Several lines of evidence have accumulated in recent years indicating that nuclear cytoplasmic interactions play an important role in the formation and fate of the developing mouse embryo. Early nuclear transplantation experiments indicated that the ability of nuclei to direct cleavage after transfer into enucleated zygotes falls abruptly with nuclei from more advanced preimplantation stages [1]. Transcriptional activation of the nuclei, which occurs during the second cell cycle probably precludes the reprogramming of nuclei from later cleavage stages [2]. Thus, when an 8-cell nucleus is transferred to an enucleated zygote, such a reconstituted zygote is blocked at the 2-cell stage. However, when identical 8-cell nuclei were transferred into both blastomeres of enucleated 2-cell embryos, they were able to support development to the blastocyst stage and even gave rise to live offspring [2-4]. This indicated the importance of the cytoplasmic environment for the ability of the incoming nucleus to support development. It should be noted that in these experiments, the nuclear cytoplasmic ratio was also an important factor in determining the development of the reconstituted embryos [2]. Similar observations were also made when monitoring the development of haploid embryos [5]. In another study, Latham and Solter [6] examined the ability of androgenones, obtained by replacing the female pronucleus of a zygote by the male pronucleus, to develop to the blastocyst stage. Androgenones generated from C57B1/6 eggs were found to be much more competent to give rise to blastocysts than were DBA/2 androgenones. However, when androgenones were constructed from (DBA/2×C57B1/6)F1, zygotes (genetic constitution of the embryos will hereafter be indicated with the female parent coming first followed by the male parent), by replacing the DBA/2 female pronucleus with a C57B1/6 pronucleus, they also developed poorly. This was not simply due to the lack of some component in DBA/2 cytoplasm, since the impaired development was also observed when C57B1/6 male pronuclei from pairs of (DBA/2×C57B1/6) F1, were transferred to an enucleated C57B1/6 egg.

The unfolding of the genetic story of Prader-Willi and Angelman syndromes provided the first recognition of human genomic imprinting. These disorders, which are clinically very distinct, are related through their genomic proximity and the inverse direction of the imprinting which affects them. Both are interesting disorders in themselves, especially in that both have distinctive behavioral patterns among their clinical features that may teach us much about normal human behavior.

Prader-Willi syndrome (PWS) is a complex multi-system condition whose major features include infantile hypotonia with decreased arousal, poor suck and failure-to-thrive; characteristic dysmorphic facial features; hypopigmentation; childhood onset of obesity due to lack of satiety; hypogonadotropic hypogonadism with genital hypoplasia and delayed and incomplete puberty; short stature for genetic background; developmental delay and usually mild mental retardation; and a characteristic behavioral disturbance with temper tantrums and obsessive-compulsive behavior. PWS occurs in about 1/15,000 people. Since its first description in 1956, it has been apparent that many of these features arise from insufficient function of the hypothalamus, and recent identification of neurosecretory growth hormone insufficiency and temperature and sleep regulation abnormalities support this. However, no visible gross or microscopic abnormalities of the hypothalamus are seen on neuropathology. The finding of a chromosome 15q 11-13 deletion in a proportion of patients with PWS by Ledbetter and colleagues in 1981 was the first window to the exciting genetic discoveries of the past decade, including recognition (initially by Butler and Palmer) that the deletion is always on the paternally-derived chromosome 15 in PWS, and the finding by Nicholls and coworkers that the vast majority of the remainder of patients had normal chromosomes but had maternal uniparental disomy (UPD). Nearly all patients with clinically typical PWS have either 15q deletion (about 75%) or maternal UPD (about 25%). This is the first human disorder that was recognized to result from uniparental disomy, and lead to many insights into imprinting.

Maternal and paternal disomies and equivalent duplications for specific chromosome regions can readily be generated in the mouse. Most do not result in abnormality but for 10 regions distributed over 6 chromosomes developmental anomalies ranging from early embryonic lethalities to characteristic phenotypic abnormalities occur. With certain chromosome regions, maternal or paternal duplication leads to different and, in some cases, opposite anomalies so that, in total, 15 different imprinting effects can be distinguished. These observations have provided key evidence on the occurrence of imprinting in mammals.

On the basis of the established homologies between mouse and human chromosomes, it is possible to predict which segments of the human genome are subject to equivalent imprinting. In this regard it is significant that candidate imprinting effects for the two classical examples of imprinting in humans, namely the Prader-Willi and Angelman syndromes, have been found in the mouse.

Current studies are aimed at reducing the size of the imprinting regions with the objective of facilitating identification of the genes involved. Furthermore, the developmental profiles of genes already identified as being subject to imprinting are being determined.

A new approach involves the analysis of a novel mutation that causes growth retardation and cranial abnormalities and which shows the inheritance pattern of an imprinted gene, as originally predicted by Hall (Am J Hum Genet 46: 857, 1990). It is anticipated that the mutation will represent a deletion and will lie within one of the recognised imprinting regions.

The following paper is concerned with potential changes in the normal epigenetic process in a diploid individual, when a chromosome pair or segment is inherited from one parent only, instead of the expected biparental contribution. This aberrant mode of transmission arises from the high rate of gamete aneuploidy in humans. It has received the name uniparental disomy (UPD), and has emerged as an important factor in the new field of nontraditional inheritance, depicted in Table 1.

The following definitions may foster a better understanding of this discussion.

UPD is the inheritance of both copies of a chromosome [or chromosomal segment(s)] from a single parent, instead of the normal biparental transmission of the pair. In isodisomy, the two uniparental copies are identical, being derived from the same parental chromosome. In heterodisomy, the two uniparental chromosomes are different, being derived from the homologues of a pair.

Genetic studies have shown that both a maternally and paternally inherited copy of mouse chromosome 7 are essential for normal embryogenesis. When the parental dosage is altered, such as in maternal or paternal uniparental disomy for chromosome 7 (UPD7), the resulting embryos die. This is due to the altered dosage of imprinted genes which are normally expressed only from the paternally or maternally inherited chromosome homologue. Several genes on mouse chromosome 7 are subject to parental imprinting. Mutant phenotypes seen in UPD7 embryos and chimaeras can be explained by the altered dosage of some of these genes.

The mechanism(s) that causes genes to be expressed in a parental origin specific manner has not yet been determined but is believed to involve germline specific modifications to DNA and/or chromatin which are acted upon after fertilisation to affect the activity of imprinted genes. Two genes, H19 and Igf2, are located 90kb apart on the distal end of chromosome 7 and are imprinted reciprocally with the maternally inherited allele of HI9 and paternally inherited allele of Igf2 being expressed. We have used UPD7 embryos to identify epigenetic modifications that distinguish the two parental alleles in the H19 and Igf2 domain by comparing DNA and chromatin from normal and maternal UPD cobceptuses. Clear cut differences in DNA methylation and chromatin compaction were observed for the H19 gene with the paternal allele exhibiting promoter methylation and nuclease insensitivity. These were not found in sperm. In addition, no major differences were noted for the Igf2 gene, although subtle parental origin specific modifications were found. These studies suggest that the two genes may share a common regulatory mechanism which controls their reciprocal imprinting.

In most wealthy industrialized countries, socioeconomic circumstances, hygiene and medical care have improved to such an extent that perinatal complications and congenital disorders have become the major causes of infant mortality and of chronic handicap in children [1].

In the past four decades, we have made impressive advances understandig the cytogenetic and molecular basis of congenital disorders. Dozens of syndromes associated with multiple physical handicaps and mental retardation have been related to specific numerical or structural chromosome aberrations. In situ hybridization and the development of a rapidly increasing number of DNA probes enable the detection of chromosomal abnormalities with a resolution down to the level of a single gene. The discovery of tumor-associated chromosomal aberrations, microdeletions, more than a hundred oncogenes, and the cloning of tumor suppressor genes are illustrations of the importance of molecular cytogenetics.

When I started my scientific career at the end of the fifties, some 1,000 syndromes were proven or suspected to be subject to Mendelian inheritance and this number has now increased to more than 6,600 [2]. Progress in elucidating the protein defects responsible for these single gene disorders has been much slower and currently about 400 protein defects have been delineated [3]. This knowledge has been important not only for basic research on biochemical pathways and cell biology, but has also provided new perspectives for early laboratory diagnosis of index patients, carrier detection, genetic counseling and in some instances, such as phenylketonuria, for newborn screening and early treatment [4]. In those instances where a protein defect is also expressed in cultured skin fibroblasts, amniotic cells or chorionic villi, prenatal diagnosis becomes possible for couples at increased risk [5, 6].

In 1875, Sir Francis Galton was the first to propose that by comparing the concordance of a specific trait or disorder in monozygotic (MZ) twins (assumed to be genetically identical) and dizygotic (DZ) twins (which were assumed to be as similar or as different as any other pair of sibs), it would be possible to distinguish between environmental and heritable effects. DZ twins are derived from the fertilization of two eggs by two sperm. MZ twins are the result of the fertilization of one egg by one sperm that then divides to form two embryos.

Genetic research has made amazing progress since 1875. Advances in molecular DNA techniques and developmental genetics have made it possible to differentiate unequivocally between MZ and DZ twins [1, 14]. It is in this new light that this manuscript will review the recent knowledge about twins and the twinning process.

The incidence of DZ twins differs between population [11, 18]. A DZ twin birth in North America is estimated to occur 9–11 times in every 1000 live births or 1 in 100 births; i.e. 1 in every 50 people is a DZ twin.

The DZ twinning rate is closely related to maternal age, parity, height, weight and gonadotropin levels. An increased DZ twinning rate is seen with increasing maternal age and peaks around 35–39 years; higher parity is also associated with a higher DZ twinning rate [19]. Tall heavy women are more likely to give birth to DZ twins than short thin women. A higher incidence of DZ twins has also been reported with the use of new reproductive techniques.

Recent developments in molecular genetics and cytogenetics have allowed for better understanding of the inheritance and expression of genes. Many newly recognized mechanisms such as genomic imprinting, mosaicism, allelic expansion, cytoplasmic inheritance and uniparental disomy have been recognized to play an important role in human heredity.

Genomic imprinting refers to differences in the phenotype which are observed depending on whether the gene was inherited from the father or from the mother. Genomic imprinting is a difficult concept to understand because imprinting has been used loosely to refer to a number of different mechanisms including psychological development, endocrinological actions of cells and protein-protein interactions. Genomic imprinting produces parent-of-origin effects. Parent-of-origin effects is a term that encompasses many of the non-traditional types of inheritance and other genetic and non-genetic mechanisms which show an effect depending on whether they were paternally or maternally derived.

Tuberous sclerosis complex (TSC) is an autosomal dominant multisystem disorder with extensive clinical variability. Present estimates of the prevalence of TSC suggest that it may exceed 1:6,000. New mutations are frequent, as about 2/3 of all cases are apparently sporadic. Locus heterogeneity has been established, with one gene on chromosome 9q34 (TSC1) and the other on chromosome 16p13.3 (TSC2). The majority of TSC2 mutations are propably subtle alterations. In some cases, somatic and germline mosaicism might be explanations for intrafamilial phenotypic variation and apparent non penetrance. A role of the predicted protein product tuberin in growth suppression would be in agreement with allelic losses observed in tumors of TSC patients. Studies on tuberin using antibodies raised against various parts of the protein can be expected to provide insight into its normal and impaired function.

WBS is an overgrowth malformation syndrome characterized by highly variable expressivity, associated with predisposition to different types of pediatric tumors including Wilms' tumor (WT), adrenocortical carcinoma (ADCC), rhabdomyosarcoma (RMS) and hepatoblastoma (HEP). Most cases are sporadic, however 15% of the cases are familial. Cytogenetic, genetic and molecular analysis of the different forms of this syndrome and associated tumors have provided increasing evidence that the gene (or genes) map to 11p15.5 and that genomic imprinting can account for the strange genetics of this syndrome/tumors [1]. Two candidate genes, H19 and Igf2, which are both imprinted, but in opposite direction, map close to but not within one of the two smallest region(s) defined both by constitutional and tumoral rearrangements. These two genes, H19 and Igf2, and their specific parental imprint, may thus account for the pattern of inheritance observed, the variable expressivity, the specific loss of alleles and the loss of imprint. However, that these genes map 400 kb away from one cluster of breakpoints observed in the cytogenetic cases of WBS suggests that other genes could be involved. Indeed, although mapping to a different subregion, a sequence wit the properties of a tumor Suppressor (rhabdomyosarcoma cell line) has recently been isolated [2].

Furthermore, neither reduplication of the active Igf2 paternal allele nor relaxation of Igf2 imprinting is sufficient for tumorigenesis, thus indicating that other mutation(s) must occur. The phenotypic consequences of these aberrant expressions will be better understood when the tissues, the stage of development or the state of differentiation are precisely identified [3-5].

Prader-Willi syndrome (PWS) and Angelman syndrome (AS) are distinct mental retardation disorders caused by paternal deficiency (PWS) or maternal deficiency (AS) of gene(s) in 15qll.2-ql3. We have constructed a 3.5 Mb yeast artificial chromosome (YAC) contig of the PWS/AS region and cosmid contigs of selected YACs at D15S13, SNRPN, S10, and S113. Cosmid clones have been used for fluorescence in situ hybridization (FISH) detection of deletions in PWS and AS patients. In addition, a total of 28 short tandem repeat polymorphisms (STRs) have been mapped to specific YACs in the contig, providing a highly informative set of markers for detection of deletion or uniparental disomy (UPD) in PWS and AS patients. Use of the 3 most informative markers in this region (S542, S128, and ASSCA-1) plus 3 markers distal on 15q (S123, S125, and S131) provide an efficient diagnostic strategy for UPD15.

A combination of FISH and STR analysis has identified small deletions in one sporadic and one familial case of PWS (family O). Both deletions involve all or part of the SNRPN gene but do not extend telomeric to PAR-5 or PAR-1, two novel transcripts expressed exclusively from the paternal chromosome. However, expression of SNRPN, PAR-5, and PAR-1 is lost in both cases, implying the presence of an imprinting control region near SNRPN. The smallest deletion in family O is estimated at approximately 30-40 kb in size and involves a newly identified CpG island at the 5′ end of SNRPN which is methylated on the maternal chromosome. This small deletion in two PWS affected siblings was present in the father and the paternal grandmother, both of whom were phenotypically normal.

In contrast to the random inactivation of either maternal or paternal X-chromosome in the somatic cells of eutherian mammals, in marsupials the paternal X-chromosome is preferentially inactivated in all cells. Similar exclusively paternal X-inactivation occurs in two extraembryonic cell lineages of mice and rats. Thus, genetic imprinting is an important feature of X-inactivation. In embryonic development the initiation of X-inactivation is thought to occur through the X-inactivation centre, located on the X-Chromosome, and thus imprinting probably acts through this centre. A candidate gene for a role in the inactivation centre is Xist (X inactive specific transcript) which is expressed only from the inactive X-Chromosome. The expression of Xist in the mouse embryo is appropriate for it to be a cause rather than a consequence of inactivation. It appears before inactivation, and only the paternal allele is expressed in the extraembryonic lineages. In the germ cells also changes in X-chromosome activity are accompanied by changes in Xist expression. Studies of methylation of the Xist gene have shown that in male tissues where Xist is not active it is fully methylated, whereas in the female the allele on the active X-chromosome only is methylated. In male germ cells, where Xist is expressed, it is demethylated and the demethylation persists in mature spermatozoa. Thus a methylation difference in germ cells could possibly be the imprint. In androgenotes, with paternally derived chromosomes, Xist is expressed at the 4-cell stage, whereas in gynogenotes and parthenogenotes expression does not appear until the blastocyst stage. Thus, Xist expression shows imprinting. When expression appears in parthenogenotes it is random, suggesting that the imprint has been lost. The Xist gene has no open reading frame and is thought to act through mRNA but its function is unknown.

Our studies aim to identify the mechanisms and genes involved in genomic imprinting in mammalian development and human disease. Imprinting refers to an epigenetic modification of DNA that results in parent-of-origin specific expression during embryogenesis and in the adult. This imprint is reset at each generation, depending on the sex of the parental gametogenesis. Prader-Willi (PWS) and Angelman (AS) syndromes are excellent models for the study of genomic imprinting in humans, since these distinct neurobehavioural disorders are both associated with genetic abnormalities (large deletions, uniparental disomy, and imprinting mutations) of inheritance in chromosome 15q11-q13, dependent on the parental origin (reviewed in ref. 1). Some AS patients have biparental inheritance, consistent with a single imprinted gene (active on the maternal chromosome), whereas similar PWS patients are not found suggesting that at least two imprinted genes (active on the paternal allele) may be necessary for classical PWS. We have previously shown that the small ribonucleoprotein associated protein SmN gene (SNRPN), located in the PWS critical region [2], is only expressed from the paternal allele and is differentially methylated on parental alleles [3]. Therefore, SNRPN may have a role in PWS. Methylation imprints have also been found at two other loci in 15q11-q13, PW71 [4] and D15S9 [5], which map 120 kb and 1.5 Mb proximal to SNRPN, respectively. We have now characterized in detail the gene structure and expression from two imprinted loci within 15q11-q13, SNRPN and D15S9, which suggests that both loci are surprisingly complex, with important implications for the pathogenesis of PWS.

Genetic imprinting implies the preferential or exclusive expression of one of the parental alleles of a subset of autosomal loci. The insulin-like growth factor II (IGF2) and H19 loci are particularly interesting examples of this phenomenon since their products appear to display growth agonistic and antagonistic properties, respectively. In addition, IGF2 and H19 are only 90 kb apart, are expressed from opposite parental alleles [1,2] and show a striking similarity in their spatial expression patterns during human prenatal development [3]. One exception is the choroid plexus and leptomeninges which express 1GF2 biallically with no detectable H19 expression [3]. Observations like these have fuelled ideas that there is an enhancer competition between the IGF2 and H19 loci [4]. The imprinting status of the H19 locus would then indirectly control the expressivity of IGF2. This model is likely to be too simple since the P1 promoter of IGF2 is not functionally imprinted during liver development in humans [4]. Moreover, while the liver P2-P4 promoters are expressed primarily from the paternally derived allele during human prenatal development, the P2-P4 promoters can be expressed from both parental alleles in complex patterns during postnatal human development [5]. The enhancer competition model might be put to the test in human and mouse uniparental embryos since the parental origin of their diploid genomes cannot be discerned. Unexpectedly, H19 which is expressed preferentially from the maternal allele in mouse [6] and human [7] placenta is expressed in both mouse and human trophoblasts (in complete hydatidiform moles) lacking the maternal genome. In the normal human placenta, the repressed paternal H19 allele is more methylated. Interestingly, the CpG methylation pattern of H19 is strikingly similar between normal placenta and complete moles. Hence, both paternal H19 alleles are similarly methylated indicating that postzygotic modification events typical of normal development have taken place in complete moles as well in spite of the absence of the maternal genome. In contrast to the normal placenta, H19 is expressed biallelically in complete moles as assessed by allele-specific in situ hybridisation analysis of dispermie moles [8]. We discuss these results in relation to current models of IGF2/H19 imprinting mechanism(s).

Fragile X syndrome is the most common cause of interited mental retardation in humans, with a frequency of approximately 1 in 1200 males and 1 in 2500 females [1]. It is second only to Down syndrome as a genetic cause of mental retardation, which has an overall frequency of 1 in 600. These frequency estimates suggest that fragile X syndrome accounts for approximately 3% of mental retardation in males, and perhaps as much as 20% in males with IQs between 30 and 55 [2]. The disease derives its name from the observation of a fragile site at Xq27.3 in cultured lymphocytes, fibroblasts and amniocytes [3].

The phenotype of the fragile X syndrome is mental retardation, usually with an IQ in the 4-70 range [4] and a number of dysmorphic features: long face, everted ears and large testicles [for review see ref. 5] (Fig. 1). Not every patient shows all the physical symptoms, which are generally more apparent after childhood. Macroorchidism is a common feature of fragile X syndrome in more than 90% of postpuberal males. Some patients show hyperactivity and attention deficits as well as avoidance behaviour similar to autism. Affected females generally have a less severe clinical presentation, and their IQ scores are generally higher, with typically borderline IQs or mild mental retardation.

No gross pathological abnormalities have been described in the brains of fragile X patients. Only a few post-mortem brain studies of fragile X males have been described and the information is very limited, presenting only non-specific findings such as brain atrophy, ventricular dilatation and pyramidal neurons with abnormal dendritic spines. It has been shown that the volume of the hippocampus was enlarged compared to controls [6], while a significantly decreased size of the posterior cerebellar vermis and increased size of the fourth ventricle was found [7]. Using magnetic resonance imaging it was shown that fragile X patients have an increased volume of the caudate nucleus [8]. The caudate volume is correlated with IQ and methylation status of the FMR1 gene.

With a population incidence of about 1 in 15.000 retinoblastoma is the most frequent intraocular tumor in infancy and early childhood. It occurs in a hereditary form due to a germline mutation in about 40% of patients (30% de novo mutation and 10% transmission from an affected parent) and in a non-hereditary form due to a somatic mutation. The retinoblastoma gene is located on chromosome 13q14. This large gene of about 180 kb, consists of 27 exons of rather different sizes and encodes a 4.7 kb transcript with important function in cell cycle regulation. Individuals with bilateral, multifocal tumors are assumed to carry a germline mutation, whereas unilateral and unifocal tumors are generally due to the somatic form. Both copies of the RB1 gene must be in inactivated before a tumor develops. In about half of patients with the germline mutation the second event inactivating the second allele can be shown by loss of heterozygosity in tumor tissues compared to surrounding somatic tissues.

Knowledge of the RB1 gene locus affords an opportunity to specify the type of mutation in many patients and arrive at a definitive molecular diagnosis. This is the basis for clinical evaluation and genetic counseling. The types of mutation are large scale deletions, small deletions and insertions, and base substitutions. There is no hot-spot for mutations.

During the last several years we have studied more than 200 patients in search for large scale and small deletions and insertions, and missense mutations. Using intragenic polymorphic DNA markers we were able to identify the mutant haplotype in all familial cases. Direct DNA analysis identified a mutation in about 25% of patients. The distribution of lesions will be described in relation to the clinical situation.

It is proposed that transgenerational modulation of gene expression might be possible, if the metabolic response of the parent to some physiological or social stress modified imprint setting. Transcription regulators could theoretically mediate this process. The nature of imprinted genes poised, as it were, between a transcriptionally active and silent state, makes them good candidates for incorporation into the evolution of transgenerational adaption systems where coordinated changes in gene expression over the generations is a selective advantage. The coordination of human fetal (head) growth with the existing size of the mother's pelvis is suggested as just such a circumstance. The reduced birth weight of Dutch babies where their grandmothers suffered acute starvation in mid pregnancy, supports the notion of transgenerational adaption to nutrition, as does the secular change (increase) in child growth over the last century. The recent indication that there may be functional polymorphism in the imprinting of the human IGF2 and IGF2R genes suggests these ideas could be explored using association studies at the population and individual level.

The origins and development of human, medical and clinical genetics are interwoven and yet each of these disciplines follows its own path.

The beginnings of a systematic human genetics can be traced to the middle of the 19th century, but it took human genetics almost 100 years to mature fully and influence medicine. Its origins can be traced to the work of three scientists: Galton, Pearson and Bateson.

In 1865, 6 years after Darwin had published Origin of Species in London, and the year when Mendel's paper, “Experiments on Plant Hybrids”, was published in Brunn, Galton, also in London, set out his first ideas on human heredity. His thinking then developed in two directions. The first laid the foundations for the scientific study of human heredity through biometrics and quantitative genetics. This part of Galton's thinking is summarized in his epitaph: “the dominant idea of his life's work was to measure the influence of heredity on the mental and physical attributes of mankind”.

Galton's other line dealt with the application of heredity through eugenics, a word that Galton coined to signify “well bred”. He wanted, I quote his words: “to produce a highly gifted race of Man by judicious marriage through several generations”. Families of merit should be identified and positively encouraged to breed; conversely, the “ weak could find a welcome … in celebrate monasteries”.

In the beginning the dogma was that sex determination in man followed the Drosophila pattern in which XO is male, XXY female, and the Y chromosome has no direct influence on the determination of sex. On the grounds of specific anomalies with which they presented, females with Turner Syndrome were sex chromatin tested and found to be chromatin negative [1]. This result, confirmed in 1956 by the male frequency of red-green colour blindness in these subjects which indicated that they carried only one X chromosome in spite of their female phenotype, suggested that therefore they might be XO, and, so, hinted that sex determination in man might not follow the then accepted pattern [2]. In 1959 chromosome studies confirmed that XOs were female [3] and showed that subjects with the symmetrical XXY sex-chromosome anomaly were with Klinefelter syndrome [4]. In the same year, by showing that XOs were females also in mice [5] it became accepted that the Y chromosome was the determiner of the formation of the testis in the mammalian embryo, and so was the key element in primary sexual differentation. It would seem appropriate to call this formal model of chromosomal sex determination the Malandrium pattern [6].

In 1966 Jacobs and Ross [7], from work on males with Y chromosome deletions narrowed down the testis determining function of the Y chromosome to its short arm. Then, in 1975, Wachtel and collaborators [8] were the first to formulate a hypothesis on the sex determining gene, or, more precisely on the nature of its product. They suggested that this developmental role might be played by the H-Y antigen, a weak histocompatibility antigen which had been known to be involved in the rejection of male skin grafted onto otherwise histocompatible female mice. The idea had run into technical difficulties and a major problem was related to the significance that should be attached to the results of two different ways for demonstrating the antigen, namely the cell-mediated cytotoxicity test or the serological test. Efforts were made to keep the H-Y hypothesis alive, largely because there was a certain elegance about it [9, 10]. However eventually XX male mice, lacking H-Y by either test, spelt the end of the candidature of H-Y as the testis determining mechanism [11, 12].

The fact that all individuals are not identical in their risk for certain cancers has been known for many years. The factors responsible for the increased risk observed in certain individuals, families, or populations may be either environmental or genetic. The genetic risk factors identified to date generally result in a greatly increased relative risk – on the order of ten to one hundred thousand-fold in the case of familial retinoblastoma, for example. The major reason for the identification of genetic factors that result in such large increases in relative risk is that the transmission of such genes in families results in an easily identifiable pattern of inheritance and the concommitant ability to locate the gene resulting in increased risk by genetic linkage analysis. From our study of sporadic and familial retinoblastoma, we have identified a genetic factor whose action appears to result in a more modest increase in relative risk – on the order of one-hundred to one-thousand-fold. We believe that a locus that is responsible for this increase in relative risk lies on the human X-chromosome. Our data indicate that aberrant alleles at this locus not only appear to increase an individual's risk for a number of cancers, but also participate in the process of genome imprinting.

The epidemiological and genetic arguments in favor of the involvement of an X-linked “imprintor, mutator” gene in the generation of new, heritable, RB-1 mutations are given in detail in Naumova and Sapienza (1994). The pertinent data may be summarized as follows:

1. there is a significant excess-of-males among patients with bilateral sporadic but not unilateral sporadic retinoblastoma;

2. there is preferential retention of paternally-derived RB-1 alleles in bilateral, but not unilateral, sporadic retinoblastoma;

3. sex ratio bias in favor of males is observed among the offspring of male founders but not female founders of retinoblastoma pedigrees (i.e. the male founders fail to transmit their X-chromosomes with the expected frequency;

4. transmission ratio distortion in favor of affected individuals is observed among the offspring of male founders but not female founders, and

5. the affected male offspring of these male founders have families in which offspring are distributed equally among affected and unaffected boys and girls, as expected for an autosomal dominant trait, i.e. sex-ratio and transmission-ratio distortion are observed only among the offspring of male founders of retinoblastoma pedigrees.