| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
,
* Departments of Pediatrics and Molecular and Human Genetics, Baylor College of Medicine, USDA Children’s Nutrition Research Center, Houston, Texas, USA;
Department of Biology and Biochemistry, University of Houston, Houston, Texas, USA; and
Department of Ecology, Evolution and Behavior, University of Minnesota, St. Paul, Minnesota, USA
1Correspondence: Departments of Pediatrics and Molecular and Human Genetics, Baylor College of Medicine, USDA Children’s Nutrition Research Center, Houston, TX 77030, USA. E-mail: waterland{at}bcm.edu,
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Key Words: transgenerational epigenetic inheritance • DNA methylation • methyl supplementation • nutrition
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
. Epigenetic mechanisms may also be meiotically heritable, potentiating transgenerational epigenetic inheritance. Epigenetic inheritance is well documented in plants and lower organisms (2
, 3)
but in mammals is prevented in most cases by two distinct waves of genome-wide epigenetic reprogramming. The epigenetic information carried by the sperm and egg genomes is largely erased during preimplantation development, then reset in a lineage-specific fashion as differentiation progresses. Later, during fetal development, a second wave of epigenetic erasure and re-establishment occurs specifically in primordial germ cells (4)
.
Despite these barriers, transgenerational epigenetic inheritance does occur in mammals (2
, 5)
. Interest in the phenomenon is motivated by animal model and human epidemiologic data indicating that the effects of nonmutagenic exposures can sometimes be conveyed for several generations. Nongenetic transmission through the male germ line, in particular, suggests epigenetic inheritance. Effects of exposure of male rats to the diabetic agent alloxan (6)
or, more recently, the pesticide vinclozolin (7)
are propagated through the male germ line for several generations. Similarly, human epidemiologic data suggest that effects of prepubertal nutrition are transmitted through the male germ line to affect cardiovascular and diabetes mortality generations later (8
, 9)
. A new paradigm is emerging in which familial inheritance of disease risk is conveyed not only by genes but also by epigenetic information accumulated over previous generations (3
, 10)
.
The first direct evidence of transgenerational epigenetic inheritance in mammals came from studies in mouse nucleocytoplasmic hybrids showing that the cytoplasmic "environment" of the early embryonic genome induces persistent (11)
and heritable (12)
changes in gene expression and CpG methylation. Mammalian transgenerational epigenetic inheritance has been best characterized in the agouti viable yellow (Avy) mouse. The Avy mutation was caused by retrotransposition of an intracisternal A particle (IAP) upstream of agouti (13)
, which normally regulates the production of a yellow pigment in hair follicles. A cryptic promoter in the IAP (Fig. 1
) drives ectopic agouti expression, causing yellow coat color and other effects. Avy is characterized as a metastable epiallele (14)
; CpG methylation and overall epigenotype at Avy are established stochastically in the early embryo, then maintained throughout life. Isogenic Avy/a mice therefore display a wide range of coat color phenotypes, from yellow (hypomethylated at Avy) to brown, or "pseudoagouti" (hypermethylated at Avy). (Epigenetic studies of Avy are facilitated by maintaining the allele in heterozygosity with the nonagouti "a" allele, which does not produce functional agouti protein.)
|
The coat color of Avy/a mice tends to resemble that of their Avy/a dams (15)
. Using elegant embryo transfer experiments, Morgan et al. showed definitively that this is due to epigenetic inheritance at Avy (16)
. In addition to epigenetic inheritance, Avy exhibits epigenetic lability to environmental influences during embryonic development. Dietary supplementation of female mice with methyl donors and cofactors induces systemic Avy hypermethylation (and a darker average coat color) in their Avy/a offspring (15
, 17)
. Together, the environmental lability and capacity for epigenetic inheritance at Avy suggest that environmental influences on Avy epigenotype may be inherited transgenerationally.
Cropley et al. recently reported evidence in support of this hypothesis (18)
. They selected female pseudoagouti Avy/a (F1) offspring of control and methyl-supplemented (F0) dams, then without further dietary manipulations, compared the coat color of their (F2) Avy/a offspring. Relative to the F2 Avy/a offspring of control pseudoagouti dams, those of pseudoagouti dams that were methyl-supplemented in utero exhibited a significantly darker coat color distribution. The authors’ interpretation was that supplement-induced alterations in Avy epigenotype in the developing F1 female germ line were transmitted to the F2 offspring (18)
. However, since the germ-line DNA of pseudoagouti Avy/a mice exhibits variable maintenance of their somatic epigenotype characterized by CpG hypermethylation at Avy (19)
, the findings of Cropley et al. do not demonstrate transgenerational inheritance of acquired epigenetic information, but rather indicate that maternal methyl supplementation helps prevent the loss of existing epigenetic information as Avy epigenotype is reset in the developing germ line.
The present study was designed specifically to test for transgenerational inheritance of acquired epigenetic information at Avy. The results of our three-generation cumulative exposure study indicate that diet-induced hypermethylation at Avy is not inherited transgenerationally. Our findings complement those of Cropley et al. (18)
and suggest that whereas environmental influences may support the maintenance of epigenetic information at Avy in the developing germ line, the allele is resistant to environmentally induced acquisition of new germ-line epigenetic information.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
, 20)
; strain background in these animals is a mixture of C57BL/6J and C3H/HeJ (17)
. This congenic colony has been maintained by sibling mating and forced heterozygosity for the Avy allele for >200 generations, resulting in an essentially invariant genetic background.
Study design
To provide optimal power by which to detect transgenerational inheritance of diet-induced hypermethylation at Avy, we conducted a three-generation cumulative exposure study (Fig. 2
). At 21 days of age, slightly mottled yellow Avy/a females (with paternally inherited Avy alleles) were assigned randomly to either the control NIH-31 diet (TD#95262, Harlan Teklad, Madison, WI, USA) or NIH-31 supplemented (per kg diet) with folic acid (5.0 mg), vitamin B12 (0.5 mg), choline chloride (5.8 g), and anhydrous betaine (5.0 g) (TD#7017, Harlan Teklad) (17)
. These F0 dams were provided their respective diets throughout pregnancy and lactation and throughout additional reproductive cycles, if required. At 8 wk of age, F0 dams were mated with a/a males. F1 offspring (21 days of age) were assessed for genotype (Avy/a or a/a) and digitally photographed. Each F1 Avy/a female was weaned onto the same diet as her mother, and mated with an a/a brother (or another a/a male) at 8 wk of age. The F2 generation was treated in a fashion identical to that of the F1 generation. In an attempt to minimize any bias due to potential association of fecundity with diet or coat color, we remated each F0, F1, and F2 Avy/a female until she either contributed 2 Avy/a females to the next generation or stopped producing litters. The experiment ended when F3 offspring were weaned and classified for genotype and coat color phenotype.
|
Coat color phenotyping
The digital photographs enabled coat color phenotype to be rated by a single observer (R.A.W.) blinded to diet group. Coat color was rated as yellow (<5% brown), slightly mottled (
5% but less than half brown), mottled (
half brown), heavily mottled (>half but 
95% brown), or pseudoagouti (>95% brown) (17)
.
Bisulfite sequencing
Quantitative bisulfite sequencing of the Avy allele was performed as described (17)
. The seven CpG sites assayed 3' of the Avy IAP insertion are located at nucleotide positions 910, 916, 934, 943, 956, 972, and 991 of GenBank accession # AF540972.
Data analysis
All data were entered independently into two Excel spreadsheets and the records were compared electronically to eliminate data entry errors. The effects of generation, supplement, and generation X supplement were analyzed by an ANCOVA, with generation as the continuous variable. An ANCOVA includes scaling of the generation factor, which is not possible in an ANOVA analysis, and is a key element of our statistical design. As our analysis did not include phenotype changes from the F0 to F1 generations, our approach is conservative with respect to the main effects and the interaction term. To examine the generational dynamics, we estimated the change in phenoscore per lineage between maternal and offspring between the F1 and F3 generations. These averaged slopes were then analyzed by ANOVA, which included generational effects, as they were slopes.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
). Supplementation did not affect litter size at weaning. On average, 82% of F1 and F2 Avy/a females contributed Avy progeny to the next generation, and this proportion did not differ significantly between the control and supplemented groups (Table 1)
.
|
Avy/a coat color as a "readout" of Avy methylation
Coat color of Avy/a mice was blindly rated according to a 5-point scale (Fig. 3
A), as described (17)
. Using quantitative bisulfite sequencing (21)
, we showed that average Avy methylation in tail DNA of Avy/a mice is highly correlated with coat color classified by our system (Fig. 3B
). In this and our earlier study (17)
, supplementation effectively shifted Avy/a mice from slightly mottled to heavily mottled; these two classes showed no overlap of average Avy methylation levels (Fig. 3B
). Moreover, there is no tissue specificity of average Avy methylation in Avy/a mice (17)
. Hence, it is valid to use coat color classification of Avy/a mice as a proxy for somatic Avy methylation.
|
Effects of supplementation on coat color
This study had two potential outcomes. If the effects of maternal diet on offspring Avy methylation are partially transmitted to the next generation, we would expect the difference in coat color between the two diet groups to increase with successive generations. If, on the other hand, the effects of methyl supplementation on Avy methylation are not transmitted transgenerationally, we would expect the coat color difference between the two diet groups to be stable across the three generations. Our findings (Fig. 4
A) are clearly consistent with the latter. The overall effect of supplementation was highly significant (P<0.01); in every generation, there were fewer slightly mottled and more heavily mottled Avy/a offspring in the supplemented relative to the control group. Neither the effect of generation (P=0.94) nor the (generation) x (supplementation) interaction (P=0.69) was significant, however, indicating that the group difference did not increase with successive generations (i.e., the effect of methyl supplementation was not conveyed transgenerationally).
|
Offspring sex was not a significant factor in the model. Nonetheless, to determine whether the lack of a transgenerational effect could be attributed to some sort of negative selection bias, we examined within each diet group the average coat color of Avy/a females and, more pertinently, Avy/a females who contributed Avy offspring to the next generation. In the supplemented group, females who contributed Avy alleles to the next generation tended to be darker than the overall population of Avy offspring (Fig. 4B
). In the F1 generation this appeared to reflect an increased fecundity of darker (higher coat color score) dams, since the average coat color of all Avy females did not differ from Avy offspring in general. In the F2 generation, however, the higher average coat color score of "contributing" Avy females was attributable more to a tendency for females to be darker than males rather than a fecundity bias (Fig. 4B
).
The tendency for darker coat color in Avy/a females who contributed to the next generation would tend to enhance, not negate, the potential to detect transgenerational inheritance of induced hypermethylation in this experiment. However, a lineage-based analysis showed that unlike the control group, in which average coat color score did not vary by generation, average coat color score in supplemented animals actually declined between the F1 and F2, and again between the F2 and F3 generations (Fig. 4B
) (P=0.02). This significant transgenerational decline in coat color score, which occurred despite continued provision of the methyl supplemented diet, suggests that the diet-induced change in Avy epigenotype was actively erased between generations.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
. However, despite the established precedent for transgenerational epigenetic inheritance at Avy (16)
, our current results indicate that Avy hypermethylation induced by dietary supplementation is not conveyed transgenerationally.
These results appear to contradict those of Cropley et al. (18)
, who recently reported transgenerational inheritance of a diet effect at Avy. One might ask if our study failed to detect a transgenerational effect merely due to insufficient statistical power. This is unlikely because our lineage-based analysis (Fig. 4B
) identified a recurrent transgenerational decline in coat color score in the supplemented group. This statistically significant effect is essentially the opposite of what we expected, and suggests that continuing the experiment for more generations (i.e., increasing the power) would yield the exact same result.
Our findings differ from those of Cropley et al. (18)
most likely due to differences in study design; most important, the initial epigenetic state of the Avy allele was different in the two studies. Cropley et al. (18)
studied the coat color distribution of Avy/a offspring born to pseudoagouti Avy/a dams who had been exposed to control or methyl supplemented diet in utero. Stochastic establishment of somatic Avy epigenotype appears to be completed before gastrulation (17)
. Hence, in Avy/a embryos fated to be pseudoagouti, the cells that will undergo germ-line differentiation are already hypermethylated at Avy. Avy CpG methylation in oocytes of pseudoagouti dams is higher than that in oocytes of yellow Avy/a dams (19)
, but still considerably lower than that in somatic tissues of pseudoagouti mice (Fig. 3B
), indicating that, in pseudoagouti Avy/a mice, the Avy epigenotype established during preimplantation development is partially lost during midgestation oocyte development.
Hence, rather than demonstrating inheritance of environmentally induced alteration in Avy epigenotype, the results of Cropley et al. (18)
suggest that maternal supplementation helps prevent the loss of existing epigenetic information during germ-line development. To test whether methyl donor supplementation can induce de novo Avy hypermethylation that is inherited transgenerationally, our study began with slightly mottled yellow Avy/a F0 dams (hypomethylated at Avy). The contrast between our results and those of Cropley et al. (18)
therefore likely reflects a mechanistic distinction between diet effects on maintenance and acquisition of epigenetic information in the germ line. We would predict that if the study of Cropley et al. (18)
was repeated, focusing on the offspring of yellow and slightly mottled Avy/a dams who had been exposed to control or methyl supplemented diet in utero, the findings would be consistent with ours.
The composition and timing of the diet exposures also differed between the two studies. In this and our previous study of methyl supplementation and Avy methylation (17)
, we used an NIH-31 diet moderately supplemented with folic acid, vitamin B12, betaine, and choline (the "MS" diet designed by Wolff et al., ref. 15
). Cropley et al. (18)
used a supplemented diet containing 3-fold higher levels of these nutrients, as well as additional zinc and methionine (the "3SZM" diet of Wolff et al., ref. 15
). We have observed high perinatal mortality in litters born to dams supplemented with the 3SZM diet (unpublished observations). Hence, it is possible that the failure of Cropley et al. (18)
to detect, as we did, an effect of supplementation in Avy dams is related to a bias introduced by toxicity of the 3SZM diet. Another important difference between the two studies is the timing of the exposure. Whereas Cropley et al. (18)
provided the experimental diets only from embryonic day 8.5 (E8.5) to E15.5 (the period of germ-line differentiation), Avy/a females in the current study were provided the experimental diets throughout life.
In considering why the effect of diet on Avy methylation is not inherited, it is noteworthy that the epigenetic mark or marks that convey maternal transgenerational epigenetic inheritance at Avy remain unknown. Any such mark must be both present in the oocyte and maintained during embryonic development. Initial data suggested that CpG methylation was the heritable mark (19
, 22)
. Blewitt et al. recently tested this hypothesis by measuring Avy methylation levels at different stages of embryonic development (19)
. When the Avy allele is passed from pseudoagouti dams, it does maintain relative hypermethylation in the zygote. Surprisingly, however, in Avy blastocysts derived from multiple Avy/a pseudoagouti females, the Avy allele was shown to be completely devoid of CpG methylation (19)
, effectively ruling out CpG methylation as the conveyer of transgenerational epigenetic inheritance at Avy.
CpG methylation appears to play a central role, however, in maintaining mitotic stability of Avy epigenotype. Phenotypic variation among isogenic Avy/a mice is highly correlated with interindividual differences in Avy CpG methylation (Fig. 3)
(16
, 17)
. More important, the permanent phenotypic change induced in Avy/a offspring by maternal methyl donor supplementation is completely explained by increased CpG methylation at Avy (17)
. Nonetheless, whereas epigenetic inheritance at Avy results in a partial recapitulation of somatic Avy CpG methylation levels across generations (16)
, the data of Blewitt et al. (19)
indicate that Avy epigenotype must involve additional mechanisms. So far, no study has reported characterization of epigenetic modifications at Avy other than CpG methylation. Future studies must 1) identify histone modifications, DNA binding proteins, etc., that correlate with Avy transcriptional activity in the soma, and 2) track their ontogeny in oocytes and Avy/a early embryos to identify marks putatively responsible for transgenerational epigenetic inheritance at Avy.
The fact that CpG methylation does not mediate transgenerational epigenetic inheritance at Avy (19)
likely explains why we did not detect transgenerational inheritance of dietary effects. If methyl donor supplementation acts specifically to induce CpG hypermethylation at Avy, this may suffice to persistently alter Avy epigenotype in the soma, but clearly would not be conveyed transgenerationally. This explanation is consistent with the recurrent transgenerational loss of Avy epigenetic information we detected in the supplemented group (Fig. 4B
). We postulate that the darker coat color of "contributing dams" in the F1 and F2 generations of the supplemented group reflects increased Avy CpG methylation that affected both soma and germ line but which, in lieu of ancillary epigenetic modifications that normally orchestrate Avy transcriptional repression, is not maintained beyond the blastocyst stage in the next generation.
Conversely, the data of Cropley et al. (18)
suggest that methyl supplementation during the period of germ cell development mitigates the loss of epigenetic information at Avy by maintaining not only CpG methylation, but also the affiliated epigenetic modifications normally associated with the transcriptionally repressed state. Since epigenetic inheritance at Avy must depend on one or more of these other modifications, an epigenetic effect of early diet could be inherited transgenerationally in this manner. Clearly, however, maintenance of existing epigenotype is fundamentally distinct from acquisition of new germ-line epigenetic information. Notably, the ability of a dietary methyl supplement to prevent tissue-specific loss of epigenetic information in midgestation is reminiscent of an effect recently reported at axin fused specifically in developing tail bud (23)
.
In summary, we have shown that unlike naturally occurring epigenetic variation at Avy, the Avy hypermethylation induced by maternal methyl supplementation is not inherited transgenerationally. The Avy mouse remains a promising mammalian system in which to study transgenerational inheritance of environmentally induced epigenetic variation. Further progress with this model, however, will require a more detailed understanding of the specific molecular mechanisms by which epigenetic information at Avy is both influenced by environment and conveyed transgenerationally.
|
ACKNOWLEDGMENTS |
|---|
Received for publication February 6, 2007. Accepted for publication May 3, 2007.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Related Article
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
