Characterization of MTAP Gene Expression in Breast Cancer Patients and Cell Lines

MTAP is a ubiquitously expressed gene important for adenine and methionine salvage. The gene is located at 9p21, a chromosome region often deleted in breast carcinomas, similar to CDKN2A, a recognized tumor suppressor gene. Several research groups have shown that MTAP acts as a tumor suppressor, and some therapeutic approaches were proposed based on a tumors´ MTAP status. We analyzed MTAP and CDKN2A gene (RT-qPCR) and protein (western-blotting) expression in seven breast cancer cell lines and evaluated their promoter methylation patterns to better characterize the contribution of these genes to breast cancer. Cytotoxicity assays with inhibitors of de novo adenine synthesis (5-FU, AZA and MTX) after MTAP gene knockdown showed an increased sensitivity, mainly to 5-FU. MTAP expression was also evaluated in two groups of samples from breast cancer patients, fresh tumors and paired normal breast tissue, and from formalin-fixed paraffin embedded (FFPE) core breast cancer samples diagnosed as Luminal-A tumors and triple negative breast tumors (TNBC). The difference of MTAP expression between fresh tumors and normal tissues was not statistically significant. However, MTAP expression was significantly higher in Luminal-A breast tumors than in TNBC, suggesting the lack of expression in more aggressive breast tumors and the possibility of using the new approaches based on MTAP status in TNBC.


Introduction
Breast cancer is the most common cancer among women worldwide [1][2]. One of the alterations involved in the development and progression of the disease is the loss of expression of tumor suppressor genes [3].
The methylthioadenosine phosphorylase (MTAP) gene is located at 9p21 and is flanked by the tumor suppressor miR-31 and the cyclin-dependent kinase inhibitor 2A (CDKN2A) gene [4], which are approximately 100 kb away [5][6][7][8][9]. MTAP acts in the polyamine biosynthesis pathway and is important to the salvage of both adenine and methionine [3]. MTAP is ubiquitously expressed in all normal tissues but frequently lost in tumors mainly due to a co-deletion with CDKN2A. In normal cells, MTAP cleaves the 5'-deoxy-5'-methylthioadenosine (MTA) substrate generated during the biosynthesis of polyamines, generating adenine and 5-methylthioribose-1-phosphate (MTR-1-P). Adenine is converted to adenosine monophosphate (AMP), and MTR-1-P is converted to methionine. Cells lacking MTAP are unable to salvage AMP or methionine and are more sensitive to inhibitors of de novo AMP synthesis or to methionine starvation than normal cells [10][11]. Because MTAP is expressed in all normal tissues and is usually lost in tumors, Kadariya et al. [12] suggested using MTAP deficiency to selectively target tumor cells that are MTAP-negative. A promising therapeutic approach to cancer was proposed in 2009 by Lubin and Lubin [13], based on the addition of MTA to the treatment of MTAP-negative tumors with toxic purine analogs, like 5-fluorouracil (5-FU). Normal cells are protected from the toxic effects of purine analogs by the AMP produced from MTA. However, MTAP-negative tumor cells are not able to produce AMP from the added MTA, so the purine analogs are metabolized and exert their toxic effects [14].
We have previously reported a 90% frequency of concordant loss of heterozygosity (LOH) for intragenic microsatellite markers for CDKN2A (D9S1748) and MTAP (D9S1749) [15]. These data indicated that in breast cancer cells, the co-deletion might play an important role, as described in other types of tumors (3)(4)(5)(6)(7)(8)(9). The aim of the present study was to characterize MTAP expression in breast cancer patients and cell lines and examine the relationship between MTAP expression and chemo-sensitivity to inhibitors of AMP synthesis.

Ethics Statement
This research was approved by the "Comissão Nacional de Ética em Pesquisa (CONEP)", from the Health Division of Brazilian Government, number 251/2003. The Ethical Committee from the Istituto Ospitalieri di Cremona (Italy) approved the use of formalin-fixed paraffin embedded (FFPE) samples. Written informed consent was obtained from all patients. All of the samples were anonymized by a pathologist staff member, and none of the researchers conducting the analysis had access to the clinico-pathological data.

Fresh tumors
Forty-six fresh primary breast tumors were obtained from 45 patients between 2009 and 2011 at the Hospital Nossa Senhora das Graças (Curitiba, Brazil). Non-compromised tissues of the contralateral breast were obtained from ten patients who underwent simultaneous breast symmetrization. Histological analysis confirmed the normality of these samples. Tumor and normal samples were conserved in an RNA stabilization solution (RNAlater 1 , Applied Biosystems, USA) immediately after surgery, and stored at 4°C until RNA isolation. Clinicopathological information of the patients are summarized in Table 1. Patients had received neither chemotherapy nor radiation prior to surgery.

Formalin fixed-paraffin embedded (FFPE) samples
For MTAP gene expression in FFPE samples, a second group of 81 TNBC and 60 Luminal-A breast tumors were retrospectively collected from patients who came to the medical observation facility at the Breast Care Unit, A.O. Istituti Ospitalieri di Cremona, Italy. We did not have access to the clinico-pathological data of these patients. The histological classification of Luminal A, Luminal B, HER2 positive and TNBC tumors was based on St. Gallen guidelines [16], which define breast tumors according to the immunohistochemical staining of hormonal receptors (ER and PR), HER2 expression, the Ki-67 marker and histological grade [17][18].

Reverse transcribed quantitative PCR (RT-qPCR) and Methylation specific PCR (MS-PCR)
RNA from fresh tumors and normal samples were isolated using an RNAeasy 1 Kit (Qiagen, Germany). RNA from FFPE samples was isolated using High Pure RNA Paraffin Kit (Roche, USA). RNA from cell lines was isolated using SV-Total RNA isolation system (Promega, USA). All mRNAs were reverse-transcribed using the High Capacity cDNA Archive Kit (Applied Biosystems, USA). The integrity of all the RNA preparations was checked on a 1% agarose gel and RNA concentrations were measured with a NanoDrop™ 1000 spectrophotometer (Thermo Scientific, USA). Optimal primer pairs (S1 Table) were chosen, spanning splice junctions, using PRIMER-3 (http://frodo.wi.mit.edu/cgi-bin/primer3/primer3_www.cgi) and Oligo Analyzer software (http://www.idtdna.com/analyzer/applications/oligoanalyzer/). The specificity was verified by detecting single-band amplicons of the PCR products. For the fresh samples, reactions were performed with 15 ng of cDNA template, 2 pmol of forward and reverse primers and 5 μl of SYBR Green PCR Master Mix (Applied Biosystems), and a dissociation curve was evaluated. Standard curves for each gene were included for efficiency reaction analysis, which is necessary to quantify expression based on the comparative method (-DDCt). Samples were then normalized using the housekeeping genes Actin (ACTB) and β-2-microglobulin (B2M). These gene were chosen as housekeeping genes based on the TaqMan 1 Human Endogenous Control Array microfluidic card (Applied Biosystems, USA). The normalized values were compared with expression in the MDA-MB-231 cell line, which is used as a calibrator sample. The data are described as fold changes in gene expression relative to the calibrator sample. Relative MTAP expression in fresh samples was determined on a Mastercycler ep RealPlex System (Eppendorf, Germany). Gene expression was performed using GoTaq qPCR Master Mix (Promega, USA) in an ABI Prism 7900 Sequence Detection System (Applied Biosystems, USA).
To perform MS-PCR, genomic DNA was isolated from six breast cancer cell lines (except SK-BR-3 cell line) using the Maxwell 1 16 Cell DNA Purification Kit (Promega, USA). Genomic DNA was modified with sodium bisulfite using the Epitect Bisulfite Kit (Qiagen, Germany) according to manufacturer specifications. MS-PCR was performed in standard conditions using GoTaq 1 Hot Start (Promega, USA) and 2 μl of modified DNA. By targeting the CpG island sequence in the promoter region of the MTAP gene, specific primers recognizing methylated or non-methylated DNA were designed (S1 Table). MS-PCR products were separated on 2.5% agarose gels. The experiment was repeated twice.

esiRNA transfection
Double-strand esiRNA for MTAP and a non-specific scrambled siRNA were obtained commercially (Eupheria Biotech, Sigma). Fifty microliters of the MTAP-positive MDA-MB-435 cells were seeded in a 96-well culture vessel at 4.0 x 10 4 cells/ml the day before transfection. Cells were transfected with esiRNA duplexes or scrambled siRNA (45 nmol each) using Lipofectamine 2000 reagents (Invitrogen) according to the vendor's protocol. MTAP expression was determined using WB.

Wound Healing assay
The effect of MTAP downregulation on the invasiveness of MDA-MB-435 cells was evaluated with a wound healing assay. Cells were seeded in 24-well culture plates at 4 x 10 4 cells/mL and after 24 hours were transfected with scramble siRNA or MTAP esiRNA in duplicate, as already described. 48 hours later a wound was made along the diameter of each well and a picture was taken. Cells were allowed to proliferate and to invade the wound for the next 96 hours and pictures of the wells were taken. Using ImageJ software the wounded area uncovered by cells was obtained for each picture, the mean of the duplicates was calculated and then expressed as percentage of the initial wounded area.

Cytotoxicity experiments
Seventy-two hours after esiRNA transfection, cells were treated for 72 hours with crescent concentrations of 5'aza-deoxycytidine (AZA), methotrexate (MTX) and 5-FU at a final concentration of 100 μM, 50 μM and 10 μM, respectively. These drugs were chosen due to their action as inhibitors of de novo AMP synthesis. Cell proliferation was measured in a TECAN 1 Infinite 200 multimode microplate reader (TECAN Group Ltd., Switzerland), using the MTS assay as already described. These experiments were performed in triplicates and repeated twice.

Statistical analysis
RT-qPCR data had a non-normal distribution and were analyzed using the Shapiro-Wilk normality test. The Mann-Whitney U test was chosen to compare clinico-pathological and gene expression data between the groups (fresh tumors and normal samples, TNBC and luminal-A tumors), and a linear regression was chosen to correlate gene expression with patient's age and tumor size. Statistical analyses were carried out in Prism 5 version 5.04 (GraphPad Software Inc., USA). All statistical tests were two-tailed and p values <0.05 were considered statistically significant.

MTAP expression in breast cancer samples
Relative mRNA MTAP expression value was 1.39 ± 0.75 and 1.98 ± 1.05 in fresh primary breast tumors and normal samples, respectively. The difference was not statistically significant (p value = 0.09). To evaluate a more homogeneous group, we analyzed Invasive Ductal Carcinoma (IDC) samples separately (35 patients). The relative expression value was 1.32 ± 0.72 and 1.98 ± 1.05 in IDC and normal samples, respectively and this difference was not statistically significant (p value = 0.065) (Fig 1a). For the IDC samples, no significant differences were found between mRNA MTAP expression and the clinico-pathological parameters (axillary lymph node metastasis, ER/PR/HER2 status, tumor grade, tumor size and age at diagnosis). When the different intrinsic subtypes were considered, no difference in mRNA MTAP level was observed between Luminal-A (n = 23) and Luminal-B (n = 14) groups (1.43 ± 0.68 and 1.26 ± 0.6, respectively, p value = 0.45) (Fig 1b). Expression of MTAP mRNA was also evaluated in FFPE breast cancer samples MTAP mRNA level was 1.62 times greater in Luminal-A than in TNBC, and this difference was statistically significant (p value< 0.0001) (Fig 2).

MTAP and CDK2A status in cell lines
MTAP mRNA levels were evaluated in seven breast cancer cell lines. We also investigated the expression of the CDKN2A gene to determine the presence or absence of a co-deletion. CDKN2A and MTAP mRNAs were not detected in MDA-MB-231, ZR-75-I or MCF-7 cells. In

esiRNA transfection and cytotoxicity experiments
Because MTAP levels have been shown to influence the cytotoxicity of some anticancer agents [14],the MDA-MB-435 cell line was chosen to perform knockdown of the MTAP gene using  the esiRNA system. Transfection of esiRNA against MTAP completely knocked down MTAP protein and no differences in cell growth and cell migration, assessed by the wound healing assay were detected. S2 Fig). However, MTAP knockdown increased the cytotoxic activity of 5FU, MTX and Aza (Fig 4a-4c).

Discussion
Curtis and colleagues [19] discovered a high frequency of MTAP deletions in an integrated analysis of copy number and gene expression with two sets of almost a thousand primary breast tumors. Nevertheless, there is a lack of information on MTAP-deficiency in primary breast cancer [4]. In a previous study [15], we found a high rate (90%) of concordant LOH between CDKN2A and MTAP genes in primary breast tumors. Here, we assessed MTAP mRNA expression in a sample of fresh breast tumors and normal breast tissue, and the difference was not statistically significant (Fig 1a). In addition, we did not find any correlation between MTAP expression and the clinico-pathological parameters, probably due the small size of our sample. Miyazaki et al. [20], in a cohort of 40 osteosarcoma samples, found a 27.5% decrease in MTAP protein expression and no correlations with the clinico-pathological parameters. Our results are similar to the findings of Alhebshi et al. [21], who reported MTAP protein expression in 20 normal human skin tissue samples and 109 cutaneous squamous cell carcinomas and found no significant correlations with the clinico-pathological parameters. The small size of our sample and contamination with normal cells after macro-dissection of the fresh tumors may be responsible for the results we obtained. We studied a second group of FFPE samples and found significantly higher expression of MTAP in Luminal-A tumors than of TNBC (Fig 2). Christopher et al. [7] observed that the loss or reduction of MTAP expression in breast tumor cells is involved in anchorage-independent growth. This process is important for the progression of the disease, allowing the tumor to spread and metastasize. These characteristics are commonly observed in the more aggressive cancers like TNBC. Crespo et al. [22] noted the potential relevance of MTAP as a tumor suppressor in glioblastomas because MTAP was the single homozygously deleted gene at chromosome 9p21 (from 11 genes analyzed at this region) for which they found a high correlation between copy number values and mRNA expression levels. However, Dou et al. [23] observed an inverse correlation between cellular differentiation and MTAP relative expression in colorectal cancer, mainly due to promoter demethylation in more malignant tumors. Tang et al. [24] demonstrated that the tumor suppressor function of MTAP in HT1080 fibrosarcoma cells is not the same as its known enzymatic function. MTAP loss can be associated with CDKN2A loss [10][11][12]25], and promoter hypermethylation has been described as an alternative mechanism for the loss of MTAP expression [26]. Here, we characterized gene expression (mRNA and protein levels) of MTAP and CDKN2A in seven breast cancer cell lines and performed a promoter methylation analysis of MTAP (Fig 3,  Table 2 and S1 Fig). MCF-7 and MDA-MB-231 cells were already known to be MTAP-deficient [4]. Our results suggest that cell lines MDA-MB-231 and ZR-75-I harbor a co-deletion of MTAP and CDKN2A genes because neither gene was amplified by RT-qPCR and MS-PCR. In addition, protein expression of neither gene was detected by Western blotting. However, the methods used cannot completely exclude this possibility. In our study, MCF-7 cells show no expression of CDKN2A and MTAP; however. the MTAP promoter was partially methylated. Bisogna et al. [27] also described a deletion of CDKN2A in this cell line but not a deletion of CDKN2B or INK4A genes, which are closely located on the chromosome. Perhaps this is not a case of co-deletion, and MTAP is not expressed in MCF-7 cells due to DNA methylation. T47-D cells show strong CDKN2A mRNA expression but no evidence of protein expression by Western blotting. Bisogna et al. [27] observed DNA methylation of the CDKN2A promoter in this cell line, which could explain the absence of protein in our study. However, the presence of mRNA suggests post-transcriptional regulation, for example, via RNA interference. Kim et al. [3] analyzed a set of gastric cancer cell lines and found a correlation between mRNA down-regulation and homozygous deletion of MTAP and CKN2A because 8 of 10 cell lines expressed both genes. However, the proteins were absent in two out of ten cell lines with a homozygous deletion. We observed no difference in mRNA MTAP expression in respect to their ER, PR and HER2 expressionamong the cell lines and these data partially constrast with the fact that in FFEE triple negative samples did express higher MTAp mRNA levels. The low number of cell lines considered can be at the basis of this discrepancy.
Hellerbrand et al. [28] showed a down-regulation of MTAP in 15 hepatocellular carcinoma (HCC) samples. Another study [29] demonstrated that down-regulation of MTAP increases MTA levels in HCC, which could be involved in HCC progression. Myiazaki et al. [20] proposed that the MTAP enzyme deficiency observed in osteosarcomas was caused by genetic and epigenetic mechanisms and that MTAP deficiency could be exploited using selective chemotherapy with inhibitors of de novo polyamine synthesis. Zimling, Jorgensen and Santoni-Rugiu [30], studying MTAP immunoreactivity in 99 malignant pleural mesotheliomas (MPMs), found that 65% of the tumors analyzed had a decreased reactivity to MTAP. They proposed that this decreased MTAP expression, in combination with other common markers, could be a potential diagnostic marker. As for MPMs, the decreased expression of MTAP in TNBC (Fig 2) could be useful as a diagnostic and therapeutic marker.
Several different approaches based on MTAP status have proposed to use inhibitors of de novo purine synthesis and the enzyme substrate MTA to selectively kill MTAP-negative cells [4,11,13,[31][32][33]. Our gene expression knockdown experiments support the therapeutic approach proposed by Lubin and Lubin [13], once our data show a significantly higher sensitivity of MTAP-negative cells to 5-FU (Fig 4a). Interestingly, our data show that TNBC cells express significantly less MTAP than the more differentiated group composed of Luminal-A breast tumors (Fig 2), which may open the possibility of this new approach to TNBC patients who lack the benefit of endocrine or targeted therapy that is largely used in Luminal and HER2 groups.

Conclusions
This work investigated MTAP expression in breast cancer patients and cell lines, and examined the relationship between MTAP expression and chemo-sensitivity to inhibitors of AMP synthesis. MTAP was found significantly less expressed in TNBC than in Luminal-A breast tumors. We observed that after gene knockdown, MTAP-negative cells were significantly more sensitive to 5-FU, MTX and AZA. The observation that TNBC tumors have lower levels of MTAP has to be corroborated in additional studies, but the observation suggests that this class of patients could benefit from treatment with antimetabolites.  Table. RT-qPCR and MS-PCR primers. Primers sequences are in the sense 5'-3'. m, methylated; u, unmethylated. (DOCX)