Full Text (PDF)

Research Article

Motexafin Gadolinium and Zinc Induce Oxidative Stress Responses and Apoptosis in B-Cell Lymphoma Lines

Philip S. Lecane,1 Mazen W. Karaman,2 Mint Sirisawad,1 Louie Naumovski,1 Richard A. Miller,1 Joseph G. Hacia,2 and Darren Magda1

1Pharmacyclics, Inc., Sunnyvale, California and 2Institute for Genetic Medicine, University of Southern California, Los Angeles, California

Abstract Motexafin gadolinium (MGd, Xcytrin), an expanded porphyrin There is an emerging appreciation of the importance of zinc containing the lanthanide cation gadolinium, is currently in in regulating cancer cell growth and proliferation. Recently, clinical trials for the treatment of several forms of cancer (16). MGd is an electron-affinic compound that mediates electron we showed that the anticancer agent motexafin gadolinium (MGd) disrupted zinc metabolism in A549 lung cancer cells, transfer from a variety of intracellular reducing species, such as leading, in the presence of exogenous zinc, to cell death. Here, ascorbate, NADPH, and thiols, to oxygen to form superoxide and we report the effect of MGd and exogenous zinc on hydrogen peroxide (17–20). Recently, we generated gene expression intracellular levels of free zinc, oxidative stress, proliferation, profiles of plateau phase A549 lung cancer cell cultures treated and cell death in exponential phase human B-cell lymphoma with MGd and observed marked induction of transcript levels of and other hematologic cell lines. We find that increased levels genes that play major roles in controlling intracellular free zinc of oxidative stress and intracellular free zinc precede and levels (6). We also reported that MGd increased intracellular free correlate with cell cycle arrest and apoptosis. To better zinc levels, modulated the cellular toxicity of zinc, and inhibited understand the molecular basis of these cellular responses, cellular bioreductive activity in several human cancer cell lines. gene expression profiling analyses were conducted on Ramos Here, we describe effects of treatment with MGd and/or zinc cell cultures treated with MGd and/or zinc acetate. Cultures acetate on intracellular levels of free zinc, oxidative stress, treated with MGd or zinc acetate alone elicited transcrip- proliferation, cell cycle status, and cell death in B-cell lymphoma cell lines. Based on gene expression profiling and other functional tional responses characterized by induction of metal response element–binding transcription factor-1 (MTF-1)–regulated analyses, we find that increased levels of oxidative stress and and hypoxia-inducible transcription factor-1 (HIF-1)–regulated intracellular free zinc lead to the expression of genes under the genes. Cultures cotreated with MGd and zinc acetate control of metal response element–binding transcription factor-1 displayed further increases in the levels of MTF-1– and HIF- (MTF-1), hypoxia-inducible transcription factor-1 (HIF-1), and 1–regulated transcripts as well as additional transcripts NF-E2-related factor 2 (NRF-2), and correlate with cell cycle arrest regulated by NF-E2–related transcription factor 2. These data and apoptotic response. Overall, these studies lead us to suggest provide insights into the molecular changes that accompany that cotreatment of Ramos cells with MGd and zinc acetate the disruption of intracellular zinc homeostasis and support a increases intracellular free zinc levels with a concomitant increase role for MGd in treatment of B-cell hematologic malignancies. in oxidative stress levels that activate adaptive survival responses but eventually lead to cell death by disrupting redox balance. (Cancer Res 2005; 65(24): 11676-88)

Introduction Materials and Methods The intracellular abundance of loosely bound or free zinc can Cells and Cell Culture Reagents have a profound effect on cellular metabolism, survival, and Ramos, Raji, and DB B-cell lymphoma lines were purchased from growth. For example, elevated levels of intracellular free zinc have American Type Culture Collection (Rockville, MD). DHL-4 and HF-1 lines been proposed to inhibit glycolysis via glyceraldehyde phosphate were obtained from Ronald Levy (Stanford University, Stanford, CA). The dehydrogenase, the citric acid cycle via the a-ketoglutarate HF-1 cell line was derived from a patient with follicular lymphoma (21). dehydrogenase complex and mitochondrial respiration (1–3). DHL-4 was derived from a patient with diffuse large cell lymphoma (22). Unless otherwise indicated, all cell culture reagents were purchased from Free zinc can also inhibit the pentose cycle–dependent enzyme, Invitrogen (Carlsbad, CA). Cells were cultured in a 5% CO2 incubator at thioredoxin reductase (4–7). Furthermore, zinc levels have been 37jC at a density between 0.2 106 and 1 106 cells/mL as previously n reported to modulate protein kinase C (8, 9), nuclear factor B described (6). MGd was prepared as a 2 mmol/L (2.3 mg/mL) formulation

(10, 11), and p53 (12) activities and signaling pathways relevant to in 5% aqueous mannitol. Zinc acetate (ZnOAc2) and cobalt acetate (Aldrich carcinogenesis (13, 14). In light of these effects, it is perhaps not Chemical, Milwaukee, WI) were used as 2 mmol/L formulations in 5% surprising that zinc possesses both antiapoptotic or proapoptotic aqueous mannitol. properties that are dependent on intracellular levels (6, 15). Apoptosis Assays Annexin V–propidium iodide. Cells from exponential phase cultures were treated with MGd, zinc, or control (5% mannitol) solution for 24 or Note: Supplementary data for this article are available at Cancer Research Online 48 hours. After incubation, cells were harvested and washed twice with a (http://cancerres.aacrjournals.org/). solution of 0.5% bovine serum albumin (BSA) in HBSS. An aliquot of cells Requests for reprints: Darren Magda, Pharmacyclics, Inc., 995 E. Arques Avenue, (1 106) was added to 500 AL diluted binding buffer from the Annexin Sunnyvale, CA 94085. Phone: 408-774-3318; Fax: 408-328-3689; E-mail: dmagda@ V–propidium iodide kit (BD Biosciences, San Jose, CA). Cells were pelleted, pcyc.com. I2005 American Association for Cancer Research. resuspended in 100 AL of diluted binding buffer, and treated with the doi:10.1158/0008-5472.CAN-05-2754 Annexin V–propidium iodide reagent as per the protocol of the

Cancer Res 2005; 65: (24). December 15, 2005 11676 www.aacrjournals.org

Downloaded from cancerres.aacrjournals.org on September 26, 2021. © 2005 American Association for Cancer Research. Motexafin Gadolinium and Zinc in B-Cell Lines manufacturer. Flow cytometry was done on a FACSCalibur instrument and ref. 24). Exponential-phase cultures were treated with control 5% mannitol data were analyzed using the CellQuest Pro software package (BD vehicle or ZnOAc2 in the presence or absence of MGd, as described above, Biosciences). for 4 hours. Following treatment, cells were isolated by centrifugation. Cell Mitochondrial membrane potential. Loss of the mitochondrial pellets were washed and resuspended in a solution of 0.5% BSA in PBS. An 6 membrane potential (DWm) of cells was measured by the use of JC-1 aliquot of 10 cells (200 AL) was removed, centrifuged, and treated with (Molecular Probes, Inc., Eugene, OR). Cells undergoing early apoptosis lose FluoZin-3 reaction buffer as described (24). An aliquot of the cell fluorescence in the 585 nm channel and gain it in the 530 nm channel. suspension was supplemented with 2 Ag/mL propidium iodide (Sigma Briefly, cells cultured as described above were washed twice with complete Biochemical), incubated for 5 minutes, and subjected to two-variable flow medium, resuspended in 0.5 mL JC-1 solution (10 Ag/mL in complete cytometric analysis as described previously (6). medium), and incubated at 37jC for 15 minutes. Cells were isolated by centrifugation, washed once, and then resuspended in 0.5 mL solution of Measurement of Reactive Oxygen Species Reactive oxygen species were measured in live cells as intracellular 0.5% BSA in PBS and assayed immediately on the flow cytometer. peroxides by monitoring the oxidation of 2V,7V-dichlorofluorescin-diacetate Cellular Proliferation to 2V,7V-dichlorofluorescein (Molecular Probes). Cells (1 106/mL) were The proliferation of exponential phase cultures was assessed by incubated in a solution of 1 Ag/mL of 2V,7V-dichlorofluorescin-diacetate in colorimetric assay. In brief, 2 105 suspension cells per well were seeded 0.5% BSA in HBSS for 15 minutes at 37jC. Two milliliters of additional 0.5% on 96-well V-bottomed microtiter plates. Stock solutions of control vehicle, BSA in HBSS were added, cells were isolated by centrifugation, and the

MGd, or ZnOAc2 in medium were added and plates were incubated at pellet was resuspended in a solution of 50 Ag/mL of 7-aminoactinomycin D 37jC under a 5% CO2/95% air atmosphere. After 24 hours, medium was (7-AAD) in 0.5% BSA in HBSS. Cell suspensions were incubated at ambient replaced with fresh medium. After 2 additional days, medium was temperature for 2 to 3 minutes and stored on ice until analysis. The exchanged with fresh medium (150 AL/well) supplemented with 3- fluorescent intensity in live (i.e., 7-AAD impermeable) cells was analyzed by (4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (0.5 mg/mL, flow cytometry. Sigma Biochemical, St. Louis, MO). Plates were incubated at 37jC and viable cells were measured as described (23). Cell Cycle Analysis Exponential phase cultures were treated with control 5% mannitol Intracellular Free Zinc vehicle or ZnOAc2 in the presence or absence of MGd as described above. The concentration of intracellular free zinc was assessed using the ion- Thirty minutes before harvest, cultures were treated with 5-bromo-2V- specific fluorescent probe, FluoZin-3-AM (FluoZin-3, Molecular Probes; deoxyuridine (BrdUrd) at a final concentration of 10 Amol/L. Cells were

Figure 1. MGd treatment alters levels of intracellular free zinc, oxidative stress, and apoptotic response of Ramos cells to zinc. Exponential phase Ramos cultures were treated with control vehicle (Mannitol), zinc acetate (Zinc,50Amol/L), MGd (10 Amol/L), or the combination for up to 24 hours in duplicate experiments. Columns: A, fold increase of FluoZin-3 fluorescence in live-gated cells. B, fold increase of dichlorofluorescein (DCF) fluorescence in live-gated cells. C, percentage of Annexin V–stained cells. D, percentage of live-gated cells exhibiting green (nonaggregated) JC-1 fluorescence characteristic of lost mitochondrial membrane potential. Bars, SD. www.aacrjournals.org 11677 Cancer Res 2005; 65: (24). December 15, 2005

Figure 2. Treatment of Ramos cells with MGd and zinc inhibits DNA synthesis. Exponential phase Ramos cultures were treated with zinc acetate (50 Amol/L) and MGd (10 Amol/L) for up to 8 hours. BrdUrd (10 Amol/L) was added 30 minutes before harvest. After the indicated time intervals, cells were washed, fixed, and stained with fluorescein-conjugated anti-BrdUrd antibody (Fluorescein) and 7-AAD. Cell cycle occupancy was analyzed by flow cytometry using fluorescein signal as a measure of DNA synthesis and 7-AAD signal as a measure of DNA content. The effect of treatment with control vehicle, hydroxyurea (HU, 250 Amol/L), ionizing radiation (XRT, 10 Gy), and 5-fluoro-2V-deoxyuridine (FUDR,10Amol/L) after 8 hours is shown for comparison. Representative of two experiments.

isolated by centrifugation, washed once with 0.5% BSA in PBS, and the Gene Expression Profiling resulting cell pellets were fixed using 0.5 mL Cytofix/Cytoperm reagent MGd (10 Amol/L), ZnOAc2 (25 or 50 Amol/L), the combinations, or (BD Biosciences). After incubation at ambient temperature for 30 control (5% mannitol) solution was added to Ramos cultures. Each minutes, cells were isolated by centrifugation, washed with 3% fetal treatment was done in triplicate. After 4 hours of incubation, all cultures bovine serum in PBS, resuspended in 10% DMSO in medium, and stored were washed twice with 0.5% BSA in HBSS and total RNA was isolated and at 20jC until analysis. Cells were stained using a fluorescein-conjugated subjected to analysis on Human Genome U133 Plus 2.0 Arrays (Affymetrix, anti-BrdUrd antibody (clone PRB1, E-Bioscience, San Diego, CA) and 7- Santa Clara, CA) as described (28). Microarray Suite version 5.0 software AAD. Cell cycle occupancy was analyzed by flow cytometry using (Affymetrix) was used to generate raw gene expression scores and fluorescein signal as a measure of DNA synthesis and 7-AAD signal as a normalize the relative hybridization signal from each experiment as measure of DNA content as described (25). For comparison, cultures described (28). All gene expression scores were set to a minimum value of were treated with 5-fluoro-2V-deoxyuridine, hydroxyurea, or irradiated 50 to minimize noise associated with less robust measurements of rare using a 137Cs irradiator (Model 40 Gammacell, J.L. Shepherd & transcripts. Both the parametric Student’s t test and the permutation- Associates, San Fernando, CA). based significance analysis of microarrays were used to determine genes differentially expressed in treatment versus control groups (29). We report Hypoxia-inducible Transcription Factor-1 ELISA data from genes that are at least 1.5-fold differentially expressed relative to a Total HIF-1 protein was detected by sandwich ELISA using the DuoSet controls using the Student’s t test (P V 0.005) because this empirically IC HIF-1a ELISA kit obtained from R&D Systems (Minneapolis, MN). All proved to be a more stringent criterion than significance analysis of incubations were done at ambient temperature. Briefly, 96-well plates were microarray analysis using the same 1.5-fold cutoff and a <1% false discovery coated with HIF-1a capture antibody overnight before blocking with rate (data not shown). All scaled fluorescent intensity values and .cel files 5% BSA in wash buffer. Protein lysates (50 Ag protein per well prepared are available at http://hacialab.usc.edu/supplement/lecane_etal_2005/. according to the instructions of the manufacturer) were added for 2 hours, whereupon plates were washed and a biotinylated detection antibody specific for HIF-1a was added. A streptavidin-horseradish peroxidase Results format was used for detection. The absorbance at 450 minus 570 nm was Intracellular free zinc is elevated in Ramos cells treated measured using a microplate reader (SpectraMax Plus, Molecular Devices, with motexafin gadolinium and zinc. We have previously shown Palo Alto, CA). HIF-1a concentrations were calculated by linear regression A A using a standard curve prepared from HIF-1a standard supplied with the that treatment of Ramos cells with 10 mol/L MGd and 50 mol/L ELISA kit. zinc led to synergistic increases in the level of free intracellular zinc within 6 hours of treatment as measured using the ion- Western Blotting specific dye FluoZin-3 (6). To further characterize this response Western blotting was done as described (26, 27). Antibodies against and relate it to other cellular phenotypes, we analyzed the kinetics heme oxygenase 1 and metallothioneins 1 and 2 (clone E9) were obtained of intracellular free zinc accumulation. Cotreatment of Ramos cells from Santa Cruz Biotechnology (Santa Cruz, CA) and DAKO (Glostrup, A A Denmark), respectively. All membranes were blotted with an anti-Hsc70 with 10 mol/L MGd and 50 mol/L zinc led to a 4-fold increase (Santa Cruz Biotechnology) antibody to control for loading and transfer. in median FluoZin-3 fluorescence within 2 hours compared with f Bands were imaged and quantified in the linear range and normalized control cells (Fig. 1A). This signal increased to 8-fold within to Hsc70 by using the Odyssey Infrared Imaging System (LICOR, Inc., 12 hours and remained constant thereafter. Treatment with Lincoln, NE). 10 Amol/L MGd or 50 Amol/L zinc alone led to small increases

Cancer Res 2005; 65: (24). December 15, 2005 11678 www.aacrjournals.org

Downloaded from cancerres.aacrjournals.org on September 26, 2021. © 2005 American Association for Cancer Research. Motexafin Gadolinium and Zinc in B-Cell Lines in FluoZin-3 fluorescence within 2 hours, which returned to baseline levels by 12 hours. As a negative control, we found no increase in cellular fluorescence at 530 nm in the absence of FluoZin-3 (data not shown). Motexafin gadolinium and zinc increase oxidative stress in Ramos cells. We measured oxidative stress in Ramos cells treated with MGd and/or zinc over time by monitoring the conversion of dichlorofluorescin acetate to dichlorofluorescein (Fig. 1B). Cultures of Ramos cell treated with both 10 Amol/L MGd and 50 Amol/L zinc displayed 1.8-fold increase in median dichlorofluorescein signal within 2 hours, which gradually diminished over the 24-hour time course of the experiment to background levels. Treatment with MGd or zinc alone also increased dichlorofluorescein signal, albeit not to the same degree as the combination. Treatment of Ramos cells with hydrogen peroxide also increased dichlorofluorescein and FluoZin-3 fluorescence (see Supplementary Data). Motexafin gadolinium and zinc treatment leads to apoptosis in Ramos cells. To determine the rate at which cotreated Ramos cells undergo cell death, Ramos cultures treated as above were analyzed using FITC-labeled Annexin V reagent to detect early and late apoptotic events (Fig. 1C). In addition, the dye JC-1 was used to assess mitochondrial function (Fig. 1D). In cultures treated with MGd and zinc, 21% of Ramos cells exhibited a positive Annexin V signal within 8 hours of treatment. This fraction increased to 30% within 12 hours and 68% by 24 hours. Analogous results were obtained using JC-1, with 38% of cells exhibiting nonaggregated (green) JC-1 fluorescence characteristic of mitochondrial dysfunction within 8 hours of combined treatment with MGd and zinc. This fraction increased to 52% by 12 hours and 74% by 24 hours. No significant change in Annexin V signal or JC-1 fluorescence was observed within 4 hours or as a result of treatment with MGd or zinc alone. Motexafin gadolinium and zinc treatment leads to cell cycle arrest in Ramos cells. Next, we sought to examine the kinetics of growth rate responses to cotreatment with MGd and zinc acetate. To do this, Ramos cultures cotreated as above were labeled with BrdUrd and 7-AAD to determine cell cycle occupancy (25). As shown in Fig. 2, cotreatment with MGd and zinc halted BrdUrd incorporation in S-phase cells actively synthesizing DNA within 8 hours. It also led to inhibition of cell entry and progression through G1 and G2-M phases (as determined by DNA content analysis; data not shown). The effects of treatment with 5-fluoro- 2V-deoxyuridine, hydroxyurea, and ionizing radiation are shown for comparison. Figure 3. MGd treatment alters levels of oxidative stress, intracellular free zinc, Motexafin gadolinium modulation of zinc activity in other and apoptotic response of B-cell lines to zinc. Exponential phase cultures were treated with control vehicle, zinc acetate (50 Amol/L), MGd (10 Amol/L), or lymphoma cell lines. The effect of MGd and zinc on the combination in duplicate experiments. Columns: A, fold increase of FluoZin-3 proliferation was examined in Ramos and several other fluorescence in live-gated cells after 4 hours of treatment. B, fold increase hematologic cell lines (see Supplementary Data). The B-cell lines of dichlorofluorescein fluorescence in live-gated cells after 4 hours of treatment. C, percentage of Annexin V–stained cells after 24 and 48 hours of treatment. (Ramos, Raji, DB, DHL-4, and HF-1) seemed to be more sensitive than the Jurkat T-cell line and the myeloid cell lines K562 and HL-60. In all cell lines, sensitivity to zinc was increased by MGd, no significant change in median dichlorofluorescein fluorescence whereas low concentrations of zinc or MGd alone had no in the DB or HF-1 lines at 4 hours. No changes in FluoZin-3, significant effect. dichlorofluorescein, or Annexin V–FITC fluorescence were The five B-cell lines were further tested for changes in observed in Jurkat, K562, or HL-60 lines under these conditions intracellular free zinc levels and oxidative stress after 4 hours (data not shown). and apoptosis after 24 and 48 hours of treatment with 10 Amol/L Gene expression profiling of motexafin gadolinium–treated MGd and 50 Amol/L zinc acetate (Fig. 3). Increases in FluoZin-3, Ramos cells. To assess the effects of MGd or zinc treatment on dichlorofluorescein, and Annexin V–FITC fluorescence relative to gene expression profiles, total cellular RNA was isolated from control were in the following order: Ramos > DHL-4 > DB > Raji > exponential phase Ramos cultures treated in triplicate with control HF-1. This roughly matches the sensitivity of these lines to vehicle (5% mannitol), 10 Amol/L MGd, 25 or 50 Amol/L zinc treatment with zinc in the proliferation assay. However, there was acetate, or the zinc and MGd combinations for 4 hours and www.aacrjournals.org 11679 Cancer Res 2005; 65: (24). December 15, 2005

Downloaded from cancerres.aacrjournals.org on September 26, 2021. © 2005 American Association for Cancer Research. Cancer Research analyzed on oligonucleotide microarrays (28). Twenty-nine tran- tially regulated in response to treatment with 25 Amol/L zinc scripts (25 of which had annotated functions) showed differential acetate, 347 transcripts were differentially expressed in response to expression in response to MGd treatment that reached our criteria treatment with the higher concentration of 50 Amol/L zinc acetate for statistical significance (z1.5-fold, P V0.005; Table 1). The most (284 up-regulated and 63 down-regulated). A total of 12 of 29 (41%) prominent consequence of MGd treatment was the up-regulation of the transcripts significantly changed by MGd treatment were of MTF-1-regulated genes (30–33), including metallothionein and also changed (z1.5-fold in the same direction; P V 0.005) by zinc transporter family transcripts, as was observed previously treatment with 50 Amol/L zinc acetate (Fig. 4A). However, 28 of 29 in A549 lung cancer plateau phase cultures (6). The levels of six (97%) MGd-responsive genes were also differentially expressed in transcripts were down-regulated, including SLC39A10,which the same direction in Ramos cultures treated with 50 Amol/L zinc encodes a transporter involved in the uptake of zinc. Interestingly, using less stringent criteria (z1.2-fold, P V 0.05). a splice variant of this transporter was significantly increased, To simplify presentation and interpretation, a selected group presumably reflecting additional mechanisms operating to regulate of the 1,226 differentially expressed transcripts in the 10 Amol/L levels of intracellular free zinc. In addition, we observed HIF-1- MGd + 50 Amol/L zinc acetate group are shown in Table 2. A total related transcripts displaying significant changes, including DDIT4, of 64% (178 of 278) of the differentially expressed transcripts in the EGLN1, and PFKFB3 (Table 1). Similar expression patterns were 10 Amol/L MGd and 25 Amol/L zinc acetate group were shared observed in response to 50 Amol/L zinc. with the 10 Amol/L MGd and 50 Amol/L zinc acetate group Depending on treatment condition, the numbers of significantly (Fig. 4B). Using less stringent criteria (1.2-fold, P V 0.05), 253 of 278 changed transcripts were as follows: 25 Amol/L zinc acetate (3) < (91%) of the transcripts differentially expressed in the former MGd alone (29) < MGd + 25 Amol/L zinc acetate (278) f 50 Amol/L group were shared with the latter group. This is especially zinc acetate (347) V MGd + 50 Amol/L zinc acetate (1,226). interesting given that changes in cell viability were observed in Whereas only one annotated transcript (SLC39A10) was differen- the 10 Amol/L MGd and 50 Amol/L zinc acetate group but not in

Table 1. Genes differentially expressed in response to MGd treatment

c b Gene ID* Gene symbol Gene description MGd Zinc MGd + zincx

FCk P FCk P FCk P

7779 SLC30A1 Solute carrier family 30 (zinc transporter), member 1 7.6 0.000 7.3 0.005 20.0 0.001 4502 MT2A Metallothionein 2A 5.5 0.000 8.2 0.004 42.4 0.003 54541 DDIT4 DNA damage–inducible transcript 4 5.2 0.000 3.7 0.006 30.1 0.001 4501 MT1X Metallothionein 1X 4.9 0.001 7.3 0.005 34.3 0.002 AF333388{ — Metallothionein 1 H-like protein 3.7 0.002 7.5 0.002 26.4 0.013 4494 MT1F Metallothionein 1F (functional) 3.5 0.000 4.7 0.001 18.7 0.003 AF078844{ — Hqp0376 protein 2.6 0.004 3.4 0.001 10.6 0.005 3303 HSPA1A Heat shock 70 kDa protein 1A 2.3 0.002 2.2 0.002 26.7 0.001 57181 SLC39A10 Solute carrier family 39 (zinc transporter), member 10 2.1 0.002 1.9 0.006 2.8 0.003 3304 HSPA1B Heat shock 70 kDa protein 1B 1.9 0.004 1.7 0.010 10.7 0.000 54583 EGLN1 Egl nine homologue1 (C. elegans) 1.8 0.000 2.0 0.001 3.3 0.000 5567 PRKACB Protein kinase, cAMP-dependent, catalytic, b 1.8 0.003 1.4 0.007 1.8 0.001 113791 MGC17330 HGFL gene 1.6 0.004 2.3 0.024 4.8 0.002 5209 PFKFB3 6-Phosphofructo-2-kinase/fructose-2,6-biphosphatase 3 1.6 0.000 1.5 0.014 2.0 0.004 2976 GTF3C2 General transcription factor IIIC, polypeptide 2, b110 kDa 1.6 0.000 1.9 0.042 1.9 0.013 55818 JMJD1A Jumonji domain containing 1A 1.6 0.001 1.4 0.001 1.8 0.001 115330 GPR146 G protein-coupled receptor 146 1.6 0.003 1.0 0.585 1.6 0.002 54407 SLC38A2 Solute carrier family 38, member 2 1.6 0.002 2.2 0.002 9.2 0.001 154743 FLJ31818 Hypothetical protein FLJ31818 1.5 0.000 1.4 0.000 3.1 0.006 55900 ZNF302 Zinc finger protein 302 1.5 0.001 1.5 0.014 1.8 0.006 AI339606{ —— 1.5 0.005 1.4 0.013 1.8 0.002 7752 ZNF200 Zinc finger protein 200 1.7 0.005 1.9 0.009 2.1 0.000 55114 ARHGAP17 Rho GTPase activating protein 17 1.8 0.001 2.2 0.000 4.7 0.000 1839 DTR Diphtheria toxin receptor 1.8 0.003 1.7 0.013 1.6 0.013 1102 CHC1L Chromosome condensation 1-like 1.8 0.003 1.5 0.031 2.6 0.001 57181 SLC39A10 Solute carrier family 39 (zinc transporter), member 10 2.4 0.005 2.4 0.005 2.4 0.005

*http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=gene. cCells treated with 10 Amol/L MGd compared with control cells treated with mannitol. bCells treated with 50 Amol/L zinc compared with control cells treated with mannitol. xCells treated with 10 Amol/L MGd and 50 Amol/L zinc compared with control cells treated with mannitol. kFold change. {Accession number.

Cancer Res 2005; 65: (24). December 15, 2005 11680 www.aacrjournals.org

Figure 4. Treatment of Ramos cells with MGd and zinc alters RNA and protein expression. Exponential phase cultures were treated with control vehicle, zinc acetate (25 or 50 Amol/L), MGd (10 Amol/L), or the combination. A, Venn diagram showing relationship between transcripts significantly altered by treatment with 10 Amol/L MGd or 50 Amol/L zinc. B, Venn diagram showing relationship between transcripts significantly altered by treatment with 10 Amol/L MGd and 25 or 50 Amol/L zinc. C, levels of HIF-1a protein relative to control vehicle after 4 hours treatment as measured by ELISA. Effect of incubation in 2% oxygen atmosphere (Hypoxia) or with cobalt acetate (Cobalt) is shown for comparison. Bars, 1 SD. D, representative Western blots showing levels of heme oxygenase 1 (HO1) and metallothioneins 1 and 2 (MT) after 16 hours treatment. Hsc70 served as a loading control. the 10 Amol/L MGd and 25 Amol/L zinc acetate group. Overall, we cell lines that might not be representative of primary tumor cells, observed a trend toward larger magnitudes of differential gene this approach has the advantage of avoiding heterogeneity that can expression in response to cotreatment with MGd and zinc relative confound analysis of drug activity in patient samples. Thus, we to individual treatments (Tables 1 and 2). first sought to understand the effect of drug treatment in cultured Levels of hypoxia-inducible transcription factor-1A, metal- B-cell lines as a prelude to further studies involving primary cancer lothioneins, and heme oxygenase-1. We showed that some of the cells. As we had previously shown that the cellular activity of transcriptional changes determined by DNA microarray analysis 10 Amol/L MGd was enhanced in the presence of 25 and 50 Amol/L were also reflected in alterations of protein expression. Cellular exogenous zinc (6), we chose to focus on these zinc concentrations levels of HIF-1a in Ramos cells were measured by ELISA following in our studies. They are also relevant given that (a) standard tissue treatment with MGd and zinc for 4 hours (Fig. 4C). Total cellular culture conditions are deficient in zinc, having an estimated 3- to HIF-1a levels were increased 1.5- to 3-fold by treatment with MGd, 6-fold lower concentration (f4 Amol/L) compared with normal zinc, or the combination. As expected, HIF-1a levels were also human plasma samples and (b) interstitial fluid zinc concen- increased by treatment with cobalt acetate or by the use of hypoxic trations can vary greatly in vivo (1, 34). Within 2 hours of MGd culture conditions. Levels of metallothioneins and heme oxygenase-1 and/or zinc treatment, Ramos cells showed significant increases in were shown by Western blot to be increased following cotreatment intracellular free zinc (Fig. 1A). These levels continued to increase with MGd and zinc for 16 hours (Fig. 4D). Metallothionein and for at least 12 hours and remained high, at least in the group heme oxygenase-1 are proteins with expression induced by MTF-1 cotreated with MGd and zinc. This could represent a catastrophic and NRF-2, respectively. loss of zinc homeostasis in these cells due to an overwhelming of cellular stress responses. Four other B-cell lines treated with MGd and zinc displayed increased intracellular free zinc levels Discussion after 4 hours (Fig. 3A) albeit to variable degrees. MGd is an electron-affinic anticancer agent that is being tested We also observed substantial increases in levels of reactive clinically in a variety of settings, including the treatment of B-cell oxygen species in Ramos cells within 2 hours of treatment with lymphoma. Although the current study is limited by its reliance on 10 Amol/L MGd and/or 50 Amol/L zinc (Fig. 1B). However, in www.aacrjournals.org 11681 Cancer Res 2005; 65: (24). December 15, 2005

Table 2. Transcriptional responses of Ramos cells cotreated with MGd and zinc

Category Gene ID* Symbol Gene 10 Amol/L 10 Amol/L 10 Amol/L 50 Amol/L description MGd + MGd + MGd zinc 50 Amol/L zinc 25 Amol/L zinc

c c c c FC P FC P FC P FC P

Transition metal cation homeostasis Metal sequestration 4493 MT1E Metallothionein 1E 3.2 0.050 1.9 0.025 1.1 0.128 1.3 0.112 (functional) 4494 MT1F Metallothionein 1F 18.7 0.003 8.5 0.001 3.5 0.000 4.7 0.001 (functional) 4495 MT1G Metallothionein 1G 9.1 0.005 5.1 0.022 1.6 0.176 2.0 0.124 4496 MT1H Metallothionein 1H 27.4 0.002 13.6 0.001 2.5 0.017 5.0 0.008 ——Metallothionein 1 H-like 26.4 0.013 24.6 0.001 3.7 0.002 7.5 0.002 4501 MT1X Metallothionein 1X 24.6 0.002 14.3 0.001 3.4 0.008 5.2 0.004 — MT2A Metallothionein 2A 42.4 0.003 20.4 0.004 5.5 0.000 8.2 0.004 ——Similar to human 10.6 0.005 7.0 0.001 2.6 0.004 3.4 0.001 metallothionein-IF Metal transporters 7779 SLC30A1 Solute carrier family 30 20.0 0.001 12.0 0.001 7.6 0.000 7.3 0.005 (zinc transporter), member 1 7779 SLC30A1 Solute carrier family 30 6.7 0.007 2.7 0.002 1.3 0.108 2.1 0.020 (zinc transporter), member 1 57181 SLC39A10 Solute carrier family 39 2.8 0.003 2.7 0.000 2.1 0.002 1.9 0.006 (zinc transporter), member 10 57181 SLC39A10 Solute carrier family 39 2.4 0.005 2.4 0.005 2.4 0.005 2.4 0.005 (zinc transporter), member 10 23516 SLC39A14 Solute carrier family 39 2.6 0.010 1.0 0.423 1.0 0.482 2.3 0.007 (zinc transporter), member 14 55334 SLC39A9 Solute carrier family 39 3.7 0.015 1.2 0.205 1.2 0.430 2.4 0.072 (zinc transporter), member 9 7037 TFRC Transferrin receptor 5.3 0.001 1.3 0.357 1.6 0.111 5.4 0.002 (p90, CD71) Oxidative stress Antioxidant 3162 HMOX1 Heme oxygenase 4.4 0.010 1.0 1.000 1.0 1.000 1.0 1.000 (decycling) 1 5034 P4HB Procollagen-proline, 2.7 0.001 1.4 0.039 1.4 0.062 2.4 0.002 2-oxoglutarate 4-dioxygenase 6415 SEPW1 Selenoprotein W, 1 2.8 0.003 1.9 0.005 1.4 0.110 1.4 0.095 83667 SESN2 Sestrin 2 4.4 0.000 2.2 0.007 1.2 0.315 1.3 0.103 7296 TXNRD1 Thioredoxin reductase 1 2.0 0.004 2.1 0.000 1.2 0.120 1.3 0.006 Chaperone related 22824 APG-1 Heat shock protein 2.6 0.000 1.6 0.003 1.0 0.910 1.4 0.102 (hsp110 family) 9531 BAG3 BCL2-associated 8.0 0.005 3.1 0.000 1.4 0.031 1.2 0.251 athanogene 3 821 CANX DnaJ (Hsp40) homologue, 2.5 0.002 1.2 0.190 1.2 0.263 2.0 0.004 subfamily B, member 1 3337 DNAJB1 DnaJ (Hsp40) homologue, 4.7 0.000 2.3 0.000 1.3 0.013 1.5 0.017 subfamily B, member 4 11080 DNAJB4 DnaJ (Hsp40) homologue, 3.0 0.002 1.5 0.003 1.1 0.354 1.3 0.030 subfamily B, member 6 10049 DNAJB6 DnaJ (Hsp40) homologue, 3.2 0.000 1.9 0.006 1.3 0.047 1.4 0.021 subfamily C, member 3

(Continued on the following page)

Cancer Res 2005; 65: (24). December 15, 2005 11682 www.aacrjournals.org

Table 2. Transcriptional responses of Ramos cells cotreated with MGd and zinc (Cont’d)

Category Gene ID* Symbol Gene 10 Amol/L 10 Amol/L 10 Amol/L 50 Amol/L description MGd + MGd + MGd zinc 50 Amol/L zinc 25 Amol/L zinc

c c c c FC P FC P FC P FC P

5611 DNAJC3 Heat shock 105 kDa/110 3.3 0.001 1.2 0.270 1.1 0.418 3.1 0.007 kDa protein 1 3303 HSPA1A Heat shock 27 kDa protein 1 26.7 0.001 6.8 0.000 2.3 0.002 2.2 0.002 3304 HSPA1B Heat shock 60 kDa 10.7 0.000 4.7 0.000 1.9 0.004 1.7 0.010 protein 1 (chaperonin) 3308 HSPA4 Heat shock 70 kDa 3.6 0.001 1.7 0.061 1.5 0.116 2.0 0.016 protein 1A/1B 3315 HSPB1 Heat shock 70 kDa 4.2 0.020 1.9 0.009 1.1 0.090 1.2 0.113 protein 1B 3326 HSPCB Heat shock 70 kDa 2.2 0.003 1.4 0.064 1.3 0.154 1.7 0.025 protein 4 3329 HSPD1 Heat shock 90 kDa 4.6 0.002 1.2 0.344 1.3 0.226 1.6 0.007 protein 1, b 10808 HSPH1 Heat shock protein 6.6 0.000 2.6 0.008 1.4 0.148 2.3 0.009 (hsp110 family) Glutathione related 1491 CTH Cystathionase 3.9 0.000 2.2 0.005 1.3 0.120 1.7 0.007 (cystathionine c-lyase) 2730 GCLM Glutamate-cysteine ligase, 2.4 0.005 1.6 0.000 1.0 0.623 1.2 0.223 modifier subunit 84706 GPT2 Glutamic pyruvate 1.8 0.001 1.2 0.081 1.1 0.471 1.3 0.046 transaminase 2 2744 GLS Glutaminase 2.0 0.001 1.7 0.001 1.2 0.027 1.2 0.066 2936 GSR Glutathione reductase 2.2 0.008 1.2 0.123 1.1 0.347 1.9 0.001 Transporters 6509 SLC1A4 Solute carrier family 1, 3.3 0.000 1.5 0.011 1.1 0.377 2.0 0.009 member 4 6520 SLC3A2 Solute carrier family 16, 2.1 0.006 1.4 0.027 1.1 0.244 1.3 0.108 member 1 6566 SLC16A1 Solute carrier family 16, 3.4 0.006 1.3 0.273 1.6 0.060 3.7 0.002 member 6 9120 SLC16A6 Solute carrier family 3, 2.8 0.000 1.6 0.016 1.1 0.552 1.5 0.013 member 2 54407 SLC38A2 Solute carrier family 38, 10.1 0.000 4.0 0.000 1.6 0.012 2.2 0.002 member 2 8501 SLC43A1 Solute carrier family 43, 2.1 0.001 1.6 0.001 1.2 0.009 1.3 0.030 member 1 6541 SLC7A1 Solute carrier family 7, 2.4 0.010 1.3 0.263 1.0 0.986 1.8 0.057 member 1 23657 SLC7A11 Solute carrier family 7, 10.2 0.003 4.8 0.003 1.8 0.018 2.2 0.107 member 11 Metabolism 178 AGL Amylo-1,6-glucosidase, 1.6 0.009 1.3 0.029 1.1 0.321 1.1 0.362 4-a-glucanotransferase 7915 ALDH5A1 Aldehyde dehydrogenase 3.5 0.000 1.6 0.003 1.3 0.026 3.5 0.019 5 family, member A1 4329 ALDH6A1 Aldehyde dehydrogenase 1.9 0.001 1.1 0.222 1.1 0.377 1.3 0.077 6 family, member A1 1717 DHCR7 7-Dehydrocholesterol 1.6 0.004 1.3 0.078 1.2 0.045 1.3 0.024 reductase 2023 ENO1 Enolase 1 (a) 1.7 0.004 1.2 0.001 1.2 0.005 1.5 0.005 2194 FASN Fatty acid synthase 2.9 0.000 1.4 0.011 1.2 0.023 1.7 0.005 65220 FLJ13052 NAD kinase 2.2 0.007 1.4 0.113 1.3 0.176 1.7 0.010 51706 NQO3A2 NAD(P)H:quinone 2.3 0.000 1.9 0.001 1.3 0.145 1.7 0.030 oxidoreductase type 3, polypeptide A2

(Continued on the following page)

www.aacrjournals.org 11683 Cancer Res 2005; 65: (24). December 15, 2005

Table 2. Transcriptional responses of Ramos cells cotreated with MGd and zinc (Cont’d)

Category Gene ID* Symbol Gene 10 Amol/L 10 Amol/L 10 Amol/L 50 Amol/L description MGd + MGd + MGd zinc 50 Amol/L zinc 25 Amol/L zinc

c c c c FC P FC P FC P FC P

5209 PFKFB3 6-Phosphofructo-2-kinase/ 2.0 0.004 1.7 0.016 1.6 0.000 1.5 0.014 fructose-2,6-biphosphatase 3 5210 PFKFB4 6-Phosphofructo-2-kinase/ 2.0 0.044 2.4 0.003 1.4 0.085 1.8 0.057 fructose-2,6-biphosphatase 4 5230 PGK1 Phosphoglycerate kinase 1 5.6 0.007 2.0 0.001 1.7 0.018 2.4 0.000 389753 — Similar to 4.0 0.000 2.1 0.005 1.8 0.018 3.3 0.001 phosphoglucomutase 5 Receptors and ligands 8772 FADD Fas (TNFRSF6)-associated 1.6 0.001 1.1 0.278 1.2 0.053 1.4 0.023 via death domain 55179 FAIM Fas apoptotic 1.7 0.001 1.3 0.010 1.1 0.102 1.4 0.006 inhibitory molecule 25816 TNFAIP8 TNF, a-induced protein 8 1.6 0.000 1.4 0.001 1.2 0.007 1.1 0.060 8795 TNFRSF10B TNF receptor superfamily, 1.6 0.003 1.0 1.000 1.0 1.000 1.0 0.423 member 10b 608 TNFRSF17 TNF receptor superfamily, 2.2 0.003 1.5 0.014 1.3 0.018 1.4 0.005 member 17 7132 TNFRSF1A TNF receptor superfamily, 2.3 0.004 1.6 0.003 1.2 0.120 1.0 0.934 member 1A 944 TNFSF8 TNF (ligand) superfamily, 1.9 0.001 1.2 0.042 1.2 0.042 1.6 0.001 member 8 Apoptosis related 332 BIRC5 Baculoviral IAP repeat- 1.8 0.000 1.3 0.058 1.2 0.326 1.6 0.044 containing 5 (survivin) 51651 BIT1 Bcl-2 inhibitor of transcription 2.6 0.000 1.4 0.044 1.2 0.004 1.2 0.010 664 BNIP3 BCL2/adenovirus 2.1 0.004 1.9 0.003 1.4 0.002 1.6 0.015 E1B interacting protein 3 665 BNIP3L BCL2/adenovirus 1.7 0.000 1.7 0.003 1.4 0.022 1.5 0.001 E1B interacting protein 3-like 836 CASP3 Caspase 3, apoptosis-related 1.6 0.000 1.4 0.072 1.2 0.150 1.2 0.131 cysteine protease 1613 DAPK3 Death-associated 1.9 0.000 1.1 0.059 1.1 0.072 1.9 0.001 protein kinase 3 162989 DEDD2 Death effector domain 4.2 0.003 2.1 0.004 1.3 0.007 1.2 0.036 containing 2 1677 DFFB DNA fragmentation factor, 1.6 0.002 1.1 0.507 1.1 0.279 1.2 0.013 b polypeptide 26355 E2IG5 Growth and transformation– 2.0 0.001 2.1 0.000 1.5 0.000 1.4 0.001 dependent protein 29923 HIG2 Hypoxia-inducible protein 2 1.6 0.005 1.6 0.009 1.2 0.164 1.1 0.244 8739 HRK Harakiri, BCL2 3.1 0.002 1.6 0.099 1.3 0.217 1.4 0.142 interacting protein 4615 MYD88 Myeloid differentiation 1.8 0.000 1.3 0.007 1.1 0.011 1.3 0.002 primary response gene 11188 NISCH Nischarin 1.5 0.003 1.1 0.115 1.2 0.096 1.3 0.016 22984 PDCD11 Programmed cell death 11 1.8 0.002 1.7 0.013 1.5 0.011 1.5 0.010 27250 PDCD4 Programmed cell death 4 2.4 0.000 1.5 0.022 1.2 0.137 1.8 0.005 9141 PDCD5 Programmed cell death 5 1.5 0.001 1.3 0.111 1.1 0.669 1.2 0.232 5366 PMAIP1 Phorbol-12-myristate-13- 2.7 0.000 1.8 0.004 1.3 0.068 1.6 0.009 acetate–induced protein 79155 TNIP2 TNFAIP3 interacting protein 2 1.6 0.002 1.5 0.013 1.3 0.003 1.5 0.000 Cell cycle control 472 ATM Ataxia telangiectasia mutated 1.7 0.000 1.0 0.450 1.1 0.046 1.6 0.103 545 ATR Ataxia telangiectasia and 1.5 0.005 1.3 0.025 1.1 0.011 1.2 0.015 Rad3 related

(Continued on the following page)

Cancer Res 2005; 65: (24). December 15, 2005 11684 www.aacrjournals.org

Table 2. Transcriptional responses of Ramos cells cotreated with MGd and zinc (Cont’d)

Category Gene ID* Symbol Gene 10 Amol/L 10 Amol/L 10 Amol/L 50 Amol/L description MGd + MGd + MGd zinc 50 Amol/L zinc 25 Amol/L zinc

c c c c FC P FC P FC P FC P

54619 CCNJ Cyclin J 2.0 0.003 1.5 0.002 1.2 0.225 1.3 0.124 8881 CDC16 CDC16 cell division cycle 1.8 0.001 1.3 0.006 1.2 0.017 1.4 0.004 16 homologue 995 CDC25C Cell division cycle 25C 2.0 0.000 1.7 0.032 1.2 0.018 1.4 0.083 996 CDC27 Cell division cycle 27 1.7 0.002 1.1 0.548 1.2 0.053 1.7 0.003 988 CDC5L CDC5 cell division cycle 1.7 0.001 1.3 0.008 1.1 0.083 1.3 0.013 5-like (S. pombe) 157313 CDCA2 Cell division cycle associated 2 1.6 0.005 1.2 0.029 1.1 0.066 1.3 0.009 23097 CDK11 Cyclin-dependent kinase (CDC2-like) 11 2.0 0.003 1.7 0.012 1.3 0.090 1.8 0.014 1017 CDK2 Cyclin-dependent kinase 2 3.3 0.002 1.2 0.306 1.4 0.122 2.9 0.000 1027 CDKN1B Cyclin-dependent kinase 2.4 0.001 1.9 0.000 1.3 0.008 1.3 0.035 inhibitor 1B (p27, Kip1) 1869 E2F1 E2F transcription factor 1 1.6 0.001 1.5 0.003 1.1 0.125 1.2 0.139 1871 E2F3 E2F transcription factor 3 1.6 0.003 1.3 0.027 1.1 0.325 1.2 0.054 4193 MDM2 Mdm2, transformed 3T3 2.6 0.003 1.3 0.131 1.1 0.423 2.2 0.002 cell double minute 2 5591 PRKDC Protein kinase, 1.5 0.002 1.1 0.480 1.1 0.117 1.5 0.002 DNA-activated, catalytic 5925 RB1 Retinoblastoma 1 1.7 0.002 1.8 0.000 1.2 0.008 1.4 0.005 (including osteosarcoma) Signaling 1649 DDIT3 DNA-damage-inducible transcript 3 6.6 0.000 1.7 0.014 1.2 0.337 1.1 0.337 54541 DDIT4 DNA-damage-inducible transcript 4 30.1 0.001 13.6 0.001 5.2 0.000 3.7 0.006 54583 EGLN1 Egl nine homologue 1 3.3 0.000 2.9 0.004 1.8 0.000 2.0 0.001 8518 IKBKAP IKB, kinase-associated protein 1.6 0.005 1.3 0.005 1.2 0.022 1.3 0.026 9641 IKBKE IKB, kinase epsilon 1.8 0.001 1.6 0.031 1.3 0.029 1.6 0.030 359948 IRF2BP2 IFN regulatory factor 2 2.7 0.001 1.6 0.023 1.2 0.073 1.7 0.004 binding protein 2 55233 MOBKL1B MOB1, Mps one binder 3.4 0.001 1.1 0.403 1.3 0.104 3.6 0.012 kinase activator-like 1B 4763 NF1 Neurofibromin 1 1.7 0.000 1.4 0.001 1.2 0.154 1.4 0.001 9975 NR1D2 Nuclear receptor subfamily 1, 3.7 0.001 1.2 0.163 1.2 0.157 2.8 0.000 group D, member 2 8554 PIAS1 Protein inhibitor 2.9 0.000 1.8 0.005 1.3 0.124 3.2 0.013 of activated STAT, 1 5567 PRKACB Protein kinase, 3.0 0.003 1.7 0.001 1.4 0.020 2.0 0.015 cAMP-dependent, catalytic, b 5580 PRKCD Protein kinase C, d 3.0 0.002 1.3 0.064 1.0 0.647 1.7 0.015 25865 PRKD2 Protein kinase D2 1.8 0.005 1.5 0.006 1.4 0.010 1.7 0.002 117289 TAGAP T-cell activation 2.9 0.001 1.8 0.066 1.3 0.159 1.3 0.198 GTPase activating protein 57761 TRB3 Tribbles homologue 3 3.9 0.003 1.2 0.020 1.1 0.445 1.5 0.204 7422 VEGF Vascular endothelial 2.2 0.000 1.0 0.838 1.3 0.035 1.6 0.003 growth factor Relevant to the function 467 ATF3 Activating transcription factor 3 3.9 0.005 1.0 1.000 1.0 1.000 1.1 0.423 of transcriptions factors 571 BACH1 BTB and CNC homology 1 3.3 0.004 1.8 0.036 1.0 0.567 1.4 0.032 865 CBFB Core-binding factor, b subunit 2.2 0.003 1.4 0.120 1.2 0.286 2.1 0.006 8452 CUL3 Cullin 3 1.7 0.000 1.2 0.001 1.0 0.088 1.1 0.007 2309 FOXO3A Forkhead box O3A 2.0 0.001 1.4 0.007 1.2 0.072 1.5 0.004 26959 HBP1 HMG-box transcription factor 1 3.0 0.001 2.3 0.002 1.5 0.031 1.4 0.029 204851 HIPK1 Homeodomain 3.1 0.000 1.2 0.205 1.1 0.437 2.8 0.000 interacting protein kinase 1 3398 ID2 Inhibitor of DNA binding 2 2.8 0.000 1.6 0.002 1.1 0.119 1.2 0.047 3727 JUND Jun D proto-oncogene 2.0 0.004 1.3 0.039 1.1 0.042 1.0 0.553

(Continued on the following page)

www.aacrjournals.org 11685 Cancer Res 2005; 65: (24). December 15, 2005

Table 2. Transcriptional responses of Ramos cells cotreated with MGd and zinc (Cont’d)

Category Gene ID* Symbol Gene 10 Amol/L 10 Amol/L 10 Amol/L 50 Amol/L description MGd + MGd + MGd zinc 50 Amol/L zinc 25 Amol/L zinc

c c c c FC P FC P FC P FC P

57157 PHTF2 Putative homeodomain 3.0 0.003 1.2 0.410 1.3 0.117 3.2 0.004 transcription factor 2 5971 RELB v-Rel reticuloendotheliosis 1.6 0.002 1.3 0.061 1.2 0.096 1.2 0.148 viral oncogene B 6596 SMARCA3 SWI/SNF–related, 1.5 0.000 1.1 0.016 1.0 0.471 1.2 0.003 regulator of chromatin, a3 6597 SMARCA4 SWI/SNF–related, 1.8 0.000 1.1 0.126 1.1 0.070 1.4 0.012 regulator of chromatin, a4 8467 SMARCA5 SWI/SNF–related, 1.6 0.000 1.3 0.001 1.1 0.129 1.2 0.002 regulator of chromatin, a5 7428 VHL von Hippel-Lindau 2.0 0.003 1.1 0.246 1.1 0.337 1.8 0.002 tumor suppressor

*Entries with identical National Center for Biotechnology Information gene ID designations represent results from different probe tilings interrogating the same gene. cFold change.

contrast to the intracellular free zinc levels described above, and zinc preceded mitochondrial dysfunction and early events of reactive oxygen species decreased over the course of the 24-hour apoptosis and thus were not a consequence of them. Furthermore, treatment. However, these data could be strongly influenced by increased intracellular free zinc and oxidative stress roughly apoptosis-related events or alterations in cellular metabolism that correlate with cell death, with Ramos the most sensitive line, occur at later times. followed by DHL-4 and the others (Fig. 3C). However, intracellular In keeping with their differences in intracellular free zinc, four free zinc levels seemed to be better predictors of proliferative and other B-cell lines displayed variable degrees of oxidative stress after apoptotic response. K562, HL60, and Jurkat lines did not exhibit 4 hours of cotreatment with MGd and zinc (Fig. 3B). Interestingly, changes in oxidative stress, intracellular free zinc, or apoptosis treatment of Ramos cultures with hydrogen peroxide also led to under these conditions (data not shown). transient increases in oxidative stress and sustained increases in To better understand the molecular changes accompanying loss intracellular free zinc (see Supplementary Data). This is consistent of zinc homeostasis before apoptosis, we examined the effect of with observations that zinc can induce oxidative stress in cultured 4-hour treatment with MGd and/or zinc on gene expression in mammalian cells, and, conversely, that thiol oxidation can mobilize Ramos cultures. There was a striking overlap in transcriptional zinc (35–37). responses to 10 Amol/L MGd or 50 Amol/L zinc (Table 1). In addition, we sought to relate changes in intracellular free zinc Depending on the stringency of our criteria, up to 97% of MGd- and oxidative stress to cellular growth rate. We observed a large responsive genes were also differentially expressed in the same reduction in the number of Ramos cells actively synthesizing direction in cells treated with 50 Amol/L zinc. This indicates that DNA in S-phase after treatment with MGd and 50 Amol/L zinc by MGd acts as a ‘‘zinc mimetic’’ with regard to the transcriptional 4 hours (Fig. 2). Other cell lines tested also displayed decreased responses induced in Ramos cells. DNA synthesis under these conditions (see Supplementary Data). Treatment with MGd or zinc both resulted in a strong and Interestingly, treatment with 50 to 100 Amol/L zinc inhibited the sustained induction of MTF-1-regulated metallothionein and zinc proliferation of four additional B-cell lines, an acute myelogenous transporter 1 (ZnT1) genes, which play major roles in regulating leukemia line (K562), and a T-cell lymphoma line (Jurkat), but not intracellular free zinc levels, as well as HIF-1-regulated genes in an acute promyelocytic leukemia line (HL60; see Supplementary (e.g., PFKB3, DDIT4, and EGLN1; Table 1). PFKFB3 is a kinase/ Data). In all lines except HL60, MGd cotreatment potentiated the phosphatase that modulates the concentration of fructose-2,6- inhibition by zinc. The effect of MGd and zinc differed strikingly bisphosphate, a key modulator of the glycolytic rate in proliferating from that of 5-fluoro-2V-deoxyuridine or ionizing radiation, both of cells (38). DDIT4 is a proapoptotic protein recently reported to be which permitted BrdUrd incorporation into DNA and changed cell a negative regulator of the mammalian target of rapamycin cycle distribution with accumulation of cells in G1-S and G2-M, pathway (39, 40). EGLN1 (also known as PHD2) is a prolyl respectively. It also differed from hydroxyurea, which inhibited hydroxylase that plays a key role in regulating HIF-1 activity BrdUrd incorporation but allowed passage through G2M. This by targeting HIF-1a for ubiquitin-mediated degradation (41). suggests that increased intracellular free zinc inhibits proliferation Although it is not as effective in this regard as the better known at multiple checkpoints. ‘‘hypoxia-mimetic’’ cobalt, zinc can inhibit the activity of HIF- Importantly, we observed that the increased oxidative stress and associated hydroxylases by displacing iron from the active site intracellular free zinc levels induced by cotreatments with MGd of these enzymes (42, 43). Indeed, we measured greater cellular

Cancer Res 2005; 65: (24). December 15, 2005 11686 www.aacrjournals.org

HIF-1a levels by ELISA in Ramos cultures treated with either zinc residues shown to be modified under oxidative stress conditions or MGd (Fig. 4C). Thus, we propose that MGd induces hypoxia- (48, 49). It has been proposed that induction of thioredoxin mimetic transcriptional responses in this system as a result of reductase (TXNRD1) and increased glutathione levels serves to HIF-1 stabilization due to increased intracellular free zinc and/or restore Keap-2 binding of NRF-2 as part of a feedback loop (48). generation of reactive oxygen species. Induction of NRF-2 response genes could therefore reflect the It is notable that levels of transcripts under the control of MTF-1 altered redox state of the cells under conditions where this (e.g., metallothionein family members) and under the control of enzyme is inhibited. HIF-1 (e.g., DDIT4) are increased synergistically in some instances Overall, our data indicate that the primary effect of mode- by MGd and zinc treatment (Tables 1 and 2). These changes could rately increased free zinc in Ramos cells is the activation of contribute to the observed biological effects of the combined MTF-1 and HIF-1. Induction of free zinc at higher levels treatment. Indeed, the increased activation of HIF-1 would be increases oxidative stress, leading to the activation of NRF-2. expected to alter cellular metabolism to favor glycolysis over Paradoxically, it is conceivable that the activation of transcrip- oxidative phosphorylation via the induction of transcripts, such as tional cascades under the control of MTF-1, HIF-1, and NRF-2 PFKFB3 and PGK1 (38). HIF-1 is often considered to be essential by MGd could be protective toward reactive oxygen species in for tumor growth and, indeed, its inhibition is the subject of some circumstances (i.e., through ‘‘adaptive resistance’’ mecha- ongoing drug development activities (44). However, under the nisms and the induction of ‘‘cell life’’ transcripts; refs. 50, 51). appropriate conditions, HIF-1 activation can have negative Cells already under oxidative stress, on the other hand, would consequences for tumor growth by induction of targets linked to be expected to display increased sensitivity to MGd as a apoptosis, such as BNIP3, E2IG5, PMAIP1, and DDIT4, or through consequence of preexisting elevated levels of intracellular free metabolic alteration of cells in the low nutrient context of the zinc, depleted stores of reducing equivalents, and dependence tumor microenvironment (45). on key antioxidant enzymes, such as thioredoxin reductase (52). In addition to the MTF-1- and HIF-1-regulated transcripts If so, this would imply that a considerable degree of selectivity discussed above, cotreatment with MGd and zinc resulted in the might be achieved in targeting this drug to diseases involving expression of NRF-2-regulated transcripts, such as GCLM, HMOX1, oxidative stress. and NQO3A2, which all have antioxidant response elements in their promoters (46, 47). Additional transcripts, such as TXNRD1, Acknowledgments CTH, GSR, and a variety of transporters (e.g., SLC7A11) pre- Received 8/4/2005; revised 9/14/2005; accepted 9/30/2005. sumably involved in cellular uptake of amino acids required for Grant support: V Foundation for Cancer Research, National Institute of glutathione synthesis, are also induced. The induction of NRF-2 Environmental Health Sciences grant P30-ES07048, and a grant from Pharmacyclics activity may be related to its nuclear translocation following (J.G. Hacia). The costs of publication of this article were defrayed in part by the payment of page disruption of the cytoplasmic Keap-1–NRF-2 complex. The charges. This article must therefore be hereby marked advertisement in accordance capacity of Keap-1 to bind NRF-2 is regulated by critical cysteine with 18 U.S.C. Section 1734 solely to indicate this fact.

References 10. Butcher HL, Kennette WA, Collins O, Zalups RK, oxygen species. Int J Radiat Oncol Biol Phys 2001;51: Koropatnick J. Metallothionein mediates the level and 1025–36. 1. Costello LC, Franklin RB. Novel role of zinc in activity of nuclear factor nB in murine fibroblasts. 20. Evens AM, Lecane P, Magda D, et al. Motexafin the regulation of prostate citrate metabolism and J Pharmacol Exp Ther 2004;310:589–98. gadolinium generates reactive oxygen species and its implications in prostate cancer. Prostate 1998;35: 11. Kim CH, Kim JH, Moon SJ, et al. Pyrithione, a zinc induces apoptosis in sensitive and highly resistant 285–96. ionophore, inhibits NF-nB activation. Biochem Biophys multiple myeloma cells. Blood 2005;105:1265–73. 2. Dineley KE, Votyakova TV, Reynolds IJ. Zinc inhibition Res Commun 1999;259:505–9. 21. Eray M, Tuomikoski T, Wu H, et al. Cross-linking of of cellular energy production: implications for mito- 12. Hainaut P, Mann K. Zinc binding and redox control surface IgG induces apoptosis in a bcl-2 expressing chondria and neurodegeneration. J Neurochem 2003;85: of p53 structure and function. Antioxid Redox Signal human follicular lymphoma line of mature B cell 563–70. 2001;3:611–23. phenotype. Int Immunol 1994;6:1817–27. 3. Sheline CT, Behrens MM, Choi DW. Zinc-induced 13. O’Halloran TV. Transition metals in control of gene 22. Epstein AL, Levy R, Kim H, Henle W, Henle G, Kaplan cortical neuronal death: contribution of energy failure expression. Science 1993;261:715–25. HS. Biology of the human malignant lymphomas. IV. attributable to loss of NAD(+) and inhibition of 14. Haase H, Maret W. Intracellular zinc fluctuations Functional characterization of ten diffuse histiocytic glycolysis. J Neurosci 2000;20:3139–46. modulate protein tyrosine phosphatase activity in lymphoma cell lines. Cancer 1978;42:2379–91. 4. Lenartowicz E, Wudarczyk J. Enzymatic reduction of insulin/insulin-like growth factor-1 signaling. Exp Cell 23. Mosmann T. Rapid colorimetric assay for cellular 5,5V-dithiobis-(2-nitrobenzoic acid) by lysate of rat liver Res 2003;291:289–98. growth and survival: application to proliferation and mitochondria. Int J Biochem Cell Biol 1995;27:831–7. 15. Fraker PJ, Telford WG. A reappraisal of the role of cytotoxicity assays. J Immunol Methods 1983;65:55–63. 5. Rigobello MP, Callegaro MT, Barzon E, Benetti M, zinc in life and death decisions of cells. Proc Soc Exp 24. Gee KR, Zhou ZL, Ton-That D, Sensi SL, Weiss JH. Bindoli A. Purification of mitochondrial thioredoxin BiolMed 1997;215:229–36. Measuring zinc in living cells. A new generation of reductase and its involvement in the redox regulation of 16. Mehta MP, Rodrigus P, Terhaard CH, et al. Survival sensitive and selective fluorescent probes. Cell Calcium membrane permeability. Free Radic Biol Med 1998;24: and neurologic outcomes in a randomized trial of 2002;31:245–51. 370–6. motexafin gadolinium and whole-brain radiation ther- 25. Gratzner HG, Leif RC. An immunofluorescence 6. Magda D, Lecane P, Miller RA, et al. Motexafin apy in brain metastases. J Clin Oncol 2003;21:2529–36. method for monitoring DNA synthesis by flow cytom- gadolinium disrupts zinc metabolism in human cancer 17. Biaglow JE, Miller RA, Magda D, Lee I, Tuttle S. etry. Cytometry 1981;1:385–93. cell lines. Cancer Res 2005;65:3837–45. Inhibition of thioredoxin reductase by motexafin 26. Naumovski L, Ramos J, Sirisawad M, et al. Sapphyrins 7. Holmgren A. Thioredoxin. Annu Rev Biochem 1985;54: gadolinium: effects on PLDR. Prog Abst Radiat Res Soc induce apoptosis in hematopoietic tumor-derived cell 237–71. 2002;107. lines and show in vivo antitumor activity. Mol Cancer 8. Forbes IJ, Zalewski PD, Giannakis C. Role for zinc in a 18. Magda D, Gerasimchuk N, Lecane P, Miller RA, Ther 2005;4:968–76. cellular response mediated by protein kinase C in Biaglow JE, Sessler JL. Motexafin gadolinium reacts with 27. Mizzen CA, Cartel NJ, Yu WH, Fraser PE, human B lymphocytes. Exp Cell Res 1991;195:224–9. ascorbate to produce reactive oxygen species. Chem McLachlan DR. Sensitive detection of metallothio- 9. Csermely P, Szamel M, Resch K, Somogyi J. Zinc can Commun (Camb) 2002;2730–1. neins-1, -2 and -3 in tissue homogenates by immuno- increase the activity of protein kinase C and contributes 19. Magda D, Lepp C, Gerasimchuk N, et al. Redox blotting: a method for enhanced membrane transfer to its binding to plasma membranes in T lymphocytes. cycling by motexafin gadolinium enhances cellular and retention. J Biochem Biophys Methods 1996;32: J Biol Chem 1988;263:6487–90. response to ionizing radiation by forming reactive 77–83. www.aacrjournals.org 11687 Cancer Res 2005; 65: (24). December 15, 2005

28. Karaman MW, Houck ML, Chemnick LG, et al. 38. Chesney J, Mitchell R, Benigni F, et al. An inducible cysteine ligase modifier subunit gene. Revision of the Comparative analysis of gene-expression patterns in gene product for 6-phosphofructo-2-kinase with an AU- ARE consensus sequence. J Biol Chem 2002;277: human and African great ape cultured fibroblasts. rich instability element: role in tumor cell glycolysis and 30730–7. Genome Res 2003;13:1619–30. the Warburg effect. Proc Natl Acad Sci U S A 1999;96: 47. Nioi P, McMahon M, Itoh K, Yamamoto M, 29. Tusher VG, Tibshirani R, Chu G. Significance analysis 3047–52. Hayes JD. Identification of a novel Nrf2-regulated of microarrays applied to the ionizing radiation 39. Shoshani T, Faerman A, Mett I, et al. Identification of antioxidant response element (ARE) in the mouse response. Proc Natl Acad Sci U S A 2001;98:5116–21. a novel hypoxia-inducible factor 1-responsive gene, NAD(P)H:quinone oxidoreductase 1 gene: reassess- 30. Giedroc DP, Chen X, Apuy JL. Metal response element RTP801, involved in apoptosis. Mol Cell Biol 2002;22: ment of the ARE consensus sequence. Biochem J 2003; (MRE)-binding transcription factor-1 (MTF-1): struc- 2283–93. 374:337–48. ture, function, and regulation. Antioxid Redox Signal 40. Corradetti MN, Inoki K, Guan KL. The stress- 48. Dinkova-Kostova AT, Holtzclaw WD, Cole RN, et al. 2001;3:577–96. inducted proteins RTP801 and RTP801L are negative Direct evidence that sulfhydryl groups of Keap1 are the 31. Lichtlen P, Wang Y, Belser T, et al. Target gene search regulators of the mammalian target of rapamycin sensors regulating induction of phase 2 enzymes that for the metal-responsive transcription factor MTF-1. pathway. J Biol Chem 2005;280:9769–72. protect against carcinogens and oxidants. Proc Natl Nucleic Acids Res 2001;29:1514–23. 41. Freeman RS, Hasbani DM, Lipscomb EA, Straub JA, Acad Sci U S A 2002;99:11908–13. 32. Lichtlen P, Schaffner W. Putting its fingers on Xie L. SM-20, EGL-9, and the EGLN family of hypoxia- 49. Wakabayashi N, Dinkova-Kostova AT, Holtzclaw stressful situations: the heavy metal-regulatory tran- inducible factor prolyl hydroxylases. Mol Cells 2003;16: WD, et al. Protection against electrophile and oxidant scription factor MTF-1. Bioessays 2001;23:1010–7. 1–12. stress by induction of the phase 2 response: fate of 33. Andrews GK. Cellular zinc sensors: MTF-1 regulation 42. Schofield CJ, Ratcliffe PJ. Oxygen sensing by HIF cysteines of the Keap1 sensor modified by inducers. of gene expression. Biometals 2001;14:223–37. hydroxylases. Nat Rev Mol Cell Biol 2004;5:343–54. Proc Natl Acad Sci U S A 2004;101:2040–5. 34. Choi DW, Koh JY. Zinc and brain injury. Annu Rev 43. Hirsila M, Koivunen P, Xu L, Seeley T, Kivirikko KI, 50. Vargas MR, Pehar M, Cassina P, et al. Fibroblast Neurosci 1998;21:347–75. Myllyharju J. Effect of desferrioxamine and metals on growth factor-1 induces heme oxygenase-1 via nuclear 35. Seo SR, Chong SA, Lee SI, et al. Zn2+-induced ERK the hydroxylases in the oxygen sensing pathway. FASEB factor erythroid 2-related factor 2 (Nrf2) in spinal activation mediated by reactive oxygen species causes J 2005;19:1308–10. cord astrocytes: Consequences for motor neuron cell death in differentiated PC12 cells. J Neurochem 44. Semenza GL. Targeting HIF-1 for cancer therapy. Nat survival. J Biol Chem 2005;280:25571–9. 2001;78:600–10. Rev Cancer 2003;3:721–32. 51. Lee JM, Calkins MJ, Chan K, Kan YW, Johnson 36. Malaiyandi LM, Dineley KE, Reynolds IJ. Divergent 45. Lee MJ, Kim JY, Suk K, Park JH. Identification of the JA. Identification of the NF-E2-related factor-2- consequences arise from metallothionein overexpres- hypoxia-inducible factor 1 a-responsive HGTD-P gene dependent genes conferring protection against oxi- sion in astrocytes: zinc buffering and oxidant-induced as a mediator in the mitochondrial apoptotic pathway. dative stress in primary cortical astrocytes using zinc release. Glia 2004;45:346–53. Mol Cell Biol 2004;24:3918–27. oligonucleotide microarray analysis. J Biol Chem 2003; 37. Maret W, Vallee BL. Thiolate ligands in metal- 46. Erickson AM, Nevarea Z, Gipp JJ, Mulcahy RT. 278:12029–38. lothionein confer redox activity on zinc clusters. Proc Identification of a variant antioxidant response 52. Miller RA, Biaglow JE. The thioredoxin reductase/ Natl Acad Sci U S A 1998;95:3478–82. element in the promoter of the human glutamate- thioredoxin system. Cancer Biol Ther 2005;4:6–13.

Cancer Res 2005; 65: (24). December 15, 2005 11688 www.aacrjournals.org

Downloaded from cancerres.aacrjournals.org on September 26, 2021. © 2005 American Association for Cancer Research. Motexafin Gadolinium and Zinc Induce Oxidative Stress Responses and Apoptosis in B-Cell Lymphoma Lines

Philip S. Lecane, Mazen W. Karaman, Mint Sirisawad, et al.

Cancer Res 2005;65:11676-11688.

Updated version Access the most recent version of this article at: http://cancerres.aacrjournals.org/content/65/24/11676

Cited articles This article cites 50 articles, 20 of which you can access for free at: http://cancerres.aacrjournals.org/content/65/24/11676.full#ref-list-1

Citing articles This article has been cited by 9 HighWire-hosted articles. Access the articles at: http://cancerres.aacrjournals.org/content/65/24/11676.full#related-urls

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://cancerres.aacrjournals.org/content/65/24/11676. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.