Molecules Induces Increased Expression of Adhesion Endothelial ...

The Journal of Immunology

Persistent Infection of Human Microvascular Endothelial Cells by Coxsackie B Viruses Induces Increased Expression of Adhesion Molecules1 Maria M. Zanone,2* Enrica Favaro,* Pier G. Conaldi,† James Greening,‡ Antonella Bottelli,† Paolo Cavallo Perin,* Nigel J. Klein,‡ Mark Peakman,§ and Giovanni Camussi* Numerous studies indicate that enteroviruses, such as the Coxsackievirus (CV) group, are linked to autoimmune diseases. Virus tropism and tissue access are modulated by vascular endothelial cells (ECs), mainly at the level of the microvasculature. Data on the permissiveness of ECs to CV are, however, scanty and derived from studies on large vessel ECs. To examine the susceptibility of microvascular ECs to infection of group B CV (CVB), human dermal microvascular ECs (HMEC-1) were infected with three CVB strains, and the immunological phenotype of the infected cells was analyzed. All CVB persistently infected the EC cultures without producing overt cytopathic effects. Infected ECs retained endothelial characteristics. Release of infectious particles in cell supernatants persisted for up to 3 mo of culture. Infection up-regulated expression of the adhesion molecules ICAM-1 and VCAM-1, with the highest values detected during the first 30 days of infection (p < 0.05 vs uninfected HMEC-1). CVB infection increased production of the proinflammatory cytokines, IL-6, IL-8, and TNF-␣, which may account for the enhanced expression of adhesion molecules. Parallel infection of macrovascular HUVEC had less evident effects on induction of ICAM-1 and did not significantly increase expression of VCAM-1. Moreover, mononuclear cell adhesion to CVB-infected HMEC-1 monolayers was increased, compared with uninfected monolayers. These results provide evidence that small vessel ECs can harbor a persistent viral infection, resulting in quantitative modification of adhesion molecule expression, which may contribute to the selective recruitment of subsets of leukocytes during inflammatory immune responses. Furthermore, our data confirm that the behavior against a viral challenge of ECs in large vessels and microvessels may differ. The Journal of Immunology, 2003, 171: 438 – 446.

E

nteroviruses, especially those of Coxsackievirus (CV)3 group B (CVB), have been linked to the induction of autoimmune diseases, such as myocarditis progressing to dilated cardiomyopathy (1, 2) and type 1 diabetes mellitus (3). The mechanisms by which viruses might induce autoimmunity remain controversial, and several hypotheses have been proposed to explain this link. These include molecular mimicry, bystander activation of autoreactive T cells by virus-specific T cells, and superantigenic activity of viral proteins activating T cells expressing a particular set of V␤ receptors (2, 4 –7). Viruses could also induce

*Department of Internal Medicine and Center of Experimental Medicine (CeRMS), University of Torino, Torino, Italy; †Department of Medicine and Public Health, University of Insubria, Varese, Italy; ‡Immunobiology Unit, Institute of Child Health, University College, London, United Kingdom; and §Department of Immunology, Guy’s, King’s, and St. Thomas’s School of Medicine, London, United Kingdom Received for publication September 27, 2002. Accepted for publication April 29, 2003. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. 1 This work was supported by a grant from Regione Piemonte (Italy), Ricerca Sanitaria Finalizzata (Grant 488; 1999, by the National Research Council (Italy)) target project biotechnology, and Istituto Superiore di Sanita` (Italy) target project AIDS, and by Diabetes U.K. (DUK). J.G. is a DUK Pediatric Research Fellow, and M.P. is a DUK Senior Clinical Research Fellow. 2 Address correspondence and reprint requests to Dr. Maria M. Zanone, I Divisione Universitaria di Medicina, Dipartimento di Medicina Interna, Corso Dogliotti 14, 10100 Torino, Italy. e-mail address: [email protected] 3 Abbreviations used in this paper: CV, Coxsackievirus; CAR, CV and adenovirus receptor; CVB, group B CV; DAF, decay-accelerating factor; DiOC18, 3,3⬘-dioctadecyloxacarbocyanine; EC, endothelial cell; HMEC-1, human dermal microvascular EC; MFI, mean fluorescence intensity; TCID50, 50% tissue culture infectious dose; vWF, von Willebrand’s factor.

Copyright © 2003 by The American Association of Immunologists, Inc.

expression of class II HLA molecules on the surface of infected cells, targeting them for immunological recognition. CVB infections are characterized by primary viral replication at the portal of entry, followed by viremia and secondary involvement of different organs. The distinct tropism of different virus strains is probably responsible for the variable clinicopathological manifestations of CVB infections. One of the major determinants of tissue tropism of CVB is the local expression of appropriate cellular receptors. Receptors important in CVB tropism include the recently identified CV and adenovirus receptor (CAR) and cell surface molecules such as integrin VLA-2, ICAM-1, and the decay-accelerating factor (DAF) (8). Viral variants and the nature of the infection, acute or persistent, are other determinants of the pathological phenotype that may be relevant to induction of chronic disease or autoimmunity. There is in vivo evidence of persistent infections of the heart (9 –11), skeletal muscle (12), CNS (13), as well as in vitro evidence of virus persistence in cultured cells, such as human kidney cells (14), vascular endothelial cells (ECs) (15), and pancreatic islets (16). In these cultured cells, persistent infection is associated with increased production of cytokines, such as platelet-derived growth factor A/B and TGF-␤1/2, TNF-␣, and IFN-␣, which, in turn, could be mediators of cytotoxic effects, disease progression, and maintenance of infection. Vascular ECs appear to be important in modulating virus tropism and tissue access in murine studies (17). Parenchymal cells of an organ are rarely in direct contact with the circulatory system. Therefore, viruses in the circulation must either circumvent or infect vascular ECs to reach secondary organs. Data on the susceptibility of human endothelium to CVB are scanty and derived from studies on large vessel ECs. HUVEC have been shown to be persistently 0022-1767/03/$02.00

The Journal of Immunology infected by different CVB strains (15, 18), but physiological and pathological events take place mainly at the level of the microvasculature. Recently, it has been shown that the behavior of ECs from large vessels and microvessels may differ (19 –24). ECs are important both in the control of leukocyte traffic and in the mediation of inflammation (25), through expression and secretion of an extensive array of key immunological accessory molecules and mediators. There is also evidence that ECs can contribute to the presentation of allo- and autoantigen to T cells, and expression of MHC class I and II molecules is essential for such interactions. In light of the potential of CVBs to infect ECs, and the possible consequences of such infection in relation to tissue inflammation, we investigated the susceptibility of a human microvascular EC line (HMEC-1) to infection by different CVB strains. The persistence of infection and the cellular and immunological phenotype of the infected cells, including adhesion and costimulatory and HLA molecules, were analyzed. Mononuclear cell adhesion and cytokine production were also evaluated to explore the biological effects of infection on the interplay between CVB, ECs, and immune cells.

Materials and Methods Viruses CVB strains were originally obtained from the American Type Culture Collection (Manassas, VA). Virus stocks were prepared in serum-free medium, as described (14), using human KB cells infected at a multiplicity of infection of 0.1. KB cultures displaying ⬎90% cytopathic effect after 24-h incubation at 37°C were disrupted by two freeze-thaw cycles. Cell debris was removed by centrifugation; cell-free supernatants were subjected to titer determination and stored at ⫺70°C. Virus titers were determined in quadruplicate by a micromethod using KB cells and expressed as 50% tissue culture infectious dose (TCID50) per milliliter. The endotoxin levels in the virus preparations were ⬍0.01 U/ml (Limulus assay).

Cell culture and infection HMEC-1, produced by transfection and immortalization of dermal microvascular ECs with the coding region for the simian virus 40 A gene product, large T Ag (26, 27), were cultured onto EC attachment factor (Sigma Aldrich, Milan, Italy)-coated tissue culture plates in MCDB131 (Life Technologies Italia, S. Giuliano Milanese, Italy) medium with 20% FCS, 10 mM L-glutamine, 12 ␮g/ml EC growth factor, 10 ng/ml epidermal growth factor, 1 ␮g/ml hydrocortisone (all products from Clonetics, San Diego, CA), and antibiotics. Cells were grown until confluent, washed twice with HBSS, and dispersed with trypsin/EDTA when subcultured (splitting cells 1:2 or 1:3). For infection, HMEC-1 were subcultured in a T25 flask at 50% confluence, and 2 days later were washed and infected with three different strains of CVB (CVB-3, CVB-4, and CVB-5) at multiplicity of infection of 5 in serum-free medium. After 2-h incubation at 37°C, cells were washed with HBSS and replenished with complete medium. A sample of supernatants was taken as the initial point for the time course of virus and cytokine production. CVB infection was monitored by evaluating the development of cytopathic effect and extracellular virus titers, as described. As human macrovascular ECs, immortalized HUVEC, HUVEC-CST, as previously described, were used (15). Parallel cultures of uninfected HMEC-1 and HUVEC-CST, cultured using the same medium and culture conditions of the infected counterparts, were generated for comparative experiments.

Phenotypic characterization and detection of surface molecules In time course experiments of cell staining, infected and uninfected cells were seeded onto a 24-well tissue culture plate at each subculture, and when at ⬃80 –90% confluence were grown in serum-free medium and harvested after 2 days. Cell monolayers were collected with nonenzymatic cell dissociation solution (Sigma-Aldrich), resuspended by gentle mechanical action, aliquoted, and washed immediately in PBS containing 5% FCS and 0.02% NaN3. For intracellular staining, cell aliquots were permeabilized by using 2 ml of FACS lysing solution (BD Biosciences, Erembodegem, Belgium) for 10 min, followed by 500 ␮l of FACS permeabilizing solution (BD Biosciences) for 30 min, according to the manufacturer’s instructions.

439 Cells were stained for 10 min at room temperature with saturating amounts of anti-human mAb against CD62E (final dilution 1/100), CD54 (1/10), CD106 (1/20) CD55 (1/200) (all from Serotec, Oxford, U.K.), CD146 FITC (clone P1H12, Chemicon, Temecula, CA), CD105 R-PE (Serotec), and CD40 FITC (Euroclone, Devon, U.K.). For intracellular staining, permeabilized cells were washed and incubated with rabbit anti-human von Willebrand’s factor (vWF) antiserum (1/200) (Sigma-Aldrich) and murine anti-enterovirus VP1 peptide mAb (1/40) (DAKO, Glostrup, Denmark) for 30 min. After washing, for unconjugated primary Abs, R-PE-conjugated F(ab⬘)2 of goat anti-mouse Ig or FITC-conjugated anti-rabbit Ig (DAKO) were added for 10 min at room temperature or 30 min for intracellular staining. After washing, cells were resuspended in wash buffer for analysis by flow cytometry using CellQuest software (BD Biosciences). Ten thousand events were collected, and results were expressed as mean fluorescence intensity (MFI) and percentage of positive gated events. For HLA class II molecule staining, cells were incubated for 30 min at 4°C with mouse anti-human Abs against HLA-DR (clone L243; BD PharMingen, San Diego, CA) and HLA-DQ (clone SPV-L3; Serotec) at baseline and after stimulation with 100 IU/ml IFN-␥ analyzing the 24-, 46-, and 72-h time points after stimulation. In all experiments, cells were also stained with the corresponding isotype control Abs (Serotec or BD PharMingen), and nonspecific staining was subtracted from the appropriate population. Each time point experiment is expressed as the mean of two separate flow cytometric analyses. To evaluate a cause-effect relationship between cytokine production and expression of adhesion molecules, infected and uninfected cells were treated with blocking polyclonal Abs against human IL-6, IL-8, and TNF-␣ (Sigma-Aldrich) (10 ␮g/ml), and the analysis of adhesion molecules by flow cytometry was performed after 4-h incubation. Infected cultures were also evaluated for expression of endothelial markers and viral capsid protein by immunofluorescence microscopy. Briefly, cells seeded onto a 24-well plate were cultured for 48 h, washed for 5 min three times with PBS containing 0.25% BSA, and fixed with 4% paraformaldehyde in PBS for 30 min at 4°C. After washing as above, cells were permeabilized with 1% Triton for 10 min, washed, and incubated with primary Abs against vWF for 30 min or VP1 for 1 h at room temperature. After further washing, cells were incubated with conjugated secondary Abs for 1 h. Nonpermeabilized cells were stained with R-PE-conjugated Abs against CD105 and FITC-conjugated Abs against CD31. After washing as above, cells were examined by inverted UV microscopy.

Detection of CD40 by Western blotting Cells were lysed at 4°C for 1 h in lysis buffer (50 mM Tris-HCl, pH 8.3, containing 1% Triton X-100, 10 ␮M PMSF, 10 ␮M/ml leupeptin, and 100 U/ml aprotinin). After centrifugation of the lysates at 15,000 ⫻ g, the supernatants of infected and uninfected HMEC-1 were quantitated for protein concentration by the Bradford technique. Protein content of the samples was normalized to 50 ␮g/sample in 20 ␮l by appropriate dilution in lysis buffer. Proteins were directly subjected to 8% SDS-PAGE and then transferred electrophoretically to nitrocellulose. The membranes were incubated with blocking solution (10% low-fat milk in 20 mmol Tris-HCl, pH 7.5, 500 mM NaCl, 0.1% Tween (TBST)) for 1 h, and subsequently with polyclonal rabbit Ab against human CD40 (Saint Cruz Biotechnology, Santa Cruz, CA) (diluted 1/200 in PBS, 0.5% Tween 20, 10% BSA) at a concentration of 500 ng/ml overnight at 4°C. After extensive washing with TBST, the blots were incubated with peroxidase-conjugated protein A (200 ng/ml) (diluted 1/4000) (Amersham, Little Chalfont, U.K.) for 1 h at room temperature. The enzyme was removed by washing as above, and blots were then incubated for 2 min with a chemiluminescence reagent (ECL; Amersham) and exposed to an autoradiography film for 1–5 min.

EC adhesion assay PBMCs were obtained by Ficoll-Hypaque centrifugation of heparinized blood from a healthy donor and washed twice in HBSS, and the pellet was resuspended in EC culture medium at a concentration of 2 ⫻ 106 cells/ml. To measure PBMC adhesion to CVB-3-infected and uninfected HMEC-1 cells, 2 ⫻ 106 PBMCs in 2 ml of culture medium were added to a 25-cm2 flask of confluent HMEC-1. Flasks were incubated for 1 h at 37°C. Nonadherent cells were removed by aspiration of the supernatant and an additional two washes. ECs and adherent PBMCs were dispersed by incubation with nonenzymatic cell dissociation solution, resuspended by gentle mechanical action, and washed in PBS containing 5% FCS and 0.02% NaN3. The cell pellet was resuspended in 100 ␮l of wash buffer and aliquoted for staining for T cells (with 10 ␮l of anti-human CD3 RPE mAb) and monocytes (anti-CD14 FITC) and isotype control for 30 min at 4°C.

440

CVB INFECTION OF HMECs

After washing, cells were analyzed by FACS using a two-parameter cytogram. In parallel experiments, PBMCs were labeled overnight at 37°C with 3 mM 3,3⬘-dioctadecyloxacarbocyanine (DiOC18) (Molecular Probes, Eugene, OR), a green fluorescent membrane stain, diluted 1/250 in culture medium (1 ml solution per 5 ⫻ 105 cells), and subsequently added to a 25-cm2 flask, as above. Adherent cells were counted by digital analysis (Windows MicroImage, version 3.4; CASTI Imaging, Venice, Italy) of images obtained using a video camera (Leica DC100, Deerfield, IL) and expressed as mean of cells counted in ten ⫻10 inverted microscope fields. To evaluate whether PBMC adhesion was mediated by ICAM-1 and VCAM-1, 30 min before addition of PBMCs, infected and uninfected cell monolayers were treated with blocking mAb against ICAM-1 and VCAM-1 (Serotec) at 20 ␮g/ml, alone or in combination.

Cytokine assays Cell culture supernatants of both infected and uninfected cultures were collected before each subculture and before each medium exchange, centrifuged at 1200 rpm for 5 min, aliquoted, and stored at ⫺80°C until assayed. IL-6, IL-8, and TNF-␣ were measured in duplicate by ELISA (R&D Systems Europe, Abingdon, U.K.), according to the manufacturer’s instructions. Briefly, in a quantitative sandwich enzyme immunoassay technique, microplate wells coated with mAb specific for the relevant cytokine are first incubated with supernatant samples and standards. After washing, the enzyme-linked polyclonal cytokine-specific Ab is added, followed by the substrate solution. Color development is stopped, and the color intensity read at the appropriate wavelength on a microplate reader (Bio-Rad, Hercules, CA). Detection limits of these assays were 3 pg/ml for IL-6, 31 pg/ml for IL-8, and 0.5 pg/ml for TNF-␣.

Statistical analysis Mean values of MFI and percentage of positive cells for surface molecules between uninfected and CVB-3-, CVB-4-, and CBV-5-infected cells were compared using the Mann-Whitney U test. Data were analyzed using the SPSS statistical package (SPSS, Chicago, IL), and p values ⬍0.05 were considered significant.

Results Acute and persistent infection of HMEC-1 cells and EC phenotype HMEC-1 cultures consistently allowed replication of CVB-3, CVB-4, and CVB-5. Virus production peaked 24 – 48 h after infection (10⫺5.5–10⫺5 TCID50/ml) and decreased thereafter with fluctuations. Despite virus replication, HMEC-1 did not show cytolytic changes, usually associated with CVB replication (28). Two-thirds of medium was replaced every 3– 4 days. Infected HMEC-1 reached confluence within 7–10 days, whereas uninfected HMEC-1 reached confluence in 4 –5 days, suggesting that infection reduced cell growth. As shown in Fig. 1, infected HMEC-1 assumed a more granular morphology, but maintained endothelial characteristics, as assessed by detection of vWF expression and EC-associated epitopes, i.e., transmembrane glycoprotein melanoma cell adhesion molecule (CD146) and endoglin (CD105) by FACS analysis and/or immunofluorescence (Fig. 1). Viability of both infected and uninfected HMEC-1 was 90 –96%, as detected by trypan blue exclusion. Infection with all the three CVB strains led to chronic release of infectious virus ⱖ10⫺2.5/3 TCID50/ml of virus up to 120 days. Replication of all CVB strains resumed after the replating of frozen infected cultures. Detection by flow cytometry of VP1 capsid protein during chronic infection indicated that replication occurred in 11.8 ⫾ 5% of cells (mean of six experiments) (Fig. 2). Similar results were obtained by immunofluorescent staining, showing a positive staining for VP1 in ⬃10% of cells per microscope field (Fig. 2), without differences between different strains. Similar results were obtained after CVB infection of HUVEC-CST (data not shown).

FIGURE 1. Microscopy and flow cytometric analysis of infected HMEC-1 at the third week of infection. Uninfected (A) and CVB-3-infected (B) HMEC-1 cell monolayers. Original magnification ⫻10. Infected cells appear more granular, but retain EC characteristics, i.e., positive cytoplasm staining for vWF, assessed by immunofluorescence (inset in B), and expression of CD105 (C) and CD146 (D), assessed by FACS analysis (dark line histogram). Thin line histograms represent the corresponding isotype control Abs.

EC surface molecule expression Acute infection of HMEC-1 with CVB-3, CVB-4, and CVB-5 upregulated expression of both ICAM-1 (CD54) and VCAM-1 (CD106) compared with uninfected cells. Such an increase persisted during the first 60 –70 days of infection, with no major differences among different CVB strains (Fig. 3). The highest values for ICAM-1 and VCAM-1 were detected during the first 30 – 40 days of infection. During the ⬃100-day time course analysis, MFI and percentage of positive cells for ICAM-1 and VCAM were consistently higher in infected cells than the mean values for each molecule in uninfected cells ( p ⬍ 0.05). Fig. 4 summarizes mean ⫾ SD MFI, and mean ⫾ SD percentage of positive cells for ICAM-1 and VCAM-1 of FACS analysis in the first 60 days of infection. In contrast, E-selectin (CD62E) expression was not detected on uninfected HMEC-1 and was not induced by CVB infection (data not shown). Expression of DAF (CD55) on infected cells was persistently decreased (mean MFI ⫾ SD of 34.5 ⫾ 2.5 for CVB-3, 22.1 ⫾ 1.1 for CVB-4, and 26.9 ⫾ 13 for CVB-5) compared with uninfected HMEC-1 (70.2 ⫾ 14) ( p ⬍ 0.05) (Fig. 5). Similarly, expression of the costimulatory molecule CD40 was consistently decreased in all three CVB strain infections throughout the time course of infection (mean MFI ⫾ SD of 9.2 ⫾ 6 for CVB-3, 9 ⫾ 4 for CVB-4, and 8.7 ⫾ 3 for CVB-5) compared with uninfected HMEC-1 (12.4 ⫾ 5), although such decrease did not reach statistical significance ( p ⫽ 0.07) (Fig. 5). These results were confirmed by Western blot analysis; lysates of CVB-infected HMEC-1 showed a reduced expression of CD40 compared with that of uninfected cell lysate (inset in Fig. 5). CVB infection per se did not induce surface expression of HLA class II molecules HLA-DR and DQ. Similarly, after IFN-␥, the magnitude of increase in expression of HLA class II molecules and the time course of up-regulation were similar compared with uninfected HMEC-1. Incubation of infected cells at the third to fourth week of infection with Abs against IL-6 and IL-8 induced a decrease of ICAM-1

The Journal of Immunology

441

FIGURE 2. Detection of VP1 protein in infected cells. Staining for VP1 capsid protein in uninfected (A, C) and chronically infected (⬃60 days) (B, D) HMEC-1 cells, assessed by immunofluorescence and FACS analysis (dark line histogram), respectively. By immunofluorescence microscopy, positive cells appear as bright spots on cell monolayer. By flow cytometry, 11.8 ⫾ 5% of cells (mean of six experiments) were positive for VP1. Dashed line histograms represent the isotype control Abs.

FIGURE 3. Time course analysis of adhesion molecule expression in infected ECs. Time course analysis of MFI of ICAM-1 (F, dark line) and VCAM-1 (E, dotted line) for HMEC CVB-3 (A), CVB-4 (B) CVB-5 (C), and HUVEC-CST CVB-3 (D). Virus titer during the time course is indicated as TCDI50 per milliliter at the top of each figure. The mean value ⫾ SD of MFI of ICAM-1 and VCAM-1 of uninfected counterparts (mean of several experiments) are shown as straight lines.

442

FIGURE 4. Adhesion molecule expression in uninfected and infected ECs. MFI (A) and percentage of positive cells (B) for ICAM-1 (filled) and VCAM-1 (open) of uninfected and infected HMEC-1 and HUVEC-CST during the first 60 –70 days of infection. Data are expressed as mean value ⫾ SD of 9 –10 experiments in duplicate. ⴱ, p ⬍ 0.05 vs uninfected cells.

expression (33.8% decrease of MFI and 11.3% decrease of percentage of positive cells for anti-IL-6; 47.9% decrease of MFI and 33.2% decrease of percentage of positive cells for anti-IL-8). This effect was increased by combined incubation with anti-IL-6 and anti-IL-8 (57% decrease of MFI and 40.5% decrease of percentage of positive cells) (Fig. 6). The effect of the blocking Abs was less evident on VCAM-1 expression. Blockade of TNF-␣ did not decrease expression of adhesion molecules on CVB-infected cells. In HUVEC-CST, CVB infection up-regulated expression of ICAM-1 only ( p ⬍ 0.05 vs uninfected cells), with a pattern similar to that of HMEC-1 (Fig. 3). In contrast, CVB infection did not increase VCAM-1 expression, although MFI values tended to fluctuate. Similarly to HMEC-1, CD55 (mean MFI ⫾ SD 4.2 ⫾ 2) and CD40 (21.9 ⫾ 3.4) expression decreased in infected cells compared with uninfected cultures (14.6 ⫾ 10 for CD55, and 36.4 ⫾ 30 for CD40), although without reaching statistical significance (Figs. 4 and 5). Adhesion assay FACS analysis of harvested adherent PBMCs incubated on HMEC-1 monolayer during the first 60 days of infection indicated

CVB INFECTION OF HMECs

FIGURE 5. CD40 and CD55 expression in uninfected and infected ECs. Mean value ⫾ SD of MFI (A) and percentage of positive cells (B) for CD40 (filled) and CD55 (open) of uninfected and infected HMEC-1 and HUVEC-CST during the first 60 –70 days of infection. Data are expressed as mean value ⫾ SD of four to eight experiments in duplicate. ⴱ, p ⬍ 0.05 vs uninfected cells. Inset, Representative Western blot analysis of CD40 expression on infected and uninfected cells. Lane 1, Uninfected HMEC-1; lane 2, HMEC-1 infected by CVB-5; lane 3, HMEC-1 infected by CVB-4. Three experiments were performed with similar results.

that CVB-3 infection increased adherence of T cells and monocytes by ⬃2-fold. Similarly, analysis of digital images of DiOC18labeled PBMCs indicated a mean of 31 adherent PBMCs per microscope field in infected HMEC-1 monolayer compared with a mean of 8 cells in uninfected cell monolayer (Figs. 7 and 8). mAbs directed against ICAM-1 and VCAM-1, in combination, decreased the number of adherent of T cells by 46.5% and monocytes by 24.8% (mean of three experiments). Fig. 8 shows mean values of adherent cells with and without blockade of adhesion molecules. Cytokine levels Endothelial-derived cytokines measured by ELISA on cell-free supernatants indicated that CVB infection induced an increased production of IL-6 and IL-8, compared with mean value of uninfected cells. The maximum increase of IL-6 ranged from 4 (in CVB-4 infection) to 13 times (in CVB-5 infection) ( p ⫽ 0.05 vs uninfected cells), and the increase of IL-8 from 3 (in CVB-4 infection) to 8 times (in CVB-3 infection) ( p ⫽ 0.07 vs uninfected cells)

The Journal of Immunology

443

FIGURE 7. Adhesion of PBMCs on cell monolayer. Inverted microscopy detection of adherent DiOC18-labeled PBMCs on uninfected (A) and infected (B) HMEC-1 monolayer. Adherent PBMCs appear as bright spots on cell monolayer. FACS analysis in the same experimental conditions of harvested adherent PBMCs on uninfected (C) and infected (D) cells. In a two-parameter cytogram, lymphocytes (CD3⫹, CD14⫺) are represented as signals in the upper left-hand quadrant, whereas monocytes (CD14⫹, CD3⫺) are represented as signals in the lower right-hand quadrant.

FIGURE 6. Representative FACS analysis of ICAM-1 expression on infected cells before (dark line) and after 4-h incubation with 10 ␮g/ml anti-IL-6 Abs (A, dotted line), or with 10 ␮g/ml anti-IL-8 Abs (B, dotted line), or with anti-IL-6 and anti-IL-8 in combination (C, dotted line). Three experiments were performed with similar results. Thin line histograms represent the corresponding isotype control Abs.

during the first 2 wk of the infection. Values tended to decrease thereafter. Fig. 9 shows mean values for repeated measurements of both cytokines during the first 2 wk of infection. Production of TNF-␣ in infected cells was variable. This cytokine was undetectable in supernatants of uninfected cells, while 3–5 pg/ml was detected in some supernatants of infected cells only during the first 2–3 wk of infection and none thereafter. Cell-free supernatants of HUVEC-CST showed a similar increase of IL-8 production in infected cells, while the increase of IL-6 was less consistent (Fig. 9). Similarly, TNF-␣ was detectable (from 2 to 10 pg/ml) during the first 4 wk of infection.

rather than simple cell lysis, as is becoming increasingly evident in clinical and experimental settings. These viruses can in fact persist for a considerable time in infected patients and cause chronic pathology or trigger immunopathological damage to infected and uninfected tissues (9 –13). In murine studies on CVB infectivity of different organs, the capacity of different mouse strains to respond with different forms of disease following CV infections indicates the essential role of host factors in developing specific diseases (17). These factors include not only the host immune response, but also types and characteristics of cells that become infected in different tissues. In line with this scenario, it is essential that, to gain access to secondary organs, viruses pass through the vascular endothelium by transcytosis or infection, or via infected circulating cells migrating into the target

Discussion In the present study, we addressed the question of whether a chronic infection of HMECs by CVB could be established in vitro, despite the fact that these enteroviruses are generally considered to be highly lytic viruses (28). Our results indicate that the establishment of a persistent CVB replication is a reproducible event with different CVB strains. All three CVB tested productively infected microvascular ECs for up to 3 mo without obvious cytolysis. In addition, we demonstrate that the infection induces modification of adhesion molecule expression that may influence the pattern of migration and extravasation of leukocytes in inflammation and immunity. These data add weight to the view that common CV infections are able to trigger complex pathophysiological processes,

FIGURE 8. Mean value ⫾ SD of adherent lymphocytes (CD3⫹) and monocytes (CD14⫹) on uninfected HMEC-1 (filled) and infected HMEC-1 without (open) and with (crosshatched) preincubation with mAb against ICAM-1 (20 ␮g/ml) and VCAM-1 (20 ␮g/ml). Data are expressed as mean value ⫾ SD of three experiments.

444

FIGURE 9. Endothelial cytokine production. Mean value ⫾ SD of IL-6 (A) and IL-8 (B) on cell-free supernatants of infected (filled) and uninfected (open) HMEC-1 and HUVEC-CST during the first 2 wk of infection. *, p ⫽ 0.05 and **, p ⫽ 0.07 vs uninfected cells.

tissues. Vascular ECs have therefore a major role in viral tropism and disease pathogenesis (17, 29). Our work was prompted by the consideration that, at present, only large vessel ECs derived from human umbilical veins have been studied and shown to constitute a reservoir for long-term CVB production (15). However, biological and functional differences exist between cells derived from large vs small vessels, and the vast majority of pathophysiological events involving ECs occurs at the level of microvasculature. These diversities include growth requirement in vitro (19, 30); PG secretory profile (20); immunologic phenotype (21); and types, amounts, and regulation of cell adhesion molecules (22, 23, 31–33). At a functional level, differential and sequential expression of adhesion molecules mediates trafficking of leukocytes to specific lymphoid and nonlymphoid tissues. The microvasculature has therefore a key role as the interface between the vascular space and organ parenchymas in physiological and pathological events. Previously, infectivity of CVB was assayed in vitro in murine microendothelial cells derived from different organs and only in short-term cultures (17). In the present study, we first describe persistent replication of CVB in HMECs. The detection of VP1 capsid protein and the release of infectious particles in the EC-free supernatants persisted for up to 3 mo without producing apparent cytopathic effect. Our results show that a small proportion of the

CVB INFECTION OF HMECs cells, ⬃10%, appeared to be involved in viral replication during chronic infection, suggesting that persistence is probably established through a mechanism of carrier-state culture, as proposed to explain CVB persistence in other cell types (15, 34). Chronic infection of ECs in vivo could provide better viral access to tissues underlying the endothelial layer and subsequent parenchymal cell infection, supporting the concept that microvascular ECs modulate virus tropism and pathogenicity. The mechanisms of CVB persistence are not clear. It is possible that the infected cells undergo cytolysis and release virions to infect more cells, thus maintaining a chronic infection of the culture without massive cell destruction. Alternatively, it could be hypothesized that the cells can cure themselves of viruses, e.g., by limiting production of cell host products required for viral replication, or by production of antiviral mediators. In our experimental conditions, cell cultures harbored a productive infection throughout the course of the experiments. Stability of the cell membrane could also be another important factor in the ability of infected cells to survive infection, without lysis. In previous studies, the distinct susceptibility of different cell types to long-term infection has been related to the production of IFNs (15, 35). It has also been suggested that the persistent infection of cultured HUVEC may be due to down-regulation of viral receptors in infected cells. However, in a recent study, expression of the specific CVB receptor, CAR, in these cells was not quantitatively altered by infection with CVB, but rather by culture confluence (18). Further studies are therefore needed to determine whether similar regulation occurs in ECs derived from other organs. In the present study, CAR expression was not examined. However, a decrease of surface expression of another CVB receptor, the DAF, was detected in infected cultures. DAF is per se insufficient to mediate cell infection, but it is known to mediate cell surface attachment and maintenance of virus in a conformationally unaltered state (36). Our study was also designed to test the hypothesis that viral infection could modify cell phenotype, and, in particular, the expression of adhesion molecules that may modulate EC interaction with leukocytes. Leukocyte extravasation involves a series of regulated adhesive interactions between leukocyte receptors and ligands on ECs. Regulation of endothelial and leukocyte adhesion molecules involves both quantitative changes in surface expression and qualitative changes in avidity; for the endothelial adhesion molecules, quantitative alterations in surface expression predominate (25). Our study shows that CVB infection is associated with enhanced expression of the major adhesion molecules ICAM-1 and VCAM-1. In the time course experiments, maximal expression was detected in the first 40 days of infection, fluctuating at lower levels thereafter. Interestingly, this increase did not correlate with titers of viruses released in supernatant, which showed fluctuations throughout the whole time course, after an initial peak. Confirming the heterogeneity of ECs from different sites, acute and persistent CVB infection of HUVEC-CST had less evident induction effects on ICAM-1 and did not increase expression of VCAM-1. This observation further highlights the differences between endothelium in large vessels vs the microendothelium, in which interactions with circulating leukocytes occur primarily. Several studies have demonstrated considerable variability of induction of adhesion molecules, VCAM-1 in particular, on ECs in different vascular beds and under inflammatory conditions or cytokine stimulation (24, 31, 32, 37). In the present study, increased production of proinflammatory cytokines, IL-6 and IL-8, indicates EC activation by virus. Such up-regulation has been documented after interaction of ECs with other infectious agents, contributing to their pathogenetic sequelae

The Journal of Immunology (38 – 40). In contrast, TNF-␣ expression was particularly low, suggesting a minor role of this cytokine in CVB infection of HMEC-1. However, it cannot be excluded that TNF-␣ produced by infected cells may synergize with the other cytokines in enhancing expression of adhesion proteins, extravasation of leukocytes, and differentiation and proliferation of lymphocytes (25, 32, 41). Cytokineblocking experiments only partially reduced expression of adhesion molecules, indicating that besides IL-6 and IL-8, other molecules, such as IFNs (15, 35) or IL-1 (42), may contribute to the up-regulation of ICAM-1 and VCAM-1 (41). The raised expression of adhesion molecules may have in vivo functional consequences, because ICAM-1 is involved in firm adhesion of all leukocyte subsets, while VCAM-1 participates in rolling as well as firm adhesion of lymphocytes and monocytes. Indeed, we show that adhesion of lymphocytes and monocytes is increased on persistently CVB-infected cell monolayers, and partially reduced by blocking Abs against the adhesion molecules, in a cause-effect relationship. Other integrins, as shown for VLA-5 in CMV infection (43), may be involved in this process. Despite only 10% of chronically infected ECs, the adhesion of leukocytes was increased by ⬃2-fold. This indicates that the enhanced adhesion was mediated by the paracrine effects of virus-induced mediators (41, 42) acting on uninfected cells, and that adhesion does not occur only on the chronically infected cells. Our data therefore provide an indication that EC infection could influence the evolution of a viral infection, as enhanced cellular recruitment may lead to enhanced and persistent tissue inflammation. Selective recruitment of leukocyte subsets could also influence the immune response to viral infection, in a complex interplay between adhesion molecules and cytokines. In line with this hypothesis, CVB-3-induced murine acute myocarditis has been shown to result in enhanced expression of ICAM-1 in myocardial cells, suggesting that the expression of ICAM-1 plays a critical role in the cell-mediated cytotoxicity in acute viral myocarditis and, potentially, in the subsequent dilated cardiomyopathy (44). In this scenario, ECs could also play a critical role in generating autoimmune reactions. The mechanisms by which viruses might act in inciting autoimmune mechanisms remain a controversial issue. It has been suggested that ECs might act as APCs. In the present study, we did not find evidence to support this hypothesis, as CVB-infected cells did not exhibit an enhanced expression of class II MHC or CD40 costimulatory molecules. In conclusion, our findings document that small vessel ECs can harbor a persistent viral infection, resulting in quantitative modification of adhesion molecule expression and cytokine production, with some differences compared with endothelium from large vessels. These factors, in turn, may contribute to the selective recruitment of subsets of leukocytes during inflammatory or immune response, and modulate pathological expression of virus-induced diseases.

References 1. Rose, N. R., A. Herskowitz, and D. A. Neumann. 1993. Autoimmunity in myocarditis: models and mechanisms. Clin. Immunol. Immunopathol. 2:95. 2. Luppi, P., W. A. Rudert, M. M. Zanone, G. Stassi, G. Trucco, D. Finegold, G. J. Boyle, P. del Nido, F. X. McGowan, and M. Trucco. 1998. Idiopathic dilated cardiomyopathy: a superantigen-driven autoimmune disease. Circulation 98:777. 3. Graves, P. M., J. M. Norris, M. A. Pallansch, I. C. Gerling, and M. Rewers. 1997. The role of enteroviral infections in the development of IDDM: limitations of current approaches. Diabetes 46:161. 4. Horwitz, M. S., L. M. Bradley, J. Harbertson, T. Krahl, J. Lee, and N. Sarvetnick. 1998. Diabetes induced by Coxsackie virus: initiation by bystander damage and not molecular mimicry. Nat. Med. 4:781. 5. Jun, H. S., and J. W. Yoon. 2001. The role of viruses in type I diabetes: two distinct cellular and molecular pathogenic mechanisms of virus-induced diabetes in animals. Diabetologia 44:271.

445 6. Conrad, B., E. Weidmann, G. Trucco, W. Rudert, R. Behboo, C. Ricordi, H. Rodriquez-Rilo, D. Finegold, and M. Trucco. 1994. Evidence for superantigen involvement in insulin-dependent diabetes mellitus aetiology. Nature 371:351. 7. Luppi, P., M. M. Zanone, H. Hyoty, W. A. Rudert, C. Haluszczak, A. M. Alexander, S. Bertera, D. Becker, and M. Trucco. 2000. Restricted TCR V␤ gene expression and enterovirus infection in type I diabetes: a pilot study. Diabetologia 43:1484. 8. Bergelson, J. M., J. A. Cunningham, G. Droguett, E. A. Kurt-Jones, A. Krithivas, J. S. Hong, M. S. Horwitz, R. L. Crowell, and R. W. Finberg. 1997. Isolation of a common receptor for Coxsackie B viruses and adenoviruses 2 and 5. Science 75:1320. 9. Muir, P., A. J. Tizley, T. A. H. English, F. Nicholson, M. Signey, and J. E. Banatvala. 1989. Chronic relapsing pericarditis and dilated cardiomyopathy: serological evidence of persistent enterovirus infection. Lancet 333:804. 10. Muir, P., F. Nicholson, S. J. Illavia, T. S. McNeil, J. F. Ajetunmobi, H. Dunn, W. G. Starkey, K. N. Reetoo, N. R. Cary, J. Parameshwar, and J. E. Banatvala. 1996. Serological and molecular evidence of enterovirus infection in patients with end-stage dilated cardiomyopathy. Heart 76:243. 11. Klinger, K., C. Hohenadl, A. Canu, M. Albrecht, M. Seemann, G. Mall, and R. Kandolf. 1992. Ongoing enterovirus-induced myocarditis is associated with persistent heart muscle infection: quantitative analysis of virus replication, tissue damage, and inflammation. Proc. Natl. Acad. Sci. USA 89:314. 12. Cunningham, L., N. E. Bowles, R. J. M. Lane, V. Dubowitz, and L. C. Archard. 1990. Persistence of enteroviral RNA in chronic fatigue syndrome is associated with abnormal production of equal amounts of positive and negative strands of enteroviral RNA. J. Gen. Virol. 71:1399. 13. Stone, R. 1994. Post-polio syndrome: remembrance of virus past. Science 264: 909. 14. Conaldi, P. G., L. Biancone, A. Bottelli, A. De Martino, G. Camussi, and A. Toniolo. 1997. Distinct pathogenetic effects of group B Coxsackieviruses on the human glomerular and tubular kidney cells. J. Virol. 71:9180. 15. Conaldi, P. G., C. Serra, A. Mossa, V. Falcone, F. Basolo, G. Camussi, A. Dolei, and A. Toniolo. 1997. Persistent infection of human vascular endothelial cells by group B Coxsackieviruses. J. Infect. Dis. 175:693. 16. Chehadeh, W., J. Kerr-Conte, F. Pattou, G. Alm, J. Lefebvre, P. Wattre´ , and D. Hober. 2000. Persistent infection of human pancreatic islets by Coxsackievirus B is associated with ␣ interferon synthesis in ␤ cells. J. Virol. 74:10153. 17. Huber, S. A., C. Haisch, and P. A. Lodge. 1990. Functional diversity in vascular endothelial cells: role in Coxsackievirus tropism. J. Virol. 64:4516. 18. Carson, S. D., J. T. Hobbs, S. M. Tracy, and N. M. Chapman. 1999. Expression of the Coxsackievirus and adenovirus receptor in cultured human umbilical vein endothelial cells: regulation in response to cell density. J. Virol. 73:7077. 19. Kubota, Y., H. Kleinman, G. R. Martin, and T. J. Lawley. 1988. Role of laminin and basement in the differentiation of human endothelial cells into capillary-like structures. J. Cell Biol. 107:1589. 20. Charo, I. F., S. Shak, M. A. Karasek, P. Davison, and I. Goldstein. 1986. Prostaglandin I2 is not a major metabolite of arachidonic acid in cultured endothelial cells from human foreskin microvessels. J. Clin. Invest. 74:914. 21. Swerlick, R. A., K. H. Lee, T. M. Wick, and T. J. Lawley. 1992. Human dermal microvascular endothelial cells but not human umbilical vein endothelial cells express CD36 in vivo and in vitro. J. Immunol. 148:78. 22. Swerlick, R. A., E. Garcia-Gonzalez, Y. Kubota, Y. Xu, and T. J. Lawley. 1991. Studies of the modulation of MHC antigen and cell adhesion molecule expression on human dermal microvascular endothelial cells. J. Invest. Dermatol. 97:190. 23. Fujimoto, T., and S. J. Singer. 1988. Immunochemical studies of endothelial cells in vivo. II. Chicken aortic and capillary endothelial cells exhibit different cellsurface distributions of the integrin complex. J. Histochem. Cytochem. 36:1309. 24. Lidington, E. A., D. L. Moyes, A. M. McCormack, and M. L. Rose. 1999. A comparison of primary endothelial cells and endothelial cell lines for studies of immune interactions. Transplant Immunol. 7:239. 25. Carlos, T. M., and J. M. Harlan. 1994. Leukocyte-endothelial adhesion molecules. Blood 84:2068. 26. Ades, W. E., F. J. Candal, R. A. Swerlick, V. G. George, S. Summers, D. C. Bosse, and T. J. Lawley. 1992. HMEC-1: establishment of an immortalized human microvascular endothelial cell line. J. Invest. Dermatol. 99:683. 27. Xu, Y., R. A. Swerlick, N. Sepp, D. Bosse, E. W. Ades, and T. J. Lawley. 1994. Characterization of expression and modulation of cell adhesion molecules on an immortalized human dermal microvascular endothelial cell line (HMEC-1). J. Invest. Dermatol. 102:833. 28. Schlegel, A., and K. Kirkegaard. 1995. Cell biology of enterovirus infection. In Human Enterovirus Infection. H. A. Rotbart, ed. ASM Press, Washington DC, p. 135. 29. Friedman, H. M., E. J. Macarak, R. R. MacGregor, J. A. Wolfe, and N. A. Kefalides. 1981. Virus infection of endothelial cells. J. Infect. Dis. 143:266. 30. Folkman, J., C. Haudenschild, and B. R. Zetter. 1979. Long term culture of capillary endothelial cells. Proc. Natl. Acad. Sci. USA 76:5217. 31. Petzelbauer, P., J. R. Bender, J. Wilson, and J. S. Pober. 1993. Heterogeneity of dermal microvascular endothelial cell antigen expression and cytokine responsiveness in situ and in cell culture. J. Immunol. 151:5062. 32. Swerlick, A. R., K. H. Lee, L. Li, N. T. Sepp, S. Wright Caughman, and T. J. Lawley. 1992. Regulation of vascular cell adhesion molecule 1 on human dermal microendothelial cells. J. Immunol. 149:698. 33. Lee, K. H., T. J. Lawley, Y. Xu, and R. A. Swerlick. 1992. VCAM-1, ELAM-1, and ICAM-1 independent adhesion of melanoma cells to cultured human dermal microvascular endothelial cells. J. Invest. Dermatol. 98:79.

446 34. Matteucci, D., A. Toniolo, P. G. Conaldi, F. Basolo, Z. Gori, and M. Bendinelli. 1985. Systemic lymphoid atrophy in Coxsackievirus B3-infected mice: effects of virus and immunopotentiating agents. J. Infect. Dis. 151:1100. 35. Heim, A., A. Canu, P. Kirschner, T. Simon, G. Mall, P. H. Hofschneider, and R. Kandolf. 1992. Synergistic interaction of interferon-␤ and interferon-␥ in Coxsackievirus B3 infected carrier cultures of human myocardial fibroblasts. J. Infect. Dis. 166:958. 36. Shafren, D. R. 1998. Viral cell entry induced by cross-linked decay-accelerating factor. J. Virol. 72:9407. 37. Haraldsen, G., D. Kvale, B. Lien, I. Farstad, and P. Brandtzaeg. 1996. Cytokineregulated expression of E-selectin, intercellular adhesion molecule-1 (ICAM), and vascular cell adhesion molecule-1 (VCAM) in human intestinal microvascular endothelial cells. J. Immunol. 156:2558. 38. Kaplanski, G., N. Teysseire, C. Farnarier, S. Kaplansky, J. Lissitzky, J. Durand, J. Soubeyrand, C. Dinarello, and P. Bongrand. 1995. IL-6 and IL-8 production from cultured human endothelial cells stimulated by infection with Rickettsia conorii via a cell-associated IL-1␣-dependent pathway. J. Clin. Invest. 96:2839.

CVB INFECTION OF HMECs 39. Filler, S. G., A. Pfunder, B. Spellberg, and J. Edwards. 1996. Candida albicans stimulates cytokine production and leukocyte adhesion molecule expression by endothelial cells. Infect. Immun. 64:2607. 40. Rose, F., S. Zeller, T. Chakraborty, E. Domann, T. Machleidt, M. Kronke, W. Seeger, F. Grimminger, and U. Sibelius. 2001. Human endothelial cell activation and mediator release in response to Listeria monocytogenes virulence factors. Infect. Immun. 69:897. 41. Mantovani, A., F. Bussolino, and E. Dejana. 1992. Cytokine regulation of endothelial cell function. FASEB J. 6:2591. 42. Dengler, T. J., M. J. Raftrey, M. Werle, R. Zimmermann, and G. Schonrich. 2000. Cytomegalovirus infection of vascular cells induces expression of pro-inflammatory adhesion molecules by paracrine action of secreted interleukin-1␤. Transplantation 69:1160. 43. Scholz, M., R. A. Blaheta, J. Vogel, H. W. Doerr, and J. Cinatl. 1999. Cytomegalovirus-induced transendothelial cell migration: a closer look at intercellular communication mechanisms. Intervirology 42:350. 44. Seko, Y., H. Matsuda, K. Kato, Y. Hashimoto, H. Yagita, K. Okumura, and Y. Yasaki. 1993. Expression of intercellular adhesion molecule-1 in murine hearts with acute myocarditis caused by coxsackie B3. J. Clin. Invest. 91:1327.