Zebularine partially reverses GST methylation in prostate cancer cells and restores sensitivity to the DNA minor groove binder brostallicin

Abstract

Brostallicin is a DNA minor groove binder that shows enhanced antitumor activity in cells with high glutathione S-transferase (GST)/glutathione content. Prostate cancer cells present, almost invariably, methylation of the GSTP1 gene promoter and, as a consequence, low levels of GST-pi expression and activity. In these cells, brostallicin shows very little activity. We tested whether pretreatment of heavily GST-methylated prostate cancer cells with demethylating agents could enhance the activity of brostallicin. Human prostate cancer cells LNCaP and DU145 were used for these studies both in vitro and in vivo. The demethylating agent zebularine was used in combination with brostallicin. Methylation specific PCR and pyrosequencing were used to determine the level of GST methylation. Pretreatment with demethylating agents enhanced the in vitro activity of brostallicin in LNCaP cells. Zebularine, in particular, induced an enhancement of activity in vivo comparable to that obtained by transfecting the human GSTP1 gene in LNCaP cells in vitro. Molecular analysis performed on tumor xenografts in mice pretreated with zebularine failed to detect re-expression of GST-pi and demethylation of GSTP1. However, we found demethylation in the GSTM1 gene, with consequent re-expression of GST-mu at the mRNA level. These results indicate that zebularine, both in vitro and in vivo, enhances the activity of brostallicin and that this enhancement correlates with re-expression of GST-pi and GST-mu. These findings highlight the potential therapeutic value of combining demethylating agents and brostallicin in tumors with GST methylation that poorly respond to brostallicin.

Keywords: :

• DNA methylation
• prostate cancer
• combination chemotherapy
• reversal of drug resistance
• DNA reactive agents

Introduction

Prostate cancer represents one of the tumors most investigated due to the aberrant methylation patterns acquired during its development, which lead to global hypomethylation of the genome and hypermethylation in the promoter regions of several genes belonging to different pathways.^1,2 In particular, GSTP1 promoter methylation is clearly detected in a very high percentage (roughly 90%) of clinical samples and many studies have investigated the possibility to exploit it as an early tumor marker.^3-5

GST-pi belongs to the GST family of detoxifying enzymes; this detoxification function is usually increased as a resistance mechanism of cancer cells against some chemotherapeutics (such as cisplatin and its derivative, melphalan).^6-9 In the past, this characteristic has prompted investigators to specifically design anticancer molecules, which could be intracellularly activated by GSH binding, catalyzed by the GST enzyme.

Brostallicin belongs to this class of anticancer drugs: it is a synthetic α-bromoacrylic derivative of dystamicin A, binding the AT rich regions in the minor groove of the DNA.^10-14 The ability to alkylate these DNA sequences is a consequence of the intracellular GSH binding to the α-bromoacrylic group of brostallicin through a nucleophilic attack. The specificity of brostallicin cytotoxicity for cells with a high GST/GSH content has been clearly demonstrated in vitro and in vivo.^15-17 This kind of activation makes brostallicin a good candidate for combination therapy with those anticancer agents whose mechanism of resistance depends on increased levels of the GSH/GST system.

Pre-clinical evaluation of brostallicin with cisplatin in the colorectal cancer model HCT 116 has demonstrated that brostallicin administration after cisplatin results in a significantly improved antitumor activity, if compared with the activity of the two agents administrated alone, and depends on the increased level of GST enzymatic activity after treatment with cisplatin.^11,17

One of the new therapeutic approaches against solid tumors is the combination of molecules specifically reverting the aberrant DNA methylation in cancer cells (through drugs inhibiting DNA methyltransferase activity) and cytotoxic drugs.^18,19

This therapeutic purpose is based on the knowledge that tumors might epigenetically silence specific genes as a mechanism of resistance to therapy, as it has been well demonstrated for hMLH1 in ovarian cancer after treatment with cisplatin.^20,21 As a consequence, the re-expression of these genes could re-sensitize tumors to therapy.

Besides the oldest-used drugs 5-azaC and its deoxyderivative 5-aza-dC, whose clinical application was hampered by a strong hematological toxicity, in the last years the cytidine deaminase inhibitor zebularine has begun to be considered for its demethylating properties.^22-24 While the first attempts of zebularine use were in combination with 5-aza-dC on hematological tumors with the goal of inhibiting the cytidine deaminase enzyme responsible for 5-aza-dC inactivation, more recently it was shown that zebularine treatment of T24 bladder cancer cells was able to revert p16 methylation in vivo, and to induce p16 expression, determining a decrease of tumor growth rate.^25-28 At the moment, few studies have investigated the feasibility of zebularine as demethylating agent in a combination approach on solid tumors.^29-33

The prostate cancer cell line LNCaP represents a good model to study a possible new therapeutic approach for the early/metastatic prostate tumors because it maintains the strong methylation of the GSTP1 gene. In particular, this work was aimed at testing the capability of zebularine to induce re-expression of GST enzymes by demethylation of their gene promoter, in order to extend the antitumor activity of brostallicin in those tumors lacking GST enzymatic activity due to methylation of the gene.

Results

The activity of brostallicin was tested in two human prostate cancer cell lines, DU145 and LNCaP. DU145 cells are sensitive to the drug, while LNCaP were only moderately responsive, with a 20% reduction in cell number at the highest concentration utilized (Fig. 1A).

Figure 1. (A) Cytotoxicity induced by increasing concentrations of brostallicin treatment in LNCaP (◻) and DU145 (●) cells. Results are reported as the percentage of inhibition relative to controls and are the mean ± SD of at least six replicates. °p < 0.05, *p < 0.01 vs LNCaP (B) GST total activity (left panel) and GST-pi protein expression (right panel) in LNCaP and DU145 cells. Actin is used as loading control in western blotting experiments. *p < 0.01 vs LNCaP (C) Pyrosequencing results showing the average of methylation percentage of 14 different CpG islands (corresponding with Y in the sequence at the bottom of the graph) in the GSTP1 promoter sequence of LNCaP (gray bars) and DU145 (white bars) cells. Results are expressed as mean ± SD of three independent evaluations. (D) GST activity (left panel) and GST-pi protein expression (right panel) in parental LNCaP and in two clones transfected with the human GSTP1 gene. Actin is used as loading control in western blotting experiments. *p < 0.01 vs LNCaP wt (E) Activity of brostallicin in parental LNCaP cells (●) and in LNCaP-derived clones 41 (■) and 54 (◻) overexpressing GST-pi. °p < 0.05, *p < 0.01 vs wt

When the activity of GST was determined in the two cell lines, a clear difference was found (Fig. 1B, left panel). In fact, the sensitive DU145 cell line showed a much higher activity than the resistant LNCaP cells. Similarly, the level of GST-pi was well detectable in DU145 by western blotting, while it was almost undetectable in LNCaP cells (right panel). These differences in GST-pi expression and GST enzymatic activity were associated with a clear quantitative difference in the methylation percentage in the two cell lines by analyzing 14 CpG dinucleotides of GSTP1 promoter by pyrosequencing (Fig. 1C).

Based on these preliminary data, LNCaP cells were transfected with the human GSTP1 gene and two clones, clone 41 and clone 54, were selected for further studies. These clones showed an increased expression of GST-pi (as determined by western blotting, Fig. 1D, right panel) and an increased GST enzymatic activity (Fig. 1D, left panel). When tested for their response to brostallicin, both clones (41 and 54) showed an increased sensitivity compared with mock transfected cells (Fig. 1E).

To further associate GST expression with brostallicin activity in LNCaP cells, pretreatment with demethylating agents was performed. First, we analyzed the effect of pretreatment of LNCaP cells with the classical demethylating agent 5′-aza deoxycytidine (5-aza dC), which resulted in an increased antiproliferative activity of brostallicin (Fig. 2A). At the effective doses, 5-aza dC per se was too toxic, as shown in Figure 2B, to allow a proper estimation of the role of GST demethylation in determining the increased activity of brostallicin. Nevertheless, in these experimental conditions, 5-aza dC was able to induce an increase in the total GST activity in LNCaP cells (Fig. 2C).

Figure 2. (A) Activity of brostallicin in the absence (left) or presence (right) of 0.05 μM 5-aza dC. Cells were treated with different concentrations of brostallicin and colonies formed stained with comassie blue. (B) Effect of 0.05 μM 5-aza dC on the growth of LNCaP cells. Black bar, controls, gray bar, 5-aza dC treated cells. Results are reported as mean ± SD of at least six replicates. *p < 0.01 vs wt (C) GST activity in untreated (black bar) or 0.05 μM 5-aza dC-treated (gray bar) LNCaP cells. Results are the mean ± SD of at three replicates. *p < 0.01 vs wt

Further studies were performed with the demethylating agent zebularine, which has been reported to be less toxic than 5-aza dC. As it can be seen (Fig. 3A), zebularine increased the activity of brostallicin at concentrations in which an increased GST activity and a partial re-expression of GST was observed (Fig. 3B). Treatment with 100 or 125 micromolar zebularine induced the appearance of a band corresponding to the unmethylated form of GSTP1 and a slight increase in the ratio of unmethylated/methylated band for GSTM1 (from 0.31, in untreated cells, to 0.42, for the highest zebularine concentration) (Fig. 3C). At the same concentrations, zebularine induced an increase in the expression of GST-pi and GST-mu mRNA measured by real time PCR (Fig. 3D). At the concentrations used, zebularine alone did not cause any appreciable cytotoxicity.

Figure 3. (A) Cytotoxicity induced by increasing concentrations of brostallicin treatment in LNCaP cells in the absence (●) or in the presence of 100 (■) or 125 μM (◻) of zebularine. Results are reported as the percentage of inhibition relative to controls and are the mean ± SD of at least six replicates. *p < 0.01 vs cells not pretreated with zebularine. (B) GST activity in LNCaP cells treated with different concentrations of zebularine. Results are the mean ±− SD of at three replicates. *p < 0.01 vs cells not pretreated with zebularine. (C) Methylation specific–PCR in the GSTP1 (upper panel) and GSTM1 (lower panel) gene performed in LNCaP cells treated with zebularine. Lane 1 MW marker; Lanes 2 and 3 control cells methylated (M) and unmethylated (U); Lanes 4 and 5 100 μM zebularine treated cells, methylated (M) and unmethylated (U); Lanes 6 and 7 125 μM zebularine treated cells, methylated (M) and unmethylated (U); Lanes 8 and 9 internal controls (blanks) (D) Real time RT PCR performed in LNCaP cells untreated or treated with 100 or 125 mM zebularine. The table reports the expression ratio between treated and control cells (arbitrarily set to 1). The values haves been normalized for cyclophillin used as internal control.

When zebularine was combined with brostallicin in the drug-sensitive cell line DU145, zebularine did not increase the activity of brostallicin (Fig. 4A), suggesting that the enhancing effects observed in LNCaP cells were due to re-expression of GST. Similarly, zebularine did not further increase the activity of brostallicin in GST expressing LNCaP clones (Fig. 4B).

Figure 4. (A) Activity of brostallicin in the absence (●) or in the presence (◻) of 100 μM zebularine in DU145 cells. Results are reported as the percentage of inhibition relative to controls and are the mean ± SD of at least six replicates. (B) Activity of brostallicin in LNCaP-GST clone (clone 41) in the absence (●) or in the presence (◻) of 100 μM zebularine. Results are reported as the percentage of inhibition relative to controls and are the mean ± SD of at least six replicates. °p < 0.05 vs cells not pretreated with zebularine.

It has to be noted that, both in clones ectopically expressing GST-pi and in parental LNCaP cells pretreated with the demethylating agent zebularine, the increased levels of GST causes an increase of brostallicin antiproliferative activity by only 60%.

We then moved to an in vivo setting (a scheme of treatment schedule is reported in Fig. 5A) where brostallicin did not shown any antitumor activity in parental LNCaP xenografts at the tolerated dose of 0.4 mg/Kg. Conversely, some antitumor activity was observed in GST-expressing, LNCaP-derived, clone 41 (40% tumor inhibition, Fig. 5B).

Figure 5. (A) Scheme of the treatment schedules used for the in vivo experiments performed in LNCaP-GST (left panel) and in LNCaP wt (right panel) xenografts. (B) In vivo antitumor activity of brostallicin in LNCaP-GST clone 41 xenografts. Mice were treated with brostallicin 0.4 mg/kg i.v. twice on days 6 and 20 after randomization (▲). Untreated mice (black diamond) °p < 0.05, 2 d after the last treatment with brostallicin. (C) In vivo antitumor activity of brostallicin in parental LNCaP wt xenografts. Mice were treated with brostallicin 0.4mg/kg i.v. twice on days 6 and 20 after randomization (●), zebularine 500 mg/kg daily ip twice for 7 d (■), first treatment: day 0 until day 6; second treatment: from day14 until day 20, or with the combination of zebularine and brostallicin (▲). Untreated mice (black diamond). *p < 0.01, combination group vs control and brostallicin groups and °p < 0.05 vs zebularine group. The tables report parameters derived from the curves in the two experimental settings and described at the bottom of the figures and in Materials and Methods.

Similar antitumor activity was observed in parental LNCaP cells when brostallicin was administered following a pretreatment with zebularine (50% tumor inhibition, Fig. 5C) with an acceptable toxicity (approximately 20% weight loss). If we also consider parameters such as tumor growth delay (T-C) or days required to reach 200% of the initial tumor volume, a significant increased activity of the drug combination over the single drugs could be found.

To try to correlate the increased activity of the combination of zebularine and brostallicin with changes in GST levels, GST-pi expression was checked in tumors pretreated with zebularine in comparison with the untreated ones. As reported in Figure 6A by IHC, we failed to observe an increase in GST-pi expression following zebularine treatment. Similar data have been obtained by western blotting (data not shown). Accordingly, when the methylation status of GSTP1 was assessed in tumors of mice untreated or treated with zebularine, we did not see a decreased level of methylated bands after treatment with zebularine (Fig. 6C). Similarly, the expression of GST-pi mRNA was undetectable by real time PCR (Ct undetermined, that means >45), either in controls or in zebularine-treated tumors (Fig. 6D).

Figure 6. (A) Immunohistochemistry analysis of GST-pi expression in parental LNCaP cells untreated (LNCaPwt) or pretreated for 7 consecutive days with zebularine 500mg/Kg (Zebularine). Positive control for GST expression is represented by untreated LNCaP-GST clone 41 growing in immunodeficient mice (LNCaP-GST). The first two panels report H&E staining. (B) Predicted CpG islands in GSTP1 gene according to MethPrimer program. (C) Methylation specific–PCR in the GSTP1 gene performed in LNCaP tumors obtained from untreated or zebularine treated mice (tumors taken after the last treatment and at autopsy) as reported in the figure. Calponin is used as internal control. The three last lanes represent the controls of methylation specific PCRs obtained by using unmethylated DNA (0), totally methylated DNA (100) or an equal mixture of them (50). (D) Real Time RT PCR performed in LNCaP tumors untreated or treated with zebularine. The table reports the cycle in which amplification occurs for GSTP1 and for cyclophillin used as internal control. The Delta Ct values calculated from the data are reported in the last column.

It has been previously showed that in the human breast cancer cell line MCF-7, presenting methylation of GST and resistance to brostallicin treatment, transfection of GST-pi and GST-mu did restore sensitivity to the drug, while transfection of GST-α did not modify the response to the compound.¹⁶ These findings highlight the potential role of both GST-pi and GST-mu in brostallicin responsiveness. The GSTM1 gene also has a putative CpG island (Fig. 7A), and we indeed found that the GSTM1 gene is also methylated in LNCaP cells. We therefore checked in LNCaP tumors whether pretreatment of zebularine was able to alter the methylation of GSTM1 in tumors growing in mice. As it can be seen in Figure 7B, zebularine was indeed able to induce, although to a limited extent, demethylation of the GSTM1 promoter in two out of three tumors analyzed. This phenomenon was associated with a slight but detectable increase in GST-mu transcription. Analysis of mRNA expression by real time RT-PCR showed that GST-mu was undetectable in samples from untreated mice (Ct > 45, meaning undetermined) and become detectable, although at relatively low levels, in all tumors treated with zebularine analyzed, with a median Ct of 40.2 (Fig. 7C).

Figure 7. (A) Predicted CpG islands in GSTM1 gene according to MethPrimer program. (B) GSTM1 gene methylation specific–PCR performed in LNCaP tumors obtained from untreated or zebularine treated mice (tumors taken after the last treatment have been analyzed). M refers to methylated band and U to unmethylated amplified DNA. (C) Real Time RT PCR performed in LNCaP tumors untreated or treated with zebularine. The table reports the cycle in which amplification occurs for GSTM1 and for cyclophillin used as internal control. The Delta Ct values calculated from the data are reported in the last column.

Discussion

The minor groove DNA binder brostallicin, differently from classical alkylating agents, needs glutathione to exert its maximal antitumor activity.^11,15 Several cancer cells do express high levels of GST and are relatively resistant to classical alkylating agents, including cisplatin.^6,8 In these tumors, brostallicin has potential therapeutic application, retaining, at least in preclinical studies, its antitumor activity.^15,17

In prostate cancer, methylation of GST represents one of the initial steps of transformation resulting in low or absent expression of the protein.^3-5 In this situation, we have shown that brostallicin is likely to be ineffective. We have shown that, in a human prostate cancer cell line not presenting GST methylation, brostallicin is effective. In contrast, in a cell line with heavily methylated GST, LNCaP, the drug does not have any cytotoxic effect. The activity is restored upon transfection of the human GSTP1 gene, further corroborating the role of glutathione/GST in the mechanism of action of brostallicin. GST-pi expression partially restored the activity to brostallicin in LNCaP cells, suggesting that other mechanisms, along with GST-methylation, could account for resistance to the drug in this cell line that is resistant to several other anticancer agents.

Nevertheless, we used this cell line to prove that the use of demethylating agents can be useful in restoring the activity of brostallicin. Zebularine, in particular, has been shown to be able to demethylate several genes.^23,25,26,34 We confirmed the ability of zebularine to revert GST methylation in vitro in LNCaP cells. In addition, we showed that this finding was associated to increased mRNA expression of both GST-pi and GST-mu and to an increased response of LNCaP cells to brostallicin when combined with zebularine. The activity of brostallicin in LNCaP cells pretreated with zebularine was similar to the one obtained in two independent clones ectopically overexpressing GST-pi, although the level of GST were considerably high, as expected, in cells transfected with the gene. These data demonstrate that partial re-expression of GST is sufficient to restore the glutathione-mediated activity of brostallicin and, hence, increases the chances of the in vivo activity of the combination.

In mice bearing LNCaP tumors, zebularine was indeed able to increase the activity of brostallicin, which was ineffective when given as single agent. When we tried to correlate the induction of antitumor activity with GST-pi expression, we failed to detect demethylation of its promoter as well as any change in its expression by real time PCR, western blotting or IHC.

It has been shown that the reaction of brostallicin with glutathione is catalyzed by GST-pi and GST-mu, but not by GST-α.¹⁶ It is interesting to note that GST-mu has a high affinity binding site for brostallicin, even higher than that of GST-pi.¹⁶

GST-mu gene is also methylated in LNCaP cells, even if its methylation level is not as high as in GST-pi and hence could, at least in vivo, be a better candidate for zebularine action. Indeed, we showed that in mice pretreated with zebularine we could see, although to a limited extent, the appearance of an unmethylated band in genomic analysis and a re-expression of GST-mu at the mRNA level.

It is also interesting to note that zebularine did not increase either the activity of brostallicin in LNCaP clones overexpressing GST-pi or the activity of the drug in the other prostate cancer cell line, DU145 (presenting lower levels of methylation percentage and a significant GST-pi protein expression and enzymatic activity), further corroborating the hypothesis that the observed synergistic activity with brostallicin in parental LNCaP cells is due to its ability to partially re-express GSTs.

In addition, pretreatment of LNCaP cells with zebularine did not modify the activity of docetaxel or cisplatin (data not shown) in conditions in which brostallicin activity was enhanced, again supporting the specific effect on GSH/GST re-expression induced by zebularine.

Our data indicate that clinical trials combining zebularine with brostallicin could represent a valid alternative for those tumors potentially responsive to the drug that display GST promoter methylation, as is the case for prostate cancer.

The evidence that only partial re-expression of GST is sufficient to increase the GSH/GST-mediated activity of brostallicin is an additional value. This would in fact prevent the interaction with the activity of other anticancer agents (such as cisplatin and melphalan), whose activity could be counteracted by the presence of high levels of GST/GSH.

Materials and Methods

Cell cultures and drug treatments

Human prostate cancer cell lines DU145 and LNCaP, obtained by ATCC, were grown in RPMI 1640 medium supplemented with 10% fetal bovine serum.

LNCaP clones overexpressing GST-pi were generated by transfecting cells with the human GSTP1 gene subcloned in pCDNA3 plasmid using lipofectamine.

Brostallicin was supplied by Pharmacia SpA, zebularine was obtained from National Cancer Institute, 5′aza-deoxycytidine was purchased from Sigma.

In vitro drug-induced cytotoxicity was determined using the MTS assay. Each experiment consisted of 6 replicates.

In vivo brostallicin was administrated intravenously (iv) and zebularine was administrated intraperitoneally (ip); the volume of administration for both drugs was 10 ml/kg of body weight.

Western blotting

For western blot analysis, total cell lysates were prepared following standard procedures and proteins separated on SDS-PAGE and transferred to nitrocellulose membranes. Membranes were reacted with primary antibodies followed by secondary antibodies (Santa Cruz Biotechnology) and developed using the ECL kit (Amersham Biosciences).

Measurement of GST total enzymatic activity

In vitro enzymatic GST total activity was determined using 1-chloro-2–4-dinitrobenzene as a substrate as previously described.¹⁷

Real time RT-PCR

Real time RT-PCR was used for relative quantification of GST-mu mRNA. Three hundred nanograms of total RNA purified with the SV40 Total RNA Isolation System (Promega) were retrotranscribed and amplified by real time PCR (ABI Prism 7900 Sequence Detection System, Applied Biosystems) using ready to use solution (Assay on Demand, Applied Biosystems).

Methylation specific-PCR (MSP)

One microgram of genomic DNA was modified with sodium bisulfite using the Epitect Bisulfite kit (Qiagen) according to the specifications of the manufacturer. To verify successful bisulfite modification of the DNA, a region of the CALPONIN promoter, with cytosines outside of CpG islands, was amplified by PCR for each DNA sample analyzed. MSP products were separated on agarose gels. PCR protocols, used for CALPONIN, GSTP1 and GSTM1 are shown in Table S1.

DNA pyrosequencing

After bisulfite treatment of DNA, PCR was performed using primers that bracket the CGI of the GSTP1 gene promoter. PCR was done in a total volume of 50 ul, containing: 200 ng of template DNA, 10 uM of each primer for GSTP1 purchased by VH Bio Ltd, UK, MgCl₂ 2 mM, 0.2 mM deoxynucleotide triphosphate (Applied Biosystems) and 2 units FastStart Taq (Roche Diagnostics) (Table S1).

The PCR product was immobilized to streptavidin-coated Sepharose beads and single stranded templates were prepared. Sequencing primer (Table S1) was annealed to the template before analysis in the PSQ96MA pyrosequencing system (Biotage). The degree of methylation at individual CpG site was then analyzed with the AQ software (PSQ96MA, version 2.1; Biotage).

Xenograft models

Six-week-old NOD/SCID male mice were obtained from Harlan, Italy. Mice were maintained under specific pathogen-free conditions. Procedures involving animals and their care were conducted in conformity with institutional guidelines that are in compliance with national (Legislative Decree 116 of January 27, 1992, Authorization n.169/94-A issued December 19,1994, by Ministry of Health) and international laws and policies (EEC Council Directive 86/609, OJ L 358. 1, December 12, 1987; Standards for the Care and Use of Laboratory Animals, United States National Research Council, Statement of Compliance A5023-01, November 6, 1998).

Exponentially growing LNCaP cells (5 × 10⁶ cells/mouse) were injected in mice subcutaneously. When tumors reached approximately 0.15–0.2 cm³, mice were randomly divided in groups and untreated or treated with brostallicin alone or in combination with zebularine. Tumor diameters were measured with a caliper twice weekly until the animals were sacrificed. The volume of tumors was calculated by the formula: Tumor volume (cm³) = (length × width²)/2. Body weights were measured weekly during the treatment period.

The percentage of tumor regression (T/C%) was determined as: T/C% = 100× (mean RTV of treated group)/(mean RTV of control group), where RTV = TVn/TV₀ (TVn is the tumor volume at day n and TV₀ is the tumor volume at day 0). The Tumor Growth Delay (T-C) was calculated as the median time (in days) required for the treated-group tumors less the median time required for the control-group tumors to reach a predetermined size (0.4 cm³).

Toxicity was evaluated on the basis of weight loss, calculated by the ratio:

100 × (mean Body Weight Treated Group/mean Body Weight Control Group).

Immunohistochemistry

Formalin fixed tumors were embedded in paraffin. Slices (4 μm) were put on super frost plus slides and hematoxylin and eosin staining was performed for morphological evaluation.

For immunohistochemistry (IHC), the Ventana Discovery XT autostainer was used: after Ventana Cell Conditioner 1 unmasking, incubation (2 h) with GST-pi monoclonal antibody (Novocastra NCL-GSTpi-438) was performed. The En-Vision labeled Polymer-HRP anti mouse (DAKO K4001) was used as secondary antibody with a 16-min incubation. Standard diaminobenzidine (DAB) chromogen from Ventana revealed the antibody binding.

Statistical analysis

Tumor volumes of treated and untreated mice were compared by analysis of variance (ANOVA) using the StatView Statistical Package (SAS Institute Inc., 3rd edition).

Disclosure of Potential Conflicts of Interest

No potential conflicts of interest were disclosed.

Supplementary Material

Supplementary materials may be found here:

http://www.landesbioscience.com/journals/epigenetics/article/24916