ARTICLES nature neuroscience Epigenetic programming by maternal behavior 8 Ian CG Weaver, 2, Nadia Cervoni, Frances A Champagne, 2, Ana C D'Alessio Shakti Sharma! Jonathan R Seckl#, Sergiy Dymov Moshe Szyf2,3& Michael J Meaney,2 Here we report that increased pup licking and grooming(LG)and arched-back nursing(ABN) by rat mothers altered the offspring LG and abn were found to have differences in DNA methylation, as compared to offspring of low-LG-ABN'mothers. These g differences emerged over the first week of life, were reversed with cross-fostering, persisted into adulthood and were associated g with altered histone acetylation and transcription factor (NGFI-A)binding to the GR promoter. Central infusion of a histone a deacetylase inhibitor removed the group differences in histone acetylation, DNA methylation, NGFI-A binding GR expression and hypothalamic-pituitary-adrenal(HPA)responses to stress, suggesting a causal relation among epigenomic state, GR expression nd the maternal effect on stress responses in the offspring Thus we show that an epigenomic state of a gene can be established through behavioral programming, and it is potentially reversible 8 Through undefined epigenetic processes, maternal effects influence inhibits CRF synthesis and release, thus dampening HPA responses to s the development of defensive responses to threat in organisms rang- stress". The adult offspring of high- compared with low-LG-ABN o ing from plants to mammals,. In the rat, such effects are mediated by mothers show increased hippocampal GR expression and enhanced mission of individual differences in stress responses from mother to high-LG-ABN mothers show decreased hypothalamic CRF expre s offspring -. Mother-pup contact in the rat primarily occurs within sion and more modest HPA responses to stress. Eliminating the dif- e litter, licks and grooms her pups, and nurses while occasionally experience on HPA responses to stress in adulthood 2, suggesting that licking and grooming the pups. There are stable individual differ- the difference in hippocampal GR expression serves as a mechanism ences in two forms of maternal behavior-LG and ABN--over the for the effect of early experience on the development of individual first week of lactation6-10. Such naturally occurring variations in differences in HPA responses to stress o maternal behavior are associated with the development of individual In vivo and in vitro studies suggest that maternal LG and ABN differences in behavioral and HPA responses to stress in the offspring. increase GR gene expression in the offspring through increased sero As adults, the offspring of high-LG-ABN' mothers are less fearful and tonin(5-HT)activity at 5-HT7 receptors, and the subsequent activa show more modest HPA responses to stress than the offspring of low- tion of cAMP and CAMP-dependent protein kinase activity3-15Both LG-ABN mothers"-. Cross-fostering studies show that the biological the in vitro effect of 5-HT and the in vivo effect of maternal behavior offspring of low-LG-ABN mothers reared by high-LG-abn dams on gR gene expression are accompanied by an increased hippocam resemble the normal offspring of high-LG-ABN mothers(and vice pal expression of nerve growth factor-inducible protein A(NGFI-A,a versa"). These findings suggest that variations in maternal behavior transcription factor also known as egr-1, krox-24, zenk and zif-268) serve as a mechanism for the nongenomic transmission of individual The non-coding exon I region of the hippocampal GR includes a pre differences in stress reactivity across generations, 5. 9. The critical moter region, exon 17, containing a binding site for NGFI-A6 question concerns the mechanisms whereby these maternal effects, or(Fig. la). Splice variants of the GR mRNA containing the exon 17 other forms of environmental ' programming, are sustained over the sequence are found predominantly in the brain, and the expression of lifespan of the animal GR mRNAs containing the exon 1, sequence is increased in the off- Maternal behavior in the rat permanently alters the development of spring of high-LG-ABN mothers or following manipulations that HPA responses to stress through tissue-specific effects on gene increase maternal licking and grooming(Weaver, L.C. G. et al., Soc. expression. The magnitude of the HPA response to acute stress is a Neurosci. Abstr. 697. 15, 2001), suggesting that the use of this promoter function of hypothalamic corticotropin-releasing factor (CRF) is enhanced as a function of maternal care. Although these findings release, which activates the pituitary-adrenal system. There are also might explain the increased GR expression in the neonate, we are left adulatory influences, such as glucocorticoid negative feedback that with the question of how the effect of maternal care might persist into Hospital Research Center, 6875 LaSalle Blvd., Montreal, Quebec H4H 1R3, Canada. 2McGill Program for the Study of Behaviour, Genes and Environment ology and Therapeutics, McGill University, 3655 Sir William Osler Promenade, Montreal, Quebec H3G 1Y6, Canada. Molecular Centre, Edinburgh University, Western General Hospital, Edinburgh EH4 2XU, UK. Correspondence should be addressed to M.J. M meaney @mcgill. ca)or M.S. ( moshe szyf@mcgill. ca). d online 27 June 2004; corrected 27 July 2004(details online); doi: 10.1038/nn1276 NATURE NEUROSCIENCE VOLUME 7 NUMBER 8 AUGUST 2004 847
ARTICLES Through undefined epigenetic processes, maternal effects influence the development of defensive responses to threat in organisms ranging from plants to mammals1,2. In the rat, such effects are mediated by variations in maternal behavior, which serve as the basis for the transmission of individual differences in stress responses from mother to offspring3–5. Mother-pup contact in the rat primarily occurs within the context of a nest-bout, which begins when the mother approaches the litter, licks and grooms her pups, and nurses while occasionally licking and grooming the pups6. There are stable individual differences in two forms of maternal behavior—LG and ABN—over the first week of lactation6–10. Such naturally occurring variations in maternal behavior are associated with the development of individual differences in behavioral and HPA responses to stress in the offspring. As adults, the offspring of ‘high-LG-ABN’ mothers are less fearful and show more modest HPA responses to stress than the offspring of ‘lowLG-ABN’ mothers6–9. Cross-fostering studies show that the biological offspring of low-LG-ABN mothers reared by high-LG-ABN dams resemble the normal offspring of high-LG-ABN mothers (and vice versa9). These findings suggest that variations in maternal behavior serve as a mechanism for the nongenomic transmission of individual differences in stress reactivity across generations4,5,9. The critical question concerns the mechanisms whereby these maternal effects, or other forms of environmental ‘programming’, are sustained over the lifespan of the animal. Maternal behavior in the rat permanently alters the development of HPA responses to stress through tissue-specific effects on gene expression. The magnitude of the HPA response to acute stress is a function of hypothalamic corticotropin-releasing factor (CRF) release, which activates the pituitary-adrenal system. There are also modulatory influences, such as glucocorticoid negative feedback that inhibits CRF synthesis and release, thus dampening HPA responses to stress11. The adult offspring of high- compared with low-LG-ABN mothers show increased hippocampal GR expression and enhanced glucocorticoid feedback sensitivity7,9. Predictably, adult offspring of high-LG-ABN mothers show decreased hypothalamic CRF expression and more modest HPA responses to stress7. Eliminating the difference in hippocampal GR levels abolishes the effects of early experience on HPA responses to stress in adulthood12, suggesting that the difference in hippocampal GR expression serves as a mechanism for the effect of early experience on the development of individual differences in HPA responses to stress5. In vivo and in vitro studies suggest that maternal LG and ABN increase GR gene expression in the offspring through increased serotonin (5-HT) activity at 5-HT7 receptors, and the subsequent activation of cAMP and cAMP-dependent protein kinase activity13–15. Both the in vitro effect of 5-HT and the in vivo effect of maternal behavior on GR gene expression are accompanied by an increased hippocampal expression of nerve growth factor-inducible protein A (NGFI-A, a transcription factor also known as egr-1, krox-24, zenk and zif-268). The non-coding exon 1 region of the hippocampal GR includes a promoter region, exon 17, containing a binding site for NGFI-A16 (Fig. 1a). Splice variants of the GR mRNA containing the exon 17 sequence are found predominantly in the brain, and the expression of GR mRNAs containing the exon 17 sequence is increased in the offspring of high-LG-ABN mothers or following manipulations that increase maternal licking and grooming16 (Weaver, I.C.G. et al., Soc. Neurosci. Abstr. 697.15, 2001), suggesting that the use of this promoter is enhanced as a function of maternal care. Although these findings might explain the increased GR expression in the neonate, we are left with the question of how the effect of maternal care might persist into 1Douglas Hospital Research Center, 6875 LaSalle Blvd., Montréal, Québec H4H 1R3, Canada. 2McGill Program for the Study of Behaviour, Genes and Environment and 3Department of Pharmacology and Therapeutics, McGill University, 3655 Sir William Osler Promenade, Montréal, Québec H3G 1Y6, Canada. 4Molecular Medicine Centre, Edinburgh University, Western General Hospital, Edinburgh EH4 2XU, UK. Correspondence should be addressed to M.J.M. (michael.meaney@mcgill.ca) or M.S. (moshe.szyf@mcgill.ca). Published online 27 June 2004; corrected 27 July 2004 (details online); doi:10.1038/nn1276 Epigenetic programming by maternal behavior Ian C G Weaver1,2, Nadia Cervoni3, Frances A Champagne1,2, Ana C D’Alessio3, Shakti Sharma1, Jonathan R Seckl4, Sergiy Dymov3, Moshe Szyf2,3 & Michael J Meaney1,2 Here we report that increased pup licking and grooming (LG) and arched-back nursing (ABN) by rat mothers altered the offspring epigenome at a glucocorticoid receptor (GR) gene promoter in the hippocampus. Offspring of mothers that showed high levels of LG and ABN were found to have differences in DNA methylation, as compared to offspring of ‘low-LG-ABN’ mothers. These differences emerged over the first week of life, were reversed with cross-fostering, persisted into adulthood and were associated with altered histone acetylation and transcription factor (NGFI-A) binding to the GR promoter. Central infusion of a histone deacetylase inhibitor removed the group differences in histone acetylation, DNA methylation, NGFI-A binding, GR expression and hypothalamic-pituitary-adrenal (HPA) responses to stress, suggesting a causal relation among epigenomic state, GR expression and the maternal effect on stress responses in the offspring. Thus we show that an epigenomic state of a gene can be established through behavioral programming, and it is potentially reversible. NATURE NEUROSCIENCE VOLUME 7 | NUMBER 8 | AUGUST 2004 847 © 2004 Nature Publishing Group http://www.nature.com/natureneuroscience
ARTICLES methylation of GR promoter. (a)Sequence map of the exon 17 GR promoter including the 17 gion b(encircled).(b, c)Methylation analysis the 17 CpG dinucleotides of the exon 17 GR promoter region from adult high-and low-LG ABN offspring (6-10 clones sequenced/animal n=4 animals/group; P<0.01).(b)Percentage of cytosine residues that were methylated (mean +s.e.m.)for the first 15 CpG dinucleotides 口 High LG-ABN (P<0.05).(c)Percentage of methylated ytosines(mean ts.e. m )for the 5'(site 16)and 3(site 17)CpG dinucleotides within the NGFl-A 1001■ LOW-LG/ABN of cross fostering the offspring of high- and low- LG-ABN mothers on cytosine methylation of the 5 and 3'CpG dinucleotides within the NGFI-A binding sequence of the exon 17 GR promot L-L: animals born to and reared by low-LG-ABN mothers H-H: animals born to and reared by high-LG-ABN mothers: H-L: animals born to 5678g10112131415 high-LG-ABN mothers and reared by low-LG-ABN mothers: L-H: animals born to low-LG-ABI mothers and reared by high-LG-ABN mothers. (e)Percentage of cytosine methylation(mean+ m)of the 5 and 3 CpG dinucleotides within 5 CpG dinucleotide 3 CpG dinucleotide 5 CpG dinucleotide 3 CpG dinucleotide the NGFl-A binding region of the exon 17 GR romoter gene in the offspring of high-or low- LGABN mothers(n= 5 animals/group; P<0.001)as a function of age. There were no differences at any postnatal age in level of ytosine methylation of the 3 CpG (site 17) L-L H-H H-L LH LL H-H H- L-H E20 P1 PB P21 P90 oLd P1 adulthood. Gene expression is controlled by the epigenome, which is sites of the exon 1, GR promoter sequence(Fig. 1b,c). A two-way comprised of chromatin structure/ and DNA methylation. We ANOVA revealed a highly significant effect of Group(F= 55.9 tested the hypothesis that maternal care alters DNA methylation of P<00001)and Region(F=27.7, P<00001), as well as a significant o the Gr exon 17 promoter, and that these changes are stably main- Group x Region interaction effect (F= 27.7,P<0.0001) tained into adulthood and associated with differences in GR expres- Importantly, the cytosine residue within the 5" CpG dinucleotide sion and HPa responses to stress (site 16) of the NGFl-A consensus sequence(Fig. Ic)is always methylated in the offspring of low-LG-ABN mothers, and rarely RESULTS methylated in the offspring of high-LG-ABN dams In contrast, the Maternal care and methylation of exon 17 promoter 3 CpG dinucleotide(site 17)remains methylated, regardless of dif- DNA methylation is a stable, epigenomic mark at CpG dinucleotides ferences in maternal care. Dissected hippocampi inevitably contain often associated with stable variations in gene transcription 8-20. Two glial cells as well as neurons. Considering the pronounced effect of kinds of changes in DNA methylation are known to affect gene maternal care on the methylation status of the 5" Cpg dinucleotide expression: regional, non-site specific DNA methylation around a of the NGFI-A response element(>90%), the effect of maternal care promoterand site-specific methylation. Hypomethylation of CpG must include neuronal as well as glial cells; both populations express dinucleotides of regulatory regions of genes is associated with active GR23,24 and NGFl-A25 genes chromatin ture and transcriptional activity.. Thus, the methylation pattern is a stable signature of the epigenomic status of a Cross-fostering reveals epigenetic marking by maternal behavior regulatory sequence. We focused on the methylation state of the exon Our findings suggest that specific sites within the exon 1, GR pro- 17 GR promoter, which is activated in the hippocampus in offspring moter are differentially methylated as a function of maternal behav- of high-LG-ABN mothers ior, but these findings are merely correlational. To directly examine To determine whether DNA methylation of specific target sites on the relation between maternal behavior and DNA methylation the gr promoter change in response to maternal care, we mapped within the exon 1, promoter, we performed an adoption study in differences in the methylation status of individual cytosines within which the biological offspring of high- or low-LG-ABN mothers the CpG dinucleotides of the exon 1, promoter from hippocampal were cross-fostered to either high-or low-LG-aBN dams within 12 h tissue from the adult offspring of high-and low-LG-ABN mothers. of birth. Cross-fostering produced a pattern of exon 1, promoter sodium bisulfite mapping. 22, with a particular interest in methylation that was associated with the rearing mother(F= 4.8, around the NGFl-A consensus sequence(Fig. la). The P<0.05; Fig. ld)and thus reversed the difference in methylation at wed significant differences in the methylation of specific specific cytosines, notably at the 5" CpG dinucleotide(site 16)of the 848 VOLUME 7 NUMBER 8 AUGUST 2004 NATURE NEUROSCIENCE
sites of the exon 17 GR promoter sequence (Fig. 1b,c). A two-way ANOVA revealed a highly significant effect of Group (F = 55.9, P < 0.0001) and Region (F = 27.7, P < 0.0001), as well as a significant Group × Region interaction effect (F = 27.7, P < 0.0001). Importantly, the cytosine residue within the 5′ CpG dinucleotide (site 16) of the NGFI-A consensus sequence (Fig. 1c) is always methylated in the offspring of low-LG-ABN mothers, and rarely methylated in the offspring of high-LG-ABN dams. In contrast, the 3′ CpG dinucleotide (site 17) remains methylated, regardless of differences in maternal care. Dissected hippocampi inevitably contain glial cells as well as neurons. Considering the pronounced effect of maternal care on the methylation status of the 5′ CpG dinucleotide of the NGFI-A response element (>90%), the effect of maternal care must include neuronal as well as glial cells; both populations express GR23,24 and NGFI-A25 genes. Cross-fostering reveals epigenetic marking by maternal behavior Our findings suggest that specific sites within the exon 17 GR promoter are differentially methylated as a function of maternal behavior, but these findings are merely correlational. To directly examine the relation between maternal behavior and DNA methylation within the exon 17 promoter, we performed an adoption study in which the biological offspring of high- or low-LG-ABN mothers were cross-fostered to either high- or low-LG-ABN dams within 12 h of birth9. Cross-fostering produced a pattern of exon 17 promoter methylation that was associated with the rearing mother (F = 4.8, P < 0.05; Fig. 1d) and thus reversed the difference in methylation at specific cytosines, notably at the 5′ CpG dinucleotide (site 16) of the Figure 1 Maternal care alters cytosine methylation of GR promoter. (a) Sequence map of the exon 17 GR promoter including the 17 CpG dinucleotides (bold) and the NGFI-A binding region16 (encircled). (b,c) Methylation analysis of the 17 CpG dinucleotides of the exon 17 GR promoter region from adult high- and low-LGABN offspring (6–10 clones sequenced/animal; n = 4 animals/group; *P < 0.01). (b) Percentage of cytosine residues that were methylated (mean ± s.e.m.) for the first 15 CpG dinucleotides (*P < 0.05). (c) Percentage of methylated cytosines (mean ± s.e.m.) for the 5′ (site 16) and 3′ (site 17) CpG dinucleotides within the NGFI-A binding sequence (*P < 0.0001). (d) The effect of cross-fostering the offspring of high- and lowLG-ABN mothers on cytosine methylation of the 5′ and 3′ CpG dinucleotides within the NGFI-A binding sequence of the exon 17 GR promoter gene in adult hippocampi (n = 5 animals/group). L-L: animals born to and reared by low-LG-ABN mothers; H-H: animals born to and reared by high-LG-ABN mothers; H-L: animals born to high-LG-ABN mothers and reared by low-LG-ABN mothers; L-H: animals born to low-LG-ABN mothers and reared by high-LG-ABN mothers. (e) Percentage of cytosine methylation (mean ± s.e.m.) of the 5′ and 3′ CpG dinucleotides within the NGFI-A binding region of the exon 17 GR promoter gene in the offspring of high- or lowLG-ABN mothers (n = 5 animals/group; P < 0.001) as a function of age. There were no differences at any postnatal age in level of cytosine methylation of the 3′ CpG (site 17). ARTICLES adulthood. Gene expression is controlled by the epigenome, which is comprised of chromatin structure17 and DNA methylation18. We tested the hypothesis that maternal care alters DNA methylation of the GR exon 17 promoter, and that these changes are stably maintained into adulthood and associated with differences in GR expression and HPA responses to stress. RESULTS Maternal care and methylation of exon 17 promoter DNA methylation is a stable, epigenomic mark at CpG dinucleotides often associated with stable variations in gene transcription18–20. Two kinds of changes in DNA methylation are known to affect gene expression: regional, non-site specific DNA methylation around a promoter19 and site-specific methylation. Hypomethylation of CpG dinucleotides of regulatory regions of genes is associated with active chromatin structure and transcriptional activity18,20. Thus, the methylation pattern is a stable signature of the epigenomic status of a regulatory sequence. We focused on the methylation state of the exon 17 GR promoter, which is activated in the hippocampus in offspring of high-LG-ABN mothers. To determine whether DNA methylation of specific target sites on the GR promoter change in response to maternal care, we mapped differences in the methylation status of individual cytosines within the CpG dinucleotides of the exon 17 promoter from hippocampal tissue from the adult offspring of high- and low-LG-ABN mothers. We used sodium bisulfite mapping21,22, with a particular interest in the region around the NGFI-A consensus sequence (Fig. 1a). The results showed significant differences in the methylation of specific 848 VOLUME 7 | NUMBER 8 | AUGUST 2004 NATURE NEUROSCIENCE L-L H-H H-L L-H 0 20 40 60 80 100 C-methylation (%) L-L H-H H-L L-H 0 20 40 60 80 100 5´ CpG dinucleotide 3´ CpG dinucleotide High-LG/ABN Low-LG/ABN Age (d) E20 P1 P6 P21 P90 0 20 40 60 80 100 E20 P1 P6 P21 P90 0 20 40 60 80 100 * * * C-methylation (%) 5´ CpG dinucleotide 3´ CpG dinucleotide 1681 ccc 1741 ctctgctagtgtgacacactt1cg2cgcaactc3cgcagttgg4cggg5cg6cggaccacccctg7c 1801 ggctctgc8cggctggctgtcaccct9cgggggctctggctgc10cgaccca11cgggg12cgggct 1861 c13cgag14cggtt ccaagcct15cggagtggg16cggggg17cgggagggagcctgggagaa 11 14 15 16 17 18 19 110 111 2 5´ 3´ 1234 5 6 7 8 9 10 11 12 13 14 15 Region C-methylation (%) Low-LG/ABN High-LG-ABN 0 20 40 60 80 100 * * * * * * * 5´ 3´ CpG dinucleotide Low LG-ABN High LG-ABN * 0 20 40 60 80 100 a b c d e C-methylation (%) © 2004 Nature Publishing Group http://www.nature.com/natureneuroscience
ARTICLES LOW-LG/ABN h-LG/ABN Figure 2 Chromatin immunoprecipitation analysis of the association between histone H3-K9 acetylation and NGFl-A binding to the exon 17 GR sequence in hippocampal tissue from adult offspring of high- and low-LG- ABN mothers(n= 4 animals/group).(a, b)Lanes were loaded with non- (middle) primary antibody immunoprecipitated (A), or non-immune IgG antibody immuno-precipitated (N) hippocampal extracts).(a)Representative Southern blot of the amplified exon l7 region from acetyl-histone H3-K9 immunoprecipitated hippocampal tissue(194 bp band )and B-actin(171 bp b band)control. (b)Representative Southern blot of the amplified exon 17 GR exon 1, ion of the GR from NGFI-A immunoprecipitated hippocampal tissue (194 bp band). DNA loading was controlled using primers specific for the ubiquitously expressed B-actin promoter-o region. Exon 1b estrogen receptor-a promoter region, which does not contain NGFl-A recogniti elements(493 bp), amplified from the same NGFI-A immunoprecipitated hippocampal tissue was run as a control for specificity and showed no signal. (c)Relative optical density(ROD; m )of exon 17 2061日 sequence amplified from acetyl-histone H3-K9 or NGFI-A immunoprecipitated hippocampal tissue of adult high-and low-LG-ABN offspring(n= 4 animals/group: *P<0.001; **P<00001) 203 to examine the methylation status of the cytosines within the exon 1 GR promoter during development( Fig. le). Statistical analysis of the Acetyl H3K9 data for the 5 CpG(site 16)revealed a highly significant effect of Group(F= 66.7, P<00001)and Age(F=21.1, P<00001)as well as a significant interaction effect(F= 13.7, P<00001). Tukey post-hoc analysis revealed that the Group effect on methylation status of the 5 CpG(site 16)was significant at P6, P21 and P90(P < 0.001),but 9 NGFI-A consensus sequence(Fig. Id, left panel). Thus, in the not at E20 or PI. Just before birth(embryonic day 20; E20)the entire low-LG-ABN offspring that were fostered to high-LG-aBN dams, region was unmethylated in both groups. Strikingly, one day after methylation of this 5 site within the exon 1, promoter was indistin- birth(postnatal day 1; Pl)the exon 1, gr promoter was de novo a guishable from that of the biological offspring of high-LG-ABN methylated in both groups. The 5 and 3 CpG sites of the exon 1,GR mothers. Likewise, the methylation of the same 5" CpG dinucleotide NGFI-A response element in the offspring of both high- and low- in the biological offspring of high-LG-ABN mothers reared by low- LG-ABN mothers, which exhibit differential methylation later in life, There was no effect or mparable to that of low-LG-ABN offspring. were de novo methylated to the same extent. These data show that Lg-abn dams was co cross-fostering at the cytosine within the 3 CpG dinucleotide(site 17; Fig. ld) 0. These findings suggest that variations in maternal care directly alter TSA TSA 100 ng/ml Vehicle 100 ng/ml the methylation status of the exon 1, promoter of the GR gene. Thus Maternal care we have demonstrated that a DNA methylation pattern can be estab lished through a behavioral mode of programming without germ line ANIANIANI A N transmission In parental imprinting, a well-established paradigm of inheritance of an epigenomic mark, the paternally and maternally (acetyl-H3K9 IP) inherited alleles are differentially methylated. These methylation pat terns are defined during maturation of spermatocytes and oocytes, (acetyl-H3K9 IP) nd are transmitted to the offspring through the germ line- Timing of the maternal effect on DNA methylation -[--- The maternal care of high- and low-LG-ABN mothers differs only during the first week of life,. Thus, we wondered whether this period ER-a exon 1b to the Ing of the difference in NGFI-A DNA methylation in the offspring. We used sodium bisulfite mapping Figure 3 HDAC inhibitior (TSA) eliminates maternal effect on histone acetylation and NGFl-A binding (a)Chromatin immunoprecipitation nalysis of the association between histone H3-K9 acetylation and NGFI-A binding to the exon 17 GR promoter sequence in hippocampal tissue from vehicle- and TSA-treated (100 ng/ml) adult offspring of high- and low-LG- ABN mothers(n= 4 animals/group; lane labels as described in Fig. 2). b)Relative optical density(ROD; mean ts.e. m )of exon 17 sequence amplified from acetyl-histone H3-K9 or NGFl-A immunoprecipitate hippocampal tissue(P<0.05; **P<0.01) NATURE NEUROSCIENCE VOLUME 7 NUMBER 8 AUGUST 2004
ARTICLES NGFI-A consensus sequence (Fig. 1d, left panel). Thus, in the low-LG-ABN offspring that were fostered to high-LG-ABN dams, methylation of this 5′ site within the exon 17 promoter was indistinguishable from that of the biological offspring of high-LG-ABN mothers. Likewise, the methylation of the same 5′ CpG dinucleotide in the biological offspring of high-LG-ABN mothers reared by lowLG-ABN dams was comparable to that of low-LG-ABN offspring. There was no effect of cross-fostering at the cytosine within the 3′ CpG dinucleotide (site 17; Fig. 1d). These findings suggest that variations in maternal care directly alter the methylation status of the exon 17 promoter of the GR gene. Thus we have demonstrated that a DNA methylation pattern can be established through a behavioral mode of programming without germ line transmission. In parental imprinting, a well-established paradigm of inheritance of an epigenomic mark, the paternally and maternally inherited alleles are differentially methylated. These methylation patterns are defined during maturation of spermatocytes and oocytes, and are transmitted to the offspring through the germ line26. Timing of the maternal effect on DNA methylation The maternal care of high- and low-LG-ABN mothers differs only during the first week of life7,8. Thus, we wondered whether this period corresponds to the timing for the appearance of the difference in DNA methylation in the offspring. We used sodium bisulfite mapping to examine the methylation status of the cytosines within the exon 17 GR promoter during development (Fig. 1e). Statistical analysis of the data for the 5′ CpG (site 16) revealed a highly significant effect of Group (F = 66.7, P < 0.0001) and Age (F = 21.1, P < 0.0001) as well as a significant interaction effect (F = 13.7, P < 0.0001). Tukey post-hoc analysis revealed that the Group effect on methylation status of the 5′ CpG (site 16) was significant at P6, P21 and P90 (P < 0.001), but not at E20 or P1. Just before birth (embryonic day 20; E20) the entire region was unmethylated in both groups. Strikingly, one day after birth (postnatal day 1; P1) the exon 17 GR promoter was de novo methylated in both groups. The 5′ and 3′ CpG sites of the exon 17 GR NGFI-A response element in the offspring of both high- and lowLG-ABN mothers, which exhibit differential methylation later in life, were de novo methylated to the same extent. These data show that NATURE NEUROSCIENCE VOLUME 7 | NUMBER 8 | AUGUST 2004 849 Figure 2 Chromatin immunoprecipitation analysis of the association between histone H3-K9 acetylation and NGFI-A binding to the exon 17 GR sequence in hippocampal tissue from adult offspring of high- and low-LGABN mothers (n = 4 animals/group). (a,b) Lanes were loaded with nonimmunoprecipitated input (I), acetylated histone H3-K9 (top) or NGFI-A (middle) primary antibody immunoprecipitated (A), or non-immune IgG antibody immuno-precipitated (N) hippocampal extracts). (a) Representative Southern blot of the amplified exon 17 region from acetyl-histone H3-K9 immunoprecipitated hippocampal tissue (194 bp band) and β-actin (171 bp band) control. (b) Representative Southern blot of the amplified exon 17 region of the GR from NGFI-A immunoprecipitated hippocampal tissue (194 bp band). DNA loading was controlled using primers specific for the ubiquitously expressed β-actin promoter-α region. Exon 1b estrogen receptor-α promoter region, which does not contain NGFI-A recognition elements (493 bp), amplified from the same NGFI-A immunoprecipitated hippocampal tissue was run as a control for specificity and showed no signal. (c) Relative optical density (ROD; mean ± s.e.m.) of exon 17 sequence amplified from acetyl-histone H3-K9 or NGFI-A immunoprecipitated hippocampal tissue of adult high- and low-LG-ABN offspring (n = 4 animals/group; *P < 0.001; **P < 0.0001). Figure 3 HDAC inhibitior (TSA) eliminates maternal effect on histone acetylation and NGFI-A binding. (a) Chromatin immunoprecipitation analysis of the association between histone H3-K9 acetylation and NGFI-A binding to the exon 17 GR promoter sequence in hippocampal tissue from vehicle- and TSA-treated (100 ng/ml) adult offspring of high- and low-LGABN mothers (n = 4 animals/group; lane labels as described in Fig. 2). (b) Relative optical density (ROD; mean ± s.e.m.) of exon 17 sequence amplified from acetyl-histone H3-K9 or NGFI-A immunoprecipitated hippocampal tissue (*P < 0.05; **P < 0.01). © 2004 Nature Publishing Group http://www.nature.com/natureneuroscience
ARTICLES at the lysine-9(K9) residue of H3 and H4 histones is a well-estab ished marker of active chromatin/.29. Acetylation of the histone tails neutralizes the positively charged s,which disrupts his- tone binding to negatively charged DNA and thus promotes tran scription factor binding. We tested the hypothesis that the maternal effect on DNA methylation results in(i)increased histone acetyl- tion at the k9 residue of the H3 histone(s)associated with the exon 1, GR promoter and (ii) increased interaction between NGFI-A and the promoter sequence. We performed a chromatin immuno 123456789101112131415 precipitation(ChIP)analysis of histone H3-K9 acetylation and NGFI-A protein binding to the exon 1, GR promoter in the native chromatin environment in vivo Intact hippocampi from adult off- b avehicle: low LG/ABN spring of high- and low-LG-ABN mothers were crosslinked in vivo by paraformaldehyde perfusion. We then selectively immunopre cipitated protein-DNA complexes with either an acetylated H3-K9 histone primary antibody or an NGFI-A primary antibody. The protein-DNA complexes were uncrosslinked, and the precipitated genomic DNA was subjected to PCR amplification with primers pecific for the exon 1, GR promoter sequence. There were signifi cant Group effects for the association of both histone H3-K9 acety lation(t= 2.1, *P<0.001)and NGFI-A (t=3. 1, **P<00001)with 17 Gr These results indicated signifi- cantly greater histone H3-K9 acetylation association and threefold Figure 4 TSA effects on cytosine methylation. (a, b)Methylation analysis of greater binding of NGFI-A protein to the hippocampal exon 1,GR e the 17 CpG dinucleotides of the exon 1, GR promoter in hippocampi of promoter in the adult offspring of high- compared with low-LG- G ABN mothers (n= 5 animals/group).(a)Percentage of cytosine residues GR promoter involves DNA methylation, histone H3-K9 acetyla that were methylated(mean+ s.e. m )for the first 15 CpG dinucleotides tion and alterations in ngFi-a binding (P<0.05).(b)Percentage of methylated cytosines for the 5(site 16)and 3(site 17)CpG dinucleotides within the NgFl-a binding region (P<0.001;**P<0.003) Reversal of maternally mediated epigenetic marking These findings suggest that maternal care influences hippocampal GR expression, and thus HPA function in the offspring, through epigenetic alterations that regulate NGFl-A binding to the exon 1, promoter. A both the basal state of methylation and the first wave of de novo critical question is whether the impact of early experience is reversible methylation after birth occur similarly in both groups. Whereas it is and whether epigenetic programming is modifiable in adult, post generally accepted that DNA methylation patterns are formed prena- mitotic tissues? The generally accepted model is that the DNA methyla- o tally and that de novo methylation occurs early in development, there tion pattern is an irreversible reaction in adult post-mitotic cells is at least one documented example of postnatal de novo methylation However, recent data from in vitro experiments suggests that in certain of the Hoxa5 and Hoxb5 genes27 Because similar analyses are not doc- instances it is possible to induce replication-independent demethylation umented for other genes, it remains unknown whether changes in of ectopically methylated genes by increasing histone acetylation us ethylation are common around birth or whether they are unique to the histone deacetylase(HDAC) inhibitor trichostatin A(TSA)29,30 Cytosine methylation attracts methylated DNA binding proteins and The differences in the methylation status of the exon 1, GR pro- HDACs that prevent histone acetylation and thus transcription factor moter between the two groups developed between PI and P6, the binding29, 30. Activation of chromatin through HDAC inhibition might period when differences in the maternal behavior of high-and low- trigger DNA demethylation by increasing the accessibility of DNA to LG-ABN dams are apparent. 8. By P6, the NGFI-A response element demethylase activity 0. We tested the hypothesis that inhibition of 5 CpG dinucleotide (site 16) was effectively demethylated in the HDACs with TSA would result in increased K9 acetylation of H3-his- high-, but not in the low-LG-ABN group. The group difference in tones associated with the exon 1, GR promoter, DNA demethylation, CpG dinucleotide methylation remains consistent through to adult- NGFl-A binding and reversal of maternal programming of str hood (P90; Fig. le). These findings, together with those of the cross- responses in the adult offspring of low-LG-ABN mothers fostering study, suggest that the group difference in DNA We first used ChIP analysis to determine whether histone H3-K9 ethylation occurs as a function of a maternal behavior over the acetylation and NGFl-A protein binding to the exon 17 gr pro first week of life. The results of earlier studies indicate that the first moter is altered in the offspring of high-and low-LG-ABN mothers week of postnatal life is a 'critical period'for the effects of early through intracerebroventricular (i.c.v. ) infusion of the adult off- experience on hippocampal GR express pring with TSA (100 ng/ml)or vehicle. Statistical analysis revealed a significant Group x Treatment interaction effect for both the his- Maternal effects on chromatin structure and NGFI-A binding tone H3-K9 acetylation(F=4.93, P<0.05)and NGFI-A(F=8.9 The next question concerns the functional importance of such dif- P=0.01). Post-hoc analysis showed that for both assays, vehicle ferences in methylation. DNA methylation is associated with treated offspring of low-LG-ABN mothers showed significantly changes in chromatin activity states. Chromatin gates the accessi- ("P<0.01)less association than any other group. These results indi- bility of promoters to transcription factors. Histone acetylation cate greater histone H3-K9 acetylation association and more bind 850 VOLUME 7 NUMBER 8 AUGUST 2004 NATURE NEUROSCIENCE
ARTICLES both the basal state of methylation and the first wave of de novo methylation after birth occur similarly in both groups. Whereas it is generally accepted that DNA methylation patterns are formed prenatally and that de novo methylation occurs early in development, there is at least one documented example of postnatal de novo methylation of the Hoxa5 and Hoxb5 genes27. Because similar analyses are not documented for other genes, it remains unknown whether changes in methylation are common around birth or whether they are unique to this GR promoter. The differences in the methylation status of the exon 17 GR promoter between the two groups developed between P1 and P6, the period when differences in the maternal behavior of high- and lowLG-ABN dams are apparent5,8. By P6, the NGFI-A response element 5′ CpG dinucleotide (site 16) was effectively ‘demethylated’ in the high-, but not in the low-LG-ABN group. The group difference in CpG dinucleotide methylation remains consistent through to adulthood (P90; Fig. 1e). These findings, together with those of the crossfostering study, suggest that the group difference in DNA methylation occurs as a function of a maternal behavior over the first week of life. The results of earlier studies indicate that the first week of postnatal life is a ‘critical period’ for the effects of early experience on hippocampal GR expression28. Maternal effects on chromatin structure and NGFI-A binding The next question concerns the functional importance of such differences in methylation. DNA methylation is associated with changes in chromatin activity states18. Chromatin gates the accessibility of promoters to transcription factors17. Histone acetylation at the lysine-9 (K9) residue of H3 and H4 histones is a well-established marker of active chromatin17,29. Acetylation of the histone tails neutralizes the positively charged histones, which disrupts histone binding to negatively charged DNA and thus promotes transcription factor binding. We tested the hypothesis that the maternal effect on DNA methylation results in (i) increased histone acetylation at the K9 residue of the H3 histone(s) associated with the exon 17 GR promoter and (ii) increased interaction between NGFI-A and the promoter sequence. We performed a chromatin immunoprecipitation (ChIP) analysis of histone H3-K9 acetylation and NGFI-A protein binding to the exon 17 GR promoter in the native chromatin environment in vivo. Intact hippocampi from adult offspring of high- and low-LG-ABN mothers were crosslinked in vivo by paraformaldehyde perfusion. We then selectively immunoprecipitated protein-DNA complexes with either an acetylated H3-K9 histone primary antibody or an NGFI-A primary antibody. The protein-DNA complexes were uncrosslinked, and the precipitated genomic DNA was subjected to PCR amplification with primers specific for the exon 17 GR promoter sequence. There were significant Group effects for the association of both histone H3-K9 acetylation (t = 2.1, *P < 0.001) and NGFI-A (t = 3.1, **P < 0.0001) with the exon 17 GR promoter sequence. These results indicated significantly greater histone H3-K9 acetylation association and threefold greater binding of NGFI-A protein to the hippocampal exon 17 GR promoter in the adult offspring of high- compared with low-LGABN mothers (Fig. 2). Thus, maternal programming of the exon 17 GR promoter involves DNA methylation, histone H3-K9 acetylation and alterations in NGFI-A binding. Reversal of maternally mediated epigenetic marking These findings suggest that maternal care influences hippocampal GR expression, and thus HPA function in the offspring, through epigenetic alterations that regulate NGFI-A binding to the exon 17 promoter. A critical question is whether the impact of early experience is reversible and whether epigenetic programming is modifiable in adult, postmitotic tissues? The generally accepted model is that the DNA methylation pattern is an irreversible reaction in adult post-mitotic cells. However, recent data from in vitro experiments suggests that in certain instances it is possible to induce replication-independent demethylation of ectopically methylated genes by increasing histone acetylation using the histone deacetylase (HDAC) inhibitor trichostatin A (TSA)29,30. Cytosine methylation attracts methylated DNA binding proteins and HDACs that prevent histone acetylation and thus transcription factor binding29,30. Activation of chromatin through HDAC inhibition might trigger DNA demethylation by increasing the accessibility of DNA to demethylase activity30. We tested the hypothesis that inhibition of HDACs with TSA would result in increased K9 acetylation of H3-histones associated with the exon 17 GR promoter, DNA demethylation, NGFI-A binding and reversal of maternal programming of stress responses in the adult offspring of low-LG-ABN mothers. We first used ChIP analysis to determine whether histone H3-K9 acetylation and NGFI-A protein binding to the exon 17 GR promoter is altered in the offspring of high- and low-LG-ABN mothers through intracerebroventricular (i.c.v.) infusion of the adult offspring with TSA (100 ng/ml) or vehicle. Statistical analysis revealed a significant Group × Treatment interaction effect for both the histone H3-K9 acetylation (F = 4.93, P < 0.05) and NGFI-A (F = 8.97, P = 0.01). Post-hoc analysis showed that for both assays, vehicle - treated offspring of low-LG-ABN mothers showed significantly (*P < 0.01) less association than any other group. These results indicate greater histone H3-K9 acetylation association and more bind- 850 VOLUME 7 | NUMBER 8 | AUGUST 2004 NATURE NEUROSCIENCE Figure 4 TSA effects on cytosine methylation. (a,b) Methylation analysis of the 17 CpG dinucleotides of the exon 17 GR promoter in hippocampi of vehicle- and TSA-treated (100 ng/ml) adult offspring of high- and low-LGABN mothers (n = 5 animals/group). (a) Percentage of cytosine residues that were methylated (mean ± s.e.m.) for the first 15 CpG dinucleotides (*P < 0.05). (b) Percentage of methylated cytosines for the 5′ (site 16) and 3′ (site 17) CpG dinucleotides within the NGFI-A binding region (*P < 0.001; **P < 0.003). © 2004 Nature Publishing Group http://www.nature.com/natureneuroscience
ARTICLES (>3 fold) of NGFI-A protein to the hippocampal exon 1, gr a Treatment - Vehicle TSA promoter in the adult offspring of TSA-treated low-LG-ABN moth ers compared with the vehicle-treated offspring of low-LG-ABN LG-ABN*Low High Low High mothers(Fig 3); there were no significant differences between TSA treated offspring of low-LG-ABN mothers and either TSA- or vehi- aGR92kDa→ cle-treated offspring of high-LG-ABN dams. As expected, TSA a-tubulin 60 kD reatment did not change histone H3-K9 acetylation or NGFI-A binding in the adult offspring of high-LG-ABN mothers, because the GR exon 1, promoter region in the offspring of high-LG-ABN g401 mothers is normally associated with acetylated histones and highly bound with NGFI-A To determine whether TSA treatment reverses the maternal effect on methylation within specific CpG dinucleotides on the exon 17 GR promoter, we mapped the differences in methylation using the 8 sodium bisulfite technique, focusing on the NGFI-A consensus Vehicl sequence within the exon 1, region(Fig. la). Statistical analysis of the data across all 17 sites revealed a significant effect of Group b S (F=93. 2, P<00001 ), Treatment(F= 52.8, P<00001)and Region (F= 30.4, P<00001), as well as a significant Group xTreatment x -O- TSA high-LG/ABN a Region interaction(F= 2.1, P=0.01), Group x Treatment interac tion(F= 19.9, P< 0.0001), Group x Region interaction(F= 4.1, P< 0.0001) and Treatment x Region interaction(F= 2.8, P<00001). The results again revealed significant differences in the of regions of the exon 1, GR promoter sequence(Fig 4)with significant differences within the 5" CpG(site 8 16)and 3 CpG(site 17) dinucleotides of the NGFI-A consensus Before1020406090 s sequence(Fig. 4b). Statistical analysis of the data from these two o sites revealed a highly significant effect of Group (F= 43.8, Time(min) P<00001), Treatment(F= 65.3, P< 00001)and Region Figure 5 TSA eliminates the maternal effect on hippocampal GR (F=113.3, P<00001), as well as a significant Group x Treatment ression and hPa res o stress. (a)Top: a represe a Interaction(F= 16.0, P 0.0001), Group x Region interaction blot showing absolute levels of electrophoresed hippocampal GR (F=37.8, P<00001)and Treatment x Region interaction(F=4.5, immunoreactivity(IR)from vehicle- and TSA(100 ng/ml)-treated adult P=0.04). Post-hoc analysis revealed that TSA treatment signifi- offspring of high-or low-LG-ABN mothers Molecular weight markers cantly decreased the degree of cytosine methylation within the 5'(SeeBlue, Santa Cruz Biotech)correspond to a single major band at 92 (site 16)CpG dinucleotide of the NGFI-A binding region of the kDa. The middle panel shows the membrane reprobed for a-tubulin I exon 1, GR promoter in the offspring of low-LG-ABN mothers illustrating absolute levels of electrophoresed hippocampal protein bound o comparison to vehicle-treated low-LG-ABN mothers('P<0.001).major band at -60 kDa and the intensity of the signal was similar in all TSA treatment produced'demethylation' of the 5 CpG(site 16) lanes. The lower panel shows quantitative densitometric analysis(relative and 3 CpG(site 17)dinucleotides in the offspring of low-LG-ABN optical density, ROD)of GR IR levels from samples(n= 5 animals/group mothers, and hypomethylation of the 3'CpG(site 17) dinucleotide P<0.001).(b)Plasma corticosterone responses(mean ts.e. m )to a in the offspring of high-LG-ABN mothers(Fig. 4b). These findings 20-min period of restraint stress (solid bar) in vehicle- and TSA suggest that TSA treatment can reverse the hypermethylated status of (100 ng/mp-treated adult offspring of high- or low-LG-ABN mothers the exon 1,GR promoter in the offspring of low-LG-ABN mothers. (n=10 animals/group:"P<0.01) TSA treatment resulted in a more extensive change in DNA methyla- tion than maternal care per se, since the 3 CpG (site 17) dinu- cleotide, which is unaffected by maternal behavior, is partially NGFI-A response element, we found that methylation of the cyto demethylated in response to TSA treatment in both cohorts sine within the 5 CpG dinucleotide(site 16)completely eliminated (Fig 4b). Also, as in the original study(Fig 1b), maternal care altered the binding of NGFI-A, whereas methylation of the cytosine within he methylation status of other CpG dinucleotides in the exon 1, the 3 CpG dinucleotide(site 17)only slightly reduced NGFI-A pro- sequence; in the case of sites 1, 2, 5, 12, 14 and 15, these effects were tein binding(IC G W, M.S.& M.J. M, unpublished data) similarly reversed with central TSA infusion. The significance of these sites for transcription factor binding is currently unknown and Reversal of maternal effect on GR expression hus a focus of ongoing studies. Thus, stable DNA methylation GR gene expression in the hippocampus is increased in the adult marking by maternal behavior is reversible in the adult offspring offspring of high- compared with low-LG-ABN mothers.We hippocampus by pharmacological modulation of chromatin struc- suggest that such differences are mediated by the differential ture. While TSa altered the methylation of the both the 5 and 3 methylation of the 5 CpG dinucleotide(site 16)of the NGFI-A CpG sites within the NGFI-A response element, the former appears consensus sequence in the exon 1, GR promoter and the subse to be critical for the effect on NGFl-A binding to the exon 1, pro- quent alteration of histone acetylation and NGFI-A binding to the moter. In a previous in vitro study using electrophilic mobility shift exon 17 sequence. If the differential epigenetic marking regulates assays(EMSA)with purified recombinant NGFI-A protein and the expression of the exon 17 GR promoter in high -versus low-LG differentially methylated oligonucleotide sequences containing the offspring, then reversal of the epigenetic marking should be accom NATURE NEUROSCIENCE VOLUME 7 NUMBER 8 AUGUST 2004 851
ARTICLES ing (>3 fold) of NGFI-A protein to the hippocampal exon 17 GR promoter in the adult offspring of TSA-treated low-LG-ABN mothers compared with the vehicle-treated offspring of low-LG-ABN mothers (Fig. 3); there were no significant differences between TSAtreated offspring of low-LG-ABN mothers and either TSA- or vehicle-treated offspring of high-LG-ABN dams. As expected, TSA treatment did not change histone H3-K9 acetylation or NGFI-A binding in the adult offspring of high-LG-ABN mothers, because the GR exon 17 promoter region in the offspring of high-LG-ABN mothers is normally associated with acetylated histones and highly bound with NGFI-A. To determine whether TSA treatment reverses the maternal effect on methylation within specific CpG dinucleotides on the exon 17 GR promoter, we mapped the differences in methylation using the sodium bisulfite technique, focusing on the NGFI-A consensus sequence within the exon 17 region (Fig. 1a). Statistical analysis of the data across all 17 sites revealed a significant effect of Group (F = 93.2, P < 0.0001), Treatment (F = 52.8, P < 0.0001) and Region (F = 30.4, P < 0.0001), as well as a significant Group × Treatment × Region interaction (F = 2.1, P = 0.01), Group × Treatment interaction (F = 19.9, P < 0.0001), Group × Region interaction (F = 4.1, P < 0.0001) and Treatment × Region interaction (F = 2.8, P < 0.0001). The results again revealed significant differences in the methylation of a number of regions of the exon 17 GR promoter sequence (Fig. 4) with significant differences within the 5′ CpG (site 16) and 3′ CpG (site 17) dinucleotides of the NGFI-A consensus sequence (Fig. 4b). Statistical analysis of the data from these two sites revealed a highly significant effect of Group (F = 43.8, P < 0.0001), Treatment (F = 65.3, P < 0.0001) and Region (F = 113.3, P < 0.0001), as well as a significant Group × Treatment interaction (F = 16.0, P < 0.0001), Group × Region interaction (F = 37.8, P < 0.0001) and Treatment × Region interaction (F = 4.5, P = 0.04). Post-hoc analysis revealed that TSA treatment significantly decreased the degree of cytosine methylation within the 5′ (site 16) CpG dinucleotide of the NGFI-A binding region of the exon 17 GR promoter in the offspring of low-LG-ABN mothers in comparison to vehicle-treated low-LG-ABN mothers (*P < 0.001). TSA treatment produced ‘demethylation’ of the 5′ CpG (site 16) and 3′ CpG (site 17) dinucleotides in the offspring of low-LG-ABN mothers, and hypomethylation of the 3′ CpG (site 17) dinucleotide in the offspring of high-LG-ABN mothers (Fig. 4b). These findings suggest that TSA treatment can reverse the hypermethylated status of the exon 17 GR promoter in the offspring of low-LG-ABN mothers. TSA treatment resulted in a more extensive change in DNA methylation than maternal care per se, since the 3′ CpG (site 17) dinucleotide, which is unaffected by maternal behavior, is partially ‘demethylated’ in response to TSA treatment in both cohorts (Fig. 4b). Also, as in the original study (Fig. 1b), maternal care altered the methylation status of other CpG dinucleotides in the exon 17 sequence; in the case of sites 1, 2, 5, 12, 14 and 15, these effects were similarly reversed with central TSA infusion. The significance of these sites for transcription factor binding is currently unknown and thus a focus of ongoing studies. Thus, stable DNA methylation marking by maternal behavior is reversible in the adult offspring hippocampus by pharmacological modulation of chromatin structure. While TSA altered the methylation of the both the 5′ and 3′ CpG sites within the NGFI-A response element, the former appears to be critical for the effect on NGFI-A binding to the exon 17 promoter. In a previous in vitro study using electrophilic mobility shift assays (EMSA) with purified recombinant NGFI-A protein31 and differentially methylated oligonucleotide sequences containing the NGFI-A response element, we found that methylation of the cytosine within the 5′ CpG dinucleotide (site 16) completely eliminated the binding of NGFI-A, whereas methylation of the cytosine within the 3′ CpG dinucleotide (site 17) only slightly reduced NGFI-A protein binding (I.C.G.W., M.S. & M.J.M., unpublished data). Reversal of maternal effect on GR expression GR gene expression in the hippocampus is increased in the adult offspring of high- compared with low-LG-ABN mothers7,9. We suggest that such differences are mediated by the differential methylation of the 5′ CpG dinucleotide (site 16) of the NGFI-A consensus sequence in the exon 17 GR promoter and the subsequent alteration of histone acetylation and NGFI-A binding to the exon 17 sequence. If the differential epigenetic marking regulates the expression of the exon 17 GR promoter in high- versus low-LG offspring, then reversal of the epigenetic marking should be accomNATURE NEUROSCIENCE VOLUME 7 | NUMBER 8 | AUGUST 2004 851 Figure 5 TSA eliminates the maternal effect on hippocampal GR expression and HPA responses to stress. (a) Top: a representative western blot showing absolute levels of electrophoresed hippocampal GR immunoreactivity (IR) from vehicle- and TSA (100 ng/ml)-treated adult offspring of high- or low-LG-ABN mothers. Molecular weight markers (SeeBlue, Santa Cruz Biotech) correspond to a single major band at 92 kDa. The middle panel shows the membrane reprobed for α-tubulin IR, illustrating absolute levels of electrophoresed hippocampal protein bound to the transfer membrane. Molecular weight markers correspond to a single major band at ∼60 kDa and the intensity of the signal was similar in all lanes. The lower panel shows quantitative densitometric analysis (relative optical density, ROD) of GR IR levels from samples (n = 5 animals/group; *P < 0.001). (b) Plasma corticosterone responses7 (mean ± s.e.m.) to a 20-min period of restraint stress (solid bar) in vehicle- and TSA (100 ng/ml)-treated adult offspring of high- or low-LG-ABN mothers (n = 10 animals/group; *P < 0.01). © 2004 Nature Publishing Group http://www.nature.com/natureneuroscience