DNA methylation and histone modifications modulate the β1,3 galactosyltransferase β3Gal-T5 native promoter in cancer cells

The International Journal of Biochemistry & Cell Biology 44 (2012) 84–90

Contents lists available at SciVerse ScienceDirect

The International Journal of Biochemistry & Cell Biology journal homepage: www.elsevier.com/locate/biocel

DNA methylation and histone modifications modulate the ␤1,3 galactosyltransferase ␤3Gal-T5 native promoter in cancer cells Anna Caretti a , Silvia M. Sirchia b , Silvia Tabano b , Aida Zulueta a , Fabio Dall’Olio c , Marco Trinchera d,∗ a

Department of Medicine, Surgery and Dentistry, San Paolo Hospital, Università degli Studi di Milano, 20142 Milan, Italy Department of Medicine, Surgery and Dentistry – Unit of Medical Genetics, San Paolo Hospital, Università degli Studi di Milano, 20142 Milan, Italy Department of Experimental Pathology, University of Bologna, 40100 Bologna, Italy d Department of Biomedical Sciences Experimental and Clinical (DSBSC), University of Insubria Medical School, 21100 Varese, Italy b c

a r t i c l e

i n f o

Article history: Received 23 August 2011 Received in revised form 19 September 2011 Accepted 27 September 2011 Available online 5 October 2011 Keywords: Glycosyltransferase Glycobiology Carbohydrate antigen Gene expression Tumor marker

a b s t r a c t The native promoter of ␤1,3 galactosyltransferase ␤3Gal-T5 contributes to the expression of the enzyme and its oligosaccharide products, such as Lewis antigens, in many tissues. It is mainly sensitive to nuclear factor NF-Y and located nearby two CpG islands. To elucidate the regulation of the native promoter, we analyzed NF-Y protein and ˇ3Gal-T5 mRNA, and found that NF-Y is scarcely modulated among various cell lines and biopsies from normal or cancerous colon. Conversely, ˇ3Gal-T5 expression levels vary in the cell lines and are strongly down-regulated in colon cancer. We also performed quantitative methylation analysis of ˇ3Gal-T5 CpG islands and found an inverse correlation between mRNA expression and DNA methylation. In particular, the methylation levels of both islands are always increased in cancer, with respect to the corresponding normal counterpart, in matched normal and tumor samples of colon and breast origin. Moreover, treatment with chromatin remodeling agents 5-aza-2 deoxycytidine and trichostatin A does not restore transcription in completely negative cells, but only increases expression in basally positive cells. However, methylation analysis after 5-aza-2 deoxycytidine treatment revealed partial demethylation of both islands in all treated cells. Finally, chromatin immunoprecipitation assays on ˇ3Gal-T5 promoter showed that histone H3K4 trymethylation, H3K79 dimethylation, and H3K9-14 acetylation are high in cells expressing the transcript, and very low in those negative, while H4K20 trimethylation and H3K27 dimethylation are the opposite. We conclude that complex epigenetic modulation underlies the regulation of ˇ3Gal-T5 native promoter. © 2011 Elsevier Ltd. All rights reserved.

1. Introduction ␤1,3 galactosyltransferase ␤3Gal-T5 is responsible for type 1 chain oligosaccharide synthesis, including the selectin ligand sialyl-Lewis a (NeuAc␣2,3Gal␤1,3[Fuc␣1,4]GlcNAc), epitope of tumor marker CA19.9, and other Lewis antigens as Lewis a (Gal␤1,3[Fuc␣1,4]GlcNAc) and Lewis b (Fuc␣1,2Gal␤1,3[Fuc␣1,4]GlcNAc) (Isshiki et al., 1999). The role of type 1 chain oligosaccharide is not known in details. On one hand, sialyl-Lewis a is a selectin ligand potentially involved in cancer metastasis (Kannagi, 2007) and considered a marker of

Abbreviations: Gal-T, galactosyltransferase; NF-Y, CAAT binding factor (also named CBF); LTR, long terminal repeat; ChIP, chromatin immunoprecipitation; 5AZA, 5-aza-2 deoxycytidine; TSA, trichostatin A; H3, histone 3; RT, reverse transcription. ∗ Corresponding author at: DSBSC, via JH Dunant 5, 21100 Varese, Italy. Tel.: +39 0332 397 160; fax: +39 0332 397 119. E-mail address: [email protected] (M. Trinchera). 1357-2725/$ – see front matter © 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.biocel.2011.09.010

malignancy (Yamashita and Watanabe, 2009). On the other hand, the synthesis of type 1 chain oligosaccharides was reported to counteract the synthesis of polylactosamine chains and sialylLewis x (NeuAc␣2,3Gal␤1,4[Fuc␣1,3]GlcNAc) (Salvini et al., 2001; Isshiki et al., 2003; Mare and Trinchera, 2004), another selectin ligand, both involved in tumor progression and metastasis. ␤3GalT5 is active in epithelia of various organs. In mammary gland, thymus, and trachea, as well as in some human cancer cell lines, transcription is mainly driven by a native promoter that was found sensitive to nuclear factor NF-Y, also in mice (Mare and Trinchera, 2007). In the organs of the gastrointestinal tract (as the colon, stomach and pancreas) another promoter is active and stronger than the native promoter (Isshiki et al., 2003; Mare and Trinchera, 2007; Dunn et al., 2003). This alternative promoter has a retroviral origin (named LTR), has been probably acquired about 10–15 millions years ago (Dunn et al., 2005), and is regulated through homeoproteins such as hepatocyte nuclear factor HNF1, and caudal-related homeobox Cdx (Isshiki et al., 2003; Dunn et al., 2003). In various cell lines of different tissue origin, but even among those derived from the same tissue, ␤3Gal-T5 transcript,

A. Caretti et al. / The International Journal of Biochemistry & Cell Biology 44 (2012) 84–90

as a whole, is widely modulated (Isshiki et al., 1999; Mare and Trinchera, 2004). Moreover, it is strongly down-regulated in colon cancer with respect to the normal mucosa (Salvini et al., 2001; Isshiki et al., 2003): the mechanisms responsible for differential expression among cell lines and down-regulation in colon cancer are not known at present. Since NF-Y is rather ubiquitous (Caretti et al., 2003) and the native promoter is located in between two CpG islands (supplemental Fig. S1), we wanted to elucidate whether modulation of NF-Y accounts for differential expression of the transcript, or whether epigenetic mechanisms, such as DNA methylation and histone modifications, are involved. To this aim we first compared the expression levels of NF-Y in various cell lines and tissues expressing different amounts of the native transcript. We then performed quantitative bisulfite DNA sequencing starting form the same sources, and treated cell lines with drugs affecting DNA methylation and histone deacethylation to rescue expression. Moreover, we performed ChIP analysis to asses the chromatin conformation in some cell lines expressing different levels of such transcript. 2. Materials and methods 2.1. Cell line, tissues, and cell treatments Human breast cancer cell lines MCF-7 and MDA-MB-231, human gastric cell line MKN-45, human bile duct carcinoma cell line HuCC-T1, and human colon cancer cell lines HT-29, HCT-15, COLO205 and SW1116 were cultured as reported (Mare and Trinchera, 2004, 2007). Human breast cancer cell line BT-474, a gift of Dr. G. Fontana (University of Milan), was cultured in DMEM containing 10% fetal bovine serum, 100 U/ml penicillin, 1 mg/ml streptomycin, and 2 mM l-glutamine. Human breast samples were collected at surgery and kindly donated by Dr. Filippo Mare (Hospital di Circolo, Varese, Italy); human colon samples were as reported (Salvini et al., 2001; Trinchera et al., 2011). For treating cells with drugs affecting DNA methylation and histone deacethylation, HCT-15, MDA-MB231, MCF-7, and MKN-45 (1–4 × 105 cells) were plated in 6-well plates, incubated 24 h with regular medium that was replaced with medium containing different amount of 5AZA (Sigma, dissolved in DMSO as 10 mM stock solution) and/or TSA (Sigma, dissolved in ethanol as 1 mg/ml stock solution). Media were replaced every 24 h with media containing freshly diluted drugs. At the end of treatment cells were harvested by trypsinization and processed for DNA or RNA preparation. 2.2. Western blot Freshly collected cell pellets, upon trypsinization and washing with PBS, were processed to obtain nuclear extracts using a commercially available kit (NEPER, Pierce) as reported (6). Frozen biopsies from colon and breast were dounce homogenized and submitted to nuclear extraction as for cell pellets. Protein concentration was determined by the Coomassie Plus Protein Assay (Pierce). Aliquots of nuclear extracts (5–10 ␮g of protein) were separated by 10% SDS-PAGE, transferred to a nitrocellulose membrane (Trans-Blot SD Semi Dry Transfer Cell, BIORAD) and blotted with rabbit polyclonal anti NF-YA (Santa Cruz sc-10779, 1:200) or rabbit polyclonal anti H3 (Cell Signaling, 1:500), according to our published protocol (Caretti et al., 2010). 2.3. Bisulfite sequencing Genomic DNA was extracted from human tissues and cell lines using a commercially available kit (QIAamp DNA, Qiagen) and 0.1–1.5 ␮g were submitted to bisulfite treatment and purification using the Epitect bisulfite Kit (Qiagen), or the MethylCode bisulfite

85

conversion Kit (Invitrogen). For cloning, the obtained material was amplified by PCR as follows. Amplifications (35 cycles) were performed in 25 ␮l using an hot start Taq (Promega) according to the manufacturer’s recommendations in the presence of an enhancer (PCRx Enhancer system, Invitrogen) with 2 ␮l of bisulfite converted DNA as template and first reaction primers (see supplemental Table S1). Amplification program included a single treatment at 94 ◦ C for 3 min followed by cycles consisting of 1 min at 94 ◦ C, 1 min at 58 ◦ C, and 1 min at 72 ◦ C, and a final extension step at 72 ◦ C for 8 min. Nested PCRs (25 cycles) were performed in a final volume of 50 ␮l using 1–2.0 ␮l of the PCR products and inner primers designed to contain restriction sites (see supplemental Table S1). Reaction mixtures and amplification programs were as above but annealing temperature was 62 ◦ C. Amplified fragments were column-cleaned, digested with appropriate restriction enzymes, column cleaned again, and cloned into pGl3 vector for sequencing (at Eurofins sequencing service). Pyrosequencing experiments were aimed to quantitatively evaluate the methylation levels of two islands, CpG 1 and CpG 2, allowing us to investigate 4 CpG sites on each island (for details on PCR and sequencing primers see supplemental Table S1 and supplemental Fig. S1). PCR were carried out starting from 20 ng of bisulfite-converted DNA, in a final volume of 50 ␮l. PCR conditions were the same for both regions and included: 48 cycles consisting of 95 ◦ C for 30 s, 56.4 ◦ C for 30 s and 72 ◦ C for 20 s; 30 ␮l of PCR products were loaded on PyroMArk ID instrument (Biotage AB, Uppsala, Sweden) and quantitative DNA methylation analyses were performed in the PSQ HS 96 System (Biotage), with the PyroGold SQA reagent kit (Biotage AB, Uppsala, Sweden) according to the manufacturer’s instructions. Raw data were analyzed using the Q-CpG software v1.0.9 (Biotage AB, Uppsala, Sweden), that calculates the ratio of converted C’s (T’s) to unconverted C’s at each CpG, giving the percentage of methylation. For each sample, the methylation value represents the mean between at least two independent PCR and Pyrosequencing experiments. 2.4. RT-PCR Quantification of ␤3Gal-T5 native transcripts was performed by competitive RT-PCR as previously described (Mare and Trinchera, 2004; Trinchera et al., 2011). Briefly, total RNA, prepared and DNase-treated using a commercially available kit (SV RNA, Promega) was quantitated by fluorometry with Qubit (Invitrogen). First strand cDNA was synthesized in a 20 ␮l volume by Moloney Murine Leukemia virus reverse transcriptase (2500 U/ml, USB-Affymetrix) in the presence of 1.0–2.0 ␮g RNA, 0.4 ␮M oligo(d)T12–18 primer, the supplied buffer, and 1000 U/ml human placental RNase inhibitor. Reactions were kept at 37 ◦ C for 45 min and then at 42 ◦ C for 45 min. Control reactions were prepared by omitting the reverse transcriptase. cDNAs were amplified in a volume of 25 ␮l in the presence of the indicated amounts of competitor, for 35 cycles (␤3Gal-T5) or 5 pg competitor for 25 cycles (␤-actin), using 2.5 U/ml of GoTaq Hot Start polymerase (Promega) in a mixture containing the supplied buffer, 2.0 mM MgCl2 , 0.1 mM dNTPs, 250 pmol/ml of each primer, and different amounts of cDNAs. Amplification program included a single treatment at 94 ◦ C for 3 min followed by cycles consisting of 1 min at 94 ◦ C (melting) and 3.5 min at 72 ◦ C (annealing plus extension) and a final extension step at 72 ◦ C for 8 min. No amplification was detected with control reactions. Competitor cDNAs were as reported (Salvini et al., 2001; Mare and Trinchera, 2007). Parallel PCR amplifications were performed on known amounts of standard cDNAs premixed with the competitors. Standard cDNAs were the original cloned sequences quantitated and diluted as for the competitors. Aliquots of PCR reactions were analyzed on 1% agarose gels stained with ethidium

86

A. Caretti et al. / The International Journal of Biochemistry & Cell Biology 44 (2012) 84–90

bromide. Quantification was performed by densitometric scanning of the gel and the amounts of amplified target cDNAs were calculated from their respective standard curves and normalized by those for ␤-actin. The target/competitor ratios were proved to make PCR results quantitative (Gilliland et al., 1990). 2.5. CHIP assay Chromatin was prepared from about 107 cells using a commercially available kit (SimpleChIP, Cell Signaling) with the following modifications: nuclei were resuspended in the provided buffer, diluted to have an optical density of 0.2 at 260 nm, and treated with a Micrococcal nuclease (Takara) for 5 min at 37 ◦ C, using 100 Unit of enzyme for each ml of suspension. Such condition was found in a preliminary experiment to provide the best DNA fragmentation (1–5 nucleosomes, 150–900 bp). About 1 ␮g of chromatin DNA, determined by fluorometry upon reversing cross linking (Qubit, Invitrogen), diluted to a final volume of 0.25 ml, was used for each immunoprecipitation. Aliquots (10–20 ␮l) were kept as input DNA. Antibody binding reactions were performed at 4 ◦ C overnight under rotation. Precipitation with agarose-bound G-protein, washing, elution, reverse of cross-linking, and DNA purification were performed with the kit reagents according to the manufacture’s protocol. The following antibodies were used: anti rabbit IgG, 2 ␮l (Cell Signaling); anti-acetyl-Histone H3 (Lys 9), named H3K9Ac, 2 ␮l (GeneSpin Milan, Italy, PAb004), antitrymethyl-Histone H3 (Lys4), named H3K4me3, 2 ␮l (GeneSpin, PAb005), anti-dimethyl-Histone H3 (Lys27), named H3K27me2, 2 ␮l (Abcam, ab24684), anti-dimethyl-Histone H3 (Lys79), named H3K79me2, 2 ␮l (Millipore #04-835), anti-trymethyl-Histone H4 (Lys20), named H4K20me3, 2 ␮l (Millipore #07-749), anti-acetylHistone H3 Lys 9/Lys 14, named H3K9-14Ac, 1 ␮l (Millipore #06-599), To quantify the immunoprecipitated DNA competitive PCR was performed using 10 fg of a 200 bp oligonucleotide (IDT) as competitor and specific primers (see supplemental Table S1 for sequences). Reaction mixture, 25 ␮l, and amplification program were as above described for RT-PCR, but annealing temperature was 68 ◦ C and cycles repeated 35 times. Parallel PCR amplifications were performed on known amounts of standard cDNAs premixed with the competitor. Standard cDNA was the original cloned sequence quantitated and diluted as for the competitor. Aliquots of PCR reactions were analyzed on 1% agarose gels stained with ethidium bromide. Quantification was performed by densitometric scanning of the gel and the amounts of amplified target cDNAs were calculated from the standard curve and normalized by input DNA amplification. 3. Results 3.1. Expression levels of ˇ3Gal-T5 native transcript and transcription factor NF-Y To evaluate the actual quantitative relationship between NFY levels and transcript expression we analyzed various cell lines and human colon biopsies by western blot and competitive RT-PCR. Using an antibody specific to the A subunit of NF-Y trimer, we found that such transcription factor was ubiquitously present in all cell lines, irrespective of the expression levels of native ␤3Gal-T5 transcript (Fig. 1). In particular, relevant amounts of NF-Y were detected in cell lines devoid of any detectable ␤3Gal-T5 native transcript, as HCT-15, MDA-MB-231, and BT-474. Similarly, in two specimens obtained from healthy human colon mucosa, that express high levels of transcript, and in 6 specimens obtained from colon cancer biopsies, expressing no or low amounts of transcript, NF-Y was also ubiquitously detected (Fig. 2). The finding that the native transcript

Fig. 1. Quantification of NF-Y protein in various cell lines expressing different levels of ␤3Gal-T5 native transcript. (A) Nuclear extracts (5–10 ␮g of protein) were separated by 10% SDS-PAGE and transferred to a nitrocellulose membrane that was blotted with anti NF-Y or anti histone H3 antibodies followed by HRP-labeled secondary antibody and chemiluminescence detection, as detailed under “Experimental procedures”. The NF-Y doublet seen in MDA-MB-231 cells corresponds to the long (upper band) and short form (lower band) as reported (Li et al., 1992). A representative western blot is shown. (B) RNA extracted from the cell lines was reverse transcribed, and normalized amounts of the resultant first-strand cDNA were mixed with competitor (truncated) cDNAs (5 pg and 5 fg for ␤-actin and ␤3GalT5, respectively) and subjected to PCR (25 and 35 cycles for ␤-actin and ␤3Gal-T5, respectively) using primers specific to ␤-actin and ␤3Gal-T5 native transcript, respectively. Twenty percent of each amplification reaction was electrophoresed in a 1% agarose gel and visualized by ethidium bromide staining. The target doublet corresponds to the alternative splicing previously reported (Mare and Trinchera, 2007). A representative gel is shown. (C) Densitometric scanning of bands obtained by western blots or gel images was performed to quantitate NF-Y and ␤3Gal-T5 native transcript, respectively. The amounts of NF-Y were calculated with respect to the corresponding bands of histone H3 (right scale); those of ␤3Gal-T5 native transcript were calculated from a standard curve and normalized to the amounts calculated for ␤-actin (left scale). Results are the mean ± standard deviation for three determinations.

is scarcely detectable in colon cancer prompted us to evaluate the expression in 10 matched pairs of human colon cancer and adjacent normal mucosa (Fig. 3). In 9 of them, the transcript appeared downregulated in cancer confirming that silencing of native promoter is part of the reported down-regulation of ˇ3Gal-T5 gene that occurs during malignant transformation (Salvini et al., 2001; Isshiki et al., 2003). ␤3Gal-T5 native promoter is located in the context of two CpG islands, CpG 1 and 2 (supplemental Fig. S1): CpG 1 is shorter (15 CG pairs) and placed upstream, CpG 2 is longer (76 CG pairs) and placed downstream, and includes the NF-Y responsive element, exon 1, and the initial part of intron 1. Taking together such data, we hypothesized that epigenetic mechanisms may affect expression. 3.2. Methylation of ˇ3Gal-T5 native promoter in cell lines and biopsies We performed quantitative analysis of methylation of CpG islands 1 and 2 (supplemental Fig. S1) using pyrosequencing assay, starting from genomic DNA of different cancer cell lines and matched normal and tumor samples. In addition, to assess the

A. Caretti et al. / The International Journal of Biochemistry & Cell Biology 44 (2012) 84–90

87

Fig. 2. Quantification of NF-Y protein and ␤3Gal-T5 native transcript in human colon samples derived from normal mucosa or cancer specimens. Detection and quantification of NF-Y by western blot (panel A) and of ␤3Gal-T5 native transcript by competitive RT-PCR (panel B) were performed and used for bar graph quantification (panel C) as in Fig. 1. The NF-Y doublet seen in sample A2 corresponds to that detected in MDA-MB-231 cells (Fig. 1). Bioptic material is listed in supplemental Table S2.

spread of methylation along the entire CpG islands, we carried out a direct bisulfite sequencing of the promoter region of 8 independent clones derived from the amplification of 855 bps, encompassing 12 and 66 CpGs of CpG islands 1 and 2, respectively. In the cell lines (Fig. 4, upper panel and supplemental Fig. S2A) we found a strong inverse correlation between transcript expression and DNA methylation status. In particular, in HuCC-T1 cells, that express the highest level of the transcript, CpG 2 appears almost unmethylated, and CpG 1 is just scarcely methylated (mean methylation levels ∼20%). In MKN-45 and MCF-7 cell lines, where the expression levels of transcript are from low to moderate, CpG 1 is more methylated (mean methylation values ∼80% in both lines) while CpG 2 is almost unmethylated in MCF-7, and mildly methylated

Fig. 3. Quantification of ␤3Gal-T5 native transcript expressed in matched pairs of colon cancer and surrounding normal mucosa. Detection and quantification of ␤3Gal-T5 native transcript by competitive RT-PCR was performed and presented as in panels B and C of Fig. 1.

Fig. 4. Methylation analysis of ␤3Gal-T5 native promoter by pyrosequencing. Genomic DNA was extracted from cell lines and biopsies, treated with bisulfite, purified, and analyzed by pyrosequencing, using amplification and sequencing primers specific for CpG 1 and 2 of the promoter sequence (see supplemental Table S1 and Fig. S1). Raw data (pyrograms) were analyzed using the Q-CpG software v1.0.9, that calculates the ratio of converted C’s (T’s) to unconverted C’s at each CpG, giving the percentage of methylation. For each sample, the methylation value represents the mean between at least two independent PCR and pyrosequencing experiments. Examples of pyrograms are reported in supplemental Fig. S3. The amounts of ␤3GalT5 native transcript in the corresponding samples are reported on the right scale as deduced from previous figures.

in MKN-45 (mean methylation levels ∼30%). In HCT-15 and MDAMB-231 cells, where the transcript is undetectable, both CpGs are hypermethylated (from 70 to 90%). In matched normal and tumor colon samples (Fig. 4 central panel and supplemental Fig. S2B), the methylation levels of both islands are increased in cancer with respect to the corresponding normal mucosa. In this latter, where the expression of transcript is detectable at different levels, mean methylation values of CpG 1 is below 50%, and that of CpG 2 below 20%. In cancer samples, it increases up to 60–70% in CpG 1 and up to 40% in CpG 2. Interestingly, in the only sample pair where transcript remains detectable in cancer sample, methylation of CpG 2 is only 4%, less than the corresponding normal mucosa. To better sound out the role of promoter methylation in cancerogenesis, we also

88

A. Caretti et al. / The International Journal of Biochemistry & Cell Biology 44 (2012) 84–90

Fig. 5. Effect of 5AZA and TSA treatment on the expression of ␤3Gal-T5 native transcript in MKN-45 cells. MKN-45 cells were treated with various amounts of DNA methyltransferase inhibitor 5AZA or of histone deacetylase inhibitor TSA, or both, for 4 days as detailed under “Experimental Procedures”. Detection and quantification of ␤3Gal-T5 native transcript by competitive RT-PCR were performed as in Fig. 1.

analyzed by pyrosequencing (Fig. 4 lower panel) four matched normal and tumor breast samples. Similarly, we found that the degree of methylation is higher in cancer cells than in normal counterpart. Moreover, we were able to perform the quantification of ␤3GalT5 native transcript only in one normal breast specimen (sample 1, 0.8 fg/pg ␤-actin) and 3 cancer specimens (samples 1, 3 and 4, undetectable in all cases). The overall results indicate that DNA methylation of the ␤3Gal-T5 native promoter occurs in cancer cells and correlates with a lower level of gene expression. 3.3. Effect of chromatin remodeling agents and chromatin conformation To further address the hypothesis of an epigenetic regulation of ␤3Gal-T5 transcription, we treated some cell lines with DNA methyltransferase inhibitor 5AZA and histone deacethylase inhibitor TSA. In two cell lines negative to native transcript (HCT15 and MDA-MB-231), treatment for different times with different concentrations of each drug, or with a combination of both, was unable to restore a detectable expression. However, the treatment was able to reduce methylation (Fig. 4 upper panel) from ∼80% to ∼60% in MDA-MB-231 and from ∼70 to ∼40% in HCT15. Indeed, both agents were able to increase expression of the transcript in MKN-45 cells, that express per se a low amount of the transcript (Fig. 5). In particular, both TSA and 5AZA treatments increased the expression of ∼80%. Combination of both drugs failed to provide any further improvement. We were unable to perform similar experiments in MCF-7 cells since they were not viable upon such treatments. In line with expression finding, pyrosequencing analysis of MKN-45 cells treated with 5AZA (Fig. 4 upper panel) showed demethylation of both islands. Altogether the results from methylation analysis and drug experiments suggest that complex epigenetic modulation underlies the regulation of this promoter. 3.4. Histone modifications A panel of histone modifications including markers (Cedar and Bergman, 2009) of active (H3K4me3, H3K79me2 and H3K9-14Ac) versus inactive (H3K27me2, H4K20me3) chromatin were tested. We performed quantitative ChIP assay from five cell lines expressing different levels of ␤3Gal-T5 native transcript, using antibodies specific to each modification (Fig. 6). High expression of transcript in HuCC-T1 cells was found together with high levels of

Fig. 6. ChIP analysis of histone modifications of ␤3Gal-T5 native promoter in various cell lines. Chromatin was cross-linked, prepared, digested, and immunoprecipitated with the indicated antibodies as reported under “Experimental procedures”. Aliquots of immunoprecipitated DNA and dilutions of non-precipitated (Input) DNA were mixed with 10 fg of competitor DNA and submitted to PCR amplification (35 cycles) using a primer pair spanning 250 bp near to the NF-Y responsive element of ␤3Gal-T5 native promoter. (A) Twenty percent of each amplification reaction was electrophoresed in a 1% agarose gel and visualized by ethidium bromide staining. A representative gel is shown. (B) The relative level of each histone modification was calculated by a standard curve and normalized to the input DNA. Results are the mean ± standard deviation for two amplifications of two independent precipitations.

modifications associated with transcriptionally competent chromatin (H3K4me3, H3K79me2, H3K9Ac and H3K9-14Ac), and low levels of those related to silenced chromatin (H3K27me2 and H4K20me3). Moderate to low expression of transcript in MCF7 and MKN-45 cells was associated with a similar pattern, with quantitative differences especially in MKN-45 cells. Absence of the transcript (HCT-15 and MDA-MB-231 cells), was related to the opposite histone code. These cell lines in fact resulted negative for H3K4me3, H3K79me2 and H3K9-14Ac and positive for H3K27me2, H4K20me3 and H3K9Ac. These results confirm that the transcriptional activity of ␤3Gal-T5 native promoter is associated with chromatin status.

A. Caretti et al. / The International Journal of Biochemistry & Cell Biology 44 (2012) 84–90

4. Discussion In this paper we found that a complex interplay of epigenetic mechanisms directs the expression of ˇ3Gal-T5 from native promoter. This promoter is known to be sensitive to NF-Y transcription factor, that we found not significantly modulated in various cell lines as well as in normal versus colon cancer matched tissues, according to the concept that it is a rather ubiquitous factor (Caretti et al., 2003). Quantitative evaluation of ˇ3Gal-T5 promoter methylation of both CpG islands indicates that the methylation levels are strongly modulated and inversely correlate with the gene expression. In particular, in a cell line characterized by high levels of ␤3Gal-T5 native transcript (HuCC-T1), both islands are scarcely methylated, while in cell lines negative for ˇ3Gal-T5 expression (HCT-15 and MDA-MD-231), both islands are almost fully methylated. This correlation is also observed in colon and breast cancer biopsies, where it is evident in matched pairs of specimens and appears as a rather constant feature. Analysis of the entire CpG regions by direct bisulfite sequencing confirmed the pyrosequencing data, indicating that the methylation levels of the pyro-targeted sequences reflect those of the whole corresponding island. Interestingly, treatment with chromatin remodeling agents does not restore transcription in completely negative cell lines, but only increases expression in mildly positive cells (MKN-45). However, methylation analysis after 5AZA treatment revealed partial demethylation of both islands in all treated cell lines. Nevertheless this demethylation is not sufficient to restore transcription where the promoter is completely silenced, but can enhance the expression from a basically active promoter. The chromatin organization of ˇ3Gal-T5 native promoter is consistent with methylation results. We found that H3K4 trymethylation, H3K79 dimethylation and H3K9-14 acetylation are high in cells expressing the transcript, and very low in those not expressing, while H4K20 trimethylation and H3K27 dimethylation are the opposite. H3K9 acetylation is instead constantly high in this promoter irrespective of the transcriptional activity. Many other glyco-genes were reported to be deregulated in cancer cells because of epigenetic alterations (Kannagi et al., 2010; Kim and Deng, 2008; Miyazaki et al., 2004), including galectins (Demers et al., 2009; Juszczynski et al., 2010), enzymes involved in the biosynthesis of sugar nucleotides (Giordanengo et al., 2004; Oetke et al., 2003), transporters (Yusa et al., 2010) and even glycosyltransferases (Chakraborty et al., 2006; Kawamura et al., 2008; Serpa et al., 2006; Tong et al., 2010; Chachadi et al., 2010). However, many of the cited studies are restricted to the investigation of the methylation status of the promoter region. In the case of ␤1,4 N-acetylgalactosaminyltransferase 4GalNAcT-II, a CpG island nearby the putative promoter region is heavily methylated in colon cancer (Kawamura et al., 2008), and treatment with 5AZA only results in partial recovery of enzyme expression (Wang et al., 2008). Interestingly, this enzyme is involved in Sda antigen (NeuAc␣2,3[GalNAc␤1,4]Gal␤1,4GlcNAc) biosynthesis and its expression counteracts that of sialylated Lewis antigens (Kawamura et al., 2005; Malagolini et al., 2007). Very recently, epigenetic regulation of N-acetylglucosaminyltransferase IX was reported in neural cells (Kizuka et al., 2011). In this case, the chromatin status appeared responsible for tissue specific transcription. The mutual interplay among different epigenetic mechanisms, more than the single processes considered per se, is emerging as a paradigm of gene regulation (Murr, 2010). In fact, increasing evidences suggest that strong interplay exists between DNA methylation and histone modifications: some patterns are associated with potential full reactivation of transcription upon demethylation, while others are not. In particular, many studies have shown that hypermethylation of CpG islands at the promoter of genes

89

is additionally associated with particular combinations of chromatin modifications. Some of them, but not all, are reverted by demethylating treatments. In fact, various hypermethylated tumor suppressor genes reactivated by DNA demethylation do not return to a fully euchromatic state, because they maintain several repressive histone modifications marks (McGarvey et al., 2006; Jacinto et al., 2009). The results of the present investigation on ˇ3Gal-T5 chromatin status are in agreement with such model of gene regulation. It is predictable that the actual relevance of epigenetic mechanisms in the regulation of glyco-genes is still underestimated and represents an intriguing challenge of future investigations (Lauc and Zoldos, 2009). In conclusion, our overall results suggest that different degrees of repression can affect ˇ3Gal-T5 promoter. Extinction of ˇ3GalT5 transcription must be determined by a stable repressive state in the chromatin structure determined by more than one mechanism, including DNA methylation and chromatin modifications. Acknowledgments This work was supported by grants from Mizutani Foundation for Glycosciences (2008) to MT, and from the University of Bologna and Pallotti Legacy for Cancer Research to F.D. AC was supported by a fellowship from the University of Milan. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.biocel.2011.09.010. References Caretti G, Salsi V, Vecchi C, Imbriano C, Mantovani R. Dynamic recruitment of NF-Y and histone acetyltransferases on cell-cycle promoters. J Biol Chem 2003;278:30435–40. Caretti A, Bianciardi P, Sala G, Terruzzi C, Lucchina F, Samaja M. Supplementation of creatine and ribose prevents apoptosis in ischemic cardiomyocytes. Cel Physiol Biochem 2010;26:831–8. Chachadi VB, Cheng H, Klinkebiel D, Christman JK, Cheng PW. 5-Aza2 -deoxycytidine increases sialyl Lewis X on MUC1 by stimulating beta-galactoside:␣2,3-sialyltransferase 6 gene. Int J Biochem Cell Biol 2010;43:586–93. Chakraborty AK, Sousa JF, Chakraborty D, Funasaka Y, Bhattacharya M, Chatterjee A, et al. GnT-V expression and metastatic phenotypes in macrophage-melanoma fusion hybrids is down-regulated by 5-Aza-dC: evidence for methylation sensitive, extragenic regulation of GnT-V transcription. Gene 2006;374:166–73. Cedar H, Bergman Y. Linking DNA methylation and histone modification: patterns and paradigms. Nat Rev Genet 2009;10:295–304. Demers M, Couillard J, Giglia-Mari G, Magnaldo T, St Pierre Y. Increased galectin-7 gene expression in lymphoma cells is under the control of DNA methylation. Biochem Biophys Res Commun 2009;387:425–9. Dunn CA, Medstrand P, Mager DL. An endogenous retroviral long terminal repeat is the dominant promoter for human beta1,3-galactosyltransferase 5 in the colon. Proc Natl Acad Sci USA 2003;100:12841–6. Dunn CA, van de Lagemaat LN, Baillie GJ, Mager DL. Endogenous retrovirus long terminal repeats as ready-to-use mobile promoters: the case of primate beta3GAL-T5. Gene 2005;364:2–12. Gilliland G, Perrin S, Blanchard K, Bunn HF. Analysis of cytokine mRNA and DNA: detection and quantitation by competitive polymerase chain reaction. Proc Natl Acad Sci USA 1990;87:2725–9. Giordanengo V, Ollier L, Lanteri M, Lesimple J, March D, Thyss S, et al. Epigenetic reprogramming of UDP-N-acetylglucosamine 2-epimerase/Nacetylmannosamine kinase (GNE) in HIV-1-infected CEM T cells. FASEB J 2004;18:1961–3. Isshiki S, Togayachi A, Kudo T, Nishihara S, Watanabe M, Kubota T, et al. Cloning, expression, and characterization of a novel UDP-galactose:␤N-acetylglucosamine ␤1,3-galactosyltransferase (␤3Gal-T5) responsible for synthesis of type 1 chain in colorectal and pancreatic epithelia and tumor cells derived therefrom. J Biol Chem 1999;274:12499–507. Isshiki S, Kudo T, Nishihara S, Ikehara Y, Togayachi A, Furuya A, et al. Lewis type 1 antigen synthase (beta3Gal-T5) is transcriptionally regulated by homeoproteins. J Biol Chem 2003;278:36611–20. Jacinto FV, Ballestar E, Esteller M. Impaired recruitment of the histone methyltransferase DOT1L contributes to the incomplete reactivation of tumor suppressor genes upon DNA demethylation. Oncogene 2009;28:4212–24.

90

A. Caretti et al. / The International Journal of Biochemistry & Cell Biology 44 (2012) 84–90

Juszczynski P, Rodig SJ, Ouyang J, O’Donnell E, Takeyama K, Mlynarski W, et al. MLL-rearranged B lymphoblastic leukemias selectively express the immunoregulatory carbohydrate-binding protein galectin-1. Clin Cancer Res 2010;16:2122–30. Kannagi R. Carbohydrate antigen sialyl Lewis a—its pathophysiological significance and induction mechanism in cancer progression. Chang Gung Med J 2007;30:189–209. Kannagi R, Sakuma K, Miyazaki K, Lim KT, Yusa A, Yin J, et al. Altered expression of glycan genes in cancers induced by epigenetic silencing and tumor hypoxia: clues in the ongoing search for new tumor markers. Cancer Sci 2010;101:586–93. Kawamura YI, Kawashima R, Fukunaga R, Hirai K, Toyama-Sorimachi N, Tokuhara M, et al. Introduction of Sda carbohydrate antigen in gastrointestinal cancer cells eliminates selectin ligands and inhibits metastasis. Cancer Res 2005;65:6220–7. Kawamura YI, Toyota M, Kawashima R, Hagiwara T, Suzuki H, Imai K, et al. DNA hypermethylation contributes to incomplete synthesis of carbohydrate determinants in gastrointestinal cancer. Gastroenterology 2008;135:142–51. Kim YS, Deng G. Aberrant expression of carbohydrate antigens in cancer: the role of genetic and epigenetic regulation. Gastroenterology 2008;135:305–9. Kizuka Y, Kitazume S, Yoshida M, Taniguchi N. Brain-specific expression of Nacetylglucosaminyltransferase IX (GnT-IX) is regulated by epigenetic histone modifications. J Biol Chem 2011;286:31875–84. Lauc G, Zoldos V. Epigenetic regulation of glycosylation could be a mechanism used by complex organisms to compete with microbes on an evolutionary scale. Med Hypotheses 2009;73:510–2. Li XY, Hooft van Huijsduijnen R, Mantovani R, Benoist C, Mathis D. Intron–exon organization of the NF-Y genes. Tissue-specific splicing modifies an activation domain. J Biol Chem 1992;267:8984–90. Malagolini N, Santini D, Chiricolo M, Dall’Olio F. Biosynthesis and expression of the Sda and sialyl Lewis x antigens in normal and cancer colon. Glycobiology 2007;17:688–97. Mare L, Trinchera M. Suppression of ␤1,3galactosyltransferase ␤3Gal-T5 in cancer cells reduces sialyl-Lewis a and enhances poly N-acetyllactosamines and sialylLewis x on O-glycans. Eur J Biochem 2004;271:186–94. Mare L, Trinchera M. Comparative analysis of retroviral and native promoters driving expression of beta1,3-galactosyltransferase beta3Gal-T5 in human and mouse tissues. J Biol Chem 2007;282:49–57.

McGarvey M, Fahrner JA, Greene E, Martens J, Jenuwein T, Baylin SB. Silenced tumor suppressor genes reactivated by DNA demethylation do not return to a fully euchromatic chromatin state. Cancer Res 2006;66:3541–9. Miyazaki K, Ohmori K, Izawa M, Koike T, Kumamoto K, Furukawa K, et al. Loss of disialyl Lewisa , the ligand for lymphocyte inhibitory receptor sialic acidbinding immunoglobulin-like lectin-7 (Siglec-7) associated with increased sialyl Lewis(a) expression on human colon cancers. Cancer Res 2004;64: 4498–505. Murr R. Interplay between different epigenetic modifications and mechanisms. Adv Genet 2010;70:101–41. Oetke C, Hinderlich S, Reutter W, Pawlita M. Epigenetically mediated loss of UDPGlcNAc 2-epimerase/ManNAc kinase expression in hyposialylated cell lines. Biochem Biophys Res Commun 2003;308:892–8. Salvini R, Bardoni A, Valli M, Trinchera M. beta 1,3-Galactosyltransferase beta 3Gal-T5 acts on the GlcNAcbeta 1→3Galbeta 1→4GlcNAcbeta 1→R sugar chains of carcinoembryonic antigen and other N-linked glycoproteins and is down-regulated in colon adenocarcinomas. J Biol Chem 2001;276: 3564–73. Serpa J, Mesquita P, Mendes N, Oliveira C, Almeida R, Santos-Silva F, et al. Expression of Le(a) in gastric cancer cell lines depends on FUT3 expression regulated by promoter methylation. Cancer Lett 2006;242:191–7. Tong WG, Wierda WG, Lin E, Kuang SQ, Bekele BN, Estrov Z, et al. Genome-wide DNA methylation profiling of chronic lymphocytic leukemia allows identification of epigenetically repressed molecular pathways with clinical impact. Epigenetics 2010;5:499–508. Trinchera M, Malagolini N, Chiricolo M, Santini D, Minni F, Caretti A, et al. The biosynthesis of the selectin-ligand sialyl Lewis x in colorectal cancer tissues is regulated by fucosyltransferase VI and can be inhibited by an RNA interference-based approach. Int J Biochem Cell Biol 2011;43:130–9. Wang HR, Hsieh Y, Twu YC, Yu LC. Expression of the human Sda ␤-1,4N-acetylgalactosaminyltransferase II gene is dependent on the promoter methylation status. Glycobiology 2008;18:104–13. Yamashita K, Watanabe M. Clinical significance of tumor markers and an emerging perspective on colorectal cancer. Cancer Sci 2009;100:195–9. Yusa A, Miyazaki K, Kimura N, Izawa M, Kannagi R. Epigenetic silencing of the sulfate transporter gene DTDST induces sialyl Lewisx expression and accelerates proliferation of colon cancer cells. Cancer Res 2010;70:4064–73.