transplanted hearts with CAV (HTx�CAV), a large number of T cellswere observed (CD4:CD8�2:1), mainly localized in the neointima and adventitia. Most of these T cells appeared to beactivated (human leukocyte antigen DR positive). Coronary arteries from transplanted hearts without CAV(HTx�CAV), HTx�CAV, and references were also analyzed for cytokine and C�CR mRNA expression with the useof quantitative polymerase chain reaction. Interferon-� was highly expressed in HTx�CAV compared with HTx�CAV.Interleukin-4 and interleukin-10 were expressed at the same level in both HTx groups and references. In HTx�CAV,all C�CR, but especially the T–helper 1 (TH1) C�CR, were more abundant than in the HTx�CAV and references.However, TH2 CCR4 expression did not differ significantly between both HTx groups.

Conclusions—In coronary arteries with CAV, most T cells are CD4� and express human leukocyte antigen DR. Theseactivated TH cells are mainly memory TH1 cells on the basis of their C�CR profile and cytokine expression.(Circulation. 2006;114:1599-1607.)

Key Words: arteriosclerosis � cardiovascular diseases � chemokines � heart failure � immunology� pathology � transplantation

Heart transplantation (HTx) is a last resort for patientswith end-stage heart failure. Short-term patient survival

of acute rejection has improved much over the last yearsowing to better immunosuppressive management, but long-term survival rates have not changed dramatically. This lackof progress is predominantly attributable to cardiac allograftvasculopathy (CAV). CAV is a concentric thickening of theblood vessel wall due to proliferation of smooth muscle cells(SMCs) in the intima layer (neointima) of the coronaryarteries.1,2 Consequently, the lumen of the arteries is nar-rowed, leading to untreatable cardiac ischemia.3,4 On average,the incidence of CAV shows an increase of 10% every year.5,6

Editorial p 1561Clinical Perspective p 1607

The pathogenesis of CAV is related to hyperlipidemia,cytomegalovirus infection, hypertension, ischemia, andreperfusion injury, and the intensity of acute rejection andthe development of CAV are related.3,4,7 Still, the exactorigin of CAV is not understood fully. Several studiesindicate that interferon-� (IFN-�) plays a central role in itsdevelopment.8,9 It may activate inflammatory cells but alsofibroblasts and SMCs to produce a variety of chemokines,eg, monokine induced by IFN (MIG), IFN-inducible T

Received October 25, 2005; revision received June 7, 2006; accepted June 21, 2006.From the Department of Pathology (J.v.L., M.F.M.v.O., A.H.B., D.F.v.W., J.v.K., E.d.K., R.A.d.W.), Heart Lung Center Utrecht (N.d.J.), and

Department of Immunology (F.H.J.G.-M.), University Medical Center Utrecht, Utrecht; and Department of Internal Medicine, Erasmus Medical Center,Rotterdam (C.C.B.), the Netherlands.

Correspondence to Roel de Weger, PhD, Molecular Pathology and Immunopathology, Department of Pathology (H04.312), University Medical CenterUtrecht, Heidelberglaan 100, PO Box 85500, 3508 GA Utrecht, The Netherlands. E-mail r.deweger@umcutrecht.nl

© 2006 American Heart Association, Inc.

Circulation is available at http://www.circulationaha.org DOI: 10.1161/CIRCULATIONAHA.105.597526

1599

Transplantation

by guest on May 31, 2013http://circ.ahajournals.org/Downloaded from

cell-� chemoattractant (ITAC), and IFN-inducible protein10 (IP-10).10,11 These chemokines and others such asRANTES (regulated on activation normal T cell expressedand secreted) and fractalkine are directly linked with CAVdevelopment in animal studies.12–14 Studies for the role ofchemokines in human CAV development are limited to 2immunohistochemical investigations15,16 and studies onhuman coronary arteries transplanted in severe combinedimmunodeficiency mice reconstituted with human periph-eral blood mononuclear cells (PBMCs).17,18 CAV developsin rodents as early as 14 days after transplantation,12 but itscourse in humans is relatively slow. Extrapolation ofanimal studies to the human situation seems treacherousand requires confirmation in human arteries. To obtain abetter insight into the relation between infiltrating cellsand CAV development, an inventory was made of infil-trating cells, cytokines, and chemokines and chemokinereceptors (C�CR) in coronary arteries obtained from HTxpatients at postmortem examination with the use of immu-nohistochemistry and quantitative polymerase chain reac-tion (PCR).

MethodsPatientsPatient and coronary artery characteristics are summarized in Table1. All patients were treated with a standard triple immunosuppressivetherapy of cyclosporin A, azathioprine, and steroids. All coronaryarteries were obtained at autopsy. Informed consent of all patientswas obtained before transplantation. The HTx patients were subdi-vided into a group that developed severe CAV (n�8) and a groupthat did not (n�7). The severity of CAV was determined by thedegree of intima thickening, amount of infiltrate, and degree oflumen occlusion. If the ratio [intima/(intima�lumen)] exceeded50%, CAV was diagnosed (measurements were performed at 3different positions in the vessel wall).

HistologyFor routine pathology, all coronary arteries were analyzed with theuse of hematoxylin-eosin (HE) or elastica van Gieson staining fordetection of the lamina elastica. The total number of infiltratingleukocytes per area was determined with the use of a grid thatcorresponds to 10 000 �m2. The mean amount of cells was deter-mined at 5 different places in the same area.

ImmunohistochemistryImmunohistochemistry was performed as previously described by deJonge et al.19 Briefly, 4-�m slides were cut from formalin-fixed,

TABLE 1. Patient Characteristics

Tissue

Patient Gender Age Months After HTx * † ‡ Cause of Death

Reference 1 Male 64 � � � � � P Hepatic cirrhosis

Reference 2 Male 39 � � � � � P Septic shock

Reference 3 Male 56 � � � � � P Massive bleeding

Reference 4 Female 35 � � � � A P Hypophosphatemia

Reference 5 Female 20 � � � � � F Cardiogenic shock

Reference 6§ Male 56 0.03 � A/� F Lung embolism

Reference 7 Male 57 0.03 � A F Sepsis

Reference 8§ Male 42 0.5 � A/� F Pump failure

Reference 9 Female 36 0.5 � � P Sepsis

HTx 1 Male 24 4 � � P Acute rejection

HTx 2 Female 17 5 � A F Sepsis

HTx 3 Male 59 5 � � F Acute rejection

HTx 4§ Male 64 12 � � P/F Acute rejection

HTx 5§ Male 56 12 � A/� P/F Pump failure

HTx 6§ Female 16 16 � � F CAV, ischemia

HTx 7§ Male 56 17 � � P CAV, heart failure

HTx 8 Female 62 18 � � F Sepsis

HTx 9§ Male 43 30 � � F Sepsis�acute rejection

HTx 10 Male 68 59 � A F Malignancy

HTx 11 Male 56 61 � � F Sepsis�acute rejection

HTx 12 Male 64 72 � � F CAV, myocardial infarct

HTx 13§ Male 60 75 � A/� F Malignancy

HTx 14 Female 58 84 � � P CAV, ischemia

HTx 15 Male 71 113 � A F Malignancy

The coronary artery characteristics varied as follows: *coronary arteries with (�) or without (�) CAV; †coronaryartery tissue with atherosclerotic plaques (A); ‡coronary arteries paraffin embedded and used for immunohistochem-istry (P) or frozen in liquid nitrogen and used for mRNA analysis (F).

§Multiple coronary arteries were used from this patient, either after different tissue preservation or from differentlocations.

1600 Circulation October 10, 2006

by guest on May 31, 2013http://circ.ahajournals.org/Downloaded from

paraffin-embedded coronary arteries. The tissue slides were depar-affinized and rehydrated by subsequent washing in xylene, 100%EtOH, 70% EtOH, and distilled water, all for 5 minutes. Subse-quently, endogenous peroxidase activity was blocked. Immunohis-tochemical procedures varied per antibody used; pretreatment, typeof antibody, and procedure are summarized in Table 2.After incu-bation, all slides were stained with diaminobenzidine. Standardcontrols (isotype controls and controls that leave out the primaryantibody) were tested on parallel tissue specimens. For control, allC�CR and marker antibodies were tested and titrated on lymphoidtissues19 (data not shown).

mRNA IsolationmRNA was extracted from frozen sections of whole coronaryarteries and isolated with the use of magnetic oligo(dT) beads(GenoPrep mRNA beads; GenoVision, Valencia, Calif), according tothe manufacturer’s instructions. The mRNA was stored at �80°C.

cDNA SynthesisOne microliter of oligo(dT)(15) primers (0.50 �g; Promega Corpo-ration, Phoenix, Ariz) and 1 �L of random primers (0.50 �g;Promega) were added to 18 �L mRNA derived from the mRNAisolation. This solution was heated in a closed Eppendorf tube for 5minutes at 70°C, and the tube was cooled down to room temperature.Then 16 �L ribonuclease-free water, 12 �L 5X RT buffer (Invitro-gen Corporation, Carlsbad, Calif), 6 �L 0.1 mol/L dithiothreitol, and4 �L deoxy nucleotide triphosphate were added. After vigorousmixing, 1 �L RNasin (Promega) was added and mixed again, andfinally 1 �L SuperScript RNase H-Reverse Transcriptase (Invitro-gen) was added and heated in a closed Eppendorf tube for 1 hour at42°C. The cDNA was stored at �20°C.

Real-Time Quantitative PCRTo study the expression levels of C�CR and cytokines, primer/probecombinations were custom-made by Applied Biosystems (FosterCity, Calif). Per well, 12.5 �L Taqman Universal Master Mix wasused, and 1.25 �L primer per probe, 6.25 �L Milli-Q, and 5 �LcDNA sample were added. The quantitative PCR reactions wereperformed by the Prism 7700 sequence detection system (Applied

Biosystems). Thermal cycling constituted a denaturation step at 95°Cfor 10 minutes, followed by 45 cycles of 95°C for 15 seconds and60°C for 60 seconds. All experiments were performed in duplicate;when the difference between duplicate experiments exceeded a meanthreshold cycle (Ct) value of 0.8 (2�SD of all duplicates), the datawere rejected and repeated.

To quantify the data, the comparative Ct method was used.Relative quantity was defined as 2���Ct, in which �Ct�Ct(target)�Ct (reference), ��Ct��Ct (sample)��Ct (calibrator). Asreference, we used the gene PBGD, a housekeeping gene stablyexpressed in heart tissue.19 The calibrator was a batch placentalmRNA used for normalization.

Laser Tissue MicrodissectionFrozen tissue sections of 10 �m were cut on PEN-foil–mounted glassslides (PALM Microlaser Technologies, Bernried, Germany). Thesections were air dried, stained with HistoGene Staining Solution(Arcturus, Mountain View, Calif), and rinsed with Milli Q water.From the coronary arteries, several layers were microdissected withthe use of the Robot-Microbeam (Palm Microlaser Technologies),and mRNA was isolated.

Statistical AnalysisQuantitative PCR statistical analysis was performed with the use ofthe Mann-Whitney test (Prism GraphPad Software, San Diego,Calif). Correlation analysis was performed with Spearman correla-tion coefficients with the use of Prism. P�0.05 was consideredstatistically significant.

The authors had full access to and take full responsibility for theintegrity of the data. All authors have read and agree to themanuscript as written.

ResultsCardiac Allograft VasculopathyCoronary arteries in normal individuals and in cardiac transplantrecipients without CAV (HTx�CAV) contain an intima definedas a cellular layer (between endothelium and lamina elasticainterna), mainly composed of SMCs (Figure 1a and 1b).

TABLE 2. Immunohistochemical Procedures

Antibody Pretreatment Primary Dilution Bridge Secondary Origin Primary Antibody

CD3 (rabbit) Citrate 1/200 Go�RaBIO Streptavidin PO Dako

CD4 (mouse) EDTA 1/100 Ho�MoBIO Streptavidin PO Novocastra

CD8 (mouse) Citrate 1/200 Ho�MoBIO Streptavidin PO Dako

CD20 (mouse) Citrate 1/200 Ho�MoBIO Streptavidin PO Dako

CD68 (mouse) Citrate 1/800 Ho�MoBIO Streptavidin PO Novocastra

HLA-DR (mouse) Citrate 1/400 Ho�MoBIO Streptavidin PO Dako

RANTES (goat) EDTA 1/100 Rb�GoPO Rabbit PV R&D Systems

CCR5 (goat) EDTA 1/100 Rb�GoPO Rabbit PV Capralogics

MIG (mouse) EDTA 1/100 � � � Mouse PV Peprotech

IP-10 (goat) EDTA 1/200 Rb�GoPO Rabbit PV Santa Cruz

ITAC (rabbit) EDTA 1/50 � � � Rabbit PV Peprotech

CXCR3 (mouse) EDTA 1/2000 � � � Mouse PV R&D Systems

Fractalkine (rabbit) EDTA 1/50 � � � Rabbit PV Santa Cruz

CX3CR1 (goat) EDTA 1/800 Rb�GoPO Rabbit PV Santa Cruz

The different bridges were biotinylated goat anti-rabbit IgG (Go�RaBIO; Vector, Burlingame, Calif), biotinylatedhorse anti-mouse IgG (Ho�MoBIO; Vector), or horseradish peroxidase–labeled rabbit anti-goat IgG (Rb�GoPO; Dako,Carpinteria, Calif). All cell markers used peroxidase-labeled (PO) streptavidin (Immonotech, Fullerton, Calif) as asecondary antibody. Other secondary antibodies used were rabbit or mouse powervision (PV; Klinipath, Duiven, theNetherlands). Origin of primary antibodies: R&D Systems, Minneapolis, Minn; Carpalogics, Hardwick, Mass; SantaCruz Biotechnology Inc, Santa Cruz, Calif; Peprotech, Rocky Hills, NJ; Dako Corp, Carpinteria, Calif; NovocastraLaboratories Ltd, Newcastle, UK. Controls were performed as previously described.19,20

van Loosdregt et al Th1 Expression in Cardiac Allograft Vasculopathy 1601

by guest on May 31, 2013http://circ.ahajournals.org/Downloaded from

For evaluation, the CAV vessels were divided into 6 areas(Figure 1e): endothelial layer (area 1), neointima (area 2),tunica media (area 3), adventitia (area 4), accumulations ofmononuclear cells (MNCs) in the neointima (area 5), andaccumulations of MNCs outside the vessel wall (area 6;Figures 1e and 2). Figures 1 and 2 clearly illustrate that theneointima in HTx�CAV is composed of 2 layers. One layer(area 2b) linked to the lamina elastica interna of the neointimais positive for anti–smooth muscle actin (Figure 1c). The

other layer, next to the endothelium (area 2a), contains noSMCs but is composed of loose connective tissue, which ispositive for anti–collagen type I (as is the total neointima;Figure 1d). Most infiltrating cells in the neointima werelocalized in area 2a, at the border of area 2a and 2b, or justbeneath the endothelium. Many infiltrating cells were alsoobserved in the adventitia. Signs of immigration of MNCswere observed at the site of the lumen of the coronary arteriesin between the endothelial cells (Figure 2e and 2f) but also inthe blood vessels in the adventitia (Figure 2e).

In some coronary arteries atherosclerosis was observed,which was defined as focal lesions, more asymmetrical thanin CAV, with or without a lipid core or calcification and withinterruption of the lamina elastica interna. The infiltratesassociated with atherosclerosis were predominantly located inthe shoulder of the lesion. These areas with atherosclerosisare described separately.

Immunohistochemistry

Cell Markers and Cell Composition of InfiltrateBecause reference and coronary arteries without CAV con-tained few if any infiltrating cells, these were not analyzed byimmunohistochemistry. For the evaluation of the infiltratingcells in HTx�CAV vessels, 8 paraffin-embedded coronaryarteries were studied. A comparison was made betweeninfiltrating cells in atherosclerotic plaques, infiltrating cells inthe myocardium of the same allografted hearts, extravascularaccumulations of MNCs (area 6), and infiltrating cells in theHTx�CAV coronary arteries (Table 3).

All immunohistochemical data are summarized in Table 3and partly illustrated in Figure 2. The majority of infiltratingcells were CD3�, and the CD4/CD8 ratio was �2. Thepercentage (indicated as percentage of total leukocytes) of Bcells (CD20) was low (5% to 15%) in HTx�CAV vessels,except for the MNCs surrounding the coronary arteries (area6); these contained about the same numbers of B and T cells.In the CAV areas 1 to 4, only low percentages of macro-phages (CD68) were present (8% to 15%). In addition, in theMNC accumulations in and outside the CAV vessels, onlysmall percentages of macrophages were observed. The num-ber of macrophages in atherosclerotic plaques was muchhigher (up to 50% and more) than elsewhere. Human leuko-cyte antigen DR (HLA-DR) expression, almost absent inreference arteries, was high in most layers of the coronaryarteries and heart tissue of the transplanted heart (Figure 2d).It was mainly expressed on infiltrating MNCs. The cellcompositions in layers 1 (endothelium), 2 (neointima), and 4(adventitia) were comparable (Table 3).

Chemokines and Chemokine Receptors on Infiltrating CellsThe C�CR expressions on MNCs (Table 3) in areas 1, 2, and4 were comparable (indicated as percentage of total leuko-cytes) for RANTES (37% to 44%), MIG (29% to 37%), IP-10(59% to 63%), ITAC (56% to 60%), fractalkine (59% to66%), CCR5 (58% to 68%), CXCR3 (73% to 78%), andCX3CR1 (24% to 26%). The expression in area 3 (tunicamedia) was relatively low, as were the cell numbers. In theother areas of the CAV vessels, the composition of the C�CRdiffered: MNC accumulations in the CAV vessels (area 5)

Figure 1. Cross section of the coronary artery wall. a,HE-stained coronary artery of a nontransplanted heart. b,�-Smooth muscle actin staining of a nontransplanted heart. c,Coronary artery with CAV stained for �-smooth muscle actin. d,Collagen type 1 staining on a coronary artery with CAV. e, Over-view of a coronary artery with CAV (HE stained). Different areasare indicated by numbers: 1, endothelium; 2a, loose �-smoothmuscle actin–negative part of the neointima; 2b, �-smooth mus-cle actin–positive part of the neointima; 3, tunica media; 4,adventitia; 5, accumulation of MNCs in the neointima; 6, accu-mulations of MNCs outside the vessel wall. The filled arrowsindicate the lamina elastica interna, and the open arrows desig-nate the lamina elastica externa.

1602 Circulation October 10, 2006

by guest on May 31, 2013http://circ.ahajournals.org/Downloaded from

showed a relatively low expression of RANTES, ITAC,CXCR3, and CX3CR1, and the mononuclear cell accumula-tions outside the vessels (area 6) showed relatively lowRANTES, MIG, and CX3CR1 expression. In both the myo-cardium of the transplanted heart and in atheroscleroticplaques, the occurrence of C�CR on infiltrating cells waslow overall compared with the neointima (Table 3).

Coronary arteries with severe atherosclerosis contained anumber of infiltrating cells, especially in the shoulders offatty plaques. These infiltrates consisted of T cells (CD4� aswell as CD8�), plus a majority of macrophages. The numberof infiltrating cells in atherosclerotic plaques was low com-pared with CAV coronary arteries. In atherosclerotic plaquesin HTx�CAV, the infiltrates showed an expression pattern inbetween that of the intima of HTx�CAV and the nontrans-planted atherosclerotic plaque, suggesting a mixture of bothCAV and atherosclerosis. These infiltrates were therefore notevaluated further.

Real-Time Quantitative Polymerase Chain ReactionData on the expression of cytokines IFN-�, interleukin-4(IL-4), interleukin-10 (IL-10), and transforming growthfactor-� (TGF-�) are shown in Figure 3a. IL-4 was weaklyexpressed, without significant differences between refer-ences, HTx�CAV, and HTx�CAV. In addition, the expres-sion of IL-10 in these 3 groups did not differ, although it

tended toward a higher expression in the HTx�CAV com-pared with the HTx�CAV group (P�0.19). By contrast, theIFN-� expression in HTx�CAV was significantly higher(P�0.02) than the expression in reference vessels andHTx�CAV. Additionally, expression of IFN-� was higher inHTx�CAV than in references. TGF-� expression in coronaryarteries after HTx was high compared with reference arteries,but coronary arteries after HTx with or without CAV did notdiffer significantly in this respect.

The expression of the chemokines RANTES (P�0.01),ITAC (P�0.002), IP-10 (P�0.001), and MIG (P�0.001) inHTx�CAV was significantly higher than in references (Fig-ure 3b). Fractalkine expression did not differ (P�0.73). InHTx�CAV, a significantly higher expression for RANTES(P�0.01), fractalkine (P�0.002), and ITAC (P�0.01) wasobserved compared with the HTx�CAV expression. Theexpression of MIG (P�0.41) and IP-10 (P�0.11) did notreach significance between the 2 HTx groups. All chemo-kines were more strongly expressed in HTx�CAV than in thereference vessels.

The expression of 4 chemokine receptors (CCR5, CCR4,CX3CR1, CXCR3) was only significantly different betweenthe references and HTx�CAV for CXCR3 (P�0.002) andCCR4 (P�0.01; Figure 3c). The difference observed forCXCR3 (P�0.055) between HTx�CAV and HTx�CAV

Figure 2. Coronary arteries with CAV. a,HE staining of 2 coronary arteries withCAV and surrounding accumulations ofmononuclear cells (area 6). b, CD20staining, coronary arteries almost devoidof CD20� B cells but strong staining inarea 6. c, CD3 staining, strong T-cellstaining for CD3 in neointima andadventitia but also in area 6. d, HLA-DRstaining, many HL-DR–positive MNCs inthe infiltrate localized in the neointima.e, Lymphocyte migration from capillariesin the adventitia (arrow). f, Lymphocytemigration through the endothelium ofthe coronary arteries (arrow). L indicateslumen of the arteries.

van Loosdregt et al Th1 Expression in Cardiac Allograft Vasculopathy 1603

by guest on May 31, 2013http://circ.ahajournals.org/Downloaded from

was nearly significant. A significantly higher expression forCCR5 (P�0.03) and CX3CR1 (P�0.001) was observed inthe HTx�CAV group compared with the HTx�CAV group.Although in Figure 3c the expression of CCR4 seems to belower in the HTx�CAV group than in the HTx�CAV group,the expression of CCR4 did not differ significantly betweenboth HTx groups (P�0.42). The chemokine receptor expres-sion was significantly higher for all receptors between theHTx�CAV and reference groups.

To exclude the possibility that the differences observed inC�CR expression were related to the time after HTx, theexpressions of all C�CR in all HTx coronary arteries (bothHTx�CAV and HTx�CAV) were plotted against time. InFigure 4 this is illustrated for CX3CR1 and CCR5; in none ofthe cases was a significant correlation observed. Likewise, nocorrelation was found between the C�CR expressions afterHTx and (acute) rejection grade (data not shown).

Laser Tissue MicrodissectionWith the use of laser tissue microdissection, 4 distinct layersof the coronary arteries with CAV were isolated: the neoin-tima (area 2a and area 2b separately), the tunica media (area3), and the adventitia (area 4). mRNA from these layers wasanalyzed for IFN-�, TGF-�, ITAC, MIG, CXCR3, CX3CR1,and CCR4. Without exception, all these genes were expressedin the neointima layer (area 2a) and the adventitia, whereasthe expression levels in the tunica media and the neointima(area 2b; containing predominantly SMCs) were very low ornot detectable (data not shown).

DiscussionCAV is a process that negatively affects survival after HTx.The concentric growth is generally believed to be caused bythe influx and growth of SMCs originating in the tunicamedia of the coronary arteries. In this report we show that thisneointima is composed of 2 layers: one adjacent to the lumenand composed of a loose connective tissue in which manyinfiltrating cells are present, and another next to the laminaelastica interna, which is composed of SMCs. A similarmorphology was described by Atkinson et al.21 Many infil-trating cells are located at the border of these 2 layers in theneointima. Clearly, the neointima of all human coronaryarteries obtained from nontransplanted hearts at autopsyalready contains a considerable number of SMCs. Thissuggests that in the process of neointima formation, asobserved in human CAV, the influx of SMCs is limited andthe increase of intima volume is mainly owing to formation ofconnective tissue (filled with MNCs). This type of intimaformation seems to differ from the more rapid neointimaformation in experimental animals, which mainly containsSMCs.4,22

Because T cells are important producers of cytokines andchemokines, we have focused on the role of T cells in the wallof coronary arteries. Most cells were present in the adventitiaand in the neointima. The tunica media contained only a smallnumber of MNCs. The cells seem to migrate from differentsites. We speculate that the cells in the adventitia enter fromthe capillary network in the adventitia, whereas the cells inthe neointima migrate directly into the arterial wall from the

TABLE 3. Immunohistochemical Data of Coronary Arteries With CAV

Area 1, % Area 2, % Area 3, % Area 4, % Area 5, % Area 6, % Heart, %Atherosclerotic

Plaque, %

Cell markers

CD3 56.0 64.2 41.1 57.7 72.2 53.0 56.5 40.6

CD4 43.3 51.0 34.8 47.6 41.8 39.9 26.7 21.5

CD8 20.9 23.6 17.6 25.6 12.2 18.1 20.8 20.1

CD4/CD8 ratio 2.0 2.1 1.9 1.8 3.4 2.2 1.2 1.1

CD20 11.6 11.9 6.8 5.9 16.3 47.2 6.0 1.0

CD68 14.1 14.4 8.2 9.7 2.1 2.9 13.2 54.9

HLA-DR 45.3 53.5 29.4 40.7 28.9 44.7 38.1 50.3

C�CR

RANTES 43.9 37.4 20.8 40.8 22.6 23.9 40.2 24.4

CCR5 58.1 67.5 44.0 59.2 52.5 53.9 28.4 1.0

MIG 34.2 37.3 17.9 29.3 43.1 16.7 9.3 1.0

IP-10 59.4 62.8 46.8 61.2 54.7 61.8 31.9 17.2

ITAC 55.7 56.2 29.1 60.4 27.2 46.4 36.9 24.3

CXCR3 72.8 74.4 59.5 77.6 49.7 67.7 49.0 33.0

Fractalkine 58.9 67.1 42.3 65.8 58.4 60.3 35.8 25.3

CX3CR1 25.8 24.3 10.6 26.4 6.3 13.2 8.1 12.0

No. of infiltrating cellsper 10 000 �m2 (n�8)

35.5 21.4 4.4 15.4 70.5 150.2 11.5 35.0

Mean percentages of cells positive for a specific antibody are shown. Coronary arteries with CAV were scored for 8 different areas:endothelium (area 1), neointima (area 2), tunica media (area 3), adventitia (area 4), accumulations of MNCs in the neointima (area5), accumulations of MNCs outside the coronary artery (area 6), and MNCs in between the cardiomyocytes (heart) and theatherosclerotic plaque.

1604 Circulation October 10, 2006

by guest on May 31, 2013http://circ.ahajournals.org/Downloaded from

lumen of the coronary artery, through the endothelial layer.However, the composition of both populations did not differin terms of lymphocyte subsets, chemokine production, andchemokine receptor expression (as determined by immuno-histochemistry and mRNA expression after laser microdis-section). This indicates that this population of cells in theblood encounters the same trigger for migration at the site ofthe adventitia and the neointima.

Most of the infiltrating MNC were T cells, at a CD4:CD8ratio of �2. Because not all HLA-DR expression (45% to55%) can be ascribed to macrophages (10% to 15%), at least

part of the T cells must be HLA-DR positive (see also themorphology; Figure 2d). Consequently, many T cells areHLA-DR positive and apparently in an activated state. Thenumbers of macrophages and B cells are relatively small,particularly when the numbers were compared with those inthe infiltrate in atherosclerotic plaques or in extranodularMNC accumulations surrounding the coronary arteries. Theformer consisted mainly of macrophages; the latter containedhigh numbers of B cells. Compared with the infiltrating cellsin the myocardium, there is a clear difference in the percent-ages of CD8� cells and C�CR expression. In brief, the

Figure 3. mRNA analysis by quantitativePCR. Mean and SE of mRNA analysis byquantitative PCR are depicted for thefollowing: cytokines: IL-10, IL-4, IFN-�,and TGF-� (a); chemokines: MIG, IP-10,ITAC, RANTES, and fractalkine (b); andchemokine receptors: CXCR3, CCR5,CX3CR1, and CCR4 (c). *Significant(P�0.05) difference compared with thereference group. †Significant differencebetween HTx�CAV and HTx�CAV.

Figure 4. mRNA expression correlatedwith time after HTx. Quantitative PCRanalysis of mRNA expression correlatedwith time of the chemokine receptorsCXCR3 (a) and CCR5 (b) is shown. Nocorrelation of expression was found withtime for CX3CR1 (P�0.08) and CCR5(P�0.10).

van Loosdregt et al Th1 Expression in Cardiac Allograft Vasculopathy 1605

by guest on May 31, 2013http://circ.ahajournals.org/Downloaded from

composition of the T-cell population in the neointima isunique compared with adjacent infiltrates.

Chemokines and cytokines play an important role in boththe process of acute heart allograft rejection23,24 and neointi-ma formation, as was shown in various experimental mod-els.11,25–27 Studies in human cardiac transplant recipients arescarce or fragmentary. In most human studies for CAV, heartbiopsies or PBMCs were analyzed.8,15,28 They have generallydemonstrated an upregulation of ITAC, IP-10, and CXCR3 inCAV patients compared with controls, whereas MIG andRANTES did not change. In 2 studies, the presence of somechemokines in human CAV has been documented with theuse of immunohistochemistry.15,16 These studies showed ahigher expression of RANTES and ITAC in coronary arterieswith CAV compared with references, but IP-10 and MIGwere completely absent.

In the present study, high numbers of infiltrating cellsexpressing cytokine receptors CCR5 and CXCR3 (50% to75%) in the intima and adventitia were observed by immu-nohistochemistry. CX3CR1 was expressed less prominentlybut still on a relatively high percentage of leukocytes (25%compared with other infiltrates at 10% to 15%). In addition,the chemokines MIG (35% to 40%), IP-10 (55% to 60%),ITAC (55% to 60%), and fractalkine (55% to 65%) wereexpressed on a considerable number of MNCs. MIG andCCR5 are implicated in T-cell attraction and activation inCAV,11,25–27 whereas CX3CR1 and IP-10 have been impli-cated in acute rejection in various animal studies.24,28 Thisindicates that within the neointima and the adventitia, anactive infiltrate of MNCs is present with a high chemoattrac-tant potential. Within the T-cell population, CCR5 andCXCR3 are mainly expressed on T–helper 1 (TH1) cells.28–32

CCR5 is also expressed on macrophages, but these are onlypresent in low numbers. CCR5 and CXCR3 are also ex-pressed on natural killer cells,33 but these could not bedetected in the CAV wall (data not shown). CX3CR1 can beexpressed by natural killer cells, dendritic cells, and macro-phages, but on T cells it is characteristic for memory cells.29

To further evaluate the infiltrate in the arterial wall,quantitative PCR was performed. These data underscore theimmunohistochemical data, a strong expression of the TH1chemokines ITAC, MIG, and fractalkine, and a less promi-nent expression of RANTES and IP-10. Interestingly, theexpression of all chemokines was stronger in the HTx�CAVvessels than in the HTx�CAV vessels. In addition, the TH1chemokine receptors (CXCR3, CCR5, and CX3CR1) arehighly expressed in the HTx�CAV group compared with theHTx�CAV group. Blockade of these receptors attenuateschronic rejection in experimental animals.25,26 Despite thelow number of leukocytes in HTx�CAV, there was asignificant expression of IL-10, TGF-�, MIG, and CCR4.This expression is most likely caused by macrophages,fibroblasts, or even SMCs.

Cytokines were not evaluated immunohistochemically be-cause the stainings were not unequivocal. However, thequantitative PCR data showed that the TH1 cytokine IFN-�was significantly elevated in the HTx�CAV group comparedwith the HTx�CAV group. IFN-� has been indicated invarious experimental and human studies as one of the most

important mediators of CAV induction.17,18 Burns et al30

performed similar techniques on human coronary arteriestransplanted in severe combined immunodeficiency micereconstituted with human PBMCs. They also described amajor involvement of TH1 cells producing IFN-�, whichwere localized mainly in the neointima or adventitia. TheTH2-linked cytokine IL-4 is only expressed at low levels inall groups, and the TH0/TH2-associated cytokine IL-10 isrelatively abundantly expressed in the HTx�CAV groupcompared with the nontransplanted and HTx�CAV group.Expression of TGF-� was significantly higher in all HTxcoronary arteries compared with references, but there was nosignificant difference between the HTx�CAV andHTx�CAV groups.

In summary, coronary arteries with CAV contain anabundant T-cell infiltrate. Most T cells are CD4� and expressHLA-DR. These activated T lymphocytes appear to be TH1cells on the basis of their C�CR profile and cytokineexpression (IFN-� versus IL-4, CCR4); most likely they arememory cells (CX3CR1�).29,31–33 A TH1 response is oftenassociated with a strong inflammatory (cytotoxic) response.Although low numbers of TH2 cells are present, this studysuggests that the activated TH1 cells in the intima are able toinduce a matrix proliferative response, perhaps by activatingmacrophages (TGF-� production), which leads to an increaseof the neointima.32,33

DisclosuresNone.

References1. Segovia J. Update on cardiac allograft vasculopathy. Curr Opin Organ

Transplant. 2002;7:240–251.2. Weis M, von Scheidt W. Coronary artery disease in the transplanted heart.

Annu Rev Med. 2000;51:81–100.3. Weis M, von Scheidt W. Cardiac allograft vasculopathy: a review. Cir-

culation. 1997;96:2069–2077.4. Waller J, Brook NR, Nicholson ML. Cardiac allograft vasculopathy:

current concepts and treatment. Transpl Int. 2003;16:367–375.5. Costanzo MR, Naftel DC, Pritzker MR, Heilman JK 3rd, Boehmer JP,

Brozena SC, Dec GW, Ventura HO, Kirklin JK, Bourge RC, Miller LW.Heart transplant coronary artery disease detected by coronary angiog-raphy: a multiinstitutional study of preoperative donor and recipient riskfactors: Cardiac Transplant Research Database. J Heart Lung Transplant.1998;17:744–753.

6. Julius BK, Attenhofer Jost CH, Sutsch G, Brunner HP, Kuenzli A, VogtPR, Turina M, Hess OM, Kiowski W. Incidence, progression and func-tional significance of cardiac allograft vasculopathy after heart transplan-tation. Transplantation. 2000;69:847–853.

7. Yamani MH, Yousufuddin M, Starling RC, Tuzco M, Ratliff NB, CookDJ, Abdo A, Crowe T, Hobbs R, Rincon G, Bott-Silverman C, McCarthyPM, Young JB. Does acute cellular rejection correlate with cardiacallograft vasculopathy? J Heart Lung Transplant. 2004;23:272–276.

8. Zhao DX, Miller GG, Luster AD, Mitchell RN, Libby P. Differentialexpression of the IFN-gamma-inducible CXCR3-binding chemokines,IFN-inducible protein 10, monokine induced by IFN, and IFN-inducibleT cell alpha chemoattractant in human cardiac allografts: association withcardiac allograft vasculopathy and acute rejection. J Immunol. 2002;169:1556–1560.

9. Nagano H, Libby P, Taylor MK, Hasegawa S, Stinn JL, Becker G, TilneyNL, Mitchell RN. Coronary arteriosclerosis after T-cell-mediated injuryin transplanted mouse hearts: role of interferon-gamma. Am J Pathol.1998;152:1187–1197.

10. Shimizu K, Mitchell RN. Chemokine-mediated recruitment of inflam-matory and smooth muscle cells in transplant-associated arteriosclerosis.Curr Opin Organ Transplant. 2003;8:55–63.

1606 Circulation October 10, 2006

by guest on May 31, 2013http://circ.ahajournals.org/Downloaded from

11. Yun JJ, Fischbein MP, Whiting D, Irie Y, Fishbein MC, Burdick MD,Belperio J, Strieter RM, Laks H, Berliner JA, Ardehali A. The role ofMIG/CXCL9 in cardiac allograft vasculopathy. Am J Pathol. 2002;161:1307–1313.

12. Yun JJ, Fischbein MP, Laks H, Fishbein MC, Espejo ML, Ebrahimi K,Irie Y, Berliner J, Ardehali A. Early and late chemokine productioncorrelates with cellular recruitment in cardiac allograft vasculopathy.Transplantation. 2001;69:2515–2524.

13. Yun JJ, Fischbein MP, Laks H, Irie Y, Espejo ML, Fishbein MC, BerlinerJA, Ardehali A. RANTES production during development of cardiacallograft vasculopathy. Transplantation. 2001;71:1649–1656.

14. Umehara H, Bloom E, Okazaki T, Domae N, Imai T. Fractalkine andvascular injury. Trends Immunol. 2001;22:602–607.

15. Kao J, Kobashigawa J, Fishbein MC, MacLellan WR, Burdick MD,Belperio JA, Strieter RM. Elevated serum levels of the CXCR3 che-mokine ITAC are associated with the development of transplant coronaryartery disease. Circulation. 2003;107:1958–1961.

16. Pattison JM, Nelson PJ, Huie P, Sibley RK, Krensky AM. RANTESchemokine expression in transplant-associated accelerated atherosclero-sis. J Heart Lung Transplant. 1996;15:1194–1199.

17. Wang Y, Burns WR, Tang PC, Yi T, Schrechner JS, Zerwes HG, SessaWC, Lorber MI, Pober JS, Tellides G. Interferon-gamma plays a nonre-dundant role in mediating T cell-dependent outward vascular remodelingof allogeneic human coronary arteries. FASEB J. 2004;18:606–608.

18. Koh KP, Wang Y, Yi T, Shiao SL, Lorber MI, Sessa WC, Tellides G,Pober JS. T cell-mediated vascular dysfunction of human allograftsresults from IFN-gamma dysregulation of NO synthase. J Clin Invest.2004;114:846–856.

19. de Jonge N, van Wichen DF, van Kuik J, Kirkels H, Lahpor JR, Gmelig-Meyling FH, van der Tweel JG, de Weger RA. Cardiomyocyte death inpatients with end-stage heart failure before and after support with a leftventricular assist device: low incidence of apoptosis despite ubiquitousmediators. J Heart Lung Transplant. 2003;22:1028–1036.

20. van Hoffen E, van Wichen DF, Leemans JC, Broekhuizen RA, BrugginkAH, De Boer M, De Jonge N, Kirkels H, Slootweg PJ, Gmelig-MeylingFH, De Weger RA. T cell apoptosis in human heart allografts: associationwith lack of co-stimulation? Am J Pathol. 1998;153:1813–1824.

21. Atkinson C, Horsley J, Rhind-Tutt S, Charman S, Phillpotts CJ, WallworkJ, Goddard MJ. Neointimal smooth muscle cells in human cardiacallograft coronary artery vasculopathy are of donor origin. J Heart LungTransplant. 2004;23:427–435.

22. Atkinson C, Southwood M, Pitman R, Phillpotts C, Wallwork J, GoddardM. Angiogenesis occurs within the intimal proliferation that characterizes

transplant coronary artery vasculopathy. J Heart Lung Transplant. 2005;24:551–558.

23. Hancock WW, Gao W, Csizmadia V, Faia KL, Shemmeri N, Luster AD.Donor-derived IP-10 initiates development of acute allograft rejection. JExp Med. 2001;193:975–980.

24. Whiting D, Hsieh G, Yun JJ, Banerji A, Yao W, Fishbein MC, BelperioJ, Strieter RM, Bonavida B, Ardehali A. Chemokine monokine inducedby IFN-gamma/CXC chemokine ligand 9 stimulates T lymphocyte pro-liferation and effector cytokine production. J Immunol. 2004;172:7417–7424.

25. Gao W, Faia KL, Csizmadia V, Smiley ST, Soler D, King JA, Danoff TN,Hancock WW. Beneficial effects of targeting CCR5 in allograftrecipients. Transplantation. 2001;72:1199–1205.

26. Yun JJ, Whiting D, Fischbein MP, Banerji A, Irie Y, Stein D, FishbeinMC, Proudfoot AE, Laks H, Berliner JA, Ardehali A. Combined blockadeof the chemokine receptors CCR1 and CCR5 attenuates chronic rejection.Circulation. 2004;109:932–937.

27. Horiguchi K, Kitagawa-Sakakida S, Sawa Y, Li ZZ, Fukashima N,Shirakura R, Matsuda H. Selective chemokine and receptor geneexpressions in allografts that develop transplant vasculopathy. J HeartLung Transplant. 2002;21:1090–1100.

28. Melter M, Exeni A, Reinders MEJ, Fang JC, McMahon G, Ganz P,Hancock WW, Briscoe DM. Expression of the chemokine receptorCXCR3 and its ligand IP-10 during human cardiac allograft rejection.Circulation. 2001;104:2558–2564.

29. Fraticelli P, Sironi M, Bianchi G, D’Ambrosio D, Albanesi C,Stoppacciaro A, Chieppa M, Allavena P, Ruco L, Girolomoni G, Sini-gaglia F, Vecchi A, Mantovani A. Fractalkine (CX3CL1) as an amplifi-cation circuit of polarized Th1 responses. J Clin Invest. 2001;107:1173–1181.

30. Burns WR, Wang Y, Tang PCY, Ranjbaran H, Iakimov A, Kim J, CuffyM, Bai Y, Pober JS, Tellides G. Recruitment of CXCR3� and CCR5�T cells and production of interferon-�-inducible chemokines in rejectinghuman arteries. Am J Transplant. 2005;5:1226–126.

31. Loetscher M, Gerber B, Loetscher P, Jones SA, Piali L, Clark-Lewis I,Baggiolini M, Moser B. Chemokine receptor specific for IP10 and MIG:structure, function, and expression in activated T-lymphocytes. J ExpMed. 1996;184:963–996.

32. Sallusto F, Lanzavecchia A, Mackay CR. Chemokines and chemokinereceptors in T-cell priming and Th1/Th2-mediated responses. ImmunolToday. 1998;19:568–574.

33. Mantovani A, Sica A, Sozzani S, Allavena P, Vecchi A, Locati M. Thechemokine system in diverse forms of macrophage activation and polar-ization. Trends Immunol. 2004;25:677–686.

CLINICAL PERSPECTIVECardiac allograft vasculopathy is an untreatable concentric enlargement of the intima of coronary arteries in the donor heartafter heart transplantation. This intima enlargement (neointima formation) causes ischemia and finally failure of thetransplanted heart. In contrast to the current opinion on cardiac allograft vasculopathy or chronic rejection, this study showsthat this intima enlargement is not caused by smooth muscle cell infiltration and proliferation but is merely a volumeincrease caused by extracellular matrix formation (collagen I and fibroblasts). This luminal layer of the neointima is in mostcases populated by a large number of T cells. These T cells have an activated helper 1 phenotype and are most likely eitherdirectly or indirectly via macrophages responsible for the induction of the intima proliferation. These data can lead to amore specific form of therapeutic approach of chronic rejection, which is an increasingly difficult problem in long-termsurvival of transplanted organs. The medication should therefore be directed against matrix formation rather than againstsmooth muscle cell infiltration and proliferation. Possible targets for this therapeutic approach may be the cytokineproduction of the T cells (eg, inhibition of growth factors like transforming growth factor-�) or the infiltration of these Tcells in the intima. This may explain the positive effect of medication with immunosuppressive drugs like everolimus andsirolimus.

van Loosdregt et al Th1 Expression in Cardiac Allograft Vasculopathy 1607

by guest on May 31, 2013http://circ.ahajournals.org/Downloaded from