Background
Treatment of estrogen receptor- (ER-) positive breast cancer with tamoxifen (TAM) has been the gold standard for the past 30 years [
1]. TAM is also the only FDA-approved breast cancer chemoprevention agent, as its use is associated with decreased occurrence of contralateral breast cancer in patients treated with adjuvant TAM [
2]. Both the therapeutic efficacy and adverse effects of TAM vary considerably among individuals [
3]. In terms of therapeutic efficacy, 4-hydroxy-TAM and endoxifen are the major active metabolites of TAM. The detoxification of 4-hydroxy-TAM is carried out by phase II enzymes such as sulfotransferases, including SULT1A1 and uridine diphosphate glucuronosyltransferase (UGTs) [
4].
Sulfotransferases catalyze the transfer of a sulfonate group from 3′-phosphoadenosine 5′ -phosphosulfate (PAPS) to a range of xenobiotic and endogenous compounds such as drugs, carcinogens, and steroid hormones [
5]. The human SULT1A subfamily consists of SULT1A1, SULT1A2, SULT1A3, and SULT1A4. SULT1A1 protein is found in many different human tissues, with the highest abundance found in the liver [
6‐
11]. SULT1A1 activity varies several-fold among individuals [
12‐
14]. It has been shown that gene expression/protein activity levels of SULT1A1 affect the efficacy of TAM treatment [
15‐
18]. SULT1A1 can also contribute to increased cancer risk (as reviewed in [
19]), including breast cancer risk [
20‐
22]. Furthermore, SULT1A1 expression is related to disease state, with little to no expression in normal breast epithelia but plentiful protein expression in most breast tumors [
23,
24]. Given the role that SULT1A1 plays in drug efficacy and in individual susceptibility to disease, it is important to elucidate the factors regulating the differential expression of SULT1A1.
Some studies have shown that single nucleotide polymorphisms (SNPs) in the human
SULT1A1 promoter, 3′-untranslated region (UTR), and coding regions contribute to SULT1A1 availability and activity [
13,
25‐
27]. However, SNPs account for only a small percentage of the variation of SULT1A1 activity. Some studies demonstrated that
SULT1A1 gene copy number variants (CNV) exist in human populations, and that CNV is associated with SULT1A1 activity [
13,
28,
29]. Nevertheless, doubling
SULT1A1 copy number does not appear to double its activity, indicating that factors other than copy number are determinants of SULT1A1 activity. Thus, other factors must play a role in regulating the expression of SULT1A1. Traditionally, SULT1A1 has been considered a non-inducible enzyme. Since its expression is variable across tissues, it could be postulated that transcription factor (TF) binding (and the tissue-specific availability of TFs) may affect gene expression in SULT1A1. There have been reports of TF regulation of SULT1A1. There is one report that demonstrated that
SULT1A1 promoter activity is dependent on the presence of ubiquitous Ets family transcription factors [
30]. While these TFs appear to be responsible for basal, constitutive expression, TFs that affect differential SULT1A1 gene expression have not been examined. In this study, TFs differentially expressed between low SULT1A1-expressing transformed epithelial mammary cells and high SULT1A1-expressing breast cancer cells have been identified for the first time using a TF Activation Profiling Plate Array assay.
Methods
Cell culture
The human breast cancer cell lines ZR-75-1, MCF-7, T-47D, MDA-MB-231, and the human transformed mammary epithelial cell line MCF-10A were purchased from American Type Culture Collection (Rockville, MD) and were sustained in a 37°C incubator containing 5% CO2. ZR-75-1 cells were cultured in RPMI 1640 medium with 10% fetal bovine serum (FBS). MCF-7 cells were cultured in Improved MEM medium with 10% FBS and 0.01 mg/ml insulin (Life Technologies, Carlsbad, CA). MDA-MB-231 cells were cultured in DMEM medium with 10% FBS. MCF-10A cells were cultured in DMEM/F12 medium supplemented with 5% chelex-treated horse serum (Life Technologies), 20 mg/ml of epidermal growth factor (Life Technologies), 10 μg/ml insulin (Sigma, St. Louis, MO), 0.5 g/ml hydrocortisone (Sigma), and 0.1 μg/ml cholera toxin (Sigma).
Human subjects
Normal human liver samples (n = 84) were obtained from the U.S. Cooperative Human Tissue Network under a University of Arkansas for Medical Sciences Institutional Research Committee-approved protocol. All specimens were snap-frozen upon collection. Normal tissue status was confirmed with histology by the Cooperative Human Tissue Network. Samples included 33 women and 51 men, all of whom were Caucasian. The average age of the study population was 59, ranging from 26 to 102.
Differential transcription factor activation assay
The differential TF activation profile in ZR-75-1 and MCF-10A cells was determined using a TF Activation Profiling Plate Array assay following the manufacturer’s instructions (Signosis, Sunnyvale, CA). Briefly, 5 μg of nuclear extract from ZR-75-1 or MCF-10A cells was hybridized with 48 different biotin-labeled probes that contain consensus sequences of TF DNA-binding sites. The TF-bound probes were captured by complementary sequences of the probes pre-coated on a 96-well plate. The captured probes were detected with streptavidin-HRP. The levels of luminescence of each corresponding well were quantitatively analyzed with Spectramax M5 (Sunnyvale, CA) and compared between MCF-10A cells and ZR-75-1 cells.
Nuclear factor I (NF-I) siRNA treatment of ZR-75-1 cells
All siRNAs and transfection reagents were purchased from Life Technologies. ZR-75-1 cells that were at the exponential growth stage were transfected with 50 pmol of pre-designed siRNA against 4 NFI family members (NFI-A, NFI-B, NFI-C, NFI-X) or 100pmol siGAPDH as a negative control using Lipofectamine™ 2000 transfection reagent following the manufacturer’s instructions. After 48 h transfection, cells were collected for total RNA isolation. The percentage of knockdown of target gene expression was determined using quantitative RT-PCR (qRT-PCR). The sequences of siRNAs are listed in Table
1.
18S | 5′ TTCGAACGTCTGCCCTATCAA 3′ | 5′ ATGGTAGGCACGGCGACTA 3′ |
GAPDH | 5′ ACAGTCAGCCGCATCTTCTT 3′ | 5′ ACGACCAAATCCGTTGACTC 3′ |
NFI-A | 5′ CCAGCGCCCGGCAGTTATGT 3′ | 5′ ATTCATCCTGGGTGAGACAGAGCGG 3′ |
NFI-B | 5′ AACCAGCCAGCCTAACGGCA 3′ | 5′ TCGCACTGCACTGGGATGGG 3′ |
NFI-C | 5′ GACATGGAAGGAGGCATCTC 3′ | 5′ GGGCTGTTGAATGGTGACTT 3′ |
NFI-X | 5′ CCACTGCCCAACGGGCACTTA 3′ | 5′ CCGTCACATTCCAGACCCCGGA 3′ |
SULT1A1 | 5′ AGGAGTTCATGGACCACAGC3′ | 5′ TGAAGGTGGTCTTCCAGTCC3′ |
siNFI-A | 5′ GGUAUUCCGCUGGAAAGUAtt 3′ | |
siNFI-B | 5′ AGUGUCAUCUCAACUCGAAtt 3′ | |
siNFI-C | 5′ GGACAGGGCGUCUUCCUAAtt 3′ | |
siNFI-X | 5′ GAAUCCGGACAAUCAGAUAtt 3′ | |
DNA and RNA isolation and quantitative real-time PCR (qRT-PCR)
Total RNA and total DNA from cultured cells and from snap-frozen human liver tissues were isolated with an AllPrep
® DNA/RNA/protein kit (QIAGEN, Valencia, CA) according to the manufacturer’s instructions. The quantity and quality of the isolated RNA and DNA were determined by an Agilent 2100 Bioanalyzer (Agilent Technology, Palo Alto, CA). qRT-PCR using SYBR-green reagent was performed and analyzed as described previously[
31]. qRT-PCR using TaqMan
® reagents (Life Technologies) was carried out as detailed previously [
13]. The primer sequences used with SYBR-green reagent are listed in Table
1.
Determination of SULT1A1 gene copy number
SULT1A1 gene was amplified by relative quantitative PCR using a pre-designed Custom Plus TaqMan® Copy Number kit following the company’s instructions (Life Technologies). SULT1A1 copy number variations were calculated using the Applied Biosystems CopyCaller™ software (Life Technologies).
Statistical analysis
Paired t tests were used to compare baseline and treatment measurements within a group. Pearson’s correlation coefficients were used to describe the linear association between variables. All data from samples were expressed as mean ± sem.
Discussion
SULT1A1 is expressed in different tissues, in different physiological states of the same tissue, and in different individuals. The factors that regulate SULT1A1 gene expression are poorly understood. Human studies show that SULT1A1 can be regulated by alternative promoter usage [
32], SNPs in the coding or promoter region [
26,
27], CNVs [
28], or variants in the 3′UTR of the gene [
13]. However, these genetic variants account for only a portion of the variation of SULT1A1 activity. Hempel
et al. [
30] identified Ets synergized with Sp1 as one of the regulators of SULT1A1 expression. It is postulated that ubiquitous Ets is utilized to ensure constant expression of SULT1A1 in tissues such as liver, skin, and gut, since SULT1A1 is important in xenobiotic metabolism. This current study elucidated NFI gene family as an important TF in regulating SULT1A1 expression in different cell types.
The data presented here suggest that NFI may play a major role in regulating SULT1A1 expression in different physiological and disease states. The NFI family proteins are associated with changes in different cell growth states and with oncogenic processes and disease states as reviewed in [
33‐
35]. TF profiling array data demonstrated that NFI in high-SULT1A1-expressing ZR-75-1 breast cells was expressed more than in low-SULT1A1-expressing non-cancer MCF-10A cells. This positive association was also observed in human liver samples and in different human cell lines. Further siRNA transfection assays suggested that SULT1A1 expression is controlled, at least partially, by NFI in breast cancer cells. While the siRNA data demonstrate NFI-B has the least effect on SULT1A1 expression, there is still an effect and the p-value approached significance at p = 0.07, while the NFI isoforms with significant effects were at p = 0.05. This effect may not be a direct one, and NFI-B may not be the only factor which has an effect on SULT1A1 expression. Rather it may be part of another mechanism that has a compound effect on SULT1A1 expression. This could explain the lower significance of NFI-B. There was no additive SULT1A1 expression inhibition observed when ZR-75-1 cells were co-transfected with siNFI-A, siNFI-B, and siNFI-C, suggesting that each gene has its own role in regulating SULT1A1 expression.
NFI siRNA knockdown experiments inhibited about 40% of SULT1A1 expression in ZR-75-1 cells, suggesting other regulatory mechanism(s) exist. Previous studies have shown that the human population possesses one to five copies of
SULT1A1, and that its enzyme activity is correlated with CNV [
13,
14,
28]. The data presented here demonstrate that there were
SULT1A1 copy number differences in human cell lines also, and that cells with more copies had higher SULT1A1 expression than cells with only one copy. However, higher copy number does not always correlate with the degree of expression. T-47D cells, which contained five copies, had similar SULT1A1 expression to that of ZR-75-1 cells, which contained three copies. MCF-7 and MDA-MB-231cells each contained one copy, yet MDA-MB-231 cells showed no SULT1A1 expression, while MCF-7 cells showed a modest amount of SULT1A1 expression. The data suggests that SULT1A1 gene expression could also be regulated by mechanisms other than NFI and copy numbers.
SULT1A1 CNV in different cell lines reflect the range of
SULT1A1 CNV reported in human populations [
13,
14,
28]. Thus, cell lines could be used as a model for human
SULT1A1 CNV-related studies.
Conclusions
In summary, our data demonstrate that the NFI family of transcription factors significantly contributes to the regulation of SULT1A1 expression in human breast cancer cell lines. Along with these TFs, SULT1A1 CNV determines gene expression. Understanding the regulation of SULT1A1 expression will facilitate the prediction of drug response in relation to breast cancer prevention and therapy.
Competing interests
The authors declare that they have no competing interests.
Authors’ contributions
SAK and AY-B participated in the study design. AY-B, LJR, RBP, XY, VKE, and SW carried out data collection. AY-B, RJP, VKE, IBD, SAK contributed in data analysis and interpretation. SAK, AY-B, RBP, and VKE were involved in manuscript preparation. All authors have read and approved the final manuscript.