Induction of the Warburg effect by Kaposi's sarcoma herpesvirus is required for the maintenance of latently infected endothelial cells

Abstract

Kaposi’s sarcoma (KS) is the most commonly reported tumor in parts of Africa and is the most common tumor of AIDS patients world-wide. KS-associated herpesvirus (KSHV) is the etiologic agent of KS. Although KS tumors contain many cell types, the predominant cell is the spindle cell, a cell of endothelial origin that maintains KSHV latency. KSHV activates many cell-signaling pathways but little is known about how KSHV alters cellular metabolism during latency. The Warburg effect, a common metabolic alteration of most tumor cells, is defined by an increase in aerobic glycolysis and a decrease in oxidative phosphorylation as an energy source. The Warburg effect adapts cells to tumor environments and is necessary for the survival of tumor cells. During latent infection of endothelial cells, KSHV induces aerobic glycolysis and lactic acid production while decreasing oxygen consumption, thereby inducing the Warburg effect. Inhibitors of glycolysis selectively induce apoptosis in KSHV-infected endothelial cells but not their uninfected counterparts. Therefore, similar to cancer cells, the Warburg effect is necessary for maintaining KSHV latently infected cells. We propose that KSHV induction of the Warburg effect adapts infected cells to tumor microenvironments, aiding the seeding of KS tumors. Additionally, inhibitors of glycolysis may provide a unique treatment strategy for latent KSHV infection and ultimately KS tumors.

Kaposi's sarcoma (KS) is the most common tumor of AIDS patients and also arises in the posttransplant setting. KS is currently the most commonly reported tumor in hospitals in parts of Africa and is seen in both HIV-positive and -negative patients. KS is a highly vascularized tumor that is predominantly populated by spindle cells, a cell of endothelial cell origin. Spindle cells in all KS tumors are infected with Kaposi's sarcoma-associated herpesvirus (KSHV). KSHV is a γ-herpesvirus and is the etiologic agent of a number of malignancies including KS, an endothelial based tumor, and two B-cell lymphoproliferative diseases, primary effusion lymphoma (PEL) and plasmablastic multicentric Castleman's disease (1). In KS and PELs, KSHV is predominantly latent and, therefore, impervious to current antiviral drug treatments that inhibit lytic replication. KSHV also establishes a predominantly latent infection in cultured endothelial cells (2). KSHV has been shown to activate many signaling pathways that may be involved in the induction of tumors. Previously, we and others demonstrated that AKT is activated by latent KSHV infection of endothelial cells (3–5). We also found that KSHV infection of endothelial cells activates hypoxia-induced factors (HIF) -1 and -2 (6). AKT and HIFs have been shown to play prominent roles in the Warburg effect.

As first noted by Otto Warburg in the 1920s, even in the presence of oxygen, most tumor cells induce glycolysis and lactic acid production as their main energy source rather than mitochondrial oxidative phosphorylation (7). Many tumor-cell types have been shown to require glycolysis and strongly down-regulate mitochondrial oxidative phosphorylation (8–10). Although lactic acid production results in less ATP per molecule of glucose, it has been proposed that the increased glycolysis and decreased oxidative phosphorylation may serve to increase the rate of ATP production without producing reactive oxygen species. Recently, it was also proposed that the Warburg effect might be involved in the avoidance of apoptosis (11). Alternatively, the Warburg effect might serve to increase biomass to provide nucleotides and lipid material necessary for rapidly dividing cells (12). This theory is supported by the fact that signaling pathways like AKT/mTOR, known to play a role in biomass production, also control aspects of the Warburg effect.

Here we find that KSHV induces aerobic glycolysis and decreases oxygen consumption in latently infected endothelial cells, thereby inducing the Warburg effect. Additionally, we find that KSHV-infected endothelial cells are extremely sensitive to glycolytic inhibitors and these inhibitors selectively induce apoptosis in latently infected cells. This finding raises the potential that glycolytic inhibitors may be unique therapeutics for the treatment of latently infected cells.

Results

KSHV Induces Aerobic Glycolysis.

To determine if KSHV induces glycolysis in endothelial cells, we examined the effect of KSHV infection on glucose uptake and regulation of key glycolytic enzymes. When cells use glycolysis as a main energy source, they require increased glucose consumption. To determine if endothelial cells latently infected with KSHV increase glucose uptake, we infected tert-immortalized microvascular endothelial cells (TIME cells) or primary dermal microvascular endothelial cells (1° hDMVECs) with KSHV. In all experiments, greater than 90% of the cells express markers of latent infection; less than 1% of the cells express markers of lytic replication as determined by immunofluorescence with antibodies to the latency associated nuclear antigen (LANA) and ORF 59 (a lytic KSHV protein). Forty-eight hours postinfection, when KSHV latency is clearly established, a radiolabeled glucose analog, 2-deoxy-D-glucose (2DG), was added to the media of mock- and KSHV-infected endothelial cells. Intracellular radiolabeled 2DG levels were then determined 10 min posttreatment. As seen in Fig. 1A, latently infected TIME cells show an ≈50% increase in glucose uptake compared with mock infected cells. This increase is in line with many cancer cells in culture (13). We previously showed that HIF transcriptional activity is increased by KSHV infection of endothelial cells (6). Hypoxia increases the stability of HIFs and HIF protein levels and activity are synergistically increased by KSHV infection combined with the hypoxia mimic, desferoxamine (Dfo) (6). When KSHV-infected cells were treated with Dfo, there was also a large synergistic increase in glucose uptake (Fig. 1A). This synergy is greater than 2-fold that of mock-infected cells treated with Dfo.

Fig. 1.

KSHV infection induces aerobic glycolysis. (A) Glucose uptake is increased in KSHV- versus mock-infected cells. Forty-eight hours after infection, mock (M) or KSHV (K) infected TIME cells were exposed to a radiolabeled glucose analog (Amersham TRL383) for 10 min followed by intracellular quantification of radioactivity. Cells were treated with Dfo (D), where indicated for the last 16 h of infection. KSHV-infected cells were greater than 90% latently infected and less than 1% of the cells expressed lytic markers. (B) Western blot analysis of GLUT3 expression in mock- and KSHV-infected TIME cells. Cells were treated with Dfo (D) where indicated for the last 8 h of infection. (C) Western blot analysis of HK2 expression in mock- and KSHV-infected TIME cells 48 h postinfection. (D) Increased lactate production in KSHV- versus mock-infected TIME cells. Cellular supernatant was collected from mock- or KSHV-infected cells 48 h postinfection and lactate production was quantified by an enzymatic colorimetric assay.

Glucose transporter 3 (GLUT3) is induced in many cancer cells lines, as well as during hypoxia, and is responsible for increased glucose uptake (14). GLUT3 protein levels are greatly increased in endothelial cells latently infected with KSHV as determined by Western blot analysis with anti-GLUT3 antibodies (Fig. 1B). As with glucose uptake, GLUT3 is synergistically increased when KSHV-infected cells are treated with the hypoxia mimic Dfo, indicating that GLUT3 induction is likely responsible for the increased glucose uptake into KSHV-infected cells. Hexokinase controls the first step of glycolysis by converting glucose to glucose-6-phosphate. Hexokinase 2 (HK2) is often a rate-limiting enzyme for the metabolism of glucose and, interestingly, is also associated with the mitochondria and apoptosis (8, 15–17). Latent KSHV infection of TIME cells also strongly increases the protein levels of HK2 (Fig. 1C). In summary, KSHV induces glucose uptake into latently infected endothelial cells and up-regulates a glucose transporter and the rate-limiting glycolytic enzyme HK2.

To determine if the increased glucose uptake leads to increased glycolysis, we examined the production of lactic acid. Once glucose is metabolized into pyruvate, it can be converted into acetyl-CoA inside the mitochondria and ultimately used for oxidative phosphorylation, or it can be converted into lactic acid by lactate dehydrogenase and secreted from the cell. The latter process occurs in anaerobic conditions and leads to the acidification of cell-culture media; however, induction of the Warburg effect can lead to lactic acid production in aerobic conditions (18, 19). Lactate levels in the media of mock- or KSHV-infected cells were harvested in aerobic conditions and quantified by an enzymatic assay. As seen in Fig. 1D, there is an ≈50% increase in lactate production in the KSHV-infected cell conditioned media versus mock-infected conditioned media. Additionally, the extracellular acidification rate was directly measured using the Seahorse XF extracellular flux analyzer. This highly sensitive analyzer quantifies the acidification rate in live cells over time. Mock- and KSHV-infected cells are seeded in quadruplicate into wells that are sealed by plungers containing probes that simultaneously measure extracellular acidification rates and oxygen consumption rates. The probes also have drug injection ports that can inject drugs into the media at specified time points. Measurements of acidification rates and oxygen consumption rates are taken repeatedly during 4-min intervals over the course of 3 h. As seen in Fig. 2 A and B, both endothelial cell lines (TIME and 1° hDMVECs) latently infected with KSHV acidify the media more rapidly than mock-infected cells, as determined by the higher baseline acidification rate. For example in 1° hDMVECs (Fig. 2B), KSHV-infected cells have an extracellular acidification rate (ECAR) of ∼29 mili pH per min (mpH/min), and mock-infected cells have an acidification rate of ∼20 mpH/min. This ≈50% increase in acidification rate is similar to recently transformed cells in culture as described previously (19).

Fig. 2.

KSHV-infected cells increase lactic acid production and decrease oxygen consumption. Seahorse Bioscience extracellular flux analyzer was used to measure the ECAR (A and B) and OCR (C and D) in mock- and KSHV-infected TIME (A and C) or 1° hDMVECs (B and D). Forty-eight hours postinfection, cells were seeded in 24-well plates in unbuffered solution, the wells were sealed by plungers, and ECAR and OCR was measured over 4 min. The wells were then released and resealed and measured again (each datapoint is a separate measurement of rate for quadruplicate samples). A and C are from the same experiment and B and D are from the same experiment. Line A indicates the injection time of oxamate (100 mM) and line B indicates the injection time of rotenone (1 μM).

Oxamate selectively inhibits glycolysis but not oxidative phosphorylation by inhibiting lactate dehydrogenase (LDH), the enzyme that converts pyruvate to lactate (Fig. 3) (20, 21). Oxamate treatment decreased the acidification rate of both mock- and KSHV-infected cells to the same rate (∼7 mpH/min) in both TIME cells and 1° hDMVECs (Fig. 2 A and B, after line A). This result indicates mock- and KSHV-infected cells have equal basal acidification rates of the media in the absence of glycolysis and that the differences in acidification were indeed a result of differences in the processing of glucose into lactic acid. Rotenone blocks oxidative phosphorylation by inhibiting complex I of the mitochondria (Fig. 3). As expected, 1-μM rotenone treatment has little effect on the acidification rates of mock- or KSHV-infected cells after treatment with oxamate (Fig. 2 A and B, after line B). These sets of experiments demonstrate that KSHV-infected endothelial cells increase lactic acid production, indicating the increased uptake of glucose is converted to lactic acid and that KSHV infection induces aerobic glycolysis.

Fig. 3.

Overview of cellular metabolism with glycolytic and mitochondrial inhibitors. 2DG inhibits glycolysis by metabolic trapping. Oxamate inhibits LDH activity. Rotenone inhibits complex I of mitochondrial oxidative phosphorylation.

KSHV Infection Decreases Mitochondrial Oxidative Phosphorylation.

When cancer cells increase aerobic glycolysis, it is often accompanied by a decrease in mitochondrial oxidative phosphorylation, leading to decreased oxygen consumption (7, 9, 22). To determine if KSHV infection decreases oxidative phosphorylation while increasing aerobic glycolysis, we used the Seahorse extracellular flux analyzer to measure oxygen consumption rates (OCR) as described above. As seen in Fig. 2 C and D, KSHV-infected TIME and 1° hDMVECs have a lower OCR compared with mock-infected cells. For example, in 1° hDMVECs (Fig. 2D), mock-infected cells have an OCR of ∼60 pMoles of oxygen per minute (pMoles/min), and KSHV-infected cells have an OCR of ∼26 pMoles/min. These data demonstrate that KSHV-infected cells use less oxygen, indicating decreased mitochondrial oxidative phosphorylation. As expected, oxamate treatment has little effect on the oxygen consumption rates of mock- or KSHV-infected cells (Fig. 2 C and D, after line A). Rotenone treatment, which inhibits mitochondrial oxidative phosphorylation, lowered oxygen consumption rates in both mock- and KSHV-infected endothelial cells to similar levels, indicating that the decreased oxygen consumption in KSHV-infected cells is caused by decreased oxidative phosphorylation (Fig. 2 C and D, after line B). In addition to increasing aerobic glycolysis, KSHV infection of endothelial cells inhibits oxidative phosphorylation, demonstrating that KSHV induces the entire Warburg effect in endothelial cells. Interestingly, latent KSHV infection of human foreskin fibroblast (HFF) cells did not induce increased glycolysis or decreased oxygen consumption (Fig. S1). This finding indicates that KSHV induction of the Warburg effect may be specific to endothelial cells, the relevant KS tumor cell type.

Glycolytic Inhibitors Induce Apoptosis in KSHV-Infected Cells.

Glycolytic inhibitors have been used to selectively eliminate cancer cells because of their requirement for high levels of glucose (20, 21, 23, 24). These inhibitors have small effects on normal differentiated cells that predominantly use oxidative phosphorylation but quickly kill cancer cells. To determine if KSHV-induced aerobic glycolysis is necessary for the maintenance of latently infected endothelial cells, we treated mock- and KSHV-infected TIME or 1° hDMVECs with glycolytic inhibitors 2DG or oxamate. 2DG is a glucose analog that inhibits the conversion of glucose to glucose-6-phosphate by competing with glucose for binding to HK2 (Fig. 3). 2DG inhibits the processing of glucose for both glycolysis and oxidative phosphorylation. Tumor cells are more sensitive to 2DG because they primarily use glycolysis for ATP generation and must increase glucose uptake to compensate for the lower yield of ATP per glucose molecule (23). KSHV-infected TIME cells have slightly higher basal-cell death levels 48 h postinfection than their mock-infected counterparts (Fig. 4A). However, treatment with 2DG for 48 h induces a much larger dose-dependent increase in cell death in KSHV- versus mock-infected cells when the basal-cell death levels are subtracted out. For example, after subtracting out the basal death rate in the absence of drug, 31% of KSHV-infected cells die because of treatment with 160 mM 2DG; only 13% of mock-infected cells die because of similar 2DG treatment. Oxamate specifically inhibits LDH and therefore prevents regeneration of NAD⁺ from NADH (Fig. 3). Inhibition of NAD⁺ generation in cells that mostly use glycolysis for their energy production is detrimental because NAD⁺ is an important precursor for continued glycolysis. Unlike 2DG, oxamate presumably does not directly affect processing of glucose for oxidative phosphorylation because the electron transport chain regenerates the NAD⁺ needed to continue the glycolytic flux. While indirect effects cannot be ruled out, oxamate treatment of mock- or KSHV-infected endothelial cells has no effect on oxygen consumption. Oxamate treatment selectively kills around 50% of the KSHV-infected TIME cells or 1° hDMVECs, but does not induce additional cell death over basal levels in mock-infected cells (Fig. 4B). These results were confirmed and visualized in TIME cells treated with oxamate for 36 h, by a microscopy-based fluorescent assay for dead cells (Fig. 4C, Image-It Dead Green Viability stain). Because glycolytic inhibitors can also inhibit cell proliferation, we treated mock- and KSHV-infected cells with 5-fluorouracil, a cell proliferation inhibitor. There was no increase in cell death when mock- or KSHV- latently infected TIME cells were treated with 5-fluorouracil, demonstrating that inhibition of glycolysis and not inhibition of cellular proliferation leads to cell death in latently infected cells (Fig. S2). This set of experiments demonstrates that endothelial cells latently infected with KSHV require induced glycolysis for survival. Therefore, the Warburg effect is essential for the maintenance of latently infected cells. As described above, KSHV did not induce the Warburg effect in HFF cells and accordingly, oxamate did not induce cell death in KSHV-infected HFF cells (Fig. S1). Interestingly, KSHV-induced cell death in ≈40% of cells from a primary effusion lymphoma cell line, JSC-1 (Fig. S3A). However, because PEL cell lines are isolated from human tumors and there is no proper uninfected control, it is difficult to differentiate the effects of viral infection and tumor selection in vivo. Our laboratory has previously created a KSHV-infected B-cell line in culture by transducing viral DNA into BJAB B-cells (25). However, oxamate induced cell death in ≈17% of uninfected BJAB cells and there was only a slight increase in cell death in KSHV-infected BJAB cells (Fig. S3B). Like PEL cell lines, BJABs were isolated from a human tumor and likely have already induced the Warburg effect during tumor selection in vivo.

Fig. 4.

Inhibitors of glycolysis selectively induce apoptosis in KSHV latently infected cells. (A) Mock- (blue) and KSHV- (pink) infected TIME cell death percentages were determined by a Trypan blue exclusion assay. At 48 h postinfection, cells were seeded into flasks at equal numbers and treated for an additional 48 h with 0, 160, and 320 mM 2DG. Cells death rates were determined by counting cells using a hemocytometer after addition of Trypan blue. Cell death rate (%) = no. of dead cells/no. of total cells. (B) At 48 h postinfection, mock- and KSHV-infected TIME (Upper) cells or 1° hDMVECs (Lower) were treated with 0 and 100 mM oxamate for an additional 48 h and cell death was measured as in A. (C) Forty-eight hours postinfection, mock- and KSHV-infected TIME cells were treated with 0 or 100 mM oxamate for 36 h and then treated with Image-It Dead Green Viability stain for 15 min. (D) Western blot analysis of caspase-3 and PARP cleavage after treatment of mock- (M) and KSHV- (K) infected cells with oxamate for 36 h. St = 1 μM stauroporin treatment of mock-infected cells (8 h). (E) Mock- and KSHV-infected TIME-cell death percentages were determined by Trypan blue assay (as in A and B). Cells were treated for 36 h with 0 mM oxamate (control), 100 mM oxamate, or 100 mM oxamate + 100 uM ZVAD (caspase inhibitor).

To determine the mechanism of cell death induced by oxamate in latently infected endothelial cells, we examined classic markers of apoptosis. Caspase 3 cleavage is an early event in many forms of programmed cell death and leads to the induction of poly(ADP-ribose) polymerase (PARP) cleavage. Antibodies that recognize the cleaved forms of these proteins can be used to determine if cells are undergoing apoptosis. TIME cells were infected with KSHV for 48 h and then treated with oxamate for an additional 36 h. Treatment of KSHV-infected cells with oxamate led to high levels of caspase 3 and PARP cleavage (Fig. 4D). However, there was no induction of caspase 3 or PARP cleavage when mock-infected cells were treated with the same doses of oxamate or in untreated mock or KSHV-infected cells. TIME cells treated with staurosporin, which strongly induces apoptosis in cells, was used as a positive control. To further show that KSHV-infected cells treated with oxamate selectively die by apoptosis, we treated mock- and KSHV-infected cells with oxamate and the caspase inhibitor ZVAD. Western blot data shows that 36-h treatment of KSHV-infected TIME cells with ZVAD along with oxamate prevents caspase-3 cleavage (Fig. S4). A Trypan blue exclusion assay was also done in TIME cells similarly treated with ZVAD and oxamate for 36 h. A greater than 2-fold inhibition of cell death was seen when the caspase inhibitor ZVAD was used with oxamate treatment in KSHV-infected cells, decreasing cell death from 61 to 28% (Fig. 4E). These results demonstrate that inhibition of glycolysis induces cell death by apoptosis specifically in KSHV-infected endothelial cells but not in mock-infected cells.

Discussion

We demonstrated that latent KSHV infection of endothelial cells induces aerobic glycolysis and decreases oxidative phosphorylation, the properties that define the Warburg effect, a common effect in over 90% of cancers (10). Additionally we show that KSHV induction of the Warburg effect is not universal but specifically occurs in endothelial cells, the relevant KS tumor-cell type. KSHV induction of the Warburg effect may specifically adapt endothelial cells for tumor formation. KSHV-infected endothelial cells would be able to withstand fluctuations in oxygen that can occur in tumor environments, and therefore infected cells would be able to grow quickly during oxygen fluctuation although other cells would necessitate adaptation over time. Even when there are high levels of angiogenesis, as occurs in KS tumors, new blood vessels often do not supply sufficient oxygen because of poor architecture and leakiness. Therefore, KSHV-infected endothelial cells might be well adapted to tumor microenvironments and viral induction of the Warburg effect may be important for seeding initial tumor growth.

Induction of glycolyis may not be unique to KSHV. Two retroviruses, feline leukemia virus and Rous sarcoma virus, were shown to induce glycolysis by increasing glucose uptake, glucose transporters, and lactic acid production (26, 27). It is possible that the induction of aerobic glycolysis may be a common trait of viral infections. A metabolomic study of cytomegalovirus (CMV), a β-herpesvirus, found that there were many changes in metabolic enzymes, including the induction of glycolytic enzymes during lytic infection of fibroblast cells (28). The switch from oxidative phosphorylation to glycolysis may benefit enveloped virus production by increasing biomass including nucleotides, fatty acid, and lipid material (12). Enveloped viruses might require increased lipid production to keep up with the needs of producing viral progeny by budding. However, the CMV studies were done during lytic replication of CMV, and the current studies analyze predominantly latent infection of an oncogenic herpesvirus. During latency, the need for lipid envelopes is less but the virus might need induced biomass for rapid cell division. Alternatively, a decrease in the production of reactive oxygen species by increasing glycolysis and shutting down the mitochondria may be advantageous for cell growth in latently infected cells. Additionally, acidification of the local environment may decrease the immune system's ability to respond to latently infected cells. Furthermore, the increased biomass production could set the latent cell up for rapid induction of lytic replication with the necessary nucleotides and lipid material in place for rapid production of enveloped viral progeny. Another possible function of the Warburg effect might be to prevent apoptosis in cancer cells (11). In cancer cells, induction of glycolysis is necessary for the maintenance of oncogene-induced transformation of cells as well. Although KSHV does not generally induce transformation of primary dermal microvascular endothelial cells in culture, it may induce oncogenes or oncogenic signaling pathways that could lead to apoptosis in specific settings (11). Therefore, induction of the Warburg effect may circumvent apoptosis caused by induction of oncogenes during KSHV latency. Alternatively, aerobic glycolysis may be necessary to provide enzymes or cofactors necessary to maintain the latent state. Interestingly, 2DG was shown to inhibit human papillomavirus transcription and could play a role in the replication of KSHV during latency (29).

Importantly, inhibitors of glycolysis could provide previously unexplored treatments for latent KSHV infection in endothelial cells. Current drugs that are used to treat herpesvirus infections all target lytic replication, while KS tumor cells are predominantly latently infected and cannot be targeted directly. Therefore, it is not possible to eliminate latently infected cells or cure herpesviral infections. Interestingly, rapid resolution of KS tumors occurred when posttransplant patients immunosuppressive drug regimen was switched to rapamycin, a drug that inhibits mTOR and can block induction of glycolysis (30). Therefore, inhibitors of glycolysis may be strong candidates for novel treatment of latent KSHV infection in endothelial cells and ultimately KS tumors.

Materials and Methods

Cell Lines, Reagents, and Antibodies.

TIME cells or 1° hDMVECs were maintained as monolayer cultures in EBM-2 media (Lonza) supplemented with a bullet kit containing 5% FBS (FBS), vascular endothelial growth factor, basic fibroblast growth factor, insulin-like growth factor 3, epidermal growth factor, and hydrocortisone. HFF cells were maintained in DMEM supplemented with 10% FBS, 1% glutamine, and 1% Pen/Strep antibiotics. JSC-1 and BJAB cells were maintained in RPMI 1640 media supplemented with 10% FBS, 1% glutamine, 1% Pen/Strep and 0.1% β-mercaptoethanol. 2DG and oxamate (Sigma) were diluted in water and used at the indicated final concentrations. Dfo (Calbiochem) was diluted in water and was used at a final concentration of 150 μM. Rotenone and staurosporin (Sigma) were diluted in DMSO and were used at a final concentration of 1 μM. 5-Fluorouracil (Sigma) was diluted in complete EBM-2 media and was used at a final concentration of 10 μg/mL ZVAD (BD Pharmingen) was diluted in DMSO and was used at a final concentration of 100 μM. Image-It Dead Green Viability stain (Invitrogen) was diluted in DMSO and was used at a final concentration of 100 nM. Glut3 (Abcam), HK2 (Santa Cruz), cleaved caspase-3 (Cell Signaling), PARP (Cell Signaling) and β-actin (Sigma) were used as specified by the manufacturer.

Viruses and Infection.

KSHV inoculum from induced BCBL-1 cells was tittered and used to infect TIME and 1° hDMVECs as previously described (31). All cells were infected in serum-free EBM-2 or DMEM media for 3.5 h, after which the medium was replaced with complete EGM-2 or DMEM media. Infection rates were assessed for each experiment by immunofluorescence and only experiments where greater than 90% of the cells expressed LANA and less than 1% of the cells expressed ORF59 were used.

Immunofluorescence.

Before harvesting cells for immunoblot or enzymatic assays, an aliquot of mock- or KSHV-infected 1° hDMVEC or TIME cells were seeded on LabTek Permanox four-well chamber slides (Intermountai Sci) and fixed with 4% (vol/vol) paraformaldehyde in PBS. Infection rates were monitored using antibodies against the latent KSHV protein LANA (a kind gift from A. Polson and D. Ganem, University of California San Francisco) and the lytic protein ORF59 (Advanced Biotechnologies Incorporated) as described previously (2). Cells were incubated with fluor-conjugated secondary antibodies (goat anti-rabbit Alexa Fluor 488 and goat anti-mouse Alexa Fluor 594, Molecular Probes/Invitrogen). Cells were mounted in medium containing DAPI (4',6'-diamidino-2-phenylindole) before being viewed under a Nikon Eclipse E400 fluorescence microscope (Nikon, Inc.).

Glucose Uptake.

Twenty-four hours after mock or KSHV infection, TIME cells were seeded in six-well plates at equal cell numbers. Complete media was replaced, 32 h postinfection, with serum-free EBM-2, with or without Dfo. Forty-eight hours postinfection, cells were washed three times with PBS with Mg and Ca, then fed 1 μCi (1 μL) of 2-deoxy-D-[1-³H]glucose (Amersham TRK383) in 1 mL PBS with Mg and Ca for 10 min at 37 °C. The cells were then washed two times with ice-cold PBS without Mg and Ca, then lysed in 1 mL 1% SDS for 15 min at room temperature. Next, 0.5 mL of lysate in 4.5 mL Ecoscint H (National Diagnostics LS-275) was counted on a scintillation device for 1 min to get counts per minute. The remaining lysate was quantitated by BCA Protein Assay Reagent Kit (Pierce 23225), and samples were normalized for counts per minute per milligram.

Lactate Production.

TIME cells were mock- or KSHV-infected and seeded into 10-cm dishes 24 h postinfection. At 32 h postinfection, media were changed to serum-free EBM-2. Cell supernatant was collected 48 h postinfection for lactate analysis. Lactate production was measured using an enzymatic kit (R-Biopharm) as specified by the manufacturer and results were normalized by cell number. Briefly, NAD⁺ is added to the media and is converted to NADH stoichiometrically by lactate in the media. The levels NADH are quantified colorimetrically as described by the manufacturer.

Immunoblot Analysis.

TIME cells were mock- or KSHV-infected, harvested using a cell scraper and pelleted at 1,000 × g for 7 min at 4 °C. An aliquot of the cells was seeded onto chamber slides for immunofluorescence analysis. Cell pellets were washed once in cold PBS and then resuspended in RIPA lysis buffer [50 mM Tris-HCl, pH 7.6, 150 mM NaCl, 1 mM EDTA, 1% Nonidet P-40, 0.5% deoxycholate, 0.1% SDS, 1 mM sodium orthovanadate, 1 mM sodium fluoride, 40 mM β-glycerophosphate, and Complete Mini protease inhibitor tablet (Roche)]. Protein concentrations were determined by the bicinchoninic acid assay (Pierce). Next, 15 to 50 μg of protein was fractionated on a SDS-polyacrylamide gel, transferred to a polyvinyl difluoride membrane, blotted with the appropriate primary antibody [dilutions were 1:1,000 anti-Glut3 (60 kDa), 1:800 for anti-HKII (100 kDa), 1:1,000 anti-cleaved caspase 3 (17/19 kDa), 1:1,000 anti-PARP (116/89 kDa), and 1:10,000 for anti-β-actin (45 kDa)], and subsequently with horseradish peroxidase-conjugated donkey anti-goat (1:20,000), goat anti-rabbit (1:10,000), goat anti-mouse (1:10,000). Immunoreactive proteins were visualized by chemiluminescence using the Amersham ECL Plus Western blotting detection reagents (GE Healthcare).

Seahorse Biosciences.

Mock and KSHV infected TIME cells were seeded in quadruplicate, 24 h postinfection, at equal densities (30,000 cells per well) into XF24 tissue culture plates in complete EGM-2 medium containing serum and supplements. Cell media was changed 8 h after cell seeding into unbuffered DMEM medium [8.3g/L DMEM (Sigma), 200 mM GlutaMax-1 (Invitrogen), 25 mM D-glucose (Sigma), 63.3 mM NaCl (Sigma), and phenol red (Sigma), adjusted pH to 7.4 with NaOH)] according to manufacturer's protocol. Measurement of cellular oxygen consumption and lactate production was performed during 4-min intervals over the course of 3 h using the Seahorse XF24 analyzer (Seahorse Bioscience Inc.). During the measurements, the wells were sealed with mechanical plungers containing probes that measure extracellular acidification (pH) and oxygen consumption. Both oxygen consumption and lactate production was measured under basal conditions and after injection of glycolytic inhibitor oxamate (100 mM) and mitochondrial inhibitor rotenone (1 μM). Oxygen consumption and lactate measurements were normalized to cell number.

Trypan Blue Exclusion Assay.

Primary hDMVECs, TIME cells, or HFF cells were mock or KSHV infected. Cells were seeded subconfluently, 48 h postinfection, into 12-well plates and treated with 2DG (0 mM, 160 mM, or 320 mM) or oxamate (0 mM or 100 mM) or 100 μM ZVAD plus 100 mM oxamate for 48 h. Cells were trypsinized and pelleted with cellular supernatant for 5 min at 1,000 × g. Pellet was resuspended in 60 μL media. Cell death rates were determined by counting cells using a hemocytometer after addition of Trypan blue, which stains the cytoplasm of dead cells but not live cells. Cell death rate (%) = number of dead cells/number of total cells (×100%). Similarly JSC-1, BJAB, and BJAB + KSHV cells were seeded at 600,000 cells/mL and treated with 0 mM or 100 mM oxamate for 48 h. Cell death rates were determined by a Trypan blue exclusion assay.

Image-It Dead Green Viability Stain.

TIME cells were mock- or KSHV-infected for 48 h. Cells were seeded subconfluently in 12-well plates and treated with 0 or 100 mM oxamate for 36 h. Cells were then treated with Image-It Dead Green Viability stain for 15 min and pictures were taken with a fluorescent microscope.