ORFK3 (K3) and ORFK5 (K5) are Kaposi’s sarcoma-associated herpesvirus-encoded E3 ubiquitin ligases that differentially reduce surface expression of various proteins in infected cells. In this study, we describe their effects on human dermal microvascular endothelial cells (ECs), a natural target of Kaposi’s sarcoma-associated herpesvirus infection. TNF-treated human dermal microvascular ECs transduced to express K5 show reduced capacity to capture effector memory (EM) CD4+ T cells under conditions of venular shear stress. K5 but not K3 transduction significantly reduces ICAM-1 expression and the inhibition of T cell capture was phenocopied by small interfering RNA knockdown of ICAM-1 and by anti–ICAM-1 Ab blocking. Cotransduction with an ICAM-1 truncated construct not subject to K5 ubiquitylation restored EM CD4+ T cell capture. K3 transductants effectively capture EM CD4+ T cells, but fail to support their transendothelial migration (TEM) in response to TCR engagement by superantigen presented by the ECs, leaving intact chemokine-dependent TEM. K3 but not K5 transduction significantly reduces PECAM-1 expression, and the effect on TCR-induced TEM is phenocopied by small interfering RNA knockdown of PECAM-1 and by anti–PECAM-1 Ab blocking. TCR-dependent TEM was restored in K3 transductants cotransduced to express a mutant of PECAM-1 not subject to K3-induced ubiquitylation. EM CD4+ T cells lack any known PECAM-1 counter receptor, but heterophilic engagement of PECAM-1 can involve glycosaminoglycans. In addition, TCR-induced TEM, but not chemokine-induced TEM, appears to involve a heparan- or chondroitin-like molecule on T cells. These results both identify specific roles of K5 and K3 in immune evasion and further differentiate the processes of inflammatory chemokine- versus TCR-dependent recruitment of human EM CD4+ T cells.

Kaposi’s sarcoma-associated herpesvirus (KSHV), also known as human herpesvirus 8, is the etiologic agent of Kaposi’s sarcoma (KS). Virally induced cancers are highly immunogenic, and KS is largely a disease of severely immunosuppressed patients (e.g., because of drugs given to transplant recipients or HIV-induced AIDS). The natural targets of KSHV are microvascular endothelial cells (ECs), a cell population that is directly exposed to circulating effector elements of the host immune system. KSHV has evolved multiple mechanisms of immune evasion to persist in immunocompetent hosts; indeed, 25% of the proteins encoded by KSHV genes have been shown to modulate different aspects of the host immune response (1). KSHV proteins encoded by ORFK3 (K3) and ORFK5 (K5), also known as modulator of immune response 1 and modulator of immune response 2, respectively, are E3 ubiquitin ligases that can selectively downregulate cell surface proteins that participate in immune responses. These include MHC class I, B7-2, ICAM-1 (CD54), CD1d, PECAM-1 (CD31), activated leukocyte cell adhesion molecule (ALCAM) (CD166), IFN-γR1, MICA/B, AICL, and vascular endo-thelial cadherin (VE-cadherin) (211). Most of these targets have been investigated in BJAB B lymphoblastoid cells or HeLa cells. EC-specific targets of K3 and K5 (e.g., CD31 and VE-cadherin) have been investigated in immortalized ECs (7, 11). Infection of primary ECs with KSHV downregulates both ICAM-1 and PECAM-1, which has been attributed to K5 because K5, but not K3, downregulates ICAM-1 and PECAM-1 in BJAB cells and immortalized ECs, respectively (5, 7, 12). However, the functional effects of overexpressing K3 and K5 in untransformed human microvascular ECs, the natural target of the virus, have not been described.

Microvascular ECs are active participants in the effector phase of the adaptive immune response. We have been interested in mechanisms by which microvascular ECs can recruit circulating effector memory (EM) CD4+ T cells. Unlike naive or central memory T cells, freshly isolated EM CD4+ T cells can rapidly transmigrate across cultured HUVECs or human dermal microvascular ECs (HDMECs) displaying TNF-induced adhesion molecules, such as ICAM-1 and VCAM-1, plus the inflammatory chemokine IP-10 (13). More recently, we have shown that EM CD4+ T cells, but not central memory or naive CD4+ T cells, will transmigrate in response to signals that engage the TCR, such as superantigen presented by TNF-treated HDMECs expressing class II MHC molecules (14). Although both modes of transendothelial migration (TEM) require venular type flow to provide shear stress, TCR-dependent (Ag-driven) TEM of EM CD4+ T cells is somewhat delayed compared with chemokine-driven TEM, and it may be further differentiated by its requirement for EC fractalkine. Moreover, cells receiving a TCR signal are unresponsive to chemokines, such as IP-10 or SDF-1. In this study, we show that overexpressing the KSHV genes K3 or K5 in HDMECs inhibit EM CD4+ T cell recruitment at different steps. Specifically, we find that K5 inhibits capture of flowing EM CD4+ T cells by HDMEC, whereas K3 inhibits TCR-dependent but not rapid chemokine-dependent TEM of EM CD4+ T cells. Using small interfering RNA (siRNA) and blocking Abs, as well as reconstitution experiments, we present evidence that K5-mediated inhibition of the capture of EM CD4+ T cells can be explained by ubiquitin-dependent downregulation of ICAM-1, whereas K3-mediated inhibition of TCR-driven TEM can be explained by ubiquitin-dependent downregulation of PECAM-1. PECAM-1 has not been implicated previously in the TEM of T cells, and we find that EM CD4+ T cells lack any known counter-receptor for PECAM-1. Heterophilic PECAM-1 engagement has been shown previously to induce binding of heparan- or chondroitin-like glycosaminoglycans, and we further show that a heparan- or chondroitin-like glycosaminoglycan present on EM CD4+ T cells is also important for TCR-initiated, but not chemokine-driven, TEM. These data identify specific mechanisms that can be used by KSHV to evade the human immune system, and they concomitantly expand our understanding of the differences between TCR- versus chemokine-driven TEM of human EM CD4+ T cells.

CIITA HDMEC were generated using a retroviral vector and characterized as described (14). GFP, K5, K5Em, and K3Em lentivirus transductants were prepared by transducing HDMEC or CIITA HDMEC with lentivirus produced from cotransfection of NIH3T3 cells with lentivirus vector plasmid pHRSIN–K3 or –K5 (based on the vector pHR'SINcPPT–SGW; a gift from Yasuhiro Ikeda, Mayo Clinic, Rochester, MN) or pHRSIN–UbEmerald-K3, -K5 or empty vector (coexpressing the Emerald variant of GFP under a separate ubiquitin promoter), packaging plasmid psPAX2 (a gift from Didier Trono, Ecole Polytechnique Federale de Lausanne, Switzerland), and envelope protein plasmid pLP/VSVG. HDMECs were transduced with retrovirus to express ICAM-1 lacking the cytoplasmic tail as described (15). Site-directed mutagenesis was used to create the PECAM-1 construct with all 11 cytoplasmic domain lysine residues changed to arginine, and one cysteine to alanine. This construct (PECAM-12 mut) or the unmodified cDNA (PECAM wild-type [wt]; a gift from Joseph Madri, Yale University, New Haven, CT) were subcloned into retroviral vector LZRSpBMN-Z (a gift from Garry Nolan, Stanford University). Retroviral supernatants, produced by transfecting phoenix cells, were used to transduce CIITA HDMEC. For most experiments, ECs were incubated in the presence of 10 ng/ml recombinant human TNF (TNF-α; R&D Systems, Minneapolis, MN) for 18–26 h. PE-conjugated mAbs to PECAM-1, ICAM-1, ALCAM, HLA-A,B,C control IgG (BD Pharmingen, San Diego, CA) and VE-cadherin (eBioscience, San Diego, CA) and a mouse anti-human VCAM-1 mAb (clone E1/6.5) or a mouse anti FLAG (clone M2) conjugated with Alexafluor 647 (Zenon Alexafluor 647; Invitrogen, Carlsbad, CA) were used to stain cells for FACS. A mAb to K5, a gift from Klaus Früh (Oregon Health and Science University, Portland, OR), was used to stain samples for immunofluorescence microscopy. For the blocking Ab experiments, ECs were incubated in the presence of 10 μg/ml blocking Abs to ICAM-1 (R&D Systems) or PECAM-1 (hec7) for 30 min prior to the flow assay. For siRNA experiments, HDMEC were treated with siRNAs targeting ICAM-1 (Dharmacon, Lafayette, CO) or PECAM-1 (Qiagen, Valencia CA) or negative control (Allstar negative control siRNA; Qiagen) as described (15). All reagents used in the glycosaminoglycan experiments were from Sigma-Aldrich (St. Louis, MO). ECs were preincubated with 100 μg/ml heparan sulfate (sodium salt, from bovine kidney), 100 μg/ml chondroitin-6-sulfate (sodium salt, from shark cartilage), and 100 μg/ml hyaluronic acid (sodium salt, from Streptococcus equi). T cells were treated with 50 mU/ml chondroitinase ABC or 100 mU/ml hyaluronidase (from Streptomyces hyalurolyticus).

CD4+ T cells were isolated by positive selection with magnetic beads and released with Detachabead (Dynal) from PBMCs prepared by a Ficoll gradient of blood collected from healthy donors. Memory (CD4+CD45RA) T cells were isolated by depletion of CD45RA+ cells from CD4+ T cells using anti-CD45RA mAb (eBiosciences) and pan-mouse IgG beads. EM cells were further enriched by depleting CD4+ memory cells with anti-CCR7 mAb (R&D Systems) and pan-mouse IgG beads. Isolated EM CD4+ T cells were cultured in RPMI 1640 medium supplemented with 10% FBS, 2 mM glutamine, penicillin/streptomycin, and nonessential amino acids overnight prior to assays.

Flow Cytometry Standard files of trypsinized ECs, stained for 30 min on ice with PE-conjugated mAb to ICAM-1, PECAM-1, ALCAM, VE-cadherin, HLA-A,B,C, and HLA-DR or anti–VCAM-1 mAb, followed by Alexafluor 647-conjugated goat anti-mouse, and then washed two times with 1% BSA in PBS, were acquired using FACSCAN or LSRII flow cytometers with Cellquest (Becton Dickinson, San Jose, CA) or FACSDiva (Becton Dickinson) software, respectively, and analyzed using Flowjo software (TreeStar, Ashland, OR). For intracellular staining of FLAG-tagged K3, fixed ECs permeabilized with 0.1% saponin were stained with anti-FLAG mAb M2 precomplexed with Alexafluor 647-conjugated Fab fragment goat anti-mouse IgG. Total CD4+ or EM CD4+ T cells were stained with FITC-conjugated mouse IgG control, mouse anti-human CD38, mouse anti-human CD31, or mouse anti-human CD177, washed two times with 1% BSA in PBS, acquired using LSRII flow cytometer with FACSDiva software, and analyzed using Flowjo software.

CIITA HDMEC and CIITA + GFP, CIITA + K3Em, CIITA + K5Em cotransductants, grown to confluence on 35 mm fibronectin-coated coverglasses, were incubated with 100 ng/ml TSST-1 (Toxin Technology, Sarasota, FL) 30 min prior to the flow assay and then 3–5 min with 3 μg/ml IP-10, washed twice with RPMI/10% FBS, and assembled with a parallel plate flow chamber apparatus (Glycotech, Gaithersburg, MD) using the 0.01-inch-high, 5-mm-wide slit gasket provided by the manufacturer. On a 37°C heating surface, CD4+CD45RACCR7low (EM) T cells (106 cells per 500 μl) suspended in the same medium were loaded onto the EC monolayer at 0.75 dyne/cm2 for 2 min, followed by medium only at 1 dyne/cm2 for 15 or 60 min. Samples were then fixed with 3.7% formaldehyde in PBS, stained with anti-Vβ2TCR mAb (Immunotech, Marseille, France) followed by Alexafluor 594- or 488-conjugated donkey anti-mouse IgG (Molecular Probes, Eugene, OR), mounted on slides using mounting medium containing DAPI (Molecular Probes), and examined by microscopy. An FITC filter was used to detect FITC or Alexafluor 488-stained cells, a TRITC filter was used to detect Alexafluor 594-stained cells, a DAPI filter used to detect DAPI-stained nuclei, and a Cy5 filter was used to detect Alexafluor 647-stained cells. Using a 40×/0.60 korr Ph2 objective, phase contrast optics were used to determine whether CD4+ T cells were either on top of or underneath the HDMEC monolayer. T cells that were captured and not spread were round and bright when viewed under phase contrast. CD4+ T cells that were spread, but still on top of the HDMEC monolayer, were surrounded by a bright corona of light in contrast to those that had transmigrated. The percentage of transmigrated CD4+ T cells was calculated for 200 cells per sample by analyzing 10 groups of 20 cells each, calculating the percentage for each group, and calculating the mean and SEM for the 10 groups. For cell capture experiments, samples were stained with FITC-conjugated anti CD45, Alexafluor 488-conjugated rabbit anti-FITC, and Alexafluor 488-conjugated goat anti rabbit IgG. The number of CD45+ cells in 10 fields (1000 × 1000 pixels) viewed with a 10× objective were counted to determine the number of cells per field.

For experiments in which more than two groups were compared, statistical significance was determined by one-way ANOVA using a 95% confidence interval and the Tukey posttest (Prism 4.0 for Macintosh, GraphPad Software, La Jolla, CA). Statistical error is expressed as SEM. For experiments in which two groups were compared, a t test was used.

Microvascular ECs are a natural target of KSHV infection. To study the immuno-evasive effects of K3 and K5 in this cell type, we transduced cultured HDMECs with lentivirus contructs encoding GFP (as control), K3, and K5; the K3 and K5 vectors also express GFP, but under the control of the ubiquitin promoter (Fig. 1A). Expression of K3 and K5 in transduced ECs was confirmed by immunostaining (Supplemental Figs. 2, 3). We then examined the effects of K3 and K5 overexpression in HDMECs on several surface proteins relevant for interactions with T cells and other leukocytes. K3 strongly downregulates MHC class I, whereas K5 has less effect on MHC class I but strongly downregulates ICAM-1 (CD54) (Fig. 1B). Both K3 and K5 downregulate ALCAM (CD166) and VCAM-1 in HDMECs to a similar extent, which indicates that the targets of these E3 ubiquitin ligases partly overlap. In contrast to previous studies that used immortalized HDMEC (7, 11), we find that K3 downregulates PECAM-1 much more effectively than K5, and that neither K5 nor K3 downregulates VE-cadherin (Fig. 1B, Supplemental Fig. 1, and data not shown).

FIGURE 1.

Phenotype of HDMEC transduced with lentivirus encoding GFP, KSHV gene K3 and KSHV gene K5 (K3 and K5 vectors have GFP driven by the ubiquitin promotor, K3 Em and K5 Em, respectively). A, Histograms showing GFP expression (thick line) compared with background fluorescence of nontransduced HDMEC (thin line). Greater than 95% of HDMEC are transduced by GFP, K3 Em, and K5 Em lentivirus. B, Expression of cell surface proteins on transduced HDMEC. K5 reduces ICAM-1 and MHC class I (HLA-A,B,C), whereas K3 predictably reduces MHC class I (HLA-A,B,C) but also severely reduces PECAM-1 as compared with K5. “Control” refers to staining with IgG. In all histogram plots shown, the x-axis scale is log10, and the y-axis scale is linear.

FIGURE 1.

Phenotype of HDMEC transduced with lentivirus encoding GFP, KSHV gene K3 and KSHV gene K5 (K3 and K5 vectors have GFP driven by the ubiquitin promotor, K3 Em and K5 Em, respectively). A, Histograms showing GFP expression (thick line) compared with background fluorescence of nontransduced HDMEC (thin line). Greater than 95% of HDMEC are transduced by GFP, K3 Em, and K5 Em lentivirus. B, Expression of cell surface proteins on transduced HDMEC. K5 reduces ICAM-1 and MHC class I (HLA-A,B,C), whereas K3 predictably reduces MHC class I (HLA-A,B,C) but also severely reduces PECAM-1 as compared with K5. “Control” refers to staining with IgG. In all histogram plots shown, the x-axis scale is log10, and the y-axis scale is linear.

Close modal

We then tested the effects of K3 and K5 overexpression on TEM of EM CD4+ T cells across HDMEC monolayers under conditions of venular flow (1 dyne/cm2). We established that there are two separate routes of TEM taken by EM CD4+ T cells, both of which require the application of venular levels of shear stress and TNF pretreatment of the HDMECs: 1) chemokine-dependent (e.g., IP-10 or SDF-1 supplied exogenously or produced by EC) TEM, which occurs within 15 min, and 2) TCR-dependent TEM, which prevents rapid chemokine responses but results in TEM with a slightly slower time course (within 60 min) (13, 14). To determine whether KSHV affects either of these two pathways, we generated CIITA plus K3 or K5 HDMEC cotransductants, which express class II MHC molecules, and overlaid a TNF-treated monolayer of these cells with TSST-1, which binds to MHC class II and is presented by the ECs to T cells expressing Vβ2 gene segment positive TCRs. The addition of exogenous chemokine was not necessary to stimulate TEM of EM CD4+ T cells, because of the high level of endogenous chemokine expression capable of stimulating spontaneous rapid (and pertussis toxin-sensitive) TEM on HDMECs, as compared with HUVEC (14). (It should be noted that although K3 blocks MHC class II expression in HDMECs in response to IFN-γ, probably via downregulation of the IFN-γ receptor, a known target of K3 [data not shown], MHC class II expression is unaffected by K3 or K5 in CIITA HDMEC.) EM CD4+ T cells that use Vβ2 gene segments to form their TCR respond to TSST-1 displayed by the CIITA HDMEC by undergoing slow TCR-dependent TEM; those EM CD4+ T cells in the same sample that do not use Vβ2 gene segments to form their TCR do not respond to TSST-1 and instead rapidly undergo TEM in response to endogenous chemokine (the Vβ2+ T cells are identified by staining with an anti-Vβ2 Ab, as described in 1Materials and Methods; T cells that do or do not stain with the anti-Vβ2 Ab are designated VB2+ and VB2−, respectively, in the figures). As previously noted, cells receiving a TCR signal will not rapidly transmigrate in response to chemokine, although they will do so if TSST-1 is excluded from the system (16). The TEM assay is conducted under conditions of venular flow and in both processes of TEM, the T cells must first adhere to the HDMEC monolayer. K5 transduction reduced the number of cells captured (Fig. 2A), but of those cells that did bind, the percentage found to transmigrate, whether by rapid chemokine-dependent or TCR-dependent mechanisms, was comparable to that seen in control HDMEC monolayers; immunostaining confirmed that TEM occurred at the junctions between EC expressing K5 (Fig. 2B, Supplemental Fig. 3). However, the reduced capture of T cells does result in a profound net reduction of T cells undergoing TEM in response to either chemokine or TCR signals. K3 transduction did not affect the number of cells captured under flow (Fig. 2A) or that undergo rapid, chemokine-dependent TEM, but did inhibit TCR-driven TEM (Fig. 2B).

FIGURE 2.

K5 reduces capture, and K3 reduces TCR-driven TEM, of EM CD4+ T cells. A, K5 reduces capture. EM CD4+ T cells captured in flow TEM assay across TNF-treated, TSST-1 preloaded CIITA HDMEC cotransduced with lentivirus encoding GFP (GFP CIITA), K3 Em (K3Em CIITA), K5 Em (K5Em CIITA), or not cotransduced (CIITA); mean ± SEM of one representative experiment of three is shown. K5Em CIITA binds 17% as many cells as controls expressing CIITA only. p < 0.001 between K5Em CIITA cotransductants and all other samples, and p > 0.05 between all other samples. B, K3 reduces TCR-driven TEM. Mean ± SEM of EM CD4+ T cell percent TEM in flow TEM assay across cells described above. VB2+ and VB2− refer to T cells that do or do not bind to TSST-1 on the CIITA HDMEC, respectively. One representative experiment of three is shown. ***p < 0.001; **p < 0.01.

FIGURE 2.

K5 reduces capture, and K3 reduces TCR-driven TEM, of EM CD4+ T cells. A, K5 reduces capture. EM CD4+ T cells captured in flow TEM assay across TNF-treated, TSST-1 preloaded CIITA HDMEC cotransduced with lentivirus encoding GFP (GFP CIITA), K3 Em (K3Em CIITA), K5 Em (K5Em CIITA), or not cotransduced (CIITA); mean ± SEM of one representative experiment of three is shown. K5Em CIITA binds 17% as many cells as controls expressing CIITA only. p < 0.001 between K5Em CIITA cotransductants and all other samples, and p > 0.05 between all other samples. B, K3 reduces TCR-driven TEM. Mean ± SEM of EM CD4+ T cell percent TEM in flow TEM assay across cells described above. VB2+ and VB2− refer to T cells that do or do not bind to TSST-1 on the CIITA HDMEC, respectively. One representative experiment of three is shown. ***p < 0.001; **p < 0.01.

Close modal

Our previous studies of TEM of EM CD4+ T cells had revealed that capture of EM CD4+ T cells to TNF-treated ECs could be fully recapitulated by ICAM-1 and VCAM-1 expressed on HUVEC (13). Because K5 but not K3 reduces T cell capture, and both K3 and K5 similarly reduce VCAM-1 whereas K5 much more profoundly reduces ICAM-1, we hypothesized that downregulation of ICAM-1 could be responsible for the reduced binding of EM CD4+ T cells to ECs transduced with K5. We investigated this possibility by examining the binding of these T cells to ECs treated with ICAM-1 blocking Ab or ICAM-1 siRNA. We included K3 transductants in this assay because two adhesion molecules, namely ALCAM and VCAM-1, were similarly reduced in K3 and K5 transductants (Fig. 1). Blocking Ab and strong reduction of ICAM-1 by one siRNA, but not a weaker downregulation by another siRNA (97% and 81% reduction by 741 and 747, respectively), inhibited capture of EM CD4+ T cells by TNF-treated HDMECs (Fig. 3A, 3B). Furthermore, binding of EM CD4+ T cells was rescued in K5 transductants engineered to express ICAM-1 lacking the cytoplasmic domain, and therefore not subject to ubiquitylation by K5 (Fig. 3C). These experiments support the hypothesis that the marked K5-mediated downregulation of ICAM-1 in HDMECs accounts for the reduced binding of EM CD4+ T cells, thereby diminishing the capacity for immunosurveillance by the CD4+ T cell population.

FIGURE 3.

K5 reduction of ICAM-1 reduces capture of EM CD4+ T cells. A, ICAM-1 blocking Ab reduces capture of EM CD4+ T cells to TNF-treated HDMEC. Blocking Ab to PECAM-1 or ICAM-1 was preincubated with TNF-treated HDMEC before 15-min flow assay. Mean ± SEM of the number of EM CD4+ T cells per field is shown. One representative experiment of three is shown. B, Strong knockdown of HDMEC ICAM-1 by siRNA inhibits capture of EM CD4+ T cells. TNF-treated HDMEC transfected with negative control siRNA (neg) and ICAM-1–specific siRNAs 741 or 747 were analyzed by FACS for ICAM-1 expression (histograms) or used in 15-min flow assays with EM CD4+ T cells (graph). Histograms show cells stained with isotype control IgG–FITC (filled) or anti–ICAM-1-FITC (line). Knockdown of ICAM-1 by 741 and 747 was consistently ~97% and 81% based on median fluorescence intensity, respectively, in three different experiments. “K5 neg” refers to K5 tranducatants transfected with negative control siRNA. The graph shows mean ± SEM of the number of EM CD4+ T cells per field. ***p < 0.001 compared with either nontransduced cells transfected with negative control siRNA or ICAM-1 747 siRNA; p > 0.05 for all other comparisons. C, Expression of ICAM-1 cytoplasmic deletion mutant in K5 transductants restores binding. Contour plots (left) of the EC used in the capture assay of EM CD4+ T cells on TNF-treated HDMECs transduced with K5 or GFP and transduced or not with ICAM-1 cytoplasmic deletion construct (ICAM). The numbers in the FACS plots refer to the median PE fluorescence intensity. The graph shows mean ± SEM of the number of cells per field from three separate samples combined, 10 fields per sample. p < 0.001 between K5 and all others; p > 0.05 between K5, ICAM and GFP, ICAM.

FIGURE 3.

K5 reduction of ICAM-1 reduces capture of EM CD4+ T cells. A, ICAM-1 blocking Ab reduces capture of EM CD4+ T cells to TNF-treated HDMEC. Blocking Ab to PECAM-1 or ICAM-1 was preincubated with TNF-treated HDMEC before 15-min flow assay. Mean ± SEM of the number of EM CD4+ T cells per field is shown. One representative experiment of three is shown. B, Strong knockdown of HDMEC ICAM-1 by siRNA inhibits capture of EM CD4+ T cells. TNF-treated HDMEC transfected with negative control siRNA (neg) and ICAM-1–specific siRNAs 741 or 747 were analyzed by FACS for ICAM-1 expression (histograms) or used in 15-min flow assays with EM CD4+ T cells (graph). Histograms show cells stained with isotype control IgG–FITC (filled) or anti–ICAM-1-FITC (line). Knockdown of ICAM-1 by 741 and 747 was consistently ~97% and 81% based on median fluorescence intensity, respectively, in three different experiments. “K5 neg” refers to K5 tranducatants transfected with negative control siRNA. The graph shows mean ± SEM of the number of EM CD4+ T cells per field. ***p < 0.001 compared with either nontransduced cells transfected with negative control siRNA or ICAM-1 747 siRNA; p > 0.05 for all other comparisons. C, Expression of ICAM-1 cytoplasmic deletion mutant in K5 transductants restores binding. Contour plots (left) of the EC used in the capture assay of EM CD4+ T cells on TNF-treated HDMECs transduced with K5 or GFP and transduced or not with ICAM-1 cytoplasmic deletion construct (ICAM). The numbers in the FACS plots refer to the median PE fluorescence intensity. The graph shows mean ± SEM of the number of cells per field from three separate samples combined, 10 fields per sample. p < 0.001 between K5 and all others; p > 0.05 between K5, ICAM and GFP, ICAM.

Close modal

PECAM-1 is important for TEM of neutrophils and monocytes, but has not been positively associated with the TEM of T cells (17, 18). However, such studies have focused on chemokine-driven T cell TEM, and the role of PECAM-1 in TCR-driven TEM has not previously been studied. To investigate whether the strong reduction of PECAM-1 by K3 could be responsible for the inhibition of TCR-dependent TEM, we treated CIITA HDMEC with anti–PECAM-1 blocking Ab (hec7) or PECAM-1 siRNA prior to TEM assays. Anti–PECAM-1 blocked TCR-driven TEM of EM CD4+ T cells, but had no effect on rapid TEM (Fig. 4A). CIITA HDMEC treated with two different siRNAs targeting PECAM-1 (P1 and P2) consistently reduced PECAM-1 expression by ~95% (like K3) and by ~80% (P1 and P2), respectively; Fig. 4B). PECAM-1 knockdown resulted in a dose response inhibition of TCR-driven TEM, but had no effect on rapid chemokine-dependent TEM (Fig. 4B). TCR-dependent TEM was rescued in K3 transductants by expressing a PECAM-1 mutated to escape targeted destruction by K3 (containing 11 arginines rather than lysines, as well as one alanine rather than cysteine, in the cytoplasmic domain; Fig. 4C). These results suggest that the strong ubiquitin-dependent reduction of PECAM-1 by K3 is sufficient to inhibit TCR-driven TEM, and they provide another example of how KSHV immunomodulatory genes can function to evade the host response. These results also identify PECAM-1 as a key EC molecule that is used by EM CD4+ T cells specifically in TCR-driven, but not rapid chemokine-dependent, TEM, further differentiating these processes.

FIGURE 4.

PECAM-1 dependence of TCR-driven TEM of EM CD4+ T cells. A, PECAM-1 blocking Ab blocks TCR-driven TEM. Blocking Ab to PECAM-1 (hec7) or control IgG was preincubated with TNF-treated, TSST-1 preloaded CIITA HDMEC before flow TEM assays. Graphs show mean ± SEM of percent TEM of Vβ2TCR− (left) and Vβ2TCR+ (right) EM CD4+ T cells. The left graph shows percent TEM at 15 min. B, Knockdown of HDMEC PECAM-1 by siRNA inhibits TCR-driven TEM of EM CD4+ T cells. TNF-treated CIITA HDMEC transfected with negative control (neg), P1, or P2 siRNA were analyzed by FACS for PECAM-1 expression (upper panels) or used in flow TEM assays with EM CD4+ T cells (lower panels). Upper panels show results of one representative experiment of cells stained with isotype control IgG–FITC (filled) or anti-PECAM–FITC (line). Knockdown of PECAM-1 by P1 and P2 was consistently ~95% and 80% for P1 and P2, respectively, in three different experiments. Bottom panels show mean ± SEM of data from three different experiments. The lower left panel shows percent TEM of the Vβ2− cells at 15 min of flow, and the lower right panel shows percent TEM of Vβ2+ cells at 15 and 60 min. ***p < 0.001. C, Expression of PECAM-1 without K3 target residues restores TCR-driven TEM to K3-transduced CIITA HDMEC. TEM assay of EM CD4+ T cells on CIITA HDMEC transduced to express K3 alone or in combination with PECAM-1 lacking cytoplasmic domain lysines and cysteines (12 mut) or nonmutant PECAM-1 (PECAM wt). Contour plots show FACS analysis of cells stained with PE-conjugated anti–PECAM-1. Graph shows mean ± SEM percent TEM of Vβ2TCR+ cells after 60 min flow, combined data from two separate experiments in duplicate. p < 0.001 for comparisons between all data sets except CIITA, K3Em versus CIITA, K3Em, and PECAM wt (p < 0.05). Mean fluorescence intensity of cells stained with PE-conjugated anti-PECAM: CIITA = 1028; CIITA, K3Em = 274: CIITA, K3Em, PECAM 12 mut = 641; CIITA, K3Em, PECAM wt = 276.

FIGURE 4.

PECAM-1 dependence of TCR-driven TEM of EM CD4+ T cells. A, PECAM-1 blocking Ab blocks TCR-driven TEM. Blocking Ab to PECAM-1 (hec7) or control IgG was preincubated with TNF-treated, TSST-1 preloaded CIITA HDMEC before flow TEM assays. Graphs show mean ± SEM of percent TEM of Vβ2TCR− (left) and Vβ2TCR+ (right) EM CD4+ T cells. The left graph shows percent TEM at 15 min. B, Knockdown of HDMEC PECAM-1 by siRNA inhibits TCR-driven TEM of EM CD4+ T cells. TNF-treated CIITA HDMEC transfected with negative control (neg), P1, or P2 siRNA were analyzed by FACS for PECAM-1 expression (upper panels) or used in flow TEM assays with EM CD4+ T cells (lower panels). Upper panels show results of one representative experiment of cells stained with isotype control IgG–FITC (filled) or anti-PECAM–FITC (line). Knockdown of PECAM-1 by P1 and P2 was consistently ~95% and 80% for P1 and P2, respectively, in three different experiments. Bottom panels show mean ± SEM of data from three different experiments. The lower left panel shows percent TEM of the Vβ2− cells at 15 min of flow, and the lower right panel shows percent TEM of Vβ2+ cells at 15 and 60 min. ***p < 0.001. C, Expression of PECAM-1 without K3 target residues restores TCR-driven TEM to K3-transduced CIITA HDMEC. TEM assay of EM CD4+ T cells on CIITA HDMEC transduced to express K3 alone or in combination with PECAM-1 lacking cytoplasmic domain lysines and cysteines (12 mut) or nonmutant PECAM-1 (PECAM wt). Contour plots show FACS analysis of cells stained with PE-conjugated anti–PECAM-1. Graph shows mean ± SEM percent TEM of Vβ2TCR+ cells after 60 min flow, combined data from two separate experiments in duplicate. p < 0.001 for comparisons between all data sets except CIITA, K3Em versus CIITA, K3Em, and PECAM wt (p < 0.05). Mean fluorescence intensity of cells stained with PE-conjugated anti-PECAM: CIITA = 1028; CIITA, K3Em = 274: CIITA, K3Em, PECAM 12 mut = 641; CIITA, K3Em, PECAM wt = 276.

Close modal

The role of PECAM-1 in TEM of EM CD4+ T cells triggered by TCR signaling was unexpected. PECAM-1 is a ligand for several molecules on leukocytes, including PECAM-1 itself, ADP-ribose cyclase (CD38), and CD177 (1921). However, PECAM-1, CD38, and CD177 are not expressed on EM CD4+ T cells (Fig. 5A). PECAM-1 does not bind directly to glycosaminoglycans, but PECAM-1 engagement has been reported to trigger heparan- or chondroitin-like glycosaminoglycan binding in PECAM-1 transfected cells (2225). We therefore tested whether TCR-driven TEM of EM CD4+ T cells involved HDMEC binding of a T cell-expressed glycosaminoglycan. To do so, we either incubated the ECs with a variety of glycosaminoglycans to block the putative binding site, or we enzymatically removed various glycosaminoglycans from the T cells with specific glycosaminoglycan degrading enzymes prior to TEM assays. TCR-dependent TEM (but not chemokine-dependent TEM) of EM CD4+ T cells was blocked by EC treatment with chondroitin and heparan, but not hyaluronate, or by T cell treatment with chondroitinase but not hyaluronidase (Fig. 5B). Although these data do not identify a specific PECAM-1 ligand, they support the idea that PECAM-1 engagement on HDMECs is functional in TCR-initiated TEM, as first suggested by the effects of K3.

FIGURE 5.

A, EM CD4+ T cells do not express known ligands of PECAM-1. Histograms showing total CD4+ T cells (top row) and EM CD4+ T cells (bottom row) stained with isotype control (filled peak) or stained with FITC-conjugated anti-CD31, CD38, and CD177 (solid lines). The x-axis scale is log10, and the y-axis scale is linear. B, Dependence of TCR-driven TEM on an EM CD4+ T cell glycosaminoglycan. CIITA HDMECs were preincubated with vehicle, heparan sulfate, chondroitin-6-sulfate, or hyaluronic acid, or T cells were pretreated with chondroitinase or hyaluronidase prior to 15 or 60 min flow TEM assay. Mean ± SEM of percent TEM from one representative experiment of three is shown. ***p < 0.001 compared with vehicle; *p < 0.05 compared with vehicle.

FIGURE 5.

A, EM CD4+ T cells do not express known ligands of PECAM-1. Histograms showing total CD4+ T cells (top row) and EM CD4+ T cells (bottom row) stained with isotype control (filled peak) or stained with FITC-conjugated anti-CD31, CD38, and CD177 (solid lines). The x-axis scale is log10, and the y-axis scale is linear. B, Dependence of TCR-driven TEM on an EM CD4+ T cell glycosaminoglycan. CIITA HDMECs were preincubated with vehicle, heparan sulfate, chondroitin-6-sulfate, or hyaluronic acid, or T cells were pretreated with chondroitinase or hyaluronidase prior to 15 or 60 min flow TEM assay. Mean ± SEM of percent TEM from one representative experiment of three is shown. ***p < 0.001 compared with vehicle; *p < 0.05 compared with vehicle.

Close modal

The focus of the current study was to determine the effect of expressing KSHV immune-modulatory proteins K3 and K5 in microvascular cells on the interactions of these cells with EM CD4+ T cells. Functionally, K5 and K3 expression each inhibit the ability of HDMECs to recruit EM CD4+ T cells, but these viral proteins act at different steps, specifically by inhibiting T cell capture under flow for K5 and by inhibiting TCR-driven TEM by K3. By probing the targets of K5 and K3 in blocking Ab and siRNA experiments, as well as by expression of proteins mutated to avoid ubiquitin-dependent proteolysis, we attribute the reduced capture of EM CD4+ T cells to the strong downregulation of ICAM-1 by K5. In addition, we identify PECAM-1 as a key molecule in TCR-dependent TEM, which is potently targeted by K3 in untransformed HDMECs.

The reduced binding of EM CD4+ T cells to K5 transductants is not surprising, since it has previously been reported to reduce ICAM-1 expression in other cell types and ICAM-1 is the principal ligand for LFA-1 on T cells. In HUVECs, we had found that VCAM-1 could also support T cell capture. However, HDMECs express lower quantities of VCAM-1 than do HUVECs, and our siRNA experiments support a nonredundant role for ICAM-1 expression in this cell type. In contrast, the finding that PECAM-1 plays a role in TCR-dependent TEM was not anticipated because PECAM-1, which is required for monocyte and neutrophil TEM, had been shown previously to be unimportant for the TEM of T cells (17, 18). However, prior experiments focused on chemokine-driven responses. The requirement for PECAM-1 in EM CD4+ T cell recruitment further distinguishes TCR-dependent TEM from rapid chemokine-mediated TEM. EM CD4+ T cells do not express any of the known counter-receptors for PECAM-1. Spurred by previous reports suggesting that PECAM-1 engagement can promote binding of heparan- or chondroitin-like glycosaminoglycans, we further demonstrate that such molecules can also participate in a nonredundant way in the TCR-driven TEM of EM CD4+ T cells. Therefore, our data suggest a role for three previously unidentified molecules in this process: a novel counter-receptor for PECAM-1 on T cells, a glycosaminoglycan-containing proteoglycan on T cells, and a glycosaminoglycan receptor on EC. Further study is needed to identify the specific molecules involved, but it is interesting to note that the glycosaminoglycan inhibition profile (heparan and chondroitin inhibit, hyaluronate does not) resembles that shown earlier for L cells, although the treatment with chondroitinase ABC did not have as robust an effect as inhibition by the competitor (23). This may be due to incomplete digestion of glycosaminoglycan on the T cells by chondroitinase ABC, which leaves a stub of chondroitin sulfate. Residual binding of chondroitinase ABC treated neurocan to transmembrane protein tyrosine phosphatase σ, a recently described receptor for chondroitin sulfate proteoglycan, has similarly been observed (26). One additional characteristic of the PECAM-1 ligand on EM CD4+ T cells that can be inferred from these studies is that, because hec7 binds to and blocks the N terminus of PECAM-1 previously implicated in homophilic adhesion (IgG domains 1 and/or 2), the receptor on EM CD4+ T cells can bind to a similar region, unlike the previously described heterophilic counter-receptors that bind to IgG domain 6 of PECAM-1 (27).

The strong downregulation of PECAM-1 by K3 in HDMEC was unexpected, since previous work using immortalized EC had instead identified PECAM-1 as a K5 target. Indeed, we have found in the course of these studies that the functions of K3 and K5 are cell type-specific. This is particularly evident in comparing various primary EC cells and EC cell lines. For example, we noticed that HUVECs transduced with K5 but not K3 stopped proliferating, whereas K5-transduced HDMECs proliferated at least as well if not better than nontransduced HDMECs or K3 and GFP transductants, and proliferation of the immortalized EaHy.926 cell line is unaffected by K5 (T.D. Manes, unpublished observations). We believe that we are the first investigators to assess the effects of K3 and K5 in untransformed microvascular ECs, the natural target of KSHV. The nature of our HDMECs also warrants some consideration. Freshly isolated HDMECs are a mixture of vascular and lymphatic ECs that converge to a common phenotype with both vascular and lymphatic features in cell culture. KS cells also display both vascular and lymphatic characteristics, possibly reflecting the common physiologic origin of these EC types, and the actual target of infection in vivo is a matter of some controversy (28, 29). We also note that we have studied K3 and K5 in cells not infected by KSHV. It is possible that other KSHV genes will modulate the actions of these gene products. Furthermore, there is no simple way to assess whether the levels of K3 or K5 expressed in our system are appropriate to those found in infected cells, but the lentivirus vectors we used tend to produce lower, more physiologic levels of expression than did adenovirus or plasmids. Despite this caveat, the reductionist approach used in our study is useful for understanding the specific properties of these viral proteins.

Our results further extend the list of mechanisms by which KSHV acts to evade the human immune system to include inhibition of T cell recruitment to an infected site. Analysis of these effects has revealed further differences between TCR-driven and chemokine-driven TEM of EM CD4+ T cells, which can aid in developing effective treatment for not just KS but a variety of conditions involving EM CD4+ T cells, such as autoimmunity, atherosclerosis, and allograft rejection.

We thank William Sessa for use of the immunofluorescence microscope, Louise Benson, Gwendoline Davis, and Lisa Gras for assistance in cell culture, and Klaus Früh (Oregon Health and Science University, Portland, OR) for the K5 mAb.

Disclosures The authors have no financial conflicts of interests.

This work was supported by National Institutes of Health Grants P01-HL070295 and HL51014 (to J.S.P. and T.D.M.), R01 HL046849 and R37 HL064774 (to W.A.M.), and by the Wellcome Trust (to S.H. and P.J.L.).

The online version of this article contains supplemental material.

Abbreviations used in this paper:

ALCAM

activated leukocyte cell adhesion molecule

EC

endothelial cell

EM

effector memory

HDMEC

human dermal microvascular endothelial cell

K3

ORFK3

K5

ORFK5

KS

Kaposi’s sarcoma

KSHV

Kaposi’s sarcoma-associated herpesvirus

mut

mutant

neg

negative control

siRNA

small interfering RNA

TEM

transendothelial migration

VE-cadherin

vascular endothelial cadherin

wt

wild-type.

1
Coscoy
L.
2007
.
Immune evasion by Kaposi’s sarcoma-associated herpesvirus.
Nat. Rev. Immunol.
7
:
391
401
.
2
Coscoy
L.
,
Ganem
D.
.
2000
.
Kaposi’s sarcoma-associated herpesvirus encodes two proteins that block cell surface display of MHC class I chains by enhancing their endocytosis.
Proc. Natl. Acad. Sci. U.S.A.
97
:
8051
8056
.
3
Stevenson
P. G.
,
Efstathiou
S.
,
Doherty
P. C.
,
Lehner
P. J.
.
2000
.
Inhibition of MHC class I-restricted antigen presentation by gamma 2-herpesviruses.
Proc. Natl. Acad. Sci. U.S.A.
97
:
8455
8460
.
4
Ishido
S.
,
Choi
J. K.
,
Lee
B. S.
,
Wang
C.
,
DeMaria
M.
,
Johnson
R. P.
,
Cohen
G. B.
,
Jung
J. U.
.
2000
.
Inhibition of natural killer cell-mediated cytotoxicity by Kaposi’s sarcoma-associated herpesvirus K5 protein.
Immunity
13
:
365
374
.
5
Coscoy
L.
,
Ganem
D.
.
2001
.
A viral protein that selectively downregulates ICAM-1 and B7-2 and modulates T cell costimulation.
J. Clin. Invest.
107
:
1599
1606
.
6
Sanchez
D. J.
,
Gumperz
J. E.
,
Ganem
D.
.
2005
.
Regulation of CD1d expression and function by a herpesvirus infection.
J. Clin. Invest.
115
:
1369
1378
.
7
Mansouri
M.
,
Douglas
J.
,
Rose
P. P.
,
Gouveia
K.
,
Thomas
G.
,
Means
R. E.
,
Moses
A. V.
,
Früh
K.
.
2006
.
Kaposi sarcoma herpesvirus K5 removes CD31/PECAM from endothelial cells.
Blood
108
:
1932
1940
.
8
Bartee
E.
,
McCormack
A.
,
Früh
K.
.
2006
.
Quantitative membrane proteomics reveals new cellular targets of viral immune modulators.
PLoS Pathog.
2
:
e107
.
9
Li
Q.
,
Means
R.
,
Lang
S.
,
Jung
J. U.
.
2007
.
Downregulation of gamma interferon receptor 1 by Kaposi’s sarcoma-associated herpesvirus K3 and K5.
J. Virol.
81
:
2117
2127
.
9
Li
Q.
,
Means
R.
,
Lang
S.
,
Jung
J. U.
.
2007
.
Downregulation of gamma interferon receptor 1 by Kaposi’s sarcoma-associated herpesvirus K3 and K5.
J. Virol.
81
:
2117
2127
.
10
Thomas
M.
,
Boname
J. M.
,
Field
S.
,
Nejentsev
S.
,
Salio
M.
,
Cerundolo
V.
,
Wills
M.
,
Lehner
P. J.
.
2008
.
Down-regulation of NKG2D and NKp80 ligands by Kaposi’s sarcoma-associated herpesvirus K5 protects against NK cell cytotoxicity.
Proc. Natl. Acad. Sci. U.S.A.
105
:
1656
1661
.
11
Mansouri
M.
,
Rose
P. P.
,
Moses
A. V.
,
Früh
K.
.
2008
.
Remodeling of endothelial adherens junctions by Kaposi’s sarcoma-associated herpesvirus.
J. Virol.
82
:
9615
9628
.
12
Tomescu
C.
,
Law
W. K.
,
Kedes
D. H.
.
2003
.
Surface downregulation of major histocompatibility complex class I, PE-CAM, and ICAM-1 following de novo infection of endothelial cells with Kaposi’s sarcoma-associated herpesvirus.
J. Virol.
77
:
9669
9684
.
13
Manes
T. D.
,
Pober
J. S.
,
Kluger
M. S.
.
2006
.
Endothelial cell-T lymphocyte interactions: IP[corrected]-10 stimulates rapid transendothelial migration of human effort but not central memory CD4+ T cells. Requirements for shear stress and adhesion molecules.
Transplantation
82
(
Suppl 1
)
S9
S14
.
14
Manes
T. D.
,
Pober
J. S.
.
2008
.
Antigen presentation by human microvascular endothelial cells triggers ICAM-1-dependent transendothelial protrusion by, and fractalkine-dependent transendothelial migration of, effector memory CD4+ T cells.
J. Immunol.
180
:
8386
8392
.
15
Clark
P. R.
,
Manes
T. D.
,
Pober
J. S.
,
Kluger
M. S.
.
2007
.
Increased ICAM-1 expression causes endothelial cell leakiness, cytoskeletal reorganization and junctional alterations.
J. Invest. Dermatol.
127
:
762
774
.
16
Manes
T. D.
,
Shiao
S. L.
,
Dengler
T. J.
,
Pober
J. S.
.
2007
.
TCR signaling antagonizes rapid IP-10-mediated transendothelial migration of effector memory CD4+ T cells.
J. Immunol.
178
:
3237
3243
.
17
Muller
W. A.
2009
.
Mechanisms of transendothelial migration of leukocytes.
Circ. Res.
105
:
223
230
.
18
Bird
I. N.
,
Spragg
J. H.
,
Ager
A.
,
Matthews
N.
.
1993
.
Studies of lymphocyte transendothelial migration: analysis of migrated cell phenotypes with regard to CD31 (PECAM-1), CD45RA and CD45RO.
Immunology
80
:
553
560
.
19
Sun
Q. H.
,
DeLisser
H. M.
,
Zukowski
M. M.
,
Paddock
C.
,
Albelda
S. M.
,
Newman
P. J.
.
1996
.
Individually distinct Ig homology domains in PECAM-1 regulate homophilic binding and modulate receptor affinity.
J. Biol. Chem.
271
:
11090
11098
.
20
Deaglio
S.
,
Morra
M.
,
Mallone
R.
,
Ausiello
C. M.
,
Prager
E.
,
Garbarino
G.
,
Dianzani
U.
,
Stockinger
H.
,
Malavasi
F.
.
1998
.
Human CD38 (ADP-ribosyl cyclase) is a counter-receptor of CD31, an Ig superfamily member.
J. Immunol.
160
:
395
402
.
21
Sachs
U. J.
,
Andrei-Selmer
C. L.
,
Maniar
A.
,
Weiss
T.
,
Paddock
C.
,
Orlova
V. V.
,
Choi
E. Y.
,
Newman
P. J.
,
Preissner
K. T.
,
Chavakis
T.
,
Santoso
S.
.
2007
.
The neutrophil-specific antigen CD177 is a counter-receptor for platelet endothelial cell adhesion molecule-1 (CD31).
J. Biol. Chem.
282
:
23603
23612
.
22
Muller
W. A.
,
Berman
M. E.
,
Newman
P. J.
,
DeLisser
H. M.
,
Albelda
S. M.
.
1992
.
A heterophilic adhesion mechanism for platelet/endothelial cell adhesion molecule 1 (CD31).
J. Exp. Med.
175
:
1401
1404
.
23
DeLisser
H. M.
,
Yan
H. C.
,
Newman
P. J.
,
Muller
W. A.
,
Buck
C. A.
,
Albelda
S. M.
.
1993
.
Platelet/endothelial cell adhesion molecule-1 (CD31)-mediated cellular aggregation involves cell surface glycosaminoglycans.
J. Biol. Chem.
268
:
16037
16046
.
24
Sun
Q. H.
,
Paddock
C.
,
Visentin
G. P.
,
Zukowski
M. M.
,
Muller
W. A.
,
Newman
P. J.
.
1998
.
Cell surface glycosaminoglycans do not serve as ligands for PECAM-1. PECAM-1 is not a heparin-binding protein.
J. Biol. Chem.
273
:
11483
11490
.
25
Newman
P. J.
1997
.
The biology of PECAM-1.
J. Clin. Invest.
99
:
3
8
.
26
Shen
Y.
,
Tenney
A. P.
,
Busch
S. A.
,
Horn
K. P.
,
Cuascut
F. X.
,
Liu
K.
,
He
Z.
,
Silver
J.
,
Flanagan
J. G.
.
2009
.
PTPsigma is a receptor for chondroitin sulfate proteoglycan, an inhibitor of neural regeneration.
Science
326
:
592
596
.
27
Liao
F.
,
Huynh
H. K.
,
Eiroa
A.
,
Greene
T.
,
Polizzi
E.
,
Muller
W. A.
.
1995
.
Migration of monocytes across endothelium and passage through extracellular matrix involve separate molecular domains of PECAM-1.
J. Exp. Med.
182
:
1337
1343
.
28
Hong
Y. K.
,
Foreman
K.
,
Shin
J. W.
,
Hirakawa
S.
,
Curry
C. L.
,
Sage
D. R.
,
Libermann
T.
,
Dezube
B. J.
,
Fingeroth
J. D.
,
Detmar
M.
.
2004
.
Lymphatic reprogramming of blood vascular endothelium by Kaposi sarcoma-associated herpesvirus.
Nat. Genet.
36
:
683
685
.
29
Wang
H. W.
,
Trotter
M. W.
,
Lagos
D.
,
Bourboulia
D.
,
Henderson
S.
,
Mäkinen
T.
,
Elliman
S.
,
Flanagan
A. M.
,
Alitalo
K.
,
Boshoff
C.
.
2004
.
Kaposi sarcoma herpesvirus-induced cellular reprogramming contributes to the lymphatic endothelial gene expression in Kaposi sarcoma.
Nat. Genet.
36
:
687
693
.