Introduction

During gestation, the fetus may be exposed to maternally derived toxicants including drugs which may pose a threat to the normal fetal development. The placenta and the fetal liver provide the primary fetal defense against toxicants. In general phase II conjugation capacities differ considerably between the adult and fetal/neonatal stages. A number of studies have shown that glucuronidation activity is significantly lower in fetal than adult human liver1, 2, 3 and therefore it is generally believed that sulfonation would compensate for that in the fetus. The role of fetal sulfoconjugation of xenobiotics, including drugs, is less investigated but has been demonstrated with 2-naphtol4 and ritodrine5 as substrates.

The human cytosolic sulfotransferases (SULT) are a large super family of enzymes that catalyze sulfate conjugation by transferring a sulfuryl (SO3) from the donor molecule 3′-phosphoadenosin 5′-phosphosulfat. To date, 12 human SULT genes have been identified. Among these SULT2A1 (also known as DHEA ST) is highly abundant in human livers and adrenals and is primarily associated with sulfate conjugation of dehydroepiandrosterone (DHEA).6, 7 DHEA sulfate (DHEAS) is the most abundant circulating steroid in humans being a precursor in the sex hormone synthesis. In adults, DHEAS is metabolized to more potent androgens in peripheral tissues expressing sterol sulfatases which are capable of reversing the sulfonation reaction8 and converting DHEA to androstenedione and testosterone. During fetal development, DHEAS from the fetal adrenals serves as the primary precursor for estrogen biosynthesis in the placenta.9 SULT2A1 is also responsible for sulfonation of other hydroxysteroids such as pregnenolone, cholesterol and bile acids.10, 11

There is a large inter–individual variation in SULT2A1 activity in human fetuses and in adults.12, 13, 14 Such a variation may be ascribed to genetic single-nucleotide polymorphisms or copy number variation (CNV) in the metabolic enzyme encoding genes. Several CNV in phase II enzymes that alter the conjugation capacity, have been characterized in e.g. UGT2B17, UGT2B28, GSTM and SULT2A1. A functionally important SULT2A1 CNV was recently described15 and subsequently shown to correlate with the excretion rate of DHEAS.16 However, whether this CNV alters the gene expression of SULT2A1 has not been investigated.

The aim of this study was to compare the gene expression of SULT2A1 in relation to the SULT2A1 CNV polymorphism in adult and fetal livers and in fetal lungs, adrenals and kidneys.

Materials and methods

Tissue samples

Human adult female (n=10) and male (n=9) liver samples or samples of unknown sex (n=1) were collected from patients between 30 and 75 years of age as surplus after diagnostic pathology. The clinical background of the donors included cancer (n=4), familial amyloidic polyneuropathy (n=6), hemorrhage (n=8), asystole (n=1) or unknown cause (n=1). The material was obtained from our liver bank at the Division of Clinical Pharmacology.

Samples from human fetal liver (n=60), adrenals (n=46), kidneys (n=43) and lungs (n=37) were obtained at legal abortions performed for socio-medical reasons at Karolinska University Hospital. The gestational ages were determined by crown-rump length and ranged from 5–12 weeks (median=10.2). None of the women reported any chronic or acute diseases, drug abuse or regular drug use. Maternal smoking was reported in 22 women, non-smoking in 21 whereas for 17 there was no information on smoking.

The study was approved by the National Board of Health and Welfare (51-9171/96) and the Ethics Review Board in Stockholm (fetus: DNR964/23, adult: DNR429/01).

SULT2A1 mRNA expression

Total RNA from 5–30 mg of fetal tissues samples and ~200 mg of adult's liver tissues was prepared using Allprep DNA/RNA and RNAeasy kit (Qiagen), respectively. RNA (0.5 μg) was reverse transcribed into cDNA with hexamer primer using Superscript III (Invitrogen) according to the manufacturer’s protocol and diluted ten times. The mRNA of SULT2A1 was determined in tissues from adult liver, as well as fetal liver, adrenals, kidneys and lungs by quantitative real-time polymerase chain reaction (Q-PCR) using 7500 Fast Real-Time PCR System, 7500 software v2.0.6 (Applied Biosystems, Foster City, CA, USA). The reactions were performed in triplicates containing TaqMan SULT2A1 expression assays (Hs00234219_m1, Life Technologies, Carlsbad, CA, USA), TaqMan 2x Universal PCR master mix (Applied Biosystems). The amount of cDNA in each well originated from 10 ng total RNA. Non-template controls were included in each experiment. Thermal cycling included activation (95 °C, 10 min) and was followed by 40 cycles of denaturation (95 °C, 15 s) and annealing/elongation (60 °C, 1 min). The mRNA expression of 18 S was used as reference gene. The relative expression was calculated using the ΔΔCT-formula as described by Livak and Schmittgen.17 For comparison of adult and fetal liver expression, one adult liver sample was used as a calibrator and for comparison of expression in different fetal liver tissues one fetal liver sample was chosen as a calibrator.

CNV genotyping

CNV for SULT2A1 was determined using genomic DNA prepared from fetal liver samples from the same cohort by Q-PCR using 7500 Fast Real-Time PCR System, 7500 software v2.0.6. Reactions (15 μl) were performed in triplicates containing TaqMan copy number assays (SULT2A1: Hs03013147_cn), TaqMan 2x Universal PCR master mix (Applied Biosystems) and 20 ng DNA per well. Albumin was chosen as a reference gene.18 The PCR setting was the same as in previous amplifications. The number of gene copies was approximated from the 2-ΔΔCT-values and one fetal liver sample was used as calibrator.

Statistical analysis

To investigate if there was any significant difference in relative hepatic expression of SULT2A1 between two groups (adult—fetal, male—female and smoking—non-smoking) the nonparametric Mann–Whitney test was used because the measurements were few and could not be proven to be normally distributed. When looking for significant difference in tissue-specific (liver, kidney, adrenal) expression of SULT2A1 in fetuses and in CNV (1, 2 or 3 copies) versus the relative expression of SULT2A1 in fetuses the nonparametric Kruskal–Wallis test was applied followed by Dunn’s multiple comparison post-test for identification of which groups differed from the other. The correlation analysis was performed by Spearman rank method. The results are presented as mean±s.d. if not stated otherwise. All statistical tests were performed using GraphPad Prism v. 4.00 (GraphPad Software, San Diego, CA, USA) and values of P<0.05 were considered statistically significant.

Results

Hepatic mRNA expression in human adults and fetuses

There was a large inter–individual variation in SULT2A1 mRNA expression in the adult liver samples (400-fold) compared with fetal liver samples (8.5-fold), while there was no significant difference in mean expression between adults (1.78±3.60) and fetal (0.60±0.28) liver samples (P=0.53, Figure 1a).

Figure 1
figure 1

Relative mRNA expression of SULT2A1 in (a) human adult (n=20) and fetal liver samples (n=60), (b) human adult female (n=10) and male liver samples (n=9) and in (c) fetal liver samples obtained from smoking (n=22) and non-smoking (n=21) women. No significant differences in SULT2A1 mRNA expression were found. The boxes represent median values and 95% confidence intervals and the lines represent the minimum and maximum values on a logarithmic scale.

PowerPoint slide

Full size image

Also there was no significant difference in mean expression between the liver samples of different sexes (P=0.5414, Figure 1b). The spread of relative mRNA values was larger in the female samples (400-fold) than in male samples (8.7-fold). There was no apparent effect of maternal smoking on the gene expression of SULT2A1 in fetal liver (P=0.51, Figure 1c).

There was no correlation between the fetal gestational age and the SULT2A1 liver mRNA expression (r=−19, P=0.18).

SULT2A1 gene expression in extra-hepatic fetal fetuses

In addition to liver samples, the SULT2A1 specific mRNA was also analyzed in fetal kidneys, adrenals and lungs. In liver and adrenals samples, 98% of the fetuses expressed SULT2A1, whereas SULT2A1 mRNA was detectable only in 65% of the fetal kidney samples. There was no SULT2A1 expression found in any of the lung samples.

The mean SULT2A1 mRNA levels were different in the tissues investigated being most abundant in the adrenals (mean 16.91±11.86) followed by the kidney (mean 5.52±4.39) and the livers (mean 2.56±4.39; P=0.01; Figure 2).

Figure 2
figure 2

Relative gene expression of SULT2A1 in human fetal liver (n=59), adrenals (n=45) and kidneys (n=28). mRNA expression was found to be highest in adrenal and lowest in liver tissue. The bars represent mean values and s.d. ***P<0.001, **P<0.01.

PowerPoint slide

Full size image

A significant correlation between SULT2A1 mRNA in adrenals and livers was observed (n=44, r=0.36, P=0.02), whereas there was no correlations found between the other tissues.

CNV in the SULT2A1 gene

SULT2A1 CNV distribution in the fetuses was dominated by samples with two copies (76%, n=45). Nine samples had one copy (15%) and five samples (8.5%) had three or more copies. A significant correlation between SULT2A1 gene expression in fetal liver and the copy number was found. There was a significant difference in mRNA expression between carriers of one copy (mean 1.07±1.72) and ⩾3 copy carriers (mean 4.16±2.83; P=0.01; Figure 3a).

Figure 3
figure 3

Relative expression of SULT2A1 in (a) fetal liver (n=58), (b) adrenals (n=45) and (c) kidneys (n=28) in relation to copy numbers of SULT2A1. A significant correlation between SULT2A1 hepatic mRNA expression and gene copy number was found in liver. The bars represent mean values and s.d. *P<0.05.

PowerPoint slide

Full size image

In contrast to the liver, no significant correlation between CNV and SULT2A1 mRNA expression was found in fetal adrenals (P=0.06) and kidneys (P=0.93; Figure 3b and c), even though a similar trend was discerned in the adrenals.

Discussion

Here we show for the first time that SULT2A1 mRNA is abundant in human fetal tissues in the first trimester and that there is conspicuous tissue specificity in the gene expression. SULT2A1 transcripts were found in all fetal organs investigated except lungs, being most abundant in the adrenals. This is consistent with the contention that the enzyme is of paramount importance for conjugation of DHEA to DHEAS which then is a precursor of the important estrogen formation in the placenta.19 The fetal-placental-maternal dependence upon a normal function in each partner is well appreciated since long. Our findings are supported by an earlier study demonstrating the highest SULT2A1 activity in fetal adrenals obtained in the second trimester.20

We found no significant difference in hepatic SULT2A1 mRNA expression between adult and fetal liver specimens. This is unique as most other studies published so far on phase II enzymes have demonstrated low activities compared with adults, or negligible or unmeasurable fetal enzyme activities. In fact, we expected relatively more hepatic SULT2A1 mRNA expression in adults because the sulfonation of DHEAS has been shown to be of larger magnitude (sixfold) in adults,14 but no such significant difference was noted. The reason is not clear. It is possible that post-transcriptional factor(s) may affect the activity in a positive direction in adults. Moreover, our study was conducted in fetuses derived from the first trimester and not later in gestation. It is possible that the SULT2A1 gene expression increases during pregnancy. In fact, the SULT2A1 protein expression and activity is increasing with gestational age and continuous to increase after birth.21, 22 Taken together, our data signify the specific role of SULT2A1 in the important intermediary metabolism of steroids in human ontogeny. The abundant expression of SULT2A1 during the first trimester indicates that this enzyme is important during early development. Although the production of DHEAS for estrogen formation is the main function of SULT2A1 in the fetus, it is possible that SULT2A1 exerts other physiological functions of importance during pregnancy.

O’Shaughnessy et al.23 found higher hepatic SULT2A1 mRNA concentrations in male fetuses of smoking mothers with abortion in the second trimester. Our results did not show any significant difference in fetal SULT2A1 mRNA expression between smoking and non-smoking mothers. The reason could be because of the difference in gestational age between the studies. Alternatively, the fetal sex was not determined in our study and therefore any potential influence of sex could not be correlated to smoking exposure.

In women, a large spread of the relative mRNA values was observed which could indicate that female hormonal profiles of estrogens and/or progesterone may have an influence on SULT2A1 transcription rate. Our study population was too small to further study the SULT2A1 expression profile during the menstrual cycle or whether any inter–individual variation is present in menopausal women. Estrogenic hormone receptors have been reported to regulate SULT2A1 transcription in vitro, and an estrogen receptor element has been identified.24, 25 However, the mechanisms controlling transcriptional regulation of the SULT2A1 gene in vivo are still largely unknown.

The inter–individual variation observed in the SULT2A1 gene expression may certainly partly be ascribed to genetic variation. Indeed, the CNV of SULT2A1 correlated with the relative mRNA level in fetal liver, and a similar trend of correlation was seen in adrenals. In a previous study we have demonstrated that individuals expressing three or more SULT2A1 gene copies excrete DHEAS at a higher rate than individuals with only one copy.16 The studied CNV results in a deletion/insertion of a non-coding 2849 bp sequence. Although it alters the mRNA expression it is likely that this polymorphism has a regulatory function. There was no significant correlation between CNV and mRNA expression in fetal kidneys and adrenals indicating that genotype correlation is tissue specific. The reason for this lack of correlation may be that some SULT2A1 trans-acting factors are not abundant in the kidneys. The fact that there was a strong correlation between hepatic and adrenal SULT2A1 expression further support co-regulatory mechanisms in these tissues. We have previously seen that a CNV in UGT2B17, another phase II enzyme, also exhibits a tissue-specific correlation with mRNA expression in the fetus.26

The studies of drug metabolizing enzymes in the human fetus may be a useful tool for predicting the metabolism and disposition of drugs in neonates. Such examples have been provided with acetaminophen and morphine as drug substrates. In adult life the major metabolizing pathway is via glucuronidation and very little is sulfated. However, acetaminophen was sulfated only whereas glucuronidation was absent.27 Conclusive data show that fetal livers have lower glucuronidation capacity than adult livers for many substrates.1, 2, 3, 4 Other substrates such as dopamine and ritodrine are mainly catalyzed by SULT1A3/A4 (also known as M-PST) which also appears to have a higher sulfonation rate in the human fetal than adult livers.5, 28 However, the opposite has also been described for the activity of SULT1A2 (P-PST) which was found to be higher in adults than in the fetal livers.28

For obvious reasons a limitation with our study is the small amount of tissue specimens available from the fetuses in the age range studied. Most of the material was utilized for RNA/DNA preparations. Because of this no enzyme protein or SULT-activity measurements were possible to investigate. However, the inclusion in our study of different tissues obtained from 60 fetuses for gene expression analysis is unique. Other limitations with our study include lack of information about the sex of the fetuses. Minor differences in collection procedure, such as time between removal of samples and freezing are also cofactors.

For the first time we have shown that there is no difference in hepatic SULT2A1 gene expression between adults and fetuses in the first trimester. Moreover, we here demonstrate that fetal SULT2A1 mRNA is present at highest levels in the adrenals, and absent in lungs. A newly identified CNV in SULT2A1 was correlated to hepatic gene expression in the fetuses, but this correlation was not found in other tissues, indicating a tissue-specific SULT2A1 gene regulation.