Adhesion of Memory Lymphocytes to Vascular Cell Adhesion Molecule-1–Transduced Human Vascular Endothelial Cells Under Simulated Physiological Flow Conditions In Vitro
The accumulation of mononuclear leukocytes is an early and persistent finding in atherosclerotic plaques. These mononuclear leukocytes are mostly monocyte-derived, but up to 20% are lymphocytes, predominantly CD4+ CD45RO+ (memory) T cells. To evaluate the potential of adenovirus vectors for studies of mononuclear leukocyte recruitment in vitro, we studied the effects of adenovirus vectors per se on human umbilical vein endothelial cells (HUVECs), a well-characterized in vitro model of vascular endothelium. A recombinant adenovirus containing the seven-domain isoform of rabbit vascular cell adhesion molecule-1 (rVCAM-1) was constructed and used to study lymphocyte adhesion under defined laminar flow conditions in transduced HUVEC monolayers. No increase in basal HUVEC surface expression of the inducible endothelial adhesion molecules and markers of activation, E-selectin and VCAM-1, was noted across a broad range of multiplicity of infection. A modest dose-dependent increase in surface intercellular adhesion molecule-1 expression was detectable by flow cytometry at an MOI of >30 plaque-forming units per cell. Under defined laminar flow from 1.5 to 0.5 dyne/cm2, the adenovirus vector carrying rVCAM-1 mediated stable adhesion of both a Jurkat T-cell line and primary human CD4+ CD45RO+ (memory) T cells. Monoclonal antibodies to α4-integrin or rVCAM-1 abolished adhesion, whereas monoclonal antibodies to CD18 or P-selectin had no effect. We conclude that adenoviral gene transfer is useful for studies of VCAM-1–dependent leukocyte adhesion in vitro and that endothelial expression of VCAM-1 alone, in the absence of overt endothelial cell activation, is sufficient under simulated physiological flow conditions to support adhesion of memory T cells, the predominant lymphocyte subset in atherosclerotic plaque.
The accumulation of mononuclear cells is an early event in the development of atherosclerotic lesions both in humans1 and in experimental animal models.2 These mononuclear cells are mostly monocyte-derived but also include lymphocytes, which constitute 5% to 20% of the mononuclear cell population.2 3 Plaque lymphocytes are predominantly CD45RO+ memory T cells3 and use a polyclonal repertoire of T-cell receptor variable chains, suggesting an amplification process in advanced lesions that is unlikely to be primarily antigen- or superantigen-driven.4 These considerations have prompted investigation of the expression within atherosclerotic plaques of adhesion molecules and cytokines associated with increased mononuclear leukocyte adhesion and infiltration.
VCAM-1 (CD106) is expressed by endothelial cells during atherogenesis in animal models5 and is also detectable in human atherosclerotic lesions.6 7 It selectively supports the adhesion of mononuclear cells5 8 but not neutrophils; this specificity reflects the restricted expression of its α4-integrin (CD49d) counterligand, which forms heterodimers with either β1- or β7-integrins. VCAM-1 is expressed in intimal neovasculature of human atherosclerotic plaques,9 and this expression correlates with increased accumulation of mononuclear cells.10 For these reasons, local VCAM-1 expression is thought to contribute to mononuclear cell recruitment and subsequent lesion formation.5 However, the role of VCAM-1 in this recruitment and the role of recruited mononuclear cells in atherogenesis remain undefined. Furthermore, it is unclear whether VCAM-1 expression alone would mediate mononuclear recruitment under flow, since prior studies with neutrophils have suggested an important role for initial selectin-mediated interactions under these conditions.11 12 In contrast, two recent publications demonstrated that VCAM-1 in the absence of selectins could support adhesion of lymphocytes under flow13 14 and, therefore, could potentially play a primary role in mononuclear leukocyte recruitment. However, the ability of specific lymphocyte subsets relevant to atherosclerosis, such as memory T cells, to adhere under flow conditions has not been addressed. Moreover, prior studies used purified immobilized VCAM-1, which may function differently from VCAM-1 expressed on the luminal surface of vascular endothelium.
The ability to manipulate expression of molecules such as VCAM-1 in vascular endothelial cells would aid efforts to test the role of these molecules in mononuclear leukocyte recruitment. Both the transfection efficiencies and the level of expression achieved in primary vascular endothelial cultures with traditional transfection techniques are generally inadequate to permit studies of leukocyte adhesion. Recombinant adenovirus vectors offer several significant advantages in this context. The viruses can be prepared at extremely high titer, infect nonreplicating cells, and confer high-efficiency and high-level transduction of endothelial cells in vitro and in vivo in both quiescent and injured vessels.15 16 17 18 19 20
The major disadvantages to adenoviral gene transfer have been its transience and the host immune response evoked by the vectors.21 Recent work in a rabbit femoral artery model demonstrated that adenovirus vectors induced upregulation of ICAM-1 and VCAM-1, an associated lymphocytic infiltrate, and neointimal formation.22 All of these observations present significant confounders in the use of first-generation adenovirus vectors for exploring the role of adhesion molecules in leukocyte recruitment in vivo.
In the present experiments, we investigated the potential application of adenovirus vectors in the study of mononuclear leukocyte adhesion in HUVECs, a well-characterized in vitro model of human vascular endothelium.23 We examined the effects of adenovirus vectors per se on expression of the endothelial adhesion molecules and markers of activation, E-selectin, ICAM-1, and VCAM-1. In addition, a recombinant adenovirus vector carrying rVCAM-15 was used to mediate surface expression in HUVECs in the absence of overt endothelial cell activation. These monolayers were shown to support adhesion of both a Jurkat T-cell line and human CD4+ CD45RO+ (memory) T cells under simulated physiological flow conditions.
Materials and Methods
The 293 cell line was obtained from the American Type Culture Collection and cultured in DMEM supplemented with 10% FCS (Hyclone) as previously described.24 The U937 cell line was cultured in RPMI-1640 containing 10% FCS, penicillin, and streptomycin antibiotics and 1 mmol/L l-glutamine. The Jurkat T-cell line (JS-10) was cultured under the same conditions. HUVECs were isolated from several normal-term umbilical veins, pooled, and cultured on 0.1% gelatin–coated tissue culture dishes as described previously23 25 26 in medium 199 with 20% FCS (Hyclone), endothelial cell growth factor (25 μg/mL, Biomedical Technologies), porcine intestinal heparin (50 μg/mL, Sigma Chemical Co), and antibiotics. Cell plating was adjusted so that HUVECs were ≈70% confluent on the day of infection. After infection with adenovirus vectors, HUVECs were cultured as described above, but the serum concentration was reduced to 10%. For stimulation of HUVECs, recombinant human TNF−α (200 U/mL, Biogen) was added as indicated. For experimental use in the flow plate apparatus, HUVECs (passages 1 and 2) were plated at confluence on 25-mm fibronectin-coated glass coverslips, as previously described.27
Two recombinant type 5 adenoviruses were used in these studies: AdRSVβ-gal and AdRSVrVCAM-1. AdRSVβ-gal is described in detail elsewhere.28 Both viruses use the dL 32729 backbone, contain E1/E3 deletions, and were generated in the following way. The RSV LTR and SV40 polyadenylation signal from pREP8 (Invitrogen) were cloned into the BgII site of pAdBgl II kindly provided by Dr Beverly Davidson30 (University of Iowa) after removal of the pREP8 Nhe I to Kpn I sequences, retaining a unique Nhe I site for future linearization, to yield pAdRSV4. Either a nuclear-targeted form of β-gal (AdRSVβ-gal) or the seven-domain isoform of rVCAM-1 (AdRSVrVCAM-1) was ligated into pAdRSV4 between the RSV LTR and SV40 sequences. The resulting plasmids were linearized at the unique Nhe I site and cotransfected with the large Cla I fragment of dL 327 DNA in 293 cells. Plaques were isolated, propagated in 293 cells, and characterized for protein expression by immunostaining in situ. Large-scale production of adenovirus was accomplished, as described previously,24 by infecting subconfluent 293 cell monolayers in 15-cm tissue culture dishes at an MOI of ≈10. Virus was purified by two sequential CsCl gradient centrifugations and desalted by chromatography on PD-10 columns in storage buffer (10 mmol/L Tris, pH 7.4, 1 mmol/L MgCl2, and 10% [vol/vol] glycerol). Viral titer of purified stocks was determined by plaque assay in 293 cells as previously described.24 Several different viral stocks were used in the course of these studies. Stock titers ranged from 109 to 1010 pfu/mL, with a particle-to-pfu ratio of ≈102.
Infection of HUVECs With Adenovirus Vectors
HUVECs were infected 18 hours after being plated. Infection was performed in 6- or 96-well plates (Costar) or on 25-mm-diameter glass coverslips by the addition of virus diluted in 0.6, 0.05, or 0.8 mL infection media (DMEM with 2% FCS), respectively, for 1 to 1.5 hours at 37°C. At that time, 1.5 or 0.15 mL of growth media was added to each well, and the cells were incubated for 1 to 3 days before evaluation.
Primary mAbs used included H18/7 and H4/1831 (to human E-selectin), Hu5/326 (to human ICAM-1), E1/632 (which recognizes both human and rabbit VCAM-1), Hu8/4 (which recognizes only the human VCAM-132 ), Rb1/95 (which is specific for rVCAM-1), HP2.1 (to α4-integrin, used at 10 μg/mL, purified IgG) (Immunotech), and TS1/18 (to CD18, used at 25 μg/mL, purified IgG) (American Type Culture Collection clone HB203). Antibodies to T-cell antigens CD45RA (H100) and CD45RO (UCHL1) were obtained from PharMingen. Antibody to CD28 (9.3) was kindly provided by Bristol-Myers Squibb. The function-blocking mAb HPDG2/3 to P-selectin was provided by Genetics Institute.
For β-Gal Activity
At the indicated intervals after infection, cells were released from the dishes by incubation in HBSS with 5 mmol/L EDTA and 4 mmol/L EGTA at 37°C for 20 minutes, followed by gentle aspiration. β-Gal activity was quantified using the FluoReporter LacZ flow cytometry kit (Molecular Probes, Inc) according to the manufacturer's specifications. The reaction was terminated by the addition of the competitive inhibitor phenylethyl-β-d-thiogalactopyranoside after the indicated time interval. Propidium iodide was used at 1 μg/mL to identify dead cells. Fluorescence was analyzed using a Becton-Dickinson FACS set to detect fluorescein, propidium iodide, forward scatter, and size.
Monolayers were released from dishes as described above and washed with RPMI/5% FCS (Hyclone). Cells were incubated with the indicated primary antibody for 30 to 60 minutes on ice, washed twice with RPMI/5% FCS, and incubated with a fluorescein-tagged secondary goat anti-mouse antibody (Caltag Laboratories). A nonbinding primary antibody was included as a control. As described above, in some experiments, cells were counterstained with propidium iodide to identify dead cells, and fluorescence was then analyzed using a Becton-Dickinson FACS set to detect fluorescein, propidium iodide, forward scatter, and size.
HUVEC monolayers in 96-well plates were incubated on ice with the indicated primary mAb in RPMI/1% FCS at 10 μg/mL for 45 minutes. Wells were washed three times with RPMI/1% FCS and then incubated with an FITC-conjugated goat anti-mouse polyclonal F(ab′)2 antibody (Caltag Laboratories) diluted 1:100 in Dulbecco's PBS on ice. After 45 minutes, wells were washed twice with DPBS/20% FCS and twice with DPBS alone. Cells were lysed with 0.01% NaOH in 0.1% SDS and fluorescence-quantified using either a Pandex or a CytoFluor 2350 (Perseptive Biosystems) fluorescent plate reader set at 485 (excitation)/535 (emission). For each flow experiment, a 96-well plate was cultured and infected in parallel at the same MOI. On the day of the flow adhesion assay, an FIA was performed on this plate to document rVCAM-1 expression and to assess nonspecific activation of the endothelial monolayer.
Leukocyte Isolation and Culture
CD4+ T cells were purified from single donor human platelet pheresis residues by sequential density gradient centrifugation, overnight culture, and then positive selection on Dynal Dynabeads M-450 CD4 and Dynal DETACHaBEAD (Dynal), as previously described.33 The CD4+ T cells isolated in this way are not activated; they do not express class II major histocompatibility locus antigens or CD25 (interleukin-2 receptor α chain), and they show a vigorous proliferative response to phytohemagglutinin in the presence of mononuclear accessory cells.33 Memory T cells were then prepared either by negative immunoselection to remove CD45RA+ cells or by in vitro conversion. Conversion of naive T cells to memory/activated T cells was accomplished by stimulating 2×105 CD4+ CD45RA+ cells with 1 μg/mL phytohemagglutinin, 1/1000 dilution of an ascites preparation of anti-CD28 mAb 9.3, and 5% human T-Stim (Collaborative Research) in 1 mL RPMI-1640 supplemented with 10% FCS, penicillin, streptomycin, l-glutamine, and 2-mercaptoethanol (60 μmol/L). Cells were expanded to larger culture volumes after day 4 and used in flow-adhesion assays at days 7 to 10. Phenotypic conversion was monitored by immunofluorescent staining and flow cytometry.33 Results were similar with CD45RO+ cells prepared by either immunoselection or by in vitro conversion and are pooled in the presented data.
Rotational Adhesion Assays
U937 cells were fluorescently labeled using BCECF-AM (Molecular Probes) according to the manufacturer's specifications, and resuspended in RPMI/1% FCS at a concentration of 2.0×106 cells/mL. Labeled U937 cells (0.8 mL) were added to HUVECs in C6 wells in triplicate under rotation (64 rpm on Hoefer Red Rotor) for 10 minutes at 22°C in RPMI/1% FCS.34 Wells were washed three times with RPMI/1% FCS, and adherent cells were released by divalent cation chelation. Fluorescence was quantified using a Pandex fluorescent plate reader set at 485 (excitation)/535 (emission), and the number of cells per well was calculated by comparison to fluorescence from known numbers of the labeled U937 cells, read in parallel.
Adhesion Assays Under Flow
The parallel-plate flow chamber used in the present study has been described previously in detail.34 Briefly, the chamber is composed of two stainless steel plates separated by a Silastic gasket (250 μm thickness, Dow Corning). The flow channel is formed by removal of a 5.0×50.0-mm rectangular section from the Silastic gasket. Defined levels of flow are applied to the HUVEC monolayer by drawing the perfusion medium (Dulbecco's PBS containing 0.7 mmol/L Ca2+, 0.7 mmol/L Mg2+, and 0.2% human serum albumin) through the channel using a syringe pump (model 44, Harvard Apparatus). A copper heating plate with two electrical heating cartridges (SC12-1, Hotwatt) is mounted on the top of the chamber to maintain temperature at 37°C. The channel flow can be approximated as two-dimensional fully developed laminar flow with a simple parabolic velocity profile, since the channel height is a linear function of the volume flow rate through the channel.34
Endothelial monolayers on coverslips were incubated with culture medium for 48 to 72 hours after viral infection. When appropriate, monolayers were stimulated with TNF-α (25 ng/mL) for 4 to 6 hours. Immediately before the assay, HUVEC monolayers were incubated with culture medium containing the indicated mAb or culture medium alone for 30 minutes at 37°C, positioned in the flow chamber, and mounted on an inverted microscope equipped with 10× and 40× phase-contrast objectives. A circular glass window in the top plate allows direct examination of the monolayer during the experiment. The monolayer was perfused for 5 to 10 minutes with perfusion medium to verify that the monolayer was confluent and intact. Concomitantly, leukocytes were incubated with the indicated mAb for 10 minutes at 4°C and diluted with perfusion medium to 106 cells/mL. The mAb concentration was adjusted to saturating levels in perfusion medium, and leukocytes were drawn through the chamber at controlled flow rates. The flow rate was 1.5 dyne/cm2 for 3 minutes and then was decreased every 3 minutes by 0.5 dyne/cm2 down to 0.5 dyne/cm2. The entire period of perfusion was recorded on videotape using a video recorder equipped with a time-date generator with a millisecond clock.27 Leukocyte adhesion was determined by counting the number of adherent T cells in five to seven randomly selected high-power microscope fields during the final minute at each level of flow.
All data are expressed as mean±SD. Statistical comparison of means was performed by two-tailed unpaired Student's t test or ANOVA, as indicated. The null hypothesis was considered rejected at P<.05.
Efficacy of Adenoviral Gene Transfer in HUVECs
Recombinant E1-deleted adenoviruses consistently produced high-efficiency and high-level gene transfer to HUVECs infected at 70% confluence in vitro, as indicated by flow cytometry using the fluorescent substrate for β-gal, fluorescein di-β-d-galactopyranoside (Fig 1⇓, top). Uninfected HUVECs typically exhibit a low background β-gal activity evidenced by the right-sided shoulder of the population histogram (Fig 1⇓, top). Some increase in fluorescence was evident at an MOI of 1.0 pfu/cell. However, beginning at an MOI of 10 pfu/cell, essentially the entire population expressed the transgene, producing a rightward shift in the fluorescence histogram and a dramatic increase in the mean fluorescence. At an MOI of 100 pfu/cell, a similar pattern is seen, with a further increase of mean fluorescence. However, the marginal increase in fluorescent conversion product as a reflection of β-gal activity is greatest between an MOI of 1.0 to 10 pfu/cell. Preliminary experiments revealed that the MOI of AdRSVrVCAM required to achieve significant surface expression of rVCAM-1 on HUVECs plated at confluence on coverslips for flow experiments was substantially higher (Fig 1⇓, bottom). Similarly, the MOI required for rVCAM-1 expression detectable by FIA in confluent HUVECs plated in microtiter plates was also higher than required for subconfluent monolayers (data not shown). For this reason, an MOI of 750 pfu/cell was used for both AdRSVβ-gal and AdRSVrVCAM in the flow experiments.
Effect of Adenovirus Infection on Expression of Endothelial-Leukocyte Adhesion Molecules
We examined the effects of adenoviral infection on endothelial expression of the adhesion molecules E-selectin, ICAM-1, and VCAM-1. Surface expression of these three molecules was examined by flow cytometry 24, 48, and 72 hours after infection over a broad range of MOI with AdRSVβ-gal. For HUVECs infected at 70% confluence at MOI below ≈30, no increase in surface expression of these molecules was evident at any of the time points. At an MOI of ≥30, a dose-dependent selective increase in ICAM-1 expression was evident beginning at 48 hours. At an MOI of >100, the mean fluorescence of ICAM-1 staining increased twofold to fourfold (Fig 2C⇓) at 48 hours. No increase in E-selectin or human VCAM-1 was evident after infection with either AdRSVβ-gal (Fig 2⇓) or AdRSVrVCAM-1 (data not shown) at any of the time points tested. The lack of E-selectin upregulation strongly suggests that the increased ICAM-1 expression is not due to endotoxin contamination in the viral preparations. Moreover, adenoviral infection does not interfere with the susceptibility of the cells to cytokine activation of these genes, because the endothelial cells still upregulated expression of each of these molecules in response to recombinant human interleukin-1β (Fig 2D⇓). A similar pattern of selective ICAM-1 induction at higher MOI was also seen with AdRSVrVCAM infection (data not shown), suggesting that these effects were not due to toxicity of β-gal. Because flow experiments necessitated infection of HUVECs at substantially higher MOI, these issues were reexamined in HUVECs infected on coverslips at an MOI of 750 pfu/cell. Interestingly, the relative increase in surface ICAM-1 expression under these conditions was similar to that seen at a lower MOI for HUVECs infected at 70% confluence, with an approximately fourfold increase in mean fluorescence (Fig 3⇓). No increase in E-selectin or VCAM-1 surface expression was evident by flow cytometry in confluent HUVECs infected on glass coverslips at an MOI of 750 pfu/cell. Of note, in confluent HUVECs infected in parallel in microtiter plates at an MOI of 750 pfu/cell, no increase in ICAM-1 surface expression was detectable by the less sensitive FIA (data not shown). This observation reinforces the impression that the increased ICAM-1 surface expression is relatively modest under the conditions studied.
AdRSVrVCAM-Transduced HUVECs Support U937 Adhesion
To assess whether the adenovirally expressed VCAM-1 was functional in HUVECs, rotational adhesion assays34 were performed as described above with the monocytic U937 cell line. HUVECs were infected in triplicate in C6 plates with AdRSVβ-gal or AdRSVrVCAM at an MOI of 30 pfu/cell and cultured for 48 to 72 hours after infection. AdRSVrVCAM-infected HUVECs supported significantly greater U937 adhesion than did AdRSVβ-gal–infected HUVECs (Fig 4⇓, P<.001). Binding was blocked by a mAb specific to rVCAM-1 (Rb1/9) but not by a nonbinding isotype-matched antibody to human E-selectin (H4/18) (Fig 4⇓). Similar results were obtained using static adhesion assays (data not shown).
AdRSVrVCAM-Transduced HUVECs SupportT-Cell Adhesion Under Laminar Flow
The behavior of AdRSVrVCAM-transduced HUVECs in an in vitro flow model was then evaluated. Expression of rVCAM-1 was confirmed before each flow experiment by FIA performed on a microtiter plate cultured and infected in parallel with the coverslips (data not shown). The JS-10 Jurkat T-cell line, which expresses high levels of the VCAM-1 counterligand, VLA-4, was evaluated first. Because preliminary experiments revealed no leukocyte adhesion at flow rates of ≥2.0 dyne/cm2, further studies were performed at flow rates of ≤1.5 dyne/cm2. The initial flow rate was 1.5 dyne/cm2 for 3 minutes and then was decreased every 3 minutes by 0.5 dyne/cm2 down to 0.5 dyne/cm2. As seen in Fig 5⇓, top, AdRSVrVCAM-infected HUVECs supported significantly greater adhesion of JS-10 cells under laminar flow between 1.5 and 0.5 dyne/cm2 than did AdRSVβ-gal–infected or uninfected HUVECs (P<.0001). Jurkat adhesion to AdRSVrVCAM-transduced monolayers was quantitatively similar to that seen with TNF-activated HUVECs. To verify specificity of the observed interaction, monoclonal blocking experiments were performed. A mAb to the VCAM-1 counterligand, VLA-4 (HP2.1), ablated Jurkat adhesion to AdRSVrVCAM-transduced HUVECs, but mAbs to the ICAM-1 counterligand, CD18 (TS1/18), or to a class I histocompatibility antigen (W6/32) did not block adhesion (Fig 5⇓, bottom). Jurkat T cells initially attached but, subsequently, did not roll downstream at any flow level examined. Rather, the lymphocytes remained firmly adherent to the apical surface of the endothelial cells and did not spread or transmigrate. This behavior is different from the interaction of lymphocytes with selectins, in which true rolling is observed.33
In order to model the behavior of one of the mononuclear leukocyte subsets present in atherosclerotic plaque, adhesion of CD4+ CD45RO+ (memory) T lymphocytes under flow was examined. Overall adhesion of the memory T cells was lower than that seen with the Jurkat T-cell line, manifesting little or no adhesion to AdRSVβ-gal–infected or uninfected controls (Fig 6⇓, top). AdRSVrVCAM-infected HUVECs supported significantly greater levels of memory T-cell adhesion than did either AdRSVβ-gal–infected or uninfected control HUVECs (P<.0001). As with the Jurkat cell line, adhesion was completely blocked by a mAb to the VCAM-1 counterligand, VLA-4 (HP2.1), but not by a function-blocking mAb to the ICAM-1 counterligand, CD18 (TS1/18), or an antibody to class I histocompatibility antigen (W6/32) (Fig 6⇓, middle). Leukocyte adhesion was also blocked by Rb1/9, a mAb specific for rVCAM-1, demonstrating that adhesion was mediated by the adenovirally expressed rVCAM-1 rather than endogenous human VCAM-1 or fibronectin (Fig 6⇓, bottom). No rolling of memory T cells was observed under any of the conditions tested, again suggesting no significant selectin-mediated interaction. Furthermore, a function-blocking mAb to P-selectin did not affect adhesion (Fig 6⇓, bottom).
Infection with AdRSVβ-gal did not significantly increase basal leukocyte adhesion compared with uninfected controls for any of the cells studied. In addition, as noted above, the anti-CD18 mAb did not affect adhesion to AdRSVrVCAM-transduced HUVECs. Therefore, although surface ICAM-1 expression is detectably increased, the increase is small and does not participate in leukocyte adhesion. The lack of CD18/ICAM involvement in initial attachment of T cells under simulated physiological flow conditions is consistent with previous publications.13 33
In the present experiments, we used recombinant adenovirus vectors to study mononuclear leukocyte adhesion to vascular endothelial cells under flow conditions in vitro. Recombinant adenovirus vectors produced efficient and titratable transgene expression in HUVECs, an in vitro model of human vascular endothelium.23 HUVECs were not globally activated by infection with adenovirus vectors: no increase in surface expression of the endothelial adhesion molecules and sensitive markers of activation, E-selectin and VCAM-1, was observed 24 to 72 hours after infection across a broad range of MOI. Moreover, the cells could still express all of these molecules normally in response to cytokine treatment after infection. A dose-dependent specific induction of surface ICAM-1 expression was detectable by flow cytometry. This relatively modest induction was seen with recombinant viruses carrying either β-gal or VCAM-1. However, ICAM-1 did not participate in the leukocyte adhesion observed in vitro. Interestingly, the plating conditions significantly influenced the efficacy of gene transfer. HUVECs infected 1 day after confluence required a significantly higher MOI to achieve comparable surface VCAM-1 expression than did HUVECs infected at 70% confluence. In either condition, however, an MOI was identified at which effective transgene expression could be achieved with minimal nonspecific perturbation of the endothelial cells, suggesting that adenoviral gene transfer may be a valuable tool for dissecting adhesion biology in vitro.
To study the role of VCAM-1 in mononuclear leukocyte recruitment, a recombinant adenovirus carrying the seven-domain isoform of rVCAM-1 was constructed. The virus produced efficient surface expression of functional rabbit VCAM-1 on HUVECs in the absence of activation by exogenous cytokine or other stimulus. In a well-characterized in vitro flow model, AdRSVrVCAM-infected HUVECs supported adhesion of a Jurkat T-cell line and CD4+ memory T cells under defined laminar flow conditions. This adhesion was blocked by an antibody to the VCAM-1 counterligand, VLA-4, or to rVCAM-1 but was unaffected by mAb blockade of the ICAM-1/CD18 pathway. Therefore, although viral infection evoked a modest increase in surface ICAM-1 expression, ICAM-1 did not participate in leukocyte adhesion under flow, consistent with the reports of several other investigators.13 14 Moreover, the virtually complete abolition of binding by antibodies to α4-integrins and the absence of leukocyte rolling further indicate that other adhesion molecules, specifically selectins, are not contributing significantly to leukocyte adhesion in our system. Consistent with this, function-blocking antibodies to P-selectin did not affect adhesion of memory lymphocytes to AdRSVrVCAM-1–transduced HUVECs. These results are the first to demonstrate that VCAM-1 alone is sufficient to support adhesion of memory lymphocytes to vascular endothelial cells under laminar flow at shear levels up to 1.5 dyne/cm2. The present study confirms prior studies demonstrating adhesion of lymphocytes under flow to purified VCAM-1 in vitro.13 14 These observations constitute an exception to the previously accepted paradigm that initial leukocyte adhesive events are selectin-mediated and that adhesion molecules of the immunoglobulin family (such as VCAM-1 and ICAM-1) only participate later in a multistep cascade.11 12 The present study extends previous observations by demonstrating adhesion to vascular endothelial cells expressing VCAM-1 rather than to immobilized protein and by focusing on the behavior of the predominant lymphocyte subset present in the atherosclerotic plaque, CD45RO+ memory T lymphocytes.3
In contrast to prior studies using purified VCAM-1 protein,13 14 Jurkat T cells and memory T cells failed to roll on the VCAM-1–expressing monolayers under any of the flow conditions examined but rather remained firmly adherent to the endothelial surface after initial attachment. This difference could reflect the contribution of other molecules present in endothelial cultures but not in purified protein preparations. The HUVECs studied express ICAM-1, which could enhance firm adhesion.11 However, blockade of ICAM-1/CD18 did not restore rolling to our system, which suggests that ICAM-1 does not account for the absence of rolling. Chemokines can enhance VCAM-dependent adhesion,35 and our preliminary results demonstrate that MCP-1 is detectable by enzyme-linked immunosorbent assay in the supernatants of HUVEC cultures and is unchanged by adenoviral transduction (data not shown). However, the relevance of this basal MCP-1 expression to relatively brief flow experiments, during which the perfusate continuously removes secreted proteins, remains unclear. Moreover, a recent publication suggests that MCP-1 does not enhance VCAM-dependent T-cell adhesion.36 Further studies will be necessary to determine whether the absence of rolling reflects modulation of adhesion by specific molecules in our system or a difference in the interaction of T cells with VCAM-1 presented in the context of the endothelial cell surface compared with purified VCAM-1 protein.
Consistent with previous in vitro flow studies,13 14 27 the shear forces that permitted leukocyte attachment are at the low end of physiological shear. In vivo multiple physical and biological factors not present in our simple in vitro system are thought to enhance leukocyte-endothelium interactions. These include the presence of erythrocytes37 and complex rheological patterns associated with vessel branch points. The shear forces studied may be more similar to those seen in intimal neovascular microvessels present in atherosclerotic plaques than to those within the arterial lumen. VCAM-1 expression is more prevalent on this neovasculature than on the arterial luminal endothelium in human atherosclerosis.9 10 Moreover, neovascular VCAM-1 expression correlates with increased intimal lymphocyte and macrophage accumulation.10 Our studies support the hypothesis that neovascular VCAM-1 could play an important role in local recruitment of lymphocytes.
Newman et al22 demonstrated increased expression in vivo of both ICAM-1 and VCAM-1, as well as marked lymphocyte infiltration in rabbit arteries after adenoviral infection, raising the possibility that adenoviral infection directly upregulated ICAM-1 and VCAM-1 expression in transduced cells, leading to T-cell infiltration. However, the time course, location, and spatial relationship to T-cell infiltrates of ICAM-1 and VCAM-1 expression suggested a primary role of T cell–derived cytokines in the induction of ICAM-1 and VCAM-1.22 We found that adenoviral transduction increased only ICAM-1 expression in vitro. Moreover, we found no increase in lymphocyte adhesion to monolayers infected with the β-gal virus. Our results, therefore, are consistent with the inference of Newman et al that in vivo ICAM-1 and VCAM-1 expression are upregulated indirectly.22
Although the increased expression of ICAM-1 we observed was relatively modest and did not affect adhesion in our in vitro model, it is difficult to extrapolate these findings directly to the in vivo setting. Estimation of the number of endothelial cells in a 1-cm-long rabbit arterial segment with a 2-mm diameter, assuming an endothelial cell surface area of 700 μm2,38 reveals ≈105 endothelial cells in the segment. Since investigators typically instill ≈108 to 109 pfu into such segments,19 39 40 vascular gene transfer in vivo is currently performed at an MOI of 103 to 104, well within the range in which we observe dose-dependent ICAM-1 induction in vitro. Since ICAM-1 supports adhesion of most leukocyte populations,26 41 42 even modestly increased surface expression of ICAM-1 could enhance adhesion of infiltrating leukocytes that elaborate cytokines capable of further activating vascular cells. However, in vivo, other factors such as far shorter periods of viral exposure compared with in vitro experiments might mitigate ICAM-1 induction despite comparable MOI. The relative contribution of individual adhesion molecules to the inflammatory cascade observed in vivo remains to be tested.
The present study did not address the utility of adenovirus vectors to study leukocyte recruitment in vivo. The findings of Newman et al22 suggest that background effects due to adenovirus-induced inflammation in vivo may represent a significant impediment to such studies in rabbit vascular segments. Alternative approaches, however, may be feasible. Other species may exhibit a less marked reaction to adenoviral infection.22 Certain immunoincompetent mouse mutants have a dramatically reduced immune response to adenoviral infection,43 have been extensively used in studies of leukocyte trafficking in vivo,44 45 46 and might serve as a valuable substrate for adenoviral transfer of adhesion molecule constructs. Finally, newer generation adenovirus vectors that inactivate other vital viral functions in addition to E1 can further reduce the host inflammatory response.47 A combination of these approaches may enable investigators to use adenovirus vectors to study adhesion biology in vivo. However, the extent of nonspecific proinflammatory vascular effects in these systems remains to be addressed.
Nevertheless, significant questions in adhesion biology remain and can be addressed in vitro using current adenovirus vectors. Unanswered and of obvious interest is whether VCAM-1/VLA-4 interactions will support monocyte adhesion under simulated physiological flow conditions. Furthermore, it remains unclear why naive (CD45RA+/RO−) T lymphocytes, which also express VLA-4, are scarce in atherosclerotic plaque.3 It is possible that CD45RA+/RO− T cells, in contrast to the CD45RO+ T cells studied here, do not adhere to VCAM-1 under flow. This hypothesis can be tested in vitro using the reagents described in the present study. Modulation of VCAM-dependent adhesion by local expression of other molecules such as selectins and chemokines may help determine the specificity of leukocyte recruitment. Such issues can potentially be addressed in vitro by using adenovirus vectors to achieve coexpression of relevant molecules.
In conclusion, the present study demonstrates that adenovirus vectors mediate relatively modest increases in endogenous endothelial ICAM-1 surface expression and no increase in endogenous E-selectin or VCAM-1 surface expression in infected cultured human endothelial cells. Using a recombinant adenovirus vector carrying rVCAM-1, we demonstrate that VCAM-1/VLA-4 interactions alone are sufficient to support adhesion of lymphocytes under simulated physiological flow conditions to vascular endothelial monolayers in the absence of cytokine activation. Adherent lymphocytes studied include CD45RO+ memory T cells, the predominant lymphocyte in atherosclerotic plaque.3 These studies support the hypothesis that the increased VCAM-1 expression observed in atherosclerotic plaque may be an important contributor to lymphocyte recruitment. In addition, adenovirus vectors appear to be useful tools for further dissecting VCAM-1–dependent adhesion biology in vitro.
Selected Abbreviations and Acronyms
|Ad (combination form)||=||adenovirus|
|HUVEC||=||human umbilical vein endothelial cell|
|ICAM-1||=||intercellular adhesion molecule-1|
|LTR||=||long terminal repeat|
|MCP-1||=||monocyte chemotactic protein-1|
|MOI||=||multiplicity of infection|
|RSV||=||Rous sarcoma virus|
|SV40||=||simian virus 40|
|TNF||=||tumor necrosis factor|
|VCAM-1||=||vascular cell adhesion molecule-1|
|VLA-4||=||very late antigen-4|
The authors gratefully acknowledge support from the National Institutes of Health (grants HL-54202 to Dr Rosenzweig, HL-36028 to Drs Luscinskas, Gimbrone, Rosenzweig, and Ding, and HL-475646 to Dr Luscinskas and a training grant to Dr Gerszten), the American Heart Association (Clinician Scientist Award to Dr Gerszten), the D.Y. and Joan Fu Fund for Cardiovascular Research (to Dr Rosenzweig), the Merrill Lynch Research Fund (to Dr Rosenzweig), and Bristol-Myers Squibb (to the Cardiovascular Research Center). The authors wish to thank Kay Case and Bill Atkinson (Vascular Research Division at the Brigham and Women's Hospital) for preparation of human endothelial cell cultures and Lori DesRoches and Dr Andrew Lichtman (Immunology Division, Brigham and Women's Hospital) for isolation and characterization of CD4+ T cells. We also thank Dr Myron Cybulsky for providing the rVCAM-1 cDNA and antibody Rb1/9, Dr Thomas Shenk for adenovirus dL 327, Dr Beverly Davidson for pAdBgl II, and Bristol-Myers Squibb for antibody 9.3 to CD28.
This manuscript was sent to Bradford C. Berk, MD, PhD, Consulting Editor, for review by expert referees, editorial decision, and final disposition.
- Received June 5, 1996.
- Accepted September 17, 1996.
Hansson GK, Seifert PS, Olsson G, Bondjers G. Immunohistochemical detection of macrophages and T lymphocytes in atherosclerotic lesions of cholesterol-fed rabbits. Arterioscler Thromb. 1991;11:745-750.
Stemme S, Holm J, Hansson GK. T lymphocytes in human atherosclerotic plaques are memory cells expressing CD45RO and the integrin VLA-1. Arterioscler Thromb. 1992;12:206-211.
Swanson S, Rosenzweig A, Seidman JG, Libby P. An analysis of T cell heterogeneity in advanced atherosclerotic plaques. Arterioscler Thromb. 1994;14:1210-1214.
Cybulsky MI, Gimbrone MA. Endothelial expression of a mononuclear leukocyte adhesion molecule during atherogenesis. Science. 1991;251:788-791.
O'Brien KD, Allen MD, McDonald TO, Chait A, Harlan JM, Fishbein D, McCarty J, Ferguson M, Hudkins K, Benjamin CD, et al. Vascular cell adhesion molecule-1 is expressed in human coronary atherosclerotic plaques: implications for the mode of progression of advanced coronary atherosclerosis. J Clin Invest. 1993;92:945-951.
O'Brien KD, McDonald TO, Chait A, Allen MD, Alpers CE. Neovascular expression of E-selectin, intercellular adhesion molecule-1, and vascular cell adhesion molecule-1 in human atherosclerosis and their relation to intimal leukocyte content. Circulation. 1996;93:672-682.
Alon R, Kassner PD, Carr MW, Finger EB, Hemler ME, Springer TA. The integrin VLA-4 supports tethering and rolling in flow on VCAM-1. J Cell Biol. 1995;128:1243-1253.
Erzurum SC, Lemarchand P, Rosenfeld MA, Yoo JH, Crystal RG. Protection of human endothelial cells from oxidant injury by adenovirus-mediated transfer of the human catalase cDNA. Nucleic Acids Res. 1993;21:1607-1612.
Chen S-J, Wilson JM, Muller DWM. Adenovirus-mediated gene transfer of soluble vascular cell adhesion molecule to porcine interposition vein grafts. Circulation. 1994;89:1922-1928.
Lemarchand P, Jones M, Yamada I, Crystal RG. In vivo gene transfer and expression in normal uninjured blood vessels using replication-deficient recombinant adenovirus vectors. Circ Res. 1993;72:1132-1138.
Willard JE, Landau C, Glamann DB, Burns D, Jessen ME, Pirwitz MJ, Gerard RD, Meidell RS. Genetic modification of the vessel wall: comparison of surgical and catheter-based techniques for delivery of recombinant adenovirus. Circulation. 1994;89:2190-2197.
Yang Y, Nunes FA, Berencsi K, Furth EE, Gonczol E, Wilson JM. Cellular immunity to viral antigens limits E1-deleted adenoviruses for gene therapy. Proc Natl Acad Sci U S A. 1994;91:4407-4411.
Newman KD, Dunn PF, Owens JW, Schulick AH, Virmani R, Sukhova G, Libby P, Dichek DA. Adenovirus-mediated gene transfer into normal rabbit arteries results in prolonged vascular cell activation, inflammation, and neointimal hyperplasia. J Clin Invest. 1995;96:2955-2965.
Graham FL, Prevec L. Manipulation of adenovirus vectors. In: Murray EJ, ed. Gene Transfer and Expression Protocols. Clifton, NJ: Humana Press Inc; 1991:109-128.
Gimbrone MA, Cotran RS, Folkman J. Human vascular endothelial cells in culture: growth and DNA synthesis. J Cell Biol. 1974;60:673-684.
Luscinskas FW, Cybulsky MI, Kiely JM, Peckins CS, Davis VM, Gimbrone MA. Cytokine-activated human endothelial monolayers support enhanced neutrophil transmigration via a mechanism involving both endothelial-leukocyte adhesion molecule-1 and intercellular adhesion molecule-1. J Immunol. 1991;146:1617-1625.
Luscinskas FW, Kansas GS, Ding H, Pizcueta P, Schleiffenbaum BE, Tedder TF, Gimbrone MA. Monocyte rolling, arrest and spreading on IL-4-activated vascular endothelium under flow is mediated via sequential action of L-selectin, beta 1-integrins, and beta 2-integrins. J Cell Biol. 1994;125:1417-1427.
Dong G, Schulick A, DeYoung MB, Dichek DA. Identification of a cis-acting sequence in the human PAI-1 gene that mediates TGF-1-responsiveness in endothelium in vivo. J Biol Chem. In press.
Bevilacqua MP, Pober JS, Mendrick DL, Cotran RS, Gimbrone MA. Identification of an inducible endothelial-leukocyte adhesion molecule. Proc Natl Acad Sci U S A. 1987;84:9238-9242.
Taichman DB, Cybulsky MI, Djaffar I, Longenecker BM, Teixido J, Rice GE, Aruffo A, Bevilacqua MP. Tumor cell surface alpha 4 beta 1 integrin mediates adhesion to vascular endothelium: demonstration of an interaction with the N-terminal domains of INCAM-110/VCAM-1. Cell Regul. 1991;2:347-355.
Luscinskas FW, Ding H, Lichtman AH. P-selectin and vascular cell adhesion molecule 1 mediate rolling and arrest, respectively, of CD4+ T lymphocytes on tumor necrosis factor alpha-activated vascular endothelium under flow. J Exp Med. 1995;181:1179-1186.
Spertini O, Luscinskas FW, Kansas GS, Munro JM, Griffin JD, Gimbrone MA, Tedder TF. Leukocyte adhesion molecule-1 (LAM-1, L-selectin) interacts with an inducible endothelial cell ligand to support leukocyte adhesion. J Immunol. 1991;147:2565-2573.
Laudanna C, Campbell JJ, Butcher EC. Role of Rho in chemoattractant-activated leukocyte adhesion through integrins. Science. 1996;271:981-983.
Guzman RJ, Lemarchand P, Crystal RG, Epstein SE, Finkel T. Efficient and selective adenovirus-mediated gene transfer into vascular neointima. Circulation. 1993;88:2838-2848.
Lee SW, Trapnell BC, Rade JJ, Virmani R, Dichek DA. In vivo adenoviral vector–mediated gene transfer into balloon-injured rat carotid arteries. Circ Res. 1993;73:797-807.
Bochner BS, Luscinskas FW, Gimbrone MA, Newman W, Sterbinsky SA, Derse ACP, Klunk D, Schleimer RP. Adhesion of human basophils, eosinophils, and neutrophils to interleukin 1-activated human vascular endothelial cells: contributions of endothelial cell adhesion molecules. J Exp Med. 1991;173:1553-1557.
Yang Y, Xiang Z, Ertl HC, Wilson JM. Upregulation of class I major histocompatibility complex antigens by interferon gamma is necessary for T-cell-mediated elimination of recombinant adenovirus-infected hepatocytes in vivo. Proc Natl Acad Sci U S A. 1995;92:7257-7261.
Yan HC, Juhasz I, Pilewski J, Murphy GF, Herlyn M, Albelda SM. Human/severe combined immunodeficient mouse chimeras: an experimental in vivo model system to study the regulation of human endothelial cell-leukocyte adhesion molecules. J Clin Invest. 1993;91:986-996.
Yan HC, Delisser HM, Pilewski JM, Barone KM, Szklut PJ, Chang XJ, Ahern TJ, Langer-Safer P, Albelda SM. Leukocyte recruitment into human skin transplanted onto severe combined immunodeficient mice induced by TNF-alpha is dependent on E-selectin. J Immunol. 1994;152:3053-3063.
Fukumura D, Salehi HA, Witwer B, Tuma RF, Melder RJ, Jain RK. Tumor necrosis factor alpha-induced leukocyte adhesion in normal and tumor vessels: effect of tumor type, transplantation site, and host strain. Cancer Res. 1995;55:4824-4829.
Engelhardt JF, Ye X, Doranz B, Wilson JM. Ablation of E2A in recombinant adenoviruses improves transgene persistence and decreases inflammatory response in mouse liver. Proc Natl Acad Sci U S A. 1994;91:6196-6200.