Narciclasine, an isocarbostyril alkaloid, has preferential activity against primary effusion lymphoma

Primary effusion lymphoma (PEL) is a subtype of non-Hodgkin lymphoma associated with infection by Kaposi sarcoma-associated herpes virus (KSHV). PEL is an aggressive disease with extremely poor prognosis when treated with conventional chemotherapy. Narciclasine, a natural product present in Amaryllidaceae family of flowering plants including daffodils, belongs to a class of molecules termed ‘isocarbostyril alkaloid’. We have found that narciclasine displays preferential cytotoxicity towards PEL at low nanomolar concentrations and is approximately 10 and 100-fold more potent than its structural analogs lycoricidine and lycorine, respectively. Narciclasine arrested cell-cycle progression at the G1 phase and induced apoptosis in PEL, which is accompanied by activation of caspase-3/7, cleavage of PARP and increase in the surface expression of Annexin-V. Although narciclasine treatment resulted in a marked decrease in the expression of MYC and its direct target genes,time-course experiments revealed that MYC is not a direct target of narciclasine. Narciclasine treatment neither induces the expression of KSHV-RTA/ORF50 nor the production of infectious KSHV virions in PEL. Finally, narciclasine provides dramatic survival advantages to mice in two distinct mouse xenograft models of PEL. In conclusion, our results suggest that narciclasine could be a promising agent for the treatment of PEL.

and lycorine are structural analogs of narciclasine. To identify whether the structural analogs of narciclasine also display preferential cytotoxicity towards PEL, we treated the same panel of hematological cancer cell lines with increasing concentrations lycoricidine and lycorine for 72 hours. Similar to narciclasine, its closely related structural analog lycoricidine also displayed preferential cytotoxicity towards PEL cell lines with IC 50 ranging from 82 to 162 nM ( Fig. 1B & Table 1). In contrast, the IC50 of lycoricidine for non-PEL cell lines ranged from 224 to 426 nM ( Fig. 1B & Table 1). Lycorine, the other structural analog of Narciclasine, also displayed a similar trend in cytotoxicity ( Fig. 1B & Table 1) although it was much less potent. Thus, even though narciclasine and its structural analogs show similar trend in preferential cytotoxicity towards PEL, the IC50 dose of narciclasine is approximately 10 and 100-fold lower than that of lycoricidine and lycorine, respectively. Narciclasine induces cell-cycle arrest in PEL. To identify the mechanism by which narciclasine inhibited the growth of PEL cells, we examined the effect of narciclasine on cell-cycle progression using flow cytometry.  Treatment of PEL cells with narciclasine resulted in G 1 arrest as observed by a marked increase in the number of cells in the G 1 phase along with a concomitant decrease in cells in the S-phase ( Fig. 2A). Narciclasine treatment also resulted in a significant increase in the number of cells in sub-G 1 phase indicative of apoptosis. Further, Narciclasine treated PEL cells stained brightly with SYTOX Green, a membrane-impermeable nuclear Narciclasine and its structural analogs induce apoptosis in PEL. To confirm whether narciclasine induces apoptotic cell death in PEL, we treated the PEL cell lines BC-1, BC-3 and JSC-1 with narciclasine for 36 hours followed by staining with AnnnexinV-FITC. Moderate and dramatic increases in annexinV-FITC staining were observed in PEL cells treated with narciclasine at the doses of 25 nM (2 to 3-fold) and 50 nM (8 to 10-fold), as compared with cells that were treated with a vehicle control (Fig. 3A). In contrast, only a mild increase in annexinV-FITC staining was observed in a non-PEL cell line (L428) that had been treated with narciclasine under similar conditions (maximum of 2-fold at 50 nM), confirming the preferential activity of narciclasine towards PEL (Fig. 3A). Consistent with the above results, narciclasine treatment activated caspase-3/7 and increased PARP cleavage, two hallmarks of apoptosis, only in PEL cell lines (Fig. 3B,C). Similar results were obtained with lycoricidine and lycorine ( Supplementary Fig. S1).
MYC is not a direct target of narciclasine. Recently, we have shown that the agents that downregulate MYC are selectively toxic to PEL 8,9 . Since, narciclasine also display preferential activity against PEL, we tested the effect of narciclasine on the expression of MYC in PEL. BC-1, BC-3 and JSC-1 cells were treated with narciclasine (25 nM) for 48 hours followed by immunoblotting for MYC and GAPDH (loading control). As shown in Fig. 4A, narciclasine treatment resulted in a dramatic decrease in the level of MYC in PEL cell lines. In contrast, narciclasine had no significant inhibitory effect on MYC expression in L428, a non-PEL cell line (Fig. 4A). We used qRT-PCR analysis to examine the effect of narciclasine on the expression of four known MYC target genes TERT, SLC19A1, MYB and PMM2. Narciclasine treatment induced a significant reduction in expression of all four MYC target genes in both BC-1 and BC-3 cell lines (Fig. 4B). The above results indicated that expression of MYC is downregulated by narciclasine in PEL. To check, whether MYC is the primary target of narciclasine, we performed time-course experiments by treating PEL cells with narciclasine for 12, 24, 36 and 48 hours followed by immunoblotting for MYC, PARP and caspase-3. As shown in Fig. 5, treatment of PEL cells with narciclasine for 12 hours resulted in cleavage of PARP and caspase-3 cleavage, indicating induction of apoptosis. However, a significant reduction in the level of MYC was seen only after 36 hours of treatment (Fig. 5). Collectively, the above results suggest that although MYC expression is downregulated in PEL upon treatment with narciclasine, it is not a direct target of narciclasine.
Inhibiting Rho pahway has no effect on the activity of narciclasine in PEL. Narciclasine has been shown to induce stress fiber formation by activating RhoA in gliobastoma cells 6 . However, in PEL cells (BC-1 and BC-3) narciclasine did not induce stress fiber formation as observed by F-actin staining (data not shown). Further, two known inhibitors of Rho pathway, CCG1423 and Rhosin-HCl 10,11 , did not have any effect on the activity of narciclasine in PEL cells (Fig. 6). These results are not surprising because activation of Rho and stress www.nature.com/scientificreports www.nature.com/scientificreports/ fiber formations are usually observed in adherent cells 12 . Collectively, these results indicate that Rho pathway has no role in the activity of narciclasine in PEL.

Lack of induction of lytic replication by Narciclasine in PEL.
The defining feature of PEL is the presence of KSHV in all tumor cells. Disruption of Myc by shRNA results in the induction of KSHV replication and transcription activator (RTA), a master regulator and marker for lytic reactivation 13 . Therefore, a potential concern is that narciclasine treatment may induce RTA expression and lytic reactivation of KSHV. Interestingly, treatment of PEL cells with narciclasine did not induce RTA expression as determined by immunoblotting (Fig. 7A). As a positive control, we used BCBL-TREx-RTA cells, which express RTA from tetracycline-inducible promoter and undergo a complete cycle of viral replication upon treatment with doxycycline (Fig. 7A). Earlier, we have made a reporter cell line termed 293A-PAN-Luc, which expresses the firefly luciferase gene under the control KSVH PAN promoter and responds to infectious KSHV virions by upregulating the luciferase activity 14 .
To test whether narciclasine treatment of PEL cells produces infectious KSHV virions, cell-free supernatants from narciclasine-treated PEL cells and doxycycline-treated BCBL-TREx-RTA (positive control) were incubated with 293A-PAN-Luc cells for 48 hours in the presence of polybrene, followed by measurement of firefly luciferase activity. A striking increase in luciferase activity was observed in wells that were incubated with supernatants from doxycycline treated BCBL-TREx-RTA cells (Fig. 7B). In contrast, the luciferase activity of the wells treated with supernatants from narciclasine treated PEL cells stayed near the base line (Fig. 7B). These results suggest that narciclasine treatment neither reactivates KSHV lytic cycle nor produces infectious virions. narciclasine displays potent in vivo growth inhibitory activity in orthotopic CLX and PDX mouse models of PEL. To check the in vivo efficacy of narciclasine against PEL, we employed two distinct mouse models of PEL -a Cell Line Xenograft (CLX) model using BC-1 cells expressing an RSV-promoter-driven firefly luciferase (BC-1-Fluc) and a Patient Derived Xenograft (PDX) model using UMPEL-1 cells 15 . In the CLX model, 1 × 10 7 BC-1-Fluc cells were intraperitoneally injected into NSG mice on day 1. Bioluminescence imaging (BLI) was performed on day 5 to confirm the presence of tumors (Fig. 8A). Animals with established tumors were randomly assigned to vehicle control (10% HPCD) or narciclasine (1 mg/kg once daily for 12 days). Narciclasine treatment inhibited the growth of established PEL as seen by significant decreases in the BLI signal (Fig. 8A,B) and body weight gain, a surrogate for tumor-induced ascites (Fig. 8C). Further, there was a striking increase in the median survival of mice that received narciclasine (median survival = 37 days) in comparison to vehicle control mice (median survival = 24 days) (Fig. 8D). BC-1 cells produce and secrete cytokines IL-6 and IL-10. As another surrogate marker for the presence of tumor cells, we measured the levels of circulating human IL-6 and IL-10 in the NSG mice bearing BC-1-Fluc cells on day 19 after tumor inoculation. As shown in Fig. 8E, circulating human IL-6 and IL-10 were readily detected in the plasma of NSG mice bearing BC-1-Fluc cells that were treated with vehicle control. In contrast, no circulating IL-6 and IL-10 were seen in mice treated with narciclasine, confirming the effective control of PEL by narciclasine in vivo.
For PDX model, 1 × 10 7 patient-derived UMPEL-1 cells were intraperitoneally injected in to NSG mice. 4 days after tumor inoculation, animals were randomly assigned to vehicle control and narciclasine (1 mg/kg once daily for 9 days). As shown in Fig. 9A,B, treatment with narciclasine inhibited the growth of patient-derived UMPEL-1 cells as seen by a significant decrease in body weight gain over the experimental period. Moreover, there was a 2-fold increase in the median survival of mice that received narciclasine (median survival = 30 days) as compared with vehicle control mice (median survival = 15 days) (Fig. 9C). In contrast to BC-1 cells, UMPEL-1 www.nature.com/scientificreports www.nature.com/scientificreports/ cells produce and secrete only human IL-10. We did not detect any circulating human IL-10 in animals inoculated with patient-derived UMPEL-1 cells that were treated with narciclasine (Fig. 9D). In contrast, significant amount of human IL-10 was detected in vehicle control treated animals bearing patient-derived UMPEL-1 cells. In the same setting, a group of animals were also treated with narciclasine analog -lycoricidine (1 mg/kg once daily for 9 days). As shown in Supplementary Fig. S2, lycoricidine has a moderate but significant inhibitory effect on the growth of patient-derived UMPEL-1 cells as determined by a significant decrease in body weight gain, increase in median survival (by 6 days) along with a significant decrease in circulating human IL-10. The moderate effect of lycoricidine in vivo is in clear contrast to the striking effect of narciclasine at similar treatment doses, which confirms the in vitro cytotoxicity data suggesting the superior activity narciclasine over its structural analogs.

Discussion
PEL accounts for approximately 4% of all HIV-related non-Hodgkin lymphoma and is universally associated with KSHV infection 16 . The median overall survival of PEL is only 4.8 months, indicating an urgent need to identify new therapeutic options for this disease 17 . In this study, we found that narciclasine and its structural analogs have preferential activity against PEL. Preclinical studies of narciclasine in other cancers have shown that its mechanism of action involves inhibition of translation elongation factor eEF1A 4 , activation of RhoA 6 , and induction of autophagy 18 . It is possible that all these pathways play a role in the activity of the narciclasine against PEL. However, all those pathways are essential for the survival of both cancer cells and normal cells. In this study, we found that all the cancer cells in the panel were susceptible to narciclasine induced cell death at higher doses, while cell lines representing PEL succumbed to narciclasine at doses that were approximately 2.5-fold lower than the cell lines representing other cancers. This led us to hypothesize that narciclasine may downregulate MYC in PEL as MYC is essential for the survival of PEL and molecules that target MYC are effective and selective against PEL 8,9 . Indeed, we found that narciclasine downregulated the expression of MYC with a concomitant decrease in the expression of its direct target genes. However, time-course experiments in PEL upon treatment with narciclasine revealed that apoptosis was initiated before the downregulation of MYC, suggesting that MYC is not a direct target of narciclasine in PEL.
Downregulating MYC expression in PEL using shRNA leads to lytic reactivation of KSHV by activating RTA/ ORF50, the master regulator of KSHV lytic cycle 13 . As lytic reactivation of KSHV is believed to promote KSHV tumorigenesis, it can be a potential safety issue with the use of narciclasine in PEL 19 . However, we found that www.nature.com/scientificreports www.nature.com/scientificreports/ narciclasine treatment neither activated RTA nor induced any infectious KSHV particles, thereby allaying those safety concerns. These results are consistent with our observation that MYC is not a direct target of narciclassine.
Narciclasine has been shown to activate Rho signaling and induce stress fiber formation in glioblastoma cells 6 . However, our study suggests that Rho pathway has no role in the activity of narciclasine in PEL. The KSHV genome contains several genes that mimics the activity of cellular genes, thereby activating multiple oncogenic signaling pathways (Akt, MAPK, NF-κB, Notch, Wnt/β-catenin) 20 . At present, it is not known whether narciclasine has any effect on any of the above-mentioned oncogenic signaling pathways present in PEL. Therefore, further studies are needed to identify the primary target of narciclasine in PEL. Apart from its potent anticancer activities, narciclasine has been shown to possess anti-inflammatory properties by inhibiting NF-κB pathway in murine models of inflammation 21 . NF-κB pathway is constitutively active in a number of PEL cell lines [22][23][24] and has been shown to be essential for their survival and proliferation 24,25 . Therefore, it is conceivable that inhibition of survival pathways, such as NF-κB, also contributes to the activity of narciclasine against PEL cells.
Even though narciclasine has shown impressive activity in various pre-clinical studies of cancers, it has never been tested in a human clinical trial. A potential bottleneck for the clinical testing of narciclasine is its limited availability. Until now, narciclasine has been isolated from the bulbs of narcissus. However, recent studies have successfully demonstrated organic routes for chemical synthesis of narciclasine and its analogs 26,27 . In this study, we show that compounds that are structurally related to narciclasine also display similar activity, albeit at higher doses. Therefore, structural modifications of narciclasine can be made to improve its efficacy and bio-availability. Finally, it is pertinent to note that concentrations as high as ~250 nM of narciclasine were readily achievable in murine pharmacokinetics studies, indicating the presence of a significant therapeutic window, which can be exploited for the treatment of PEL 3 .  www.nature.com/scientificreports www.nature.com/scientificreports/ at 37 °C and 5% CO 2. After the cell lines were received they were expanded and multiple vials were frozen in liquid nitrogen. None of the cell lines used in this study were cultured continuously for longer than 3 months. After 3 months, a new frozen vial was thawed and used for experiments. PEL cells were authenticated by their expression of KSHV-LANA. No further authentication of cell lines characteristics was done.
Cell viability, cell-cycle and apoptosis analysis. Cells from exponentially growing cultures were plated in untreated flat-bottom 96 well plates at a density of 1 × 10 4 cells/well, treated with an increasing concentration of the drugs and subsequently assessed for cell viability using the MTS reagent (3-4,5-dimethylthiazol-2yl)-5-(3 -carboxy -methoxyphenyl)-2-(4-sulfophenyl)-2H-tet razolium, inner salt) following the manufacturer's instructions. Percent cell survival was calculated based on the reading of cells grown in the presence of DMSO control, as described previously 29 . IC 50 values were calculated using Graphpad prism. Experiments were performed in triplicate. Cell-cycle analysis was done as described previously 30 . Apoptosis was analyzed using Apo-ONE ® Homogeneous Caspase-3/7 Assay kit and FITC AnnexinV apoptosis detection kit I as per manufacturer instructions, as described previously 28 . Luciferase assay for the presence of KSHV infectious virions. 293A-PAN-Luc cells engineered to express a stably integrated copy of PAN promoter-driven luciferase reporter construct were plated in 24 well plates and treated with cell-free supernatants from PEL cell lines treated with narciclasine or doxycycline (500 ng/ ml) treated BCBL-TREx-RTA supernatants for 48 hours in the presence of polybrene (8 µg/ml). Post incubation, cells were lysed to assay the firefly luciferase activity as described earlier 32 . Cell line xenograft (CLX) mouse model of PEL. A total of 1 × 10 7 BC-1 cells stably expressing firefly luc were injected intraperitoneally into approximately 6-week-old female NOD.Scid-Gamma (NSG) mice (purchased from Jackson laboratory or from the breeding colony maintained in house). To assess establishment of tumors, mice were imaged 5 days post inoculation using an IVIS Spectrum Imaging system (Perkin Elmer's, Waltham, MA, USA). Animals were randomly assigned into two groups that were treated with once daily intraperitoneal injections of vehicle control (10% HPCD) or narciclasine (1 mg/kg) for 12 days, continuously. The animals were monitored for survival and body weight gain was measured every 3 days as a surrogate measure of tumor progression 9 . All animal handling procedures were performed with the approval of the University of Southern California (USC) Institutional Animal Care and Use Committee (IACUC), in accordance with ethical guidelines and regulations.

Patient derived xenograft (PDX) mouse model of PEL.
A total of 1 × 10 7 patient derived UMPEL-1 cells that were passaged only in mice were injected intraperitoneally into approximately 6-week-old female NSG mice 15 . 4 days after tumor cells inoculation, mice were randomly divided into three groups and were treated with a once daily intraperitoneal injections of vehicle control (10% HPCD) or narciclasine (1 mg/kg) or lycoricidine (1 mg/kg) for 9 days. The animals were monitored for survival and body weight gain over the experimental period. All animal handling procedures were performed with the approval of the University of Southern California (USC) Institutional Animal Care and Use Committee (IACUC), in accordance with ethical guidelines and regulations.
Statistical analysis. Two-tailed unpaired Student t test was used to test for differences between 2 groups.
Differences with a P ≤ 0.05 were considered statistically significant. All experiments were repeated a minimum of two times.

Data availability
The data and reagents will be available up on request to senior author P.M.C. and/or R.G.