Neurotoxicant exposure during pregnancy is a confounder for assessment of iodine supplementation on neurodevelopment outcome

Neurotoxicology and Teratology 51 (2015) 45–51

Contents lists available at ScienceDirect

Neurotoxicology and Teratology journal homepage: www.elsevier.com/locate/neutera

Neurotoxicant exposure during pregnancy is a confounder for assessment of iodine supplementation on neurodevelopment outcome Françoise Brucker-Davis a,b,⁎, Fleur Ganier-Chauliac c, Jocelyn Gal d, Patricia Panaïa-Ferrari e, Patricia Pacini f, Patrick Fénichel a,b, Sylvie Hiéronimus a a

Department of Endocrinology, Diabetology and Reproductive Medicine, CHU de Nice, Nice, France Institut National de la Recherche Médicale, UMR U895, Université Nice-Sophia Antipolis, Nice, France Department of Neurology, Centre d'Action Médico-Sociale Précoce (CAMPS), CHU-Lenval, Nice, France d Departments of Clinical Research and Innovation and Statistics, Epidemiology and Biostatistics Unit, Centre Antoine Lacassagne, Nice, France e Department of Biochemistry, CHU de Nice, Nice, France f Observatoire du Développement Durable, Métropole Nice-Côte d'Azur, Nice, France b c

a r t i c l e

i n f o

Article history: Received 1 April 2015 Received in revised form 8 June 2015 Accepted 31 July 2015 Available online 4 August 2015 Keywords: Iodine Neuro-cognitive development Pregnancy PCB 118 Thyroxine-Binding-Globulin Thyroid tests Heavy metals

a b s t r a c t Context: The developing brain is vulnerable to iodine deficiency (ID) and environmental neuro-toxicants. Objectives: To assess neurocognitive development of children whose mothers have received (or not) iodine supplementation during pregnancy, in an area of borderline ID, while assessing in utero exposure to environmental neuro-toxicants. Design/patients: Among 86 children born from normal euthyroid women who participated in our prospective interventional study on iodine supplementation (150 μg/day) started early in pregnancy, 44 (19 with iodine supplementation, 25 controls) were assessed at two years using the Bayley test. Information on parents' education and habits (smoking), and on child development was recorded. Thyroid tests at each trimester of pregnancy and on cord blood (CB) were available, as well as milk concentrations of selected environmental compounds known for their neurotoxicity, including heavy metals and PCBs. Results: There was no difference in Bayley tests for children born to mothers with and without iodine supplementation, but sample size was small. Language and Social-Emotional Scales were negatively correlated with TBG at all times tested, while PCB 118 correlated negatively with all Language scales. Among maternal and CB thyroid tests, only CB thyroglobulin, the best marker of iodine status, correlated (negatively) with neurodevelopment scales (Motor and Expressive Language). Conclusions: This pilot study suggests that PCB118 has a negative impact on neurocognitive development, possibly mitigating the benefit of iodine supplementation in an area of borderline ID. We propose that exposure to environmental neurotoxicants should be taken into account when designing studies on the benefit of iodine supplementation in pregnancy. The potential interactions between TBG, environmental neurotoxicants and brain development warrant further studies. © 2015 Elsevier Inc. All rights reserved.

1. Introduction Impairment of neurodevelopment has long been known in case of severe iodine deficiency (ID) (Zimmermann, 2009). Beyond historical, striking reports, subtle defects have been described with mild ID (Taylor et al., 2014). ID is recognized as the first cause of preventable neurodevelopmental defect. Despite national programs of iodinesupplementation, ID has not been eradicated, especially in women of reproductive age, although the situation has improved (Stagnaro⁎ Corresponding author at: Department of Endocrinology, Diabetology and Reproductive Medicine, Archet Hospital 2, CHU Nice, 151 route de Saint-Antoine, 06200 Nice, France. E-mail address: [email protected] (F. Brucker-Davis).

http://dx.doi.org/10.1016/j.ntt.2015.07.009 0892-0362/© 2015 Elsevier Inc. All rights reserved.

Green and Pearce, 2013). The way ID can cause neurological development impairment is not fully elucidated, but is likely to involve a disturbance of maternal and fetal thyroid economy (Zimmermann, 2009; Gilbert et al., 2013). Although the focus on ID is important, it should not overshadow other environmental parameters affecting child neurocognitive development. Fetal exposure to environmental neuro-toxicants, which is emerging as a public health issue (Rodier, 2004), has so far not been investigated in the endocrine field of ID. Their neurotoxicity may involve several types of mechanisms, including, among others, a disruption of thyroid or neurotransmitter pathways (Gilbert et al., 2012), a direct toxic effect on the brain through oxidative damage (Wang and Du, 2013), or mixed mechanisms (Fischer et al., 1998; Boucher et al., 2009). Among the many neuro-toxicants, PCBs have been most studied

46

F. Brucker-Davis et al. / Neurotoxicology and Teratology 51 (2015) 45–51

(Boucher et al., 2009) and are also known to disrupt thyroid pathways through various mechanisms (Brucker-Davis, 1998; Giera et al., 2011). We took advantage of a prospective, randomized study of iodinesupplementation of euthyroid pregnant women to assess the neurodevelopment of their children at the age of two, with a focus both on iodine status and supplementation, and on exposure to environmental neurotoxicants with known thyroid-disrupting properties. 2. Material and methods 2.1. Patients Among 86 children from our prospective, randomized, interventional study on early iodine-supplementation in normal, euthyroid (initial FT4 N 12 pmol/l, TSH b 2.5 mUI/l, TPO negative) pregnant women (Hiéronimus et al., 2012), 44 participated in the follow-up neuropsychological study. Among the 65 reachable families, eight were excluded (two neonatal deaths, six ineligibles) and 13 declined to participate. Thus, 44 children were tested (19 in the iodine-supplemented group, 25 in the control group), after parents signed a consent form. Compared to the 86 of the original cohort, the 44 children were similar, although the children from the iodine-group were born to slightly older mothers, at an earlier gestational age with consequently smaller birthweight, and with a lower rate of C section (Hiéronimus et al., 2012). This study was performed between July 2007 and July 2008, one group receiving iodine-enriched pregnancy vitamins (150 μg/day, Oligobs Maxiode, Laboratoire C.C.D, Paris, France), the other one receiving the same vitamins mix, but without iodine (Oligobs Grossesse). Treatment was given from the day of enrolment (before 10 week gestation) throughout pregnancy, until three months post-partum. Compliance with vitamin supplementation was assessed by hospital pharmacists, based on the number of pills brought back at each visit. Using parental questionnaires and child medical records, information on parents' education, health, smoking and alcohol consumption, family issues (language spoken at home, type of child care), and child development (health, growth curve) was recorded. 2.2. Hormonal and chemical assays Comprehensive maternal thyroid tests were performed at each trimester, at delivery and at the 3-month post-partum visit. At delivery, cord blood (CB) samples were collected and analyzed for thyroid tests, including thyroglobulin (Tg), anti-Tg antibodies, fT4, fT3, TSH, and rT3. Maternal milk was collected in glass tubes within the five first days postpartum for measurement of selected neuro-toxicants. Spot ioduria was measured by mass spectrometry ICP/MS (Pasteur-Cerba Laboratory, Cergy Pontoise, France, detection threshold 15 μg/l; intra- and inter-series coefficient of variation, CV b 10%). Tg was measured by Immunoradiometric assay (Thyroglobulin IRMA, CisBio International, Gif-sur-Yvette, France). FT4, fT3, total T4, TSH, and anti-Tg antibodies were measured by chemiluminescence (ADVIA Centaur, Siemens Healthcare Diagnostics, France); rT3 was measured by RIA: RIA rT3 (Pasteur Cerba Laboratory, Cergy Pontoise, France). ThyroxineBinding-Globulin (TBG) was measured by RIA (RIA-gnost-TBG, CisBio International, Gif-sur-Yvette, France). Reference ranges were established in our laboratory for fT4 and TSH during the first trimester of pregnancy (2.5–97.5 percentiles): fT4 11.47–19.23 pmol/l; TSH 0.053–3.23 mUI/l. The other reference ranges were provided by the manufacturer outside pregnancy: Tg 5–50 ng/ml; anti-Tg antibodies b 60 UI/ml; fT3 3–7 pmol/l; rT3 0.14–0.54 nmol/l; TBG 10– 42 μg/ml. CV were: for fT4 (intra-assay CV 2.31%; inter-assay CV 1.95%); TSH (intra-assay CV 2.67%; inter-assay CV 3.97%); Tg (intra-assay CV 2.4%; inter-assay CV 4.5%); anti-Tg antibodies (intra-assay CV 5.5%; inter-assay CV 1.8%); fT3 (intra-assay CV 2.35); rT3 (intra-assay CV 8.54%; inter-assay CV 6.21%).

We measured milk concentrations of 15 environmental compounds known for their neuro-toxicity and thyroid disruption properties (Boucher et al., 2009): three heavy metals (lead, mercury, cadmium), Dichloro-diphenyl-dichloroethylene (DDE), six selected PCBs (PCB77, 118, 126, 138, 153 and 180), four polybrominated biphenyls (PBDEs: BDE47, 99, 100, 153), and hexachlorobenzene. Analysis was performed by gas chromatography coupled with mass spectrometry, at the Laboratoire de l'Observatoire du Développement Durable (Nice, France), a laboratory certified by the French Ministry of Environment. Threshold of quantitation was 0.3 ng/g of fat, except for heavy metals 0.2 ng/g of milk. Percentage of fat was determined for each sample. Data are expressed in ng/g of fat and ng/g of milk for lipophilic compounds and ng/g of milk for heavy metals. 2.3. Neuropsychological assessment Cognitive and psychomotor development of children was assessed using the Bayley Scales of Infant and Toddler Development (Third Edition, 2006, by Nancy Bayley, Editor Harcourt Assessment Inc., San Antonio, TX, USA). The Bayley test was administered to children aged two in presence of one of the parents, by the same investigator, blinded for the iodine-supplementation status of the mother, at a date as close as possible from the child's second birthday. Test administration was standardized (same location, time of adjustment, office set up to accommodate children and families). Testing was re-scheduled in case of interfering child health issue. Raw scores were normalized for the actual age of the child at testing. Four scales were studied: Cognitive, Language, Motor and Social-emotional, with two subscales for Language (Receptive and Expressive communication) and two subscales for Motor (Fine Motor and Gross Motor). Social-emotional scores were established based on a parent-filled questionnaire. Data were expressed as Composite scores and Percentile ranks. Composite scores, derived from sums of subtest scaled scores, were generated for the Language scales and Motor scales. For Cognitive and Social-Emotional Scales (which have a single test score), composite score equivalents were available. The composite scores were scaled to a metric, with a mean of 100 and a Standard Deviation of 15, and ranged from 40 to 160. Percentile rank ranged from 1 to 99, with 50 as the mean and median. Percentile rank indicates the standing of a child relative to children among the standard sample group. 2.4. Statistical analysis Data were entered and stored in a database file, and transferred into R3.0.2 software for statistical analysis. We first compared the groups with (n = 19) or without (n = 25) iodine-supplementation. Quantitative variables were expressed as means, Standard Deviation, medians, and range. Qualitative variables were described by counts and percentages. Chi-square or Fisher's exact tests were used to establish differences in the distribution of discontinuous variables, student's test or Mann–Whitney's U-test to compare continuous variables. For the whole group of 44 children, we studied the parameters that could influence the four composite scores of neuro-development, including: Iodine (interventional iodine-supplementation, or ioduria at each trimester), maternal (at each trimester) and CB thyroid function tests, neurotoxicant concentrations in milk, and potential clinical confounders: children parameters (sex, birth weight), delivery complications, maternal age, parental education, parental alcohol or tobacco consumption. The correlations between continuous variables were determined using the Pearson's test or Spearman's rank test. Parametric test of analysis of variance or nonparametric test of Kruskall and Willis were performed to test variables expressed as categories versus continuous variables. In this pilot study, all significant independent variables in the univariate analysis were introduced in the model for multivariate analysis. When two variables were correlated, we excluded one of them from the model. Subsequently, we designed a final “stepping-down” model,

F. Brucker-Davis et al. / Neurotoxicology and Teratology 51 (2015) 45–51

keeping only the variables from the initial model that remained significant. Test of significance was two-tailed and considered significant with an alpha level of p b 0.05.

3. Results The iodine-supplemented and control groups were similar, except for the shorter term at birth in the iodine-group, resulting in smaller and lighter babies (Table 1). Table 2 shows the hormonal maternal and CB profile of each group. In CB, there was no significant difference, although Tg tended to be lower in the iodine-group. In mothers, ioduria was higher and Tg lower in the iodine-supplemented group, starting at the second trimester. There was no difference in thyroid tests (Table 2) between the two groups, including antithyroid antibody titers. Out of 44 children, 32 were breastfed (duration of nursing: 5.5 months). There was no significant difference in neurotoxic compounds concentrations in colostrum (Table 3). Exposure to PCB153, PCB138, DDE, PBDE47, cadmium and lead was quite ubiquitous (N90% of samples above quantitation threshold). Regarding neuro-cognitive testing, most children scored in the normal range or above, whatever their group (Table 4). There was no difference in Bayley test results between breast-fed and bottle-fed babies (data not shown). There was no statistical difference between the iodine and the control groups, whether we looked at the composite scores (Table 4), or at the percentiles (data not shown). Significant results in univariate analysis of continuous variables known to influence neurodevelopment are shown in Table 5. Several non-continuous variables were significant as well, mainly parental education, paternal smoking, and sex of the child (data not shown). Results of the multivariate analysis are detailed in Table 6. Some parameters significant in univariate analysis were correlated: milk PCB118 concentrations and maternal blood TBG concentrations (R = 0.41, p = 0.02, e.g. at 2d trimester), maternal and paternal education (R = 0.5, p = 0.0005). We tested different models using either TBG or PCB118, and either maternal or paternal education. Regarding the

Table 1 Patients characteristics. Iodine (n = 19)

Controls (n = 25)

p value

31 (21–37)

28 (21–39)

NS

5 (26%) 4 (22%) 3 (17%) 16 (84%) 5.5 (0.75–24)

7 (28%) 7 (28%) 2 (8%) 17 (68%) 5.5 (0.3–24)

NS NS NS NS NS

Pregnancy & delivery Term at birth (WA) C section n (%)

38.6 (35–40.5) 2 (10.5%)

40.4 (37.4–41.6) 3 (12%)

b0.005 NS

Child Sex Birth weight (g) Birth length (cm) Age at 1st sitting (month) Age at 1st walk (month) Age at 1st word (month) First baby in family n (%)

8F/11M 3060 (2170–3590) 48.5 (45.5–52) 6 (4.5–9) 12 (9–18) 10 (5–21) 7 (37%)

8F/17M 3530 (2380–4330) 50 (45.5–52.5) 6 (5–9.5) 12 (9–18) 12 (7–21) 14 (56%)

NS b0.005 b0.05 NS NS NS NS

Family issues Parental education mother % university educ. father % university educ. Day Care by family n (%) French-only spoken at home

N = 11 (58%) N = 7 (37%) 9 (47%) 16 (84%)

N = 14 (56%) N = 10 (40%) 9 (36%) 21 (84%)

NS NS NS NS

Mother Age (years) Smoking n (%) During pregnancy At testing Alcohol during pregnancy Breast feeding (N1 month) n (%) Duration of nursing (months)

Results are expressed as medians (with range in parenthesis) educ.=education. Alcohol during pregnancy was always occasional and in small quantity.

47

Table 2 Biological tests. Iodine (n = 19)

Controls (n = 25)

P value

Cord blood TSH mUI/l FT4 pmol/l Tg ng/ml

8.0 (2.51–16.3) 13.4 (11.7–17.4) 66.8 (12.7–176.1)

6.2 (1.8–24.5) 12.8 (9.5–17.2) 96.1 (15.4–219.8)

NS NS NS

Maternal thyroid tests TSH 1st trimester mUI/l TSH 2dtrimester mUI/l TSH 3dtrimester mUI/l FT4 1st trimester pmol/l FT4 2dtrimester pmol/l FT4 3d trimester pmol/l Tg 1st trimester ng/ml Tg 2d trimester ng/ml Tg 3d trimester ng/ml Ioduria 1st trim μg/l Ioduria 2d trim μg/l Ioduria 3d trim μg/l TBG 1st trim μg/ml TBG 2d trim μg/ml TBG 3d trim μg/ml

1.17 (0.51–2.37) 1.21 (0.51–3.11) 1.13 (0.44–3.05) 14.1 (12.0–16.9) 12.1 (10.2–13.8) 11.5 (9.2–15.7) 17.0 (2.4–45.1) 12.4 (0.4–38.4) 14.1 (1.2–49.7) 124 (28–399) 128 (43–470) 169 (18–358) 28.2 (19.5–46.2) 46.8 (28.3–62.2) 54.6 (31.9–65.5)

1.17 (0.43–2.06) 1.08 (0.48–2.66) 1.33 (0.44–2.39) 14.0 (12.2–18.5) 11.5 (9.8–17.39) 11.6 (8.4–14.6) 18.3 (1.3–84.1) 22.7 (1.8–59.1) 29.4 (5.4–98.6) 117 (14–257) 105 (21–227) 103 (19–199) 27.2 (16.9–48.3) 46.0 (23.7–60.3) 48.2 (28.8–64.5)

NS NS NS NS NS NS NS 0.01 0.01 NS 0.005 0.005 NS NS NS

Data are expressed as medians (range).

Cognitive Scale, paternal education and second trimester anti-Tg antibody titers remained significant. Regarding the Language scale, results were similar when using maternal or paternal education, while there were slight differences when using PCB118 or TBG (Table 6). Overall, parental education, paternal smoking, PCB118 or TBG at 2d trimester remained significant depending on the models. Regarding the Motor scale, only CB Tg remained in the model. For the social-emotional scale, only second trimester TBG remained in the model. 4. Discussion This is a pilot study on the neurodevelopment at age two of children born to euthyroid mothers enrolled in a prospective, randomized study testing the benefit of iodine supplementation during pregnancy, in an area of borderline ID. We report no direct benefit of iodine supplementation per se, although correlations of CB Tg were encouraging. In contrast, we report some negative correlations of language development with PCB 118 concentrations in maternal milk and with maternal TBG. 4.1. Effect of iodine-supplementation Using the Bayley test, we found no difference in neurodevelopment for children from the iodine-supplemented group compared to controls, thus demonstrating no measurable direct benefit, at least not when iodine is given from the end of first trimester throughout pregnancy. This was reported by some authors (Rebagliato et al., 2013; Santiago et al., 2013), while other studies seemed encouraging (Berbel et al., 2009; Velasco et al., 2009). The discrepancies could be due to: 1/the small number of children tested; 2/ the timing of substitution: it might be necessary to start iodine before pregnancy (Berbel et al., 2009); 3/ the selected population: normal thyroid tests, TPOnegative women vs unselected population including women with predisposition to thyroid disease (Williams et al., 2013); 4/ the age at testing: two year-old vs school-age children; or 5/ the neurological test studied (Trumpff et al., 2013). Recent meta-analyses have not brought a definitive answer on the benefit of maternal iodinesupplementation on offspring neurodevelopment in areas of mild ID, since prospective interventional studies are lacking (Taylor et al., 2014; Trumpff et al., 2013). In our small cohort, ioduria was a poor parameter of iodine status with even paradoxical negative correlations in the first trimester. Ioduria is

48

F. Brucker-Davis et al. / Neurotoxicology and Teratology 51 (2015) 45–51

Table 3 Environmental compounds concentrations in maternal milk. Maternal milk

% of detection

Total (n = 32)

Iodine (n = 15)

Controls (n = 17)

p value

Selected POP PCB153 ng/g fat PCB 138 ng/g fat PCB 180 ng/g fat PCB 118 ng/g fat PCB 77 and 126 ng/g fat DDE ng/g fat HCB ng/g fat PDBE 47 ng/g fat PBDE 99 ng/g fat PBDE 100 ng/g fat PBDE 153 ng/g fat

100 94 87.5 19 0 97 56 100 41 32 3

94.1 (32.8–512.8) 79.2 (0–256.4) 52.8 (0–307.7) 0 (0–118.6) 0 (0–0) 162.8 (0–5000) 15.4 (0–82.2) 254.1 (44.6–1317.6) 0 (0–179.8) 0 (0–53.3) 0 (0–8.3)

114.4 (32.8–512.8) 95.4 (0–256.4) 78 (0–307.7) 0 (0–118.6) 0 (0–0) 163.5 (0–5000) 16.4 (0–67.8) 256.9 (102.2–1179.5) 13.3 (0–179.8) 0 (0–53.3) 0 (0–8.3)

66.1 (33.8–391.9) 72.5 (0–237.5) 48.4 (0–240) 0 (0–49) 0 (0–0) 149.4 (32.2–392) 14.5 (0–82) 250 (44.6–1317.6) 0 (0–43.8) 0 (0–24.2) 0 (0–0)

NS NS NS NS NS NS NS NS NS NS NS

Heavy metals Mercury ng/g milk Lead ng/g milk Cadmium ng/g milk

31 93 100

0 (0–0.7) 27 (12–100) 3.3 (0.8–30.6)

NS NS NS

0 (0–9) 25.2 (0–100) 3.2 (0.8–30.6)

0 (0–9) 22.2 (0–52.2) 2.4 (0.9–10)

Data are expressed as: medians (range) 32 women nursed their babies (15 out 19 women receiving iodine supplementation and 17 out of 25 controls). Medians for the environmental compounds are for all measured samples including undetectable (by convention reported as “0”).

known to be a tricky marker in small groups (Zimmermann, 2009), while CB Tg is emerging as a better marker of fetal iodine status (Zimmermann, 2009; Ma and Skeaff, 2014). CB Tg was negatively correlated with three developmental scales (Expressive Language, Motor and Gross Motor scales), suggesting indirectly a benefit of iodine-supplementation for the children, while we had previously reported a benefit for the mother, with lower maternal Tg in the iodine-supplemented group (Brucker-Davis et al., 2013).

Table 4 Results of neuro-cognitive testing with the Bayley Scales of Infant and Toddler Development Test.

Age at testing (days) Cognitive scale Composite score Percentile Number of children below the 85 percentile Language scale Global Composite score Percentile of global score Number of children below the 85 percentile Receptive communication Composite score Expressive communication Composite score Motor scale Global composite score Global percentile Number of childrenbelow the 85 percentile Fine motor subtest Composite score Gross motor subtest Composite score Social-emotional scale Composite score Percentile Number of children below the 85 percentile

Iodine (n = 19) Controls (n = 25)

p value

734 (706–763)

737 (717–759)

NS

110 (90–145) 75 (25–99.9) 0

110 (90–145) 75 (25–99.9) 0

NS NS NS

104.5 (71–121) 62 (3–92) 2

100 (77–127) 50 (6–96) 3

NS NS NS

100

100

NS

105

100

NS

110 (97–124) 75 (42–95) 0

110 (94–133) 75 (34–99) 0

NS NS NS

115

115

NS

100

100

NS

100 (65–110) 50 (1–75) 2

90 (75–120) 25 (5–91) 2

NS NS NS

4.2. Impact of thyroid tests We found no correlation of child neurodevelopment with extensive maternal and CB thyroid hormone tests, including fT4. Our selected population of women with normal free T4 does not allow drawing conclusion about the potentially deleterious effect of isolated hypothyroxinemia, which was reported by others as a significant parameter (Henrichs et al., 2010; Craig et al., 2012). There are strong experimental and clinical evidence supporting the crucial role of thyroid hormones in early brain development, cognitive and sensorial functions (Rovet, 2014). However, circulating maternal thyroid hormone levels at one time-point doesn't necessarily reflect fetal brain thyroid hormone concentrations, partly explaining discrepancies in the literature (Oken et al., 2009). Transport of thyroid hormones, mainly T4, through the placenta is not well elucidated, but is likely to include the role of protein carriers, with transthyretin (TTR) the most studied (Richard et al., 2012). The whole process of thyroid hormone delivery to the brain is extraordinary complex (Wirth et al., 2014), especially in the fetus, requesting coordination between multiple players, in order to deliver the right amount of hormone, at specific time-point and location, for the right group of cells (Wirth et al., 2014; Zoeller and Rovet, 2004; Preau et al., 2014). We report here for the first time an association of maternal TBG levels at all trimesters with children neurodevelopment: lower were the TBG levels, better were the scores of Language and SocialEmotional scales. TBG, the main thyroid hormone carrier in human, is physiologically increased during pregnancy due to high estrogen levels. Surprisingly, we find an association with TBG, but not with circulating maternal thyroid hormone levels (free or total). We could speculate that higher levels of TBG could result in relative less fT4 available for crossing the placenta, and eventually reaching the fetal compartment.

4.3. Impact of environmental neuro-toxicants

Results are expressed as medians of composite scores and in percentiles with the range in parenthesis. Social-Emotional scale scoring is based on questionnaire filled by the parent present on the day of testing.

We had chosen 15 compounds known for their thyroid and/or neurotoxic effect in vivo or in vitro (Wang and Du, 2013; Boucher et al., 2009; Brucker-Davis, 1998; Fritsche et al., 2005; Goldey and Taylor, 1992; Ribas-Fito et al., 2006; Herbstman et al., 2010; Farina et al., 2011; Mason et al., 2014). Most are lipophilic and persistent, like PCBs, PBDE, hexachlorobenzene and DDE, while heavy metals (mercury, lead and cadmium) are not. Lipophilic compounds, persistent in biological tissues, are easily excreted in milk. Measurements in milk collected within the first five days post-partum reflect in utero exposure to

F. Brucker-Davis et al. / Neurotoxicology and Teratology 51 (2015) 45–51

49

Table 5 Univariate correlations of neuro-developmental scales with clinical, thyroid and environmental parameters (continuous variables). Cognitive

Language

RL

EL

Motor

FM

GM

Clinical features Birth weight Thyroid tests Anti Tg AB 2d T Anti Tg AB 3d T TBG 1st T TBG 2d T TBG 3d T TBG post-partum Tg in CB TSH 1st T TSH 2d T TSH 3d T TT4 1st T TT4 3d T UIE 1st T Neurotoxicants PCB 118 (ng/g milk) PCB 118 (ng/g fat) PBDE 99 (ng/g fat) PBDE 99 (ng/g milk) Cadmium

Social-Emotional

0.34a −0.43c −0.30a

−0.44b

−0.30a −0.31a −0.40b −0.33a −0.48b

−0.31a −0.32a

−0.34a −0.45c −0.3a −0.53c −0.31a

−0.53c −0.38b −0.56c −0.35a

−0.30a

−0.45b −0.32a −0.36a

−0.40b −0.31a −0.34a −0.37a −0.42b −0.42a

−0.48c −0.46b −0.45a

−0.38a

−0.38b

−0.35a

0.42a 0.38a

0.35a −0.35a

RL = Receptive Language; EL = Expressive Language; FM = Fine Motor; GM = Gross Motor. AB = antibodies; T = trimester; CB = cord blood. Results are expressed as R and p values. a p values b 0.05. b p values b 0.01. c p values b 0.005.

persistent pollutants, but also post-natal exposure in breast-fed babies. In this small cohort, neurodevelopmental scores were similar in nursed and bottle-fed babies, suggesting that any deleterious effect of those persistent pollutants could be counter-balanced by the well-known benefits of long-term breast feeding (Ribas-Fito et al., 2007). We report a strong association of PCB118 milk concentrations with the Language developmental scales, including Receptive and Expressive communication, while no other chemicals reached significance, in contrast with other human studies which found a negative effect of PBDE (Herbstman et al., 2010), DDE (Ribas-Fito et al., 2006) and heavy metals (Kim et al., 2012; Boucher et al., 2010). Although the production of PCBs has been banned in the 1970s, they are persistent environmental contaminants, still largely present in human tissues. PCB congeners are classified according to their specific modes of action (dioxin vs non-dioxin-like compounds) (Fischer et al., 1998; Hansen, 1998). Many congeners are known to be neurotoxic, through multiple mechanisms, including thyroid disruption (Fischer et al., 1998; Hansen, 1998). There are some discrepancies in the literature on PCB neurotoxicity, partly due to the end-point studied, the type of congener, or the period of exposure (Boucher et al., 2009). The most consistent effects are on executive functioning, processing speed, verbal abilities and visual recognition memory (Boucher et al., 2009). PCB118 is a mono-ortho

congener (Boucher et al., 2010; Dallaire et al., 2009; Doi et al., 2013; Hisada et al., 2013), with limited dioxin-like activity and pleiotropic actions pertaining to thyroid and brain function (Hansen, 1998). Our data showing a negative association with Language scales are consistent with those of Doi et al. who reported that PCB118 may also affect “sociocognitive” function very early in life, with potential consequences later in life (Doi et al., 2013). PCB118 easily crosses the placenta and is largely distributed into fetal tissues, including fetal brain (Berg et al., 2010). PCB118 may disrupt the thyroid economy through different mechanisms, including, for some, a possible trend for mild hypothyroidism (Dallaire et al., 2009). Importantly, it is now recognized that thyroid disruption may be due to entirely different mechanisms, independently of circulating thyroid hormone concentrations. PCB118 may have some tissue-specific “T3-like activity”. Like T3, it induces a dose-dependent increase in oligodendrocyte formation in an in vitro model of human neural progenitor cell (Fritsche et al., 2005). This action was blocked by NH3, a T3-receptor antagonist. It has been proposed that this “T3mimicking action” involves the thyroid hormone receptor complex, independently of receptor binding (Fritsche et al., 2005). This effect could be deleterious for brain development if the developmental timing is not respected. Furthermore, PCB118 was positively correlated to 2d trimester TBG. TBG is negatively correlated to the Language scales and

Table 6 Multivariate analysis of correlations between developmental scales and clinical, thyroid and environmental parameters. Scale studied

Significant model

Cognitive scale Language scalea Receptive languagea Expressive languagea Motor scale Fine motor Global motor Social–Emotional scale

Father education and anti Tg at 2d trimester Mother (or father) education, paternal smoking, PCB 118 Mother (or father) education, paternal smoking, PCB118 Mother (or father) education, PCB118, and CB Tg CB Tg Delivery problem, Child sex, TSH and ioduria at 1st trimester CB Tg TBG at 2d trimester

R2 = 0.207 R2 = 0.590 R2 = 0.577 R2 = 0.370 R2 = 0.120 R2 = 0.400 R2 = 0.162 R2 = 0.300

p b 0.01 p b 0.0001 p b 0.0001 p = 0.005 p = 0.02 p = 0.001 p b 0.01 p = 0.0001

a PCB 118 in the model. If using TBG at 2d trimester in the model (instead of PCB 118), we observed the following result: for Language scale, paternal smoking and TBG at 2d trimester (R2 = 0.372, p 0.0002); for Receptive Language subscale, paternal smoking, TBG at 2d trimester and ioduria at 1st trimester (R2 = 0.415, p = 0.0005); for Expressive Language subscale, TBG at 2d trimester and CB thyroglobulin (R2 = 0.30, p = 0.001).

50

F. Brucker-Davis et al. / Neurotoxicology and Teratology 51 (2015) 45–51

the Social-Emotional scale. Data on TBG concentrations and pollutant contamination are very scarce and conflicting (Dallaire et al., 2009; Hisada et al., 2013). Interestingly, PCBs, including PCB118, have a structure similar to thyroid hormones and compete with them for binding to thyroid hormone carriers, mainly TTR. We speculate that the fraction of free hormones available for transfer through the fetal blood–brain-barrier is altered according to the degree of PCB contamination, with less thyroid hormones reaching the fetal brain. We also speculate that PCB118 enters the brain via its binding to TTR, which functions as a thyroid hormone distributor (Alsherhri et al., 2015). The choroid plexus has the ability to synthetize TTR, which is unilaterally secreted into the cerebrospinal fluid, and is believed to allow the movement of T4 into the brain. It may well be that pollutants like PCBs, able to bind to TTR, could compete with T4 for their delivery to the brain (Alsherhri et al., 2015), thus disturbing thyroid hormone pathways and causing neurotoxicity. The main limitation to our study is the small number of children who ended up being tested, only 50% of the original cohort. Thus, our statistical analysis lacks power, and our results should be considered preliminary. However, we were able to confirm the impact of wellestablished clinical parameters influencing neurodevelopment, such as parental education and smoking, bringing plausibility to our data on PCB 118 and TBG. Our data suggest that PCB118 may have a negative impact on early language development, possibly mitigating the benefit of iodine supplementation. Those results remain to be confirmed on a larger population. Thus, we propose that exposure to environmental neurotoxicants, such as PCB118, should be considered as potential confounding factors when designing studies on iodine and neurodevelopment. The potential interactions between TBG, environmental neurotoxicants and brain development warrant further studies. Conflicts of interest None. Acknowledgments Research was founded by a grant from the University Hospital of Nice (2009-A01166-51), as an extension to a first grant from the French Ministry of Health (DGS2006-0103), and was promoted by the Direction of Clinical Research of the University Hospital of Nice. We wish to thank the families for their participation and the laboratory CCD (France) for providing the vitamins (iodinated or not) free of charge. References Alsherhri, B., D'Souza, D.G., Lee, J.Y., Petratos, S., Richardon, S.J., 2015. The diversity of mechanisms influenced by transthyretin in neurobiology: development, disease and endocrine disruption. J. Neuroendocrinol. http://dx.doi.org/10.1111/jne.12271. Berbel, P., Mestre, J.L., Santamaria, A., Palazon, I., Franco, A., Graells, M., Gonzalez-Torga, A., Morreale de Escobar, G., 2009. Delayed neurobehavioral development in children born to pregnant women with mild hypothyroxinemia during the first month of gestation: the importance of early iodine supplementation. Thyroid 19, 511–519. Berg, V., Lyche, J.L., Gutleb, A.C., Lie, E., Skaare, J.U., Aleksandersen, M., Ropstad, E., 2010. Distribution of PCB 118 and PCB 153 and hydroxylated PCB metabolites (OH-CBs) in maternal, fetal and lamb tissues of sheep exposed during gestation and lactation. Chemosphere 80, 1144–1150. Boucher, O., Muckle, G., Bastien, C.H., 2009. Prenatal exposure to PCBs: a neuropsychological analysis. Environ. Health Perspect. 117, 7–16. Boucher, O., Bastien, C.H., Saint-Amour, D., Dewailly, E., Ayotte, P., Jacobson, J.L., Jacobson, S.W., Muckle, G., 2010. Prenatal exposure to methylmercury and PCBs affects distinct stages of information processing: an event-related potential study with Inuit children. Neurotoxicology 31, 373–384. Brucker-Davis, F., 1998. Effects on environmental synthetic chemicals on thyroid function. Thyroid 8, 827–856. Brucker-Davis, F., Panaia-Ferrari, P., Gal, J., Fenichel, P., Hieronimus, S., 2013. Iodine supplementation throughout pregnancy does not prevent the drop in FT4 in the second and third trimesters in women with normal initial thyroid function. Eur. Thyroid J. 2, 187–194.

Craig, W.Y., Allan, W.C., Kloza, E.M., Pulkkinen, A.J., Waisbren, S., Spratt, D.I., Palomaki, G.E., Neveux, L.M., Haddow, J.E., 2012. Mid-gestational maternal free thyroxine concentration and offspring neurocognitive development at age two years. J. Clin. Endocrinol. Metab. 97, E22–E28. Dallaire, R., Dewailly, E., Pereg, D., Dery, S., Ayotte, P., 2009. Thyroid function and plasma concentrations of polyhalogenated compounds in Inuit adults. Environ. Health Perspect. 117, 1380–1386. Doi, H., Nishitani, T., Nagai, T., Kakeyama, M., Maeda, T., Shinohara, K., 2013. Prenatal exposure to a PCB congener influences fixation duration on biological motion at 4months-old: a preliminary study. PLoS ONE 8, e59196. http://dx.doi.org/10.1371/ journal.pone.0059196. Farina, M., Rocha, J.B.T., Aschner, M., 2011. Mechanisms of methylmercury-induced neurotoxicity: evidence from experimental studies. Life Sci. 89, 555–563. Fischer, L.J., Seegal, R.F., Ganey, P.E., Pessah, I.N., Kodavanti, P.R.S., 1998. Symposium overview: toxicity of non-coplanar PCBs. Toxicol. Sci. 41, 49–61. Fritsche, E., Cline, J.E., Nguyen, N.-H., Scanlan, T.S., Abel, J., 2005. PCBs disturb differentiation of normal human neural progenitor cells: clue of involvement of thyroid hormone receptors. Environ. Health Perspect. 113, 871–876. Giera, S., Bansal, R., Ortiz-Toro, T.M., Taub, D.G., Zoeller, R.T., 2011. Individual PCB congeners produce tissue- and gene-specific effects on thyroid hormone signalling during development. Endocrinology 52, 2909–2919. Gilbert, M.E., Rovet, J., Chen, Z., Koibuchi, N., 2012. Developmental thyroid hormone disruption: prevalence, environmental contaminants and neurodevelopmental consequences. Neurotoxicology 33, 842–852. Gilbert, M.E., Hedge, J.M., Valentin-Blasini, L., Blount, B.C., Kannan, K., Tietge, J., Zoeller, R.T., Crofton, K.M., Jarrett, J.M., Fisher, J.W., 2013. An animal model of marginal iodine deficiency during development: the thyroid axis and neurodevelopmental outcome. Toxicol. Sci. 132, 177–195. Goldey, E.S., Taylor, D.H., 1992. Developmental toxicity following premating maternal exposure to hexachlorobenzene in rats. Neurotoxicol. Teratol. 14, 15–21. Hansen, L.G., 1998. Stepping backward to improve assessment of PCB congener toxicities. Environ. Health Perspect. 106 (Suppl. 1), 171–189. Henrichs, J., Bongers-Schokking, J.J., Schenk, J.J., Ghassabian, A., Schmidt, H.G., Visser, T.J., Hooijkaas, H., de Muinck Keizer-Schrama, S.M.P.F., Hofman, A., Jaddoe, V.V.W., Visser, W., Steegers, E.A.P., Verhulst, F.C., Rijke, Y.B., Tiemeier, H., 2010. Maternal thyroid function in early pregnancy and cognitive functioning in early childhood: the generation R study. J. Clin. Endocrinol. Metab. 95, 4227–4234. Herbstman, J.B., Sjödin, A., Kurzon, M., Lederman, S.A., Jones, R.S., Rauh, V., Needham, L., Tang, D., Niedzwiecki, M., Wang, R.Y., Perera, F., 2010. Prenatal exposure to PBDEs and neurodevelopment. Environ. Health Perspect. 118, 712–719. Hiéronimus, S., Ferrari, P., Gal, J., Berthier, F., Azoulay, S., Bongain, A., Fénichel, P., BruckerDavis, F., 2012. Relative impact of iodine supplementation and maternal smoking on cord blood thyroglobulin in pregnant women with normal thyroid function. Eur. Thyroid J. 1, 264–273. Hisada, A., Shimodaira, K., Okai, T., Watanabe, K., Takemori, H., Takasuga, T., Noda, Y., Shirakawa, M., Kato, N., Yoshinaga, J., 2013. Serum levels of hydroxylated PCBs, PCBs and thyroid hormone measures of Japanese pregnant women. Environ. Health Prev. Med. 18, 205–214. Kim, Y., Ha, E.-H., Park, H., Ha, M., Kim, Y., Hong, Y.-C., Kim, E.-J., Kim, B.-N., 2012. Prenatal lead and cadmium co-exposure and infant neuro-development at 6 months of age: the Mothers and Children's Environmental Health (MOCEH) study. Neurotoxicology 35, 15–22. Ma, Z.F., Skeaff, S.A., 2014. Thyroglobulin as a biomarker of iodine deficiency: a review. Thyroid 24, 1195–1209. Mason, L.H., Harp, J.P., Han, D.Y., 2014. Pb neurotoxicity: neuropsychological effects of lead toxicity. Biomed. Res. Int. http://dx.doi.org/10.1155/2014/840547. Oken, E., Braverman, L.E., Platek, D., Mitchell, M.L., Lee, S.L., Pearce, E.N., 2009. Neonatal thyroxine, maternal thyroid function, and child cognition. J. Clin. Endocrinol. Metab. 94, 497–503. Preau, L., Fini, J.B., Morvan-Dubois, G., Demeneix, B., 2014. Thyroid hormone signaling during early neurogenesis and its significance as a vulnerable window of endocrine disruption. Biochim. Biophys. Acta http://dx.doi.org/10.1016/ jbbagrm.2014.06.015. Rebagliato, M., Murcia, M., Alvarez-Pedrerol, M., Espada, M., Fernandez-Somoano, A., Lertxundi, N., Navarrete-Munoz, E.M., Forns, J., Aranbarri, A., Llop, S., Julvez, J., Tardon, A., Ballester, F., 2013. Iodine supplementation during pregnancy and infant neuropsychological development INMA mother and child cohort study. Am. J. Epidemiol. 177, 944–953. Ribas-Fito, N., Torrent, M., Carrizo, D., Munoz-Ortiz, L., Julvez, J., Grimalt, J.O., Sunyer, J., 2006. In utero exposure to background concentrations of DDT and cognitive functioning among preschoolers. Am. J. Epidemiol. 164, 955–962. Ribas-Fito, N., Julvez, J., Torrent, M., Grimalt, J.O., Sunyer, J., 2007. Beneficial effects of breastfeeding on cognition regardless of DDT concentrations at birth. Am. J. Epidemiol. 166, 1198–1202. Richard, K., Li, H., Landers, K.A., Patel, J., Mortimer, R.H., 2012. Placental transport of thyroid hormone and iodide. In: Zheng, J. (Ed.), Recent Advances in Research on the Human Placenta, pp. 309–334 (chapter 15). Rodier, P.M., 2004. Environmental causes of central nervous system maldevelopment. Pediatrics 113 (Suppl. 4), 1076–1083. Rovet, J., 2014. The role of thyroid hormones for brain development and cognitive function. Paediatr. Thyroidol Endocr. Dev. 26, 26–43. Santiago, P., Velasco, I., Muela, J.A., Sanchez, B., Martinez, J., Rodriguez, A., Berrio, M., Gutierrez-Repiso, C., Carreira, M., Moreno, A., Garcia-Fuentes, E., Soriguer, F., 2013. Infant neurocognitive development is independent of the use of iodised salt or iodine supplements given during pregnancy. Br. J. Nutr. 110, 831–839.

F. Brucker-Davis et al. / Neurotoxicology and Teratology 51 (2015) 45–51 Stagnaro-Green, A., Pearce, E.N., 2013. Iodine and pregnancy: a call to action. Lancet 382, 292–293. Taylor, P.N., Okosieme, O.E., Dayan, C.M., Lazarus, J.H., 2014. Impact of iodine supplementation in mild to moderate iodine deficiency: systematic review and meta-analysis. Eur. J. Endocrinol. 170, R1–R15. Trumpff, C., De Schepper, J., Tafforeau, J., Van Oyen, H., Vanderfaeillie, J., Vandevijvere, S., 2013. Mild iodine deficiency in pregnancy in Europe and its consequences for cognitive and psychomotor development of children: a review. J. Trace Elem. Med. Biol. 27, 174–183. Velasco, I., Carreira, M., Santiago, P., Muela, J.A., Garcia-Fuentes, E., Sanchez-Munoz, B., Garriga, M.J., Gonzalez-Fernandez, M.C., Rodriguez, A., Caballero, F.F., Machado, A., Gonzalez-Romero, S., Anarte, M.T., Soriguer, F., 2009. Effect of iodine prophylaxis during pregnancy on neurocognitive development of children during the first two years of life. J. Clin. Endocrinol. Metab. 94, 3234–3241.

51

Wang, B., Du, Y., 2013. Cadmium and its neurotoxic effects. Oxidative Med. Cell. Longev. http://dx.doi.org/10.1155/2013/898034. Williams, F.L.R., Watson, J., Ogston, S.A., Visser, T.J., Hume, R., Willatts, P., 2013. Maternal and umbilical cord levels of T4, FT4, TSH, TPOAb, and TgAb in term infants and neurodevelopmental outcome at 5.5 years. J. Clin. Endocrinol. Metab. 98, 829–838. Wirth, E.K., Schweizer, U., Köhrle, J., 2014. Transport of thyroid hormone in brain. Mini review article. Front. Endocrinol. 5, 98. http://dx.doi.org/10.3389/fendo.2014.00098. Zimmermann, M.B., 2009. Iodine deficiency. Endocr. Rev. 30, 376–408. Zoeller, R.T., Rovet, J., 2004. Timing of thyroid hormone action in the developing brain: clinical observations and experimental findings. J. Neuroendocrinol. 16, 809–818.