Effective Human Herpesvirus 8 Infection of Human Umbilical Vein Endothelial Cells by Cell-Mediated Transmission

Abstract

Cell-free transmission of human herpesvirus 8 (HHV-8) to human cells in vitro has been reported to be difficult, if not impossible. The present experiments were conducted with the idea that cell-cell contact may produce much more effective transmission, so-called cell-mediated transmission. Primary human umbilical vein endothelial cells (HUVECs) were cocultured with an HHV-8-infected lymphoma cell line, BCBL-1 cells. When a ratio of 12-O-tetradecanoylphorbol-13-acetate (TPA)-treated BCBL-1 cells to HUVECs of 10:1 was used, more than 20% of HUVECs were found to express the HHV-8 latency-associated nuclear antigen (LANA) 48 h after the start of coculturing; this value increased to more than 30% after 72 h. HHV-8-encoded ORF26, K8, K8.1, K10, K11, ORF59, and ORF65 proteins were not detected in these HHV-8-infected HUVECs until 72 h. The HHV-8 antigens were not observed in HUVECs cocultured with TPA-treated BCBL-1 cells separated by a membrane. Thirty days after removal of the BCBL-1 cells from the cell-mediated transmission experiment, the HUVECs still expressed LANA and the HHV-8 genome was detected by PCR in these cells. Moreover, the ORF59 protein, a DNA replication-associated protein of HHV-8, was expressed in such HUVECs in the presence of TPA stimulation. These results indicated a far more effective transmission mechanism, cell-cell contact, suggesting the possibility that such a mechanism works in vivo.

Human herpesvirus 8 (HHV-8) is associated with Kaposi's sarcoma (KS), primary effusion lymphoma (PEL), and a subset of multicentric Castleman's disease (2, 19, 28). Serological examinations have revealed that almost all KS patients have anti-HHV-8 antibodies (3, 12, 22, 30). Immunohistochemical studies have directly demonstrated that the proliferating spindle-shaped cells of KS lesions express a latency-associated nuclear antigen (LANA) (6, 15, 16, 21), strongly suggesting that HHV-8 is the pathogenic agent of KS.

While there are some controversies about the origin of the spindle-shaped cells, endothelial cells appear to be the primary candidate, as inferred from the expression of endothelium-specific molecules (1, 4, 11, 25–27, 29). Thus, attempts to transmit HHV-8 to endothelial cells have been conducted (7, 8, 18, 20, 23). In general, there are two modes of viral transmission: direct contact between target and provider cells (so-called cell-mediated transmission) and cell-free transmission. In previously reported HHV-8 infection experiments, cell-free transmission has been investigated; no detailed work on cell-mediated transmission has been reported. Meanwhile, cell-free transmission was reported for human B cells in vivo in SCID-hu mice (5) and in vitro assays. However, viral transmission to endothelial cells with a cell-free supernatant was reported to be far more difficult (23).

Flore et al. demonstrated that primary human umbilical vein endothelial cells (HUVECs) can be infected with HHV-8 using purified viral particles (7). They prepared purified and concentrated viral particles from the supernatant of the 12-O-tetradecanoylphorbol-13-acetate (TPA)-treated BC-3 cell line, an HHV-8-positive- and Epstein-Barr virus-negative PEL cell line, and added the medium of HUVECs at 5 to 10 genome equivalents per cell. Moreover, they also reported that HHV-8 infection caused long-term proliferation and survival of these cells (7). Moses et al. also succeeded in transmitting HHV-8 to dermal microvascular endothelial cells transfected with human papillomavirus (HPV) E6 and E7 genes by exposing these cells to the nonconcentrated culture supernatant of BCBL-1 cells, another HHV-8-positive PEL cell line (18, 24). Dermal microvascular endothelial cells thus infected with HHV-8 were transformed, lost contact inhibition, and proliferated in soft agar (18). Another group succeeded in transmitting HHV-8 to human neonatal brain endothelial cells using a highly concentrated suspension of HHV-8 particles derived from the culture supernatant of BCBL-1 cells (23). However, these conditions were too artificial and do not appear to reflect the conditions occurring in vivo because of the use of highly concentrated viral particles or transformed target cells.

To our knowledge, neither data on cell-cell contact transmission of HHV-8 in vitro nor evidence for the presence of cell-mediated transmission in vivo has ever been reported. Thus, in this study, we attempted to transmit HHV-8 to human endothelial cells by cell-cell contact using primary cultures of HUVECs and BCBL-1 cells.

Transmission of HHV-8 to HUVECs by coculturing with BCBL-1 cells.

To determine the possibility of HHV-8 infection, we cocultured HUVECs and TPA-treated BCBL-1 cells. TPA treatment is known to increase the production of HHV-8 particles (24). HUVECs were obtained from healthy donors with their informed consent and were cultured in an RPMI 1640-based conditioned medium containing 10% fetal calf serum and 30 μg of endothelial cell growth supplement (Upstate Biotechnology, Lake Placid, N.Y.)/ml on chamber glass slides coated with 1.5% gelatin (Wako Chemicals, Osaka, Japan). BCBL-1 cells pretreated with 20 ng of TPA/ml for 48 h were then added to the HUVEC culture at a ratio of 10:1 for cocultivation. Virus antigens expressed in HUVECs were investigated by an immunofluorescence assay (IFA) using rabbit polyclonal antibodies against LANA (ORF73), which is expressed in the latent phase of HHV-8 infection. We also investigated the expression of von Willebrand factor (VWF; factor VIII-related antigen), a marker of HUVECs, and CD45, a leukocyte common antigen (LCA), using mouse monoclonal antibodies (Dako, Copenhagen, Denmark) to confirm the transmission of HHV-8 to HUVECs.

For the IFA, chamber slides were washed thoroughly with phosphate-buffered saline (PBS) three times to remove BCBL-1 cells. The remaining, adherent cells were fixed in 4% paraformaldehyde–PBS for 10 min, permeabilized with 0.5% Triton X-100–PBS for 20 min, blocked with 2% bovine serum albumin in PBS for 60 min at room temperature, and incubated with a primary antibody at appropriate dilutions for 60 min at 37°C. Alexa-488-conjugated goat anti-mouse immunoglobulin G (IgG) and Alexa-568-conjugated goat anti-rabbit IgG (Molecular Probes, Eugene, Oreg.) were applied as secondary antibodies for 30 min at 37°C. Imaging was performed using a confocal microscope equipped with an argon-krypton laser (LSM-MicroSystem; Carl-Zeiss, Jena, Germany). The emission patterns of two types of fluorescence were collected separately and overlapped using a computer to form two-color images. Antigen expression was determined 0, 6, 12, 18, 24, 48, and 72 h after the start of coculturing.

The IFA revealed that all adherent cells on the chamber slides, after coculturing with BCBL-1 cells for 24 h and washing, expressed VWF, and some of the cells expressed both VWF in the cytoplasm and LANA in the nucleus (Fig. 1A). We also performed double staining of LANA and LCA; however, LANA-positive cells were negative for LCA (Fig. 1B). On the other hand, centrifuged BCBL-1 cells were stained by the anti-LCA antibody but not by the anti-VWF antibody (Fig. 1C and D). These data indicate that almost all adherent cells on the chamber slides were HUVECs and that some of the HUVECs were infected with HHV-8 in this system.

FIG. 1.

Immunodetection of HHV-8-encoded proteins. (A and B) Adherent cells on chamber slides on which HUVECs were cocultured with TPA-treated BCBL-1 cells. All adherent cells express VWF (in panel A, Alexa 488, green), and some of them express LANA (in panels A and B, Alexa 568, red). These adherent cells do not express LCA (in panel B, Alexa 488, green). (C and D) Centrifuged BCBL-1 cells. BCBL-1 cells express LANA (in panels C and D, Alexa 568, red) and LCA (in panel D, Alexa 488, green) but not VWF (in panel C, Alexa 468, green). (E to I) Expression of LANA (in panels E to G, Alexa 488, green) and vIL-6 (in panels H and I, Alexa 488, green) in HUVECs cocultured with BCBL-1 cells. Propidium iodide (red) was used as a counterstaining agent, and yellow indicates overlap (E to I). (J) No staining of LANA (Alexa-488, green) in HEp-2 cells cocultured with TPA-treated BCBL-1 cells. (K) LANA expression in HUVECs cocultured with BCBL-1 cells 30 days after infection. Green indicates Alexa-488 staining of LANA, and red indicates Alexa-568 staining of VWF. (L) ORF59 protein expression in TPA-treated HUVECs cocultured with BCBL-1 cells 30 days after infection. Green indicates Alexa-488 staining of ORF59 protein, and red indicates Alexa-568 staining of VWF.

Expression of viral antigens in HUVECs.

We investigated the time course of viral antigen expression using rabbit polyclonal antibodies against K2 (viral interleukin-6 [vIL-6]), ORF26 (capsid protein [CP]), K8, K8.1, K10, K11, ORF59 (processivity factor [PF8]), ORF65 (viral minor CP [vMCP]), and LANA in the cocultured HUVECs. All of these antibodies were established by our colleagues and us (13–15). The Alexa-488-conjugated anti-rabbit IgG antibody was used as the secondary antibody. Nuclear counterstaining was performed with propidium iodide (0.5 mg/ml). For the determination of the percentage of stained cells, several photographs were taken blindly at a low magnification (×4) with a confocal microscope, and the numbers of positively and negatively staining cells in each photograph were counted. The positivity rates were determined from the averages of three experiments (Fig. 2). The IFA revealed that LANA was detected 6 h after the start of coculturing at a positivity rate of less than 1%. The number of LANA-positive cells increased over time to 30% in 72 h (Fig. 1E to G and Fig. 2A). However, viral antigens other than LANA and vIL-6 were not detected within this period (data not shown). Anti-vIL-6-antibody-reactive cells were observed 12 h after the start of coculturing in a small population (less than 0.1%), and the positivity rate did not change by 72 h after the start of coculturing (Fig. 1H and I). No stained cells were detected in identically treated HEp-2 cells, an epithelial cell line derived from a laryngeal epidermoid carcinoma (Fig. 1J). These data show that at least 6 h is required for the expression of viral antigens in cocultures of primary cultured HUVECs with BCBL-1 cells and that the latent phase of infection is the predominant form.

FIG. 2.

Percentages of LANA-positive HUVECs cocultured with BCBL-1 cells. Vertical bars indicate standard deviations from three experiments. (A) Time course experiment. Solid, broken, and gray lines indicate the LANA positivity rates in HUVECs with cell-mediated transmission using TPA-stimulated BCBL-1 cells and nonstimulated BCBL-1 cells and cell-free transmission using TPA-stimulated BCBL-1 cells, respectively. BCBL-1 cell/HUVEC ratio, 10/1. (B) Titration of BCBL-1 cells and HUVECs. The ratio of BCBL-1 cells to HUVECs varied from 10⁻⁴ to 10¹. Solid, broken, and gray lines indicate culturing times of 48, 24, and 0 h, respectively.

Efficiency of HHV-8 transmission to HUVECs depends on the number of TPA-treated BCBL-1 cells.

HUVECs were cocultured with BCBL-1 cells, pretreated with or without TPA, for 72 h. HHV-8 infection was monitored by immunostaining with the anti-LANA antibody. BCBL-1 cells pretreated with TPA showed 7- to 10-fold the infectivity of untreated cells (Fig. 2A). When the ratio of the number of TPA-treated BCBL-1 cells to that of HUVECs was changed from 1:10,000 to 10:1, HHV-8 infection occurred at a ratio of 1:10 or more (Fig. 2B). These data suggest that the efficiency of HHV-8 transmission to HUVECs depends on the number of TPA-treated BCBL-1 cells.

Cell-cell contact is important in the present transmission system.

To evaluate the role of direct contact between provider and target cells, we used the filtered culture supernatant obtained from TPA-pretreated BCBL-1 cells as the source of a cell-free viral supernatant. A Transwell pore membrane system (pore size, 0.4 μm; Corning-Costar, Cambridge, Mass.) was used for this purpose. This experiment was performed with a BCBL-1 cell/HUVEC ratio of 10:1. We confirmed the presence of viral particles in the culture supernatant of BCBL-1 cells by negative-stain electron microscopy and PCR analysis (data not shown). In the IFA, no LANA-positive cells were detected after coculturing with BCBL-1 cells separated by a membrane until 72 h, whereas more than 20% of HUVECs were found positive for LANA by cell-mediated transmission at 48 h (Fig. 2A). To confirm the results, adherent cells were collected after thorough washing and used for DNA extraction. PCR analysis targeting KS330233 (2) revealed that HHV-8 DNA was detected in HUVECs cocultured with BCBL-1 cells by cell-mediated transmission but not by cell-free transmission (Fig. 3). These results indicate that a direct interaction between provider and target cells is required for the effective transmission of HHV-8 to HUVECs in the present transmission system.

FIG. 3.

PCR analysis of HHV-8 infected HUVECs. HHV-8 DNA was detected in HUVECs infected with HHV-8 by cell-mediated transmission but not in those infected by cell-free transmission. HHV-8 DNA was also detected in HUVECs 30 days after infection. Lanes: M, 100-bp ladder marker; 1, HUVECs (not cocultured); 2, HUVECs infected by cell-free transmission for 48 h; 3, HUVECs infected by cell-mediated transmission for 48 h; 4, HUVECs 30 days after cell-mediated transmission; 5, HUVECs 30 days after cell-mediated transmission plus TPA; 6, BCBL-1 cells; 7, THP-1 cells; 8, no DNA

Persistent expression of LANA in HUVECs.

LANA-positive cells were first detected 6 h after the start of coculturing; this value increased to more than 30% after 72 h (Fig. 2A). To determine the duration of LANA expression, BCBL-1 cells were removed from the HUVEC culture 24 h after the start of coculturing, and HUVECs were washed thoroughly and continuously cultured for 30 days. We confirmed by microscopy that BCBL-1 cells were not present in each well after washing. The culture medium was changed every 2 days, and subculturing was performed once a week. LANA expression in each subculture was monitored by the IFA. The percentage of LANA-positive cells as well as the absolute number of these cells increased until day 5 after the start of coculturing (up to 40 to 50%). After passage 5 (30 days after removal of BCBL-1 cells), we added TPA to the culture medium and continuously cultured HUVECs for 48 h. DNA was extracted from these HUVECs before and after the addition of TPA. PCR analysis revealed that the HUVECs still contained a DNA fragment of the HHV-8 genome 30 days after removal of BCBL-1 cells. The IFA demonstrated that about 40% of the HUVECs expressed LANA, and TPA treatment induced the expression of ORF59, a DNA replication-associated protein of HHV-8, in 20% of all the HUVECs (Fig. 1K and L). We also confirmed the expression of LANA in passage 14 of HUVECs (3 months after removal of BCBL-1 cells), and the positivity rate (about 40%) did not change (data not shown). These data indicate that HUVECs cocultured with BCBL-1 cells carry the HHV-8 genome as a result of cell-mediated transmission and continue to express LANA for up to at least 3 months after infection.

Discussion.

In the present study, we showed that coculturing of TPA-treated BCBL-1 cells and HUVECs results in de novo HHV-8 infection of HUVECs. We also showed that such cell-mediated transmission was far more efficient than the previously reported cell-free system (7, 18, 20, 23). We also showed that the latent phase of infection was the predominant form for HUVECs and that the viral genome was carried in HUVECs during the latent phase of infection. To our knowledge, this is the first report describing an in vitro model of HHV-8 infection via cell-mediated transmission.

Previous in vitro transmission experiments with HHV-8 were performed conventionally using immortalized endothelial cells (18, 23). There is only one study that demonstrated successful transmission of HHV-8 to primary cultured HUVECs (7). In that system, the viral particles were present at a high concentration (5 to 10 genome equivalents per cell). Another group reported HHV-8 infection of untreated human endothelial cells from a neonatal brain using highly concentrated viral particles (5); however, even under these conditions, they failed to infect HUVECs. In contrast, we showed in this study that HHV-8, via cell-mediated transmission, could infect HUVECs without further treatments, such as transfection using transforming genes. Thus, we can conclude that cell-mediated transmission is far more effective in HHV-8 infection of normal human endothelial cells than cell-free transmission.

It remains to be determined which of the transmission modes, cell mediated or cell free, occurs in HHV-8 infection of endothelial cells in vivo. Although the number of viral particles in the sera of infected individuals is important for cell-free transmission, their titers in sera have not been determined (10). Based on the present results, we speculate that cell-mediated transmission may be the more predominant mode. If this is the case, HHV-8-infected peripheral blood B cells may play the role of a reservoir, as suggested by previous reports (10, 17), considering that BCBL-1 cells, an HHV-8-infected B-cell line, functioned as virus providers in this study.

We observed that HHV-8-infected HUVECs could proliferate without the further addition of HHV-8. Although the number of uninfected HUVECs also increased under these conditions, the number of LANA-positive cells increased until confluence (Fig. 2A). In addition, 40% of HUVECs expressed LANA even 30 days after infection. A similar observation was reported by another group (7). They reported that HHV-8 infection caused the transformation of HUVECs and that the transformed HUVECs exhibited long-term survival compared to uninfected HUVECs. To investigate whether transformation occurred in our HHV-8-infected HUVECs, we performed colony assays in the coculture dishes with the use of soft agar. However, at up to 3 months after infection, no evidence of transformation, such as focus formation, loss of contact inhibition in culture dishes, or anchorage-independent growth in soft agar, was noted, even though LANA was expressed in the cells (data not shown). Thus, we could not determine in the present study whether transformation occurred in HHV-8-infected HUVECs, but the results of the present study suggest that such transformation of HUVECs also may be caused by cell-mediated transmission of HHV-8. Most of the spindle-shaped cells observed in KS lesions express LANA, while few of these cells express lytic proteins (14, 21). It was recently reported that LANA could inhibit p53-mediated cell death, resulting in prolonged cell life and sustaining the persistence of HHV-8 in endothelial cells (9). Therefore, the prolonged cell life and much more rapid cell growth mediated by HHV-8 latent-phase infection may contribute to the mechanism of multistep tumorigenesis, in combination with the effects of putative viral oncogenes and/or cellular events present in the KS microenvironment. These additional factors may be necessary for the complete transformation of HUVECs that was not observed in the present study.