In the mid-nineteenth century, Gregor Mendel discovered that when round garden peas were crossbred with wrinkled garden peas, all of the offspring were round, regardless of the parental origin of each trait in the cross. The resultant principle of equivalence – that genes are expressed equally, no matter what the parental origin is – and its corollary, the biparental inheritance of autosomal genes, have become a central dogma of genetics. However, in recent decades, researchers have discovered several phenomena that challenge the conventional Mendelian notions of equal, biparental inheritance. Genomic imprinting, the unequal expression of alleles depending on the parent of origin, is perhaps the most clinically significant exception to Mendel’s laws of inheritance. The clinical consequences of genomic imprinting may be unmasked when a pair of homologous chromosomes are abnormally inherited from a single parent. This situation is termed uniparental disomy (UPD).
Eric Engel first suggested UPD as a mechanism for human genetic disease in 1980 on the basis of observations made by Searle and others (Searle et al., 1971; Lyon et al., 1975) that mice with translocations were susceptible to nondisjunction (see Article 16, Nondisjunction, Volume 1), in which the homologous chromosomes comprising the translocation would malsegregate during gametogenesis (see Article 13, Meiosis and meiotic errors, Volume 1). Engel hypothesized that by mating these translocation carriers, a subset of offspring would receive a nullisomic gamete from one parent and a disomic gamete from the other parent, resulting in a chromosomally balanced individual with both chromosome homologs coming from one parent (Engel, 1980). Further research in the mouse by Cattanach and others (1985) provided the first evidence of the clinical effects of mammalian uniparental disomy; however, the existence of UPD in humans did not come until several years later, when the development of DNA-based polymorphic markers allowed the parental origin of chromosome homologs to be determined (Spence et al., 1988).
The first documented case of human UPD arose from the investigation of a child with cystic fibrosis and short stature (Spence et al., 1988). Marker analysis revealed that the child inherited both chromosomes 7 from her mother. The cystic fibrosis was likely the result of the presence of a recessive disease allele on both copies of identical homologs of chromosome 7 inherited by the child. Spence et al. (1988) proposed three mechanisms in addition to the gametic complementation theory hypothesized by Engel (1980), including postfertilization error and the “rescue” of a conception either by the loss of an extra chromosome in a trisomy or the duplication of a single chromosome in a monosomy. The UPD can contain one copy of each of the contributing parent’s homologs (heterodisomy) or two copies of one of the parent’s chromosomes (isodisomy). Recessive disease alleles, such as that for cystic fibrosis, are exposed through isodisomies, which often result from a monosomy rescue or postfertilization recombination (Spence et al., 1988). UPD for nearly every chromosome has been documented as a result of the investigation of the abnormal inheritance of recessive disorders, including spinal muscular atrophy type III, osteogenesis imperfecta, and Bloom syndrome (reviewed in Cassidy, 1995; Shaffer, 2003).
Because of the high lethality of monosomic embryos and the requirement of a duplication event early in development, cases of UPD arising from a monosomy rescue are far rarer than those UPDs that arise after resolution of a trisomy. Mosaicism (see Article 18, Mosaicism, Volume 1) may result because of the presence of a mixture of trisomic cells and cells that have lost the extra chromosome copy. The trisomic cells in some conceptuses may be restricted only to the placenta, whereas the fetal cells contain normal chromosomal complements (termed confined placental mosaicism). Several UPD cases have been reported following a discrepancy between karyotyping performed on chorionic villus samples (CVS) from the placenta and amniotic fluid samples in which cells are derived from the fetus (Kalousek et al., 1993; Jones et al., 1995; Ledbetter and Engel, 1995).
Several structural chromosome abnormalities may prompt molecular investigations that may lead to the identification of UPDs, including Robertsonian transloca-tions, nonacrocentric isochromosomes, marker chromosomes, derivative chromosomes, and reciprocal translocations (Shaffer et al., 2001a). Of these, Robertsonian translocations are the most likely to be involved in cases of UPD, probably due to their relatively high incidence in the human population and the increased risk of malsegregation during gametogenesis, which would result in aneuploid gametes (Berend et al., 2000). Resolution of trisomies may result in UPD. Malsegregation of structural abnormalities may result in gametic complementation and UPD; however, because gametic complementation requires two independent nondisjunction events, one in each parent, UPDs resulting from this mechanism are relatively rare.
Of the four mechanisms that may result in UPD, all but postfertilization error (somatic recombination) result in whole-chromosome uniparental disomies. Mitotic crossing-over may result in partial UPDs; sporadic cases of Beckwith-Wiedemann syndrome (BWS) (see Article 30, Beckwith-Wiedemann syndrome, Volume 1) are often caused by a partial paternal disomy of the distal short arm of chromosome 11 (Henry et al., 1991; Bischoff et al., 1995).
Studies of UPD in mouse and human have highlighted the differential maternal and paternal genetic contribution to development. Androgenotes, mouse embryos that contain only paternally derived chromosomes, exhibit poor growth while the extra-embryonic tissues develop relatively normally; in contrast, the embryonic tissues of gynogenotes, which contain two copies of the maternal chromosomes, develop well, but the extra-embryonic structures fail to develop, resulting in embryonic death shortly after implantation (Surani et al., 1984; see also Article 28, Imprinting and epigenetics in mouse models and embryogenesis: understanding the requirement for both parental genomes, Volume 1). This phenomenon is the result of genomic imprinting and the developmental genetic need for contributions from both parents.
In humans, abnormal pregnancies in which the placenta appears as a cyst known as a hydatidiform mole demonstrate the same abnormal development: the moles, which contain all paternal chromosomes, consist entirely of placental tissue. The maternal counterpart, teratomas, which contain two copies of the maternal genome, contain only embryonic tissues (Jacobs et al., 1982). These examples suggest that, rather than the maternal and paternal genomes being equivalent, certain regions of the genome are not expressed equally from the maternal and paternal contributions. Because the differential contributions are not marked by changes to the DNA sequence, epigenetic modifications (chemical changes to the DNA or the surrounding proteins) must result in differential gene expression, depending on whether the genes are inherited from the mother or father.
UPDs may unmask additional phenotypic effects beyond those attributed to recessive disease through the duplication and subsequent overexpression, or deletion and lack of expression, of imprinted genes (see Article 26, Imprinting and epigenetic inheritance in human disease, Volume 1). For example, 25-30% of patients with Prader-Willi syndrome (PWS) (see Article 29, Imprinting in Prader-Willi and Angelman syndromes, Volume 1) have maternal UPD for maternal chromosome 15. The other 70% have deletions of a small part of paternal chromosome 15, indicating that the loss of a paternally expressed gene results in the syndrome. Conversely, maternal deletion of 15q in ~70% of cases or paternal disomy for chromosome 15 in ~5% of cases results in a clinically distinct disorder, Angelman syndrome (AS; Nicholls et al., 1989). The disparity between frequencies of UPD in the two syndromes is likely a result of higher rates of female nondisjunction than male nondisjunction (Abruzzo and Hassold, 1995). However, most cases of AS caused by UPD are isodisomy, likely due to a rescue of a monosomic conceptus, whereas maternally derived UPD cases in PWS are usually heterodisomic, due to a rescue of a trisomic conceptus. Because monosomies are less viable than trisomies and tend to result in miscarriage sooner, there is a smaller “window of opportunity” in which the monosomic conceptus can be rescued, likely resulting in fewer cases of isodisomy (Shaffer, 2003). In addition to maternal and paternal disomies 15, several other chromosomes show clinically distinct phenotypes due to imprinting, including paternal disomy 6, maternal disomy 7, partial paternal disomy 11, and maternal and paternal disomies 14 (Ledbetter and Engel, 1995; Shaffer et al., 2001b; Kotzot, 1999). Phenotypes suggestive of imprinting effects are also found in maternal disomy 2, maternal disomy 6, paternal disomy 9, maternal disomy 16, paternal disomy 16, and maternal disomy 20; however, either because of insufficient cases reported, conflicting reports in the literature, or the confounding effects of mosaicism, the imprinting effects for these chromosomes remain uncertain (Shaffer et al., 2001b; Shaffer, 2003).
Imprinting disorders often involve cognitive and growth problems. Patients with PWS, for example, have short stature and are frequently obese (Cassidy, 1984). In contrast, the major clinical features of BWS include overgrowth (Henry et al., 1991). The differential growth effects of some imprinted syndromes – particularly PWS, AS, BWS, and Silver-Russell syndrome (SRS; maternal UPD7 and small stature) – may indicate a conflict between the paternal and maternal genomes over the amount of resources demanded of the mother by her offspring; paternally expressed genes would encourage embryonic growth at the expense of the mother, whereas maternally expressed genes would encourage decreased offspring size (Moore and Haig, 1991). However, contradictory evidence exists for a number of chromosomes, although this may simply indicate that the differential maternal and paternal contributions to development are more complex than what the original models predict (Hurst and McVean, 1997).
Although it is a relatively newly described phenomenon, genomic imprinting has been shown to play a key role in several developmental processes. Disruption of these normal dosage imbalances (Shaffer et al., 2001b) by uniparental disomy has permitted key insights into the possible mechanisms and functions of these epigenetic modifications, the ramifications of which continue to challenge the seemingly unshakable concepts that Mendel observed in peas over 100 years ago.
