Dr. Tycko offered two comments, although he noted that he was restricted to his own expertise.
He indicated that one area of further emphasis should be investigating the hypothesis that there may be interindividual variation in epigenetic marks that affect the expression of imprinted genes in the placental trophoblast, and thereby lead to variations in placental growth. He noted that such variation may account for some cases of IUGR.
He added that a second area of emphasis may be studies to elucidate the biochemical and cell biological functions of imprinted genes in the human placenta and in mouse models. He indicated that it is known that imprinted genes affect "supply and demand" for maternal nutrients, but added that the pathways mediating this balance were only beginning to be worked out. He stated that this work may ultimately yield novel approaches, perhaps even nutritional (although this was speculative), that could improve pregnancy outcome.
The growth program in mammals is controlled by genetic and epigenetic mechanisms and environmental factors. Epigenetic marks are those that are heritable through cell divisions, but are not dependent on changes in DNA sequence. These marks act like switches that help to control gene activity. Signals from the environment may change the epigenetic program that may ultimately modulate the genetic program.
In recent years most studies on growth programming have been focused on the identification of growth control genes. In contrast, little is known about how epigenetic mechanisms orchestrate the genetic program to control growth. The latter is an exciting new frontier of research.
Genomic imprinting is regulated by epigenetic mechanisms. Imprinted genes are crucial for the control of fetal growth, and many are expressed both in the placenta and in fetal tissues. It is important to identify novel imprinted genes and study their function, particularly in the placenta where these genes play important roles in supplying maternal nutrients to the growing fetus. The placenta is a preferential organ target for imprinted action, and the function of imprinting in this organ should be fully exploited. Study of the epigenetic mechanisms that regulate imprinted gene expression (e.g., DNA methylation, histone modifications, and nucleosomal organization) is fundamental to understand the link between the genetic program and epigenetic modifications that control growth.
The placenta is highly "plastic" to genetic or epigenetic insult, as compared to the fetus. Indeed, placentas show viability in certain cases of confined placental mosaicism and can withstand epigenetic deregulation during somatic cloning. Albeit more plastic, placentas may be prone to epigenetic errors (epimutations), which may alter their capacity to transfer nutrients to the growing fetus. Most IUGR cases in humans are idiopathic. It is conceivable that a percentage of these cases may result from placental dysfunction due to epimutations. Epimutations may affect particular loci that are regulated by epigenetic mechanisms (e.g., imprinted genes) or alternatively be a genome wide event. High-throughput analysis of epigenetic modifications (e.g., methylation of DNA and histone modifications) could be investigated in placental DNA from IUGR cases.
Several cases of early pregnancy losses have a normal chromosomal constitution and a non-discernable genetic/environmental cause. The possibility that genome-wide epigenetic errors may underlie the pathogenesis of these cases has not been explored. High-throughput analysis of epigenetic modifications (e.g., methylation) could be investigated in those particular cases.
The mammalian fetus relies exclusively on its mother for its supply of nutrients, provided through the placenta. To ensure normal fetal growth, the supply of nutrients across the placenta must meet the requirements (the demand) made by the developing fetus. The identification of signalling interactions from fetus to placenta (and vice versa) is central to understanding of the balance between nutrient supply and demand. The study of mouse models for IUGR that mimic human IUGR are invaluable tools to identify supply and demand signals in response to in utero adaptations. The identification of such signalling molecules can be obtained by micro-array expression profiling and proteomics in fetal tissues (the demand center), in placenta (the supply center) and cord blood. The investigation of epigenetic profiles in these cell types may also shed light between the interaction of the genetic and epigenetic programs that result from in utero adaptations.
Adaptations in utero to adverse environmental conditions (e.g., malnutrition, toxic agents, etc.) may have profound consequences in later life, such as increased risk of chronic diseases. The environmental signals may affect the epigenetic program and alter the pattern of expression of genes relating to the metabolic pathways of these diseases. An analysis of the epigenetic profile in diseased tissues either in humans or mice models may provide clues of the involvement of epigenetics in fetal programming of adult diseases.
Dr Jackson indicated that he did not have much to add to what was already said at the meeting. He noted that he was intrigued by the molecular work and appreciated Dr. Beaudet's take on things. He added that the examination of ICSI as suggested sounded worthwhile. However, Dr. Jackson noted that, like the CPM, the effect of ICSI requires more targeted data collection as an initial step. To that end, he and Dr. Wolstenholme were in correspondence and were working toward a letter of intent and a proposal to gather available data from the United States, parts of Europe, the United Kingdom, and Australia to have an ongoing and accessible CPM database. He indicated that he had developed a web-accessible interactive database function for prenatal diagnosis and clinical genetic information, which contain the rudiments of the needed database for the CPM and associated stories. Expanding this database to gather the appropriate information would be relatively simple. Convincing contributors and maintaining the reliability and quality of the data was not simple, but was doable. A reasonable resource would, he believed, generate useful projects. He also acknowledged Dr. Kalousek's work and graphics, which he used in his presentation.
Although the link between CPM and IUGR is now well established, knowledge is largely restricted to the effects of CPM resulting from corrections of trisomies 2, 7, and 16 of meiotic origin. For these three subgroups of CPM, a strong correlation exists between placental trisomy and significant fetal growth restriction. In some cases of trisomy 16 CPM in particular, this correlation can be the most severe form of IUGR and onset may be sufficiently early to result in a non-viable pregnancy. The effects of CPM for other trisomies are much less well investigated. Some abnormalities, e.g., trisomy 6 CPM, are quite rare, so accumulated experience is limited. Others, e.g., trisomy 18 CPM, are relatively common, and although clearly the majority of cases do not suffer from severe IUGR, we actually do not have very good information on birthweights of an unselected series. It is quite possible that subsets of cases do have such problems, or that less dramatic alterations of birthweight or other adverse outcomes may be occurring.
There is only patchy information on the origins of most CPM and much of this is probably unpublished. A particular problem is that investigations of origin on mosaicism are often only undertaken when IUGR or other adverse outcome is apparent. Correction of trisomy without adverse outcome is thus likely to be highly under-reported. As most experience with CPM comes from prenatal detection during CVS, for the majority of pregnancies, where CVS does not form part of antenatal care, CPM is never detected. Furthermore, unless placental investigations are undertaken during pregnancy or in the immediate postnatal period when the placenta is still available, the link between CPM and IUGR is currently, essentially undetectable. Only children in whom UPD is present would give an indication that CPM may have been an issue, but this fact is of limited practical value as karyotype-wide screening for UPD is expensive and time-consuming. Postnatal detection of UPD usually only happens fortuitously, or where a syndrome or some other prior indication of particular chromosome involvement is present.
The major source of quality data about CPM, UPD, and IUGR is scattered across the world in hundreds of diagnostic cytogenetic laboratories performing routine cytogenetic analysis of CVS, none of which have the time or resources to put it all together. Series of cases in the United States, the United Kingdom, and the rest of Europe were collected about 10 years ago, but these were primarily aimed at collecting data on the reliability of CVS in terms of feto-placental discordance. Only the later cases in these series have much information on birthweight and UPD; for the European study, much of this data remains unpublished and is currently inaccessible directly. A prospective data collection, in conjunction with centralization of existing published and unpublished series of cases, would provide a powerful resource for both research and diagnostic purposes. Individual laboratories would probably be more likely to contribute data if, in return, the combined database is relatively freely accessible and in a searchable form, for use for diagnostic purposes. The full extent of unpublished series of well-studied cases that may be available is unclear, but one series from 45,000 CVS procedures is in the United Kingdom, held on a database that may be a suitable prototype for a larger study. Further thoughts in this area may form part of an application for funding.
There is an impressive lack of information about the long-term effects of CPM, with and without UPD and/or IUGR. CPM-related IUGR may not follow exactly the same pattern of disease susceptibility seen in later life in IUGR cases as a whole, and, indeed, each chromosome may well have specific features superimposed on a more common background picture. It may be possible to identify sufficiently large cohorts of existing cases to set up retrospective collaborative studies. If not, prospective studies or a combination of the two approaches will need to be considered. Studies should concentrate on trisomies 2, 7, and 16 initially, moving out to other abnormalities as knowledge increases. A subset of CPM cases is undoubtedly at risk of having gonadal mosaicism. This outcome will have reproductive implications for the children concerned, the oldest of whom are now in their late teens. It is likely that neither these children, or their parents, will be aware of these risks, as they were not appreciated at the time when counselling took place.
Major changes in placental pathology and maternal serum biochemistry have been reported in conjunction with the more severe forms of CPM-related IUGR, but in general this whole area has been lightly studied. Further work in this area should prove instructive and may result in improved protocols for the identification and subsequent handling of CPM/IUGR pregnancies. As indicated above, most are currently never recognized.
The risk of imprinting errors in IVF pregnancies, particularly ICSI, needs full assessment. The rapid expansion of the use of assisted reproduction technology without proper studies can, at best, be described as showing a lack of caution. At worst, it could be described as reckless.
CPM is just one component of the wider problem of chromosome abnormality, so virtually all aspects of data collection of incidences of abnormality, such as that derived from pre-implantation genetic diagnosis research, will help to set it in context. Funding this sort of work is probably beyond the scope of this current research initiative, but information about CPM may be an attractive "spin-off" from projects funded from other budgets, and its value should be recognized.
Areas of continuing research could include the following:
The role of imprinting in fetal growth, both fetal growth restriction and overgrowth
Assisted reproduction and congenital malformations/inherited diseases
Either of these areas of research is well suited to a collaborative mult-site effort to optimize clinical population, clinical genetic expertise, and gene analysis. As a geneticist and perinatologist, Dr. Wilkins-Haug noted that examination of imprinting and its role in fetal growth is crucial.
Dr. Beaudet indicated that he was very concerned by recent reports that ICSI may cause imprinting defects. Given the rarity of these defects and the fact that the molecular basis for these cases was of a very specific and rare type, he felt that it is was quite likely that ICSI could cause imprinting defects. Although some initial reports were somewhat reassuring (PMID 11044469), more recent reports were very worrisome. These reports include PMID 12016519, related to Angelman syndrome; PMID 12525545, related to Beckwith-Wiedemann syndrome; PMID 12549484, regarding Angelman syndrome; and PMID 12559867, regarding retinoblastoma. The use of ICSI for convenience has grown rather rapidly in recent years as far as far as he understands. He felt that this growth seems like a particularly bad idea. If ICSI is the only option for fertilization, he felt it may be justified. But, he added, it seems difficult to justify using ICSI for convenience or somewhat greater efficiency when there was worrisome evidence that the procedure may cause imprinting defects. AS is a very severe disorder, and the life-long care and treatment for such a patient is extremely burdensome.
He noted that his presentation focused primarily on autism, and that he and his colleagues were intensively studying the possibility that imprinting defects may contribute to the occurrence of autism.
Dr. Beaudet indicated that science was just beginning to uncover the importance of epigenetic changes in disease processes. For genomic imprinting in particular, the time of gametogenesis and the time around fertilization are particularly important. He mentioned that he had published a commentary regarding epigenetics and disease (PMID 12394355). One review (PMID 11498579) is particularly relevant as regards imprinting around the time of fertilization; he felt the participants may find it helpful.