Maternal Smoking during Pregnancy Specifically Reduces Human ...

ORIGINAL E n d o c r i n e

ARTICLE R e s e a r c h

Maternal Smoking during Pregnancy Specifically Reduces Human Fetal Desert Hedgehog Gene Expression during Testis Development Paul A. Fowler, Sarah Cassie, Stewart M. Rhind, Mark J. Brewer, J. Martin Collinson, Richard G. Lea, Paul J. Baker, Siladitya Bhattacharya, and Peter J. O’Shaughnessy Departments of Obstetrics and Gynaecology (P.A.F., S.C., S.B.) and Biomedical Sciences (J.M.C.), Institute of Medical Sciences, University of Aberdeen, Aberdeen AB25 2ZD, United Kingdom; Biomathematics and Statistics Scotland (M.J.B.), Animal Ecology (S.M.R.), Macaulay Institute, Aberdeen AB15 8QH, United Kingdom; School of Veterinary Medicine and Sciences (R.G.L.), University of Nottingham, Loughborough LE12 5RD, United Kingdom; and Department of Veterinary Preclinical Studies (P.J.B., P.J.O.), Division of Veterinary Physiology and Pharmacology, University of Glasgow Veterinary School, Glasgow G61 1QH, United Kingdom

Context: Maternal cigarette smoking during gestation increases cryptorchidism and hypospadias and reduces testis size and fertility in sons by unknown mechanisms. Objective: The objective of the study was to determine whether maternal smoking is linked with changes in male human fetal endocrinology, testis gene expression, and liver concentrations of cigarette smoke chemicals. Design: This was an observational study of the male fetus, comparing pregnancies during which the mothers either did or did not smoke. Setting: The study was conducted at the universities of Aberdeen, Glasgow, and Nottingham and Macaulay Institute (Aberdeen). Patients/Participants: Testes, blood, and livers were collected from 69 morphologically normal human male fetuses of women undergoing elective termination of normal second-trimester pregnancies. Main Outcome Measures: Testosterone, human chorionic gonadotropin, LH, and cotinine; expression of 30 reproductive/developmental genes; liver concentrations of 16 polycyclic aromatic hydrocarbons; and Leydig, Sertoli. and germ cell numbers were determined. Results: There were no significant differences in fetal size, testis weight, cell numbers, seminiferous tubule diameter, or circulating LH and testosterone. Fetuses from smoking mothers had smoking range cotinine levels and liver concentrations of polycyclic aromatic hydrocarbons that were significant predictors of maternal smoking (P ⬍ 0.001). Only the Sertoli cell-specific gene, desert hedgehog (DHH), was significantly altered by maternal smoking (reduced 1.8-fold, P ⫽ 0.013). Conclusions: The consequences of reduced DHH signaling in men and mice are consistent with epidemiology for effects of gestational maternal smoking on sons. Given the absence of other observed effects of maternal smoking, we concluded that reduced DHH is part of a mechanism linking maternal gestational smoking with impaired reproductive development in male offspring. (J Clin Endocrinol Metab 93: 619 – 626, 2008)

aternal cigarette smoking during gestation has been increasingly associated with rising incidences of cryptorchidism (1– 4) and hypospadias (5) and reductions in testis

M

size (6), sperm counts/quality (7–9), and fertility (2, 10, 11) in male offspring. Although not all studies have replicated all these findings (12, 13), the rising incidences of these reproductive

0021-972X/08/$15.00/0

Abbreviations: AhR, Aryl hydrocarbon receptor; DHH, desert hedgehog; hCG, human chorionic gonadotropin; PAH, polycyclic aromatic hydrocarbon; PTC1, patched 1; qPCR, quantitative PCR; SF1, steroidogenic factor 1.

Printed in U.S.A. Copyright © 2008 by The Endocrine Society doi: 10.1210/jc.2007-1860 Received August 20, 2007. Accepted November 7, 2007. First Published Online November 13, 2007

J Clin Endocrinol Metab, February 2008, 93(2):619 – 626

jcem.endojournals.org

619

620

Fowler et al.

Smoking and Fetal Testicular Gene Expression

problems are associated with testis dysgenesis syndrome (14 – 16), which has its origin during fetal development. The only component of testis dysgenesis syndrome that does not seem to be associated with gestational maternal smoking (17, 18) is testicular cancer (also increasing in incidence), although a strong geographical association exists between the two (19). Maternal smoking during gestation disturbs reproductively important developmental events, including testosterone production (20), the fetal hypothalamic-pituitary-adrenal axis (21), and neonatal prolactin and GH (22) levels. It should be noted that some of these data (e.g. Ref. 20) have not yet been replicated and, interestingly, adult sons of gestational smokers do not appear to have significantly disturbed endocrine profiles (23). However, gestational maternal smoking is also associated with many developmental problems, including impaired fetal growth (24, 25), which is largely assumed to be due to poor placental development; childhood respiratory ill health (26); and abnormalities of the nervous system and cognition (27) as well as possible increased mortality risk for male fetuses (28). Cigarette smoke contains a complex mix of chemicals that can affect fetal development, including metals (29), nicotine (30), and polycyclic aromatic hydrocarbons (PAHs) (31). The latter are a large class of toxic pollutants that are mutagenic (32) and cross the placenta, thereby increasing fetal susceptibility to DNA damage (33–35). Linking reduced fertility or increased incidences of reproductive developmental abnormalities with specific exposures and chemicals is particularly difficult in humans. Therefore, mechanisms underpinning the effects of gestational maternal smoking on the developing fetus remain very poorly understood. Around half of the women having elective second-trimester terminations of normally progressing pregnancies smoke through the pregnancy. We therefore combined ongoing studies of fetal testis development (e.g. Ref. 36) with the in utero exposure of fetuses to cigarette smoke chemicals to address serious gaps in current knowledge of the effects of environmental chemicals on fetal human tissues and expression of fetal testicular genes and associated fetal endocrinology.

Subjects and Methods Study design To test the hypothesis that maternal smoking affects the fetal testis in utero, all eligible women (see Sample collection) were recruited prospectively, with maternal and fetal morphological data recorded during consent and sample processing. In a subset of cases in which fetal cardiac blood was obtained, circulating hormones and cotinine were assayed and, in another subset of cases, the expression levels of key developmental genes were determined. Cases were matched for gestational stage between smoking and control (nonsmoking) groups.

Sample collection The collection of fetal material (detailed in Ref. 36) was approved by the National Health Service Grampian Research Ethics Committees (REC 04/S0802/21). Women seeking elective, medical terminations of pregnancy were recruited with full written, informed consent by nurses working independently at Aberdeen Pregnancy Counseling Service. There was no change in patient treatment or care, and women were able


to withdraw from the study at any point. Only normally progressing pregnancies (determined at ultrasound scan) from women older than 16 yr of age with a good grasp of English and between 11 and 21 wk of gestation were collected. Maternal morphological data and numbers of cigarettes smoked per day during pregnancy were recorded. Fetuses were transported to the laboratory within 30 min of delivery, weighed, crownrump length recorded, and sexed. Where possible, an ex vivo blood sample was collected by cardiac puncture and the plasma stored at ⫺20 C. The gonads were weighed and either snap frozen in liquid nitrogen and stored at ⫺85 C or fixed for 5.5 h in Bouins, transferred to 70% ethanol, and processed for histology (37, 38). The fetuses used for quantitative PCR (qPCR) and stereology are the same as those reported in the serial study of gene expression changes across the second trimester reported by O’Shaughnessy et al. (36). The present study focused on the effects of maternal smoking rather than expression changes across the second trimester and, in addition, also reports on the fetal testis gene expression of desert hedgehog (DHH), steroid acute regulatory protein, androgen receptor, and P450 oxidoreductase.

Endocrine and cotinine assays Testosterone, a key marker of masculinization, was determined using a DELFIA kit (PerkinElmer Life Sciences, Cambridge, UK) with a detection limit of 0.3 nmol/liter and intra- and interassay coefficients of variation of 3.2 and 8.6%, respectively. Because of potential interference with the testosterone assay owing to the many hemolyzed samples, all 34 fetal plasma samples available were extracted on C18 spin columns (GE Healthcare Bucks, UK; C18 Self Pack POROS 10 R2; Applied Biosystems Ltd., Warrington, UK). Extracted samples (see supplemental data, published as supplemental data on The Endocrine Society’s Journals Online Web site at http://jcem.endojournals.org) were reconstituted in assay buffer (PerkinElmer Life Sciences) and used for testosterone assay. The recovery of testosterone was 122 ⫾ 9.7%. LH and intact human chorionic gonadotropin (hCG) were measured using DELFIA kits robust for hemolysis, with detection limits and inter- and intraassay coefficients of variation of 0.05 U LH/liter, 2.4 and 4.2%, and 0.5 U hCG per liter, 4.1 and 4.6%. There was a 0.02% cross-reaction of intact hCG in the LH assay. Cotinine, a metabolite of nicotine and a marker of smoking, was determined with a kit (Cozart Plc., Abingdon, Kent, UK) using a 25-ng cotinine per milliliter cutoff to maintain cross-reaction with nicotine at 0.6%. The determination was semiquantitative, based on calibrators at 0, 10, 25, and 50 ng cotinine per milliliter, with values between 0 and 12 ng cotinine per milliliter being considered negative.

Morphology, stereological determination of cell numbers, and immunohistochemistry For immunohistochemistry 5-␮m testis sections were stained with hematoxylin and eosin or mounted onto ChemMate slides (DakoCytomation Ltd., Ely, UK). Sections were dewaxed and endogenous peroxidase activity quenched (3% H2O2, PBS, 15 min). After blocking (PBS, 4% serum, 0.3% BSA, 2 h), primary antibody was added overnight at 4 C [anti-Dhh (N-17), 1:100 in blocking buffer; sc-33940; Santa Cruz Biotechnology, Santa Cruz, CA]. Slides were washed and secondary antibody (biotinylated rabbit antigoat IgG; B-7014; Sigma, Poole, UK. 1:200 in blocking buffer) added for 1 h, 22 C. After washing, color reaction was developed by 30 min incubation in Vectastain R.T.U. Elite ABC reagent (VectorLabs, Peterborough, UK) followed by 20 min in 800 ␮g/ml diaminobenzidine, 50 mM Tris (pH 7.6), 0.5% H2O2. Sections were washed in water and counterstained with hematoxylin. For stereology, testes were embedded in Technovit 7100 resin, cut into 20-␮m sections, and stained with Harris’ hematoxylin. Total testis volumes were estimated [Cavalieri principle (39)] from the same slides used to determine cell number. Leydig, Sertoli, and germ cell numbers were counted using the optical dissector technique (40). Sertoli and germ cells were identified by their distinctive nuclei and position within the tubule (41, 42), whereas the Leydig cells were identified by their interstitial location, round nuclei, and prominent nucleoli (41, 42). Total germ cell number


jcem.endojournals.org

was measured without categorization as gonocytes, intermediate cells, or prespermatogonia (43– 45). The numerical density of each cell type was estimated using an Olympus BX50 microscope fitted with a motorized stage (Prior Scientific Instruments, Cambridge, UK) and Stereologer software (Systems Planning Analysis, Alexandria, VA). Cross-sectional tubule diameters were measured (six fields of view per two sections per testis, ⫻20) using an Olympus BX41 with a ProgRes C5 digital camera and ProgRes CapturePro (Jentopik Laser Optic Systeme GmbH, Jena, Germany).

Real-time qPCR For quantification of 30 specific mRNA species in testes (see supplemental data), real-time PCR was used after reverse transcription of the isolated RNA. To allow specific mRNA levels to be expressed per testis and to control for RNA extraction efficiency, RNA degradation, and reverse transcription step, 5 ng external standard (luciferase mRNA; Promega UK, Southampton, UK) was added to each testis at the start of the RNA extraction (46). Testis RNA was extracted using TRIzol (Life Technologies, Paisley, UK), and the RNA was reverse transcribed using random hexamers and Moloney murine leukemia virus reverse transcriptase (Superscript II; Life Technologies) (47). The real-time PCR approach used the SYBR green method in a 96-well plate format using a MX3000 cycler (Stratagene, Amsterdam, The Netherlands). Reactions contained 5 ␮l 2⫻ SYBR master mix (Stratagene), primer (100 nM), and template in a total volume of 10 ␮l. At the end of the amplification phase, a melting curve analysis was carried out on the products formed. All primers were designed by Primer Express 2.0 (Applied Biosystems) using parameters previously described (46).

PAH determinations The fetal liver concentrations of 16 PAHs known to be in cigarette smoke (e.g. Ref. 48) were determined. Seven internal standards (see supplemental data) were added at 0.05 ␮g to freeze-dried livers and PAHs extracted into ethanoic potassium hydroxide (90 C, 8 h) followed by 3 ⫻ 10 ml isohexane and evaporation to 3 ml. After absorption chromatography to remove lipids, samples were loaded onto 10-g silica columns (Merck, Nottingham, UK) previously conditioned with 40 ml isohexane (Rathburn Chemicals, Walkerburn, UK). Fractions were eluted with 75 nM isohexane/dichloromethane (1:1 vol/vol) and concentrated under nitrogen. PAH content was determined using gas chromatography linked to mass spectrometry operated in the single ion monitoring mode (Trace MS; Thermo Electron, Hemel Hempstead, UK) linked to a Trace 2000

621

GC fitted with an AS2000 autosampler). Separations were effected on a Zebron ZB5 fused silica capillary column coated with 95% dimethylpolysiloxane/5% phenyl with a phase thickness of 0.25 mm (Phenomenex, Macclesfield, UK). The operating temperature for each PAH analysis started at 70 C for 3 min, was then ramped at 5 C/min to 250 C and held for 1 min, ramped to 300 C at 6 C/min and held for 6 min, and then ramped to 325 C at 10 C/min and held for 5 min. The carrier gas was helium and samples were injected in splitless mode with a surge. The mass spectrometer was operated in the electron ionization mode at 70 electron volts and a source temperature of 250 C. The ions monitored for each compound were between 128 and 278 m/z (see supplementary data). Response factors were calculated relative to the seven internal standards. Certified reference materials could not be found for sample matrix under investigation, and reproducibility was therefore monitored by repeated analysis of spiked liver tissue. The recovery values for PAH analysis were between 81 and 116% (see supplementary data).

Statistical analysis Analyses were performed using JMP (5.1; Thompson Learning, London, UK) and Minitab (14.12; Minitab UK Ltd., Coventry, UK). Normality of data distribution was tested with the Shapiro-Wilk test and nonnormally distributed data were log transformed and rechecked for normality before analysis. Care was taken to ensure that no comparison involved groups with any difference between smokers and nonsmokers in terms of stage of gestation, thus avoiding bias. Because most of the genes investigated showed changes in expression across the second trimester (36), two-way ANOVA was used to test the combined effects of gestational age (weeks) and maternal smoking (yes/no) on morphological and biochemical data and gene expression levels. To ensure that statistically significant differences were robust, these analyses were repeated with the weeks of gestation divided into 11–13, 14 –16, and 17–19 wk and less than 16 and more than 16 wk, and the statistical findings were the same. Relationships between variables were explored by linear discriminant analysis to classify observations, with a cross-validation procedure to avoid bias. We assessed the importance of genes and PAHs in their ability to classify mothers into smoking or nonsmoking categories via a simple test of a binomial proportion on the number of patients classified correctly; the null hypothesis of no information corresponds to a classification probability of one half for the binomial proportion, enabling us to obtain a P value for the actual number classified correctly. After initial data analysis, the DHH qPCR was repeated twice to confirm the findings, and the results for DHH are expressed as a mean of three

TABLE 1. Morphological and endocrine data for mothers and male fetuses (mean ⫾ SEM) Endocrine subseta

All fetuses n Maternal characteristics Age (yr) Body mass index (kg/m2) Cigarettes per day Fetal characteristics Age (wk) Weight (g) Crown-rump length (mm) Testis weight (g) Testosterone (nmol/liter) Intact hCG (U/liter) LH (U/liter) Cotinine (ng/ml)

Control

Smoker

34

35

23 ⫾ 1 25 ⫾ 1 0

26 ⫾ 2 26 ⫾ 1 12 ⫾ 1

14.2 ⫾ 0.4 66 ⫾ 6 91 ⫾ 2 14 ⫾ 1

14.9 ⫾ 0.4 87 ⫾ 6 101 ⫾ 2 19 ⫾ 1

P value

Control

Smoker

18

25

0.147 0.192

24 ⫾ 2 25 ⫾ 2 0

24 ⫾ 1 25 ⫾ 1 11 ⫾ 1

0.280 0.801 0.526 0.127

15.2 ⫾ 0.6 96 ⫾ 9 105 ⫾ 3 18 ⫾ 2 9.4 ⫾ 1.6 145 ⫾ 17 16 ⫾ 5 0.6 ⫾ 1.9

15.4 ⫾ 0.5 102 ⫾ 7 108 ⫾ 2 21 ⫾ 1 8.6 ⫾ 1.3 88 ⫾ 15 19 ⫾ 4 45.2 ⫾ 2.5

qPCR subseta P value

0.960 0.806

0.778 0.916 0.783 0.373 0.788 0.049 0.682