Sulfonoquinovosyl diacylglyceride selectively targets acute lymphoblastic leukemia cells and exerts potent anti-leukemic effects in vivo

DNA topoisomerase II inhibitors e.g. doxorubicin and etoposide are currently used in the chemotherapy for acute lymphoblastic leukemia (ALL). These inhibitors have serious side effects during the chemotherapy e.g. cardiotoxicity and secondary malignancies. In this study we show that sulfonoquinovosyl diacylglyceride (SQDG) isolated from Azadirachta indica exerts potent anti-ALL activity both in vitro and in vivo in nude mice and it synergizes with doxorubicin and etoposide. SQDG selectively targets ALL MOLT-4 cells by inhibiting catalytic activity of topoisomerase I enzyme and inducing p53 dependent apoptotic pathway. SQDG treatment induces recruitment of ATR at chromatin and arrests the cells in S-phase. Down-regulation of topoisomerase I or p53 renders the cells less sensitive for SQDG, while ectopic expression of wild type p53 protein in p53 deficient K562 cells results in chemosensitization of the cells for SQDG. We also show that constant ratio combinations of SQDG and etoposide or SDQG and doxorubicin exert synergistic effects on MOLT-4 cell killing. This study suggests that doses of etoposide/doxorubicin can be substantially reduced by combining SQDG with these agents during ALL chemotherapy and side effects caused can be minimized. Thus dual targeting of topoisomerase I and II enzymes is a promising strategy for improving ALL chemotherapy.

used in combination with other anti-ALL agents during remission induction and consolidation phases of the chemotherapy, respectively. Doxorubicin treatment shows severe cardiotoxicity while etoposide treatment often induces therapy related secondary malignancies 3,4 . Therefore, development of effective, selective and safe anti-ALL chemotherapeutic agents is justified to improve the ALL chemotherapy.
DNA topoisomerases are well established molecular targets of several anticancer and antibacterial drugs and currently many topoisomerase inhibitors are used in the chemotherapy of different cancers 5,6 . Topoisomerases control DNA topology by transiently cleaving and religating DNA strands, during various DNA metabolic processes. In the nucleus of mammalian cells two principal types of topoisomerases are present: topoisomerase I (topo I) and topoisomerase II (topo II) 7,8 . Enzymatic reaction cycle of topo I proceeds through four different steps-(a) DNA binding: the enzyme binds on its preferential binding site, (b) DNA nicking: nucleophilic attack on phosphodiester backbone by the active site tyrosine-723 residue and formation of transient 3′ phosphotyrosyl bond, (c) controlled strand rotation and (d) religation of DNA strand. Under normal conditions the nicking of DNA strand is transient and nick is rapidly sealed after strand rotation. Trapping of the enzyme and DNA in nicked condition by topo inhibitors i.e. stabilization of covalent enzyme-DNA complexes, generates DNA lesions, initiates cell cycle arrest and ultimately induces apoptosis. Several inhibitors are known for topoisomerase enzymes and are classified into two categories: (a) the catalytic topoisomerase inhibitors which inhibit catalytic activity of the enzymes prior to DNA binding and (b) the topoisomerase poisons which inhibit either controlled strand rotation or religation step of enzymatic reaction cycle and thus stabilizes covalent enzyme-DNA complexes 9 . Camptothecin (CPT) is an alkaloid isolated from the bark of the Chinese tree, Camptotheca acuminate 10 . CPT stabilizes topo I-DNA covalent complexes by inhibiting the religation step 11 .
To attain high proliferation, cancer cells necessitate higher activity of topoisomerases and indeed these enzymes have been shown to overexpress in several tumors and cancer cell lines 12,13 , e.g. topo I is overexpressed in ALL cells and also in MOLT-4 cells 12,14 . p53 is involved in multicellular processes e.g. cell cycle arrest, senescence, apoptosis and DNA repair 15 . During DNA damage p53 initiates two specific responses viz. cell cycle checkpoint activation and cell death via apoptosis. The prerequisite for each of these functions appears to depend on: the cell type, the cell environment and other cellular genetic alterations 16 . p53 plays important role in chemosensitivity and chemoresistance of cancer cells 17 . Mutations in the TP53 gene often lead to drug resistance, while wild type p53 protein plays important role in chemosensitivity of anti-cancer agents 18 .
Sulfonoquinovosyl diacylglyceride (SQDG) is a member of plant sulfolipids. SQDG was first reported in photosynthetic bacteria and higher plants by Benson and coworkers 19 . SQDG used in the present study was isolated by chromatographic separation of methanolic extract of the leaves of A. indica and characterized by extensive 2D-NMR and mass spectroscopy (Fig. 1a) 20 . SQDG has been reported for its anti-leukemic, anti-bacterial and anti-viral activities 20,21 . In this study we show that SQDG inhibits topo I enzyme of MOLT-cells, generates DNA replication stress, arrests the cells in S-phase and induces p53 dependent apoptotic pathway. Combinations of SQDG with etoposide and doxorubicin exert synergism and SQDG treatment reduces tumor growth in the nude mice xenografted with MOLT-4 cells.

SQDG catalytically inhibits topo I enzyme and prevents camptothecin mediated formation of topo I-DNA covalent complexes in vitro.
In order to check the inhibitory effect of SQDG on topo I enzymatic activity, in vitro DNA relaxation assay was performed using topo I enzyme and supercoiled pBS DNA in the presence of different concentrations of SQDG. At 3 μ M SQDG concentration, complete inhibition of the topo I relaxation activity was observed (Fig. 1b). Preincubation of the enzyme with SQDG for 5 minutes, before adding supercoiled pBS DNA, markedly enhanced the inhibition and relaxation activity was completely inhibited at 1.5 μ M SQDG (Fig. 1c). Preincubation dilution assay was also performed to make sure that SQDG bound form of the enzyme is inactive. After 5 minutes preincubation of the enzyme with SQDG, reaction mixture was diluted to 10 folds with the reaction buffer. After the dilution supercoiled DNA was added and relaxation assay was performed ( Supplementary Fig.  S1). Dilution of the reaction mixtures did not affect topo I inhibition caused by SQDG suggesting that SQDG bound form of the enzyme remains inactive. On the other hand SQDG did not inhibit DNA topoisomerase II alpha enzyme at 5, 10 and 15 μ M concentrations (Fig. 1d).
To find whether SQDG is catalytic inhibitor or topo I poison, in vitro oligonucleotide cleavage assay was performed using synthetic oligonucleotide with topo I binding site (Fig. 2a). Camptothecin (CPT) is a topo I poison which inhibits religation step of the enzymatic reaction cycle and stabilizes covalent topo I-DNA complexes 6,11 . Oligonucleotide cleavage assay in the presence of 10 μ M and 20 μ M CPT substantially increased the amount of cleaved 12 mer oligo compared to control, as CPT stabilizes topo I-oligo covalent complexes (Fig. 2b). Betulinic acid (BA) is a catalytic inhibitor of topo I 22,23 and was used as another control. In presence of 10 μ M and 20 μ M SQDG, we did not observe any increase in the amount of cleaved 12 mer oligo compared to the control reaction, indicating that SQDG is not a topo I poison like CPT but it is a catalytic topo I inhibitor like BA. When 20 μ M BA or 20 μ M SQDG was added in the reactions and incubated for 5 minutes prior to addition of 20 μ M CPT, the amount of cleaved 12 mer oligo substantially reduced in comparison to CPT alone reactions, indicating that SQDG/BA abrogates CPT mediated topo I-oligo covalent complex formation (Fig. 2b). Besides, DNA cleavage assay was also performed to revalidate the results of oligonucleotide cleavage assay. SQDG (10 μ M and 20 μ M) did not Scientific RepoRts | 5:12082 | DOi: 10.1038/srep12082 stabilize the topo-DNA complexes while CPT (10 μ M and 20 μ M) substantially increased formation of the topo-DNA complexes (Fig. 2c). The assay further confirmed that SQDG is a catalytic topo I inhibitor.
In order to study effect of SQDG on the DNA binding of topo I enzyme electrophoretic mobility shift assay was performed using synthetic oligonucleotide containing topo I binding site (Fig. 2a,d). Results of EMSA indicated that in the presence of 5 and 10 μ M SQDG, binding of topo I to oligonucleotide was decreased and in the presence of 20 μ M SQDG, binding of topo I to oligonucleotide was almost completely abolished. In the presence of 5, 10 and 20 μ M CPT binding of topo I to oligonucleotide was increased which may be due to the stabilization of covalent topo I-oligonucleotide complexes by CPT. Thus, EMSA results clearly suggest that SQDG inhibits DNA binding activity of topo I. In order to understand comparative affinities of CPT and SQDG for the enzyme, competition cleavage assays were performed ( Supplementary Fig. S2). Upon preincubation of the enzyme with 5, 10 and 20 μ M SQDG for 5 minutes, CPT mediated cleavage was decreased in 5 and 10 μ M SQDG preincubated conditions and the cleavage was completely abolished in 20 μ M SQDG preincubated condition ( Supplementary Fig. S2a). On the other hand upon preincubation of the enzyme with 5, 10 and 20 μ M CPT for 5 minutes, addition of 20 μ M SQDG did not affect the CPT mediated cleavage complex formation ( Supplementary Fig. S2b).  Besides, upon simultaneous addition of SQDG and CPT, CPT mediated cleavage was not increased in 10 and 20 μ M CPT conditions and was same as 5 μ M CPT condition ( Supplementary Fig. S2c). These results suggest that both SQDG and CPT have similar affinities for the enzyme. Together the results of cleavage assays and electrophoretic mobility shift assay ( Fig. 2 and Supplementary Fig. S2) suggest that SQDG inhibits topo I upstream in the enzymatic reaction at the DNA binding step whereas CPT inhibits topo I downstream in the enzymatic reaction at relegation step. Thus, upon preincubation of the enzyme is with SQDG, CPT mediated cleavage is not observed and opposite to this, upon preincubation of the enzyme with CPT, SQDG can not affect CPT mediated cleavage complex formation. However, upon simultaneous addition of SQDG and CPT, CPT mediated cleavage is observed but the cleavage does not increase with increasing concentrations of CPT, suggesting that in the presence of SQDG the enzyme becomes unable to bind to DNA and therefore CPT can not stabilize covalent enzyme-DNA complexes anymore. SQDG selectively kills acute lymphoblastic/lymphocytic leukemia cell lines and the cell killing is topo I dependent. Cell Table S1). Jurkat cell line was also found to be resistant for SQDG treatment with a very high IC 50 value of 75.67 ± 6.4 μ M. One difference among MOLT-4, MOLT-3, Reh and the other leukemic cells used in this study is p53 status. Studies have shown that MOLT-4, MOLT-3 and Reh cell lines express wild type p53 [24][25][26][27] ; Jurkat, RAJI, THP-1 and U937 cells express mutant p53 28-30 while K562 and HL-60 are devoid of p53 protein [31][32][33] . Therefore, the effects of SQDG treatment were also checked on three p53 wild type solid tumor cell lines and three p53 mutant solid tumor cell lines. Surprisingly, SQDG did not affect viability of the cell lines and only A549 cells were partially affected at 50 μ M concentration of SQDG. SQDG also did not affect viability of WI-38 cells (normal lung fibroblast cell line) and peripheral blood mononuclear cells (PBMC) (Fig. 3b,c and Supplementary Table S1).
To assess the role of topo I inhibition in SQDG mediated killing of MOLT-4 cells, siRNA silencing of TOP1 gene was performed by using a pool of three different siRNAs. Knockdown of TOP1 gene in MOLT-4 cells rendered the cells less sensitive for SQDG treatment ( Fig. 4a and Supplementary Fig. S3a). IC 50 value of SQDG for control siRNA transfected cells was found to be 14.04 ± 0.71 μ M while IC 50 value of SQDG for topo I siRNA transfected cells was found to be 29.09 ± 2.08 μ M (Supplementary Table S2). Fold resistance was calculated by the ratio of IC 50 value of siRNA transfected cells to IC 50 value of control siRNA transfected cells. Fold increase in resistance for topo I siRNA transfected cells was found to be 2.07 folds. Knockdown of TOP1 gene was corroborated by one additional siRNAs pool to mitigate off-target effects (Supplementary Fig S4a. and Supplementary Table S2). siRNA silencing of TOP1 gene was also performed in Reh cell line. Similar to MOLT-4 cells in Reh cells knockdown of TOP1 gene rendered the cells less sensitive for SQDG treatment ( Supplementary Fig S4b and S4c). IC 50 value of SQDG for control siRNA transfected cells was found to be 14.48 ± 0.28 μ M while IC 50 value of SQDG for topo I siRNA transfected cells was found to be 27.51 ± 0.41 μ M (Supplementary Table S2). Fold increase in resistance for topo I siRNA transfected cells was found to be 1.89 folds. These results indicate that SQDG mediated MOLT-4 and Reh cell killings are topo I dependent.  Table S3). Since SQDG treatment arrested cells in S-phase and studies have shown that inhibition of topo I impairs DNA replication as topo I has profound role during cellular DNA replication 34,35 , therefore we suspected that SQDG treatment may induce DNA replication stress as a result of topoisomerase I inhibition. In order to test the induction of replication stress we checked recruitment of ATM-and-Rad3-related kinase (ATR) at chromatin after SQDG treatment. ATR is known to be recruited at chromatin during the replication stress as stalled replication forks recruit ATR [36][37][38] . In response to SQDG treatment we observed ATR enrichment at chromatin, reflecting the generation of DNA replication stress (Fig. 4e,f and Supplementary Table S4). After 4 hours of the treatment, 23 folds increased in ATR recruitment at chromatin was observed which increased to 315 folds after 16 hours of SQDG treatment, indicating the generation of replication stress in SQDG treated  To study whether SQDG induces apoptosis in MOLT-4 cells, we performed annexin-V and propidium iodide staining of SQDG treated and untreated cells followed by flow cytometric analysis. CPT (2 μ M) was used as a positive control for apoptosis induction. Treatment of MOLT-4 cells with 15 μ M or 20 μ M SQDG for 48 hours markedly induced apoptosis in MOLT-4 cells (Fig. 5a). Since MOLT-4 cells express wild type p53, therefore we thought that p53 might play role in induction of apoptosis in these cells. In order to study p53 dependent apoptotic pathway, MOLT-4 cells were treated with 15 μ M SQDG for different time points and induction of the proteins involved in p53 dependent apoptotic pathway was checked (Fig. 5b,c). A 34 fold increase in p53 activation was observed after 8 hours of SQDG treatment, which reached to 60 folds at 24 hours post-treatment. Bcl-associated X protein (Bax) activation is associated with induction of p53 dependent apoptosis. Bax is a transcriptional target of p53 and it induces cytochrome c release from mitochondrial membrane 39,40 . Bax was activated by 2 folds after 8 hours of SQDG treatment and increased to 13 folds at 24 hours post-treatment. Cytochrome c together with apoptotic protease activating factor-1 and procaspase-9 forms a complex called apoptosome, which cleaves procaspase-9 and converts it in to caspase-9 41,42 . Cleavage of procaspase-9 was increased by 9 folds after 8 hours of SQDG treatment and reached to 50 folds at 24 hours post-treatment. Caspase-9 cleaves procaspase-3 into caspase-3, which initiates process of cell death 41 . Caspase-3 is an executive caspase which activates various endonucleases and proteases. During apoptosis induction PARP-1, a 113 kDa protein, is cleaved into 89 and 24 kDa fragments. In our study caspase-3 and PARP-1 cleavages started at 12 hours of SQDG treatment and continued up to 24 hours of the treatment. Together these findings suggest that SQDG treatment induces a p53 dependent apoptotic pathway in MOLT-4 cells. To further study p53 dependency of SQDG mediated cell killing, siRNA gene silencing was performed to knockdown TP53 gene expression in MOLT-4 cells ( Fig. 5d and Supplementary Fig. S3b). Cell viability experiments with TP53 silenced cells indicated that knockdown of p53 renders the cells less sensitive for SQDG treatment.  Table S5). Fold increase in resistance for p53 siRNA transfected cells was found to be 1.77 folds. Knockdown of TP53 gene was corroborated by one additional siRNAs pool to mitigate off-target effects ( Supplementary Fig S4d. and Supplementary Table S5). siRNA silencing of TP53 gene was also performed in Reh cell line. Similar to MOLT-4 cells in Reh cells knockdown of TP53 gene rendered the cells less sensitive for SQDG treatment ( Supplementary Fig S4e and S4f). IC 50 value of SQDG for control siRNA transfected cells was found to be 14.31 ± 1.04 μ M while IC 50 value of SQDG for p53 siRNA transfected cells was found to be 22.35 ± 2.42 μ M (Supplementary Table S5). Fold increase in resistance for p53 siRNA transfected cells was found to be 1.56 folds. Knockdown of TP53 gene was corroborated by one additional siRNAs pool to mitigate off-target effects (Supplementary Fig S4f). On the other hand ectopic expression of p53 protein in p53 deficient K562 myeloid leukemia cells amended these cells relatively more sensitive for SQDG treatment (Fig. 5e and Supplementary Fig. S3c). IC 50 value of SQDG for K562 cells transfected with empty vector was found to be 56.51 ± 1.14 μ M while IC 50 value of SQDG for K562 cells ectopically expressing wild type p53 was found to be 19.54 ± 2.51 μ M (Supplementary Table S6). Fold increase in sensitivity for K562 cells ectopically expressing p53 was found to be 2.89 folds. p53 protein was also ectopically expressed in an another p53 deficient cell line: HL-60. Similar to K562 cells ectopic expression of p53 protein in HL-60 cells amended these cells relatively more sensitive for SQDG treatment (Supplementary Fig S5). IC 50 value of SQDG for the cells transfected with empty vector was found to be 92.65 ± 6.35 μ M while IC 50 value of SQDG for the cells ectopically expressing wild type p53 was found to be 24.01 ± 2.0 μ M (Supplementary Table S6). Fold increase in sensitivity for HL-60 cells ectopically expressing p53 was found to be 3.85 folds. Together these observations suggest that SQDG mediated leukemic cell killing is p53 dependent.

SQDG synergizes with etoposide and doxorubicin. Dose-effect curves for etoposide-SQDG
and doxorubicin-SQDG drug combinations indicated that the combinations exert synergism as all the data points for the two combinations fell very near to Fa value of 1.0, (Supplementary Fig. S6a and S6b). As dose-effect curves do not provide any quantitative measurement of synergy therefore using CompuSyn software median-effect analysis was performed to quantitatively evaluate etoposide-SQDG and doxorubicin-SQDG drug combinations [43][44][45] . In median-effect analysis a combination index value (CI) is calculated for each drug combination. CI = 1.0, is additive effect; CI > 1.0, is antagonism and CI < 1.0, is synergism. CI values for etoposide-SQDG and doxorubicin-SQDG combinations ranged from 0.4 to 0.7 and from 0.1 to 0.6, respectively (Tables 1 and 2). These CI values are indicative of synergism and demonstrate that etoposide-SQDG and doxorubicin-SQDG combinations exert synergistic effects on MOLT-4 cell killing. In the Table 2, for combination of doxorubicin (1.4 μ M) and SQDG (35 μ M), a sudden drop in the CI value (CI value = 0.136 ± 0.036) was observed. It has been shown that topo IIα isoform shows sensitivity for both doxorubicin and etoposide whereas topo IIβ isoform is more sensitive for doxorubicin than etoposide 46 . Thus, sudden drop in the CI value for the combination of doxorubicin (1.4 μ M) and SQDG (35 μ M) is due to inhibition of both the isoforms of topo II enzyme by doxorubicin and inhibition of topo I by SQDG. In the isobolograms, generated using CompuSyn software Fa values

SQDG inhibits tumor growth in nude mice.
In vivo study in nude mice demonstrated that SQDG exerts potent antitumor activity and significantly inhibits growth of tumors. In control group, MOLT-4 xenografted mice (n = 4) were injected with PBS only. In SQDG treated group, MOLT-4 xenografted mice (n = 4) were injected with 2 mg/kg body weight SQDG once per week. In the SQDG treated mice, tumor volumes were significantly reduced compared to PBS treated mice (Fig. 6a-c). Mean tumor doubling time (TD) for the PBS treated mice was 4 ± 0.5 days, which upon SQDG treatment significantly increased to 11.5 ± 2 days (values are mean ± SEM; P < 0.05, SQDG vs. PBS treated). Specific tumor growth delay (SGD), which represents the number of tumor doubling times delayed by the treatment 47 , was found to be 1.87 in response to SQDG treatment, indicating that the treatment delayed TD by 1.87 times compared to PBS treatment.
In order to assess tumor regression pharmaco-dynamic (PD) markers in the tumor tissues from vehicle and SQDG treated mice were analyzed (Supplementary Table S9). For assessment of apoptosis in the tumor cells, cleavage of PARP-1 was analysed by immunohistochemical method using specific antibody to detect cleaved PARP-1. Cleavage of PARP-1 was found to be at very low levels in the tumor tissues from vehicle treated mice whereas in the tumor tissues from SQDG treated mice PARP-1 cleavage increased to the intermediate levels (Supplementary Table S9), suggesting the induction of apoptosis in the SQDG treated tumor cells. For assessment of S phase arrest in the tumor cells, cyclin A2 expression was analysed by immunohistochemical method. Cyclin A2 expression was found to be at intermediate level in the tumor tissues from vehicle treated mice whereas in the tumor tissues from SQDG treated mice cyclin A2 expression increased to the high levels (Supplementary Table S9), suggesting S phase arrest in the SQDG treated tumor cells. For assessment of DNA damage response in the tumor cells, γ -H2AX levels were analysed by immunohistochemical method. γ -H2AX levels were found to be low  Table 2. Doxorubicin-SQDG combination effect analysis. MOLT-4 cells were treated with either SQDG or doxorubicin or doxorubicin-SQDG constant ratio combinations and cell viability data were analyzed using CompuSyn Software. Values are presented as mean ± SD, for the three independent experiments. 'DOX' is abbreviation for doxorubicin.
in the tumor tissues from vehicle treated mice whereas in the tumor tissues from SQDG treated mice γ -H2AX levels were increased at intermediate levels (Supplementary Table S9), suggesting induction of apoptotic DNA damage in the SQDG treated tumor cells. In addition to PD markers, expression of Ki-67 (cell proliferation marker) was also analysed in the tumor cells. Ki-67 expression was found to be high in the tumor cells from vehicle treated mice whereas in the tumor cells from SQDG treated mice Ki-67 expression was found to be low (Supplementary Table S9), suggesting that tumor cell proliferation is decreased upon SQDG treatment. Together these results suggest that SQDG treatment induces S phase arrest, apoptosis and apoptotic DNA damage and thereby decreases cell proliferation in the tumor tissues. These results also substantiate the in vitro observations for S-phase arrest and apoptosis induction in MOLT-4 cells. Together these results provide substantial evidence that SQDG exerts potent anti-leukemic activity in vivo and thus SQDG is a promising anti-leukemic agent.

SQDG docks in the DNA binding region of topo I by forming non-covalent interactions.
Molecular docking was performed between human topoisomerase I (PDB ID: 1A36) 48 and SQDG (Pubchem, CID 50898453) 20 (Fig. 6d). The best fit docking solution predicted that the head group of SQDG forms three hydrogen bonds with the active site residues (ARG488, ARG590, LYS532) and the ring oxygen of SQDG forms an electrostatic interaction with the side chain of LYS532 (Fig. 6e). Hydrocarbon tails of SQDG span in the DNA binding groove of the enzyme, probably by forming hydrophobic interactions with various hydrophobic residues in this region (F353, I355 in vicinity with the C15 side chain of SQDG and V256, F259, G359, L360, L373 in vicinity with the C14 side chain of SQDG). Together these results predict that SQDG non-covalently interacts with the enzyme and blocks DNA binding of the enzyme. Docking results also corroborate EMSA results.

Discussion
This study demonstrates that SQDG isolated from the leaves of Azadirachta indica exerts selective and potent anti-ALL effects by inhibiting DNA relaxation activity of topo I. Silencing of TOP1 gene in MOLT-4 and Reh cell lines increased resistance of these cell lines for SQDG treatment, suggesting that SQDG mediated MOLT-4 and Reh cell killing is mediated by topo I inhibition. Studies have shown that topo I is over expressed in ALL and MOLT-4 cells 12,14 . In our study SQDG selectively killed ALL cell lines (MOLT-4, MOLT-3 and Reh) without affecting the growth of other leukemic cells and solid tumor cells. This selective cell killing may be reflected due the higher expression of topo I in these cells. One difference among MOLT-4, MOLT-3, Reh and the other leukemic cells used in this study is p53 status. p53 is well known for inducing cell cycle arrest and/or apoptosis in response to various cellular stresses 49 and p53 is activated upon genotoxic stresses like DNA damage and DNA replication stress 50,51 . During DNA replication stress enrichment of ATR at chromatin activates ATR and activated ATR phosphorylates checkpoint kinase-1 which subsequently activates and stabilizes p53 protein 50 . In our study we also found the enrichment of ATR at chromatin and thus we hypothesized that SQDG treatment may stabilize p53 protein and induce p53 dependent apoptotic pathway following topo I inhibition, DNA replication stress and S-phase arrest in MOLT-4 cells. To test the hypothesis two different experiments were performed. First, TP53 gene of MOLT-4 cells was knocked down using siRNA. Knocking down p53 increased the resistance of MOLT-4 cells for SQDG. In the second experiment MOLT-4 cells were treated with SQDG for different time points and induction of p53 dependent apoptotic pathway was checked. SQDG treatment stabilized p53 and activated the proteins involved in the p53 dependent apoptotic pathway. These experiments confirmed the involvement of p53 dependent pathway in SQDG mediated cell killing. Involvement of p53 in leukemia cell chemosensitivity has been previously described by Trepel et al. 52 . K562 myeloid leukemia cells do not express p53 protein 31,33 . In our study K562 cells were merely partially affected by SQDG treatment at more than 50 μ M concentration but ectopic expression of wild type p53 protein rendered these cells relatively more sensitive to SQDG treatment. This observation further supported the p53 dependency for SQDG mediated leukemic cell killing. Altogether these observations suggest that overexpression of topo I and expression of wild type p53 in MOLT-4 cells participate to the higher sensitivity of these cells for SQDG. One question which still remains is: why and how SQDG did not affect the growth of solid tumor cell lines which express both topo I and wild type p53 e.g. HCT116, HepG2? One possible reason could be involvement of one or more factors associated with leukemic cells which facilitate(s) the SQDG mediated cell killing. Although, A549 cells, with wild type p53, were partially (~25%) killed at 50 μ M SQDG.
In vivo study in nude mice demonstrated that SQDG treatment delays tumor doubling time and reduces the expression of cell proliferation marker: Ki67, indicating in vivo anti-tumorigenic activity of SQDG. Treatment of mice with SQDG also induced S phase arrest, apoptosis and apoptotic DNA damage in the tumor tissues. Combinations of SQDG with etoposide and doxorubicin exerted synergistic effects on killing of MOLT-4 cells. Thus we suspect that using these combinations, doses of etoposide and doxorubicin can be reduced during ALL chemotherapy and side effects caused by the chemotherapeutic agents can be minimized. Altogether, our study suggests that SQDG is a novel drug candidate for use in ALL chemotherapy and combining the anti-ALL drugs with SQDG might improve the chemotherapeutic regimen for ALL. Selective killing of MOLT-4 cells by SQDG and its in vivo anti-tumor activity in nude mice makes it attractive and promising anti-ALL agent. This study indicates the prospect of SQDG as a lead molecule for future ALL chemotherapy.

Methods
Reagents, cell culture and cell viability assays. SQDG was isolated as described previously 20 .
Etoposide, doxorubicin, betulinic acid and camptothecin were purchased from Sigma (St. Louis, MO, US). Cell lines were cultured in the ATCC recommended culture media supplemented with 10% fetal bovine serum and antibiotics (Invitrogen, Carlsbad, CA, US). Cells were seeded in 96 well plates and treated with respective compounds. After treatment cells were incubated with 3-(4,5-dimethylthiazol-2-yl)-2,5diphenyltetrazolium bromide (MTT) and colorimetry was performed on Thermo MultiskanEX plate reader at 595 nm.
Ethics statement. Experiments with nude mice were conducted at National Institute of Immunology (NII). Mice were obtained from the Small Animal Facility of the NII. All animals were housed and used as per the national guidelines provided by the Committee for the Purpose of Control and Supervision of the Experiments on Animals. Protocols for the experiments with nude mice were approved by the Scientific RepoRts | 5:12082 | DOi: 10.1038/srep12082 Institutional Animal Ethics Committee of NII. All the methods were carried out in accordance with the approved guidelines. DNA relaxation assay. Human DNA topoisomerase I and topoisomerase II enzymes were purchased from TopoGEN Inc (Port Orange, Florida, USA). For DNA relaxation assay, 100 fmol supercoiled pBluescript SK(+ ) (pBS) DNA per reaction was used and the assay was carried out as described previously 22 . Oligonucleotide cleavage assay. Synthetic oligonucleotides containing topo I binding site 53 were purchased from IDT Inc. (San Jose, CA, USA). CL25 oligonucleotide single strand was 5′ -end labeled using γ -32 Phosphate ATP by T4-polynucleotide kinase reaction and CP25 oligonucleotide single strand was 5′ -end labeled using ATP. The strands were annealed in annealing buffer [10 mM Tris-HCl (pH 7.5), 1 mM EDTA, and 100 mM NaCl] and oligonucleotide cleavage assay was performed as described previously 23 .
DNA cleavage assay. DNA cleavage assay was performed as described previously 22 .
Electrophoretic mobility shift assay. Electrophoretic mobility shift assay for analysis of topo I-oligonucleotide binding was performed as described previously 54 . The same oligonucleotide used for oligonucleotide cleavage assay was used for electrophoretic mobility shift assay.
Immunoband depletion assay. MOLT Nuclear and chromatin fractionations. Nuclear and chromatin fractionations were performed as described previously 56 . For ATR detection, equal amounts of nuclear and chromatin proteins were electrophoresed and immunoblotting was performed for ATR and histone H3 proteins. For quantitative measurement of ATR enrichment at chromatin, densitometry was performed using ImageJ software. ATR levels were normalized with histone H3 levels and fold change was calculated. Drug combination effect analysis. MOLT-4 cells were treated with individual agents or constant ratio combinations of etoposide-SQDG and doxorubicin-SQDG for 72 hours and cell viability assays were performed. Fraction affected (Fa) values were calculated as described previously 45 . Fa values were used to evaluate drug combinations by CompuSyn software from ComboSyn, Inc. (Paramus, NJ, USA) 43-45 . Mice and tumor xenografts. For tumor xenografts, approximately, 1 × 10 6 MOLT-4 cells in 100 μ l RPMI-1640 were mixed with equal volume of Matrigel (BD Biosciences) and implanted subcutaneously into 6-week-old athymic nude mice bearing the nu/nu gene [NIH(s) (nu/nu)]. Mice were maintained in pathogen-free conditions. When the tumor sizes reached 150-200 mm 3 , mice were injected intra peritoneally with SQDG dissolved in PBS (2 mg/kg body weight/week), for two weeks. Tumor length (L) and width (W) were measured using Vernier calipers and tumor volume was calculated by the equation: Volume = L × W 2 /2 (mm 3 ). Tumor doubling time (TD) was calculated and the overall response rate of xenografts to treatment was assessed by measuring specific tumor growth delay (SGD) as described previously 47 . Data were analyzed by student's unpaired t-test and values were considered significantly different at P < 0.05 and expressed as the mean ± SEM.
Immunohistochemistry. Immunohistochemistry of tumor tissues from SQDG and vehicle treated mice was performed as described by Jalava et al. 57 . Diaminobenzidine tetrahydrochloride (DAB) staining was done using DAB substrate kit (BD Pharmingen) and nuclei were stained with Harris' hematoxyline solution for 30 seconds. Scoring of immunostaining was done as described by Perrone et al. 58 .

Molecular docking.
Crystal structure of the receptor molecule i.e. human topoisomerase I complexed with its substrate DNA and inhibitor camptothecin was collected from Protein Data Bank (PDB) (PDB ID: 1A36) 48 . Three-dimensional (3D) coordinates of DNA and camptothecin were removed from the 1A36 structure for docking of SQDG. Structure of the ligand sulfonoquinovosyl diacylglyceride (SQDG) was derived from Pubchem (CID 50898453) 20 and 3D coordinates of the molecule were generated using Openbabel 2.3.1 59 . 3D structure of SQDG was optimized through energy minimization using AMBER force field of VEGAZZ 3.0.0.52 package 60,61 . Probable binding cavities of 1A36 were identified using the CASTp web server 62 and molecular docking of SQDG with topo I was performed using the GOLD v5.0.1 package from Cambridge Crystallographic Data Centre 63 . Following parameters were used in the docking cycles: population size (100), selection pressure (1.100000), number of operations (100,000), number of islands (5), niche size (2), crossover weight (95), mutate weight (95), and migrate weight (10). 100 docking calculations were run for the ligand and the best docking solutions were identified based on lowest energy-docking mode and critical manual inspection satisfying favorable interactions between the ligand and the protein molecule.