Characterization of susceptible chiasma configurations that increase the risk for maternal nondisjunction of chromosome 21
Characterization of susceptible chiasma configurations that increase the risk for maternal nondisjunction of chromosome 21Neil E. Lamb1, Eleanor Feingold2, Amanda Savage1, Dimitris Avramopoulos3, Sallie Freeman1, Yuanchao Gu1, Anni Hallberg5, Jane Hersey1, Georgia Karadima3, Dorothy Pettay1, Denise Saker4, Joe Shen4, Lisa Taft1, Margareta Mikkelsen5, Michael B. Petersen3,5, Terry Hassold4 and Stephanie L. Sherman1,*
1Department of Genetics, Emory University School of Medicine, Atlanta, GA, USA, 2Department of Biostatistics, Emory School of Public Health, Atlanta, GA, USA, 3Department of Genetics, Institute of Child Health, Athens, Greece, 4Department of Genetics and The Center for Human Genetics, Case Western Reserve University School of Medicine, Cleveland, OH, USA and 5Danish Center for Human Genome Research, J.F. Kennedy Institute, Glostrup, Denmark
Received March 12, 1997;Revised and Accepted June 23, 1997
Recent studies of trisomy 21 have shown that altered levels of recombination are associated with maternal non-disjunction occurring at both meiosis I (MI) and meiosis II (MII). To comprehend better the association of recombination with nondisjunction, an understanding of the pattern of meiotic exchange, i.e. the exchange of genetic material at the four-strand stage during prophase, is required. We examined this underlying exchange pattern to determine if specific meiotic configurations are associated with a higher risk of non-disjunction than others. We examined the crossover frequencies of chromosome 21 for three populations: (i) normal female meiotic events; (ii) meiotic events leading to MI non-disjunction; and (iii) those leading to MII non-disjunction. From these crossover frequencies, we estimated the array of meiotic tetrads that produced the observed crossovers. Using this approach, we found that nearly one-half of MI errors were estimated to be achiasmate. The majority of the remaining MI bivalents had exchanges that clustered at the telomere. In contrast, exchanges occurring among MII cases clustered at the pericentromeric region of the chromosome. Unlike the single exchange distributions, double exchanges from the non-disjoined populations seemed to approximate the distribution in the normal population. These data suggest that the location of certain exchanges makes a tetrad susceptible to non-disjunction. Specifically, this susceptibility is associated with the distance between the centromere and closest exchange. This result challenges the widely held concept that events occurring at MII are largely independent of events occurring at MI, and suggests that all non-disjunction events may be initiated during MI and simply resolved at either of the two meiotic stages.
Trisomy 21, the chromosomal abnormality responsible for >95% of individuals with Down syndrome (DS), is the most common identified cause of mental retardation, with an incidence of 1 in 600 live births (1 ). It is thought that 1 in 150 of all conceptions involve a DS fetus, but up to 80% of these are lost during early pregnancy (2 -4 ). As with many aneuploid syndromes, DS shows a marked relationship with advancing maternal age (5 ), exerted without respect to race, geography or socioeconomic factors (6 ). There is a surprising lack of knowledge about the cause of trisomy 21, i.e. the processes that underlie the mechanism of non-disjunction. Recently, however, various studies have shed light on some factors that may be involved. In ~90% of trisomy 21 individuals, the additional chromosome is maternal in origin (7 -15 ). Using pericentromeric markers to infer the meiotic stage of origin for this error, ~70% of the maternal errors have been found to occur during meiosis I (MI), while the other 30% occur during meiosis II (MII). These studies have also identified the first molecular correlate of non-disjunction in humans: altered recombination. We and others have generated genetic linkage maps for chromosome 21 from maternal MI and MII trisomies (7 ,8 ,10 ,15 ). The map for maternal MI trisomy was found to be significantly shorter than the normal female map, indicating a reduction in recombination. The distribution of recombination across the chromosome was also altered for this population, with the reduction confined to proximal regions of the chromosome. Most of the exchange that did occur in these individuals was located near the telomere. In contrast, the genetic length of the MII trisomy-based map was significantly longer than the normal female map, with recombination especially increased near the centromere. Thus, the absence of proximal recombination appears to be associated with MI non-disjunction, while the presence of exchange in this same region is associated with MII non-disjunction. Certain chiasma configurations, therefore, appear to confer increased risk of non-disjunction. These configurations, initiated during MI, may mal-segregate at either MI or MII, or possibly at both divisions. On the basis of these results, a two-hit model for non-disjunction has been put forward (15 ,16 ). The first hit, which occurs during fetal meiosis, establishes bivalents with `susceptible' meiotic configurations. The second hit involves an age-related degradation of a meiotic process which increases the risk of improper segregation for these susceptible bivalents. Under normal meiotic conditions, the presence of a single chiasma-regardless of its location-may be sufficient for proper chromosome segregation. However, as the ovary ages, a decay or breakdown in the meiotic apparatus (e.g. a spindle component or sister chromatid cohesion protein) may occur, disturbing the meiotic process. At this point, certain exchange configurations may be more likely to undergo improper segregation and non-disjunction. In this manner, as the age of a woman increases, so too does her chance of a meiotic disturbance.
It is unclear if these differences in overall patterns of recombination are due to the prevalence of one specific type of chromosomal exchange on chromosome 21 or rather a shift in the global exchange patterns for the chromosome. Genetic maps are a composite of the achiasmate, single and multiple tetrads and are not easily separated into each component part. Although genetic maps can represent locations and clusters of exchanges, they cannot, for example, determine if the proximal clustering found among the MII trisomy 21 cases is due primarily to single or higher order exchanges occurring in this region. To answer this type of question, other methods must be employed.
In most organisms, a chiasma can occur anywhere along the euchromatin of the chromosome arm. The probability of a chiasma, however, is not a constant value across the chromosome, but rather follows specific patterns of placement (17 ). In human males, exchange patterns have been obtained from direct chiasma counts of spermatocytes (18 -21 ). Similar frequency and distribution studies have been difficult to obtain for females, due to technical problems in procuring appropriately staged oocytes for study. Recently, we have presented analytical methods to estimate the array of meiotic tetrads that give rise to a sample of meiotic events based upon the crossover frequencies observed in family data (22 ). Such techniques are based upon methods used to study tetrad exchange patterns in Drosophila (29 ,30 ). The chromosome of interest is divided into several intervals, and the recombinant status (the observed number of recombinations and the interval where each is located) for each chromosome is determined. Chromosome exchange distributions are then estimated from these recombination patterns using maximum likelihood methods. Such an analysis can be applied to both male and female samples, circumventing many technical problems inherent to cytological chiasma counting. In addition, the frequency and distribution of exchanges from abnormal meiotic events, like non-disjunction, can be analyzed by these methods.
We employed these techniques to examine populations of chromosome 21 where either MI or MII non-disjunction has occurred, extending the work of previous recombination-based studies. For this study, we have utilized a more extensive set of chromosomal markers to provide more complete coverage along chromosome 21, particularly near the telomere. The estimated exchange patterns obtained from the non-disjoined populations were then compared with the exchange pattern of a sample of normal chromosome 21 female meiotic events to understand better the significant differences in genetic maps observed between both non-disjunction populations and the normal population. Of specific interest was the determination of the distribution of exchange underlying the telomeric cluster in the MI non-disjunction population and the proximal cluster in the MII non-disjunction population. Our results indicate that susceptible meiotic tetrads are associated with the distance between the centromere and the closest meiotic exchange. Pericentromeric exchanges appear to predispose a tetrad to a MII non-disjunction. On the other hand, if the closest exchange is near the telomere, a great distance from the centromere, a MI non-disjunction susceptibility appears to be established.
The frequencies of each type of recombinant event were calculated for maternal MI and MII trisomic events, as well as for normal female meiotic events. An `observed recombinant' resulting from an exchange occurring during the four-strand stage of meiosis is based on more information when two chromatids are contributed by a parent due to non-disjunction than when only one chromatid is contributed after normal meiosis (see Materials and Methods). When two chromatids are contributed, a transition from heterozygosity to homozygosity (or vice versa) between genetic markers indicates an observed recombinant event. When only one chromatid is contributed, a switch in parental phase between genetic markers indicates an observed recombinant. From these observed data, the overall frequency and placement of exchanges along the tetrad were estimated for each population (Table 1 and Fig. 1 ). On average, the normal population contained 1.38 exchanges per meiotic event for chromosome 21. The average number of exchanges per tetrad for MI non-disjunction was 0.73, confirming the reduction in meiotic exchange for this population.
Our previous studies had estimated that 10% of normal meioses had no exchanges and that these represented, in part, the lack of a fully informative marker set (8 ). With the use of a more extensive set of chromosome markers, we now estimate that none of the normal tetrads are achiasmate. In contrast, ~45% of MI non-disjunction tetrads failed to engage in chromosomal exchange. An odds ratio was computed from this value as a measure of association between exchange and non-disjunction. Since the frequency of achiasmate tetrads in the normal population was zero, we estimated an infinite odds ratio for achiasmate tetrads. In other words, an achiasmate tetrad confers an infinite risk for a MI non-disjunction.
Meiotic events with at least one exchange were examined to determine if certain configurations were susceptible to non-disjunction. The placement of the single exchange events was found to account for nearly all of the deviation in the trisomic distributions (Fig. 2 ). The MI population of single exchanges (n = 74) showed a strong shift towards the telomere. More than 80% of all single exchanges were located in the telomeric third of the chromosome, compared with only 40% of those that segregated normally (n = 141). In contrast, the distribution of single exchanges for the MII population (n = 40) was shifted in the opposite direction, towards the centromere. Nearly 63% of single exchanges occurred in the proximal third of the chromosome for this population, compared with only 35% in the normal population. To investigate this centromeric shift more closely, we subdivided the most proximal interval on the chromosome into four smaller segments and re-analyzed the data (Fig. 3 ). We first examined the location of single exchange events. In the normal population, nearly all of the exchange was located at the distal end of the interval. A marked shift towards the centromere was observed when the MII population was examined; exchanges were widely distributed across the interval and, in general, located more closely to the centromere than in the normal population. Pericentromeric exchange therefore, appears to be a distinguishing feature of MII non-disjunction.
An additional method to compare the exchange distributions identifies the location of the most proximal exchange for each meiotic event among the different populations (Fig. 5 ). This serves as a measure of the distance between the centromere and closest meiotic exchange. Additionally, this type of comparison combines the contributions from both single and double exchange events. For this analysis, the chromosome was divided into thirds (proximal, medial and distal), and the location of the most proximal exchange was identified. From these values, we constructed odds ratios to measure the association between exchange placement and non-disjunction (Table 2 ). We used the medial region as the `referent' (non-susceptible) region. If the most proximal exchange in a meiotic tetrad occurred in the proximal region of the chromosome, that tetrad was ~2.8 times as likely to undergo a MII non-disjunction than if it were in the medial region. In contrast, if the most proximal exchange was in the distal region of the chromosome, that tetrad was ~4.9 times as likely to undergo a MI non-disjunction than if the exchange occurred in the medial region.
To examine the possible association of maternal age and altered exchange, both the MI and MII populations were divided into two groups based upon the age of the woman at the time of delivery/termination of the trisomic offspring (Table 1 ). There were 95 MI cases involving `younger' women (age <= 31 years) and 80 MI cases involving `older' women (age >= 32 years). The average number of exchanges per meiotic tetrad for chromosome 21 were 0.77 and 0.65 for the younger and older MI groups, respectively. By t-test analysis, these values were not significantly different (P = 0.471). Using the same age groupings, the MII population yielded 32 `younger' and 44 `older' cases. The average number of exchanges per meiotic tetrad were 1.63 for the younger and 1.42 for the older group, also not a statistically significant difference (P = 0.075).
For chromosome 21, previous genetic studies have shown that recombination differs between disjoined and non-disjoined chromosome populations. Decreased recombination, except for the most distal regions, is seen among maternal MI errors (7 ,8 ,10 ,15 ), while increased recombination, especially near the centromere, is observed among maternal MII errors. Previous studies using genetic maps for comparison were unable to determine if these altered recombination patterns occur among all meiotic events, or if they represent a specific subgroup of aneuploid events.
We have presented an analysis of the underlying exchange patterns of both normal and trisomic populations to examine the amount and patterns of meiotic exchange that occur in the oocyte prior to chromosome segregation utilizing a more extensive set of chromosome markers. We found that a large proportion of the bivalents involved in MI non-disjunction have no evidence of any meiotic exchange. It is possible that these represent chromosomes that failed to pair during prophase I. Alternatively, it may be that these bivalents did pair, but the tetrads did not participate in exchange. Regardless of which of these occurred, these achiasmate bivalents failed to engage in normal segregation.
Of those MI bivalents that do participate in exchange, most only engage in a single exchange, initiated in the telomeric one-third of the chromosome. In contrast, exchange is divided almost evenly between single and double events in the MII non-disjunction population. More than half of the single exchange events are initiated in the centromeric one-third of the chromosome, many in the pericentromeric region. To summarize, it appears that the location of certain exchanges may make a tetrad susceptible to non-disjunction. Specifically, this susceptibility is associated with the distance between the centromere and closest exchange. The presence of an exchange in the proximal or medial region of the chromosome appears to stabilize the tetrad. If the closest exchange is near the telomere, the tetrad is associated with an MI non-disjunction. In contrast, a pericentromeric exchange confers a susceptibility for MII non-disjunction.
Current theories involving chiasma placement and disjunction are consistent with these findings. For example, Hawley et al. (16 ) proposed that the reduced exchange rates observed for chromosomes non-disjoining at MI reflected the tendency for certain configurations of exchange, specifically single distal exchanges, to undergo non-disjunction more frequently than others. These distal exchanges less effectively link homologs and orient each kinetochore to opposing spindle poles. Even so, in a fully functional oocyte, these susceptible configurations would still disjoin normally. Hawley proposes, however, that the ability of a woman's oocytes to form normal spindles diminishes as the age of the woman increases. Under these impaired conditions, the weakness of susceptible exchanges would become evident. In the presence of a damaged spindle, distally linked homologs would be less likely to orient correctly at the metaphase plate than their counterparts with proximal exchanges. Subsequently, these bivalents, along with achiasmate tetrads, would be ejected from the spindle and, as a tetrad, randomly migrate into the secondary oocyte or polar body.
Models have also been advanced to explain the increased proximal exchange observed in chromosomes that undergo MII non-disjunction. Possibly, proximal chiasmata predispose to `chromosome entanglement' at MI, with the bivalent being unable to separate, passing intact to the MII metaphase plate (15 ). Upon MII division, the bivalent divides reductionally, resulting in a disomic gamete with identical centromeres. In this manner, proximal recombination, an event which occurred during MI, is resolved and visualized as an MII error. This model also depends upon some type of age-dependent meiotic disturbance in addition to a susceptible meiotic tetrad for the occurrence of a non-disjunction event.
An alternate theory proposes that the resolution of proximal chiasmata leads to premature sister chromatid separation just prior to anaphase I (23 ). Resolution of chiasmata requires the release of sister chromatid cohesion distal to the site of the exchange (16 ). Attempts to resolve chiasmata that are very near the centromere could result in premature separation of the chromatids. If the sister chromatids migrate to a common pole during MI, they have a 50% chance of randomly traveling into the same product of meiosis during MII, resulting in an apparent MII non-disjunction. Studies of non-disjunction in both humans (24 ) and Drosophila (25 ) have provided preliminary support for this model.
A common thread which runs through many of these models requires the summation of specific meiotic events to occur before a non-disjunction takes place. In essence, this is identical to the two-hit model discussed earlier. Although specific patterns of exchange may be susceptible to non-disjunction, on their own these patterns are not sufficient to ensure mal-segregation. Some additional meiotic agitation (abnormal spindle formation, premature release of sister chromatid adhesion, hormonal imbalance, etc.) is required for non-disjunction. As there is no maternal age effect on meiotic exchange, it is this meiotic agitation that must be maternal age-dependent.
Whatever mechanism is responsible for the non-disjunction event, it appears that specific patterns of chromosomal exchange are associated with chromosome 21 aneuploidy. The trisomic population, however, does contain a subgroup of tetrads which mirror the normal population in number and placement of exchanges. For this group, circumstances in the oocyte, unrelated to genetic recombination, may be responsible for chromosome mal-segregation. For example, the compromised microcirculation hypothesis of Gaulden (26 ) suggests that non-disjunction arises from a series of cascading events. Initially, a poorly developed microvasculature is created around the ovarian follicle, resulting in a reduced blood flow through the area. Gaulden suggests this could be due to a hormonal imbalance occurring during follicular development. An alternative model, proposed by Avramopoulos et al. (27 ) examines the possible role of apolipoprotein E alleles in the developing oocyte. A significantly higher frequency of allele [epsilon]4 was identified in young mothers with an MII error. Carriers of this allele have a higher plasma cholesterol, which is strongly related to atherosclerosis. Avramopoulos et al. suggest that in these mothers, atherosclerosis develops in the microvasculature around the maturing follicle, leading to an oxygen deficit inside the follicle. Regardless of the mechanism, the resulting oxygen deficit leads to an increase in the concentration of carbon dioxide and lactic acid inside the follicle, which in turn decreases the intracellular pH of the oocyte. This pH drop reduces the size of the mitotic spindle, with the subsequent displacement and non-disjunction of a chromosome. A similar cascade of events, observed in grasshopper neuroblasts, usually leads to the non-disjunction of the smallest chromosome. In humans, chromosome 21 is the smallest chromosome, comprising 1.4% of the genome (28 ) and would be likely to be the first chromosome ejected from the spindle under this type of model. In this manner, such a cascade of events could lead to non-disjunction of an otherwise normal meiotic tetrad.
It seems likely that several pathways exist which lead to non-disjunction. These data indicate that at least two mechanisms, one involving achiasmate and the other involving chiasmate tetrads, are dependent on meiotic exchange. In addition, other mechanisms may exist in which meiotic exchange is not a factor. It remains to be seen if these exchange distributions are restricted to chromosome 21 or instead represent a genome-wide pattern. Such studies are critical to a complete understanding of the possible role of meiotic exchange in proper chromosome segregation.
We previously have described an analysis that can be used to construct exchange maps based upon the recombination frequencies observed in a population of normal meiotic events (22 ). Such techniques are based upon methods used to study tetrad exchange patterns in Drosophila (29 ,30 ). Briefly, we observe frequencies, r, of each recombination type and use these to estimate frequencies, e, of each corresponding exchange pattern that occurred during the four-strand stage of meiosis. The chromosome of interest is divided into several intervals, and the recombinant status (the observed number of recombinations and the interval where each is located) for each chromosome is determined. In this way, the overall recombination pattern and frequency of each recombinant class can be observed, e.g. a single recombinant in the second interval, a double recombinant in intervals 1 and 6. Chromosome exchange distributions are then estimated from these recombination patterns using maximum likelihood methods.
Such an analysis can be extended to study populations arising from abnormal meiotic events, such as non-disjunction. The equations which transform recombination frequencies into exchange values for normal meiotic events [as developed in Lamb et al. (22 )] must, however, be modified. For example, among normal individuals, only one of the four chromatid strands from the parental meiotic tetrad is passed into the gamete. Subsequently, in any given interval, a meiotic exchange has a 50% chance of appearing in the chromatid that is observed, independently of other intervals. Among trisomic individuals, however, two of the four tetrad strands are available for study. As a result, more information is obtained about the tetrad. The amount of information available differs depending upon the origin of the non-disjunction error and the number of exchanges. For example, errors which arise at MI produce progeny with one chromatid from each of the two homologs. Assuming that the parent in which the non-disjunction error occurred was heterozygous at the centromere, trisomic progeny will retain this heterozygosity. The most proximal exchange is detected 50% of the time and is observed as reduction to homozygosity at loci distal to the exchange.
In contrast, errors that arise at MII produce progeny with both sister chromatids from the same homolog. Again, assuming a parent is heterozygous at the centromere, these trisomic individuals show loss of the heterozygosity (reduction to homozygosity) at the centromere. Under this scenario, for a MII non-disjunction population, the most proximal exchange will always be detected and is observed as a transition from homozygosity to heterozygosity at a distal locus. Thus, the underlying exchange equations must account for the additional chromatid recovered, as well as the type of non-disjunction error that has occurred. The following example illustrates this procedure. Suppose we divide the chromosome into just two intervals, and observe r00 = frequency of non-recombinants, r10 = frequency of chromatids with a single crossover in the first interval but not the second, r01 = frequency of chromatids with a single crossover in the second interval but not the first, and r11 = frequency of chromatids with a crossover in both the first and second interval. The equations relating rs and es are:
For normal meiosis:
r00 =e00 + 1/2 e10 + 1/2 e01 + 1/4 e11
r10 = 1/2 e10 + 1/4 e11
r01 = 1/2 e01 +1/4 e11
r11 = 1/4 e11
For MI non-disjunction:
r00 =e00 + 1/2 e10 + 1/2 e01 + 1/4 e11
r10 = 1/2 e10
r01 = 1/2 e01 + 1/4 e11
r11 = 1/2 e11
For MII non-disjunction:
r00 =e00
r10 = e10 + 1/2 e11
r01 = e01
r11 = 1/2 e11
The maximum likelihood estimates of the exchange frequencies, es, can be obtained by solving these equations, using the invariance property of maximum likelihood estimation.
The present study population consists of 292 trisomy 21 conceptions of maternal origin and their parents. In all instances, cytogenetic studies were consistent with non-mosaic trisomy 21. The majority of cases were ascertained because of clinical features of DS. This dataset includes trisomy cases analyzed in previous studies (7 -9 ,14 ,15 ). DNA was prepared from either blood samples or frozen fetal tissue as described elsewhere (31 ,32 ).
Physical locations for markers based upon data presented by Lawrence et al. 1993 (34).
We used a panel of 45 chromosome 21 DNA markers, grouped into 20 tightly linked regions to determine the parental origin of trisomy, the meiotic stage of error and recombination status. As previously noted, only maternal errors were included in this study. Among these cases, 199 (68.1%) were attributed to MI non-disjunction and 81 (27.7%) to MII non-disjunction. The remaining 12 (4.1%) were reduced to homozygosity at all markers examined and subsequently classified as mitotic in origin.
The ability to distinguish mitotic from maternal MII meiotic errors is important when examining tetrad exchange rates. If such cases in fact are meiotic in origin, their exclusion will erroneously increase the average number of exchange events per tetrad. On the other hand, if such cases are actually mitotic, their inclusion as meiotic will decrease the average number of exchanges per tetrad. As previously discussed (13 ,15 ), it seems likely that most, if not all, of these cases are truly mitotic in origin.
The genotype data for the normal population was obtained from the 40 families in the CEPH database. The chrompic option of the mapping program CRIMAP (33 ) was applied to these data to determine recombination status across each chromosome arm, assuming the most likely phase.
For each population, the chromosome was divided into six roughly equal physical intervals with an average interval size of 6.3 Mb (Table 3 ). Alternatively, the chromosome can be divided into intervals based upon genetic lengths. The choice of physically equal intervals allows us to gather information with respect to the physical boundaries of the chromosome, i.e. how many megabases separated the centromere and first exchange. Each interval was examined for the presence or absence of a recombinant event. If a recombinant occurred at the junction of two intervals, the recombination was recorded in both intervals, each with a probability of occurrence of 1/2. This allowed determination of the overall recombination pattern for each chromosome, which was then converted into frequencies for each class of recombination. Once these frequencies were known, the exchange distributions could be estimated.
No statistical comparisons between the different exchange distributions were made for two reasons. First, previous work found statistically significant differences between genetic maps of MI cases, MII cases and normal female meiotic events using likelihood ratio tests (7 ,8 ,15 ). In this new analysis that estimates the number of exchanges, our aim was simply to describe those differences more fully. Statistical hypothesis tests are not particularly useful for describing differences between distributions, particularly distributions on as many as six points such as those in Figure 2 . Second, to test the hypothesis that the distributions in Figure 2 are different, the `obvious' analysis would be a [chi]2 test. However, use of that test would be incorrect because the proportions shown are estimated, not observed directly. Such a test would be strongly biased towards showing a difference. More sophisticated bootstrap or permutation methods are possible, but there are some unresolved technical issues in the correct application of them to these data. The issue of appropriate statistical analysis of this type of data is discussed further in Lamb et al. (22 ).
We thank W. Robinson for insightful discussion; P.A. Jacobs, I. Uchida and C. Torfs for contributing cases to the study; H. Willard, S. Schwartz and A. Chakravarti for valuable discussions; and A. Lynn for chromosome 21 genotyping data from the CEPH families. This work was supported by NIH PO1 HD32111, and E.C. grant GENE-CT93-0015 and BMH4-CT96-0554 to the European Chromosome 21 Consortium (M.B.P.).
2 Bous, J., Boué, A. and Lazar, P. (1975) Retrospective and prospective epidemiological studies of 1,506 karyotyped spontaneous abortions. Teratology, 12, 11-26.
3 Hassold, T. and Jacobs, P. (1984) Trisomy in man. Annu. Rev. Genet., 18, 69-97.MEDLINE Abstract
4 Freeman, S., Grantham, M., Hassold, T., Herbert, M., Hersey, J., Nuccio, J., Pettay, D., Takaesu, N. and Phillips, C. (1991). Cytogenetic and molecular studies of human spontaneous abortions. Am. J. Hum. Genet. Suppl., 49, 916A.
5 Penrose, L.S. (1933) The relative effects of paternal and maternal age in Mongolism. J. Genet., 27, 219-224.
6 Matsunga, E. and Maruyama, T. (1969) Human sexual behavior, delayed fertilization and Downs syndrome. Nature, 221, 642-644.
7 Sherman, S.L., Takaesu, N., Freeman, S.B., Grantham, M., Phillips, C., Blackston, R.D., Jacobs, P.A., Cockwell, A.E., Freeman, E., Uchida, I., Mikkelsen, M., Kurnit, D.M., Buraczynska, M., Keats, B.J. and Hassold, T.J. (1991) Trisomy 21: association between reduced recombination and non-disjunction. Am. J. Hum. Genet., 49, 608-620.MEDLINE Abstract
8 Sherman, S.L., Petersen, M.B., Freeman, S.B., Hersey, J., Pettay, D., Taft, L., Frantzen, M., Mikkelsen, M. and Hassold, T.J. (1994) Non-disjunction of chromosome 21 in maternal meiosis I: evidence for a maternal age dependent mechanism involving reduced recombination. Hum. Mol. Genet., 2, 1529-1535.
9 Yoon, P.W., Freeman, S.B., Sherman, S.L., Taft, L.F., Gu, Y.,Pettay, D., Flanders, W.D., Khoury, M.J. and Hassold, T.J. (1996) Advanced maternal age and the risk of Down syndrome characterized by the meiotic stage of the chromosomal error: a population based study. Am. J. Hum. Genet., 58, 628-633.MEDLINE Abstract
10 Warren, A.C., Charkravarti, A., Wong, C., Slaugenhaupt, S.A., Halloran, S.L., Watkins, P.C., Metaxotou, C. and Antonarakis, S.E.. (1987) Evidence for reduced recombination on the non-disjoined chromosome 21 in Down syndrome. Science, 237, 652-654.MEDLINE Abstract
11 Antonarakis, S.E. and Down Syndrome Collaborative Group (1991) Parental origin of the extra chromosome in trisomy 21 as indicated by analysis of DNA polymorphisms. New Engl. J. Med., 324, 872-876.MEDLINE Abstract
12 Antonarakis, S.E., Petersen, M.B., McInnis, M.G., Adelsberger, P.A., Schinzel, A.A., Binkert, F., Pangalos, C. et al. (1992) The meiotic stage of non-disjunction in trisomy 21: determination by using DNA polymorphisms. Am. J. Hum. Genet., 50, 544-550.MEDLINE Abstract
13 Antonarakis, S.E., Avramopolos, D., Blouin, J., Talbot, C.C. and Schinzel, A.A. (1993) Mitotic errors in somatic cells cause trisomy 21 in about 4.5% of cases and are not associated with advanced maternal age. Nature Genet., 3, 146-150.MEDLINE Abstract
14 Mikkelsen, M., Hallberg, A., Poulsen, H., Frantzen, M., Hansen, J. and Petersen, M.B. (1995) Epidemiology study of Down's syndrome in Denmark, including family studies of chromosomes and DNA markers. Dev. Brain Dysfunct., 8, 4-12.
15 Lamb, N.E., Freeman, S.B., Savage-Austin, A., Pettay, D., Taft, L., Hersey, J., Gu, Y., Shen, J., Saker, D., May, K., Avramopoulos, D., Petersen, M.B., Hallberg, A., Mikkelsen, M., Hassold, T.J. and Sherman, S.L. (1996) Non-disjunction of chromosome 21: evidence for initiation of all maternal errors during meiosis I. Nature Genet., 14, 400-405.MEDLINE Abstract
16 Hawley, R.S., Frazier, J. and Rasooly, R. (1994) Separation anxiety: the biology of non-disjunction in flies and people. Hum. Mol. Genet., 3, 1521-1528.MEDLINE Abstract
18 Hultén, M.A. (1974) Chiasma distribution at diakinesis in the normal human male. Hereditas,76, 55-78.
19 Hultén, M.A., Lawrie, M. and Laurie, D.A. (1990) Chiasma-based genetic map of chromosome 21. Am. J. Med. Genet. Suppl., 7, 149-154.
20 Laurie, D.A. and Hultén, M.A. (1985) Further studies on bivalent chiasma frequency in human males with normal karyotypes. Ann. Hum. Genet., 49, 189-201.MEDLINE Abstract
21 Laurie, D.A., Hultén, M.A. and Jones, G.H. (1981) Chiasma frequency and distribution in a sample of human males: chromosomes 1, 2, and 9. Cytogenet. Cell Genet., 31, 153-166.MEDLINE Abstract
22 Lamb, N.E., Feingold, E. and Sherman, S.L. (1996) Estimating meiotic exchange patterns from recombination data: an application to humans. Genetics, in press.
23 Koehler, K.E., Boulton, C.L., Collins, H.E., French, R.L., Herman, K.C., Lacefield, S.M., Madden, L.D., Schuetz, D. and Hawley, R.S. (1996) Spontaneous X chromosome non-disjunction events occurring at MI and MII in Drosophila melanogaster oocytes have different recombinational histories. Nature Genet., 14 406-413.
24 Angell, R.R., Xian, J., Keith, J., Ledger, W. and Baird, D.T. (1994) First meiotic division abnormalities in human oocytes: mechanisms of trisomy formation. Cytogenet. Cell Genet., 65, 194-202.MEDLINE Abstract
25 Miyazaki, W.Y. and Orr-Weaver, T.L. (1992) Sister chromatid misbehavior in Drosophila ord mutants. Genetics, 132, 1047-1061.
26 Gaulden, M.E. (1992) Maternal age effect: the enigma of Down syndrome and other trisomic conditions. Mutat. Res., 296, 69-88.MEDLINE Abstract
27 Avramopoulos, D., Mikkelsen, M., Vassilopoulos, D., Grigoriadou, M. and Petersen, M.B. (1996) Apolipoprotein E allele distribution in parents of Down's syndrome children. Lancet, 347, 862-865.MEDLINE Abstract
28 Antonarakis, S.E., Warren, A.C., McCormick, M.K. and Lewis, J.G. (1989) Molecular mapping of chromosome 21 and the region responsible for Down's syndrome. In Hassold, T.J. and Epstein, C.J. (eds), Molecular and Cytogenetic Studies of Non-disjunction. Liss, New York, pp. 29-43.
29 Weinstein, A. (1936) The theory of multiple strand crossing over. Genetics, 21, 155-199.
30 Merriam, J.R. and Frost, J.N. (1964) Exchange and non-disjunction of the X chromosomes in female Drosophila melanogaster. Genetics, 49, 109-122.
31 Hassold, T.J., Kumlin, E., Takaesu, N. and Leppert, M. (1985) Determination of the parental origin of sex-chromosome monosomy using restriction fragment length polymorphisms. Am. J. Hum. Genet., 37, 965-972.
32 Grimberg, J., Nawoschik, S., Belluscio, L., McKee, R., Turck, A. and Esienberg, E. (1989) A simple and efficient non-organic procedure for the isolation of genomic DNA from blood. Nucleic Acids Res.,17, 8390.MEDLINE Abstract
33 Lander, E.S. and Green, P. (1987) Construction of multilocus genetic linkage maps in humans. Proc. Natl Acad. Sci. USA, 84, 2363-2367.MEDLINE Abstract
34 Lawrence, S., Collins, A., Keats, B.J., Hulten, M.A. and Morton, N.E. (1993) Integration of gene maps: chromosome 21. Proc. Natl Acad. Sci. USA, 90, 7210-7214.MEDLINE Abstract
*To whom correspondence should be addressed. Tel: +1 404 727 5862; Fax: +1 404 727 3949; Email: ssherman@gmm.gen.emory.edu
-->
This page is maintained by OUP admin. Last updated Wed Aug 13 15:52:16 BST 1997. Part of the OUP Journals World Wide Web service.
Copyright
Oxford University Press, 1996
