Repurposing Cytarabine for Treating Primary Effusion Lymphoma by Targeting Kaposi’s Sarcoma-Associated Herpesvirus Latent and Lytic Replications

ABSTRACT Oncogenic Kaposi’s sarcoma-associated herpesvirus (KSHV) is etiologically linked to primary effusion lymphoma (PEL), an aggressive and nontreatable malignancy commonly found in AIDS patients. In this study, we performed a high-throughput screening of 3,731 characterized compounds and identified cytarabine, approved by the FDA for treating numerous types of cancer, as a potent inhibitor of KSHV-induced PEL. We showed the high efficacy of cytarabine in the growth inhibition of various PEL cells by inducing cell cycle arrest and apoptosis. Cytarabine inhibited host DNA and RNA syntheses and therefore induced cellular cytotoxicity. Furthermore, cytarabine inhibited viral DNA and RNA syntheses and induced the rapid degradation of KSHV major latent protein LANA (latency-associated nuclear antigen), leading to the suppression of KSHV latent replication. Importantly, cytarabine effectively inhibited active KSHV replication and virion production in PEL cells. Finally, cytarabine treatments not only effectively inhibited the initiation and progression of PEL tumors but also induced regression of grown PEL tumors in a xenograft mouse model. Altogether, our study has identified cytarabine as a novel therapeutic agent for treating PEL as well as eliminating KSHV persistent infection.

PEL is a B-cell neoplasm involving body cavities of pleural, pericardial, and peritoneal spaces usually without extracavitary tumor masses (3). All PEL cells harbor multiple copies of the KSHV genome, which are required for their survival. Up to 70% of PEL cases are also associated with Epstein-Barr virus (EBV) infection (3). Most PEL cells are latently infected by KSHV, but a small number of them undergo spontaneous lytic replication. Several KSHV latent genes, including LANA (latency-associated nuclear antigen) (ORF73), vCyclin (ORF72), vFLIP (ORF71), and a cluster of microRNAs (miRNAs), drive the proliferation and survival of PEL cells (3,4). Numerous viral lytic genes such as viral interleukin-6 (vIL-6) (ORF-K2) also play a role in PEL growth and survival (3).
PEL usually occurs in HIV-infected patients, of whom half have KS or a history of KS (3). PEL accounts for about 4% of non-Hodgkin's lymphomas (NHLs) in HIV patients (5,6). Rare cases of PEL have been described in HIV-negative immunocompromised patients after solid organ transplantation or elderly men living in areas with a high KSHV prevalence, such as Mediterranean and Eastern European regions (7,8).
The prognosis of PEL patients is usually poor, with a median survival time of 6.2 months (9). There is currently no efficient and specific treatment for PEL (2,3). Because of its rarity, there has been so far no large prospective clinical trial to investigate the proper therapy for PEL. Hence, finding a treatment for PEL remains a challenge. In this context, repurposing old drugs is an attractive strategy for identifying potential treatment options for PEL.
Here, we performed a high-throughput screening (HTS) of 3,731 characterized compounds to identify inhibitors targeting KSHV-induced oncogenic addiction and malignancies. We used a model of KSHV-induced cellular transformation of rat primary mesenchymal stem cells. This model offers both parallel uninfected (MM) and transformed (KMM) cells for comparative screening (10). We identified cytarabine as a promising candidate for targeting KSHV-induced oncogenic addiction. Cytarabine is currently approved by the FDA for treating acute myeloid leukemia (AML), acute lymphocytic leukemia (ALL), and NHLs (11). We demonstrated that cytarabine was effective in inducing cell cycle arrest and apoptosis in PEL cells. Furthermore, cytarabine effectively inhibited the initiation and progression of PEL and induced the regression of grown PEL tumors in a xenograft mouse model. Importantly, cytarabine not only did not trigger, but rather inhibited, the KSHV lytic replication program, preventing virion production. Mechanistically, cytarabine inhibited both cellular and viral DNA and RNA syntheses and triggered the degradation of KSHV major latency-associated nuclear antigen (LANA), hence inducing cell stress and inhibiting KSHV persistent infection.

Identification of inhibitors of KSHV-transformed cells.
To identify inhibitors of KSHV-transformed cells, we conducted an HTS with libraries of small molecules using MM and KMM cells (10). The libraries consist of 3,731 individual compounds, including the 2,320 compounds of the Spectrum collection from MicroSource Discovery System Inc., covering a wide range of biological activities and structural diversities that are suitable for HTS programs; the NIH clinical collection, consisting of 781 compounds previously tested in clinical trials; the EMD Millipore kinase collection, consisting of 327 well-characterized, pharmacologically active, potent protein kinase and/or phosphatase inhibitors; and the EMD Millipore StemSelect small-molecule regulator collection, con-sisting of 303 pharmacologically active compounds, including extracellular domaintargeting reagents as well as cell-permeant reagents that regulate intracellular targets. We treated MM and KMM cells with 5 M concentrations of each compound for 48 h and counted the surviving cells following staining with 4=,6-diamidino-2-phenylindole (DAPI). We selected 50 compounds representing 1.3% of the libraries that induced cytotoxicity in Ͼ50% of KMM cells and Ͻ10% of MM cells ( Fig. 1A and B). Interestingly, most of the selected compounds (54%) are anti-inflammatory, followed by antibacterial (10%), antihypertensive (10%), antioxidant (4%), antiviral (2%), antiarthritic (2%), antiangiogenic (2%), and catecholaminergic (2%) (Fig. 1C). Among them, cytarabine, a cytidine analogue with a modified sugar moiety, i.e., arabinose instead of ribose, is currently used for treating leukemia and lymphomas (11) and hence is an interesting candidate that could be repurposed for KSHV-induced malignancies.
To confirm the inhibition efficacy of cytarabine on KSHV-transformed cells, we treated MM and KMM cells with 0, 2.5, 5, and 10 M cytarabine over a period of 120 h. Cytarabine preferentially inhibited the proliferation of KMM cells in a dose-dependent manner ( Fig. 1D and E). Consistently, cytarabine induced cell cycle arrest in KMM but not MM cells ( Fig. 1F and G). Furthermore, extended treatment with cytarabine for up to 15 days completely eliminated live KMM cells by day 9 at 10 M and by day 12 at both 2.5 and 5 M (Fig. 1I). In contrast, while cytarabine at 5 and 10 M had a slight inhibitory effect on MM cells, most cells survived up to day 12; however, most of the cells detached by day 15, albeit they remained alive (Fig. 1H).
Cytarabine inhibits the proliferation of diverse PEL lines. We tested the inhibitory efficacy of cytarabine on different PEL lines, including BCBL1, BC3, JSC1, and BCP1 cells, which are singly infected by KSHV, and BC1 cells, which are dually infected by KSHV and EBV. As there is no appropriate control for PEL cells, we included BJAB, a KSHV-and EBV-negative Burkitt's lymphoma cell line as a reference. At 0.5, 2, and 5 M, cytarabine effectively inhibited the proliferation of all PEL lines tested, which manifested a greater sensitivity than BJAB cells ( Fig. 2A to F). As a result, PEL lines had 50% inhibitory concentrations (IC 50 s) ranging from 0.44 M to 1.29 M while BJAB cells had an IC 50 of 2.34 M (Fig. 2G). Extended treatment of BCBL1 and BCP1 cells with 1 M cytarabine for up to 15 days completely killed all cells of both lines by day 9, indicating the lack of any emerging resistance ( Fig. 2H and I).
Cytarabine induces cell cycle arrest and apoptosis in PEL cells. To identify the mechanism of cytarabine-mediated cytotoxicity, we examined the effect of cytarabine on the cell cycle. Cytarabine at 5 M effectively induced cell cycle arrest in BCBL1 and BC3 cells after as little as 4 h of treatment (Fig. 3A). A similar result was observed with BJAB cells. Cytarabine at 5 M induced apoptosis in 63% of BCBL1 cells and 49% of BC3 cells following 4 h of treatment (Fig. 3B). Furthermore, treatment with 5 M cytarabine for 24 h induced apoptosis markers, including cleaved PARP1 (c-PARP1) and cleaved caspase 3 (c-caspase 3) in BCBL1, BC3, BC1, JSC1, and BCP1 cells (Fig. 3C). In contrast, cytarabine did not induce any apoptosis in BJAB cells ( Fig. 3B and C).

Cytarabine inhibits PEL initiation and progression and induces regression of grown PEL.
To examine the efficacy of cytarabine for PEL treatment, we employed a xenograft mouse model. We induced PEL in Nod/Scid mice by engrafting BCBL1-Luc cells. At day 3 postengraftment, we treated the mice with either phosphate-buffered saline (PBS) or a liposome form of cytarabine, DepoCyt, at 50 mg/kg of body weight every other day for 3 weeks. Mice treated with PBS started to gain weight at as early as 1 week as a result of PEL development, while those treated with cytarabine maintained relatively constant weight (Fig. 4A). At week 3 posttreatment, we performed live bioluminescence imaging and detected strong signals in all mice treated with PBS (mice a to j in Fig. 4B and C). However, mice treated with cytarabine (mice k to t) had no detectable signal at this time point, indicating that cytarabine completely inhibited PEL growth.
Next, we examined if cytarabine could control or induce regression of grown PEL. The mice engrafted with BCBL1-Luc cells for 3 weeks were randomly separated into two groups and treated with PBS or DepoCyt at 50 mg/kg every other day for 4.5 weeks. Mice treated with cytarabine had reduced weights (mice f, b, c, i, and e in Fig. 4D), indicating PEL regression, while those treated with PBS continued to gain weight (mice a, g, h, d, and j in Fig. 4D), indicating continuous PEL growth. At week 5 and 5.5 postengraftment, mice h and d and mice a, g, and j of the PBS group, respectively, died of PEL. Bioluminescence imaging at week 5 postengraftment showed that the remaining 3 mice in the PBS group had strong luminescent signals ( Fig. 4E and F). In contrast, the luminescent signals in all mice treated with cytarabine were dramatically reduced, with those in mice f, b, and i being reduced to almost undetectable levels, indicating that cytarabine effectively induced regression of most of the grown tumors ( Fig. 4E and F). We compared the survival rates for the two groups and observed a statistically significant increase in survival rate for mice treated with cytarabine compared to mice treated with PBS (100% survival rate at week 8 postengraftment for the cytarabine group versus 0% survival rate at week 6.5 for the PBS group) (Fig. 4G). These results demonstrated that cytarabine could be an effective drug for inhibiting PEL initiation and progression and inducing regression of grown PEL.
Cytarabine inhibits KSHV latent replication. Since the survival of PEL cells depends on KSHV latent infection and multiple copies of the viral episome, we determined if cytarabine might induce cytotoxicity by inhibiting KSHV latent infection. Treatment of BC1 and BCBL1 cells with cytarabine for 72 h decreased the level of intracellular KSHV DNA by at least half (Fig. 5A). Simultaneously, LANA transcript was reduced by at least half (Fig. 5B). Interestingly, LANA protein, which is essential for the replication and persistence of KSHV episome, was undetectable after 24 h of cytarabine treatment (Fig. 5C). Accordingly, Ͻ1 copy of KSHV genome per cell was detected after 3 days of cytarabine treatment compared to approximately 50 and 100 copies of KSHV genome per cell in the untreated BCP1 and BCBL1 cells, respectively (Fig. 5D). Since cytarabine can be incorporated into RNA and DNA by competing with intracellular nucleotides (12,13), we determined if it might directly inhibit KSHV latent replication and expression of KSHV latent genes. We pulsed BCBL1 cells with bromodeoxyuridine (BrdU) or 4-thiouridine (4sU) for 4 h and then immunoprecipitated the BrdU-labeled DNA or 4sU-labeled RNA to monitor de novo DNA or RNA synthesis, respectively. Whereas the amounts of newly synthesized total DNA and RNA continued to increase over a period of 4 h in control cells treated with dimethyl sulfoxide (DMSO), those of cytarabine-treated cells did not increase, indicating that cytarabine inhibited the de novo syntheses of total DNA and RNA (Fig. 5E). By using quantitative real-time PCR (qPCR) for BrdU-labeled cellular DNA (18S and ␤-actin) and KSHV DNA (LANA), we detected inhibition of both cellular and KSHV DNA syntheses as early as 30 min following cytarabine treatment (Fig. 5F). Interestingly, inhibition of ␤-actin and KSHV DNA syntheses seemed to be more efficient than 18S DNA synthesis. Similarly, by using reverse transcription-qPCR (RT-qPCR) for 4sU-labeled cellular RNA and KSHV RNA, we detected inhibition of ␤-actin and LANA RNA syntheses by cytarabine (Fig. 5G). Inhibition of LANA RNA synthesis, which could be observed as early as 15 min following cytarabine treatment, seemed to be more efficient than ␤-actin RNA synthesis. Interestingly, cytarabine treatment for up to 4 h had minimal effect on 18S RNA synthesis. Taken together, these results indicated that cytarabine inhibited the syntheses of cellular and KSHV DNA and RNA, which might account for its inhibitory effect on PEL cells and KSHV latent infection.

DISCUSSION
There is currently no specific and efficient therapy for PEL. The common recommendation is CHOP chemotherapy, which is a combination of cyclophosphamide, doxorubicin, vincristine, and prednisone (6). However, the outcome is dismal with a 1-year overall survival rate at 39.3% and an aggressive clinical course (9).
An anti-CD20 antibody has been developed for treating CD20 ϩ B-cell NHLs. This approach can be considered for treating some rare cases of PEL expressing CD20, as it has shown some effect on MCD (14). Autologous stem cell transplantation in combination with high-dose chemotherapy is another strategy for treating PEL (15,16). Finally, in HIV-infected PEL patients, targeting HIV infection and restoring immune functions by combined antiretroviral therapy (cART) alone has shown excellent outcomes (17). However, cART remains problematic because of potential adverse effects due to drug-drug interactions, which increase chemotherapy toxicities (18).
We have identified numerous "known" small molecules that inhibit the proliferation of KSHV-transformed cells but have minimal cytotoxicity to uninfected primary cells. These compounds are well characterized, with most of them currently in clinical use or trials, and hence are potential candidates for repurposing for KSHV-induced malignancies. Among them, cytarabine is effective in inducing cytotoxicity to PEL cells.
Cytarabine, or cytosine arabinoside, was first synthesized in 1959 and approved for clinical usages by the FDA in 1969 (11). This drug has been used in therapy for several blood cancers such as AML, ALL, and NHLs (11). Cytarabine has recently been utilized for treating meningeal leukemias, lymphomas, and recurrent embryonal brain tumors (19). Because of these multiple indications of treatment, the mechanism of action, which is mainly based on its inhibitory effect on DNA synthesis, the pharmacokinetics, and the toxicity of this compound in patients are well described (20). We have shown that cytarabine has a strong effect on PEL cells, inducing arrest and death in vitro, and that it completely abrogated tumor progression and induced regression of grown tumors in a PEL mouse model. These results indicate that cytarabine might be an ideal candidate for repurposing for PEL therapy. Numerous studies have recently demonstrated that targeting pathways involved in differentiation and survival is a promising strategy for treating PEL. Activation of p53 with Nutlin-3 disrupted p53-MDM2 interaction, induced apoptosis of PEL cells, and inhibited PEL progression (21). Silencing BLIM-1, a transcription factor involved in B-cell differentiation, led to PEL cell death (22). Triptolide inhibited cell proliferation and PEL progression by suppressing STAT3 activity, IL-6 secretion, and LANA expression (23). Chloroquine, an inhibitor of autophagy, induced a caspase-dependent apoptosis (24). Finally, treatment with the thymidine analogue azidothymidine (AZT) sensitized PEL cells to Fas ligand and TRAIL-mediated apoptosis and might be sufficient to restore T cell control via KSHV-specific T CD4 ϩ response (25).
We have demonstrated that cytarabine induces cell cycle arrest and apoptosis in PEL cells by inhibiting DNA and RNA syntheses. Cytarabine can theoretically be incorporated into the DNA of any proliferating cells to induce DNA damage (12). It can also be incorporated into RNA to inhibit its polymerization and therefore has an inhibitory effect on resting cells as well (13). In cells, cytarabine competes with the endogenous cytidine during nucleic acid synthesis after conversion into its triphosphate. Because the arabinose sugar of cytarabine sterically hinders the rotation of the molecule in DNA, it inhibits DNA replication. We have shown that cytarabine induces cell cycle arrest but not cell death in BJAB, an immortalized proliferating Burkitt's lymphoma B-cell line. This effect could be due to the incorporation of cytarabine into cellular DNA during S phase, resulting in the cytotoxicity on "normal" proliferating cells. However, in PEL cells, we have observed that cytarabine induces both cell cycle arrest and apoptosis, indicating that an alternative mechanism might be involved in its cytotoxic effect. Indeed, PEL cells are universally associated with KSHV infection and highly depend on KSHV latent proteins (3). We have shown that, in addition to cellular DNA and RNA, cytarabine inhibits KSHV DNA and RNA, resulting in the inhibition of replication of the KSHV latent genome. This might explain the higher sensitivity of PEL cells than BJAB cells to cytarabine.
Nucleoside analogues are known to have an efficient inhibitory effect on herpesviral persistent infections. By impairing the interaction between herpes simplex virus 1 (HSV-1) DNA and the cellular nucleosomes, cytarabine can inhibit HSV-1 replication (26). However, HBV is resistant to this drug but sensitive to its close relative the adenosine analogue 9-beta-D-arabinofuranosyladenine (ara-A) (27). The chemically close fluoroiodocytosine analogue 1-(2-deoxy-2-fluoro-beta-D-arabinofuranosyl)-5-iodocytosine (FIAC) is efficient against HSV-1, HSV-2, and EBV (28,29). Because of its antitumor and antiviral effects, it would be interesting to test if cytarabine is effective against cancers associated with other oncogenic viruses.
KSHV latently infected cells constitute a viral reservoir in the host. Upon stimulation by stress, inflammatory cytokines, calcium ionophores, phorbol ester, or histone deacetylase inhibitors (trichostatin A and NaB), KSHV can be reactivated from latency, expressing cascades of lytic genes and producing progeny virions (30)(31)(32). KSHV lytic replication in a small number of cells is essential for the spread and progression of early-stage KS (33). Numerous chemotherapies can induce reactivation of HSV-1 and EBV from latency, raising concerns about these therapeutic approaches (34,35). Our results show that cytarabine not only does not induce KSHV reactivation but also robustly inhibits the viral lytic program induced by NaB. In addition to KSHV, over 70% of PEL cases are coinfected by EBV, which itself is associated with several cancers (36). Similarly to KSHV, EBV lytic replication participates in the spreading and pathogenesis of EBV-associated malignancies (37). Since cytarabine does not reactivate EBV from latency, it can be used for PEL treatment independently of the EBV status (38). Compared to other chemotherapies, the advantage of using cytarabine in PEL treatment is its multiple cellular and viral effects without increasing the spread of KSHV and tumor cells.
Taken together, we have identified cytarabine as an excellent candidate for reposition for treating PEL patients. It displays a strong inhibitory effect on PEL through multiple mechanisms. Since cytarabine is currently in clinical use and approved for treating other blood malignancies, it would be interesting to carry out advanced clinical trials to evaluate its usage and identify the optimal doses for treating PEL.  Targeting KSHV Latency for Treating PEL ® BJAB and PEL cell lines (JSC1, BCBL1, BC3, BC1, and BCP1) were maintained in RPMI 1640 supplemented with 20% FBS and antibiotics (39).

MATERIALS AND METHODS
High-throughput screening. MM and KMM cells were seeded in 96-well plates at 5,000 cells/well for 16 h and then treated with small molecules at 5 M final concentrations in 0.1% DMSO for 48 h. Cells treated with 0.1% DMSO and Bay11 were used as a negative control and a positive control, respectively. Cells were washed 2 times with 1ϫ PBS and fixed with 4% paraformaldehyde for 15 min at room temperature prior to DAPI staining. Live cells were automatically counted with the Cellomics ArrayScan VTI HCS reader (Thermo Scientific), and the results were analyzed with the HCS Studio cell analysis software (Thermo Scientific). The number of live cells in the DMSO control was set as 100% and used to normalize cells treated with different compounds. A total of 73 hits that gave at least 50% cytotoxicity to KMM cells but less than 10% cytotoxicity to MM cells were selected from the first round of screening and then validated in a secondary screening, resulting in the selection of 50 final compounds. Secondary screening was carried out on the ImageXpress Micro System (Molecular Devices), and the surviving cells were automatically quantified using MetaXpress software (Molecular Devices). Compounds that showed effects similar to those of the first-round screening with less than 20% variation were selected.
Cell proliferation assay. MM, KMM, and PEL cells plated at a density of 200,000 cells/well and treated with different reagents, including DMSO, cytarabine, or sodium butyrate (NaB), were counted daily using a Malassez chamber.
De novo RNA synthesis. RNA was extracted from cells labeled with 500 M 4-thiouridine (4sU) (T4509; Sigma-Aldrich) for the indicated times in the presence of DMSO or 5 M cytarabine in complete medium. RNA at 20 g was biotinylated with 0.2 mg/ml of EZ-Link HPDP-biotin (N-[6-(biotinamido)hexyl]-3=-(2=-pyridyldithio)-propionamide) (catalogue no. 21341; Thermo Fisher Scientific) for 2 h at room temperature and then subjected to phenol-chloroform extraction and isopropanol precipitation to remove the unlabeled HPDP-biotin. The biotinylated-RNA fraction pellet was resuspended in 100 l nuclease-free water and incubated with an equal volume of washed Dynabeads MyOne streptavidin C1 beads (catalogue no. 65001; Thermo Fisher Scientific) for 30 min at room temperature. After being washed 3 times with washing buffer, the labeled RNA was eluded with 100 l of 100 mM dithiothreitol (DTT), extracted by phenol-chloroform, and precipitated with isopropanol. The final pellet was resuspended in 10 l of nuclease-free water. cDNA was synthesized using 50 ng of RNA and specific primers for ␤-actin, 18S, and LANA.
De novo DNA synthesis. DNA was extracted from cells labeled with 10 M BrdU (catalogue no. B5002; Sigma-Aldrich) for the indicated times in the presence of DMSO or 5 M cytarabine in complete medium. DNA at 1 g was resuspended in 100 l nuclease-free water and fragmented with the DpnII restriction enzyme for 90 min at 37°C. Digested DNA was denatured at 99°C for 10 min, cooled instantly in ice for 5 min, and incubated with 5 g of a biotinylated anti-BrdU antibody (ab2284; Abcam) in 100 l of RIPA buffer at 4°C for 16 h. The biotinylated DNA was extracted with phenol-chloroform and precipitated with isopropanol. The biotinylated DNA pellet was resuspended in 50 l nuclease-free water and mixed with an equal volume of Dynabeads MyOne streptavidin C1 beads (catalogue no. 65001; Thermo Fisher Scientific) in 50 l of RIPA buffer for 2 h at 4°C. After 3 washes with washing buffer, the labeled DNA was eluted by boiling for 10 min in 100 l 0.1% SDS solution. The DNA was extracted with phenol-chloroform, precipitated with isopropanol, and resuspended in 10 l of nuclease-free water. The DNA was examined by qPCR with specific primers for ␤-actin, 18S, and LANA.
Animal experiments. For the tumor initiation experiment, 20 female Nod/Scid mice at 5 weeks old were each intraperitoneally injected with 10 7 BCBL1 cells expressing luciferase (BCBL1-Luc). At day 3 postinoculation, mice were treated with PBS or a liposome form of cytarabine (DepoCyt) at 50 mg/kg every other day for 3 weeks and weighed twice a week. At week 3 postinoculation, mice were injected with luciferin at 50 mg/kg and imaged with an IVIS Spectrum in vivo imaging system (PerkinElmer). The region-of-interest (ROI) signals based on the (p/s)/(microwatts/square centimeter) formula were analyzed with Living Image software (PerkinElmer).
For the tumor regression experiment, 10 Nod/Scid mice each engrafted for 3 weeks with 10 7 BCBL1-Luc cells were randomly split into 2 groups. One group was treated with PBS, the other group was treated with DepoCyt at 50 mg/kg every other day for 4.5 weeks, and both groups were weighed twice a week. At week 5 postinoculation (i.e., week 2 posttreatment), live imaging was performed.
The protocols for the animal experiments were approved by the University of Southern California Institutional Animal Care and Use Committee under protocol number 11722.
Statistical analysis. Statistical analysis was performed using the two-tailed t test, and a P value of Յ0.05 was considered significant. Single, double, and triple asterisks in figures represent P values of Յ0.05, Յ0.01, and Յ0.001, respectively, while NS indicates not significant. For the survival study, Kaplan-Meier survival analysis was performed and statistical significance was calculated using the log rank test.