Worldwide, assisted reproductive technologies (ART), namely in vitro fertilisation (IVF) and intra cytoplasmic sperm injection (ICSI), are increasingly used to overcome various types of fertility disorders. In developed countries, live birth rates following such procedures account for 1 to 3% of pregnancies, and these figures are postulated to escalate. However, with increasing demand for ART, in lieu of apparently normal outcomes, there have been concerns regarding the safety of children born through such means.
Recent studies on the possible consequences of human ART on epigenetic modifications, particularly genomic imprinting, have fuelled the concerns of scientist and non-scientists alike. Such consequences are both more common and more severe in animal species subject to ART, due to yet undetermined reasons. Nevertheless, animal models are useful for investigations into the possible role of each step of ART (ovarian stimulation, gamete manipulation, in vitro fertilisation, embryo culture and embryo transfer) on epigenetic reprogramming and developmental abnormalities.
Here, by using evidence from both human and animal studies, we will look at the means by which ART may affect the health of the offspring, and attempt to delineate the mechanism through which such effects may occur.
ART and Prenatal Development
Figure 1 | Assisted reproductive technologies at a glance (adapted from Winston & Hardy 2002).
The presence of intricate bidirectional communication between the cleavage stage embryo and adjacent uterine epithelial cells allows for normal pre-implantation development and establishment of pregnancy. The uterine epithelial cells serve to coordinate embryonic development with changes in uterine function; therefore, interrupting this seemingly key relationship, for instance by asynchronous embryo transfer, can result in both short and long-term tribulations to the development and growth of the fetus.
In 1981 an experiment by Wilmut et al. was a testament to this notion, where by demonstrating that the temporary exposure of day 3 sheep embryos to an advanced uterine environment for 3 days, leads to augmented fetal size at day 37 of gestation in the same species (Wilmut & Sales 1981). Later works by Sinclair et al. (1998) which were based on the experimental approach of Wilmut et al., revealed a pattern of late gestation fetal growth, where the late gestation fetuses derived from embryos exposed to advanced uterine environment, grew to have a significantly higher ratio of both secondary to primary muscle fibres (Sinclair et al. 1998; Maxfield et al. 1998).
These atypical outcomes were linked to a temporal change in the expression of a member of the MyoD gene family known as Myf-5. This gene, under the influence of Sonic hedgehog and Wnt-1, is responsible for inducing the differentiation of mesodermal precursor cells into myoblasts (Maltin et al. 2001). Based on the existing studies of the interaction between myogenic regulatory proteins and chromatin modifying enzymes during skeletal muscle differentiation (Maltin et al. 2001); these observations suggest that this temporary asynchrony in the uterus must have affected the pluripotent cell population of the pre-implantation embryo.
In addition, mouse offspring produced by IVF and somatic cell nuclear transfer (SCNT), also exhibit increased body weight at birth (Scott et al. 2010). This contradicts the studies showing reduced birth weight in human children conceived via ART when compared to their naturally conceived counterparts (Ceelen et al. 2009). Although, it is important to bear in mind that the majority of animal models of ART used in such studies, are of young age and in good reproductive and physical health; the same does not apply to individuals who seek ART, individuals whom usually suffer from conditions that affect fertility or are of older age. It has been suggested that the multiple embryo transfer in women subject to ART, could be a factor in low birth weight, since children conceived via ART by single embryo transfer do not show reduced birth weight (Ceelen et al. 2009).
Epigenetics and Imprinting disorders
Figure 2 | Dynamic reprogramming of the epigenome during development (Niemitz & Feinberg 2004).
Recent studies report of the existence of links between ART (e.g. IVF, ICSI etc.) and syndromes involving epigenetic alterations, including, Beckwith-Wiedemann syndrome (BWS), Angelman syndrome (AS) and cancerous tumours such as retinoblastoma (RB). Imprinting disorders and aberrant DNA methylation are of specific importance in these syndromes.
Data from studies of sheep and mouse indicate that antrum formation in the ovarian follicle coincides with a rapid phase of global DNA methylation (Kageyama et al. 2007), which in relation to human ART, happens near the time of initiation of pituitary down regulation and ovarian stimulation prior to egg recovery and IVF. In addition periods of in vitro zygote and embryo culture during ART, coincide with global DNA demethylation, which takes place in the pre-implantation embryo, where the paternal genome is demethylated in the first cycle, whilst the maternal genome is passively demethylated at each cell division until the blastocyst stage (Figure 2) (Mayer et al. 2000) (Grace & Sinclair 2009).
The abnormalities seen in BWS are associated with multiple genetic and epigenetic modifications of the chromosomal band 11p15, which houses a large cluster of genes that are imprinted (DeBaun et al. 2003). This region contains two imprinted subdomains: a centromeric domain including P57KIP2, LIT1, and a telomeric subdomain including IGF2 and H19 (Lee et al. 1999).
Imprinting disorders of the maternal alleles of the H19 and IGF2 gene account for 15% of cases in BWS; where the normally expressed maternal allele of the H19 gene is differentially methylated and silenced; whereas IGF2 is abnormally activated, a change that is linked to increased risk of cancer in BWS patients. In addition, 40% of BWS patients show loss of imprinting of the maternal copy of LIT1, an antisense RNA which is normally expressed paternally (Smilinich et al. 1999). Under normal conditions, based on studies of mice, LIT1 is believed to regulate silencing of the cell cycle inhibitor and tumour suppressor, p57KIP2 (Fitzpatrick et al. 2002).
Molecular studies of children with BWS, who were conceived through ART, showed that some of these children had imprinting defects in LIT1 and/or IGF2-H19 subdomains. These children were conceived via various methods of ART, thus indicating the absence of a link between BWS and a specific form of ART. Analysis of prevalence of BWS in United States and United Kingdom showed a 4.5% and 4.0% prevalence of BWS in ART patients respectively, compared to the population rate of 0.76% (USA) and 0.997(UK) (Maher et al. 2003).
On the other hand, Angelman syndrome is a disorder caused by mutations in the UBE3A gene on chromosome 15. UBE3A is normally expressed from the maternal allele and encodes an ubiquitin protein ligase (Scheffner et al. 1993). In a number of AS patients, the maternally inherited imprinting control center (ICC) on chromosome 15 (proximal to SNRP), is aberrantly hypomethylated (3% of cases) or there are microdeletions within chromosome15q (70% of cases) (Buiting et al. 1998); this modification disturbs the paternal to maternal imprint switch and disrupts the establishment of the maternal imprint leading to AS.
In a study of children with AS who were conceived though ICSI, it was shown that the majority had hypomethylation of the ICC region on chromosome 15, with no sign of interstitial deletions within chromosome 15q (which accounts for most cases of AS). Thus the majority of these cases had sporadic imprinting disorders; this is interesting since most ART linked defects in imprinting are linked to the maternal allele, e.g. hypomethylation of LIT1 in BWS and ICC in AS (Cox et al. 2002). This suggests that there is a greater effect on the oocyte than the sperm, either due to the environmental changes that the oocyte is subjected to during ART, or the effects of the underlying cause of infertility on the oocyte.
Epigenetic modifications have also been linked to human cancer. Imprinting disorders, particularly LOI, have been observed in both childhood and adult tumours. Retinoblastoma (RB) is a childhood cancerous tumour of the retina; it is caused due to mutations in the tumour suppressor gene RB1. There are reports of RB cases with differential methylation of the region of chromosome 13 around the RB1 gene (Leach et al. 1990). A study in Netherlands discovered five cases of RB in children conceived by IVF, but regrettably their RB1 methylation pattern were not reported. In addition many other studies demonstrated no linked between ART and increased risk of retinoblastoma and other childhood cancers (Klip et al. 2001) (Bradbury & Jick 2004). In fact, studies have shown that twins are more likely to develop cancer than IVF children (Pinborg 2004).
Exogenous gonadotrophin is often used for the induction of superovulation in individuals seeking ART. Despite the lack of certainty as to whether exogenous gonadotrophin has an impact on the maturation process of the oocyte or the physiological environment of the uterus, this practice has shown to negatively influence embryo development and implantations rates of pups in mice (Van der Auwera & D’Hooghe 2001). In addition, a number of human studies of implantation rates have shown that higher rates of implantations are observed on occasions when large amounts of gonadotrophin are avoided.
Imprinted genes are key for normal embryonic development and placental function. During oocyte growth, many imprinted genes acquire their epigenetic marks, it has been suggested that hormonal stimulations during this phase may trigger epigenetic disruptions. The examinations of oocytes from stimulated and non-stimulated cycles, have demonstrated that the methylation statuses of several imprinted genes are altered in stimulated oocytes in both human and mouse (Sato et al. 2007).
The germinal vesicle and metaphase II oocytes from infertile women, subjected to superovulation, display a gain of methylation at the H19 locus, accompanied with a loss of PEG1 methylation. However, due to lack of certainty as to whether these altered methylation pattern are due to ovarian stimulation or other factors such as age or underlying health problems, the investigators were lead to use oocyte from superovulated mice.
Similarly, these mice also showed a slight increase in H19 methylation, suggesting that the altered methylation of the H19 locus in human may be, at least in part, due to ovarian stimulation.
Studies using immunostaining approaches have been successful in demonstrating a difference in methylation patterns of two-cell mouse embryos derived from non-stimulated and superovulated females (Shi & Haaf 2002). Moreover, abnormal patterns of DNA methylation have been associated with comprises in in vitro development of pre-implantation embryo. In a recent study by Fortier et al (2008) it was demonstrated that superovulation can also affect methylation patterns in superovulated embryos with in vivo development (i.e. using pseudopregnant female to host the embryo) and can even lead to loss of imprinting at specific loci in mouse placenta, with or without embryo transfer (Fortier et al. 2008).
Recent literature on maternal imprint acquisition in oocytes, indicate that trophectoderm derived tissues may be more prone to loss of imprinting than the embryo proper. More importantly, the emerging consensus suggests that ovarian stimulation may have a more prominent effect on the maintenance of imprinting rather than its establishment. However, these findings must be treated with caution when extrapolated to the human condition, since the majority of studies of superovulation and imprinting defects are reports from studies on mice. More long-term investigations into the effects of ovarian stimulation in mono-ovular species are needed to reach convincing evidence in relation to humans and ART.
A follow up study of children (8-18 yrs.) conceived by in vitro fertilization, reports of higher fasting glucose levels, altered adipose tissue deposition and higher systolic and diastolic blood pressure in these children compared to their naturally conceived counterparts (Ceelen et al. 2008) (Ceelen et al. 2007).
Whilst the majority of children conceived through ART appear healthy, the reports of metabolic abnormalities like those mentioned earlier, necessitate further investigations into factors that may lead to such abnormalities. A recent study by Scott et al. (2010), investigated the effects of IVF, ICSI and somatic cell nuclear transfer (SCNT) on glucose clearance, body weight and body composition of young adult mice.
Consistent with previous findings of in vitro culture and manipulations resulting in fetal overgrowth in utero (especially in animals), both male and female pups created via ART had a birth weight higher than the control mice; although, this increase in birth weight was transient, and by 1 week of age, all mice had similar body weights. Furthermore, glucose clearance took longer in ART females, especially in those generated via SCNT; meanwhile, insulin production was highest in IVF females, who displayed the highest insulin peak amongst all groups, after glucose administration, suggesting lower glucose tolerance level in IVF mice (Scott et al. 2010). In addition, glucose clearance in IVF males required more insulin compared to control mice.
In this study, aberrant glucose clearance seems more evident in mice generated by IVF than those derived from ICSI. This is counter to the expectations of the authors, who postulated that the greater mechanical manipulations associated with ICSI, will result in more substantial changes. This difference may be linked to the culture media where the embryos can be exposed to oxidative damage. The level of damage correlates with time, such that longer periods of time spent in culture medium increases the extent of damage to the embryo, therefore, since IVF requires more incubation time than ICSI, the gametes may be exposed to more oxidative damage than those derived from ICSI, which may alter their development (Enkhmaa et al. 2009).
Nevertheless, these findings show altered glucose parameters in young mice conceived through ART, but more interestingly, since the ART mice in this study were of similar weight to the control mice (i.e. non were obese), the observed insulin resistance is likely to be independent of obesity and may even cause it later in development. This is important with regards to investigations for long-term effects of ART on glucose metabolism and obesity in embryos exposed to in vitro manipulations (Scott et al. 2010).
Worldwide, there are many types of culture media used to cultivate embryos in IVF clinics. Some of which are based on classical and empirical recipes used in animal experimentations, and more recently those with formulations that try to replicate the in vivo environment in humans. However, many of the factors present in such culture mediums are unknown, and with regards to recent reports of imprinting disorders in children conceived with ART, this issue needs to be mechanistically investigated.
Animal studies have repeatedly displayed that in vitro culture, particularly in presence of serum; can compromise the early developmental potential of embryos, as well as affecting the establishment and outcome of pregnancy. Khosla et al. (2001) found that the modification of the culture medium via the addition of serum, results in differential methylation of the imprinted genes H19, Igf2, Grb10 and Grb7 in mice, this was complemented with reduced fetal weight (Khosla et al. 2001). In addition, the paternal allele of H19 was also found to be hypomethylated and aberrantly expressed (Doherty et al. 2000). This is of interest since a similar imprinting pattern of the H19 gene is seen in human cases of BWS that were conceived through ART.
Studies of cattle have shown that embryos subject to 8 days of in vitro culture in presence of serum, display loss of methylation at the maternally imprinted SNRP gene, as discussed previously, this gene is associated with Angelman and Prader-Willi syndromes in humans (Rooke et al. 2007). Moreover this change is also observed in mice, and is most prominent in extraembryonic and placental tissues (Gosden et al. 2003). Therefore, it is evident that the genomic imprinting of the mammalian gamete and embryo are both susceptible to environmental changes, the severity of which is determined by type and timing of the change.
On the other hand, despite the lack of strong evidence and inconclusive studies (mainly due to small sample sizes), there have been reports of deleterious effects in DNA and telomeres (Honda et al. 2001) as well as altered embryonic gene expression (Emiliani et al. 2000) in cryopreserved embryos.
Despite the current association between ART and human imprinting disorders, there have only been three out of nine recognised human imprinting syndromes concerned with ART, and even so their frequency remains particularly low. Moreover, It is of interest as to why such imprinting disorders are far more severe and frequent in animals compared to human ART pregnancies; a mystery, which maternal age and infertility alone cannot explain; although, higher reproductive rates and extended periods of culture embryo in animal studies, may be the reason for the high frequency of such anomalies.
Studies of assisted reproductive technologies and their putative consequences on the epigenetic signature of the genome, will in time allow us to fill major gaps in our understanding of how ART can affect pregnancy outcome and long-tem development. Animal studies of ART induced imprinting disorders will pave the way for identifying the extent to which different ART procedure affect genomic imprinting.
Ultimately, the fixation of the majority of studies on imprinting disorders following ART, may have served to divert attentions from epigenetic alteration to non-imprinted loci in both gametes and pre-implantation embryo, alterations which may harbour health related consequences that only become apparent later in adult life. Although there is yet no apparent human subject with said long term consequences, but evidence from animal studies show that such effects exist.
Bradbury, B.D. & Jick, H., 2004. In vitro fertilization and childhood retinoblastoma. British Journal of Clinical Pharmacology, 58(2), pp.209–211.
Buiting, K. et al., 1998. Sporadic Imprinting Defects in Prader-Willi Syndrome and Angelman Syndrome: Implications for Imprint-Switch Models, Genetic Counseling, and Prenatal Diagnosis. The American Journal of Human Genetics, 63(1), pp.11–11.
Ceelen, M. et al., 2009. Growth during infancy and early childhood in relation to blood pressure and body fat measures at age 8-18 years of in vitro fertilisation children and spontaneously conceived controls born to subfertile parents. Human reproduction (Oxford, England), 24(11), pp.2788–2795.
Ceelen, M.M. et al., 2007. Body composition in children and adolescents born after in vitro fertilization or spontaneous conception. Journal of Clinical Endocrinology & Metabolism, 92(9), pp.3417–3423.
Ceelen, M.M. et al., 2008. Cardiometabolic differences in children born after in vitro fertilization: follow-up study. Journal of Clinical Endocrinology & Metabolism, 93(5), pp.1682–1688.
Cox, G.F.G. et al., 2002. Intracytoplasmic Sperm Injection May Increase the Risk of Imprinting Defects. The American Journal of Human Genetics, 71(1), pp.3–3.
DeBaun, M.R., Niemitz, E.L. & Feinberg, A.P., 2003. ScienceDirect.com – The American Journal of Human Genetics – Association of In Vitro Fertilization with Beckwith-Wiedemann Syndrome and Epigenetic Alterations of LIT1 and H19. The American Journal of Human ….
Doherty, A.S. et al., 2000. Differential effects of culture on imprinted H19 expression in the preimplantation mouse embryo. Biology of Reproduction, 62(6), pp.1526–1535.
Emiliani, S.S. et al., 2000. Comparison of ethylene glycol, 1,2-propanediol and glycerol for cryopreservation of slow-cooled mouse zygotes, 4-cell embryos and blastocysts. Human reproduction (Oxford, England), 15(4), pp.905–910.
Enkhmaa, D., Kasai, T. & Hoshi, K., 2009. Long-Time Exposure of Mouse Embryos to the Sperm Produces High Levels of ROS in Culture Medium and Relates to Poor Embryo Development – Enkhmaa – 2008 – Reproduction in Domestic Animals. Reproduction in Domestic ….
Fitzpatrick, G.V.G., Soloway, P.D.P. & Higgins, M.J.M., 2002. Regional loss of imprinting and growth deficiency in mice with a targeted deletion of KvDMR1. Nature Genetics, 32(3), pp.426–431.
Fortier, A.L. et al., 2008. Superovulation alters the expression of imprinted genes in the midgestation mouse placenta. Human Molecular Genetics, 17(11), pp.1653–1665.
Gosden, R. et al., 2003. Rare congenital disorders, imprinted genes, and assisted reproductive technology. Lancet, 361(9373), pp.1975–1977.
Grace, K.S. & Sinclair, K.D., 2009. Assisted reproductive technology, epigenetics, and long-term health: a developmental time bomb still ticking. Seminars in reproductive medicine, 27(5), pp.409–416.
Honda, S.S. et al., 2001. Induction of Telomere Shortening and Replicative Senescence by Cryopreservation. Biochemical and Biophysical Research Communications, 282(2), pp.6–6.
Kageyama, S.-I.S. et al., 2007. Alterations in epigenetic modifications during oocyte growth in mice. Reproduction (Cambridge, England), 133(1), pp.85–94.
Khosla, S.S. et al., 2001. Culture of preimplantation embryos and its long-term effects on gene expression and phenotype. Human Reproduction Update, 7(4), pp.419–427.
Klip, H.H. et al., 2001. Risk of cancer in the offspring of women who underwent ovarian stimulation for IVF. Human reproduction (Oxford, England), 16(11), pp.2451–2458.
Leach, R.J. et al., 1990. Preferential retention of paternal alleles in human retinoblastoma: evidence for genomic imprinting. Cell growth & differentiation : the molecular biology journal of the American Association for Cancer Research, 1(9), pp.401–406–406.
Lee, M.P. et al., 1999. Loss of imprinting of a paternally expressed transcript, with antisense orientation to KVLQT1, occurs frequently in BWS and is independent of insulin-like growth factor II imprinting. Proceedings of the National Academy of Sciences of the United States of America, 96(9), pp.5203–5208.
Maher, E.R. et al., 2003. Beckwith-Wiedemann syndrome and assisted reproduction technology (ART). Journal of Medical ….
Maltin, C.A. et al., 2001. Impact of manipulations of myogenesis in utero on the performance of adult skeletal muscle. Reproduction (Cambridge, England), 122(3), pp.359–374.
Maxfield, E.K.E. et al., 1998. Temporary exposure of ovine embryos to an advanced uterine environment does not affect fetal weight but alters fetal muscle development. Biology of Reproduction, 59(2), pp.321–325.
Mayer, W. et al., 2000. Demethylation of the zygotic paternal genome. Nature, 403(6769), pp.501–502.
Niemitz, E.L. & Feinberg, A.P., 2004. Epigenetics and Assisted Reproductive Technology: A Call for Investigation. The American Journal of Human Genetics, 74(4), pp.11–11.
Pinborg, A., 2004. Neurological sequelae in twins born after assisted conception: controlled national cohort study. BMJ, 329(7461), pp.311–0.
Rooke, J.A.J. et al., 2007. Ovine fetal development is more sensitive to perturbation by the presence of serum in embryo culture before rather than after compaction. Theriogenology, 67(3), pp.9–9.
Sato, A. et al., 2007. Aberrant DNA methylation of imprinted loci in superovulated oocytes. Human reproduction (Oxford, England), 22(1), pp.26–35.
Scheffner, M. et al., 1993. The HPV-16 E6 and E6-AP complex functions as a ubiquitin-protein ligase in the ubiquitination of p53. Cell, 75(3), pp.495–505.
Scott, K.A. et al., 2010. Glucose parameters are altered in mouse offspring produced by assisted reproductive technologies and somatic cell nuclear transfer. Biology of Reproduction, 83(2), pp.220–227.
Shi, W. & Haaf, T., 2002. Aberrant methylation patterns at the two-cell stage as an indicator of early developmental failure. Molecular reproduction and development, 63(3), pp.329–334.
Sinclair, K.D. et al., 1998. Fetal growth and development following temporary exposure of day 3 ovine embryos to an advanced uterine environment. Reproduction, fertility, and development, 10(3), pp.263–269.
Smilinich, N.J.N. et al., 1999. A maternally methylated CpG island in KvLQT1 is associated with an antisense paternal transcript and loss of imprinting in Beckwith-Wiedemann syndrome. Proceedings of the National Academy of Sciences of the United States of America, 96(14), pp.8064–8069.
Van der Auwera, I. & D’Hooghe, T., 2001. Superovulation of female mice delays embryonic and fetal development. Human reproduction (Oxford, England), 16(6), pp.1237–1243.
Wilmut, I. & Sales, D.I., 1981. Effect of an asynchronous environment on embryonic development in sheep. Journal of reproduction and fertility, 61(1), pp.179–184.
Winston, R.M.L. & Hardy, K., 2002. Are we ignoring potential dangers of in vitro fertilization and related treatments? Nature Cell Biology, 4, pp.S14–S18.