Epigenetics and Beyond
ISIS Report 21/01/09
Epigenetic Toxicology
Environmental toxins modify our genes and affect our children and grandchildren;
enormous implications for risk assessment of synthetic chemical and other xenobiotics
Dr. Mae-Wan Ho
A fully
referenced version of this article is posted on ISIS members’ website. Details
here
An electronic version of the full report can be downloaded from the ISIS online
store. Download Now
Bisphenol A in baby bottles and other plastics cause developmental defects
across generations
A common chemical bisphenol A (BPA) that leaches out from baby bottles and
other containers of food and drink, when given to pregnant mice, caused gross
chromosomal defects in the egg cells of her female foetuses. These female
foetuses grew into adults and produced large proportions of defective embryos
[1]. It was a case of the grandmother effect spanning three generations.
BPA is a weak oestrogen mimic, and one of the most abundant synthetic
chemicals manufactured by the chemical industry; some 2 to 3 million tonnes
made every year .It is the monomer base of polycarbonate plastic food and
drinks containers, including baby bottles, the resin lining of tins and dental
sealants; it is also found in ‘carbonless’ paper used for receipts as well
as a wide range of other common household products. BPA has been linked to
a range of reproductive defects in lab rodents, including reduced sperm production,
and increased susceptibility to prostate cancer in the male, and abnormalities
in mammary gland, brain development and the oestrous cycle in the female.
Foetal exposure to BPA in rats was shown to induce the development of precancerous
and cancerous lesions in the mammary gland [2]. In humans, an association
between BPA levels and recurrent miscarriages has been suggested [3], and
recently, a significant relationship was found between urine concentrations
of BPA and cardiovascular disease, type 2 diabetes, obesity, and liver enzyme
abnormalities in a representative sample of the adult US population [4].
The US research group led by Patricia Hunt at Washington State
University, Pullman, made headlines around the world when they published the
results in 2007 [1]. They also found that knocking out the oestrogen receptor
ERb in transgenic mice reproduced precisely those defects caused
by BPA, suggesting that BPA interfered with the binding of oestrogen to its
receptor ERb. And indeed, no further
abnormalities could be induced by BPA in the ERb
knock-out mice.
These results provide evidence for a multigenerational
effect of BPA exposure mediated by the dysregulation of ERb,
so the daughters of exposed pregnant females have an increased risk of producing
offspring with abnormal chromosomes. The BPA exposure also adversely influences
the reproductive lifespan of the females, as these abnormalities typically
result in the loss of a significant proportion of oocytes in developing female
foetuses prior to sexual maturation.
The findings of Hunt’s team [1] that really attracted the attention
of the scientific community were the epigenetic effects (resulting from interference
of oestrogen binding to its receptor) transmitted across generations, which
have become increasingly important in toxicology.
During foetal development, the germ cells in both sexes undergo massive proliferation
by ordinary cell division. Thereafter, germ cells in the testis of the male
foetus stop dividing and remain in developmental arrest until after birth, while
those in the ovary of the female foetus start meiosis (the special cell division
that leads to the formation of egg cells with half the chromosome number). By
the time of birth, the oocytes (as they are now called) have entered a protracted
period of meiotic arrest where they remain until just prior to ovulation after
an extensive period of growth in the adult ovary.
Since the publication of these findings, the Canadian federal
government has formally declared BPA a hazardous substance, and announced
they would seek to restrict imports, sales and advertising of polycarbonate
baby bottles containing BPA [5].
In contrast, a battle is still raging in the US, where the Federal
Drugs Administration (FDA) is widely criticised for ignoring its own panel
of scientific advisors and siding with the American Chemistry Council (industry’s
lobby group) in insisting that BPA is not harmful at current exposure levels.
In Europe, similarly, the European Food Safety Authority has
repeatedly reaffirmed its opinion that BPA is ‘safe’, dismissing all scientific
evidence on grounds of ‘reliability’ and ‘consistency’.[6].
The toxic effects of BPA should be considered along with
other hormone disruptors such as DES and a host of pesticides in the context
of the new discipline of epigenetic toxicology. Human epidemiologic evidence
reveals that individuals exposed to DES in utero during the first 3
months of pregnancy showed increased incidences of reproductive disorders
and the rare cancer, clear cell adenocarcinoma of the vagina. Increased incidences
of uncommon disorders were also found in the granddaughter and grandsons of
DES-exposed women, suggesting epigenetic inheritance [7]. This has been confirmed
in experiments on rats (see later).
The evidence that adult health and disease is influenced by the environment
of foetal or early neonatal development is not unique to endocrine disruption.
Low birth size and poor nutrition have been associated with increased risk of
heart disease, type 2 diabetes, osteoporosis and metabolic dysfunction [8].
Gestation represents a developmental window of vulnerability to epigenetic changes
by nutritional and environmental factors [9] (see Epigenetic
Inheritance, “What Genes Remember”, SiS 41). There is a mounting
body of evidence from both gene-specific and genome-wide studies that environmental
exposures particularly in early development can induce epigenetic changes that
may be transmitted to subsequent generations and/or lead to diseases in later
life.
Toxicogenomics’ referred to a combination of conventional toxicology,
the study of poisons especially their effects of organisms and the ecosystem,
with genomics, the study of the function of nucleotide sequences in the genomes
of organisms [10]. It has contributed immensely to defining the adverse biological
effects of environmental stressors, toxins, drugs and chemicals; but is rapidly
being transformed into ‘epigenetic toxicology’; here defined as potentially
heritable changes in gene expression with, or without accompanying
alterations in the DNA sequence of the genome [9] .
Epigenetic mechanisms are necessary for normal development and
differentiation, but these can be misdirected, leading to diseases, notably
cancers.
The epigenetic progenitor model of cancer and mechanisms
Inherited and spontaneous or environmentally induced epigenetic alterations
are increasingly recognized as early molecular events in cancer development,
as summarised in a review published in Cancer Journal in 2007 [11],
which also described a new epigenetic model of cancer, involving stem cells.
Stem cells, present in all multi-cellular organisms, are capable of renewing
themselves through cell division and are also pluripotent, i.e., capable of
giving rise to a diverse range of differentiated specialized cells [12] (see
Hushing Up Adult Stem Cells, SiS
13/14).
Cancer arises in three steps [11]. First, an epigenetic alteration
of stem/progenitor cells within a tissue, affecting genes that regulate the
expansion of progenitor cells and increase their capacity for self-renewal
and pluripotency. For example, loss of regulation of IGF2 promotes
an expansion of the progenitor cell compartment, increasing the probability
of tumour formation. This alteration can be due to events within the stem
cells themselves, the influence of the stromal compartment (the supporting
framework of connective tissues in and around organs), or environmental damage
or injury.
Second, a gatekeeper mutation occurs in a tumour-suppressor gene
(in solid tumours), or rearrangement of an oncogene. At the third and final
stage, genetic and epigenetic instability set in, leading to increased tumour
evolution.
Epigenetic alterations are potentially more damaging than
nucleotide mutations because their effects on regional chromatin structure
can spread, thereby affecting multiple genetic loci (units, usually of transcription).
Furthermore, they tend to affect a high proportion of those exposed, unlike
conventional gene mutations, which are relatively rare. The most commonly
described changes in cancer are alterations in the methylation pattern of
DNA, but epigenetic modifications of histone proteins are also implicated.
DNA methylation, the covalent addition of a methyl group to the
C5 position of cytosine, is involved in many key developmental processes,
deregulation of which would result in major developmental abnormalities. These
include the inactivation of alleles in imprinted genes, genes that are expressed
or not at different times according to whether they come from the mother or
the father; inactivation of one of the two X chromosomes in the female to
compensate for the male that has only one X chromosome; and in suppressing
transposable elements, which would otherwise become expressed, jump around
and disrupt genes and genome function in general.
DNA methylation occurs most frequently in CpG islands, and results
in a conformational change in the major groove of the DNA that alters protein
binding. CpG islands are regions in the genome with a high GC content and
frequent CpG occurrence. The human and mouse genome projects identified ~15
500 and ~29 000 CpG islands respectively. Hypermethylation of CpG-rich regions
of gene promoters inhibits expression by blocking the initiation of transcription.
Transcription of a number of tumour suppressor genes such as
p16 INK4a, BRCA1, p53, and hMLH-1 were
found to be inhibited by promoter hypermethylation in cancers.
Although silencing of tumour suppressor gene by DNA methylation
occurs frequently in cancer, genome-wide hypomethylation is one of
the earliest events in the genesis of cancer. Demethylation of the genome
can lead to the reactivation of transposable elements, thereby altering the
transcription of adjacent genes, the activation of oncogenes such as H-RAS,
and inappropriate expression of imprinted genes. Furthermore, genomic instability
associated with the hypermethylation of the DNA mismatch repair enzyme gene
MLH1 may deregulate not only critical genes involved in the initial
stages of carcinogenesis, but also those involved in the later invasion and
metastasis stages.
Two distinct sets of genes that potentially link environmental
exposures during pregnancy to adult disease susceptibility are imprinted genes
and genes with metastable expression states. Imprinted genes are inactivated
according to the parent of origin, so that only one of the two alleles (forms
of a gene) should be functional. Metastable epialleles (epigenetic
forms of an allele) have highly variable expression because of stochastic
changes in their epigenetic state, in at least some cases such as the mouse
Agouti gene, due to the insertion of a virus like intracisternal A
particle.
Imprinted genes can be deregulated in both germ cells and somatic
cells. Because imprinted genes are frequently clustered and their expression
co-ordinately regulated by imprinted control regions, a single genetic or
epigenetic change in an imprinted control region can result in disrupting
many genes. Consequently, imprinted genes are associated not only with severe
developmental disorders such as Angelman, Beckwith-Widemann, and Prader-Willi
syndromes, but also with cancer. Imprinted genes are at much greater risk
of inactivation by mutation and epigenetic alteration because one allele is
already inactivated by genomic imprinting. Because imprinted genes normally
encode either positive or negative growth effectors, they are frequently involved
in the formation of a wide range of tumours.
The epigenome is particularly susceptible to disturbance of regulation
by environmental factors during gestation, neonatal development, puberty,
and old age. Age-correlated increases in DNA promoter methylation occur in
a number of genes involved in cancer, including IGF2. Versican,
and PAX6. Alterations in epigenotype have also been observed after
adult exposure to xenobiotic chemicals (synthetic chemicals not normally produced
or metabolised by organisms).
Studies in animals provide clear evidence of epigenetic inheritance
of disease states for generations after the initial exposure, especially in
two cases involving hormone disrupting chemicals.
Fungicide caused high incidence of illness up to four generation after exposure
The anti-androgenic compound vinclozolin is a fungicide mainly used on oilseed
rape and peas in the UK and on vines, fruit and vegetables worldwide [13].
It was first introduced by BASF in Germany in 1976 and sold under a number
of trade names including Ronilan and Flotilla. Vinclozolin is linked to testicular
tumours in rats, and reproductive toxicity, and the UK Advisory Committee
on Pesticides has kept it under review since 1991, while US Environment Protection
Agency considers it an endocrine-disrupting chemical. Vinclozolin is metabolized
into more active compounds with higher binding affinity to the androgen receptor.
Researchers at Washington State University Pullman in the United
States led by Michael Skinner found that exposure of rats during embryonic
development at the time of sex-determination resulted in adult animals from
the F1, and subsequent generations up to F4, developing a number of diseases
or abnormalities of the prostate, kidney, immune system, testis, and tumours
in various tissues. In addition, several blood abnormalities developed including
high cholesterol levels.
High incidences of trans-generational disease states were found,
consistently across all generations, and appeared to be due in part to epigenetic
alterations in the male germ-line, and successive studies carried out confirm
the effects [14].
High doses were administered at 100mg/kg/day from embryonic day
8-14 of gestation, with no further treatment in subsequent generations, and
inbreeding was avoided. So only the F1 embryos and F2 germ cells (in the F1
embryos) were exposed.
The diseases and abnormalities were scored ‘blind’, i.e., the
scorer did not know if the rats were in the group that had been exposed to
the chemical.
Testis abnormalities included atrophied tubules, vacuoles or
failure of germ cell formation at a rate of 20 percent or greater. Kidney
lesions include tubular damage at 30 percent or more, tubular changes involved
extreme dilation with protein-rich fluids, fluid-filled cystic tubules, thickening
of the Bowman’s capsule surrounding the glomerulus, as well as reduced glomerular
area. Ventral prostate tissues were considered abnormal if more than 30 percent
of prostate ducts atrophied and contained no columnar secretary epithelial
cells. Immune related abnormalities include excessive macrophage and lymphocyte
invasion into multiple organs and generally accompanied by bacterial infection.
Several types of inflammation were identified in the inner ear, lower limbs,
and lower respiratory tract. Sub-dermal abscesses were found which grew in
size and caused septicaemia (widespread infection).
There were no tumours in control animals while tumours developed
in 12-33 percent of the animals in the F1-F4 generation exposed animals: 4
breast adenomas, 2 breast carcinoma, 1 lung sarcoma and 2 skin Merkel cell
melanoma together with the breast adenoma; only the lung sarcoma, Merkel cell
tumour and breast carcinoma were malignant. Prostate lesions were detected
in 45-55 percent of males. Renal lesions appeared in 20 to 50 percent of males
of all vinclozolin-exposed generations, and were also found in females. Abnormal
testis function and morphology were found in 15-38 percent of the F1-F4 generations
of exposed animals. The pathology included increase in spermatogenic cell
death, gross morphological defects in spermatogenesis, and a complete lack
of spermatogenesis.
Inflammation of the inner ear, sub-dermal abscesses and bacterial
infection were found in 12-33 percent of exposed animals, with no inflammation
in controls. Significant increase in cholesterol was found in 35 percent of
exposed animals at 6-14 months old, but not at 3 months.
Up to 50 percent of exposed animals also developed an apparent premature ageing
with appearance and behaviour at 6-14 months similar to controls older than
18 months.
The frequency of disease in the F1 generation was often less than subsequent
generations. Many animals had multiple abnormalities, and 85 percent of all
F1-F4 vinclozolin-exposed animals developed a transgenerational disease state.
Over 90 percent of all males in F1-F4 vincozolin-exposed group had a reduced
capacity of making sperms.
The disease phenotype was primarily transmitted through the male
germ-line; a vinclozolin treated F2 male out-crossed to a wild-type female
gave an increase in disease prevalence over controls in the F3 generation
(though less than the vinclozolin F3), whereas the reverse cross did not.
But the female germline also contributed to disease.
Although tumours, renal lesions and prostate lesions are observed
in aged (24 mo) rats, none of these were observed in the controls at 6-14
mo, when the two groups of animals were compared.
The results were highly significant, and the frequencies of defects
were much higher than could be accounted for by mutations rates, which are
typically 1 to 5 percent. The researchers had used an exposure level that
is almost 10 times the official lowest level at which adverse effects were
observed, i.e., 11 mg/kg/day; but biological effects have been demonstrated
at doses around 1mg/kg/day. Environmental levels of vinclozolin have not yet
been rigorously determined.
Analysis of the sperm from the F2 and F3 generations identified
25 candidate DNA sequences with altered methylation patterns in the vinclozolin-exposed
animals, with 15 sequences confirmed to have specific hypermethylation [15].
These sequences mapped to specific genes and non-coding DNA regions. The expression
pattern of several of the genes during embryonic development was altered in
the vinclozolin F1 and F2 generation testis, with some decreased in expression
while others were increased.
All candidate genes altered in the vinclozolin samples were not
in the controls. They are present on various autosomes with no major hot spot
regions, and none was present on the sex chromosomes.
Synthetic oestrogen causes cancer in descendants of exposed female animals
The synthetic oestrogen diethylstilbestrol (DES) is a potent prenatal endocrine
disruptor. Exposure in humans and experimental animals during critical periods
of reproductive tract development in embryogenesis permanently alters oestrogen
target tissues and results in long-term abnormalities such as cancer of the
uterus later in life.
For almost 30 years, DES was prescribed to women with high-risk
pregnancies to prevent miscarriages and other complications. In 1971, a clinical
report associated DES with a rare form of reproductive tract cancer, vaginal
adenocarcinoma, which was detected in a small number (0.1 percent) of adolescent
daughters of women who had taken DES while pregnant [16]. Subsequently, DES
was also linked to more frequent benign reproductive tract problems in an
estimated 95 percent of the DES-exposed daughters; reproductive organ dysfunction,
abnormal pregnancies, reduced fertility and disorders of the immune system.
Similarly, male offspring of exposed females showed structural, function,
and cellular abnormalities in the reproductive organs, inflammation, and decreased
fertility [17].
DES is no longer used clinically to prevent miscarriage, but
the exposed are still to reach an age where higher incidence of cancers are
expected, and the possibility of second-generation effects has been reported,
i.e., in the grandchildren of women prescribed DES, as mentioned earlier.
Recent research in the United States led by Retha Newbold at
the National Institutes of Environmental Health Sciences of the National Institute
of Health, North Carolina, found that in mice, exposure to DES leads to altered
gene expression that includes an oestrogen-regulated component [18]. DES is
a model endocrine disruptor and even low doses increase the incidence of cancer
in the uterus, as do low doses of other environmental oestrogen. The increased
propensity to develop tumours is transmitted through the mother to subsequent
generations of male and female descendants. The mechanisms include both genetic
and epigenetic events.
Neonatal mice (day 1-5) treated with DES at 2mg/pup/d
developed a high incidence of cancer (90-95 percent) at 18-24 months of age.
These tumours rarely spread throughout the body, but in aged animals (24 mo
or older) the lesions spread to para-aortic lymph nodes or extended to contiguous
organs. Similar findings were obtained in rats and hamsters. This may be predictive
of carcinogenic potential of environmental estrogens in women as they age,
as both the histology of the tumours and the progression to cancer are similar
to those seen in women.
In a dose response study, DES was found to cause tumours even
at a dose of 0.0002mg/pup/day. Uterine cancer followed a linear
dose-dependent response from 0 in controls to 65 percent at 2mg) as did increase in uterine weight/body weight
(estrogenicity) (01 in control to 0.3 at 2 mg/pup/day)
Other environmental oestrogens also caused uterine lesions: 17b-estradiol,
tamoxifen, hexestrol, tetrafluorodiethylstilbestrol, ethinyl estradiol, 2-hydroxyestradiol,
4-hydroxyestradiol, genistein, nonylphenol, bisphenol A, methoxychlor. Most
were tested at 2mg/pup/day, except
the weaker estrogens, genistein, nonylphenol, bisphenol A and methoxychlor,
which were administered at 200 mg/pup/d.
All except methoxychlor caused uterine lesions in aged mice. Methoxychlor
failed to cause lesions possibly because neonate liver cannot yet convert
it into oestrogen.
Prenatal DES exposure was found to delay the expression of Hox
genes involved in the development of the reproductive tract. Wnt genes
were also affected. Neo-natal exposure caused demethylation of the oestrogen-responsive
gene LF in the mouse uterus, which may be involved in tumour induction.
In hamsters exposed to DES, similarly, uterine carcinoma develop
at a high frequency, and imbalances in the oestrogen-regulated uterine expression
of c-jun, c-fox, c-myc, bax,
bcl-2 and bcl-x proto-oncogenes probably played a role.
Microarray studies with mouse uterus revealed similar altered
gene expression pathways that included an oestrogen-regulated component.
In a further series of experiments, prenatal or neonatal treatment
with DES led to tumours in the female and male genital tract, and in addition,
the susceptibility for tumours was transmitted to the descendants through
the maternal germ cell lineage. Transmission via the DES-exposed male was
not studied.
The mice were treated with DES prenatally at 2.5, 5 or 10 mg/kg/dy
on day 9-16 of gestation, or at 0.002 mg
on day 1-5 after birth; these were the highest doses that did not drastically
interfere with fertility later in life. When F1 female mice reached sexual
maturity, they were bred to control untreated males. Female and male offspring
were aged to 17-24 months and examined for genital tract abnormalities. An
increased incidence of proliferative lesions of the rete testis (a network
of ducts in the centre of the testis associated with the production of sperms
and an oestrogen target tissue in the male), and tumours of the reproductive
tract were observed in the male offspring. In female offspring, an increased
incidence of uterine adenocarcinoma was seen. The incidence was lower in DES
descendants than in their parent, (31 percent in F1 at 18 months at neonatal
dose of 0.002 compared with 11 percent in their F2 descendants).
Several genes were permanently dysregulated after DES treatment. The oestrogen-responsive
proteins LF and c-fos were permanently up-regulated in the uterus, and the
promoter regions of these genes were hypomethylated. LF was also over-expressed
in uterine tissues from female offpring of the DES exposed females.
It is important, therefore, to follow the grandchildren of women
exposed to DES.
Large implications for public health and risk assessment of xenobiotics
These findings in epigenetic toxicology have large implications for public
health and the risk assessment of xenobiotics. It is estimated that more than
100 000 xenobiotics are on the market in the European Union, some 70 000 are
potentially hazardous for human health and/or the ecosystem [19]. The vast
majority of these chemicals have not been adequately tested for safety before
they were released. Many of them are endocrine disruptors similar to those
whose epigenetic effects have been reviewed in this article. Common household
products – detergents, disinfectants, plastics and pesticides – contain endocrine
disruptors. For example, 56 pesticides in use have been identified as known
or suspected endocrine disruptors by the European Union and the scientific
community [20], but are still largely unregulated.
The new findings call for urgent action on two fronts. First, all known and
suspected endocrine disruptors and carcinogens should be banned or phased out
where there is overwhelming scientific evidence of harm, or the precautionary
principle, where there is reasonable suspicion of harm. Pesticides and chemical
fertilizers are good candidates for phasing out, if not a total ban, as there
is now substantial evidence that organic agriculture, which dispenses with pesticides
and other chemical inputs, works on all scales, and compost and green manure
can maintain or indeed, increase yields over chemical fertilizers [21]
(see Food Futures Now *Organic
*Sustainable *Fossil Fuel Free,ISIS publication).
Second, an adequate protocol involving transgenerational studies
with microarray analyses for genetic and epigenetic effects should be used
in all toxicological investigations.
One class of xenobiotics that must be included are genetically modified food
and feed, for which epigenetic effects have already been demonstrated in exposed
animals, as well as health impacts in both laboratory feeding studies and on
livestock and farm workers in farmers’ fields [22-24} (GM is Dangerous and Futile.
SiS 40; GM
Maize Reduces Fertility & Deregulates Genes in Mice, and GM Maize Disturbs
Immune System of Young and Old Mice, SiS 41).
| 
