SB225002 Induces Cell Death and Cell Cycle Arrest in Acute Lymphoblastic Leukemia Cells through the Activation of GLIPR1

Acute Lymphoblastic Leukemia (ALL) is the most frequent childhood malignancy. In the effort to find new anti-leukemic agents, we evaluated the small drug SB225002 (N-(2-hydroxy-4-nitrophenyl)-N’-(2-bromophenyl)urea). Although initially described as a selective antagonist of CXCR2, later studies have identified other cellular targets for SB225002, with potential medicinal use in cancer. We found that SB225002 has a significant pro-apoptotic effect against both B- and T-ALL cell lines. Cell cycle analysis demonstrated that treatment with SB225002 induces G2-M cell cycle arrest. Transcriptional profiling revealed that SB225002-mediated apoptosis triggered a transcriptional program typical of tubulin binding agents. Network analysis revealed the activation of genes linked to the JUN and p53 pathways and inhibition of genes linked to the TNF pathway. Early cellular effects activated by SB225002 included the up-regulation of GLIPR1, a p53-target gene shown to have pro-apoptotic activities in prostate and bladder cancer. Silencing of GLIPR1 in B- and T-ALL cell lines resulted in increased resistance to SB225002. Although SB225002 promoted ROS increase in ALL cells, antioxidant N-Acetyl Cysteine pre-treatment only modestly attenuated cell death, implying that the pro-apoptotic effects of SB225002 are not exclusively mediated by ROS. Moreover, GLIPR1 silencing resulted in increased ROS levels both in untreated and SB225002-treated cells. In conclusion, SB225002 induces cell cycle arrest and apoptosis in different B- and T-ALL cell lines. Inhibition of tubulin function with concurrent activation of the p53 pathway, in particular, its downstream target GLIPR1, seems to underlie the anti-leukemic effect of SB225002.


Introduction
Acute lymphoblastic leukemia (ALL) is the most common cancer in childhood. We previously reported that the leukemic bone marrow (BM) microenvironment has increased levels of chemokine (C-C motif) ligand 2 (CCL2) and interleukin-8 (IL- 8), and that these chemokines have a positive impact on BM mesenchymal stem cells, but no direct effect on ALL cells [1]. In our efforts to evaluate the function of the IL-8 receptor CXCR2 in ALL cells, we made use of SB225002 (N-(2-hydroxy-4-nitrophenyl)-N'-(2-bromophenyl)urea), a drug initially described as a CXCR2 antagonist [2]. Although the CXCR2 receptor was found to be non-functional in ALL [1], B-and T-ALL cell lines were sensitive to SB225002.
We learned with interest throughout the course of this project that SB225002 is not a specific inhibitor of CXCR2 as it was stated in its first description [2]. Now, SB225002 is known to have many cellular targets and effects, including the inhibition of microtubule polymerization, hyperphosphorylation of Bcl2 and BclxL, prometaphase cell cycle arrest, promotion of mitotic catastrophe and inhibition of gamma-secretase [3][4][5]; amelioration of acute experimental colitis in vivo [6]; management of both acute and chronic pain in vivo [7]; angiogenesis inhibition in vivo [8]; among others.

Ethics Statement
Institutional Review Board approval for the animal study was obtained from the Ethics Commission for Animal Use from the Institute of Biology at the University of Campinas (CEUA/ UNICAMP, protocol 3624-1). The use of a patient ALL sample in this study was approved by the Centro Infantil Boldrini Ethics Committee (CAAE 0004.0.144.000-05). The patientderived sample corresponded to frozen patient-derived xenograft cells, whose primary tumors were obtained in the early 1990s. The ethics committee has exceptionally waived the informed consent for those leukemia samples collected prior to the start of the study because it could not be obtained due to death or loss to follow-up.

Western Blot analysis
Treated and control whole cell lysates were prepared in lysis buffer (Cell Signaling, Danvers, MA, USA). Eighty to one-hundred μg of protein were electrophoresed in a 10% SDS-polyacrylamide gel (Bio-Rad, Hercules, CA, USA). Proteins were electro-blotted onto PVDF membrane in a 50 mM Tris-base, 20% methanol, and 40 mM glycine electrophoresis buffer. Membranes were incubated in 5% non-fat dry milk in PBST (Phosphate 100 mM, KCl 27 mM, NaCl 1.37 M pH 7.4 after 1X dilution; 0.2% Tween-20) for 1 h. Blots were probed with primary antibody overnight at 4°C in 2% BSA in PBST, and then incubated with a horseradish peroxidase-conjugated secondary antibody (Cell Signaling) in 5% dry milk in PBST for 1 h at room temperature. Bound antibodies were detected by Super Signal West Pico Chemiluminescent Substrate detection reagent (Pierce/Thermo Scientific, Rockford, IL, USA) and visualized by autoradiography. The primary antibodies used for Western blot analysis were: anti-GLIPR1 (Novus Biologicals, Littleton, CO, USA), p-AKT Ser473 (Cell Signaling), p-GSK3beta Ser9 (Cell Signaling), p-PDK1 Ser241 (Cell Signaling), anti-CX3R1, GAPDH and β-actin (all from Santa Cruz Biotechnology, Santa Cruz, CA, USA).

Microarray analysis
Total RNA was obtained from Jurkat cells treated with 12.5 μM SB225002 or 0.1% DMSO for 6 h and 9 h using QIAshredder (Qiagen) and RNeasy Mini Kit (Qiagen). Treatments were performed in duplicates. Antisense biotinylated cRNA was prepared on the Affymetrix GeneChip Array Station using the GeneChip HT One-Cycle cDNA Synthesis and Gene Chip HT IVT Labeling kits (Affymetrix, Santa Clara, CA, USA). Biotinylated cRNAs were hybridized to the Affymetrix HAT HG-U133A and HAT HG-U133B Array Plate. Array washing and staining were performed on the GeneChip GCAS Array Station following a robotic protocol according to the manufacturer's instructions (Affymetrix). Arrays were scanned on the GeneChip HT scanner (Affymetrix). Scanned image output files were visually examined for major chip defects and hybridization artifacts and then analyzed with Affymetrix GeneChip Microarray Analysis Suite 5.0 (MAS5) software (Affymetrix). All high-quality arrays were analyzed using the Probe Logarithmic Intensity Error (PLIER) algorithm. Genes were considered to be differentially expressed in a given group, if the 90% lower confidence bound (LCB) of the fold change (FC) between the two groups was above 1.2 [24]. Microarray data along with information from this study have been deposited in the NIH Gene Expression Omnibus database at www.ncbi.nlm.nih.gov/geo under the accession number GSE71212. To understand the biological mechanisms affected by the transcripts that were counter-regulated by the treatment, interactive networks, pathways, and functions analysis was performed using the commercial Systems Biology oriented package Ingenuity Pathways Analysis (IPA 4.0, www.ingenuity.com, QIAGEN, Redwood City, CA). Furthermore, to identify compounds that may have similar or opposite effects compared to SB225002, the differentially expressed genes, after removal of non-HG-U133A probe sets, were used to query the Connectivity Map database build 02 (www. broadinstitute.org/cmap). Details of the Connectivity Map dataset and analytics have been previously described [25].

Proliferation assays and cell viability assays
Proliferation assay was performed in 96-well micro-titer plates containing 10,000 ALL cells per well. Viable cells were counted on BD FACSCanto flow cytometer at different time points. Cell viability assays were performed in 96-well micro-titer plates containing 30,000 cells per well for the ALL cells lines and 200,000 cells per well for the PHA-stimulated normal lymphocytes using the MTT reagent (Sigma-Aldrich). The formazan dye formed by the viable cells was quantified by measuring the absorbance of the dye solution at 590 nM.

Xenograft model
Patient-derived xenograft ALL cells were thawed, washed with PBS and 1x10 7 cells were injected in NOD/SCID (NOD.CB17-Prkdcscid/J) mice (The Jackson Laboratory, Bar Harbor, ME, USA) by the tail vein. ALL engraftment was monitored as previously described [26] and outlined below. Successfully engrafted mice were sacrificed, ALL cells were collected from spleen and liver and 1x10 7 cells were immediately injected by the tail vein in eight secondary non-irradiated recipient mice for the subsequent experiments. Animals were monitored every 7 days for ALL engraftment as follows: blood was collected by retro-orbital bleeding into EDTA containing tube, mononuclear cells were isolated by ficoll centrifugation and the presence and quantity of ALL cells was analyzed by flow cytometry in a FACSCanto II equipment (Becton Dickinson, Franklin Lakes, NJ), using anti-hCD45-PE (clone HI30, BD Pharmingen, San Diego, CA or EXBIO, Prague, Czech Republic) and anti-mCD45-FITC (clone 30F-11, BD Pharmingen). When human CD45(+) cells reached 0.5% of peripheral blood cells in half of the animals, mice were randomly distributed into the different treatment groups (n = 4/each group). Mice were treated intraperitoneally with 10 mg/Kg of SB225002 or vehicle once a day, 5 days a week, for 4 weeks. Kaplan-Meier survival curves were compared using the Log-rank test.

SB225002 inhibited the proliferation of ALL cell lines at micromolar concentrations
As shown in Fig 1A, micromolar concentrations of SB225002 consistently inhibited the proliferation of the different ALL cell lines tested. In general, T-ALL cell lines (Jurkat, TALL-1, CEM and Molt-4) appeared to be more sensitive to SB225002 than precursor-B ALL cell lines (Nalm-6, REH, RS4;11 and 697). To investigate the potential cytotoxic effects of SB225002 on normal non-cancerous cells, we treated PHA-stimulated normal lymphocytes with SB225002 [5 μM and 10 μM]. As shown in Fig 1B, no cytotoxic effects were observed in the normal lymphocytes at these doses.
Primary ALL cells die rapidly when cultured in vitro [27]. Therefore, the effect of SB225002 on primary ALL was evaluated in a xenograft model of B-ALL. Balanced cohorts (n = 4 mice per treatment group) with established disease (% ALL cells 0.5% in peripheral blood) were treated with vehicle (control group) or SB225002 (10 mg/Kg intraperitoneally once a day, 5 days a week, during 4 weeks). Mice treated with SB225002 as single agent demonstrated a very modest trend towards prolonged overall survival compared to vehicle-treated controls (S1 Fig).

SB225002 treatment resulted in cell cycle arrest at G2/M and apoptosis of ALL cell lines
Recently, SB225002 was shown to possess a microtubule destabilizing activity, accompained by suppression of microtubule polymerization and induction of a prometaphase arrest [4]. It was also shown to promote mitotic catastrophe in ovarian cancer cells [5]. Cell cycle analysis in ALL cells treated with SB225002 (REH and RS4;11 [10 μM], Jurkat and TALL-1 [3.125 μM]) resulted in G2/M arrest (Fig 1C). In contrast, no cell cycle effects were observed in the PHAstimulated lymphocytes upon treatment with 10 μM SB225002 (Fig 1C). Analysis of mitotic ALL cells under SB225002 treatment (IC 50 dose) showed none of the abnormalities suggestive of mitotic catastrophe (spindle abnormalities, chromosome mis-segregation, multi-polar cell division, multiple nuclei, aneuploidy/polyploidy; data not shown).
To determine if the deleterious effects of SB225002 in ALL were caused by the induction of apoptosis, we performed Annexin-V/Propidium Iodide (PI) analysis in B-ALL (REH and RS4;11) and T-ALL (Jurkat and TALL-1) cells treated with SB225002 [10 μM] for 24 h. Most of the untreated ALL cells were non-apoptotic (Annexin-V negative/PI negative), while upon treatment with SB225002 a marked increase in the number of early (Annexin-V positive/PI negative) and late apoptotic (Annexin-V positive/PI positive) cells was observed in both B-and T-ALL cells (Fig 1D).
The transcriptional profile elicited by SB225002 is similar to that of tubulin inhibitors and involves the JUN, p53 and TNF pathways To get more insights into the mode of action of SB225002, gene expression profiling analysis was performed in Jurkat cells treated with SB225002 [IC 50 dose] for 6 h and 9 h. Transcriptional profiling analysis revealed 174 induced and 41 repressed genes that were commonly modulated after both 6 h and 9 h of treatment (S1 Table). The gene expression signature of Jurkat cells treated during 9 h with SB225002 was compared to the Connectivity Map (C-Map) database (build 02), which included 6,100 genomewide expression profiles representing 1,309 compounds. As shown in Fig 2A, the transcriptome effect of SB225002 showed high similarity to inhibitors of the PI3K/mTOR pathway (LY-294002, sirolimus, and wortmannin), inhibitors of the HSP90 chaperone (tanespimycin, 5255229, and monorden) and tubulin binding agents (5252917, rotenone, colchicine, podophyllotoxin, fenbendazole, and vinburnine). Importantly, 6 out of the top 17 compounds positively associated with the SB225002 signature were tubulin inhibitors. Jurkat and REH cells treated with SB225002 showed no alterations in p-PDK1 Ser241, p-AKT Ser473, and p-GSK3beta Ser9 levels ( S2 Fig). On the other hand, cell cycle arrest at G2/M (Fig 1C), and suppression of microtubule polymerization [4] are in agreement with SB225002 targeting of tubulin.
The differentially expressed genes after 6 h of treatment (before cell death was observed) were also analyzed using Ingenuity Pathway Analysis. As shown in Fig 2B, SB225002 treatment interfered with the cellular metabolism of cholesterol, p53 signaling, cell cycle progression, nuclear receptor signaling, TGF-β signaling and oxidative stress response. Interactive network analysis of the differentially expressed genes suggested that SB225002 triggered a transcriptional program of genes related to the activation of JUN and p53 pathways and inhibition of TNF pathway (Fig 2C-2E).
To validate these findings, we performed quantitative PCR (Q-PCR) analysis on selected candidate genes that appeared to be modulated by SB225002 treatment. The JUN pathway is directly and indirectly involved in apoptosis induction [28], and the transcription factor BACH2 is one of the most important effectors of JUN pathway apoptosis response [29]. Q-PCR assays confirmed that both c-JUN and BACH2 transcripts are up-regulated in Jurkat cells after 6 h and 9 h of treatment with SB225002 ( S3 Fig). Since Western blot analyses of c-JUN and BACH2 in REH and Jurkat cells rendered inconclusive results (data not shown), this pathway was not further explored in this work.
As shown in Fig 2E, SB225002 inhibited several targets of the TNF pathway. CX3CR1, a TNF target found to be down-regulated by SB225002 treatment, has been implicated in chronic lymphocytic leukemia attraction and adhesion to bone marrow stromal cells [30]. Q-PCR and Western blot analyses demonstrated a consistent inhibition of CX3CR1 transcripts after SB225002 treatment (Fig 3A). Interestingly, while Q-PCR demonstrated CX3CR1 transcripts more strongly down-regulated at the 6 h timepoint, Western blot analysis showed that protein levels were mostly down-regulated at the 9 h timepoint, but in both cases demonstrating that CX3CR1 was modulated by the SB225002-treatment.
As shown in Fig 2D, GLIPR1 was among the highest up-regulated genes in the p53 network modulated by SB225002 treatment. Q-PCR and Western blot analyses confirmed that treatment of Jurkat cells with SB225002 induced activation of GLIPR1 both at the mRNA and protein level (Fig 3B).

SB225002-mediated cell death is at least in part dependent on the activation of GLIPR1
GLIPR1 plays a pro-apoptotic role in prostate and bladder cancer cells [31]. In contrast, in other tumor types such as glioblastoma, GLIPR1 over-expression is associated with an increase in cellular proliferation and tumor invasion [32]. These data suggest that GLIPR1 effects in cancer cells are dependent on the tumor type. To investigate the relevance of GLIPR1 up-regulation on SB225002-mediated cell death in ALL, B-ALL (REH and RS4;11) and T-ALL (Jurkat and TALL-1) cells were transduced with GLIPR1-shRNA or control-shRNA (scramble)  (Fig 4B and S6 Fig). In addition, GLIPR1-KD resulted in a significant attenuation of the inhibitory effect of SB225002 on cell proliferation ( Fig 4C).
Apoptosis induction mediated by GLIPR1 is reported to be at least in part dependent on the production of reactive oxygen species (ROS) [33]. To investigate the potential increase on ROS generation upon SB225002 treatment, B-ALL (REH and RS4;11) and T-ALL (Jurkat and  To determine whether ROS production was promoted, at least in part, by GLIPR1 expression, GLIPR1-KD cells were analyzed concerned the production of ROS. Surprisingly, ROS production was higher in the GLIPR1-KD than in the scramble control cell lines, and SB225002 treatment resulted in increased ROS generation, even in GLIPR1-KD cells (Fig 5B). Incubation of ALL cells in presence of a ROS scavenger, N-Acetyl Cysteine (NAC), resulted in short term (6 h) decrease in ROS generation (S8 Fig) but no attenuation of SB225002 effects in terms of cell viability (Fig 5C). All together, these data suggest that the pro-apoptotic function of GLIPR1 in ALL cells, upon SB225002 treatment, is not linked to ROS generation.

Discussion
This study shows for the first time an anti-proliferative and pro-apoptotic activity of SB225002 against acute lymphoblastic leukemia. SB225002 treatment of leukemia cells induced ROS generation and cell cycle arrest at G2/M. Microarray expression analysis of SB225002-treated Jurkat cells revealed a transcriptional program typically triggered by tubulin binding agents, with some degree of overlap to the gene expression signature derived from PI3K/mTOR and HSP90 inhibitors. Suppression of microtubule polymerization by SB225002 has been experimentally demonstrated [4] and was therefore favored in our interpretation of possible molecular mechanisms for the connection between SB225002 and PI3K/mTOR or HSP90 inhibitors. For instance, HSP90 protects tubulin, keeping it in a state compatible with microtubule polymerization [34], thus the transcriptional program elicited by HSP90 inhibitors would be expected to overlap, at least in part, the one elicited by tubulin inhibitors.
Two drugs may elicit similar transcriptional profiles acting on different cellular targets. We speculate that the connection between SB225002 and PI3K/mTOR or HSP90 inhibitors may be related to the effects of anti-microtubule agents on the endoplasmic reticulum (ER) and/or translation machinery. Consistent with this hypothesis, Puromycin, an inhibitor of protein translation was also strongly linked to SB225002 in the C-Map analysis (Fig 2A).
Since ribosomal proteins, translation initiation factors, and other components of the translation machinery are associated with the cytoskeleton [35,36], destabilization of microtubules have a drastic effect on protein synthesis. Both inhibition of the PI3K/mTOR pathway [37] and inhibition of microtubule polymerization [36] result in the phosphorylation and inactivation of the translation initiation factor eIF2α. Inactivation of eIF2α causes a generalized translation repression, except for some few transcripts, including the ATF4 transcription factor, that are able to recruit ribosome binding to internal ribosome entry sites (IRES). Likewise, tubulin inhibition and disruption of the actin cytoskeleton by Cucurbitacin E were shown to increase eIF2α phosphorylation, inhibiting protein synthesis [38].
Microtubule dynamics is important to ER homeostasis. Diverse stressful conditions, including ER stress, converge to translation attenuation via eIF2α phosphorylation [39]. Microtubule disruption by colchicine leads to ER collapse and deposition of large perinuclear protein TALL-1) cell lines. B-ALL cells were treated with SB225002 [10 μM] and T-ALL with SB225002 [5 μM]. Treatment control was DMSO 0.1%. Cells were incubated for 48 h. S = scramble transfection control; G-KD = cells infected with GLIPR1-shRNA lentiviral particles (Sigma-Aldrich). P values were calculated using two-tailed Student's t-test.
SB225002 treatment of ALL cells resulted in remarkable downregulation of cholesterol biosynthesis genes (Fig 2B and S1 Table). This transcriptional output of SB225002 could be also attributed to ER stress. In fact, ER stress has been reported to cause significant transcriptional down-regulation of genes encoding key enzymes in cholesterol biosynthesis [45]. Modulation of cholesterol biosynthesis genes may also underlie the connection between SB225002 and PI3K/ mTOR inhibitors. We have recently found that the top biological functions downregulated by the PI3K inhibitor AS605240 in T-ALL (cell lines and primary cells) are related to cholesterol biosynthesis [46]. Likewise, the mTOR inhibitor Everolimus downregulated several lipid and fatty acid biosynthesis genes and induced ER stress genes in different ALL cell lines [47].
Transcriptome network analysis revealed that SB225002-induced cell death was associated to the activation of JUN and p53 pathways, and inhibition of the TNF pathway. Although the human T-ALL Jurkat cell line contains a non-sense mutation in its p53 gene, it still preserves a functional p53 protein [48,49]. GLIPR1, a transcriptional target of p53 [31,50] that was significantly upregulated upon SB225002 treatment, was chosen for further investigation.
GLIPR1 has been associated with variable functions in humans, such as cell proliferation, apoptosis and tumor growth [51]. In prostate cancer, increased expression of GLIPR1 is associated with apoptosis induction [31]. In contrast, in astrocytes-derived tumors over-expression of GLIPR1 is associated with an enhancement in cellular proliferation and tumor invasion, while GLIPR1 silencing is associated with an elevated level of apoptosis [32]. Recently, activation of GLIPR1 with an adenoviral clinical vector, in association with radiotherapy, significantly suppressed tumor growth and extended survival in prostate and bladder cancer in vivo models, suggesting that GLIPR1 activation should be explored as a potential therapeutic strategy at least in prostate and bladder tumors [52]. GLIPR1 is significantly underexpressed in ALL when compared to normal controls [53], which is compatible with a pro-apoptotic role of GLIPR1 activation upon SB225002 treatment. This is the first time functional assays were performed to evaluate the role of GLIPR1 in ALL. Importantly, down-regulation of GLIPR1 in Band T-ALL cells resulted in decreased proliferation of ALL cells and in a significant increase in cellular resistance to SB225002 treatment. Since SB225002 has an anti-mitotic effect, increased resistance to SB225002 upon GLIPR1 silencing could simply reflect the lower rate of cell proliferation. However, cytoskeleton disruption by tubulin binding agents do not spare interphase cells [54]. In addition, it has been shown that in prostate cancer cells there is an inverse correlation between the expression of GLIPR1 and c-Myc, where restoration of GLIPR1 expression downregulates c-Myc and induces cell-cycle arrest [55]. It is possible that GLIPR1 knockdown in ALL cells elevates c-Myc levels, leading to an increased resistance to SB225002's treatment. As discussed above, SB225002-treated cells revealed a gene expression profile suggestive of significant ongoing stress. Therefore, it is not possible to exclude that the cellular effects of SB225002 are at least partially dependent on GLIPR1 activation. For instance, GLIPR1 overexpression has been shown to cause inactivation of Bcl-2, dysfunction of mitochondria, and activation of a wide spectrum of caspases [33].
Apoptosis induction upon GLIPR1 overexpression in a tet-on stable clone of bladder carcinoma cell line was reported to be dependent on the production of ROS [33]. Here, we detected [5 μM] (Jurkat and TALL-1). S = scramble transfection control; G-KD = cells infected with GLIPR1-shRNA lentiviral particles (Sigma-Aldrich). P values were calculated using two-tailed Student's t-test.
doi:10.1371/journal.pone.0134783.g005 a significant increase in ROS levels in ALL cells after treatment with SB225002. However, upon GLIPR1 knockdown and treatment with SB225002, none of the cell lines showed a reduction in the production of ROS. In addition, pre-incubation of the cells with N-Acetyl Cysteine, a known ROS scavenger, prior to the treatment with SB225002, did not attenuate cell death. These data suggest that GLIPR1 gene plays a role in the apoptosis induction mediated by SB225002, but likely not through the modulation of ROS levels in ALL cells. Overall, we believe that SB225002 might have the potential to exert its activity through distinct mechanisms depending on the cellular type. Further studies including systems biology approaches should be considered and will be useful for better understanding of SB225002's effects and downstream molecular events in ALL.
Finally, the treatment of a xenograft model of ALL with SB225002 demonstrated a trend towards prolonged overall survival compared to vehicle-treated controls. These are preliminary, but encouraging results, which suggest further studies with SB225002 in vivo, with larger animal cohorts, increased treatment dose(s) and leukemia cells from different patients to better represent the heterogeneity of the disease, should be considered.
In conclusion, our results demonstrate that SB225002 has anti-proliferative and pro-apoptotic effects against precursor B-and T-ALL cell lines, at the micromolar concentration range. Cells treated with SB225002 undergo cell cycle arrest at G2/M and exhibited a transcriptional gene expression profile typically elicited by tubulin binding agents. SB225002-mediated cell death is at least partially dependent upon GLIPR1 up-regulation, irrespective of ROS generation.