Long-term parental methamphetamine exposure of ...

1 downloads 0 Views 1MB Size Report
Feb 18, 2014 - Keywords: behavior; cross-fostering; epigenetics; methamphetamine; prenatal exposure. INTRODUCTION. Methamphetamine (METH) is a ...
Molecular Psychiatry (2014), 1–11 © 2014 Macmillan Publishers Limited All rights reserved 1359-4184/14 www.nature.com/mp

ORIGINAL ARTICLE

Long-term parental methamphetamine exposure of mice influences behavior and hippocampal DNA methylation of the offspring Y Itzhak1, I Ergui2 and JI Young2,3 The high rate of methamphetamine (METH) abuse among young adults and women of childbearing age makes it imperative to determine the long-term effects of METH exposure on the offspring. We hypothesized that parental METH exposure modulates offspring behavior by disrupting epigenetic programming of gene expression in the brain. To simulate the human pattern of drug use, male and female C57Bl/6J mice were exposed to escalating doses of METH or saline from adolescence through adulthood; following mating, females continue to receive drug or saline through gestational day 17. F1 METH male offspring showed enhanced response to cocaine-conditioned reward and hyperlocomotion. Both F1 METH male and female offspring had reduced response to conditioned fear. Cross-fostering experiments have shown that certain behavioral phenotypes were modulated by maternal care of either METH or saline dams. Analysis of offspring hippocampal DNA methylation showed differentially methylated regions as a result of both METH in utero exposure and maternal care. Our results suggest that behavioral phenotypes and epigenotypes of offspring that were exposed to METH in utero are vulnerable to (a) METH exposure during embryonic development, a period when wide epigenetic reprogramming occurs, and (b) postnatal maternal care. Molecular Psychiatry advance online publication, 18 February 2014; doi:10.1038/mp.2014.7 Keywords: behavior; cross-fostering; epigenetics; methamphetamine; prenatal exposure

INTRODUCTION Methamphetamine (METH) is a psychostimulant and a major drug of abuse in many parts of the world. Amphetamines inhibit the re-uptake of dopamine (DA), leading to reverse transport of DA from the cytoplasm to the extracellular space, resulting in a massive increase in extracellular DA.1,2 METH users suffer from depression, suicidal behavior and anxiety,3–7 and METH-induced psychosis has been the focus of several studies.8,9 The consequence of METH on cognitive function is long lasting; both abstinent and current users show significant memory impairments.10 Based on 2004–2005 and 2006–2007 surveys of past-month illicit drug use among females aged 15–44 years, 5.3–7% of these women were pregnant and continued to use drugs during all three trimesters of pregnancy; 2000–3000 abused METH.11 Although serious concerns were raised about pregnant women using METH,12 the literature on METH-exposed children is sparse compared with that on cocaine-exposed babies.13 Reported effects of METH exposure in children include reduced birth weight and size,14–17 poor growth,18 neuroanatomical changes,19–21 increased stress,22 and learning and memory deficits.23,24 These outcomes may result from direct uterine exposure to METH and also from maladaptive parenting and caregiving.25,26 Further, poor maternal care by METH-exposed women may arise not only from drug use per se but also from other socioeconomic factors.27 Owing to the scarcity of longitudinal studies on METH-exposed children, the long-term consequences of in utero METH exposure

on neurobehavioral effects in adolescence and adulthood are currently unknown. A few studies have investigated F1 rat progeny of dams exposed to METH during gestation. Male F1 rat progeny were sensitized to the psychomotor stimulating effect of METH.28,29 METH caused oxidative DNA damage in embryonic and fetal brain, leading to long-term postnatal neurodevelopmental deficits unrelated to dopaminergic neurotoxicity.30 However, aside from these reports, the consequences of prenatal METH exposure are unknown. Moreover, studies on animal models of the effect of in utero psychostimulant exposure on offspring were routinely carried out by administering the drug from gestational day 8 through 19. Although this short exposure of rats to drugs of abuse may reflect the first and second trimester in human,31 it does not simulate the human drug addict experience that starts usually in adolescence and continues through adulthood, and then through pregnancy. The goal of this study was to simulate the human pattern of parental METH exposure and its effects on the offspring. To this end, male and female mice (F0) were exposed to an intermittent escalating regimen of METH or saline from adolescence through adulthood. In adulthood, parental F0 METH produced F1 METH progeny, and parental F0 saline produced F1 saline progeny. To investigate the effect of maternal care on F1 progeny, half of F1 METH progeny was cross-fostered to saline dams and half of the F1 saline progeny was cross-fostered to METH dams. In adolescence, a subset of F1 mice was subjected to several

1 Department of Psychiatry and Behavioral Sciences, Cellular and Molecular Pharmacology and Neuroscience Division, University of Miami Miller School of Medicine, Miami, FL, USA; 2John P. Hussman Institute for Human Genomics, University of Miami Miller School of Medicine, Miami, FL, USA and 3Dr John T. Macdonald Foundation Department of Human Genetics, University of Miami Miller School of Medicine, Miami, FL, USA. Correspondence: Professor Y Itzhak or Dr JI Young, Department of Psychiatry and Behavioral Sciences, Cellular and Molecular Pharmacology and Neuroscience Division, University of Miami Miller School of Medicine, 1011 NW 15th Street, Gautier Buillding, Room 503, Miami, FL 33136, USA. E-mail: [email protected] or [email protected] Received 12 September 2013; revised 6 January 2014; accepted 9 January 2014

Consequences of parental methamphetamine exposure Y Itzhak et al

2 behavioral assessments and others were set aside for analysis of DNA methylation profile in the hippocampus. We report that both in utero METH exposure and maternal care influence adolescent offspring response to novelty and aversive and appetitive stimuli. These phenotypes are accompanied by significant changes in hippocampal DNA methylation. Our findings suggest that longterm parental METH exposure including in utero drug exposure lead to long-lasting changes in DNA methylation and behavior of the offspring, resulting from direct drug exposure and altered postnatal maternal influence.

MATERIALS AND METHODS Subjects Drug treatments and generation of F1 progeny. Eight-week-old C57Bl/6J mice (Jackson Laboratory; Bar Harbor, Maine, USA) were mated and produced 22 offspring that were born on the same day. Following weaning on postnatal day (PD) 28, mice were housed in same sex groups of 4–5 per cage. On PD32, mice were randomly assigned to two treatment groups: METH (n = 6 males and n = 5 females) and saline (n = 5 males and n = 5 females). METH-HCl (Sigma-Aldrich, St Louis, MO, USA) was dissolved in saline. Starting on PD33, mice of both sexes received every-other-day intraperitoneal (i.p.) injections of either saline or an escalating dose of METH (0.1 ml per 10 g weight) as follows: PD33–39, 0.5 mg kg − 1 (four injections per week); PD40–46, 1mg kg − 1 (three injections per week); PD47–53, 2 mg kg − 1 (four injections per week); PD54–60, 4 mg kg − 1 (three injections per week). On PD60, which was injection-free day, saline males were mated with saline females and METH males were mated with METH females. From PD61 and onward, saline and a fixed dose of METH (4 mg kg − 1) were administered to saline and METH mating pairs every-other-day, except that females now received the injections subcutaneously (s.c.). The pharmacokinetics of i.p. and s.c. METH administration are quite similar.32 Females were examined for vaginal plug every morning; by PD63–64 vaginal plugs were observed in both saline and METH females. Gestational length was 19 days for both groups; all injections stopped on gestation day 17. Body weight of saline and METH dams during gestation were recorded and subsequently pups’ weights were recorded. The average litter size of saline and METH pups was 7.8 and 7.4, respectively. To confirm the results of the first set of experiments, a second set included F0 males and females that received the same schedule of METH (n = 3/sex) and saline (n = 3/sex). Because results of the two sets of experiments were similar, the final analysis included offspring of eight METH and eight saline dams. Cross-fostering. Several studies suggest that early rearing conditions, and in particular, the quality of mother–infant interactions, may influence brain development resulting in long-lasting implications for behavioral and physiological responses through epigenetic changes in gene expression that persist into adulthood.33,34 One of our goals was to investigate the effect of fostering of METH pups to saline dams and saline pups to METH dams. We followed a fostering paradigm that described both partial and full replacement of pups.35 Within 24–36 h of parturition (PD0), litters were culled into 6–8 pups and half of the pups from each litter were transferred to an opposite-treatment dam. Consequently, each dam reared 3–4 of her own biological pups and 3–4 fostered pups. The weight of the pups was recorded from PD3 to PD27. Subsequently, saline pups (total n = 62) and METH pups (n = 60) yielded four groups (comprising both sexes) as follows: (1) saline pups reared by biological saline dams (SpSd group; males n = 14 and females n = 16); (2) saline pups fostered by METH dams (SpMd group; males n = 15 and females n = 17); (3) METH pups reared by biological METH dams (MpMd group; males n = 15 and females n = 16); (4) METH pups fostered by saline dams (MpSd group; males n = 14 and females n = 15). The four F1 groups are described as SpSd, SpMd, MpMd and MpSd; the first letter signifies the pup origin and the second letter signifies the nursing dam.

of each offspring group (METH, n = 4 and saline, n = 4) were combined. Concentrations of DA and its metabolites, 3,4-dihydoxyphenylacetic acid and homovanillic acid, in the striatum were quantitated by a modified method of high-performance liquid chromatography combined with electrochemical detection as we previously described.36

Maternal behavior Maternal behavior of METH dams (n = 5) and saline dams (n = 5 ) was monitored by an observer blind to the treatments. On PD6, PD7, PD10, PD12, PD14 and PD17 three major activities were recorded for 15 min and summarized as percent time dams spent in (1) crouch (arched back) and nursing, (2) self-grooming and (3) away from the pups.

Behavioral tests of F1’s Cocaine-induced conditioned place preference and locomotor activity. The conditioned place preference (CPP) paradigm is a useful tool for evaluating the motivational effect of drugs of abuse, drug discrimination studies and the reinstatement of drug-seeking behavior.37 It was recently reported that the male progeny of sirs who were subjected to cocaine selfadministration showed reduced propensity to cocaine selfadministration.38 Although our experimental design of parental drug exposure was different, we sought to investigate cross-sensitivity of F1 METH progeny to the rewarding effect of cocaine. Custom-designed Plexiglas cages (42(length) × 20(width) × 20(height) cm; Opto-Max Activity Meter v2.16; Columbus Instruments, Columbus, OH, USA) were used to monitor place preference. The training context consisted of two compartments separated by a divider. One compartment had blackand white-striped (2.0 cm apart) walls and a white floor covered with stainless-steel grid, whereas the other compartment had black walls and smooth black floor, thus providing distinct visual and tactile cues. Each compartment was scanned by seven infrared beams at a rate of 10 Hz (2.54 cm intervals). Time spent in each compartment and locomotor activity were recorded and analyzed by the Opto-max interface and software (Columbus Instruments, Columbus, OH, USA). Experiments were performed as we previously described.37,39 On the first day, mice were habituated (20 min) to a two-compartment apparatus. During the next 4 days, conditioning was performed in an unbiased design (cocaine was paired with either the striped or the black compartment). A morning session involved saline injection (i.p.) and confinement to one compartment (30 min); the afternoon session involved cocaine injection (i.p.; 10 mg kg − 1) and confinement to the other compartment. Time spent in each compartment was recorded (20 min) 72 h following the last training session. Half of the subjects from each of the four groups (n = 13–14) were trained for CPP during late adolescence (PD46–51) and the other half (n = 13–14) were trained in mid-adulthood (PD87–94). Routinely, each group (n = 13–14) contained a similar number of males and females. To counterbalance the order of conditioning experiments, the late-adolescent group was tested for fear conditioning in mid-adulthood; mice that were tested for CPP in mid-adulthood were first tested for fear conditioning in late adolescence. Data analysis showed neither significant ‘previousconditioning effect’ nor significant age effect, and thus results of the two sub-groups were collapsed.

Striatal DA levels

Fear conditioning. The goal of this experiment was to investigate whether in utero METH exposure and maternal care influence formation of longterm memory of an aversive stimulus. Fear conditioning training and testing occurred in Plexiglas chambers (30.5 × 30.5 × 43.5 cm; Noldus Information Technology, Leesburg, VA, USA). Each chamber was equipped with a stainless-steel rod floor through which an electric shock was delivered, and an upper control panel containing a video camera, a sound emitter and a white light illuminating one corner of the chamber. Mice were fear conditioned to an auditory cue that was followed by a single footshock (0.75 mA; 2 s) in context A, as we previously described.40 After 24 h, mice were placed in context B that differed from context A, and percent freezing during the sound of the auditory cue (2 min) was analyzed. The freezing magnitude in response to the auditory cue 24 h after training represents expression of long-term memory.

We investigated whether administration of METH to dams and offspring in utero METH exposure resulted in depletion of striatal DA and its metabolites. METH dams (n = 3) and saline dams (n = 3) were killed following pups’ weaning and the striatum was extracted. F1 METH male (n = 2) and female (n = 2) offspring of the MpMd group were killed around PD40–PD45. Saline offsprings were used as controls. Results of both sexes

Spontaneous locomotor activity. The locomotor activity cages were standard transparent rectangular rodent cages (42 × 24 × 20 cm high). Each mouse was placed in the cage for 30 min and locomotion was recorded by an activity meter (Opto-Varimex-Mini Model B; Columbus Instruments) that consisted of an array of 15 infrared emitter/detector

Molecular Psychiatry (2014), 1 – 11

© 2014 Macmillan Publishers Limited

Consequences of parental methamphetamine exposure Y Itzhak et al

3 pairs, spaced at 2.65 cm intervals, measuring activity along a single axis of motion. Spontaneous locomotor activity was recorded on PD41 and PD42. Preference for black and white compartments. The two-compartment black/dark and white/light box is used to assess anxiety-like behavior in rodents.41,42 Usually, mice spend 60–65% of their time in a dark compartment;43 strong preference for the black or white compartment suggests anxiogenic or anxiolytic response, respectively. This experiment was performed on PD42–PD43. The number of F1 subjects tested for spontaneous locomotor activity and preference for black and white compartment was as follows: SpSd, males n = 14, females n = 13; SpMd, males n = 15, females n = 14; MpMd, males n = 15, females n = 13; MpSd, males n = 14, females n = 12.

DNA methylation Adolescent (PD40–PD45) F1 females from the four groups SpSd, SpMd, MpMd and MpSd who were not subjected to behavioral tests were killed by cervical dislocation and the hippocampus was removed, snapped frozen and kept at −80 °C. Genomic DNA was extracted using a DNeasy Blood and Tissue kit (Qiagen, Valencia, CA, USA) according to manufacturer's instructions and sonicated to 200–1000 bp with a Bioruptor sonicator (Diagenode, Denville, NJ, USA). One microgram of sheared DNA was denatured and used for immunoprecipitation using a mouse monoclonal anti-5-methylcytosine antibody (Diagenode) overnight at 4 °C. Anti-mouse IgG magnetic beads (Dynabeads, Life Technologies, Gaithersburg, MD, USA) were added and incubated for additional 2 h at 4 °C. After immunoprecipitation, five washes were performed with ice-cold buffer (0.5% bovine serum albumin in PBS) and beads resuspended in Tris-EDTA buffer with 0.25% SDS and 0.25 mg ml − 1 proteinse K and incubated for 2 h at 65 °C. Methylated DNA immunoprecipitation (MeDIP)-enriched DNA was purified using Qiagene MinElute columns (Qiagen), amplified using a whole-genome Amplification kit (GenomePlex Complete WGA2, SigmaAldrich) and repurified with Qiaquick PCR purification kit (Qiagen). The NimbleGen Dual-Color DNA labeling kit (Roche NimbleGen) was used to label equal amounts of pull down DNA and input samples with Cy5 (control) or Cy3, respectively, according to manufacturer’s protocol. Microarrays (Roche NimbleGen MM8 Meth 385 K CpG plus Promoter; Roche NimbleGen), which contain ~385 000 probes covering 15 936 University of California Santa Cruz-annotated CpG islands and all RefSeq gene promoter regions, were hybridized at 42 °C for 16–20 h according to the standard procedure by Roche NimbleGen. Sample labeling, hybridization and processing were performed at Arraystar (Rockville, MD, USA). Array raw data were extracted as pair files by NimbleScan software (Roche NimbleGen). Median-centering quantile normalization and linear smoothing by Bioconductor packages Ringo, limma and MEDME were performed. From the normalized log2 ratio data, a sliding-window peak-finding algorithm provided by NimbleScan v2.5 (Roche NimbleGen) was applied to find the enriched peaks with specified parameters (sliding-window width: 750 bp; mini probes per peak: 2; P-value minimum cutoff: 2; maximum spacing between nearby probes within peak: 500 bp). The identified peaks were mapped to genomic features such as transcripts and CpG Islands. A list of the differentially enriched peaks (DEPs) that reached statistical significance is provided in Supplementary Table 1. Bisulfite sequencing. Bisulfite PCR primers were designed with the BiSearch program44 (Supplementary Table 1). Bisulfite-modified genomic DNA from each sample was amplified using AmpliTaq Gold (Applied Biosystems, Foster City, CA, USA) at annealing temperatures of 50–62 °C. The obtained PCR products were cloned into pCR2.1 TOPO vector (Invitrogen, Carlsbad, CA, USA). Colonies were screened by colony PCR and eight positive clones were picked from each single bisulfite-treated DNA. A minimum of five clones per sample (n = 15–31 clones per group) were sequenced and the data were analyzed for statistical significance.

Statistics Dams’ and pups’ weights were analyzed by two-way repeated measure analysis of variance (ANOVA) and the Greenhouse–Geisser test. Behavioral tests were usually analyzed by three-way ANOVA (sex × pup: saline/ METH × dam: saline/METH) that allowed us to determine differences between sexes, saline vs METH-exposed offspring and maternal care, and saline vs METH dams. When applicable, the Bonferroni post hoc analysis was used. Significance was set at a P-value o0.05. © 2014 Macmillan Publishers Limited

For the MeDIP studies, to compare differentially enriched regions between samples, the log2 ratio values were averaged and then used to calculate the M′ value (M_ = average(log2 MeDIPE/inputE) − average(log2 MeDIPC/ inputC)) for each probe. NimbleScan sliding-window peak-finding algorithm was run on these data to find the DEPs. The DEPs, called by the NimbleScan algorithm, were filtered according to the following criteria: (1) at least one of the groups had a median value of log2 MeDIP/Input ⩾0.3 and a median value of M′ >0 within each peak region; (2) at least half of the probes in a peak had a median value of coefficient of variability ⩽0.8 in all groups within each peak region. The bisulfite PCR sequencing data was analyzed by the Mann–Whitney analysis (VassarStats) with a significance level of α = 0.05 (Po0.05, two tailed).

RESULTS METH had no effect on body weight of dams and nursing pups Body weight of saline and METH (4 mg kg − 1) dams was recorded from gestation day 3 through 17. Two-way repeated measure ANOVA and the Greenhouse–Geisser test resulted in significant time effect F = 1087.58; P o 0.0001 and non-significant group effect F = 3.09; P = 0.122, suggesting that METH had no significant effect on body weight of dams (Supplementary Figure 1A). Therefore, the inclusion of a pair-fed group was unnecessary. The lack of weight loss in the METH dams may be because of the following factors: (a) the prolonged exposure to the drug from adolescence through gestation; and (b) the relatively low dose of METH. Likewise, it has been reported that 14 days cocaine selfadministration in adolescent rats had no effect on body weight.45 The cross-fostering design yielded four groups of pups such as SpSd (n = 30), SpMd (n = 32), MpMd n = 31 and MpSd (n = 29). Each group contained about similar number of males and females. We posit that METH dams who fostered saline pups did not expose them to METH or amphetamine (METH metabolite) through the milk because (a) the t1/2 of METH is 51–70 min and the t1/2 of amphetamine is 79–88 min (s.c. and i.p.);32,46and (b) the last METH injection was given 48 h before parturition. Pups’ weights were recorded from PD3 through PD27. Because there was no significant sex-dependent difference in pups’ weight, results of both sexes were collapsed. Two-way repeated measure ANOVA and the Greenhouse–Geisser test resulted in significant time effect F = 676.89, P o 0.0001, but non-significant group effect F = 3.89, P = 0.156, suggesting that neither in utero METH exposure nor maternal care had an effect on pups’ body weight (Supplementary Figure 1B). Striatal DA levels Levels of striatal DA in saline and METH dams were 1304 ± 123 and 1227 ± 134 ng per 100 mg tissue, respectively. Levels of striatal 3,4dihydoxyphenylacetic acid in saline and METH dams were 105 ± 21 and 128 ± 19 ng per 100 mg tissue, respectively, and levels of striatal homovanillic acid in saline and METH dams were 155 ± 21 and 141 ± 19 ng per 100 mg tissue, respectively. Because there were no statistical significant differences between saline and METH dams, results suggest that the current METH regimen did not induce dopaminergic depletion (for example, neurotoxicity). Results of the levels of DA and its metabolites in METH offspring were not significantly different from saline offspring, suggesting that in utero METH exposure did not show evidence of dopaminergic neurotoxicity in adulthood. Maternal behavior Maternal behavior was observed on PD6, PD7, PD10, PD12, PD14 and PD17. Three major activities were recorded, such as (a) crouch over the pups including nursing; (b) self-grooming; and (c) away from the pups. Time spent in each of these activities during a 15-min period was recorded and expressed as % time (Supplementary Figure 2); results were analyzed by two-repeated Molecular Psychiatry (2014), 1 – 11

Consequences of parental methamphetamine exposure Y Itzhak et al

4 measure ANOVA (treatment × time). There was no significant treatment effect for each of the activities; however, there was a significant time effect (Supplementary Figure 2). Saline (n = 8) and METH (n = 8) dams spent between 80% (PD6) and 60% (PD17) of the time in crouch position and nursing, respectively. Dams of both groups spent about 15–25% of the time self-grooming or away from the pups. In both groups, the amount of time spent with the pups decreased over time, whereas the amount of time spent away from the pups increased over time (Supplementary Figure 2). Based on the three phenotypes we observed, results suggest no significant differences between saline and METH maternal behavior. Cocaine-induced CPP and hyperactivity in F1 offspring Conditioned place preference. Results were analyzed by three-way ANOVA (sex × pup: saline/METH × dam: saline/METH). There was a significant pup effect F(1,102) = 18.5, P o 0.001; and sex effect F (1,102) = 21.8, P o 0.001. Also there was significant sex × dam

interaction F(1,102) = 29.45, P o 0.001; significant pup × dam interaction F(1,102) = 15.13, Po 0.001; and significant sex × pup × dam interaction F(1,102) = 13.45, P o 0.001. The finding that MpSd and MpMd males had significantly higher CPP than SpSd and SpMd males (Figure 1a) suggests that heightened response to cocaine-conditioned reward is the result of in utero METH exposure (pup effect) that was not influenced by maternal care. However, the response of female offspring, both saline and METH in utero exposed, was influenced by maternal care of both saline and METH dams (Figure 1a). Hence, it appears that maternal care influenced female but not male offspring response to cocaine reward. Cocaine-induced hyperactivity. Locomotor activity (30 min) was recorded during the cocaine (10 mg kg − 1) conditioning session, daily, for 4 days. A two-way ANOVA (group × time) revealed a significant group effect F(7,407) = 45.97, P o 0.0001, but nonsignificant time effect F(3,407) = 3.56, P = 0.086 (for example, lack of sensitization over time). Results of cocaine-induced hyperlocomotion are similar to results of cocaine-induced place preference. Post hoc analysis showed that MpMd and MpSd males developed significantly higher hyperlocomotion compared with SpSd and SpMd males (Figure 1b). This finding suggests again pup effect that is not influenced by maternal care. Sexual dimorphism was also observed because the response of MpMd males was higher than the female counterparts (MpMd; Figure 1b). Fear conditioning Mice were fear conditioned to an auditory cue by a single footshock (0.75 mA) in context A, and tested for cued-freezing in context B after 24 h. Differences in freezing magnitude among the groups was analyzed by three-way ANOVA (sex × pup: saline/ METH × dam: saline/METH). There was a significant (a) pup effect F

Figure 1. Cocaine-induced conditioned place preference (CPP) and hyperlocomotion and fear conditioning. Four groups such as SpSd, SpMd, MpSd and MpMd (n = 26–28 per group) each containing a similar number of males and females were tested. (a) Results of CPP were analyzed by three-way analysis of variance (ANOVA; sex × pup: saline/METH (methamphetamine) × dam: saline/METH). There was a significant pup effect F(1,102) = 18.5, P o0.01, and sex effect F (1,102) = 21.8, Po0.001. Also, there was significant sex × dam interaction F(1,102) = 29.45, P o0.001; significant pup × dam interaction F(1,102) = 15.13, P o0.001; and significant sex × pup × dam interaction F(1,102) = 13.45, Po 0.001. MpSd and MpMd males had significantly higher CPP than SpSd and SpMd males (a), which suggests that heightened CPP response was the result of in utero METH exposure (pup effect) that was not influenced by maternal care. Female offspring, both saline and METH in utero exposed, was influenced by maternal care of both saline and METH dams (*Po 0.05 between group comparison). (b) Cocaine-induced hyperactivity. Results of cocaine-induced hyperlocomotion are similar to results of cocaine-induced CPP. MpMd and MpSd males developed significantly higher hyperlocomotion compared with SpSd and SpMd males (b; *Po 0.05). Sexual dimorphism is shown by the finding that MpDp females had lower cocaine-induced locomotor activity compared with male counterparts (*Po0.05). (c) Fear conditioning. Three-way ANOVA (sex × pup: saline/METH × dam: saline/METH) resulted in significant (1) pup effect F(1,104) = 33.48, Po0.001; (2) dam effect F(1,104) = 54.76, Po 0.001; and (3) interaction of pup × dam F(1,104) = 63.39, Po 0.001. Post hoc analysis showed significantly low-freezing response in METH pups reared by METH dams (MpMd of both sexes) compared saline counterparts (gray bars *P o0.05). However, METH pups fostered by saline dams showed normal freezing response compared METH pups fostered by METH dams (#P o0.05), suggesting saline maternal effect. However, because saline pups (of both sexes) fostered by METH dams (SpMd) were not significantly different than controls SpSd (c), it appears that METH dams had not effect on offspring fear conditioning. Molecular Psychiatry (2014), 1 – 11

© 2014 Macmillan Publishers Limited

Consequences of parental methamphetamine exposure Y Itzhak et al

5

(1,104) = 33.48, P o 0.001; (b) dam effect F(1,104) = 54.76, P o0.001; and (c) interaction of pup × dam F(1,104) = 63.39; P o0.001. Post hoc analysis showed significantly low-freezing response in METH pups reared by METH dams (MpMd of both sexes) compared with all other groups (Figure 1c). The finding that METH pups fostered by saline dams did not show deficit in freezing response (Figure 1c) suggests that saline dams may restore the dampened response of METH pups. However, because saline pups (of both sexes) fostered by METH dams (SpMd) were not significantly different than controls SpSd (Figure 1c), it appears that METH dams had no effect on offspring fear conditioning. F1 spontaneous locomotor activity and time in black compartment Locomotion. On PD41 and PD42, spontaneous locomotor activity was recorded. Because sex-dependent differences in locomotor activity were not observed, results of both sexes were collapsed and analyzed by two-way ANOVA (pup saline/METH × dam saline/ METH). There was a significant dam effect F(1,110) = 35.57, Po0.0001 but not pup effect F(1,101) = 1.51, P = 0.221. Pair-wise comparisons showed that locomotor activity of the MpMd and the SpMd groups was significantly lower than SpSd and MpSd groups (Figure 2a). Results suggest that maternal care of METH dams induced lower spontaneous locomotor activity, regardless whether pups were exposed to METH or saline in utero. If we assume that METH pups do not have motor impairment, then it appears that saline dams had no effect on motor behavior because locomotor activity of SpSd and MpSd groups was similar (Figure 2a). Time in black compartment. Mice usually spend 60–65% of their time in a dark/black compartment.43 On PD45 and PD46, percent time spent in the black compartment (15 min period) was analyzed by three-way ANOVA (sex × pup: saline/METH × dam: saline/METH). There was a significant sex effect F(1,96) = 85.00, P o0.0001; significant interactions of sex × dam F(1,96) = 37.80, P o0.0001; sex × pup F(1,96) = 39.11, P o0.0001; and sex × dam × pup F(1,96) = 34.50, P o 0.0001. Results depicted in Figure 2b shows that SpMd males spent significant amount of time in the black zone, whereas SpMd females spent the least amount of time in the black zone. Results suggest that maternal care of METH dams induced significant sex-dependent effect on anxiety-like behavior of saline offspring. DNA methylation One possible mechanism by which prenatal exposures affect longterm phenotypes later in life involves alterations of DNA methylation marks in the genome. To determine the role of DNA methylation in regulating behavioral responses following in utero METH exposure, we compared the DNA methylome of the hippocampus of the F1 SpSd, MpMd, SpMd and MpSd mice. Although the nucleus accumbens is the major reward substrate in brain, we chose to investigate the hippocampus because it has a border role in cognition and emotional behavior, and little is known about the role of the hippocampus in the motivational effects of psychostimulants (see Discussion for more details). DNA methylation patterns were assayed by MeDIP-chip. The MeDIP procedure is based on the enrichment of methylated DNA with an antibody that specifically binds to 5-methylcytidine that was followed by hybridization to a microarray containing all annotated CpG Islands and promoter regions (from about −1.3 to 0.5 kb of the transcriptional start site). We identified an average of 5135.5 methylation-enriched peaks in gene promoters and 6279.5 methylation-enriched peaks in CpG islands per sample (Supplementary Table 3). The group that exhibited the highest number of identified peaks was MpMd, suggesting that the combination of in utero exposure to METH and METH-induced maternal care mostly promotes DNA methylation. A comparison of © 2014 Macmillan Publishers Limited

Figure 2. Maternal care influenced spontaneous locomotor activity and time spent in black compartment. Four groups such as SpSd, SpMd, MpSd and MpMd (n = 26–28 per group) each containing a similar number of males and females were tested. (a) Reduced locomotor activity was observed in the SpMd and MpMd groups compared with SpSd and MpSd (*P o0.05) groups, suggesting METH maternal effect on spontaneous locomotor activity. Because no significant sex effect was observed, results of both sexes were collapsed. (b) SpMd males spent significantly more time in the black zone compared with SpSd males (*P o0.05), whereas SpMd females spent less time in the black zone compared with SpSd females (*P o0.05), suggesting that METH maternal care induced significant sex-dependent effect on anxiety-like behavior of saline offspring.

MpMd vs SpSd samples identified 1822 methylation peaks aligning to gene promoter regions and 1808 peaks aligning to CpG islands as differentially methylated regions (DMR). The majority of these DMR showed hypermethylation in MpMd when compared with SpSd (Supplementary Table 4). When applying the extremely stringent criteria of looking at peaks that were present in all three samples of either MpMd or SpSd and absent in all three samples of the comparison group, we restricted the list of DMR to 545 CpG islands and 156 promoter regions hypermethylated in MpMd and 209 CpG islands and 46 promoters with highest methylation in SpSd (Table 2); 75% of the DMR identified by these stringent criteria showed methylation differences by bisulfite sequencing concordant with the array data (Figure 3). Of note, there is more variability in methylation levels in samples from maternally METH-exposed mice than in SpSd samples, suggesting that the response of hippocampal cells to maternal METH exposure is heterogeneous. As the physiological relevance of intergenic CpG islands is unknown and the promoter-associated CpG islands showed a methylation profile similar to the promoter Molecular Psychiatry (2014), 1 – 11

Consequences of parental methamphetamine exposure Y Itzhak et al

6 hippocampal transcriptional landscape of the progeny; promoters that are usually active become methylated, whereas inactive promoters are demethylated. The most enriched gene ontology terms in the DMR were ‘cerebral cortex GABAergic interneuron differentiation’ for the hypermethylated (P = 0.002 after Bonferroni correction for multiple comparisons) and ‘embryonic development’ for the hypomethylated (P = 0.01 after Bonferroni), suggesting that the observed phenotypes could be the result of abnormal brain development. Classification according to the Kyoto Encyclopedia of Genes and Genomes pathway placing identified enrichment in transcriptional regulation (P = 0.0003) and the mitogen-activated protein kinase (Mapk) signaling pathway (P = 0.04) for hypermethylated and hypomethylated DMR in MpMd as compared with SpSd, respectively.

Figure 3. Methylation analysis by bisulfite sequencing. (a) Representative alleles for Bcl7c and Col24a1 for bisulfite sequencing of DNA from a SpSd and MpMd hippocampus sample. Each circle along a horizontal row depicts a CpG site in a cloned allele with white representing unmethylated and black representing methylated. (b) Percent methylation (y axis) per clone and loci on the x axis. An average of 18 alleles were sequenced for each group. *Po 0.05, **Po 0.01.

in the vast majority of analyzed DMR, we focused our report on promoter DMR. To differentiate between the effects that arise from direct METH exposure and maternal care, we compared the methylation profiles from all experimental groups (Supplementary Figure 2). We identified 62 and 35 promoter regions in which DNA methylation was elevated and reduced, respectively, in the F1 mice due to in utero METH exposure that was insensitive to maternal-fostering effects: DMR found in MpMd vs SpSd that were also found in MpMd vs SpMd and MpSd vs SpSd but were not observed in the MpMd vs MpSd comparison (Table 1). We also identified DMR induced by METH-related maternal effects: 70 hypermethylated and 39 undermethylated loci identified by their presence in MpMd vs SpSd and also in SpMd vs SpSd (methylation), and in SpSd vs SpMd and also in SpSd vs MpMd (demethylation) (Table 1). We performed bioinformatic analysis of the SpSd vs MpMd DMR associated to gene promoters using the software application Enrichr.47 Notably, enrichment analysis using histone modifications chromatin immunoprecipitation-seq gene-set libraries showed that the most enriched terms were those associated with H3K4me3 for the parental exposure to METH-induced methylation events and H3K27me3 when the input was the demethylated DMR (Table 2). These signature patterns of the unbiased promoter lists are rather interesting, as the H3K4me3 and H3K27me3 modifications are usually associated with transcribed and silenced promoters, respectively, whereas promoter methylation is generally non-permissive for transcription.48 Thus, these data suggest that parental METH exposure has important consequences on the Molecular Psychiatry (2014), 1 – 11

DISCUSSION Current use of METH surpasses the use of cocaine and opiates,49 and ~7% of METH users are pregnant women.50 However, little is known about the impact of prenatal METH exposure on developmental outcomes of the progeny.51,52 This is mostly because of the difficulty of investigating the effects of maternal drug exposure in humans that are often multidrug users, and their response could be influenced by a multitude of factors such as frequency, dose and route of drug exposure, and social and genetic determinants. We sought to conduct preclinical studies in mice to investigate the long-term behavioral consequences induced by parental METH exposure and the molecular alterations associated with the behavioral phenotypes. We used a method of drug exposure that mimics human drug use: escalating doses starting at adolescence and continuing during pregnancy. The extrapolation of daily METH dosage from human drug users to animals may be a challenge because human drug use varies upon the degree of addiction, route of drug intake, availability, cost and so on. Among METH users in the United States, selfreported drug use was 13–22 days within a 30-day period,53 averaging 3.1 g a week,54 which equals about 6.1 mg kg − 1 per day. In another study, self-reported METH use frequency was 5.3 days per week at a dose of 0.53 g per day (median), which equals about 7.5 mg kg − 1 per hit. 55 Based on these reports and a direct mg per kg dose comparison, the doses of METH given to mice (0.5–4.0 mg kg − 1 METH every-other-day) do not exceed the quantities of METH abused by humans. Exposure to drugs of abuse before and during pregnancy could potentially affect the two developmental periods in which epigenetic reprogramming occurs genome wide, such as germ cells and preimplantation embryos. Our results indicate that parental METH exposure, in a manner that resembles human use, produces alterations in offspring’s behavior and DNA methylation patterns in the hippocampus. The data demonstrate that both, the epigenotype and the behavioral phenotype of the offspring, are vulnerable to METH exposure during embryonic development, a period when wide epigenetic reprogramming occurs. In addition, the cross-fostering experiments suggest that this vulnerability extends into the postnatal period as both, the DNA methylome and the behavior of the pups, were modulated by postnatal maternal influences. Thus, we show that METH exposure before and during pregnancy results in significant lasting effects on DNA methylation that consequently may influence gene expression and produce abnormal phenotypes across the life course. Our results suggest that the aberrant behavioral phenotypes and DNA methylation we observed in F1 offspring are not due to dopaminergic neurotoxicity. Levels of striatal DA and its metabolites 3,4-dihydoxyphenylacetic acid and homovanillic acid were not significantly different between (a) METH and saline dams, and (b) F1 METH and saline offspring. In our previous studies, METHinduced dopaminergic neurotoxicity was observed following acute administration of three injections of METH (5 mg kg − 1) in © 2014 Macmillan Publishers Limited

Consequences of parental methamphetamine exposure Y Itzhak et al

7 Table 1.

Promoter methylation changes induced by either METH exposure or maternal care

Methylation induced by METH exposure

Demethylation induced by METH exposure

Methylation induced by METH-altered maternal care

Demethylation induced by METH-altered maternal care

Methylation induced by interaction of METH exposure and METHaltered maternal care

Demethylation induced by interaction of METH exposure and METH-altered maternal care

1110054O05Rik 1700020O03Rik 4933433P14Rik Adora1 Aimp1 Akap5 Atg2b Atp5s B4galnt2 BC048679 Bcl7c Cacna1g Cdc23 Chrm4 Cnga3 Cnst Cpz Dhx16 Elk3 Eme2 Exosc6 F11r Fuca1 Gatad2a Gdap1l1 Ggct Glyr1 Gm266 Gnb1l Gpatch3 Grinl1a Hrk Hspb8 Ilk Kcnab2 L2hgdh Mir762 Mrps34 Mst1 Nphp4 Nudt16l1 Pcdhgc3 Pcf11 Pgam1 Rrp8 Sep15 Six6 Snx7 Sorbs1 Srsf5 Tatdn2 Tbck Tex261 Tfb2m Tia1 Tsen34 Txnrd3 Ube1y1 Ubn1 Vbp1 Zc3h7a Zic3

1110008L16Rik 7420426K07Rik Accn4 Aldh3b1 Allc Ankrd23 Armc7 Atp2c1 BC020402 Bex1 C1ql4 Camkk2 Car9 Cldn19 Crhr2 Cyp4f15 Ddhd2 Dlgap2 Dpp6 Fam100a Gas2l3 Lsp1 Mpped2 Mysm1 Nell1 Pdia5 Pnck Ppp2r3c Rarg Rusc2 Smurf1 Socs1 Timm13 Trpm4 Wdr12

1700019N12Rik 1700037C18Rik 4921504E06Rik 5930434B04Rik Abhd11 Acot13 Adcy8 Ahsg Aph1b AW209491 Btg2 Cacng4 Cd2ap Cep57 Cib1 Cln8 Cluap1 Ctnnd2 Cygb D330012F22Rik Daam1 Dennd2a Depdc6 Dusp7 Ednra Eef1d Fam171b Fam76b Gdf11 Gpd2 Gpr35 H2-Ke6 Hadha Hadhb Irak4 Lhx2 Lhx4 Lox Mdc1 Nat10 Ndrg3 Pabpc5 Papss1 Pax8 Pcdh8 Plrg1 Pomc Prss28 Ptma Pus7l Rarres2 Rtn1 Sec24d Shfm1 Slc38a4 Stat6 Stoml2 Tbkbp1 Tbx5 Tdp2 Tkt Tmem147 Tmem57 Tpm4 Trabd Vstm2a Zfp191 Zfp316 Zfp346 Zfp384

4930539J05Rik Abca3 Add3 Atp5l B3galtl BC022687 C030019I05Rik Camk2a Catsperg1 Cbfa2t3 Celf1 Dnajc13 Fbrsl1 Fbxl3 Fndc4 Hdac11 Lrrc4b Lrrc8b Mapk12 Mier2 Ndufb4 Nkx2-1 Oas1a Slc46a1 Slc6a1 Slc9a7 Taf13 Timm50 Tmem176a Tmem176b Trhr2 Uba6 Uqcrq

1110002L01Rik 1300001I01Rik 2610204G22Rik 3010026O09Rik 3110001I22Rik Aagab Aasdhppt Anapc7 Ankrd9 Arhgap29 Asxl2 Atad3a Atl1 Atxn7l1 Bfar Cdca4 Cenpp Ctdsp2 E2f8 Edem3 Ehmt1 F730043M19Rik Fam19a2 Fastkd3 Gatsl3 Gsk3b Insm2 Iqch Irak1 Jmy Kbtbd3 Kcnmb4 Ldoc1l Lemd1 Map4k5 Mdfi Mga Mir718 Mtrr Ncapd3 Nol8 Nrarp Nup50 Pcx Pds5b Pgls Raver1 Sesn2 Sorcs1 Srrm1 Ssr4 Tgfb3 Tifab Tmem39b Ubxn4 Vamp4 Vps26b Zfp191 Zfp346

4833439L19Rik Acer1 Arhgap23 Btbd11 Cacna1c Col24a1 Gyltl1b Hdac5 Hipk2 Hras1 Limch1 Lmna Rbp1 Rps8 Snord38a Snord55 Spink2 Spint2 Srgap1 Strn Zdhhc2

Abbreviations: DMR, differentially methylated regions; Mapk, mitogen-activated protein kinase; METH, methamphetamine. Last two columns contain top 5% of DMR.

© 2014 Macmillan Publishers Limited

Molecular Psychiatry (2014), 1 – 11

Consequences of parental methamphetamine exposure Y Itzhak et al

8 Table 2.

Enriched histone modification signatures of the SpSd vs MpMd promoter DMR P-value

Z-score

Combined score

1 H3K27me3 fetal lung 8.73e–4 2 H3K27me3 H1 BMP4 derived 3.51e–3 trophoblast cultured cells 3 H3K27me3 H1 0.01 4 H3K27me3 colonic mucosa 0.02 5 H3K27me3 stomach smooth 0.03 muscle 6 H3K27me3 CD4 memory 0.04 primary cells 7 H3K27me3 CD4 naive primary 0.03 cells 8 H3K27me3 CD3 primary cells 0.06 9 H3K27me3 penis foreskin 0.04 keratinocyte primary cells 10 H3K27me3 CD8 memory 0.06 primary cells

−1.52 −1.65

10.68 9.34

−1.87 −1.69 −1.40

8.59 6.08 4.74

−1.46

4.60

−1.29

4.48

−1.45 −1.27

4.05 4.00

−1.39

3.89

1 H3K4me3 H1 2 H3K4me3 H1 derived mesenchymal stem cells 3 H3K4me3 H9 4 H3K4me3 iPS DF 6.9 5 H3K4me3 CD8 naive primary cells 6 H3K4me3 iPS DF 19.11 7 H3K4me3 CD19 primary cells 8 H3K9ac H1 9 H3K4me3 H1 BMP4 derived trophoblast cultured cells 10 H3K4me3 IMR90

3.93e–22 1.10e–14

−2.26 −2.60

98.08 70.01

1.68e–12 4.61e–12 1.59e–13

−2.84 −2.79 −2.39

64.02 60.87 58.81

3.63e–10 7.63e–10 4.42e–8 7.84e–8

−2.81 −2.64 −3.10 −2.84

49.50 44.87 40.60 35.95

1.77e–6

−2.95

30.24

Index

Name

A

B

Abbreviation: DMR, differentially methylated regions. (A) Top 10 enriched histone terms ranked based on the level of significance for the hypermethylated DMR. (B) Top 10 enriched histone terms ranked based on the level of significance for the demethylated DMR.

a single day.56 It has also been reported that pretreatment with intermittent low doses of METH protects against dopaminergic and serotonergic neurotoxicity caused by higher doses of METH.57 Hence, it appears that our schedule of METH administration did not produce dopaminergic neurotoxicity. The behavioral characterization showed several aberrations in the F1 offspring that could be a direct consequence of germ cell, in utero METH exposure, indirect result of altered maternal behavior or the combination of all. Although our investigation of maternal behavior did not reveal significant differences between METH and saline dams with reference to nursing, selfgrooming and time spent away from the pups, results of the crossfostering studies suggest that some behavioral phenotypes of F1 progenies were influenced by maternal care. The influence of cross-fostering within the same strain or between different mouse strains is a matter of debate; some suggest no significant differences in pup development,58 whereas others suggest some increased nursing of C57Bl/6 fostered pups by C57Bl/6 dams.59 In the current study, we did not attempt to investigate the effect of cross-fostering per se. It should be noted, however, that body weight of all pups from PD3 through adolescence (PD27), regardless of sex, in utero METH exposure and fostering dam, was the same in all four groups (Supplementary Figure 1B), suggesting that in utero METH exposure and maternal care had no effect on body weight. The response of offspring to cocaine reward was determined by the CPP paradigm. The findings that MpMd and MpSd males Molecular Psychiatry (2014), 1 – 11

showed significant higher CPP (Figure 1a) and cocaine-induced hyperlocomotion (Figure 1b) compared with SpSd and SpMd males suggest that prenatal METH exposure of males render them highly sensitive to cocaine effect. The enhanced response of F1 METH males to cocaine-conditioned reward is again suggestive of a lack of METH-induced neurotoxicity, which significantly diminishes cocaine CPP in mice.56 Sex-dependent differences were observed as MpMd males were more sensitive to cocaine than the female counterparts in both CPP (Figure 1a) and locomotor activity (Figure 1b) experiments. Maternal influence of METH and saline dams on female offspring response to cocaine was observed. Together, these findings suggest that the response of female— but not male— offspring to cocaine reward was modulated by maternal care. The enhanced sensitivity of F1 METH males to cocaine reward is seemingly different than the reduced response to cocaine of male offspring of sirs who were exposed to cocaine self-administration.38 However, because in the latter study offspring were not exposed to the drug in utero it is difficult to compare the results of these two studies. DNA methylation is an important epigenetic mechanism regulating cocaine-induced structural plasticity.60 Although the nucleus accumbens is the major reward substrate in brain, we chose to investigate the hippocampus because it has a border role in cognition and emotional behavior, and little is known about the role of the hippocampus in the motivational effects of cocaine.61 The dorsal hippocampus is primarily involved in spatial learning and the ventral hippocampus is associated with motivational and emotional behavior.62,63 Relevant to CPP studies, ablation or blockade of dorsal hippocampus precluded formation of cocaine CPP,64,65 and evidence supports the role of hippocampal DA in long-term memory of motivationally significant events.66 Ventral CA1 projections to the medial nucleus accumbens shell was predominately and selectively potentiated after cocaine exposure,67 supporting the role of VHP in motivational effects of cocaine. Ventral CA1 innervates regions of the prefrontal cortex, that is, the prelimbic and medial orbital areas that project to the hypothalamus, supporting the role of ventral hippocampus in stress response.68 Therefore, we posit that changes in DNA methylation in the hippocampus are relevant for appetitive and aversive learning and memory. Acquired fear response, following classical Pavlovian fear conditioning, is primarily dependent on the integrity of the hippocampus and amygdala. We found that both male and female METH offspring fostered by METH dams showed significant lower cued-freezing response (Figure 1c). The dampened freezing response was ‘rescued’ by saline dams. The finding that saline pups fostered by METH dams (SpMd) were not different than control pups (SpSd) suggests that the reduced freezing response of the METH pups is primarily because of in utero METH exposure that was not influenced by maternal care of METH dams. Thus, cognitive deficits in METH offspring could be the result of in utero METH exposure. Although it is well established that cued fear memory is primarily amygdala dependent, several studies have shown that hippocampal lesions may also impair cued fear memory.69–72 Given the role of hippocampal DNA methylation in learning and memory,73 the deficits in fear conditioning of MpMd offspring could have a relationship with aberrant hippocampal DNA methylation we observed. It appears, however, that METHinduced aberrant DNA methylation may be offset by maternal care, because METH pups fostered by saline dams did not show the deficits in cued fear memory. Additional studies are needed to investigate the profile of DNA methylation in the amygdala and other brain regions involved in acquisition of fear memory (for example, prefrontal cortex). Although maternal effect of METH dams on offspring response to cocaine-conditioned response and hyperlocomotion as well as © 2014 Macmillan Publishers Limited

Consequences of parental methamphetamine exposure Y Itzhak et al

9 fear conditioning was unnoticeable, the influence of METH dams on spontaneous locomotor activity (Figure 2a) and anxiety-like behavior (Figure 2b) is evident. Both saline pups and METH pups fostered by METH dams showed significantly lower spontaneous locomotor activity compared with saline and METH pups fostered by saline dams (Figure 2a). In the choice between black and white compartments of a cage, METH dams had significant effect on saline pups but not on METH pups. The former suggests that METH dams induced anxiogenic-like phenotype in saline males (preferring black) and anxiolytic-like phenotype in saline females (preferring white). It has been shown that early-life experience and maternal care influence hippocampal DNA methylation profile in rats and humans.74 Thus, the detected maternal METH effects on saline offspring behavior may be related to the effect of METH dams on hippocampal DNA methylation of the offspring we observed (Supplementary Figure 2). The literature on METH-exposed children is sparse compared with that on cocaine-exposed babies;13 and therefore it is difficult to identify specific behavioral phenotypes in our animal model that may correspond to abnormalities in children exposed to METH in utero. However, reported effects of METH exposure in children include reduced birth weight and size,14–17 poor growth,18 neuroanatomical changes,19–21 increased stress,22 and learning and memory deficits.23,24 The latter may relate to the deficits in fear conditioning we observed (Figure 1c). However, longitudinal studies on children that were exposed to METH in utero and their susceptibility to drug abuse in adolescence are missing but undoubtedly essential. Examination of the methylation status of hippocampal DNA of female subjects identified several interesting candidates as mediators of the aberrant behaviors. The DMR identified showed a remarkable different profile depending on the methylation change. METH-induced methylation, which should produce gene silencing, occurs in promoters that have a histone modification signature of active transcription. Conversely, METH-induced demethylation was observed in promoters associated with an epigenetic signature of transcriptional silencing. The mechanism by which prenatal METH exposure alters the epigenotype is not known. However, our data suggest that it likely involves alterations at the levels of wide-range chromatin conformation, rather than modifying the expression of specific methyltransferases, demethylases or their regulators. Concurrent with this premise is the finding that altered hippocampal DNA methylation in F1 offspring following in utero cocaine exposure (gestational days 8–19) was observed in the absence of alterations in expression of DNMTs.75 Functional categorization of the DMR identified in the SpSd vs MpMd comparison, regardless of whether the effects are because of a direct METH exposure or maternal care, revealed enrichment of genes involved in embryonic development and pattern specification. This suggests that the observed phenotype may result from abnormal development. The phenotypic response to prenatal exposure of METH is likely the result of the compound activity of a large set of dysregulated genes, rather than the product of a handful of genes. Our observation of promoter DMR enrichment in terms such as transcriptional regulation and MAPK signaling pathways supports this idea. Hypermethylation of CpG islands in gene promoter of MAPK1 was also detected in the hippocampus of prenatally cocaine-exposed offspring.75 Also, some of the identified DMR are associated with genes that could be central to the alterations of these pathways. These include histone modifiers such as deacetylases Hdac5 and Hdac11, methyltransferase Ehmt1, H2A deubiquitinase Mysm1, and hub kinases of the Mapk pathway such as Hras1 and Mapk12. Currently, it is unclear how exactly these DNA methylation changes correlate with the behavioral observations; further studies are necessary to determine the role they have in the development of the aberrant behavioral phenotypes. Changes in © 2014 Macmillan Publishers Limited

DNA methylation may not be the sole cause for the changes in behavior we observed. Numerous changes in synaptic plasticity because of in utero drug exposure may have long-term effect on the offspring; additional studies are necessary to investigate the consequences of parental METH exposure on the offspring. We identified DMR located in promoters of genes already linked to drug response, such as Adora1 (METH and cocaine76), ILK (integrin-linked kinase; cocaine77), Akap5 (A-kinase anchoring protein-5; cocaine78), Camkk2 (METH79), Hdac5 (cocaine and METH80), Pomc (heroin81), Camk2a (METH and cocaine,82 Gsk-3-β (glycogen synthase kinase 3 beta; cocaine83,84) and PCDH8 (cocaine85). These findings suggest that in utero METH exposure may induce changes in several genes that are modulated by direct exposure to other drugs of abuse. Some of these genes have critical role in synaptic plasticity and learning and memory. Alteration in ILK–glycogen synthase kinase 3 beta (Gsk-3-β) signaling was identified in an animal model associated with memory loss and Alzheimer's disease.86 A-kinase anchoring proteins (APAKs) constitute a family of scaffolding proteins that bind to protein kinase A and regulate the phosphorylation of various proteins that have been implicated in synaptic plasticity and memory consolidation.87 Likewise calcium/calmodulin-dependent kinases (Camkk2 and Camk2a) and histone deacetylase 5 (Hdac5) have been implicated in synaptic plasticity and learning and memory. Therefore, we postulate that DMR located in promoters of these genes may have profound implications on behavioral phenotypes associated with appetitive- and aversiveassociative learning we observed. We observed that the methylation of a large number of gene promoters was changed as a result of fostering by METH dams. This finding suggests that METH maternal behavior significantly contributes to the offspring epigenetic response. Interestingly, we identified DMR in gene promoters in which the methylation or expression pattern was already reported to be sensitive to maternal influences, making them good candidates to mediate the maternal METH influences. For instance, the methylation status of the Pomc (pro-opiomelanocortin) and Cln8 (claudin 8) genes is altered by intrauterine exposure to alcohol88 and maternal diabetes,89 respectively. Maternal effect on expression of Rtn1 in the offspring was reported for maternal hypothyroidism,90 and exposure to maternal diabetes resulted in dysregulation of TBX5.91 The data suggest that maternal effects on the offspring induced by a variety of stimuli (exposure to drugs, disease states and nutritional status) may be mediated by a common set of genes. A very likely candidate for this mediation is Pomc that is capable of sensing homeostatic perturbations and participates in the modulation of behavioral responses such as anxiety, cognition and response to drugs.92,93 In summary, the current study demonstrates that prenatal METH exposure produces long-lasting changes in the offspring brain epigenome that could contribute to the initiation and maintenance of the observed behavioral phenotypes. Moreover, our experiments indicate a significant influence of maternal effects on the epigenotype and behavioral phenotype of the F1 progeny. Future studies will determine whether sex-dependent behavioral phenotypes parallel sex-dependent changes in hippocampal DNA methylation, and whether changes in epigenome and behavior due to METH exposure are trans-generationally transmitted to F2 and F3 generations. CONFLICT OF INTEREST The authors declare no conflict of interest.

ACKNOWLEDGMENTS This work was supported in part by grant RO1DA026878 and R21DA029404 from the National Institute on Drug Abuse, National Institutes of Health (to YI) and award from

Molecular Psychiatry (2014), 1 – 11

Consequences of parental methamphetamine exposure Y Itzhak et al

10 the University of Miami Specialized Center of Research on Addiction and Health in Women, Children and Adolescence (to YI & JIY). We are thankful for the excellent technical support of Karen L Anderson, Michael W Kaplan and Shervin Liddie.

REFERENCES 1 Kuczenski R. Effects of phospholipases on the kinetic properties of rat striatal membrane-bound tyrosine hydroxylase. J Neurochem 1983; 40: 821–829. 2 Sulzer D, Chen TK, Lau YY, Kristensen H, Rayport S, Ewing A. Amphetamine redistributes dopamine from synaptic vesicles to the cytosol and promotes reverse transport. J Neurosci 1995; 15: 4102–4108. 3 Darke S, Kaye S, McKetin R, Duflou J. Major physical and psychological harms of methamphetamine use. Drug Alcohol Rev 2008; 27: 253–262. 4 Homer BD, Solomon TM, Moeller RW, Mascia A, DeRaleau L, Halkitis PN. Methamphetamine abuse and impairment of social functioning: a review of the underlying neurophysiological causes and behavioral implications. Psychol Bull 2008; 134: 301–310. 5 Pennay AE, Lee NK. Putting the call out for more research: the poor evidence base for treating methamphetamine withdrawal. Drug Alcohol Rev 2011; 30: 216–222. 6 Marshall BD, Werb D. Health outcomes associated with methamphetamine use among young people: a systematic review. Addiction 2010; 105: 991–1002. 7 Ellison G. Stimulant-induced psychosis, the dopamine theory of schizophrenia, and the habenula. Brain Res Brain Res Rev 1994; 19: 223–239. 8 Grant KM, LeVan TD, Wells SM, Li M, Stoltenberg SF, Gendelman HE et al. Methamphetamine-associated psychosis. J Neuroimmune Pharmacol 2012; 7: 113–139. 9 Grelotti DJ, Kanayama G, Pope HG Jr. Remission of persistent methamphetamineinduced psychosis after electroconvulsive therapy: presentation of a case and review of the literature. Am J Psychiatry 2010; 167: 17–23. 10 Volkow ND, Chang L, Wang GJ, Fowler JS, Leonido-Yee M, Franceschi D et al. Association of dopamine transporter reduction with psychomotor impairment in methamphetamine abusers. Am J Psychiatry 2001; 158: 377–382. 11 Substance Abuse and Mental Health Services Administration (SAMHSA), Office of Applied Studies. Results from the 2007 National Survey on Drug Use and Health: National Findings (NSDUH Series H-34, DHHS Publication No. SMA 08-4343). Rockville, MD, USA, 2008. 12 American College of Obstetricians and Gynecologists Committee on Health Care for Underserved Women. Committee Opinion No. 479: Methamphetamine abuse in women of reproductive age. Obstet Gynecol 2011; 117: 751–755. 13 Lester BM, LaGasse LL. Children of addicted women. J Addict Dis 2010; 29: 259–276. 14 Chomchai C, Na Manorom N, Watanarungsan P, Yossuck P, Chomchai S. Methamphetamine abuse during pregnancy and its health impact on neonates born at Siriraj Hospital, Bangkok, Thailand. Southeast Asian J Trop Med Public Health 2004; 35: 228–231. 15 Dixon SD, Bejar R. Echoencephalographic findings in neonates associated with maternal cocaine and methamphetamine use: incidence and clinical correlates. J Pediatr 1989; 115: 770–778. 16 Little BB, Snell LM, Gilstrap LC3rd. Methamphetamine abuse during pregnancy: outcome and fetal effects. Obstet Gynecol 1988; 72: 541–544. 17 Smith LM, LaGasse LL, Derauf C, Grant P, Shah R, Arria A et al. Prenatal methamphetamine use and neonatal neurobehavioral outcome. Neurotoxicol Teratol 2008; 30: 20–28. 18 Eriksson M, Jonsson B, Steneroth G, Zetterström R. Cross-sectional growth of children whose mothers abused amphetamines during pregnancy. Acta Paediatr 1994; 83: 612–617. 19 Chang L, Alicata D, Ernst T, Volkow N. Structural and metabolic brain changes in the striatum associated with methamphetamine abuse. Addiction 2007; 102: 16–32. 20 Chang L, Smith LM, LoPresti C, Yonekura ML, Kuo J, Walot I et al. Smaller subcortical volumes and cognitive deficits in children with prenatal methamphetamine exposure. Psychiatry Res 2004; 132: 95–106. 21 Cloak CC, Ernst T, Fujii L, Hedemark B, Chang L. Lower diffusion in white matter of children with prenatal methamphetamine exposure. Neurology 2009; 72: 2068–2075. 22 Smith L, Yonekura ML, Wallace T, Berman N, Kuo J, Berkowitz C. Effects of prenatal methamphetamine exposure on fetal growth and drug withdrawal symptoms in infants born at term. J Dev Behav Pediatr 2003; 24: 17–23. 23 Struthers JM, Hansen RL. Visual recognition memory in drug-exposed infants. J Dev Behav Pediatr 1992; 13: 108–111. 24 Chang L, Cloak C, Jiang CS, Farnham S, Tokeshi B, Buchthal S et al. Altered neurometabolites and motor integration in children exposed to methamphetamine in utero. Neuroimage 2009; 48: 391–397. 25 Derauf C, LaGasse LL, Smith LM, Grant P, Shah R, Arria A et al. Demographic and psychosocial characteristics of mothers using meth-amphetamine during pregnancy: preliminary results of the infant development, environment, and lifestyle study (IDEAL). Am J Drug Alcohol Abuse 2007; 33: 281–289.

Molecular Psychiatry (2014), 1 – 11

26 Shah R, Diaz SD, Arria A, LaGasse LL, Derauf C, Newman E et al. Prenatal methamphetamine exposure and short-term maternal and infant medical outcomes. Am J Perinatol 2012; 29: 391–400. 27 Good MM, Solt I, Acuna JG, Rotmensch S, Kim MJ.. Methamphetamine use during pregnancy: maternal and neonatal implications. Obstet Gynecol 2010; 116: 330–334. 28 Bubenikova-Valesova V, Kacer P, Syslova K, Rambousek L, Janovsky M, Schutova B et al. Prenatal methamphetamine exposure affects the mesolimbic dopaminergic system and behavior in adult offspring. Int J Dev Neurosci 2009; 27: 525–530. 29 Šlamberová R, Yamamotová A, Schutová B, Hrubá L, Pometlová M. Impact of prenatal methampheta-mine exposure on the sensitivity to the same drug in adult male rats. Prague Med Rep 2011; 112: 102–114. 30 Jeng W, Wong AW, Ting-A-Kee R, Wells PG. Methamphetamine-enhanced embryonic oxidative DNA damage and neurodevelopmental deficits. Free Radic Biol Med 2005; 39: 317–326. 31 Quinn R. Comparing rat’s to human’s age: How old is my rat in people years? Nutrition 2005; 21: 775–777. 32 Gentry WB, Ghafoor AU, Wessinger WD, Laurenzana EM, Hendrickson HP, Owens SM. (+)-Methamphetamine-induced spontaneous behavior in rats depends on route of (+)METH administration. Pharmacol Biochem Behav 2004; 79: 751–760. 33 Meaney MJ. Maternal care, gene expression, and the transmission of individual differences in stress reactivity across generations. Annu Rev Neurosci 2001; 24: 1161–1192. 34 Weaver IC, Cervoni N, Champagne FA, D’Alessio AC, Sharma S, Seckl JR et al. Epigenetic programming by maternal behavior. Nat Neurosci 2004; 7: 847–854. 35 Hickman DL, Swan MP.. Effects of age of pups and removal of existing litter on pup survival during cross-fostering between multiparous outbred mice. J Am Assoc Lab Anim Sci 2011; 50: 641–646. 36 Itzhak Y, Gandia C, Huang PL, Ali SF.. Resistance of neuronal nitric oxide synthasedeficient mice to methamphetamine-induced dopaminergic neurotoxicity. J Pharmacol Exp Ther 1998; 284: 1040–1047. 37 Itzhak Y, Martin JL. Cocaine-induced conditioned place preference in mice: Induction, extinction and reinstatement by related psychostimulants. Neuropsychopharmacology 2002; 26: 130–134. 38 Vassoler FM, White SL, Schmidt HD, Sadri-Vakili G, Pierce RC. Epigenetic inheritance of a cocaine-resistance phenotype. Nat Neurosci 2013; 16: 42–47. 39 Itzhak Y, Liddie S, Anderson KL. Sodium butyrate-induced histone acetylation strengthens the expression of cocaine-associated contextual memory. Neurobiol Learn Mem 2013; 102: 34–42. 40 Kelley JB, Balda MA, Anderson KL, Itzhak Y. Impairments in fear conditioning in mice lacking the nNOS gene. Learn Mem 2009; 16: 371–378. 41 Bourin M, Hascoët M. The mouse light/dark box test. Eur J Pharmacol 2003; 463: 55–65. 42 Crawley J, Goodwin FK. Preliminary report of a simple animal behavior model for the anxiolytic effects of benzodiazepines. Pharmacol Biochem Behav 1980; 13: 167–170. 43 Onaivi ES, Martin BR. Neuropharmacological and physiological validation of a computer-controlled two-compartment black and white box for the assessment of anxiety. Prog Neuropsychopharmacol Biol Psychiatry 1989; 13: 963–976. 44 Tusnády GE, Simon I, Váradi A, Arányi T. BiSearch: primer-design and search tool for PCR on bisulfite-treated genomes. Nucleic Acids Res 2005; 33: e9. 45 Frantz KJ, O'Dell LE, Parsons LH. Behavioral and Neurochemical Responses to Cocaine in Periadolescent and Adult Rats. Neuropsychopharmacology 2007; 32: 625–637. 46 Cho AK, Melega WP, Kuczenski R, Segal DS. Relevance of pharmacokinetic parameters in animal models of methamphetamine abuse. Synapse 2011; 39: 161–166. 47 Chen EY, Tan CM, Kou Y, Duan Q, Wang Z, Meirelles GV et al. Enrichr: interactive and collaborative HTML5 gene list enrichment analysis tool. BMC Bioinformatics 2013; 14: 128–134. 48 Voigt P, Tee WW, Reinberg D. A double take on bivalent promoters. Genes Dev 2013; 27: 1318–1338. 49 United National Office on Drugs and Crime. World Drug Report, Analysis Vol. 1. United Nations Publication: Vienna, Austria, 2004. 50 Substance Abuse and Mental Health Services Administration. Treatment Episode Data Set (TEDS). 1999–2009. National Admissions to Substance Abuse Treatment Services, DASIS Series, S-56, HHS Publication No. (SMA) 11-4646, 2011. 51 LaGasse LL, Derauf C, Smith LM, Newman E, Shah R, Neal C et al. Prenatal methamphetamine exposure and childhood behavior problems at 3 and 5 years of age. Pediatrics 2012; 229: 681–688. 52 LaGasse LL, Wouldes T, Newman E, Smith LM, Shah RZ, Derauf C et al. Prenatal methamphetamine exposure and neonatal neurobehavioral outcome in the USA and New Zealand. Neurotoxicol Teratol 2011; 33: 166–175.

© 2014 Macmillan Publishers Limited

Consequences of parental methamphetamine exposure Y Itzhak et al

11 53 Johnson BA, Roache JD, Ait-Daoud N, Wells LT, Wallace CL, Dawes MA et al. Effects of acute topiramate dosing on methamphetamine-induced subjective mood. Int J Neuropsychopharmacol 2007; 10: 85–98. 54 Ghahremani DG, Tabibnia G, Monterosso J, Hellemann G, Poldrack RA, London ED. Effect of modafinil on learning and task-related brain activity in methamphetamine-dependent and healthy individuals. Neuropsychopharmacology 2011; 36: 950–959. 55 Alicata D, Chang L, Cloak C, Abe K, Ernst T. Higher diffusion in striatum and lower fractional anisotrophy in white matter of methamphetamine users. Psychiatry Res 2009; 174: 1–8. 56 Achat-Mendes C, Ali SF, Itzhak Y. Differential effects of amphetamines-induced neurotoxicity on appetitive and aversive Pavlovian conditioning in mice. Neuropsychopharmacology 2005; 30: 1128–1137. 57 Stephans S, Yamamoto B. Methamphetamines pretreatment and the vulnerability of the striatum to methamphetamine neurotoxicity. Neuroscience 1996; 72: 593–600. 58 Van der Veen R, Abrous DN, De Kloet ER, Piazza PV, Koehl M. Impact of intra- and interstrain cross-fostering on mouse maternal care. Genes Brain Behav 2008; 7: 184–192. 59 Curley JP, Rock V, Moynihan AM, Bateson P, Keverne EB, Champagne FA. Developmental shifts in the behavioral phenotypes of inbred mice: the role of postnatal and juvenile social experiences. Behav Genet 2010; 40: 220–232. 60 LaPlant Q, Vialou V, Covington HE3rd, Dumitriu D, Feng J, Warren BL et al. Dnmt3a regulates emotional behavior and spine plasticity in the nucleus accumbens. Nat Neurosci 2010; 13: 1137–1143. 61 Torregrossa MM, Taylor JR. Learning to forget: manipulating extinction and reconsolidation processes to treat addiction. Psychopharmacology 2013; 226: 659–672. 62 Fanselow MS, Dong HW. Are the dorsal and ventral hippocampus functionally distinct structures? Neuron 2010; 65: 7–19. 63 Swatt JD. Hippocampal function in cognition. Psychopharmacology 2004; 174: 99–110. 64 Meyers RA, Zavala AR, Neisewander JL. Dorsal, but not ventral, hippocampal lesions disrupt cocaine place conditioning. Neuroreport 2003; 14: 2127–2131. 65 Meyers RA, Zavala AR, Speer CM, Neisewander JL. Dorsal hippocampus inhibition disrupts acquisition and expression, but not consolidation, of cocaine conditioned place preference. Behav Neurosci 2006; 120: 401–412. 66 Shohamy D, Adcock RA. Dopamine and adaptive memory. Trends Cogn Sci 2010; 14: 464–472. 67 Britt JP, Benaliouad F, McDevitt RA, Stuber GD, Wise RA, Bonci A. Synaptic and behavioral profile of multiple glutamatergic inputs to the nucleus accumbens. Neuron 2012; 76: 790–803. 68 Dong HW, Swanson LW, Chen L, Fanselow MS, Toga AW. Genomic-anatomic evidence for distinct functional domains in hippocampal field CA1. Proc Natl Acad Sci USA 2009; 106: 11794–1179. 69 Yin H, Bardgett ME, Csernansky JG. Kainic acid lesions disrupt fear-mediated memory processing. Neurobiol Learn Mem 2002; 77: 389–401. 70 Bardgett ME, Boeckman R, Krochmal D, Fernando H, Ahrens R, Csernansky JG. NMDA receptor blockade and hippocampal neuronal loss impair fear conditioning and position habit reversal in C57Bl/6 mice. Brain Res Bull 2003; 60: 131–142. 71 Martin MV, Dong H, Bertchume A, Csernansky JG. Low dose quetiapine reverses deficits in contextual and cued fear conditioning in rats with excitotoxin-induced hippocampal neuropathy. Pharmacol Biochem Behav 2005; 82: 263–269. 72 Hunsaker MR, Kesner RP. Dissociations across the dorsal-ventral axis of CA3 and CA1 for encoding and retrieval of contextual and auditory-cued fear. Neurobiol Learn Mem 2008; 89: 61–69. 73 Day JJ, Sweatt JD. DNA methylation and memory formation. Nat Neurosci 2010; 13: 1319–1323. 74 Suderman M, McGowan PO, Sasaki A, Huang TC, Hallett MT, Meaney MJ et al. Conserved epigenetic sensitivity to early life experience in the rat and human hippocampus. Proc Natl Acad Sci USA 2012; 109: 17266–17272.

75 Novikova SI, He F, Bai J, Cutrufello NJ, Lidow MS, Undieh AS. Maternal cocaine administration in mice alters DNA methylation and gene expression in hippocampal neurons of neonatal and prepubertal offspring. PLoS ONE 2008; 3: e1919. 76 Justinova Z, Ferre S, Segal PN, Antoniou K, Solinas M, Pappas LA et al. Involvement of adenosine A1 and A2A receptors in the adenosinergic modulation of the discriminative-stimulus effects of cocaine and methamphetamine in rats. J Pharmacol Exp Ther 2003; 307: 977–986. 77 Chen Q, Zhu X, Zhang Y, Wetsel WC, Lee TH, Zhang X. Integrin-linked kinase is involved in cocaine sensitization by regulating PSD-95 and synapsin I expression and GluR1 Ser845 phosphorylation. J Mol Neurosci 2010; 40: 284–294. 78 Reissner KJ, Uys JD, Schwacke JH, Comte-Walters S, Rutherford-Bethard JL, Dunn TE et al. AKAP signaling in reinstated cocaine seeking revealed by iTRAQ proteomic analysis. J Neurosci 2011; 31: 5648–5658. 79 Krasnova IN, Chiflikyan M, Justinova Z, McCoy MT, Ladenheim B, Jayanthi S et al. CREB phosphorylation regulates striatal transcriptional responses in the self-administration model of methamphetamine addiction in the rat. Neurobiol Dis 2013; 58: 132–1343. 80 Renthal W, Maze I, Krishnan V, Covington HE3rd, Xiao G, Kumar A et al. Histone deacetylase 5 epigenetically controls behavioral adaptations to chronic emotional stimuli. Neuron 2007; 56: 517–529. 81 Niikura K, Zhou Y, Ho A, Kreek MJ. Proopiomelanocortin (POMC) expression and conditioned place aversion during protracted withdrawal from chronic intermittent escalating-dose heroin in POMC-EGFP promoter transgenic mice. Neuroscience 2013; 236: 220–232. 82 Robison AJ, Vialou V, Mazei-Robison M, Feng J, Kourrich S, Collins M et al. Behavioral and structural responses to chronic cocaine require a feedforward loop involving ΔFosB and calcium/calmodulin-dependent protein kinase II in the nucleus accumbens shell. J Neurosci 2013; 33: 4295–4307. 83 Miller JS, Tallarida RJ, Unterwald EM. Cocaine-induced hyperactivity and sensitization are dependent on GSK3. Neuropharmacology 2009; 56: 1116–1123. 84 Xu CM, Wang J, Wu P, Zhu WL, Li QQ, Xue YX et al. Glycogen synthase kinase 3beta in the nucleus accumbens core mediates cocaine-induced behavioral sensitization. J Neurochem 2009; 111: 1357–1368. 85 Mash DC, ffrench-Mullen J, Adi N, Qin Y, Buck A, Pablo J. Gene expression in human hippocampus from cocaine abusers identifies genes which regulate extracellular matrix remodeling. PLoS ONE 2007; 2: e1187. 86 Shonesy BC, Thiruchelvam K, Parameshwaran K, Rahman EA, Karuppagounder SS, Huggins KW et al. Central insulin resistance and synaptic dysfunction in intracerebroventricular-streptozotocin injected rodents. Neurobiol Aging 2012; 33: e5–18. 87 Moita MA, Lamprecht R, Nader K, LeDoux JE. A-kinase anchoring proteins in amygdala are involved in auditory fear memory. Nat Neurosci 2002; 5: 837–838. 88 Govorko D, Bekdash RA, Zhang C, Sarkar DK. Male germline transmits fetal alcohol adverse effect on hypothalamic proopiomelanocortin gene across generations. Biol Psychiatry 2012; 72: 378–388. 89 West NA, Kechris K, Dabelea D. Exposure to maternal diabetes in utero and DNA methylation patterns in the offspring. Immunometabolism 2013; 1: 1–9. 90 Dowling AL, Iannacone EA, Zoeller RT. Maternal hypothyroidism selectively affects the expression of neuroendocrine-specific protein A messenger ribonucleic acid in the proliferative zone of the fetal rat brain cortex. Endocrinology 2001; 142: 390–399. 91 Bohuslavova R, Skvorova L, Sedmera D, Semenza GL, Pavlinkova G. Increased susceptibility of HIF-1α heterozygous-null mice to cardiovascular malformations associated with maternal diabetes. J Mol Cell Cardiol 2013; 60: 129–141. 92 Grisel JE, Bartels JL, Allen SA, Turgeon VL. Influence of beta-Endorphin on anxious behavior in mice: interaction with EtOH. Psychopharmacology 2008; 200: 105–115. 93 Nguyen AT, Marquez P, Hamid A, Kieffer B, Friedman TC, Lutfy K. The rewarding action of acute cocaine is reduced in β-endorphin deficient but not in μ opioid receptor knockout mice. Eur J Pharmacol 2012; 686: 50–54.

Supplementary Information accompanies the paper on the Molecular Psychiatry website (http://www.nature.com/mp)

© 2014 Macmillan Publishers Limited

Molecular Psychiatry (2014), 1 – 11