Ran GTPase-Activating Protein 1 Is a Therapeutic Target in Diffuse Large B-Cell Lymphoma

Lymphoma-specific biomarkers contribute to therapeutic strategies and the study of tumorigenesis. Diffuse large B-cell lymphoma (DLBCL) is the most common type of malignant lymphoma. However, only 50% of patients experience long-term survival after current treatment; therefore, developing novel therapeutic strategies is warranted. Comparative proteomic analysis of two DLBCL lines with a B-lymphoblastoid cell line (LCL) showed differential expression of Ran GTPase-activating protein 1 (RanGAP1) between them, which was confirmed using immunoblotting. Immunostaining showed that the majority of DLBCLs (92%, 46/50) were RanGAP1+, while reactive lymphoid hyperplasia (n = 12) was RanGAP1+ predominantly in germinal centers. RanGAP1 was also highly expressed in other B-cell lymphomas (BCL, n = 180) with brisk mitotic activity (B-lymphoblastic lymphoma/leukemia: 93%, and Burkitt lymphoma: 95%) or cell-cycle dysregulation (mantle cell lymphoma: 83%, and Hodgkin’s lymphoma 91%). Interestingly, serum RanGAP1 level was higher in patients with high-grade BCL (1.71 ± 2.28 ng/mL, n = 62) than in low-grade BCL (0.75 ± 2.12 ng/mL, n = 52) and healthy controls (0.55 ± 1.58 ng/mL, n = 75) (high-grade BCL vs. low-grade BCL, p = 0.002; high-grade BCL vs. control, p < 0.001, Mann-Whitney U test). In vitro, RNA interference of RanGAP1 showed no effect on LCL but enhanced DLBCL cell death (41% vs. 60%; p = 0.035) and cell-cycle arrest (G0/G1: 39% vs. 49%, G2/M: 19.0% vs. 7.5%; p = 0.030) along with decreased expression of TPX2 and Aurora kinases, the central regulators of mitotic cell division. Furthermore, ON 01910.Na (Estybon), a multikinase inhibitor induced cell death, mitotic cell arrest, and hyperphosphorylation of RanGAP1 in DLBCL cell lines but no effects in normal B and T cells. Therefore, RanGAP1 is a promising marker and therapeutic target for aggressive B-cell lymphoma, especially DLBCL.


Introduction
Tumor biomarkers are pivotal for screening, diagnosing, and following-up cancers. Lymphoma-specific markers also contribute to treatment strategy, prognostic stratification, and the study of tumorigenesis. Current clinically useful biomarkers for lymphoma management are both scarce and non-specific. For example, serum lactate dehydrogenase (LDH) is a widely used biomarker in lymphoma patients and is linked to prognosis [1]. However, its low specificity limits its clinical application because, in addition to tumor progression, an elevated LDH level is also found in other non-neoplastic conditions, such as myocardial damage [2]. Moreover, LDH provides no insight into tumor biology [3]. Serum beta-2 microglobulin, an established prognostic factor for multiple myeloma, has also been used for non-Hodgkin's lymphoma patients as a prognostic factor [4]. Similarly, its low specificity and sometimes low sensitivity diminish its clinical utility [5].
However, none of these markers is specific for detecting lymphomas because they are also elevated in other cancers and even in non-neoplastic diseases [13][14][15]. Thus, all the markers emphasize prognostic correlation rather than lymphoma treatment or insights into tumorigenesis.
Diffuse large B-cell lymphoma (DLBCL) is the most common subtype of non-Hodgkin's lymphoma and accounts for 30-40% of all lymphoma cases worldwide [16,17]. The mainstay strategy for treating DLBCL is multidrug immunochemotherapy. However, long-time survival is achieved in only 50% of patients, which underscores the need to develop innovative therapeutic strategies [17,18]. The present study used the comparative proteomics approach to search for candidate lymphoma biomarkers as the proper targets for treatment and study of lymphoma biology.

Culturing DLBCL and B-lymphoblastoid cell lines
Two DLBCL cell lines HT (ACC 567) and SU-DHL-5 (ACC 571) were purchased from DSMZ (Braunschweig, Germany). For comparison, we used a B-lymphoblastoid cell line (LCL), which was derived from human blood B cells immortalized by Epstein-Barr virus infection [19,20]. The culture protocol is described in Supporting Materials and Methods in File S1.

Proteomic analysis
The procedures were done as previously described [21,22], and consisted of three steps: protein separation and in-gel digestion, LC LTQ-FT ICR MS analysis, and Mascot search and label-free quantitative analysis. Proteomic analysis was done in duplicate. The details are provided in Supporting Materials and Methods in File S1.

Immunofluorescent staining
After cytospinning the cells and fixing them in acetone, the slides containing DLBCL cells and LCL were washed with phosphate buffer solution (PBS, pH 7.4), and then incubated with primary antibodies against RanGAP1 (1:100, C-5, sc-28322; Santa Cruz) for 2 hours at room temperature in the dark. After the cells had been washed with PBS, they were incubated with dye-labeled secondary antibodies. Nuclear DNA was stained with 4'-6-diamidino-2-phenylindole (DAPI; 1:1000; Invitrogen, Carlsbad, CA, USA).

Immunohistochemical staining
Immunohistochemical staining was done on deparaffinized tissue sections of formalin-fixed material after microwaveenhanced epitope retrieval as previously described [24]. Staining intensity recognizable in a low-power field (×40) for more than 30% of the tumor cells was deemed positive. The primary antibody, RanGAP1 was from Santa Cruz Biotechnology (1:100, C-5, sc-28322; Santa Cruz).
The cases enrolled for RanGAP1 staining consisted of primary DLBCL (n = 50), lymphoid hyperplasia (n = 12), and other B-cell lymphomas (BCL, n = 180) from the archival files at National Cheng Kung University Hospital. The germinal center (GC) vs. activated B-cell immunophenotype and Ki-67 (MIB-1) proliferation index were determined for DLBCL as described previously [24,25]. Double staining of Ki-67 and RanGAP1 was performed with an automated stainer (Bond-Max; Leica Biosystems Melbourne Pty Ltd, Melbourne, Australia).

Enzyme-linked immunosorbent assay (ELISA) for RanGAP1
To determine whether RanGAP1 reflects the disease status, the serum level of RanGAP1 was measured using ELISA for DLBCL patients at diagnosis. A colorimetric noncompetitive (immunometric sandwich assay) ELISA was done; procedure details are provided in Supporting Materials and Methods in File S1.

Transfecting RANGAP1-specific shRNA into cell lines
Short hairpin RNAs (shRNAs) were designed against the target sequence of RANGAP1 (5′-CAAGAGTGAAGACAAGGTCAA-3′, bases 1834-1854, NM_002883.2). Other two sets of small interfering RNA (siRNA) for RANGAP1 knockdown were also performed in duplicate. The sequences and detailed procedures are described in Supporting Materials and Methods in File S1. The inhibition of RanGAP1 expression was evaluated using immunoblotting. The cell lines were cultured and then collected for further analysis 48 h after they had been transfected.

Cell death and cell cycle assays by flow cytometry
Apoptosis and other forms of cell death were evaluated by measuring the DNA content using annexin V and propidium iodide (PI) affinity as previously described [26]. Briefly, each sample of 2.6 × 10 6 cells was transfected with RANGAP1specific shRNA (shRANGAP1) or control vector, and then cultured in 6 ml of medium. Each sample of 1.5 ml was collected after 48 h. The samples were then centrifuged, and the pellet was incubated with staining solution (PI [50 μg/ml]; 0.1% sodium citrate; 0.1% triton) overnight at 4°C in the dark. Core DNA content was measured using a logarithmic amplification in the FL2 (for annexin V) and FL3 (for PI) channels of the flow cytometer (FACSCalibur with CellQuest Pro 4.0.2; Becton Dickinson, Franklin Lakes, NJ, USA) [27].
Cell-cycle analysis was also measured using flow cytometry. The distribution of the DNA content of individual cells was stained with PI and measured with CellQuest Pro 4.0.2 using a linear amplification in the FL3 channel.

Quantitative real-time polymerase chain reaction (Q-PCR)
The Q-PCR assay was done as previously described [28]. Briefly, total RNA was isolated using an RNA extraction kit (TRIzol; Invitrogen, Carlsbad, CA, USA). Three micrograms of RNA was used to generate cDNA with reverse transcriptase (SuperScript III; Invitrogen). The primers used for Q-PCR are described in Supporting Materials and Methods in File S1.

Statistical analysis
Appropriate statistical tests-t-test, Kendall tau (τ) correlation coefficient, and Mann-Whitney U tests-were used to examine the relationships and correlations between variables. Overall survival was measured from the initial diagnosis until death from any cause; follow-up data of surviving patients were assessed at the last contact date. Estimates of overall survival distribution were calculated using the Kaplan and Meier method [29]. Time-to-event distributions were compared using the log-rank test [30]. The analyses were done using SPSS 13.0 (SPSS, Inc., Chicago, IL, USA).

Ethics Statement
The study was approved by our institutional review board (Institutional Review Board, National Cheng Kung University Hospital-HR-95-72) and was done in accord with the Helsinki Declaration of 1975 as revised in 1983. The written consent was given by the patients for their information to be stored in the hospital database and used for research.

Proteomic analysis yielded 20 proteins overexpressed in DLBCL cell lines compared with the Blymphoblastoid cell line
Two DLBCL cell lines (HT and SU-DHL-5) and a Blymphoblastoid cell line (LCL) were used for comparison to search for candidate biomarkers from cell lysates. Eighty-nine proteins were identified in total with 20 proteins up-expressed and 69 proteins down-expressed in tumor cell lines. These 20 highly-expressed proteins are listed in Table 1 and Table S1 in File S1. We chose RanGAP1 as the interesting candidate for further studies because of its dual subcellular localization. Besides, RanGAP1 is a key regulator of the Ran GTP/GDP cycle and a mitosis coordinator [31,32].

Western blotting confirmed higher RanGAP1 expression in DLBCL lines
Western blotting was used to compare the expression levels of RanGAP1 on neoplastic and reactive B cells. SU-DHL-5 cells expressed 1.6 times (unmodified form, 70 kDa) and 2.1 times (small ubiquitin-related modifier [SUMO]-1 modified form, 90 kDa) more RanGAP1 than did the LCL cells ( Figure 1A). HT cells showed a similar result (1.1 and 4.8 times for unmodified and SUMO form, respectively, Figure 1A). Differential subcellular fractions showed RanGAP1 present in both cytoplasm and nucleus ( Figure 1B). Immunofluorescence demonstrated the cytoplasmic and perinuclear localization ( Figure 1C) as described previously [33].

RanGAP1 immunohistochemically stained the majority of DLBCL cases but only germinal centers of lymph nodes
Immunohistochemical staining of nodal hyperplasia cases (n = 12) showed RanGAP1 positivity mainly in germinal centers with occasionally in dark zone only (Figure 2A). Scattered histiocytes in interfollicular areas were also positive, but the RanGAP1 in Diffuse Large B-Cell Lymphoma PLOS ONE | www.plosone.org other cells, including T cells, were negative. In contrast, the majority of DLBCL cases (46/50, 92%) were positive for RanGAP1 staining in perinuclear and cytoplasmic regions ( Figure 2B).

Higher serum level of RanGAP1 in patients with highgrade BCL than in low-grade BCL and healthy controls
Since there is no commercial assay for RanGAP1, the colorimetric noncompetitive (immunometric sandwich) ELISA has been established using two different anti-RanGAP1 antibodies (RanGAP1 C-5 and RanGAP1 N-19; Santa Cruz) with different antigen-binding sites. The former was raised against amino acids 408-587 and the latter against a peptide mapping at the N-terminus of RanGAP1 of human origin.
Preliminary experiments showed that the detection range was 0-20 ng/mL ( Figure S1 in File S1). The ELISA signal was not influenced by the presence of lipid. ELISA showed relatively high levels of RanGAP1 in the conditioned media of SU-DHL-5 cells (4.51 ng/mL) compared with LCL cells (1.09 ng/mL). For clinical samples, serum levels of RanGAP1 were higher in patients with high-grade BCL (1.71 ± 2.28 ng/mL, n = 62) than in low-grade BCL (0.75 ± 2.12 ng/mL, n = 52) and healthy controls (0.55 ± 1.58 ng/mL, n = 75) (high-grade BCL vs. lowgrade BCL, p = 0.002; high-grade BCL vs. control, p < 0.001, Mann-Whitney U test, Figure S2 in File S1). However, the RanGAP1 serum level was not so sensitive, since half (n = 31) cases of high-grade BCL were not elevated.

RanGAP1 had no prognostic significance for patients with DLBCL
We next evaluated the correlation between RanGAP1 serum level and other clinicopathologic factors in DLBCL cases: tumor stage, treatment response, and patient survival. In our DLBCL cohort, there was positive correlation between high IPI score (≥ 3) and B symptoms (Kendall tau (τ) correlation coefficient: 0.260, p = 0.041), between high LDH level and B symptoms (τ correlation coefficient: 0.247, p = 0.039), and high stage disease (τ correlation coefficient: 0.389, p < 0.001). On survival analyses (Table 3), the poor prognostic factors were old age (p = 0.001), B symptoms (p = 0.009), and a high IPI score (p = 0.003). RanGAP1 serum level had no prognostic significance and showed no correlation with other clinicopathologic factors,

Knockdown of RanGAP1 mRNA increased DLBCL cell death and cell cycle arrest but had no effect on nonneoplastic LCL cells
To test the function of RanGAP1 protein, we knocked down RanGAP1 mRNA to see the effects on cell survival and the cell cycle in DLBCL and LCL cells. The transfection rates for each cell line were as follows: LCL: 41-43%; SU-DHL-5: 76-86%; HT: 51-60%. In contrast to no effect on LCL (9.5% [vector] vs.

RanGAP1 knockdown reduced expression of Aurora kinases and TPX2 in DLBCL lines
To decipher the mechanism underlying RanGAP1knockdown-induced cell-cycle arrest and tumor cell death in DLBCL lines, we tested the effects of RanGAP1 siRNA on the expression of Aurora kinases and TPX2. The former are key regulators of mitotic cell division [34], and the latter is central in spindle assembly [35]. The RanGAP1-specific siRNAs (siRNA1 and siRNA2) downregulated the expression of TPX2 and Aurora-A, -B, and -C kinases in DLBCL lines ( Figure 4A). Q-PCR analysis showed that there was no significant decrease in the mRNA levels of Aurora kinases ( Figure 4B). These data indicated that the RanGAP1-specific siRNAs inhibited the expression of Aurora kinases but did not affect kinase transcription.

ON 01910.Na induced cell death, mitotic cell arrest and hyperphosphorylation of RanGAP1 in DLBCL cell lines but mild effects in non-neoplastic LCL
The ID50 of ON 01910.Na was around 0.031 μM for DLBCL lines by the MTT assay ( Figure 5A). ON 01910.Na showed relatively selective cytotoxicity to DLBCL by flow cytometry analysis ( Figure 5B), and induced more evident mitotic cell arrest in DLBCL lines than in LCL at the concentration between  Figure 5D). Interestingly, ON 01910.Na showed no cytotoxicity for normal CD3 + T cells and CD19 + B cells ( Figure S3 in File S1).

Discussion
Using proteomic analysis to compare DLCBL cell lines with LCL cells, we found the differential expression of RanGAP1, a cell-cycle regulator between DLBCL and reactive lymphoid hyperplasia. As a cell-cycle regulator, RanGAP1 was also frequently overexpressed in other BCL with brisk mitotic activity (lymphoblastic lymphoma/leukemia, and Burkitt lymphoma) or cell-cycle deregulation (mantle cell lymphoma and Hodgkin's lymphoma) [36,37], but only occasionally in low-grade BCL. Interestingly, serum levels of RanGAP1 were higher in patients with high-grade BCL than in low-grade BCL and healthy controls. In vitro, RNA interference with RanGAP1 showed no effects on non-neoplastic LCL cells but induced DLBCL cell death and cell-cycle arrest, by inhibiting the expression of Aurora kinases and TPX2, the crucial regulators of mitosis and cytokinesis. Interestingly, ON 01910.Na, a styryl benzylsulfone capable of multikinase inhibition was selectively cytotoxic for Table 2. Results of RanGAP1 immunostaining in B-cell lymphomas.

Lymphoma Types No. Positivity(%) Gender M/FAge (mean)
Diffuse large B-cell lymphoma 50   DLBCL. Our findings suggest that RanGAP1 is a promising therapeutic target for DLBCL. Ran is a nuclear Ras-like GTPase involved in nuclear transport, RNA processing, cell-cycle progression, and mitotic spindle formation [32]. The nuclear import cycle is orchestrated by the GTP-and GDP-bound states of Ran, which is regulated by nuclear guanine nucleotide-exchange factor (RanGEF, also known as RCC1, regulator of chromosome condensation 1) and cytoplasmic RanGAP1 [38,39]. During mitosis, Ran is involved in mitotic spindle assembly, and RanGAP1 is associated with mitotic spindles that are particularly concentrated near kinetochores [32]. RanGAP1 conjugation with SUMO-1 is required for mitotic localization and is important for spatially regulating the Ran pathway during mitosis [32,40]. Taking all these findings together, RanGAP1 appears to be a key regulator of the Ran GTP/GDP cycle and involved in cell-cycle control.
The RANGAP1 knock-out mice were embryonically lethal, which highlights the pivotal function of the RANGAP1 gene in cell survival [41]. Animal models with a conditional knockout of RanGAP1 in B cells are needed to clarify its function in B-cell development during different stages. Because RanGAP1 is expressed in germinal centers of lymph nodes and BCL with a high-growth fraction, it seems likely to be involved in cell-cycle progression [42]. Indeed, we demonstrated that inhibiting RanGAP1 expression increased DLBCL cell death and cellcycle arrest. Moreover, RanGAP1-specific siRNAs also inhibited the expression of Aurora kinases and TPX2, the key regulators of mitotic cell division, and clinical indicators of aggressive cancers. We thus suggest that downregulation of RanGAP1 induces DLBCL cell-cycle arrest and death by inhibiting the expression of Aurora kinases and TPX2. TPX2, targeting protein for XKLP2 (Xenopus kinesin-like protein 2), is a multifaceted protein for mitosis, including microtubule nucleation and targeting Aurora-A to the spindle [35]. TPX2induced activation of Aurora-A is essential for Ran-stimulated spindle assembly [43]. Aurora kinases, a novel family of serine/ threonine kinases, are substantially involved in mitotic cell division, overexpressed in many human cancers, and correlated with chromosomal instability and clinically aggressive disease [28,34]. The signals for mitotic spindle assembly contain at least two parts: one is the RanGTP signal where Aurora-A acts downstream; the other is the Aurora-B signal generated by localization of Aurora-B kinase [35,44]. Our finding that RanGAP1 knockdown inhibited the expression of Aurora-A and -B suggests that RanGAP1 may be more important than previously thought. Therefore, RanGAP1 is not merely a marker of cell division, because it is also highly expressed in mantle cell and Hodgkin's lymphomas, both of which have relatively lower proliferation activity ( Figure S4 in File S1). Aurora kinases are expressed and active at the highest level during the G2/M phase of the cell cycle [34]. We also found that RanGAP1 knockdown downregulated the expression of Aurora kinases that was correlated with cell-cycle arrest in the G0/G1 phase in DLBCL cells. In contrast, the cytotoxic effect of ON 01910.Na was through prolonged phosphorylation/hyperphosphorylation of RanGAP1.SUMO1 followed by M-phase arrest and the consequent induction of cell death [45].
Many articles have addressed aspects of the molecular biology of RanGAP1, such as the interacting molecules and the regulatory mechanisms [31,32,38]. However, its role in reactive and neoplastic B lymphocytes has not been addressed. In the literature, there is only one article showing RanGAP1 expression in LBCL cell lines [46]. In general, BCL can be divided into low-and high-proliferation fraction categories. The primary pathogenesis of the former depends on inhibiting apoptosis, such as the overexpression of BCL2 and API2, whereas the latter is characterized by brisk proliferation through the dysfunction of cell-cycle regulators [17,37]. Here, we demonstrated that RanGAP1 was highly expressed in BCL with brisk mitotic activity or cell-cycle deregulation [36,37]. Furthermore, inhibiting RanGAP1 expression increased DLBCL tumor cell death and cell-cycle arrest but showed no effect on LCL cells. The selective overexpression of RanGAP1 in aggressive B-cell and Hodgkin's lymphomas may shed light on innovative targeted therapy. Oussenko et al [45] recently reported that ON 01910.Na, an inhibitor of RanGAP1, prolongs the hyperphosphorylation of RanGAP1, which consequently induces apoptosis rather than direct DNA damage. We found a similar effect of ON 01910.Na on DLBCL cells. Furthermore, ON 01910.Na showed absent or only minimal cytotoxicity for normal B and T cells ( Figure S3 in File S1). Given that ON 01910.Na is currently under a randomized phase III trial for patients with refractory myelodysplastic syndrome [47], this drug would be very promising for the RanGAP1-targeted lymphoma therapy.
Proteomic analysis that compares tumorous and nontumorous cells is a powerful tool for discovering tumor-specific proteins [48]. By comparing whole lysates of tumorous cells with those of non-tumorous cells, we found that RanGAP1 was differentially expressed in reactive and neoplastic B-cell proliferations, as well as in BCL with low-and high-proliferation fractions. Interestingly, the serum level of RanGAP1 in patients with high-grade BCL was higher than in low-grade BCL and healthy controls. Because RanGAP1 is present in cytoplasm and perinucleus, and no secreted form is found [49], it is likely that the serum level arises from the death of tumor cells. Thus, the higher serum level might represent more tumor cell death at diagnosis, and might have no prognostic significance and show no correlation with other clinicopathologic factors [50,51]. Although serum RanGAP1 level was significantly higher in patients with high-grade BCL, its poor sensitivity may limit the clinical utility.
In conclusion, using comparative proteomic analysis, we found that RanGAP1, a cell-cycle coordinator, was present in the tumor tissues and patient serum of high-grade BCL. In  vitro, by inhibiting Aurora kinases and TPX2, knockdown of RanGAP1 increased tumor cell death and cell-cycle arrest but had no effect on non-neoplastic cells. Besides, ON 01910.Na induced hyperphosphorylation of RanGAP1.SUMO1, mitotic cell arrest and consequent cell death. Therefore, RanGAP1 is an appropriate lymphoma marker with the potential for tumortargeted therapy [45].

Supporting Information
File S1. Supporting Materials and Methods. Culturing DLBCL and B-lymphoblastoid cell lines. Proteomic analysis. Enzyme-linked immunosorbent assay (ELISA) for RanGAP1. Transfecting RANGAP1-specific shRNA into cell lines. Quantitative real-time polymerase chain reaction (Q-PCR). Figure S1. The standard curve of RanGAP1 serum level. Figure S2. Higher serum level of RanGAP1 in patients with high-grade BCL than in low-grade BCL and healthy controls. Figure S3. No cytotoxic effect on B or T cells from healthy donors at 48 hours. Figure S4. Double staining of RanGAP1 and Ki-67 in mantle cell and Hodgkin lymphomas. Table S1. Protein identification data expressed according to Paris guidelines. (PDF)