Inflammatory Cytokines Inhibit Kaposi's Sarcoma-Associated Herpesvirus Lytic Gene Transcription in In Vitro-Infected Endothelial Cells

Abstract

The response of Kaposi's sarcoma (KS)-associated herpesvirus (KSHV) to inflammatory cytokine treatment of experimentally infected endothelial cells was investigated. The cytokines inhibited spontaneous KSHV lytic gene expression but not the level of infection. The data suggest that if inflammatory cytokines present in KS lesions contribute to KSHV pathogenesis, they do so in part by promoting latent KSHV infection of the endothelial cells.

Kaposi's sarcoma (KS)-associated herpesvirus (KSHV) is the cause of KS (6) and is closely associated with primary effusion lymphoma (PEL) and multicentric Castleman's disease (reviewed in reference 21). In KS, the virus infects spindle-shaped cells lining the vascular structures characteristic of all clinical forms (3, 10). KSHV encodes a variety of genes with transforming potential, most of which are expressed during lytic replication. However, in the majority of infected KS tumor cells, KSHV replication is latent, with a restricted profile of latency-associated genes expressed (9, 25, 31, 34). Certain of these genes may contribute to KSHV pathogenesis either by transforming or promoting the cell cycle or by inhibiting apoptosis (7, 10, 15, 24, 26, 37).

Cytokines produced by infiltrating inflammatory cells, in the presence of an etiologic agent, have long been considered to contribute substantially to KS pathogenesis, at least in the AIDS setting (18, 29, 30). Indeed, the presence of abundant inflammatory cytokines, including gamma interferon (IFN-γ), in AIDS KS and in individuals at high risk of developing AIDS KS is well documented (reviewed in reference 11). In this regard, we and others have found that KSHV replication can be induced in PEL cells by IFN-γ (2, 5, 20). Moreover, inflammatory cytokines induce phenotypic and functional features in endothelial cells that are consistent with those of KS tumor cells (13). For these reasons, inflammatory cytokines, particularly IFN-γ, have been proposed to cooperate with KSHV in promoting the pathogenesis of the virus (11). Other cytokines abundant in KS include interleukin-1β (IL-1β), IL-6, and tumor necrosis factor alpha (TNF-α) (35).

In the present study, we investigated the influence of some of these inflammatory cytokines on the biology of KSHV in experimentally infected human transformed microvascular endothelial cells (tDMVEC) (22), an in vitro model of KS. The cytokines suppressed KSHV lytic gene expression in this cell type, likely owing in part to the repression of basal transcriptional activity from the putative promoter of immediate-early (IE) open reading frame 50 (ORF50), encoding replication and transcription activator RTA. This IE protein is necessary and sufficient to drive lytic replication of KSHV (19, 36).

Inflammatory cytokines in KS.

The expression of inflammatory cytokines in KS was confirmed by performing immunocytochemical assays for IFN-γ in paraffin-embedded sections taken from 38 biopsy samples from separate individuals with AIDS and KS. In all cases, abundant IFN-γ-producing cells were identified at tumor sites (Fig. 1), indicative of the presence of other inflammatory cytokines in KS lesions and confirming the findings of others (14). Surrounding, uninvolved tissue also contained IFN-γ-producing cells, but they were less abundant than at the site of the tumor.

FIG. 1.

Immunohistochemical detection of IFN-γ-producing cells in a paraffin-embedded KS liver biopsy sample. A representative cell producing IFN-γ is indicated by the arrow. Similar data were found in analyses of a further 37 KS biopsy samples from a variety of tissue types (data not shown). (A) tissue stained with anti-IFN-γ primary antibody; (B) negative control with no primary antibody. IFN-γ antigen was retrieved by microwaving in citrate buffer (pH 6.0) and detected by staining with a mouse anti-IFN-γ monoclonal antibody that was detected with the Vectastain Elite and VIP kits (Vector Laboratories). Cells were counterstained with hematoxylin.

KSHV lytic gene expression.

Temporal studies of virus gene expression in the naturally infected PEL cell line BCBL-1 were performed by real-time reverse transcription (RT)-PCR with TaqMan chemistry (PE Biosystems). KSHV lytic replication in these cells was induced as previously described (28). Real-time RT-PCR provided relative quantification of KSHV IE ORF50 and late ORF29 transcripts and was performed on cDNA synthesized from whole-cell RNA (1). Viral transcript levels were normalized relative to cellular 18S rRNA levels detected with the TaqMan rRNA control reagents (PE Biosystems). PCR was carried out in duplicate on cDNA synthesized from 6 ng of RNA with TaqMan Universal PCR Master Mix reagents (PE Biosystems) in accordance with the manufacturer's instructions. The amplification was performed under the TaqMan universal cycle conditions on an ABI Prism 5700 sequence detector. The sequences of KSHV-specific primers and probes were as follows: ORF50 5′ primer, GCGCAAGATGACAAGGGTAAG; ORF50 3′ primer, CGAGAGGCCGACGAAGC; FAM-labeled probe, TTCCACACAGGACCGCCGAAGCT; ORF29 5′ primer, CCCGGAGGACGGTCCA; ORF29 3′ primer, TGTCCCCGAATGCTGTTCTTA; FAM-labeled probe, CTCGCTGATGTGCGCAACATGCT. These KSHV primers span splice junctions. Statistical analyses were performed on normalized cycle threshold values with Student's t test. The data indicated that peak expression of ORF50 occurred 24 h after treatment with phorbol ester (phorbol 12-myristate 13-acetate [PMA] at 20 ng/ml), with 25-fold ± 9-fold induction compared with untreated cells. The peak level of induction of ORF29 expression was 41-fold ± 19-fold 48 h after treatment. These experiments were performed three times independently. In comparison, KSHV in JSC-1 PEL cells (4) showed much higher levels of induction of gene expression when treated with sodium butyrate (1 mM). The ORF50 transcript level was induced 427-fold ± 295-fold, and the ORF29 transcript level was induced 86-fold ± 56-fold. These experiments were performed 10 times independently, and the variation between experiments is a characteristic of this biological system. From these data, KSHV isolated from JSC-1 cells [KSHV(JSC-1)] was selected for the subsequent infection studies.

Endothelial cell infection.

Concentrated KSHV(JSC-1) was prepared essentially as previously described (4). Transmission electron microscopy performed on a negatively stained preparation revealed the presence of enveloped and tegumented virions (Fig. 2); the particle count was approximately 10¹⁰ virions/ml. This virus was infectious to tDMVEC as determined by immunofluorescence assay (IFA) and real-time RT-PCR. Latency-associated nuclear antigen 1 (LANA-1) IFA analyses revealed the characteristic punctate nuclear staining pattern (Fig. 3) that is also seen in PEL cells. Many tDMVEC could be infected with concentrated KSHV(JSC-1) as determined by IFA (Table 1), but the level decreased with passage of the cells in all studies to date (more than 15 independent experiments). Nevertheless, by 7 days postinfection, pronounced phenotypic changes were evident in the infected tDMVEC compared with uninfected cells (Fig. 4).

FIG. 2.

Electron microscopy of cell-free KSHV(JSC-1). (A) capsids; (B) tegumented virions. Virus was concentrated from sodium butyrate-treated cells before negative staining and visualization by transmission electron microscopy. Scale bar = 150 nm.

FIG. 3.

Immunofluorescence detection of KSHV infection of tDMVEC. (A) low-power magnification showing approximately 68% of cells infected; (B) high-power magnification of nuclei of tDMVEC from an infection separate from that shown in panel A (approximately 50% of the cells are infected). tDMVEC were infected with concentrated KSHV(JSC-1) for 2 days before staining with a rat monoclonal antibody against human herpesvirus 8 LANA-1. Nuclei were revealed by costaining with propidium iodide. Images were collected by confocal microscopy.

TABLE 1.

Inflammatory cytokine treatment of infected tDMVEC cultures^a

Treatment (time [h])	% of cells LANA-1 positive (mean ± SD)
None	56 ± 6
Phorbol ester	52 ± 3
Cytokine combination (17)	58 ± 10
Cytokine combination (48)	45.3 ± 6

^aThe percentage of infected cells (staining for LANA-1) was determined by IFA. Cells were stained at the time of harvest, 10 random fields were collected by confocal microscopy, and the number of infected cells in each field was determined and compared with the total number of cells per field. The data represent the means of at least three independent experiments. PMA was used at 20 ng/ml. Recombinant cytokines were used at concentrations similar to those used by others in endothelial cell treatment studies (12, 14) (TNF-α, 2 ng/ml; IL-1β, 5 ng/ml; IFN-γ, 500 U/ml; IL-6, 100 U/ml). There was no statistically significant difference in the number of infected cells following the different treatments compared with untreated cells (Student's t test).

FIG. 4.

Infection of tDMVEC with KSHV(JSC-1). (A) uninfected tDMVEC (magnification, ×400); (B) tDMVEC at 7 days postinfection (magnification, ×400). Cells were infected as described in the text, and infection was confirmed by anti-LANA-1 IFA.

Inhibition of KSHV(JSC-1) lytic gene expression in infected tDMVEC by inflammatory cytokines.

Cultures of infected tDMVEC were treated either with phorbol ester for 48 h or with recombinant human inflammatory cytokines overnight (17 h) or for 48 h. The cytokines were IFN-γ, TNF-α (obtained from the NIBSC Centralised Facility for AIDS Reagents, to which they had been donated by Genentech Inc., San Francisco, Calif.), IL-6 (Sigma), and IL-1β (R&D Systems). Virus gene expression was quantified by real-time RT-PCR, and the level of infection was determined by anti-LANA-1 IFA. Phorbol ester induced both ORF50 and ORF29 transcripts by sixfold compared with untreated cells. This increase in expression, compared with that in untreated cells, was statistically significant (ORF50, P ≤ 0.04; ORF29, P ≤ 0.01; Student's t test). In contrast, treatment of the cells with IFN-γ, TNF-α, IL-1β, and IL-6 in combination suppressed KSHV(JSC-1) lytic gene expression in the tDMVEC (Fig. 5A). Suppression was detectable after 17 h of treatment and was more pronounced after 48 h, when the inhibition of both viral transcript species by the cytokines compared with untreated cells was statistically significant (ORF50, P ≤ 0.001; ORF29, P ≤ 0.03; Student's t test). When infected tDMVEC were treated with each recombinant cytokine individually, inhibition of KSHV(JSC-1) gene expression was inhibited in the order TNF-α > IFN-γ > IL-1β > IL-6, where TNF-α was the most effective and IL-6 was the least effective, having little or no effect (Fig. 5A). Thus, treating KSHV(JSC-1)-infected tDMVEC with inflammatory cytokines suppressed spontaneous transcription of the lytic genes ORF50 and ORF29. IFA analyses of the infected, cytokine-treated tDMVEC suggested that this effect was not due to selective elimination of virus-infected cells (Table 1), and the cytotoxicity of the cytokines, as determined by the MTT [3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl tetrazolium bromide] cell viability assay (16, 23), was not substantial (data not shown).

FIG. 5.

Modulation of KSHV gene expression by inflammatory cytokines. (A) tDMVEC cultures were infected with KSHV(JSC-1) and then expanded into replicate culture dishes that were treated with either phorbol ester (PMA at 20 ng/ml) or each cytokine individually for 48 h or with all four cytokines in combination for either 17 or 48 h before the effect on ORF50 and ORF29 transcript levels was measured in duplicate samples by real-time RT-PCR. Induction of gene expression was calculated relative to that in parallel cultures of untreated, infected tDMVEC. The data represent the mean values obtained from at least three independent experiments, except the ORF29 transcript measurements on cells treated with each cytokine individually, which were performed in duplicate. Recombinant cytokine concentrations were as follows: TNF-α, 2 ng/ml; IL-1β, 5 ng/ml; IFN-γ, 500 U/ml; IL-6, 100 U/ml. (B) Effect of inflammatory cytokines on the basal transcription activity of the region encompassing 250 bp upstream of the ORF50 translation initiation codon (cloned into a luciferase promoter reporter vector as pORF50/250-luc). Recombinant cytokine concentrations were as follows: TNF-α, 2 ng/ml; IL-1β, 5 ng/ml; IFN-γ, 500 U/ml; IL-6, 100 U/ml. (C) Effects of inflammatory cytokines on the transcription activity of an IFN-γ-responsive promoter in plasmid pGAS-luc. In panels B and C, modulation of firefly luciferase activity by cytokine treatment was calculated relative to the firefly luciferase activity of transfected, untreated cells cultured in parallel, which was assigned an activity of 100%. Each experiment was performed in duplicate and repeated at least three times, and all firefly luciferase data were normalized to Renilla luciferase activity expressed from a separate, constitutively active plasmid that was cotransfected with the firefly luciferase reporter plasmid. This normalization compensates for differences in transfection efficiency between replicate cultures. Recombinant cytokine concentrations: TNF-α, 2 ng/ml; IL-1β, 5 ng/ml; IFN-γ, 500 U/ml; IL-6, 100 U/ml.

Inhibition of ORF50 promoter activity by inflammatory cytokines.

To determine the mechanism of the inhibition of KSHV lytic gene expression, the inflammatory cytokines were evaluated for their effects on the transcriptional activity of the sequences upstream of the ORF50 gene, containing the putative promoter. Since the protein encoded by ORF50 (RTA) is necessary and sufficient to drive lytic replication of KSHV (19, 36), we reasoned that the inhibitory effect of the recombinant cytokines was mediated at the level of the promoter of this gene. However, this promoter has been incompletely characterized. To determine if recombinant inflammatory cytokines can modulate putative transcriptional regulatory elements upstream of ORF50, the 250-bp 5′ untranslated region (5′UTR) sequence upstream from the ORF50 translation initiation codon (19, 33, 36) of KSHV(BCBL-1) was PCR amplified and cloned into a firefly luciferase reporter vector (pGL-3-Basic; Promega) to create pORF50/250-luc. The primer sequences were as follows: Forward, 5′-GCGCCTCGAGATCTCCAATACCCGGAAT; Reverse, 5′-CCCAAGCTTTTGTGGCTGCCTGGACAGTATTC. This recombinant vector demonstrated spontaneous transcriptional activity from the ORF50 5′UTR sequence. The functionality of the 5′UTR of pORF50/250-luc was verified in cotransfection studies when an ORF50 expression vector activated basal promoter activity of pORF50/250-luc by almost twofold (data not shown), consistent with the levels seen by others for a comparable ORF50 5′UTR reporter plasmid in cells not infected with KSHV (32). The effects of recombinant inflammatory cytokines on this ORF50 5′UTR construct were then determined in transient transfection assays. For these studies, the human cell line 2fTGH was selected, in which the cascade of the IFN response pathway components is intact (reviewed in reference 8). However, 2fTGH cells probably differ from tDMVEC in many ways since they were derived from the HT1080 cell line, which was established from a poorly differentiated fibrosarcoma tumor (27). Nevertheless, since HT-1080 is most likely a connective tissue cell line, it is probably derived from the embryonic mesoderm, from which the endothelial cells of the tDMVEC line were also derived. The 2fTGH cells were transiently transfected with pORF50/250-luc and treated with the recombinant cytokines before harvesting for luciferase activity assays (Fig. 5B). Parallel studies were performed with a positive control vector (pGAS-luc; Stratagene) in which the firefly luciferase reporter gene was under the control of a promoter containing the IFN-γ activation site (Fig. 5C). The pGAS-luc reporter gene was activated in response to treatment either with the combination of four inflammatory cytokines or with IFN-γ alone (Fig. 5C). However, the ORF50 5′UTR activity was suppressed significantly following treatment with the combination of four cytokines (P ≤ 0.002; Student's t test; Fig. 5B), with TNF-α having the greatest effect of each cytokine individually, and the pattern of effect of the other three cytokines individually was consistent with the data obtained for the virus-infected tDMVEC (Fig. 5A).

Therefore, our data indicate that treating KSHV(JSC-1)-infected tDMVEC with inflammatory cytokines suppresses spontaneous lytic gene expression (ORF50 and ORF29) of the virus. By extrapolation, lytic replication of the virus is also likely to be suppressed. This finding is consistent with reports in the literature that lytic replication is tightly restricted in most KSHV-associated tumors (10, 25). Inhibition of ORF50 promoter activity could be one mechanism by which KSHV lytic gene expression is suppressed by inflammatory cytokines, since basal transcription from the ORF50 5′UTR is repressed by these cytokines in reporter gene assays. In contrast, the ORF50 promoter is presumably induced by recombinant IFN-γ in PEL cells of B-cell origin (2, 5, 20). Hence, the biology of KSHV is likely to differ in two tumors with which it is associated, PEL and KS, because of differences in the interactions between the virus and the environments of the different cell lineages of these tumors.

In conclusion, these data suggest that latent KSHV infection in KS lesions is due in part to production by the immune system of inflammatory cytokines within the lesions, presumably in response to the virus infection. Importantly, latent virus might drive the proliferation in the tumor of the component endothelial cells (38; see reference 17), whereas replicating virus infects new endothelial cells.