Meiosis in the Human Female. Meiosis in the human female starts in the 3rd month of fetal life, when each ovary contains several million oocytes [Fig. 2(a)]. By the 6th month of fetal life, the oocytes reach the dictyotene stage of prophase I, in which they remain arrested until sexual maturity. However, most of the oocytes undergo programmed cell death (apoptosis). At puberty, their num­ber is less than 200,000. Of these, only about 400 become ovulated. Thus, an oocyte which is ovulated by a 40-year-old woman was arrested at dictyotene for about 40 years and then completes the first meiotic division. The second meiotic division is not finished until after fertilization. In other words, the first and second meiotic division with its high rate of non-disjunction, takes place around concep­tion [Fig. 2(b)]. At least from a theoretical point of view, chromosomal segregation should be most sensitive to envi­ronmental factors at this time, i.e., about 2 weeks after the last menstruation period.

Endogenous Risk Factors for Trisomy 21. Down’s syndrome (DS) is the most common birth defect known to be associated with an abnormal karyotype. Most trisomy 21 is due to an error in maternal meiosis, whereby about 70% originate during meiosis I (MI) and about 20% dur­ing meiosis II (MII); a defective paternal meiosis is found in up to 8% of all cases [10,11]. More than 95% of all cases are due to free trisomy 21, only a small proportion arises by a mitotic error after fertilization. Both aspects, a conspicuous phenotype and a high proportion of new mu­tants, make surveillance of trisomy 21 particularly suitable for assessing mutagenic hazards and endogenous risk fac­tors. Therefore, trisomy 21 is registered in nearly all moni­toring programs for congenital malformations as a para­digm for aneuploid mutations. Since trisomy 21 has also been extensively analyzed at the molecular level, this article will focus on risk factors for trisomy 21. Its aim is not to give a comprehensive review but a brief, exemplary selection (Table 2).

Trisomy 21 is associated with advanced maternal age, which is the established and most significant risk factor for its occurrence. A minor, generally accepted, risk factor is the birth of a previous child similarly affected with free trisomy 21. These cases might be due to gonadal mosai­cism or premature centromere division (PCD). However, it has been suggested that in some families with multiple trisomy 21 conceptions, a genetic predisposition to non-disjunction 21 may have been unmasked.

Evidence for genetic risk factors in meiotic non-dis­junction are, amongst others, abnormal folate and methyl metabolism in mothers with DS infants, which are corre­lated with polymorphisms in the MTHFR and/or MTRR genes [12]. In other cases with recurrent aneuploidy, a role for mitochondrial DNA (mtDNA) mutations was sug­gested, which could also be relevant for the maternal age affect [13]. In addition, a number of gene mutations af­fecting the germ line or somatic cells have been identified that increase the risk of mitotic non-disjunction (Table 3).

We have studied a consanguineous Arab family with 10 children, of whom three had free trisomy 21. The pa­rental karyotypes were normal and did not exhibit PCD. Based on 22 evenly distributed microsatellites, it was shown that the extra chromosome was of maternal origin, resulting from an error in MI (two cases) and MII. In addi­tion, we found that the rate of somatic malsegregation of chromosome 21, but not of 18, 22 or X, was significantly increased in the mother. It was in the normal range in her mother, grandmother and, also in her three trisomic chil­dren who carry the same chromosomes 21. Thus, the high rate of maternal meiotic non-disjunction is paralleled by an increase of mitotic non-disjunction, however, confined to chromosome 21 [14]. The genetic basis of this phenom­enon is obscure. It could be due to a recessive gene in combination with a specific chromosomal risk factor (Ta­ble 2).

There is clear evidence that in DS the recombination frequency of chromosome 21 was reduced for chromo­somes with MI errors and was higher for MII errors, and both types were associated with advanced maternal age [15-18]. These altered levels of recombination represent associations, they are neither necessary nor sufficient causes for trisomy 21, since a number of control individu­als arising by regular meioses (without non-disjunction) have recombinations in the same chromosomal areas as have their aneuploid counterparts, and the distribution of double exchanges are nearly the same for normal and aber­rant meioses.

Exogenous Risk Factors for Trisomy 21. Numerous studies have tried to establish exogenous risk factors for trisomy 21 [reviewed in 5,10,19-21] (Table 2). Maternal age distribution and selective abortion after prenatal diag­nosis have the strongest influence on its frequency in west­ern countries. If these variables remain constant, then any sudden increase in frequency must be due either to chance or to an environmental factor. Nonetheless, even despite decades of research, apart from maternal age, no single exogenous factor responsible for trisomy 21 has been un­ambiguously identified.

Two convincing correlations were reported to explain two local clusters of trisomy 21 by hazards occurring around the time of conception: the ingestion of a chemical employed against fish parasites [22] and the inhalation of iodine-131 from the Chernobyl reactor accident [23]. There are two other epidemiological studies in which parents were exposed to a high dose of irradiation at the time of conception. These concerned the inhabitants of Kerala, India [24], and Yangjiang Province, China [25], who received high background radiation from monazite soil. The only relevant health problem was a significant increase in trisomy 21. Altogether, these studies imply that chromosomal segregation during human oogenesis around conception can easily be disturbed. Perhaps this is one explanation, apart from reporting bias, that a number of correlations between exogenous and endogenous factors and trisomy 21 were reported but could not be confirmed in other studies.

1. True non-disjunction	Both homologues segre­gate together
2. Achiasmatic non-disjunction	None or prematurely resolved chiasmata
3. Premature separation of sister chromatids	Segregation of a whole chromosome and a chromatid
4. Anaphase lagging	Loss of one chromosome (chromatid)
5. Secondary non-disjuntion	Somatic non-disjunction in maternal germ cells

Figure 1. Incidence of aneuploidies in man.

Figure 2.  Meiosis in the human female. a) Meiosis starts in early fetal life and at birth all oocytes are arrested at the dictyotene stage. Most of the oocytes degenerate. b) Meiosis continues in individual oocytes after puberty. The first meiotic division occurs immediately before fertilization, the second a few hours thereafter (after [28]). In case of a zygote with trisomy 21, environmental effects during early oocyte development occurred in the grandmother, while effects during meiotic divisions operate in the mother.