- Email: [email protected]
Imprinting makes an impression
413
EDITORIALS
Imprinting makes an impression A compelling body of evidence from widely diverse areas of research suggests that expression of some
depends on their parental origin-ie, on whether they have been inherited from the mother or from the father.’-3 The importance of the concept known as genomic imprinting is now beginning to be appreciated by physicians in relation to human development and disease. Developmental geneticists have recognised for some time that the nuclear genetic contribution of mammalian mothers differed from, but was complementary to, that of fathers. Pronuclear transplantation (in which zygotes are constructed with only paternally derived chromosomes) and parthenogenic work (in which both haploid sets of4 chromosomes have come only from a female) in mice,4 and examination of the origins of complete hydatidiform moles2 and triploidss in human beings, suggest that the paternal contribution to the placenta is unique and essential. However, none of these results seemed especially remarkable until it was shown that parental origin of genetic material contributed to chromosome deletion syndromes6 and many human cancers.7 The discovery that Prader-Willi and Angelman syndromes are often associated with interstitial deletions of the long arm of chromosome 15deletions that seem to lie in exactly the same region6,8-led to consternation and disbelief. How could this be? Can two completely different clinical conditions really result from the same chromosomal deletion? Use of molecular markers allowed the parental origin of the deleted chromosomes to be identified and provided the answer. In Prader-Willi syndrome the deletion is of the paternally derived chromosome 15 whereas in Angelman syndrome the maternal chromosome is deleted. This observation has so surprised cytogeneticists that they have begun to examine parental derivation in other chromosomal abnormalities, and to wonder about the parental origin of chromosome segments that are involved in producing the clinical features of various syndromes. Additional support for the importance of parental origin comes from other cases of Prader-Willi and Angelman syndromes in which no cytogenetically visible deletion has been found. In many of these cases uniparental disomy (ie, two chromosomes 15 from the same parent) has been observed instead-two copies of mother’s chromosome 15 in Prader-Willi;9 two genes
copies of father’s in Angelman. 10.11 Thus it seems to be the lack of that part of chromosome 15 coming from father or mother which produces Prader-Willi or Angelman clinical features, respectively. Both uniparental isodisomy (two identical copies of the same chromosome from one parent) and uniparental heterodisomy (two different copies of the particular chromosome from one parent) have been observed. Heterodisomy is especially remarkable since the two chromosomes 15 function normally in the parent from whom they were derived, but when both are passed on to the zygote they fail to provide a factor essential for embryonic wellbeing and development. Both parental contributions are therefore necessary and complementary for normal growth and development. Some of the earliest imprinting work on uniparental disomy in mice suggested that factors produced by the genes inherited from one parent enhanced growth and that these factors might balance growth-suppressing gene products inherited from the other parent.12 Results of chimeric work in mice, in which pluripotent stem cells containing chromosomes from only one inserted into a zygote and become incorporated into the tissues of the developing fetus, also accord with the view that there may be a tissue-specific interaction between paternal and maternal factors.13 In mice, only the paternally inherited gene is expressed for insulin-like growth factor II14 and only the maternally derived gene for insulin-like growth factor II receptor. is This observation further supports the concept of checks and balances between the parental contributions.16 The relevance to human disease comes from the observation that many cancers are associated with loss of a specific chromosome derived from a particular parent, usually the mother. These chomosomes presumably contain tumour suppressor genes, and loss of the expressing (ie, maternal) copy results in loss of growth suppression and subsequently in development of a tumour.7 In several sporadic cases of the rare overgrowth syndrome known as Wiedeman-Beckwith syndrome, cells carry two paternally derived chromosomes 11 (paternal uniparental disomy)1’-yet more evidence of a role for imprinting in growth disorders, as well as in human cancer, since patients with this condition frequently get several types of cancer. Basic biology teaching is that we receive one chromosome of each pair from the mother and one from the father-so how could uniparental disomy come about? Work on human reproduction has shown that probably a third of all pregnancies abort, and that half of all early spontaneous abortuses carry chromosomal anomalies, mainly trisomies. Most trisomies are lethal; it is only when a disomic cell arises through loss of one of the extra chromosomes that a viable cell line will be produced. Loss of the extra chromosome will result in uniparental disomy (two chromosomes from the same parent) on one in three
parent
are
414
occasions.l Chorionic villus sampling has shown placental mosaicism in 2-3% of all pregnancies studied prenatally at 9-11 weeks.19 This finding suggests that at least 3% of pregnancies which survive to 9-11 weeks began as a trisomies and lost the extra chromosome, thereby allowing survival of the pregnancy and producing uniparental disomy in up to 1 % of the population. (This assumes, of course, that uniparental disomies of all chromosomes are equally
viable.) An interesting corollary
of uniparental disomy is that if the chromosomes from the one parent are identical copies (isodisomy), and that chromosome unfortunately carries an abnormal recessive gene, the individual with uniparental disomy for the chromosome now carries two abnormal copies of that gene and will get an autosomal recessive disease that has been inherited from only one carrier parent. This sequence of events has been shown to occur in cystic fibrosis.2O,21 The percentage of cases of autosomal recessive disorders that are uniparental rather than familial is unknown, but the fact that it occurs at all is
startling. The
imprinting implies that genetic material (chromosome regions or individual genes) is "marked" or designated in some way so as to produce differences in expression, and therefore differences in the clinical features or phenotype, depending upon whether it is inherited from mother or father. Marking is most likely to occur during meiosis (germ-line production),22 when the maternally derived and paternally derived chromosomes come together and pair and then go on to produce sperm or ova. During term
this time the genes that a female inherited from her father must be redesignated as "maternal" so that they will act in a maternal manner as she passes them on to her offspring. Similarly, the genes that a male inherited from his mother must be changed to a "paternal" designation. Several lines of evidence suggest that DNA methylation is involved in
imprinting processes, and the presence of methylation off or keep turned off genes so that they are expressed .13-14 If imprinting or part of the imprinting process occurs during meiosis there may be transgenerational effects, since the meiosis producing an egg which will eventually become a new individual occurs during the in-utero life of the mother (ie, during her own embryonic development). Consequently, expression of imprinting in an individual could be modified by the grandmother’s exposures or illnesses during her pregnancy with the mother of that individual. Now that molecular markers are available, parental origin can be easily traced, and differences dependent on the parent of origin are being increasingly recognised. Disorders whose transmission and inheritance have been poorly understood must now be examined for parent-of-origin differences and possible imprinting effects. It is no longer sufficient simply to report a chromosomal anomaly, since the may not
turn
identification of the parent from which it was inherited may result in a radically different prediction of outcome. Not only academic researchers but also
general practitioners and members of every medical specialty will probably be affected by these new concepts. They force a critical re-examination of the traditional dogma about patterns of inheritance, and may provide an explanation and accurate prediction of outcome in many disorders that were previously dismissed with vague terms such as "non-penetrant" or "multifactorial", or designated as examples of "variable expressivity". The concept of genomic imprinting is undeniably an important addition to classic genetic theory and its clinical applications. 1. Monk M, Surani A, eds. Genomic imprinting. Cambridge: Company of Biologists, 1990. 2. Hall JG. Genomic imprinting: review and relevance to human disease. Am J Hum Genet 1990; 46: 857-73. 3. Hall JG. Genomic imprinting. Curr Opin Genet Devel 1991; 1: 34-39. 4. Suram MAH. Evidence and consequences of differences between maternal and paternal genomes during embryogenesis in the mouse. In:
Rossant J, Pederson RA, eds.
Experimental approaches to mammalian embryonic development. Cambridge: Cambridge University Press, 1986: 401-35. 5. MacFadden DE, Kalousek DK. Two different phenotypes of fetuses with chromosomal triploidy: correlation with parental origin of the extra haploid set. Am J Med Genet 1991; 38: 535- 58. 6. Magenis RE, Toth-Fejel S, Allen LJ, et al. Comparison of the 15q deletion in Prader-Willi and Angelman syndromes: specific regions, extent of deletions, parental origin and clinical consequences. Am J Med Genet 1990; 35: 333-49. 7. Ponder B. Is imprinting to blame? Nature 1989; 340: 264. 8. Knoll JHM, Nicholls RD, Magenis RE, et al. Angelman and PraderWilli syndromes share a common chromosome 15 deletion but differ in parental origin of deletion. Am J Med Genet 1989; 32: 285-90. 9. Nicholls RD, Knoll JHM, Butler MG, et al. Genetic imprinting suggested by maternal heterodisomy in non-deletion Prader-Willi syndrome. Nature 1989; 349: 281-85. 10. Knoll JHM, Nicholls RD, Magenis RE, et al. Angelman syndrome: three molecular classes identified with chromosome 15q1 1q13-specific DNA markers. Am J Hum Genet 1991; 47: 149-55. 11. Malcolms S, Clayton-Smith J, Nichols M, et al. Uniparental paternal disomy in Angelman syndrome. Lancet 1991; i: 694-97. 12. Cattenach BM, Kirk M. Differential activation of maternally and paternally derived chromosome regions in mice. Nature 1985; 315: 496-98. 13. Ferguson-Smith AC, Cattanach BM, Barton SC, et al. Embryological and molecular investigations of parental imprinting on mouse chromosome 7. Nature 1991; 351: 667-70. 14. DeChiara TM, Robertson EJ, Efstratiadis A. Parental imprinting of the mouse insulin-like growth factor II gene. Cell 1991; 64: 849-59. 15. Barlow DP, Stoger R, Herrmann BG, et al. The mouse insulin-like growth factor type-2 receptor is imprinted and closely linked to the Tme locus. Nature 1991; 349: 84-87. 16. Moore T, Haig D. Genomic imprinting in mammalian development: a parental tug of war. Trends Genet 1991; 7: 45-49. 17. Henry I, Bonaiti-Pellie C, Chehensse V, et al. Uniparental disomy in a genetic cancer-predisposing syndrome. Nature 1991; 351: 665-67. 18. Hall JG. How imprinting is relevant to human disease. In: Monk M, Surani A, eds. Genomic imprinting. Cambridge: Company of Biologists, 1990: 141-48. 19. Kalousek DK. The role of confined chromosomal mosaicism in placental function and human development. Growth Genet Horm 1988; 4: 1-3. 20. Spence JE, Perciaccante RG, Greig GM, et al. Uniparental disomy as a mechanism for human genetic disease. Am J Hum Genet 1988; 42: 217-26. 21. Voss R, Ben-Simon E, Avital A, et al. Isodisomy of chromosome 7 in a patient with CF: could uniparental disomy be common in humans? Am J Hum Genet 1989; 45: 373-80. 22. Hulten MA, Hall JG. On mechanisms of genomic imprinting. Chrom Today 1990, 10: 157-62. 23. Reik W, Collick A, Norris ML, et al. Genomic imprinting determines methylation of parental alleles in transgenic mice. Nature 1987; 328: 248-51. 24. Surani MA, Reik W, Allen ND. Transgenes as molecular probes for genomic imprinting. Trends Genet 1988; 4: 59-62.
EDITORIALS
Imprinting makes an impression A compelling body of evidence from widely diverse areas of research suggests that expression of some
depends on their parental origin-ie, on whether they have been inherited from the mother or from the father.’-3 The importance of the concept known as genomic imprinting is now beginning to be appreciated by physicians in relation to human development and disease. Developmental geneticists have recognised for some time that the nuclear genetic contribution of mammalian mothers differed from, but was complementary to, that of fathers. Pronuclear transplantation (in which zygotes are constructed with only paternally derived chromosomes) and parthenogenic work (in which both haploid sets of4 chromosomes have come only from a female) in mice,4 and examination of the origins of complete hydatidiform moles2 and triploidss in human beings, suggest that the paternal contribution to the placenta is unique and essential. However, none of these results seemed especially remarkable until it was shown that parental origin of genetic material contributed to chromosome deletion syndromes6 and many human cancers.7 The discovery that Prader-Willi and Angelman syndromes are often associated with interstitial deletions of the long arm of chromosome 15deletions that seem to lie in exactly the same region6,8-led to consternation and disbelief. How could this be? Can two completely different clinical conditions really result from the same chromosomal deletion? Use of molecular markers allowed the parental origin of the deleted chromosomes to be identified and provided the answer. In Prader-Willi syndrome the deletion is of the paternally derived chromosome 15 whereas in Angelman syndrome the maternal chromosome is deleted. This observation has so surprised cytogeneticists that they have begun to examine parental derivation in other chromosomal abnormalities, and to wonder about the parental origin of chromosome segments that are involved in producing the clinical features of various syndromes. Additional support for the importance of parental origin comes from other cases of Prader-Willi and Angelman syndromes in which no cytogenetically visible deletion has been found. In many of these cases uniparental disomy (ie, two chromosomes 15 from the same parent) has been observed instead-two copies of mother’s chromosome 15 in Prader-Willi;9 two genes
copies of father’s in Angelman. 10.11 Thus it seems to be the lack of that part of chromosome 15 coming from father or mother which produces Prader-Willi or Angelman clinical features, respectively. Both uniparental isodisomy (two identical copies of the same chromosome from one parent) and uniparental heterodisomy (two different copies of the particular chromosome from one parent) have been observed. Heterodisomy is especially remarkable since the two chromosomes 15 function normally in the parent from whom they were derived, but when both are passed on to the zygote they fail to provide a factor essential for embryonic wellbeing and development. Both parental contributions are therefore necessary and complementary for normal growth and development. Some of the earliest imprinting work on uniparental disomy in mice suggested that factors produced by the genes inherited from one parent enhanced growth and that these factors might balance growth-suppressing gene products inherited from the other parent.12 Results of chimeric work in mice, in which pluripotent stem cells containing chromosomes from only one inserted into a zygote and become incorporated into the tissues of the developing fetus, also accord with the view that there may be a tissue-specific interaction between paternal and maternal factors.13 In mice, only the paternally inherited gene is expressed for insulin-like growth factor II14 and only the maternally derived gene for insulin-like growth factor II receptor. is This observation further supports the concept of checks and balances between the parental contributions.16 The relevance to human disease comes from the observation that many cancers are associated with loss of a specific chromosome derived from a particular parent, usually the mother. These chomosomes presumably contain tumour suppressor genes, and loss of the expressing (ie, maternal) copy results in loss of growth suppression and subsequently in development of a tumour.7 In several sporadic cases of the rare overgrowth syndrome known as Wiedeman-Beckwith syndrome, cells carry two paternally derived chromosomes 11 (paternal uniparental disomy)1’-yet more evidence of a role for imprinting in growth disorders, as well as in human cancer, since patients with this condition frequently get several types of cancer. Basic biology teaching is that we receive one chromosome of each pair from the mother and one from the father-so how could uniparental disomy come about? Work on human reproduction has shown that probably a third of all pregnancies abort, and that half of all early spontaneous abortuses carry chromosomal anomalies, mainly trisomies. Most trisomies are lethal; it is only when a disomic cell arises through loss of one of the extra chromosomes that a viable cell line will be produced. Loss of the extra chromosome will result in uniparental disomy (two chromosomes from the same parent) on one in three
parent
are
414
occasions.l Chorionic villus sampling has shown placental mosaicism in 2-3% of all pregnancies studied prenatally at 9-11 weeks.19 This finding suggests that at least 3% of pregnancies which survive to 9-11 weeks began as a trisomies and lost the extra chromosome, thereby allowing survival of the pregnancy and producing uniparental disomy in up to 1 % of the population. (This assumes, of course, that uniparental disomies of all chromosomes are equally
viable.) An interesting corollary
of uniparental disomy is that if the chromosomes from the one parent are identical copies (isodisomy), and that chromosome unfortunately carries an abnormal recessive gene, the individual with uniparental disomy for the chromosome now carries two abnormal copies of that gene and will get an autosomal recessive disease that has been inherited from only one carrier parent. This sequence of events has been shown to occur in cystic fibrosis.2O,21 The percentage of cases of autosomal recessive disorders that are uniparental rather than familial is unknown, but the fact that it occurs at all is
startling. The
imprinting implies that genetic material (chromosome regions or individual genes) is "marked" or designated in some way so as to produce differences in expression, and therefore differences in the clinical features or phenotype, depending upon whether it is inherited from mother or father. Marking is most likely to occur during meiosis (germ-line production),22 when the maternally derived and paternally derived chromosomes come together and pair and then go on to produce sperm or ova. During term
this time the genes that a female inherited from her father must be redesignated as "maternal" so that they will act in a maternal manner as she passes them on to her offspring. Similarly, the genes that a male inherited from his mother must be changed to a "paternal" designation. Several lines of evidence suggest that DNA methylation is involved in
imprinting processes, and the presence of methylation off or keep turned off genes so that they are expressed .13-14 If imprinting or part of the imprinting process occurs during meiosis there may be transgenerational effects, since the meiosis producing an egg which will eventually become a new individual occurs during the in-utero life of the mother (ie, during her own embryonic development). Consequently, expression of imprinting in an individual could be modified by the grandmother’s exposures or illnesses during her pregnancy with the mother of that individual. Now that molecular markers are available, parental origin can be easily traced, and differences dependent on the parent of origin are being increasingly recognised. Disorders whose transmission and inheritance have been poorly understood must now be examined for parent-of-origin differences and possible imprinting effects. It is no longer sufficient simply to report a chromosomal anomaly, since the may not
turn
identification of the parent from which it was inherited may result in a radically different prediction of outcome. Not only academic researchers but also
general practitioners and members of every medical specialty will probably be affected by these new concepts. They force a critical re-examination of the traditional dogma about patterns of inheritance, and may provide an explanation and accurate prediction of outcome in many disorders that were previously dismissed with vague terms such as "non-penetrant" or "multifactorial", or designated as examples of "variable expressivity". The concept of genomic imprinting is undeniably an important addition to classic genetic theory and its clinical applications. 1. Monk M, Surani A, eds. Genomic imprinting. Cambridge: Company of Biologists, 1990. 2. Hall JG. Genomic imprinting: review and relevance to human disease. Am J Hum Genet 1990; 46: 857-73. 3. Hall JG. Genomic imprinting. Curr Opin Genet Devel 1991; 1: 34-39. 4. Suram MAH. Evidence and consequences of differences between maternal and paternal genomes during embryogenesis in the mouse. In:
Rossant J, Pederson RA, eds.
Experimental approaches to mammalian embryonic development. Cambridge: Cambridge University Press, 1986: 401-35. 5. MacFadden DE, Kalousek DK. Two different phenotypes of fetuses with chromosomal triploidy: correlation with parental origin of the extra haploid set. Am J Med Genet 1991; 38: 535- 58. 6. Magenis RE, Toth-Fejel S, Allen LJ, et al. Comparison of the 15q deletion in Prader-Willi and Angelman syndromes: specific regions, extent of deletions, parental origin and clinical consequences. Am J Med Genet 1990; 35: 333-49. 7. Ponder B. Is imprinting to blame? Nature 1989; 340: 264. 8. Knoll JHM, Nicholls RD, Magenis RE, et al. Angelman and PraderWilli syndromes share a common chromosome 15 deletion but differ in parental origin of deletion. Am J Med Genet 1989; 32: 285-90. 9. Nicholls RD, Knoll JHM, Butler MG, et al. Genetic imprinting suggested by maternal heterodisomy in non-deletion Prader-Willi syndrome. Nature 1989; 349: 281-85. 10. Knoll JHM, Nicholls RD, Magenis RE, et al. Angelman syndrome: three molecular classes identified with chromosome 15q1 1q13-specific DNA markers. Am J Hum Genet 1991; 47: 149-55. 11. Malcolms S, Clayton-Smith J, Nichols M, et al. Uniparental paternal disomy in Angelman syndrome. Lancet 1991; i: 694-97. 12. Cattenach BM, Kirk M. Differential activation of maternally and paternally derived chromosome regions in mice. Nature 1985; 315: 496-98. 13. Ferguson-Smith AC, Cattanach BM, Barton SC, et al. Embryological and molecular investigations of parental imprinting on mouse chromosome 7. Nature 1991; 351: 667-70. 14. DeChiara TM, Robertson EJ, Efstratiadis A. Parental imprinting of the mouse insulin-like growth factor II gene. Cell 1991; 64: 849-59. 15. Barlow DP, Stoger R, Herrmann BG, et al. The mouse insulin-like growth factor type-2 receptor is imprinted and closely linked to the Tme locus. Nature 1991; 349: 84-87. 16. Moore T, Haig D. Genomic imprinting in mammalian development: a parental tug of war. Trends Genet 1991; 7: 45-49. 17. Henry I, Bonaiti-Pellie C, Chehensse V, et al. Uniparental disomy in a genetic cancer-predisposing syndrome. Nature 1991; 351: 665-67. 18. Hall JG. How imprinting is relevant to human disease. In: Monk M, Surani A, eds. Genomic imprinting. Cambridge: Company of Biologists, 1990: 141-48. 19. Kalousek DK. The role of confined chromosomal mosaicism in placental function and human development. Growth Genet Horm 1988; 4: 1-3. 20. Spence JE, Perciaccante RG, Greig GM, et al. Uniparental disomy as a mechanism for human genetic disease. Am J Hum Genet 1988; 42: 217-26. 21. Voss R, Ben-Simon E, Avital A, et al. Isodisomy of chromosome 7 in a patient with CF: could uniparental disomy be common in humans? Am J Hum Genet 1989; 45: 373-80. 22. Hulten MA, Hall JG. On mechanisms of genomic imprinting. Chrom Today 1990, 10: 157-62. 23. Reik W, Collick A, Norris ML, et al. Genomic imprinting determines methylation of parental alleles in transgenic mice. Nature 1987; 328: 248-51. 24. Surani MA, Reik W, Allen ND. Transgenes as molecular probes for genomic imprinting. Trends Genet 1988; 4: 59-62.