 

 

;.
f

:32; . I1
37;, ‘“ I

‘ a... .2)...
Ju.u.u.«r .
;. i

Elia-5 .
2....f'avi 1.: LI...-
ft?

33:13“

Eﬁgas

 

 

 

.1 >
r .2
7‘...

{L- ‘1‘!-

»rWua I 4.3.UUM

“wit... ..

zirizxssx 2
foxxﬁioh.
gust! .

.i

.35.,
. ,
:11} .35 AI hﬁoiﬁkis.‘
.3. .ni. _ 4..-... I. 2.. .32.;
‘ he... a... 2.1.:
ﬂawgk.
.

(’1'!
, ll

 

n‘b
. uvtn.nl~
.Hwti... a
A)... tutu?
52...... L
.

513... p).

. 1.. {an
.53 E.
Esnfikztr
{1...} .
33.2....
.e

litannﬂ:rav

.: £339... 4;:
. t;

.\ fixvzln‘
Ll I7!
1 . Izt...

32‘

z .‘\...r,
\I .25 .xl’ii. n

 

     

’1‘ 00/11

 

LIBRARY ‘
Michigan State
University

 

 

 

 

This is to certify that the
thesis entitled

APOLIPOPROTEIN E AS AN HEREDITARY RISK FACTOR FOR NON-
DISJUNCTION - A FEASIBILITY STUDY .

presented by

Nicole M Jones

has been accepted towards fulﬁllment
of the requirements for

 

.MaaLEL—degree in Sci pnpp

Major profes

Date

0.7639 MS U is an Afﬁrmative Action/Equal Opportunity Institution

 

PLACE IN RETURN Box to remove this checkout from your record.
TO AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE

DATE DUE

DATE DUE

 

g

 

 

 

 

 

 

 

 

 

 

 

 

 

6/01 c:/CIRC/Dat90ue.p65op.15

 

APOLIPOPROTEIN E AS AN HEREDITARY RISK FACTOR FOR NON-
DISJUNCTION - A FEASIBILITY STUDY

By

Nicole M. Jones

A THESIS

Submitted to
Michigan State University
in partial fulﬁllment of the requirements
for the degree of

MASTER IN SCIENCE
Department of Epidemiology

2002

ABSTRACT

APOLIPOPROTEIN E AS AN HEREDITARY RISK FACTOR FOR NON-
DISJUNCTION - A FEASIBILITY STUDY

By

Nicole M. Jones

Chromosomal trisomy is a major contributor to pregnancy loss. Although it has
been 40 years since the discovery of the ﬁrst human trisomy, maternal age is the only
well documented risk factor. There is a large variation in the frequency of different types
of chromosomal trisomy sampled at different times in pregnancy. Through the use of
molecular markers, it is possible to determine the parent in which the nondisjunction
event occurred and the cell division of error.

Both Alzheimer’s disease and Down syndrome have been associated with the
allele Apolipoprotein 84. We conducted a feasibility study aimed at developing methods
for a larger study guided by the hypothesis that Apolipoprotein s4 is a risk factor for non-
disjunction and Alzheimer’s disease. Our feasibility study was designed to develop
methods for measuring family history of Alzheimer’s disease and stage of non-
disjunction error among parents of trisomy pregnancies. We designed a case-control
study with cases matched to controls on ethnicity and frequency matched on age. A total
of 29 cases and 61 controls participated in our feasibility study. During our feasibility
study, we identiﬁed a collection of potential cases, archived trisomy DNA samples,
reﬁned our interview and laboratory instruments through ﬁeld-testing, and developed and

debugged a Microsoft-Access database capable of storing our interview data.

ACKNOWLEDGMENTS

I would like to acknowledge the large contribution that my advisory committee
made to the completion of my thesis project, especially my advisor Dr Claudia Holzman.
I’d also like to thank the entire Epidemiology department for creating a supportive
environment and my family for all of their encouragement. Finally, I would like to thank

my husband for learning how to am the copier at the library faster than anyone I know.

iii

TABLE OF CONTENTS

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

LIST OF TABLES v
LIST OF FIGURES vi
CHAPTER 1: CHROMOSOMAL TRISOMY I
INTRODUCTION I
HISTORY -_ - .............. .2
PREVALENCF 2
PARENTAL ORIGIN .- - .............. 4
ORIGIN OF ERROR S
ETTOLOGY ...... 3
MATERNAL AGE - 9
MODELS FOR THE MATERNAL AGE EFFECT l l
PATERNAL AGE 13
RECURRENCE STUDIES 13
TRISOMY 2 1 -PARENTAL ORIGIN 20
TRISOMY 21-STAGE 0F ERROR 20
TRISOMY21-EEEECT OF RBCOMBINATION 21
TRISOMY l8-EPIDEMIOLOGY 21
TRISOMY l3-EPIDEMIOLOGY 22
TRISOMY l6-EPIDEMIOEOGY: 22
SUMMARY OF PART ONE 24
CHAPTER 2: APOLIPOPROTEIN F- 25
BACKGROUND 25
LONGEvrIY STUDIES 26
CHOLESTEROL LEVELS, CARDIOVASCULAR DISEASE, AND STROKE ....... 26
OTHER ASSOCIATIONS .............. 27
LINK TO ALZHEIMER’S DISEASE ______ 27
APOLIPOPROTEIN E, ALZHEIMER’S DISEASE, AND DOWN SYNDROME ......................... 28
SUMMARY OF PART “NO: 3 I
CHAPTER 3: LESSONS FROM A FEASIBILITY STUDY: An Investigation of
Apolipoprotein E as an hereditary risk factor for non-disjunction and Alzheimer ’3
disease. 32
RATIONALE AND SPECIFIC AIMS 32
FEASIBIerY STUDY GOALS 33
STUDY DESIGN 33
STUDY SAMPLE .. -- - 34
PROTOCOL FOR CONTACT or CASES AND CONTROLS - 37
INTERVIEme METHODS ...... 38
INTERVIEW CONTENT ..... 39
DNA SAMPLE COLLECI'ION 39
APOE LABORATORY ASSAY METHODS - .......................................... 40
RESULTS GOAL 1 - 42
RESULTS GOAL 2 .......... -- 46
RESULTS GOAL 3 ...... 48
RESULTS GOAL 4 ........................... 49
DISCUSSION - ..... . .................................................... 53
MAJOR FEASIBILITY STUDY ACCOMPLISHMENTS ..... - - 55
Appendix: Interview Content -- 57

 

iv

LIST OF TABLES

 

 

 

 

 

 

 

 

 

 

 

Table 1: Frequency of chromosomal trisomy 4
Table 2: Cell Division of Nondisjunctional Error

Table 3: The risk for Down syndrome by maternal age and stage of pregnancy. 10
Table 4: Population based studies which analyze the risk of trisomy recurrence. 18
Table 5: Proportion of meiotic errors for trisomies 21, 18, 16, and l3.* ................... 21
Table 6: Summary of trisomies 21, 18, 13, and 16 information 23
Table 7: Number of Caucasian woman interviewed by age category 42
Table 8: Ascertainment of Caucasian Cases 42
Table 9: Number of Women Able to Estimate Relatives’ Age of Death Compared to
the Number of Women who stated that the relative had died 45
Table 10: Number of Women Unable to Report about Alzheimer’s Disease Among,
Their Parents and Grandparents 46
Table 11: Gene Frequencies for Caucasian Mothers 48
Table 12: Ascertainment of Potential Cases - 48
Table 13: Control Contact Results by Age Groups 49
Table 14: Summary of suggested changes in Protocol 52
Table 15: Summary of Methodology Strengths 53

 

LIST OF FIGURES

Figure l: Meiosis and Nondisjunction

 

Figure 2: Hha I Restriction pattern of different ApoE genotypes

 

Figure 3: Origin of Cases

 

vi

26

ABSTRACT

APOLIPOPROTEIN E AS AN HEREDITARY RISK FACTOR FOR NON-
DISJUNCTION - A FEASIBILITY STUDY

By

Nicole M. Jones

Chromosomal trisomy is a major contributor to pregnancy loss. Although it has
been 40 years since the discovery of the ﬁrst human trisomy, maternal age is the only
well documented risk factor. There is a large variation in the ﬁequency of different types
of chromosomal trisomy sampled at different times in pregnancy. Through the use of
molecular markers, it is possible to determine the parent in which the nondisjunction
event occurred and the cell division of error.

Both Alzheimer’s disease and Down syndrome have been associated with the
allele Apolipoprotein 84. We conducted a feasibility study aimed at developing methods
for a larger study guided by the hypothesis that Apolipoprotein s4 is a risk factor for non-
disjunction and Alzheimer’s disease. Our feasibility study was designed to develop
methods for measuring family history of Alzheimer’s disease and stage of non-
disjunction error among parents of trisomy pregnancies. We designed a case-control
study with cases matched to controls on ethnicity and frequency matched on age. A total
of 29 cases and 61 controls participated in our feasibility study. During our feasibility
study, we identiﬁed a collection of potential cases, archived trisomy DNA samples,
reﬁned our interview and laboratory instruments through ﬁeld-testing, and developed and

debugged a Microsoft-Access database capable of storing our interview data.

are deﬁned by a cutoff value. In a population with a normal distribution Of maternal age,
a cutoff of 1:250 will detect 59% of DS pregnancies with a false positive rate of 5%[7].
The quadruple test is designed to improve the screening detection rate over the

triple test. The quadruple test measures the sub-units of hCG (free a-hCG and free [3-
hCG), AF P and uE3l8]. This combination of four markers can detect 65% of DS

pregnancies with a false positive rate of 5%[7]. Women who screen positive on the triple
or quadruple test may chose to have a diagnostic amniocentesis or chorionic villus
sampling (CVS).

History
In 1912, the number of human chromosomes was ﬁrst reported as 48I9]. This

falsity held for over forty years until Tjio and Levan observed 46 chromosomesllol.

Later that year, Ford and Hamerton conﬁrmed that 46 was the true numberl1 1]. Today it
is known that 46 is the normal number of chromosomes for a human to possess.

In 1959, it was discovered that some deviations from 46 chromosomes were

compatible with survivalllzl. Lejeune et al. showed that trisomy of a small acrocentric
autosome was the cause of Down syndrome, a previously well—described syndrome.

Later, in 1960, trisomy 18 and trisomy 13 were described (Edwards and Patau syndrome

respectively)[13a 14].

Prevalence
The prevalence of trisomy varies at different stages of pregnancy. Trisomy is

more ﬁ'equent among the earlier stages of pregnancy with 26.8% of spontaneous

abortions, 3.8% ofstillbirths, and 0.3% of live births being t1isomicI15]. A total of 4.3%
of all clinically recognized pregnancies are t1isomicI15].

There is a large variation in the frequency of type of chromosomal trisomy
sampled at different times in pregnancy (Table 1). This inequality could either reﬂect a
difference in selective disadvantage or it could represent differences in the frequency of
non-disjunction among the chromosomes.

Upon examination of table 1, three main points are apparent. First, across the
different chromosomes the prevalence of speciﬁc trisomies among spontaneous abortions
varies greatly. The frequency ranges from zero for chromosomes 1 and 19 to 7.5% for
chromosome 16. Second, there is a large amount of selection that occurs before birth.
The only autosomal trisomies compatible with survival to term are 13, 18, and 21. Third,
even among these three trisomies, selective intrauterine mortality occurs. Only 3% of

trisomy 13, 5% of trisomy 18, and 22% of trisomy 21 pregnancies survive to birth.

Table 1: Frequency of chromosomal trisomy

 

 

 

Population
Chromosome Spontaneous Stillbirths Livebirths Probability of
abortions n=624 n=56952 survival to
n=4088 term“
1 - - - -
2 1.1 - - 0
3 0.3 - - O
4 0.8 - - 0
5 0.1 - - 0
6 0.3 - - 0
7 0.9 - - 0
8 0.8 - - 0
9 0.7 0.1 - 0
10 0.5 - - 0
11 0.1 - - O
12 0.2 - - O
13 1.1 0.3 0.005 2.8
14 1.0 - - 0
15 1.7 - - 0
16 7.5 - - 0
17 0.1 - - 0
18 1.1 1.1 0.01 5.4
19 - - - O
20 0.6 - - 0
21 2.3 1.3 0.13 23.8
227 2.7 0.2 - 0
Double 0.8 - - 0
Trisomy

 

[161*Assuming 15% spontaneous abortion and 1% stillbirth of clinically recognized
pregnancies

Parental Origin
It is possible to determine the parent in which the nondisj unctional event occurred

through the use of Menedelian inheritance patterns of genes, cytogenetic
heteromorphisms, or molecular polymorphisms. A cytogenetic heteromorphism is a
stable heritable alteration in size and shape of heterochromatic regions of certain
chromosomes. Cytogenetic heteromorphisms rely upon an individual’s ability to

interpret subtle differences at the limit of resolution of the light microscope. A molecular

polymorphism is a heritable DNA sequence that is highly variable in the population.
Molecular polymorphisms are a less subjective and easier method of tracing paternal
origin of non-disj unction. Examples of molecular polymorphisms include restriction
fragment length polymorphisms, very numerous tandem repeat polymorphisms (VNTRS)
detected with a Southern Blots, and GT repeat polymorphisms ampliﬁed with polymerase
chain reaction.
Origin of Error

It is also possible to determine the cell division of non-disjunction through the use
of molecular polymorphisms. The non-disjunctional error resulting in a trisomy can
occur in either the ovum, the sperm, or in early postzygotic division. The error can occur
during a premeiotic mitotic division of the oogonia or spennatagonia, during the ﬁrst or
second meiotic divisions of the oocyte or spermatocye, or during early division of the
zygote with a postzygotic mitotic (PZM) error. A premeiotic mitotic division cannot be
differentiated from a meiotic error because the additional chromosome will be paired and
segregated at future meiotic divisions. The cell division of error can be assigned by
determining the pattern of polymorphisms along the parents’ and trisomy’s non-disj oined
chromosome. A polymorphism is reduced when the non-disjunctional event results in an
individual that is homozygous for a single polymorphism. Figure one illustrates the
difference between a meiosis I and meiosis 11 error (MI and MII). If non-disjunction
occurs during meiosis, the trisomic conception will receive either two copies of the same
chromosome from a single parent (meiosis II non-disjunction) or two copies of different
chromosomes from a single parent (meiosis I non-disjunction). Centromeric and distal

polymorphisms distinguish whether the error occurred during meiosis or PZM.

Centromeric markers prevent misclassﬁcation of MI and MII errors because of crossing
over events which occur in distal regions of the chromosome. In order to differentiate
between MI and M11 errors, it is necessary to have a centromeric polymorphism that is
heterozygous in the parent whose chromosome is duplicated in the trisomy. For both a
PZM error and a MII error the polymorphism is reduced to homozygosity (T able 2). For
a MI error the polymorphisms is non-reduced. In order to distinguish a PZM error from a '
MII error it is necessary to look at polymorphisms that are distal to the centromere.
During a PZM error all polymorphisms will be reduced, for a MII error the distal

polymorphisms will be non-reduced.

Hmﬁmoaz RE 38.52
< x
3

H mmmomoz :cuogmmmeeoz

5695—3252 can £832 3 PEER

Table 2: Cell Division of Nondisjunctional Error

 

 

 

 

Status of polymorphism
Cell Division Centromeric Proximal Distal

MI

Nullichiasmate Nonreduced Nonreduced Nonreduced

Chiasmate Nonreduced Nonreduced Reduced
MII Reduced Reduced Nonreduced
PZM Reduced Reduced Reduced
Etiology

Extrinsic risk factors for chromosomal trisomy that have been previously
investigated include: exposure to ionizing radiation, oral contraceptive use, fertility drug
use, alcohol use, caffeine use, and cigarette smokingI17]. In addition, there has been a

focus on intrinsic factors such as thyroid autoimmunity, decreased parental HLA

heterogeneity, persistent nucleolar associations, and cytogenetic heteromorphismlwl.
The only clear and consistent risk factor for chromosomal trisomy is advanced maternal
age. One major problem with previous etiologic studies is that they failed to separate
maternal, paternal, MI and MII errors. Each of these different errors may be triggered by
different intrinsic or extrinsic risk factors.

Much of what is known about trisomy comes from one large epidemiologic

studyl1 8]. This study sampled women from hospitals in New York City and Honolulu.
The New York City cases were selected at hospital admission for spontaneous abortion or
fetal death from April 1974 to May 1984. Prior to 1981, patients were sampled from
three Manhattan hospitals. From 1981 to 1984 patients were sampled from only one
hospital. The total sample consists of 2,587 karyotyped cases with known maternal age.

The New York City sample has eight cases of inherited trisomy, 2,312 women with one

karyotyped abortion, 125 women with two karyotyped abortions, and seventeen women
with three karyotyped abortions.

The Honolulu data set sampled cases from one hospital from April 1976 to May
1985. A total of 2,921 karyotyped samples with known maternal age were available.
The Honolulu sample consists of 21 cases with inherited translocations, 2,594 women
with only one karyotyped abortion, 148 women with two karyotyped abortions, and 10
women with three karyotyped abortions. Multiple authors have analyzed data from these
two samples. The results from these studies will be presented and discussed in later
sections.
Maternal Age

In 1933, twenty-six years before the chromosomal basis of the disease was

known, Penrose discovered an association between advanced maternal age and an

increase in the risk for having a child with D3119]. In 1983, Hook et al published
population frequencies of DS by maternal age category based on data from prenatal
cytogenetics studies. They calculated regression-smoothed maternal age-speciﬁc rates of
DS abnormalities and multiplied them by a fetal selection coefﬁcient to adjust for the
excess likelihood of loss of cytogenetically abnormal fetuses. The result was estimated
maternal age-speciﬁc rates of DS in live-bom infants. These rates apply to women whose
only risk factor is advanced maternal age. Hook et al. found that the risk for DS increases
moderately in young mothers and much steeper after age 30 (Table 3). These rates were

calculated prior to the implementation of prenatal screening.

Table 3: The risk for Down syndrome by maternal age and stage of pregnancy.

 

 

 

 

Incidence
Maternal Age At CVS (9-11 At Amniocentesis At birth
(years) weeks) (16 weeks)
15-19 - - 1/1250
20-24 - - l/ 1400
25-29 - - 1/1 100
30 - - 1/900
31 - - 1/900
32 - - 1/750
33 - 1/420 1/625
34 - 1/325 1/500
35 1/240 1/250 1/350
36 1/175 1/200 1/275
37 1/130 1/150 1/225
38 l/lOO 1/120 1/175
39 1/75 1/100 1/140
40 1/60 1/75 1/ 100
41 1/40 1/60 1/85
42 1/30 1/45 1/65
43 1/25 1/35 1/50
44 1/20 1/30 1/40
45 and over 1/10 1/20 1/25
[20]

Penrose and others have concluded the increased risk of trisomy with age can be

separated into two components. One component increases in a linear fashion with

chronologic maternal age. The second component of the age effect increases in a

curvilinear with maternal age. Based on spontaneous abortion studies from the Honolulu

and New York data increased maternal age is a risk for trisomy for all autosomesl7-1'24].

However, the effect of maternal age varies by chromosome being more pronounced for

small chromosomes and less pronounced for larger ones. In addition, for chromosomes

16 and 2, the maternal age effect is strictly linear. Two trisomy studies of the cell division

of origin and the parent of origin found a maternal age effect for both maternal MI and

maternal MII errorsl25, 26].

10

Models for the maternal age effect
There are three popular models that have been suggested to explain the maternal

age effect. They are the “relaxed selection,” “Older egg,” and “production line,”
hypotheses.

The “relaxed selection” model suggests that the age—dependent increase in trisomy
is due to a decreased likelihood of aborting a trisomy and not an increased frequency of
trisomy at conception127l. This model has little support in the literature. It predicts a
maternal age effect regardless of parental origin or stage of error. However, a study of

trisomy 21 in 1992 showed that the increase in maternal age effect was present in cases of

maternal not paternal originl26l. Another molecular study of trisomy 21 found that

errors of mitotic origin showed no increase in maternal age effect regardless of parent of
originlzs]. In general, the increase in maternal age effect is restricted to cases involving
maternal meiotic non-disj unction129]. Also, the miscarriage frequency of trisomic
fetuses increases with maternal age and the miscarriage frequency of fetuses with other

types of chromosome imbalance shows no relation to maternal agel3 0].
The ‘Older egg’ theory suggests that the maternal age effect is related to a

declining quality of oocyte pool. This hypothesis is supported by the fact that DS cases

resulting from translocated chromosomes do not Show a maternal age effectl31: 32].
According to this hypothesis, factors that affect the availability of oocytes should affect
the risk of trisomy. For example, data that show that unilateral ovarectomy is a risk
factor for trisomy support the ‘Older egg’ hypothesisl33]. If the oocyte pool is reduced
due to unilateral ovarectomy, then the risk of a nondisj oined oocyte becoming fertilized

increases. If reproductive age is viewed as a continuum from menarche through

11

menopause, then early onset of menarche and menopause may be indicators of an
increased risk for chromosomal trisomy. A retrospective case-control study in 1995
hypothesized that a woman who has a child with trisomy 21 at younger than 30 years of

age would be more likely to undergo premature menopause (menopause at less than 35

years of age)[34]. In addition, they suggested that women over 30 who had a child with
trisomy 21 would be closer to menopause than an age matched control that had a normal
child. They analyzed data from interviews. They found no cases of premature
menopause among 35 women who had delivered trisomy 21 children under 30 years of
age. Also, they found no difference between the mean age of menopause among 106
case and control women over 35. This study did not support the predictions of the ‘older
egg’ hypothesis.

The “production line” hypothesis postulates that those oocytes produced last in
fetal life would form fewer chiasmata, making nondisjunction more likely[3 5]. They
would ovulate later in adult reproductive life than those oocytes produced earlier in fetal
development. The literature indicates that reduced recombination may play a role in
nondisjunction. Three investigators have reported decreases in recombination in the non-

disj oined chromosomes through the use of DNA polymorphic markers along the

chromosome [3 6'3 8]. Sherman found that older mothers had fewer recombinational
events in the non-disj oined chromosome than younger mothers. According to Sherrnan’s
work, reduced recombination seems to play an important role in trisomy 21 non-
disjunction especially for young mothers. Currently the “production line” hypothesis has

the most support in the literature.

12

Paternal Age
Paternal age has been extensively studied as a risk factor for chromosomal

trisomy. In 1977, Stene et al reported an increased risk of DS in fathers over the age of
55139]. The result was supported by similar ﬁndings of Matsunaga et al. (1978) and
Erickson and Bjerkdal (1981 )140, 41]. In addition, Stene et al. in 1981 found an
increased risk of trisomy 21 in fathers over age 41 based on amniocentesis datal42].

Other studies have found no link between DS and increased paternal age [43'501. All of -
these studies did not separate maternal, paternal, MI and MII errors. This is important
because, one would not expect to ﬁnd a paternal age effect among maternal errors. At
this point in time the weight of evidence suggests that paternal age is not an important
factor for chromosomal trisomy, and only a small proportion of trisomy nondisjunctional
errors are paternal in origin. It is necessary to wait for larger numbers of cases that have
been identiﬁed via DNA polymorphisms to be paternal errors before any conclusions can
be made as to the contribution of paternal age.
Recurrence Studies

Studies that quote recurrence risks hint that trisomy may not be a purely random
event (Table 4). A number of studies have looked at data from live births. Initial studies

by Oster and Carter found that hospitalized DS patients had a higher than expected

number of siblings with DS (compared to population data)[51: 52]. They did not exclude
translocation cases from their analysis. Stene reanalyzed this data in 1970 excluding
translocation cases. This analysis found an increased recurrence risk for mothers under

30, and the population risk for mothers over 30. Richards repeated this ﬁnding in 1977 in

sibships of institutionalized DS patients[53l. Data from trisomy 18 and trisomy 13 was

13

collected by Baty in l994l54]. Among families in a support group for trisomy 18 and 13
the risk for recurrence of trisomy 18 or 13 among siblings was not increased. This study
found a recurrence risk of 0.55% (95% CI 0-1 .63%) but had limited power to ﬁnd a
difference due to small sample size.

Mikkelsen and Stene looked at data from multiple European Centers and found

that mothers below 25 had a recurrence risk signiﬁcantly greater than the population

riskl55]. They did not indicate reason for amniocentesis among these women or separate

out translocation cases. Daniel looked at amniocentesis data from women who had an

amniocentesis performed because of a previous child with DSI56]. They found an
overall recurrence rate of 1% but did not specify age-speciﬁc rates.

Caron in 1999 looked at amniocentesis data for women referred for amniocentesis

due to advanced maternal agel57]. They found a risk of recurrence of trisomy of 1.3%
for women under 35 and 4.8% for mothers 35 or older. This recurrence risk is one and a
half times the population risk for women under 35 and over twice the population risk for
women 35 or older.

The strengths of this study were its comparisons to several reference groups. In
addition, the authors looked at other chromosome abnormalities than trisomy. The
sample came from tissues over a long period of time. The weaknesses include that the
population was sampled at hospital admission. Therefore, the population may over-
represent trisomies which present with later fetal loss and under-represent trisomies
which present with loss earlier in pregnancy. Also, the authors didn’t separate out the
maternal, paternal MI, and MII errors. Therefore, this study does not Show if one speciﬁc

type of non-disjunction could be genetic.

14

Other studies have looked at data from spontaneous abortions. In initial studies, a
total of 87 women with two karyotyped spontaneous abortions were looked atl53‘601

[61]. Among women who had a trisomic abortion the second abortuses tended to have a
trisomic karyotype as well. However there were problems with this data. Women with
their ﬁrst trisomic abortion are on average older than women with their ﬁrst normal
abortion and older at their second karyotyped abortion. This leads to an apparent
increase in the rate of trisomy when the comparison group is women whose ﬁrst abortion
had a non-trisomic karyotype. Also, women with a previous chromosomally normal
spontaneous abortion have a lower rate of trisomic abortions than do unselected women

and do not make a good comparison group.

These problems were addressed by a study by Warburton et al., in 1987l52].
They looked for an association between the karyotype of a previous spontaneous abortion
with the karyotype of subsequent spontaneous abortions in the New York City and
Honolulu data. The data were analyzed by city and combined. The authors performed an
unconditional and a conditional maximum-likelihood logistic regression analysis. They
adjusted for the potential confounders of maternal age, payment status in the New York
City sample (private versus public facilities), prior abortions, and location (New York
City versus Honolulu) in the combined analysis. The authors used all women with only a
single karyotyped abortion irrespective of reproductive history as the reference group.
They repeated their analysis using two other reference groups: (1) women who were
primigravida at the time of the ﬁrst abortion and (2) women who had a prior term
delivery but no prior spontaneous or induced abortions. The authors were concerned that

their ﬁrst reference group may have over-represented women at high risk for spontaneous

15

abortion who are known to have an increased rate of chromosomally normal abortions.
All three analyses yielded Similar results.

The authors deﬁned the ‘index abortion’ as the second karyotyped abortion for
women with two karyotyped abortions and the only karyotyped abortion for all other
women. In the adjusted analysis, the odds of trisomy at the index abortion among women
with a previous trisomic abortion were similar to those among women without a previous
karyotyped abortion. The combined estimation for the odds of trisomy at the index
abortion relative to prior trisomy abortion was 1.3 (95% conﬁdence interval 0.7 to 2.1).

The authors performed separate analyses for women under thirty and for women
greater than or equal to 30 years of age. This analysis did not Show an increased risk for
women in the younger age category. The adjusted Odds ratios were 1.3 (95% conﬁdence
interval 0.4 to 4.5) for the under thirty women and 1.2 (95% conﬁdence interval 0.7 to
2.1) for the women who were thirty and older. Since the risk of trisomy increases with
age, the authors had small numbers of women (n=1 7) in the under 30 group and limited
power to ﬁnd a 20-30% increase in risk. The results indicate that karyotype of
spontaneous abortion is not a good predictor for future trisomy. The authors suggested
possible explanations for the disconcordant ﬁndings as compared to live birth and
amniocentesis data. The authors suggested that they may have found no association
because trisomy proneness could be conﬁned to certain trisomies or only women under
30. With an effect so restricted, this study would not have had the power to ﬁnd an
association. A second possibility they proposed was that the increased recurrence rate
among live births and amniocentesis data is due to parental mosiacism. Thus trisomies

which were compatible with survival would appear to have an increased recurrence risk

16

among livebirths. The current literature does not rule out either of these two possibilities.
In addition due to the rarity of trisomy it is difficult to ﬁnd studies with enough power to

answer these questions.

17

2 >885 00 8 >885 00.5
85 000080000 m838 $3.0

08G 0080008000
800508-880

00.5580
2 035 >885 00.5 0008

80008 0 8 80588

.82 .08

 

8m>88
8 508808 £34.. .50 080 00008600 0 500 300500 mm
500 8800 800500 ma .«0 080 0.8508— 005 0800005300 5088005388 08858
005800888. 05 cm 00500 80508 w008< 800508-880 8 mm .50 00008000M 1.35 858803
8800 00>0 58 cm 80508
50800885 00.5 0.05 00580000 .2.” 00500 085 m8>m 800080
005285 80508 00.5 085 508208 .0080 .50 80>880M dong .0005
.0800 mm 00500 80508 00.5
005000885 38 .820 do 00w000> 5050098
058008 500 05 85 0.88 858058me 800508-880 800500 mm .50 8855 .203 £83m a. 08.80
8000
0050020005 350 805 .88— 58 005.80
058008 500 55 000800 5050098 85 005.8 800508-880 800500 ma .50 8835 .mmﬁ .880
.0802 $850K 00.82 0889 0058000m >505

 

00008000.. >885 50 0.8.. 05 08.000 5053 80508 588— 0058090m ”v 030,—.

18

:2-.. $2-0 002.0 002-0 00.3.0 $2-0 000-0 0.0.2-0 000-0 :0- 0 M090 0000

00500 00 mm 80508 00.0 80.0

 

 

500 mm 00500 00803 00.0 gm; 000 8000008
>885 0 55> 0000.0 00 5:00 _ 5000050 00 005 50000000
0050 0 803055 0000000000 000050880000 00803 00.0 0009 000020 .0000 .0080
00.8w 000 5500 mm 0005600
00000 0000000000 000G 808< 00 005 000008005 000000000
050000-80V 0000 0000000000 900 50005 0000500080000 00.0 50000000 00803 amm— .00000Q
0800 00000 0005000000 05 085
005000000000 0000000 800000-0880 mm 00.0 005 0508<
505885 0000000000 mm 30—05 80502 0000500080000 >500m 0000005002 0.0000 08000552
0050050 00885 00000850 00000000
000000000 00.0 00000 0000008 0 Z 0>50000000 00 0050000 0000000000m 0500 0000083
0000 00 - 0.030
080 0 - 800:3
8885 05 00 505000 8000 mm - 0080050.
0000050 50m 05 .0885 003 0800 m0 -000m 8 000m 00009
0000050 000 05 .00 85 50080.0 800508-8000 00050000 5080000 .00 880 00000500 50000000
0000 Z 008580 00.002 088G 005050000 >500ml

 

0000000000 >885 00 0000.0 05 002000 0053 0005000 50005 0050—00m "50005000 0. 030,—-

19

 

Trisomy 21 -parental origin
Trisomy 21 is the most common trisomy at birth. Chromosome 21 has

heteromorphisms located at the centromere and close to the centromere on the short arm.
Very little recombination occurs on the short arm so it is possible to extrapolate the origin
of the extra chromosome. Early studies looked at over 1500 families using chromosome
heteromorphisms. Problems with this method included: the subjective evaluation of size
and staining intensity of bands, large numbers of uninformative families, and the
heteromorphisms were located on only one side of the centromere so crossovers between
the centromere and the short arm went undetected. Based on heteromorphism studies, the
observed level of paternal non-disjunction ranged ﬁom 0-57%.

It became clear that the estimates of ”non-disjunction ﬁ'om chromosome
heteromorphisms were not reliable. However, two pieces of valuable information were
learned. First, most of trisomy 21 occurs from non-disjunction events occurring at
maternal MI. Second, some paternal and MI, paternal MII and maternal MII errors
occur. DNA polymorphisms that have been identiﬁed near the centromere on

chromosome leuggest that 91% of the nondisjunctional errors leading to a trisomy 21

conceptus are maternal in origin126a 38].

Trisomy 21-stage of error
Among the 500 maternal errors that have been classiﬁed, 75% occur during

meiosis I, 22% during meiosis II, and 3% during PZM (Table 4)[64]. Among 30 paternal
errors classiﬁed, 50% occurred during meiosis II, 23% during meiosis I, and 27% during

PZM.

20

Table 5: Proportion of meiotic errors for trisomies 21, 18, 16, and 13.*

Trisomy 21 Trisomy 18 Trisomy 13 Trisomy 16
Proportiona Proportionb Proportionc Proportiond

 

 

Cell Division

 

Paternal MI 13% 0 12% 0%
Paternal MII 7% 0 4% 0%
Maternal MI 68% 29% 68% 100%
Maternal MII 13% 62% 16% 0%

 

WM errors occur less than 5% of the time for all trisomies
a.[25]samp1e size = 500
b [6518ample size = 63
c-[66lsample size = 30
d-[67lsample size = 62

Trisomy 21-eﬁect of recombination
In 1987, Warren showed that the level of recombination was reduced along the

trisomic chromosome among trisomy 21 casesl36]. Sherman repeated this result in

1991 [3 3]. The unit of genetic map distance is the Morgan. The Morgan is deﬁned as the
length of choromsomal segment which on average undergoes one exchange per
individual chromatid strand. Sherman found that the average genetic map at maternal
MI is 39 centimorgans for a trisomic 21 versus 72 centimorgans for a normal 21. In other
words, mothers of children with trisomy 21 experience far fewer recombinational events
on chromosome 21 than mothers with non-trisomic children

' Trisomy 18-Epidemiology
Trisomy 18 is the second most common autosomal trisomy among live births.

There is a strong association with maternal agelzl]. Recent molecular studies of live
births and abortus tissue indicate that 87-95 % of trisomy 18 occurs as a result of a

21

There are no polymorphic centromere markers available for chromosome 18.
Therefore, the stage of cell division for nondisjunction must be determined through the
use of pericentromeric markers. There may be some abnormal recombination among

trisomy 18 cases. Fisher found that one third of maternal MII errors were associated with

absence of recombination. The rest appeared to be normall68].

Trisomy 13-Epidemiology
Trisomy 13 is compatible with survival to term. In 1987, Jacobs et al. presented

data on the trisomy 13 cases from their Honolulu sammel69]. Trisomy 13 was the fourth
most common trisomy in their sample. The mean maternal age for the non-translocation
trisomies in the Honolulu sample was signiﬁcantly greater than that for the whole study
population (t=3. 14, p<0.05).

By using both cytogenetic and molecular techniques, Hassold et al., analyzed the

parent and cell division of error in 30 cases of trisomy 13 from their Honolulu

samplel66]. They were able to determine the parent in which the error occurred in 20
cases with 17 (85%) being maternal and three (15%) being paternal in origin (Table 3).
The most common mechanism of origin was maternal MI non-disj unction that accounted
for 68% of cases. A trend towards increased maternal age was seen for the maternal MI
and MII errors but not for the paternal errors. This suggests that increased maternal age
is a risk factor for trisomy 13. The authors were unable to determine if recombination was
reduced or enhanced.

Trisomy 16-Epidemiology:
Trisomy 16 is the most common trisomy in humans. It occurs in over 1% of

clinically recognized conceptionsl64]. Trisomy 16 conceptuses rarely survive to term.

22

The risk of trisomy 16 increases linearly with maternal age[24s 70]. Parean origin has

been determined in 62 trisomy 16 casesl67]. In all cases the additional chromosome was
maternal in nature. The stage of error was studied in 58 trisomy 16 conceptuses. A
single centromeric marker was informative in 54 cases and all were due to a maternal MI
error. Preliminary data suggests that trisomy 16 is associated with a reduction in
recombination. In addition, this reduction is restricted to pericentromeric regions with

the distal portions having normal amounts of recombination.

Table 6: Summary of trisomies 21, l8, l3, and 16 information

 

 

Trisomy 21 Trisomy 18 Trisomy 13 Trisomy 16
Syndrome Down Edward Patau syndrome None
syndrome syndrome
Frequency Most common Second most Third most Most common
Rank at birth common at common at during
birth birth pregnancy
Recombination Reduced Reduced! ? Normal
Normal
Maternal Age Curvilinear & ? ? Linear Effect
‘ Linear Effects Only

 

23

Summary of Part One

It has been 40 years since the discovery of the ﬁrst human trisomy.

Trisomy contributes signiﬁcantly to pregnancy loss.

There is a large variation in the frequency of different types of chromosomal trisomy
sampled at different times in pregnancy.

Through the use of molecular markers, it is possible to determine the parent in which
the nondisjunctional event occurred and the cell division of error.

Maternal age is the only well documented risk factor for chromosomal trisomy.

24

CHAPTER 2: APOLIPOPROTEIN E

Background
Apolipoprotein E (ApoE) is a 299 amino acid plasma glycoprotein involved in

cholesterol transport and metabolism. ApoE is synthesized mainly in the liver but also in
small amounts in most organs including the brain and ovaries. Three different alleles

give rise to the three most common isoforms: E2, E3, and E4. The 83 allele is the most

common form among whites with an allele frequency of 78.5% while 84 and 22 have

allele frequencies of 13.5% and 8% respectivelyl71]. The frequency of the 84 allele
varies among population and has been found to be higher in particular Aﬁ'ican (~20-40%)
[72], Finnish (~20%) [73, 74], and Swedish (~20%)[75] populations, and lower among
several Asian populations (~8%)[76]. ApoE genotype can be determined by polymerase
chain reaction (PCR), restriction enzyme digestion, and gel electrophoresis. The three
isoforms have variations in sequence that results in differing locations of Hha I restriction
sites. Each digested DNA sequence results in a unique restriction fragment pattern (see
Figure 1). ApoE genotype has been investigated as a risk factor for numerous health
conditions including longevity, cholesterol level, cardiovascular disease, stroke, recovery
from head trauma, presence of gallstones, hip fractures among the elderly, and retinitis

pigrnentosa.

25

Figure 2: Hha I Restriction pattern of different ApoE genotypes
4/4 3/4 3/3 2/4 2/3 2/2

91 bp — — -— —- --
83 bp — _ __
72 bp — — . —
48 bp — — — — —
35 bp — — — -— —
bp=base pairs
Longevity Studies

In a study of 325 French centenarians, the 84 frequency was decreased to 5.8%
compared to 12.1% among controls and the 82 frequency was elevated to 12.8 % among
the centenarians compared to 6.8% among controlsl77]. Similar ﬁndings were seen
among 179 centenarians and 95 nonagenarians in Finlandl78: 79], in healthy Swedes over
60 years oldl75], among American femaleslgo], and Asian and Italian subjects[81'83].
The limitation of these case-control studies is that they do not tell us why 32 carriers

more frequently survive to very old ages and s4 carriers do not. Unlike ApoE, common

polymorphisms in other genes involved in lipoprotein metabolism, thrombosis, or

homocysteine metabolism have not been consistently associated with longevity134'36].

Cholesterol Levels, Cardiovascular Disease, and Stroke
In addition to being associated with longevity, 82 is associated with decreased

levels of total cholesterol and low density lipoprotein (LDL), and s4 is associated with

increased levels of total cholesterol and LDL. Alleles 82 and 84 are also associated with

26

increased and decreased plasma ApoE levels, respectively. A 1996 meta-analysis of nine
studies found that 24 was associated with a mild increased risk for coronary heart disease
(CHD) (odds ratio =1 .26 versus reference e3)[87]. On the other hand, a 3.5 year
prospective study of 1067 elderly F inns failed to ﬁnd an association between the 34 allele
and CHD188]. A ﬁve-year study of 666 elderly Finnish men found a twofold increase in
the 84 allele frequency among those who died from CHD[33].

A 1999 meta-analysis looked at the association of 84 with cerebrovascular disease
or stroke among nine published studies. The 84 allele was associated with an increased

risk with an odds ratio of 1.68 compared with 83. In summary, 84 is associated with a

more atherogenic lipoprotein proﬁle and moderately increased risk for CHD and stroke.

Other Associations
ApoE genotype has been investigated as a risk factor for recovery from head

trauma, presence of gallstones, hip ﬁactures among the elderly, and retinitis pigrnentosa.
In two studies of head trauma, 34 was a negative risk factor for recovery [89, 90] and 84

is associated with an increased and 32 with decreased prevalence of gallstones in

womenl91a 92]. In addition, 84 may be a risk factor for injury in the elderly. In a 7-year

longitudinal study of 1750 women over 65 e4 carriers had higher rates of bone loss and
were at increased risk to have hip ﬁacturesl93]. Finally, homozygosity for 82 or 64 has

been associated with having retinitis pigrnentosal94a 95].

Link to Alzheimer's Disease
Alzheimer’s disease (AD) is the most common form of dementia after age 40.

Prevalence increases from 0.3% in 60 to 69 age category up to 10.8% after age 80 [96].

27

Differential diagnosis is made at autopsy. AD is characterized by the presence of
neuroﬁbrillary tangles composed of hypophosphorylated tau in the neurons of the
cerebral cortex and hippocampus along with the deposits of B—amyloid within senile
plaques and cerebral blood vessels. Clinically, patients experience a slow progressive
loss of memory and cognitive abilities. There is a genetic predisposition to AD
demonstrated by an increased prevalence in ﬁrst degree relatives of AD subjects.

Other than age, ApoE genotype is the strongest established risk factor for AD.
According to a meta-analysis in 1997, compared to 83/83 subjects the odds for having
AD among whites is 3.2 for 83/84 subjects and 14.9 for 84/84 subjectsl97]. On the other

hand, the 82 allele is protective. Among whites the odds for having AD is 0.6 among the

82/83 and 82/82 carriers as compared to 83/83 genotypesl97]. ApoE 84 is also associated
with the severity of AD. Compared with non-84 carriers, AD subjects with an 84 allele
have an increased number of senile plaques, increased brain B-amyloid levels, decreased
entorhinal cortex volume, decreased choline acetyltransferase activity, and increased
neuronal degeneration in the basal nucleusl98]. In a study of newly diagnosed AD
patients 84/84 patients had the most rapid decline of cognition, while 82 carries had the
slowest ratel99]. The mechanism by which the different allelic forms of ApoE affect the

pathology of AD is not understood although it may have to do with differential binding to

the proteins of the neuroﬁbrillary tangles.

Apolipoprotein E, Alzheimer’s Disease, and Down Syndrome
AD and DS have been linked together in several ways. First, adults over 40 with

D8 are more likely to develop symptoms of AD and have the same neuropathological

28

lesionslloo]. The similarity of brain lesions could suggest that the underling pathogenic
pathways leading to AD and DS may have some features in common and perhaps could
be caused by the same genetic risk factors.

Epidemiologic data also supports the idea of shared etiologic or pathogenic

factors for DS and AD. In one study, women who had a DS child before the age of 35

were at an increased risk of developing ADIlOl]. Furthermore, it has been shown that

among ﬁrst-degree relatives, there is an increased prevalence of AD in families with DS
relatives [102, 103] and the prevalence of DS is higher than expected among the relatives
of AD patients.[104: 105]. Other studies, while not statistically signiﬁcant, have results

- that point towards a higher rate of DS in the families of AD patients[106s 107]. There
has been some suggestion that mothers that give birth before 19 years of age are at an
increased risk for having a DS child and for AD.

A third line of evidence that supports a link between AD and DS is clinical
evidence. Fingerprint dermatoglyphic patterns observed in AD patients are similar to DS
patients. AD patients much like DS patients have an increased frequency of ulnar loops
on ﬁngertips, Simian creases on the palms, pahnar hypothenar patterns, and large distal

loops in the hallucal region. This similarity may be restricted to early onset AD patients

[108, 109]. These clinical similarities suggest that common genetic factors inﬂuence the
developmental processes in DS and AD.

The evidence for a genetic risk factor for AD linked to chromosome 21 has been
variedllOS]. An initial study linked chromosome 21 to familial AD in four AD families.
In addition, at the same time it was detected that the. gene for B-amyloid precursor protein

(APP) maps to chromosome 21. This supported the theory that APP was one of the genes

29

for AD, and the AD-like symptoms of DS patients were explained by the extra dose of
the APP gene. However, other studies did not ﬁnd an association between chromosome
21 and AD. When divided into early-onset and late-onset cases of AD, the association,
although not universal, was strongest with early-onset cases. The interest in the
association between chromosome 21 and AD later decreased after sequencing of the APP
exons in AD affected individuals revealed that mutations in APP were very rare and
explain only 1-3% of familial AD cases.

Other evidence suggests that the similarities between AD and DS are not due
solely to over- expression of the APP gene located on chromosome 21. Although APP is

over-expressed in some tissues from DS patents, substantial variability exists in the B-

arnyloid deposition within DS patients from the same age groups[105]. Not all DS
patients develop AD-type dementia although B-amyloid deposits are found in the brain at
autopsy. In addition AD-type neuropathology is detectable in the brains of DS patients by
age 35 while the average of onset of clinical dementia is between 51 and 54 years (range
39-69years). These ﬁndings suggest other factors may be contributing to the severity and
timing of B-amyloid deposition and that the accumulation of B-amyloid is not enough to
develop AD-type dementia.

A study by Avramopoulos in 1996 proposed that since AD and DS have many

similarities and ApoE 84 is associated with AD, perhaps ApoE 84 was also associated

with DS[1 10]. The authors found that ApoE 84 was signiﬁcantly more common among
young mothers of DS children. This correlation was speciﬁc to MII errors. They
theorized that ApoE 84 may predispose an indiVidual to chromosome non-disj unction and

potentially to trisomy 21 mosaicism and AD. Individuals with AD have been found to

30

have increased numbers of cells trisomic for chromosome 21 in their circulation“ 11]. In

1996, Potter suggested that the correlation is speciﬁc to MII because MII most closely

resembles mitosisl1 12]. During MII and mitosis, centromeres divide and separate and
correct chromosome segregation depends on maintaining a balanced bi-directional

tension on each pair of kinetochores.

Summary of Part Two:

0 ApoE is a glycoprotein involved in cholesterol transport and metabolism.

0 ApoE genotype has been associated with many health conditions including, but not
limited to, longevity, cholesterol level, coronary heart disease, stroke, gallstones, hip
ﬁactures among the elderly, retinitis pigmentosa, and AD.

0 Age and ApoE genotype are the strongest established important risk factors for the
development of AD.

0 AD and DS have similar pathologic, clinical, and epidemiological ﬁndings which
support the existence of a underlying genetic link.

0 AD and DS have been associated in a recent study with Apolipoprotein E 84.

31

CHAPTER 3: LESSONS FROM A FEASIBILITY STUDY: An
Investigation of Apolipoprotein E as an hereditary risk factor for
non-disjunction and Alzheimer's disease.

Rationale and Specific Aims
There are two main reasons to study a potential link between ApoE 84 and

chromosomal trisomy. The ﬁrst is to increase the understanding of the mechanisms
leading to trisomy and the potential for preventive measures. The second reason to is to
look for risk factors to provide more precise genetic screening and counseling regarding
the risks of trisomy in offspring. We hypothesized that the genetic factor(s) which have
been shown to link AD and DS actually link AD and MII non-disjunction in general.
Therefore, the association with ApoE e4 should apply to all types of chromosomal
trisomy. Previous studies have failed to combine epidemiologic data about family history
of disease, ApoE genotype, and data about non-disjunctional stage of error to describe the
risk factors for trisomy. We propose that the information provided by ApoE genotype
and family history of AD could be used in a general population to augment present
screening protocols for trisomy.

To answer these two questions we would need to build on the previous research
by Schupf and Avramopoulos who formd associations between AD and DS. We would
expand the ApoE hypothesis to MII non-disjuntion in general in a study that would
incorporate the following speciﬁc hypotheses and aims:

Hypothesis 1 : The prevalence of AD is increased in trisomic families.

Speciﬁc Aim 1: We will compare the prevalence of AD in the families of women
with a history of trisomy (cases) and with no history of trisomy (controls).

32

Hypothesis 2: ApoE 84 is more prevalent in young mothers of trisomy
pregnancies than in controls.

Speciﬁc Aim 2: We will compare the 84 gene frequency in case mothers and
controls under 35.

Hypothesis 3: The association with the ApoE 84 allele is speciﬁc to maternal MII
errors.

Speciﬁc Aim 3: We will compare the frequency of the ApoE 84 allele in each
group (Mat MI & MII, Pat MI & MII)

Feasibility Study Goals
Prior to launching a large-scale study we chose to do a feasibility study with the

following goals:
1) Develop an interview which could be used to collect demographic and health

information ﬁ'om case (trisomy positive pregnancy) and control (trisomy negative
pregnancy) women

2) Field test a sample collection protocol and laboratory assay that could identify
Apolipoprotein E genotype

3) Test the potential of using the MSU Prenatal Screening Program as a population
of cases and controls.

4) Identify potential strengths and weaknesses with our case-control study design

Study Design
Our feasibility work was a case-control study. For this initial study, a case-

control design best suits our hypothesis because our exposure variable is a genetic risk
factor and trisomy is a rare disease. Our case deﬁnition was women who: 1)had
experienced a karyotype conﬁrmed trisomic pregnancy and 2)were identiﬁed by the
MSU Genetics program during the years 1995 to 1997. Our controls were matched to
cases on ethnicity, and frequency matched on age category. We chose to match by
ethnicity to control for potential confounding and maternal age in order to have sufﬁcient

age-speciﬁc strata for our analysis. We did not have any non-Caucasian controls in our

33

feasibility study. Controls consisted of women who had not experienced a trisomic
pregnancy and were ascertained by the MSU Prenatal Screening Program during the
years 1995 to 1997. Third variables that we planned on including in our analysis were:
parity, trisomy type, parent of error, and cell division of error. Parity is important since
the more pregnancies a woman has, the more opportunities she has to have a trisomic

pregnancy. In addition, higher parity was found to be a risk factor for DS in a recent

study1113]. The authors found a 15% higher risk for DS with both age and parity
considered above the age related risk. However, they did not take into consideration that
higher parity is associated with a negative attitude about termination.

Study Sample

Potential cases for the Trisomy Project were identiﬁed through three sources; the
MSU Prenatal Screening Program, the MSU Cytogenetics Laboratory, and the MSU
Genetics Clinics (see ﬁgure three).

Our case population from the MSU Prenatal Screening Program consisted of
mothers who were screened for maternal alpha fetoprotein (AF P) during the years 1995
to 1997 and who indicated a history of a prior trisomic pregnancy on the test requisition
form, had a positive screen for trisomy which was not a false positive, or a negative
screen which was later found upon follow up to be a false negative. For these cases, the
mother’s, father’s, and (if living) child’s DNA had to be self-sampled and mailed to MSU
laboratories.

The MSU Cytogenetics Laboratory case population consisted of abnormal
pregnancy material sent to the laboratory for testing during the years 1995 to 1997. For

these cases, the fetal/child DNA had already been collected by the lab, and the parent

34

DNA samples were self-sampled and delivered by mail. Three types of biologic samples
were available through the MSU Cytogenetics Laboratory, amniocentesis ﬂuid, abortus
tissue, and peripheral blood. Amniocentesis is performed primarily for reasons of
advanced maternal age, prior trisomic pregnancies, other family history of chromosomal
abnormalities, unusual ﬁndings on ultrasound, or elevated AF P or DS risk or triple test.
Blood is drawn from a live birth to perform chromosome analysis to rule out the
diagnosis of chromosomal trisomy if a child’s features are suggestive of trisomy.
Karyotyping is performed on abortus tissue primarily when there is a history of multiple
spontaneous abortions. Compared with our other sources for cases and controls, the
MSU Cytogenetics population is the least representative of the general population of
pregnant women.

The MSU Genetics Clinics’ cases consisted of families who came to MSU for
prenatal or informational counseling during the years 1995 to 1997. For these cases,
DNA ﬁ'om the mother, father, and fetus/child was collected at the time of counseling.

The majority of these patients overlapped with the MSU Cytogenetics program.

35

ﬂammam

EEU o>3

 

sense 25 25 seem

mﬂmmucou .8502

 
 

 

 

 

 

:96an 838mm
«>3

 

 

 

 

 

 

 

@005
83950
H.850
mommﬂi gem
83230
BEBE Bog
8320 £50.50 moﬂoaowognv mama—030m Emaobm

830 he EMEO an charm

36

The controls were selected ﬁom the MSU Prenatal Screening database during the
years 1995 to 1997. Of the populations we used for case ascertainment, the Prenatal
Screening Program most closely represents the general population of pregnant women.
Controls were frequency matched to cases based on maternal age at their estimated date
of conﬁnement (EDC). Case ages were calculated at the time of karyotype analysis for
abortus tissue, at EDC for amniocentesis samples, and at delivery of child for livebirth
cases. Maternal age was divided into six categories: below 19, 20 to 24, 25 to 29, 30 to
34, 35 to 39, and above 40. The ethnicity and age of the case mothers was unknown until
after the interview for cases from MSU Cytogenetics and MSU Genetics’ clinics. The
ethnicity and age of the mother was known prior to interviewing for the control mothers.
Protocol for Contact of Cases and Controls

Case women were selected for participation in chronological order beginning with
cases ascertained in 1995. One of three letters was generated for each woman based on
the method of ascertainment. Letters were mailed to the most recent address available.
Control women were randomly selected from all women screened during the years 1995
to 1997 by the MSU Prenatal Screening database and their letters were mailed to the
address on the laboratory requisition form.

Letters sent to eligible women stated that “we are conducting a study on the
causes of chromosomal abnormalities.” They were informed that at the time of the
interview we would be requesting a DNA sample. No mention of Alzheimer’s disease or
our hypothesis was included in the letters. Women were given a letter to return which

had two options: to request not to be in the study, or to inform us of their new telephone

37

number. For the case women we listed the most recent telephone number available. For
the control women, we used listed telephone numbers available at
www.5witchboard.com. Women who wished to participate in the study and for whom we
had updated telephone information were not required to return the letter. We later
included an option of returning the letter with an indication of preferred times to be
contacted for the interview. We also included a 1-800 phone number for them to call to
update a telephone number or to ask questions. The letter instructed women that we
would be calling them in two weeks.

Interviewing Methods

Four different interviewers were used. Interviewers were undergraduate students ‘ .
in their ﬁnal year of the MSU zoology bachelors in science four-year program.
Interviewers were not masked to the hypothesis and practiced administering the interview
with volunteer women not in our study population.

Two weeks after mailing the letter to our study women, our interviewers began
contacting all women who had not refused to participate. Our twelve-page interview was
identical for cases and controls and took approximately forty-ﬁve minutes to administer.
Each woman was asked at the beginning of the interview if she had a pregnancy with a
chromosomal abnormality. Control women who stated that they did have a pregnancy
with a chromosomal abnormality were allowed to become part of the case population.
We allowed this cross-over because rrrisclassiﬁcation of women into case and control

categories would cause us to calculate the genetic risk incorrectly.

38

Interview Content
The interview was broken down into ﬁve sections (see appendix). Section I

collected basic demographic information about age, race, education, and occupation.
Section II collected health information about the woman’s biological mother, father,
maternal grandparents, and paternal grandparents. We included any conditions that have
been associated with Apolipoprotein E in the literature, as well as conditions that may be
associated with premature aging. We asked about age and cause of death for each family
member as well as a list of medical conditions that included high blood pressure, stroke,
heart attack, high cholesterol, diabetes, thyroid disease, Parkinson’s disease, Alzheimer’s
disease or senile dementia, and premature graying. For the female relatives we asked
about age of menopause. In Section II we recorded information about family history of
chromosomal abnormalities including trisomy. Finally, we asked about history and
Alzheimer’s disease or dementia among biological great grandparents.

Section III asked the woman to answer the same list of health questions for herself
with the addition of a question about what age menstruation began and if she ever had
one ovary removed. Section IV collected details about reproductive background
including history of fertility problems, use of hormonal birth control methods, and
pregnancies with chromosomal abnormalities.

DNA Sample Collection

Following the interview, we requested a DNA sample ﬁ'om the mother, father,
and, if living, trisomic child. When a subject agreed to donate a sample, a collection kit
was sent through the mail. Each kit included two cytology brushes for each participant,
an informed consent sheet, an instruction sheet for collecting the samples, a postage paid

return envelope, and the 1-800 number to call with any questions. Participants were

39

instructed that the results of the testing would be conﬁdential and not available to them.
Due to limited resources, women who did not return their collection kits were contacted a
maximum of one time to remind them about the study. The reason for not returning the
collection kits was not recorded. Anecdotally a number of women noted that they were
too busy.

APOE Laboratory Assay Methods

Participants were instructed to collect cheek brush samples by “vigorously
rubbing” a sterile cytology brush against the inside of each cheek. Upon receiving the
samples, the brushes were prepped immediately or stored at 4°C for up to two days. The
two brushes were placed into a single tube containing 400uls of 50mM NaOH. The
tubes was heated to 95°C for ten minutes and immediately placed on ice for ten minutes.
The brushes were discarded and 40uls of Tris base pH 8.0 was added to each tube and the
samples were mixed. The prepared DNA was stored at -20°C.

DNA was prepared from cultures of abortus tissue and cultured amniocytes
following a standard with Gentra® DNA kit reagents. Two coverslips were used for
amniocentesis cultures and one ﬂask was used for abortus tissue cultures.
Coverslips/ﬂasks were rinsed in phosphate buffered saline. The cultured cells were
trypsinized and transferred into a 10-ml tube. The cells were centrifuged at 2.5 K for ten
minutes. The supernatant was discarded and the cell pellet was transferred to a 1.5 ml
microfuge tube with 300 pl Cell Lysis Solution. Aﬁer pipetting the solution up and down
a few times, 12 pl 1 M DTT and 3 pl 10 mg/ml proteinase K were added to each tube.
The cells were incubated at 55°C overnight in a water bath. After cooling to room

temperature, 200 pl of Protein Precipitation Solution was added and the mixture was

40

vortexed vigorously for 30 seconds. The mixture was iced for ﬁve minutes and
microfuged at 12K RPM for three minutes. The supernatant was transferred to a new
tube and 300 pl isopropanol was added. The tubes were mixed by inversion and
microfuged at 12K RPM for one minute. The supernatant was discarded and the pellet
was dried. The dried pellet was dissolved in 250p] 50 mM NaOH. The mixture was
heated to 95°C for 10 minutes and 25 ul of 1M Tris pH 8.0 was added to each tube. The
prepared DNA was stored at -20°C.

White cells were isolated from blood samples within one month of the initial draw
date. One hundred rnicroliters of blood was added to 500ul of cell lysis buffer. The
solution was vortexed and microfuged at 12K RPM for 30 seconds. The supernatant was
discarded and lOOpl red cell lysis buffer was added to the pellet. The solution was
vortexed and heated at 95°C for 10 minutes. The solution was placed on ice for 10
minutes and 10ul 1 M Tris pH 8.0 was added. The prepared DNA was stored at -20°C.

The DNA was ampliﬁed by polymerase chain reaction in a DNA Thermal Cycler
( Perkin Elmer Cetus model 9600) using oligonucleotide primers. The forward primer
sequence was 5’-ACAGAAT1’CGCCCCGGCCTGGTACAC-3’ and the reverse primer
sequence was 5’-TAAGCT'I‘GGCACGGCTGTCCAAGGA-3’. Each ampliﬁcation
reaction contained: 5 ul prepared DNA, 25pmol of each primer, 2.5mmol magnesium
chloride, 10% dirnethyl sulfoxide, 0.5mmol dinucleotide triphosphates, Perkin Elmer 10x
PCR buffer, and 0.625 units Taq polymerase in a ﬁnal volume of 25 ul. Each
ampliﬁcation reaction was subjected to an initial denaturing period of 95°C for 5 minutes.
The samples were ampliﬁed for 40 cycles of 95°C for 30 seconds, 60°C for 30 seconds,

and 70°C for 30 seconds. The products were subjected to a ﬁnal extension period of ﬁve

41

minutes at 72°C. Following the ampliﬁcation, the products were digested with 10 U of
Hha I at 37°C for at least three hours. The digested products were separated on a 12%
non-denaturing acrylamide gel at 30 mA current for two hours. The resulting gel was
stained in ethidium bromide and viewed on an ultraviolet light box. The separated bands
were photographed with a Polaroid camera.

Results Goal 1

To develop an interview which could be used to collect demographic and health
information about cases and controls

A total of 26 Caucasian case women and 56 Caucasian control women were
interviewed (Table 6). The frequency matching by age was not as close as desired. Three
case women of non-Caucasian ethnic backgrounds were interviewed. Due to the small
number of non-Caucasian case women in our feasibility sample, we did not attempt to
ﬁnd frequency matched age controls for these women. The majority of cases came from

the MSU Cytogenetic Laboratory (Table 7).

Table 7: Number of Caucasian woman interviewed by age category

 

Mother’s Age 19 & Under 20-24 25-29 30-34 35-39 40& Over Total

 

Cases 2 1 6 6 6 5 26
(7%) (3%) (21%) (21%) (21%) (17%)
Controls 5 4 8 17 13 9 56

(8%) (7%) (13%) (28%) (21%) (15%)

 

Table 8: Ascertainment of Caucasian Cases

 

 

Method of Ascertainment Number of Cases DNA Available
Cytogenetics 21 Fetus and Parents
Follow up of Positive AF P 5 Living-Child and Parents
Genetic clinics 2 Living Child and Parents
Prior pregnancy indicated 0 Living Child and Parents
on A’F P Test Requisition

 

42

Our hypothesis that the prevalence of AD is increased in trisomic families in a
larger study hinges on the collection of reliable data about AD among case and control
family members. The presence of AD is a censored variable because some family
members will die before they have the opportunity to express the disease characteristics.
In order to properly analyze this censored variable, it is necessary to have data about age
of death for family members. Therefore, one of the important results of our feasibility
study is the analysis of the quality of data that was collected for age of death. Table 8
shows the number of women who reported that a relative had died and were able to report
an estimated age of death. The maj ority of women were able to report an age of death for
the relatives that we asked about (84%). Woman reported information about their parents
more completely than about their grandparents. One hundred percent of women who
stated that their parent had died were able to estimate the age of death compared to 83%
of women who stated that their grandparent had died. The number of women in each
category is too small to test whether maternal age is correlated with knowledge about age
of parents or grandparents.

The amount of information that women are able to share about a diagnosis of
Alzheimer’s disease among their relatives is also key to our hypothesis (Table 9). Table
9 presents data for the women who responded “ I don’t know” to the question about a
diagnosis of AD. Once again, the women were more able to report information about
their parents than their grandparents. F iﬂeen percent of women were unable to report
about AD among their parents compared to 40% percent of women who were able to
report about AD among their grandparents. None of the women reported that they had

experienced symptoms of dementia themselves. The information about great-

43

grandparents cannot be compared to the information about parents and grandparents

because the question was asked in a different way.

44

Table 9: Number of Women Able to Estimate Relatives’ Age of Death Compared to

the Number of Women who stated that the relative had died

 

 

 

    
 

Mother’s 19 & 20-24 25-29 30-34 35-39
Age Under
(100%) (100%). (100%)
Control - - - .-l/l ' ‘1/1" I
000/) 000/)
Dad Case - - 1/1 1/1 4/4
(100%) (100%) (100%)
Control - - - 3/3 4/4
(100%) (100%)
Mother’s Icas'e‘“ . ,, 2134/4375
'Mother (67%) (100%) ' (60%).
. Control - 2/2 — . ’ 3/4 ; 10/12-7, ﬁll/12‘
(100%) (75%) (83%) . (92%)
. , lrdk ........ . .
Mother’s Case 1/1 1/1 3/4 5/6 4/6
Father (100%) (100%) (75%) (83%) (67%)
3=d.k.
Control 3/3 1/3 5/6 10/11 1 1/12
(100%) (33%) (83%) (91%) (92%)
3=d.k.

Above

(100%) .

 

408:

 

3/3 ‘

2/2
(100%)

,- (100%)

8/8

. (100%) g

5/5
(1 00%)

8/9
(89%)

 

Father’s Case 0/1 0/1 2/5 5/5
Father (0%) (0%) (40%) (100%)
Control 3/4 0/1 6/7 1 1/ 14
(75%) (0%) (86%) (79%)
3=d.k.

2/5
(40%)
l=d.k.
9/10
(90%)
2=d.k.

4/5
(80%)

7/8
(88%)
l=d.k.

 

d.k.= Woman does not know if relative is stiTl living

45

Table 10: Number of Women Unable to Report about Alzheimer’s Disease Among
Their Parents and Grandparents

 

 

 

Field test a sample collection protocol and laboratory assay which could identijj/

Apolipoprotein E genotype

Mother’s l9 & 20-24 25-29 30-34 35-39 40 &
Age Under Above
Mom Case ' ‘ ’- - l- l - " ‘- ’
Control - - - - - 1/9
‘ (11%)
Dad Case - - 2/6 - 1/6 -
(33%) (17%)
Control 1/5 - - 2/17 1/13 1/9
(20%) (12%) (8%) (1 1%)
i-Mothér’s’icase " ’1/2"’“-‘ ' - " "176"“ 5"”“1767‘ " ”‘4/6” 1”“ "“7175”
_ Mother . (50%) (17%) 1 (17%) (67%) (20%)
‘ 7 Control - - 1/8 3/17 2/13 , 8/9
11.2%) (1.8%)- .(l.5.%),.... (89%)
Mother’s Case 2/5 - 3/6 2/6 4/6 1/5
Father (40%) (50%) (33%) (67%) (20%)
Control 1/5 1/4 3/8 8/17 4/13 4/9
(20%) (25%) (3 8%) (47%) (3 1%) (44%)
f'Fathé’r’s” " ‘CaSe ‘ ” - “171 376 ""“3/6’ ‘ 376°" " 1/5
Mother . . (100%) (50%) (50%) (50%) (20%)
Control 2/5 1/4 1/8 7/ 17 ‘ 3/13 3/9
_ 140%)... _. (25%) ., 112%). 441%) _. (23%») . ,_ (33%)
Father’s Case 1/5 1/1 3/6 3/6 5/6 35
Father (20%) (100%) (50%) (50%) (83%) (60%)
Control 2/5 3/4 3/8 8/17 6/13 6/9
(40%) (75%) (38%) (47%) (46%) (67%)
Results Goal 2

A total of 24 DNA buccal swab collection kits were mailed out to case women.

Two case women declined to give a DNA sample after the telephone interview and

before their kit had been sent. Two case women declined to give a DNA sample upon
receiving their kit in the mail. Fourteen kits were returned from case women for a return
rate of 64% (14/22). Fifty-ﬁve DNA collection kits were mailed out to control women.
One control woman declined to give a DNA sample after completing the interview.
Twenty collection kits were returned for a return rate of 37% (20/54) for the control
women.

Four of the 14 samples collected ﬁ'om the case mothers failed to amplify under
our PCR conditions. An additional extraction procedure using phenol was attempted to
improve the quality of DNA. This attempt was unsuccessful, 4/4 did not amplify aﬁer
the additional extraction procedure.

Gene frequencies in our feasibility sample are presented in table 10. Though our
feasibility study was not designed to test the hypothesis linking trisomy to the ApoE e4
allele, we did calculate the sample size that would be required to test the hypothesis in a
larger study. In order to detect a two-fold difference in the 84 allele frequency (30% vs
15% as reported in Avramopoulos) between women under 30 with maternal MII trisomy
and women under 30 with no trisomy at an alpha equal to 0.05 with 80% power, 86 case
women under thirty with maternal MII errors would be needed and 344 control women
over 30 would be needed. The total population that would be needed would depend on
the percentage of errors that are maternal MII in nature. The gene ﬁ'equencies we found
using our laboratory assay among the case and control women for 82, 83 ,and 84 are

presented in table 10.

47

Table 11: Gene Frequencies for Caucasian Mothers

 

82 83 84

 

Case Mothers 0.10 0.80 0.10
Control Mothers 0.11 0.78 0.11

 

Results Goal 3

To test the feasibility of our study using the MSU prenatal screening program as a
population of cases and controls.

During our two years of case ascertainment we were able to identify 255 women
who ﬁt our case deﬁnition and could potentially have been included in our feasibility
study. In order to test our hypothesis 86 case women under thirty would be needed. The
rate of maternal MII errors ranges for different chromosomes. For example, the
published rate for trisomy 16 is 0%, trisomy 21 is 13% and trisomy 18 is 62%. The total
number of cases needed would depend upon the average rate of meiosis II non-
disjunction in our case population. The majority of our potential cases came through the
Cytogenetic laboratory (Table 12). Our potential cases include examples of trisomy 4, 6,
7, 9, 13, 15, 16, 17, 18, 21, and 22. Our control population contained enough women to

randomly sample and still have a 4:1 ratio.

Table 12: Ascertainment of Potential Cases

 

 

 

Ascertainment Method Number of Potential Cases
Cytogenetics laboratory 233
AF P Test - Positive Screen 81
AF P — Prior History on Test Requistion 5
AF P — Follow-up Negative Screen 1
Genetics Clinic 17
Prenatal Clinic 4
TOTAL 341

Cases in our study were identiﬁed in a retrospective manner. We attempted to

contact by mail 57 cases for our study. One case woman refused to be in the study by a

48

postcard, 3 woman refused over the telephone, and we never made contact with 24
potential case woman. Women were classiﬁed as never made contact if we mailed a
letter to an address but we were unable to contact a person at a telephone number due to a
missing or non-valid phone number. We could not verify that the address that we mailed
the letter to was valid, therefore we do not know if the woman ever received an invitation
into the study. Twenty-nine case women were interviewed for an enrollment of 51% and
a direct reﬁrsal rate of 7%.

We attempted to contact by mail 128 controls. Nine women refused to be in the
study by post card, 9 women refused to be in the study over the phone, and 61 women
were interviewed. Our enrollment rate for controls was 48% and our direct refusal rate
was 14%. The control women in the youngest two age groups were the most difﬁcult to

contact (Table 12).

Table 13: Control Contact Results by Age Groups

 

Mother’s Age 19 & below 20-24 25-29 30-34 35-39 40 & above

 

 

 

Never Made Contact 35 19 13 7 17 14
. , (81%) (79%) (50%) (23%) (57%) (54%)
Refusal 3 1 5 6 O 3
(7%) (4%) (19%) (20%) (0%) (12%)
Interviewed 5 4 8 17 13 9
(12%) (16%) (31%) (57%) (43%) (35%)
Total 43 24 26 30 3O 26
Results Goal 4

Identijy potential strengths and weaknesses with our case-control study design

Our methodology would need a number of improvements in order to repeat this
study on a larger scale. First, we spent a large amount of resources on ﬁnding women. A
number of women had changed addresses and telephone numbers. We attempted to

contact women by mail up to two years after the index pregnancy. One way to resolve

49

this problem would be to contact women and enroll them on a prospective basis, but this
would greatly increase the timeframe of the study. The most mobile and difficult group
of women to locate were the youngest age category (under 20 years). Unfortunately, the
younger women are crucial to support our hypothesis. In the future, extra recruiting
resources would have to be spent on targeting these women. Our controls were more
labor intensive to contact than our cases. The primary reason was that we did not have
their telephone numbers recorded in the database. The MSU Prenatal Screening Program
has subsequently started collecting and recording the telephone numbers of women they
screen. This addition could increase our enrollment rates that were very low overall.

Once we were able to contact women on the telephone, we had high participation
rates for the interview. In addition, the majority of women 94% stated they were willing
to donate a DNA sample. However, 64% of the case women and 37% of the control
women mailed back their collection kits. Maybe improvements in our strategies to re-
contact women would improve our collection kit retrieval rates. It is anticipated that
allocating more personnel time and resources towards this process would assist. The case
women may have felt more motivated to complete their participation in the study because
of their personal experience with a trisomic pregnancy. Unfortunately, the return of the
DNA kit was crucial to two of our three speciﬁc aims. In the future we could restrict the
interview to women who are willing to donate a sample ﬁrst in order to save on
resources.

Some of the returned DNA samples failed to amplify. This could be due to delays
in mailing samples. Some women indicated that they had let their sample sit before

mailing it. We could modify the instructions with the DNA collection kit to suggest that

50

women mail their sample immediately aﬂer collecting it. A second option is to pilot
other non-invasive DNA collection methods. Thirdly, we could call and go collect the
DNA sample in-person. Due to our limited resources, we were unable to perform the
DNA microsatellite analysis that could be used to identify the cell division and parent of
error. Development of protocols for each individual chromosome is a difﬁcult and time-
consuming process. In a larger study, it may be more cost-efﬁcient to contract an outside
individual to analyze cell division and parent of error.

A few other minor improvements could be made to our study protocols. We
could attempt to frequency match the cases and controls on interviewer so that each
interviewer interviews the same percentage of cases and controls in each age category.
Also, we could set limits on the number of calls made to an individual woman and ask
that women identify people in their household that we can leave messages with regarding
the study. Identifying a household contact person allows us to leave messages without
violating an individual’s conﬁdentiality while still assisting us with our follow-up data

collection calls.

51

Table 14: Summary of suggested changes in Protocol

 

 

Methodological Difﬁculty Suggestive Corrective Action(s)
Difﬁculty in ﬁnding participants Contact on a prospective basis
Low numbers of young women interviewed Target young women

Low DNA kit return rate 1)Offer money

2)Only interview patients with samples
3)Go to home and collect sample

DNA sample failure 1)Suggest immediate mailing
2)Pilot other non-invasive sample
collection methods
3)Go to home and collect sample
Limited resources for microsatellite Contract outside individual to analyze cell
analysis division and parent of error.
Cases and Controls not matched on Match on interviewer
interviewer
Interviewers not blinded to hypotheses Blind interviewers to hypotheses
Numerous calls made to few women Limit number of calls made to individual
Unable to leave telephone message with Get women’s permission on consent form
household members to speak with household members

 

Our methodology had a number of strengths. Our study population had an excess
number of women to sample from. We found that our letter sent out to women initially
was successful at recruiting women into the study. Both the 1-800 telephone number and
the returned letter were used by women as ways of contacting us to update us on their
telephone number. In general, we found that the notiﬁcation by mail two weeks prior to
telephoning allowed women time to contact us by telephone or mail if they wished to
decline participation. Also, the letter adequately introduced the study and motivated
women to participate. Women were familiar with study when we telephoned. We had a
low refusal rate for the interview that suggests that this format is very acceptable to our
study population. The non-invasive method of DNA collection was easily exchanged

through the mail and successfully used with young children.

52

Table 15: Summary of Methodology Strengths

Study population

Initial recruitment letter
1-800 phone number

Letter for women to return
Two week waiting period
Non-invasive method of
DNA collection

0 Telephone interview format

 

 

Discussion
The number of participants in our study is small which limits our ability to make

any strong conclusions about data trends. The majority of women reported information
about age at death of relatives (84%). In addition women were better at reporting
information about their parents (100%) than their grandparents (83%). Since we were
relying on self-reported data, we would need to validate this information by getting death
certiﬁcates to comment on its accuracy.

There was a large amount of missing data for the AD questions (36%), especially
for the grandparents (40%) as compared to the parents (15%). One possibility is that the
women who do not know this information about their grandparents have the greatest age
gap differences between their grandparents and them. If we collected information on this
age gap (via birth date of the mother and grandparents or estimated age difference) we
could test this hypothesis. An additional dilemma is that younger women are at the
center of our hypothesis, but their parents could be too young to reach the peak AD age.
A study could expand the deﬁnition of AD to include symptoms of AD when the clinical
diagnosis was unknown to the interviewee. One possible symptomatic deﬁnition would

be “ Did your relative ever experience a slow progressive loss of memory, cognitive

53

abilities, and functioning on intellectual tasks?” This deﬁnition could include non-AD
conditions resulting from other causes of dementia. In order to keep the less reliable
symptomatic diagnosis separate from the physician diagnosed cases, we could classify
cases into categories of deﬁnite -— self-report of physician diagnosed AD conﬁrmed by
medical record, probable — self-report of AD unable to be conﬁrmed by medical record,
and suspicious - report of AD-like symptoms.

Some the buccal brush DNA samples that were collected failed to amplify. Once
again these samples were crucial to our hypothesis. A recent study by Garcia-Closes.
found that a single mouthwash sample collection resulted in higher yields and better
quality DNA than two cytobrush samples (in press). This sample collection procedure is
feasible for the adults in our study. However, the young children in our study would not
be able to follow the mouthwash sample collection protocol. The most feasible solution
for the children is to continue to use the buccal brush collection procedure with an
increased emphasis in our instruction materials to participants on timely return of the
samples. In addition, modifying our extraction techniques or number of PCR cycles for
the buccal samples may be necessary.

Enrollment rates were 51 % for cases and 48% for controls of the women we sent
a mailing. It is difﬁcult to classify the number of women who were lost to study because
we have no way of knowing what percentage of women actually got the letter that we
sent them. Women in the youngest age category were most difficult to contact. Our
contact rates could be improved if women were contacted on a more prospective basis.

In addition, the recent addition of collecting women’s telephone number by MSU

54

Prenatal Screening Program could assist in the contact of controls. Motivation for
involvement in the study was not recorded.

We would have liked to match our controls to cases based on ethnicity and age.
For this feasibility study we were only able to include Caucasian controls. In a larger
study it would be feasible to sample controls from different ethnic background and match
on ethnicity as we would have liked. In addition, our recruitment strategies did not
efﬁciently recruit case and control women into the corresponding age categories. Based
on this and other problems this feasibility study aided in determining resources needed
for a larger study. Speciﬁcally resources would need to be included to spend time re-
contacting women for sample collection, over-sampling the youngest population, and
reﬁning the laboratory collection techniques.

Major Feasibility Study Accomplishments

There were four major accomplishments made by our feasibility study. We
identiﬁed a collection of potential cases (Table 12), archived trisomy DNA samples,
reﬁned our interview and laboratory instruments through ﬁeld-testing, and developed and

debugged a Microsoft-Access database capable of storing our interview data.

55

APPENDIX

56

Appendix: Interview Content

SECTION I
The first set of questions ask some background information.

1. What is your date of birth?

__ __ 19__
Month Day Year

2. What is your race or ethnic background?

White/Caucasian ........................................ l
Black/African-American .............................. 2

Asian ........................................................... 3
Hispanic ...................................................... 4

Other (specify) ............................................. 5

3. What is the highest grade you have finished in school?

Elementary 1 2 3 4 5
High School 9 10 11 12
College 13 14 15 16

Post College 17+

No formal schooling 0

GED

4. What is your usual occupation? (include home maker, student)

 

57

SECTION 11

Now, I would like to ask you some questions about your family. Please answer the
questions in this section as they relate to your biological relatives. If you are unsure
of your answer to any question, please feel free to respond “I don’t know”.

First your biological parents:
5. Is your (biological) mother alive?
Yes_ No_ Don’t know_
If Yes or Don ’t know, go to question 8.
If No, go to question 6.
6. At what age did she die?

7. What was the cause of her death?

 

8. Now, I would like to read you a list of medical conditions. Please indicate if your
mother has/had been physician diagnosed with any of the following conditions.
Respond with yes, no, or I don’t know.

 

a High blood pressure ............................ Yes_ No_ Don’t know_

b. Stroke ....................................................... Yes_ No_ Don’t know_

c. Heart attack ............................................ Yes_ No_ Don’t know_

(1. High cholesterol level (over 240) ....... Yes_ No_ Don’t know_

e. Diabetes ................................................ Yes_ No_ Don’t know_
if Yes: Age of onset? yrs Don’t know

f. Thyroid disease ........................................ Yes_ No_ Don’t know_
if Yes: Overactive?_Underactive?_ Don’t know_

g. Parkinson’s disease Yes_ No__ Don’t know_

h. Alzheimer’s disease or senile dementia..Yes_No_ Don’t know_

i.Cancer ........................................................ Yes_ No_ Don’t know_
Kind

j. Premature graying (which is a significant amount of gray hair before age 25)
............................................ Yes_ No_ Don’t know_

9. At approximately what age did she reach menopause? Don’t Know

If known, Was this the result of a hysterectomy? Yes No Don’t
Know

58

10. Is your (biological) father alive?
Yes_ No__ Don’t know_
If Yes or Don ’t know, go to question 13.
If No, go to question 11.

l 1. At what age did he die?

12. What was the cause of death?

 

13. I will again read you the same list of medical conditions. Please indicate if your
father has/had been diagnosed with any of the following:

 

a. High blood pressure ............................ Yes_ No_ Don’t know_
b. Stroke ....................................................... Yes_ No_ Don’t know_
c. Heart attack ............................................ Yes_ No_ Don’t know_
(1. High cholesterol level (over 240) ....... Yes_ No_ Don’t know_
e. Diabetes ................................................ Yes_ No_ Don’t know_
if Yes: Age of onset? yrs Don’t know
f. Thyroid disease ........................................ Yes_ No_ Don’ tknow_
if Yes. Overactive? __Underactive?___ Don’ t know_
g. Parkinson’s disease“ Yes_ No_ _Don’t know_
h. Alzheimer’ s disease or senile dementia. .Yes_ No_ Don’t know_
i. Cancer ........................................................ Yes_ No_ Don’t know_
Kind
j. Premature graying (which is a significant amount of gray hair before age 25)
............................................ Yes_ 1 No_ Don’t know_

14. Now I would like to ask you about your biological grandparents, starting with
your mother’s parents.
Is your mother’s mother alive?
Yes_ No_ Don’t know_

15. If yes, How old is she?
If no, At what age did she die?

16. What was the cause of her death?

 

59

17. Has/Had she been diagnosed with any of the following conditions:

 

a. High blood pressure ............................ Yes_ No_ Don’t know_
b. Stroke ....................................................... Yes_ No_ Don’t know_
c. Heart attack ............................................ Yes_ No_ Don’t know_
d. High cholesterol level (over 240) ....... Yes_ No_ Don’t know_
e. Diabetes ................................................ Yes_ No__ Don’t know_
if Yes: Age of onset? yrs Don’t know
f. Thyroid disease ........................................ Yes__ No_ Don’ t know
if Yes. Overactive? _Underactive?_ Don’ t know
g. Parkinson’s disease. Yes_ No_ Don’t know_
h. Alzheimer’s disease or senile dementia. .Yes_ No_ Don’t know_
i. Cancer ........................................................ Yes_ No_ Don’t know_
Kind
j. Premature graying (which is a significant amount of gray hair before age 25)
............................................ Yes_ No_ Don’t know_
18. At approximately what age did she reach menopause Don’t know
If known, Was this the result of a hysterectomy? Yes No Don’t
know

19. Now your maternal grandfather:
Is your mother’s father alive?
Yes_ No_ Don’t know_

20. If yes, How old is he?
If no, At what age did he die?

21. What was the cause of his death?

 

22. Has/Had he been diagnosed with any of the following conditions:

 

a. High blood pressure ............................ Yes_ No_ Don’t know_
b. Stroke ....................................................... Yes_ No_ Don’t know_
c. Heart attack ............................................ Yes_ No__ Don’t know_
d. High cholesterol level (over 240) ....... Yes__ No_ Don’t know_
e. Diabetes ................................................ Yes_ No_ Don’t know_
if Yes: Age of onset? yrs Don’t know
f. Thyroid disease ........................................ Yes_ No__ Don’ t know_
if Yes. Overactive? _Underactive?_ Don’ t know
g. Parkinson’s disease" Yes_ No_ Don’t know_
h. Alzheimer’ s disease or senile dementia. .Yes_ No_ Don’t know_
i. Cancer ........................................................ Yes_ No__ Don’t know_
Kind
j. Premature graying (which is a signiﬁcant amount of gray hair before age 25)
............................................ Yes_ No_ Don’t know_

60

23. Now I will ask you about your father’s parents:
Is your father’s mother alive?
Yes_ No_ Don’t know_

24. If Yes, How old is she?
If No, At what age did she die?

25. What was the cause of her death?

 

26. Has/Had she been diagnosed with any of the following conditions:

 

a. High blood pressure ............................ Yes__ No_ Don’t know_
b. Stroke ....................................................... Yes_ No__ Don’t know_
c. Heart attack ............................................ Yes_ No_ Don’t know_
d. High cholesterol level (over 240) ....... Yes__ No_ Don’t know_
e. Diabetes ................................................ Yes_ No_ Don’t know_
if Yes: Age of onset? yrs Don’t know___
f. Thyroid disease ........................................ Yes_ ' No_ Don’ t know_
if Yes. Overactive? _Underactive?_ Don’ t know
g. Parkinson’s disease. Yes_ No_ —Don’t know_
11. Alzheimer’s disease or senile dementia. .Yes_ No_ Don’t know_
i.Cancer ........................................................ Yes: No_ Don’t know_
Kind
j. Premature graying (which is a signiﬁcant amount of gray hair before age 25)
............................................ Yes_ No__ Don’t know_
27. At approximately what age did she reach menopause? Don’t know

Was this the result of a hysterectomy? Yes_ No__ Don’t know
28. Now your paternal grandfather. Is your father’s father alive?
Yes_ No_ Don’t know_

29. If yes, How old is he?
If no, At what age did he die?

30. What was the cause of his death?

 

61

31. Does/Did he have any of the following conditions:

 

a. High blood pressure ............................ Yes_ No_ Don’t know_
b. Stroke ....................................................... Yes__ No__ Don’t know_
c. Heart attack ............................................ Yes_ No_ Don’t know_
d. High cholesterol level (over 240) ....... Yes_ No__ Don’t know_
e. Diabetes ................................................ Yes_ No_ Don’t know_
if Yes: Age of onset? yrs Don’t know
f. Thyroid disease ........................................ Yes_ No_ Don’ t know
if Yes Overactive?_ Underactive?_ Don’ t know——
g. Parkinson’ s disease. Yes_ No_ —Don’t know_
h. Alzheimer’s disease or senile dementia. .Yes_ No_ Don’t know_
i. Cancer ........................................................ Yes_ No_ Don’t know_
Kind
j. Premature graying (which is a signiﬁcant amount of gray hair before age 25)
............................................ Yes_ No_ Don’t know_

32. Now I’m going to ask you about other members of your family; your brothers,
sisters, cousins, uncles, aunts, nieces and nephews. Is there anyone you know of in
your biological family who has had a child or a pregnancy with:

a.Trisomy 21 or Down syndrome ................................. Yes__ No__ Don't know_
b.Trisomy 18 or Edward syndrome .............................. Yes_ No_ Don’t know_
c.Trisomy 13 or Patau syndrome .................................. Yes_ No_ Don’t know_

d. Trisomy 16 .................................................................. Yes__ No_ Don’t know_

c. Any other chromosomal trisomy ............................... Yes__ No_ Don’ t know

f. Another chromosomal abnormality........ .. .Y..e_s No__ Don’ tknow_

If No or Don ’t know, go to Question 34.
If Yes go to Question 33.

33. Please tell me how that person (s) with the chromosome abnormality is (was)
related to you?

 

34. Do you know of any twins in your biological family? Yes No
If No, go to Question 3 7.
If Yes, go to Question 35 .

35. Identical Twins Fraternal/unlike Twins Don’t know

36. Please tell me how they are related to
you

62

37. Did any of your biological great grandparents develop Alzheimer’s disease or

dementia?
Yes_ No_ Don’t know

If No, go to question 39.
If Yes, go to Question 38.

38. How was that great grandparent related
you?

 

63

SECTION III
39. Now I would like to ask some questions about your own health.

When you were not pregnant, have you ever been treated for, or been told you
have, any of the following:

 

a. High blood pressure ............................ Yes_ No__ Don’t know_
b. Stroke ....................................................... Yes__ No_ Don’t know_
c. Heart attack ............................................ Yes__ No_ Don’t know_
d. High cholesterol level (over 240) ....... Yes_ No_ Don’t know_
e. Diabetes ................................................ Yes__ No_ Don’t know_
if Yes: Age of onset? yrs Don’t know
f. Thyroid disease ........................................ Yes_ No_ Don’ t know
if Yes: Overactive? _Underactive?_ Don’ t know
g. Parkinson’ s disease" Yes_ No_ _Don’t know_
11. Alzheimer’ s disease or senile dementia. .Yes_ No_ Don’t know_
i. Cancer ........................................................ Yes_ No_ Don’t know_
Kind
j. Premature graying (which is a significant amount of gray hair before age 25)
............................................ Yes_ No_ Don’t know_

40. At what age did you begin menstruation?

41. Have you reached menopause? Yes_ No_
If No, go to Question 42.
If Yes,. At what age did your periods stop?
Was this the result of hysterectomy?—

42. Prior to your trisomic pregnancy, have you had surgery to remove either of
your ovaries?
Yes _ No _ Both—
If Yes, Why?

 

When?

 

SECTION IV
Now I would like to ask you some questions about your pregnancies.

43. How many times have you been pregnant including any losses?

1 2 3 4 5 6 7 8 9 10 ll 12

64

44. When you were trying to get pregnant, did it (ever) take more than three
months?

Yes_ No_ Doesn’t apply—

If No or Doesn ’t apply, go to Question 4 6.

If Yes, go to Question 44.

45. More than six months?
Yes__ No__
If No, go to Question 46.
If Yes, go to Question 45.

46. Have you ever had a physician prescribe medication to help you get pregnant?
Yes_ No_ '

47. Prior to your trisomic pregnancy, Have you ever used a hormonal contraceptive
method, including birth control pills, Depo Provera, and Norplant? Yes

No
If yes, At what age did you begin this method?
For how many years (total) ?
48. Were any of your pregnancies twins or triplets? Yes_ No_

If No, go to Question 52.
if Yes, go to Question 49.

49.Whichpregnancies?_l 2 3 4 5 6 7 8 9 10 ll 12

50. Please tell me if the twins/ triplets were identical (alike) or fraternal (unlike)

Pregnancy_ Identical_ F raternal_ Different sex_ Same sex_ Don’t know_
Pregnancy_ Identical“ Fratemal_ Different sex_ Same sex_ Don’t know______
Pregnancy___ Identical_ Fratemal_ Different sex_ Same sex_ Don’t know_

51. Were you having treatment for infertility at the time you conceived these
twins/triplets?
Yes_ No

52. Now I would like to ask you some specific questions about your pregnancy
history.

Beginning with your ﬁrst pregnancy....(read questions off table)

65

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

£8 90 3828 o». v65 AHMWNML—g \32 3.83»: (<8 9w” Son £8 Ea max»
Ga. mnn. ...v oamnnnow m 235 0:5... €80 @3853 $5 QRSER 9.3» unﬁ.
385? 888380. E 38 93 So oEoBomoBo 9a 880 898 8 Q48 33. new: .5.
wanna econmon. on vamnnnov. 8% «8&8 gonna 9a 9.8.38 NR...»
3553 3.??? El: SR 88 9o Edmnnnov. .V . . . . .
. Qeaemuiwba 98 035
8% mos BBQ 398 g3 29: bonm 88:3 . no 3 BE
95% moi Bonus 828 AB «on c5 mow.“ woBomo
€80 3=ng .N a. no.5 .5. . . .
. 85. a on can an: gems
_ _
mam Bo Hanna: o». be an a. > Eon: manna mam—98 Gannon. gonna Zenonnﬁma.
a and. o». 0.8Bo 5&8 GEE mm: :Snm
Toma—now
v.8 OPEC. ﬁg
lioowm no 3 m. <8 <8
name: 2o 2o
Banana 0?»?
v8 v8 Ont}? ESnm
28.8 no no 2— m. <8 <8
83» 5.895. 2o 2o
Bonn...“ a???
v8 v8 0?»? Slalom
$88 no no 3 m. <8 <8
28:: nnwnoﬁn 2o Zo
Bonn; 9?»?

 

 

66

10.

11.

12.

13.

14.

REFERENCES

Jacobs P, Hassold T, Henry A, Pettay D, Takaesu N. Trisomy 13 ascertained in a
survey of spontaneous abortions. J of Med Genet. 1987;24:721-724.

Gaulden M. Maternal age effectzThe enigma of Down syndrome and other
trisomic conditions. Mat Res. 1992;296:69-88.

Benacerraf BR, Nadel A, Bromley B. Identiﬁcation of second-trimester fetuses
with autosomal trisomy by use of a sonographic scoring index. Radiology.

1994 ,:193(1) 135-40.

DiMaio M, Baumgarten A, Greenstein R, Saal H, Mahoney M. Screening for fetal
Down's syndrome in pregnanacy by measuring maternal serum alpha-fetoprotein
levels. N Eng J Med. 1987;317:342-346.

Cuckle H, Wald N, Lindenbamn R. Maternal serum alpha-fetoprotein
measurement: A screening test for Down syndrome. Lancet. 1984;i:926-929.

Pueschel S. Maternal alpha-fetoprotein screening for Down's syndrome. N Eng J
Med. 1987;317:376-378.

Wald NJ. Antenatal screening for Down syndrome. Prog Clin Biol Res.
1995;393:27-42.

Ryall RG, Staples AJ, Robertson EF, Pollard AC. Improved performance in a
prenatal screening programme for Down's syndrome incorporating serum-ﬂee
hCG subunit analyses. Prenat Diagn. l992;12(4):251-61.

von Winiwarter H. Etudes sur la spermatogenese htunaine. Arch Biol.
1912;27:97-189.

Tjio J, Levan H. The chromosome number in man. Hereditas. 1956;42:1-6.
Ford C, Hamerton J. The chromosomes in man. Nature. 1956;178:1020-23.

Lej eune J, Gautier M, Turpin R. Etudes des chromosomes somatique de neuf
enfants mongoliens. C R hebd Seannces Acad Sci Ser D. 1959;248:1721-1722.

Edwards J, Hamden D, Cameron A, Cross V, Wolf O. A new trisomic syndrome.
Lancet. 1960;9:787-89.

Patau K, Smith D, Therman E, Inhorn S, Wagner H. Multiple congenital anomaly
caused by an extra autosome. Lancet. 1960;1:790-93.

67

15.

16.

17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

27.

28.

Jacobs P, Hassold T. The origin of numerical chromosome abnormalities. In: Hall
J, Dunlap J, eds. Advances in Genetics. New York: Academic Press; 1995.

Hassold T, Jacobs P. Trisomy in Man. Ann Rev Genet. 1984;18:69-97.

Kline J, levin B, Shrout P, Stein Z, Susser M, Warburton D. Maternal smoking
and trisomy among spontaneously aborted conceptions. Am J Hum Genet.
1983;35:421-432.

Hassold T, Chen N, F unkhouser J, et al. A cytogenetic study of 1000 spontaneous
abortions. Ann Hum Genet. 1980;44(Pt 2):151-78.

Penrose L. The relative effects of paternal and maternal age in mongolism. J
Genet. 1933;27:21 9-224.

Hook EB, Cross PK, Schreinemachers DM. Chromosomal abnormality rates at
amniocentesis and in live-bom infants. Jama. 1983;249(15):2034-8.

Hassold T, Chiu D. Maternal age-speciﬁc rates of numerical chromosome
abnormalities with special reference to trisomy. Hum Genet. 1985;70:11-17.

Hassold T, Chiu D, Yamane J. Parental origin of autosomal trisomies. Ann Hum
Genet. 1984;48:129-144.

Risch N, Kline J, Stein Z, Warburton D. Maternal age dependent and independent
factors in autosomal trisomies. Am J Hum genet. 1986;39:68-79.

Morton N, Jacobs P, Hassold T, Wu D. Maternal age in trisomy. Ann Hum Genet.
1988;52:227-235.

Sherman S, Petersen M, Freeman S, et a1. Non-disjunction of chromosome 21 in
maternal meiosis IzEvidence for a matemal-age dependent mechanisms involving
reduced recombination. Hum Mol Genet. 1994;3:1529-1535.

Antonarakis S, Petersen M, McInnis M, et al. The meiotic stage of non-
disjunction in trisomy 21 :Determination by using DNA polymorphisms. Am J
Hum Genet. 1992;50:544-550.

Ayme S, Lippman-Hand A. Matemal-age effect in aneuploidyzDoes altered
embryonic selection play a role? Am J Hum Genet. 1982;34:558-565.

Antonarakis S. Human chromosome 21 :Genome mapping and exploration, circa
1993. Trends Genet. 1993;9zl42-148.

68

29.

30.

31.

32.

33.

34.

35.

36.

37.

38.

39.

40.

41.

42.

Yoon P, Freeman S, Sherman S, et a1. 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. 1996;58:628-633.

Warburton D. The effect of maternal age on the ﬁ'equency of trisomyzchange in
meiosis or in utero selection? In: Hassold T, Epstein C, eds. Molecular and
Cytogenetic Studies of Non-disjunction. New York: Liss; 1989.

Kikuchi Y, Oishi H, Tonomura A, et a1. Translocation Down's syndrome in Japan:
It's Frequency, mutation rate of translocation and parental age. Jpn J Hum Genet.
1969;14:93-106.

Hamerton J. Clinical Cytogenetics. Human cytogenetics. New York: Academic
Press; 1971 .

Hasshold T. unpublished results. .

Phillips 0, Cromwell S, Rivas M, Simpson J, Elias S. Trisomy 21 and maternal
age of menopausezDoes reproductive age rather than chronological age inﬂuence
risk of nondisjunction? Hum Genet. 1995;95:117-118.

Henderson S, Edwards R. Chiasma ﬁequency and maternal age in mammals.
Nature. 1968;218:22-28.

Warren A, Chakravarti A, Wong C, et al. Evidence for reduced recombination on
the nondisjoined chromosomes 21 in Down syndrome. Science. 1987;37:652.

Perroni L, Dagna Bricarelli M, Grasso M, Pierluigi M, al. e. Crossing over and
chromsome 21 nondisjunction: a study of 60 families. Am J Med Genet.
1990;Suppl. 7: 141-147.

Sherman S, Takaesu N, Freeman S, et a1. Trisomy 21 :Association between
reduced recombination and non-disjunction. Am J Hum Genet. 1991 ;49:608-620.

Stene J, Fischer G, Stene E, Mikkelsen M, Petersen E. Paternal age effect in
Down's syndrome. Ann Hum Genet. 1977;40:299-306.

Matsunaga E, Tonomura A, Oishi H, Kikuchi Y. Reexamination of paternal age
effect in Down Syndrome. Hum Genet. 1978;40:259-268.

Erickson J, Bjerkdal T. Down Syndrome associated with father's age in Norway. J
Med Genet. 1981;18:22-28.

Stene J, Stene E, Stengel Rutkowski S, Murken J. Paternal age and Down
SyndromezData from prenatal diagnosis. Hum Genet. 1981;59:119-124.

69

43.

45.

46.

47.

48.

49.

50.

51.

52.

53.

54.

55.

Roth M, Feingold J, Baumgarten A, Bigel P, Stoll C. Reexamination of paternal
age effect in Down syndrome. Human Genet. 1983;63:149-152.

Roecker G, Huether C. An analysis for paternal age effect in Ohio's Down
Sydrome births, 1970-80. Am J Hum Genet. 1983;35: 1297-1306.

Cohen B, Lilienﬁeld A, al. e. Prenatal factors in Down Syndrome: results of the
second Baltimore case-control study. In: H E, P I, eds. Population cytogenetics:
studies in humans. New York: Academic Press; 1977.

Erickson J. Down syndrome, paternal age, maternal age and birth order. Ann Hum
Genet. 1978;41:289-298.

Regal RR, Cross PK, Lamson SH, Hook EB. A search for evidence for a paternal
age effect independent of a maternal age effect in birth certiﬁcate reports of
Down's syndrome in New York state. Am J Epidemiol. 1980;112(5):650-5.

Hook E, Cross P, Lamson S, Regal R, Baird P, Uh S. Paternal age and Down
syndrome in British Columbia. Am J Hum Genet. 1981;33:123-128.

Hook EB, Cross PK. Paternal age and Down's syndrome genotypes diagnosed
prenatally: no association in New York state data. Hum Genet. l982;62(2):167-
74.

Ferguson-Smith M, Yates J. Maternal age speciﬁc rates for chromsome
abberations and factors inﬂuencing them: report of a colllaborative European
study on 52,965 arnniocenteses. Prenat Daign. 1984;4z5-44.

Oster J. The causes of mongolism. Danish Med Bull. 1956;3:158-64.

Carter C, Evans K. Risk of parents who have had one child with Down's
Syndrome (mongolism) having another similarly affected. Lancet. 1961 ;i:926-29.

Richards B. The recurrence of mongolism in sibships. J Ment Deﬁc Res.
1977;21:5-21.

Baty BJ, Blackburn BL, Carey JC. Natural history of trisomy 18 and trisomy 13:
1. Growth, physical assessment, medical histories, survival, and recurrence risk.
Am J Med Genet. 1994;49(2):175-88.

Mikkelsen M, Stene J. Previous child with Down-syndrome and other
chromosome aberration: group report. In: Murken J, Stengel-Rutkowski S,
Schwinger E, eds. Prenatal diagnosis'proceedings of the 3rd European
Confernce on Prenatal Diagnosis of Genetic Disorders. Stuttgart: Enke; 1979.

70

56.

57.

58.

59.

60.

61.

62.

63.

64.

65.

66.

67.

68.

Daniel A, Stewart L, Saville T, et a1. Prenatal diagnosis in 3,000 women for
chromosome, X-linked, and metabolic disorders. Am J Med Genet.
l982;11(1):61-75.

Caron L, Tihy F, Dallaire L. Frequencies of chromosomal abnormalities at
amniocentesis: over 20 years of cytogenetic analyses in one laboratory. Am J Med
Genet. 1999;82(2):l49-54.

Boue I, Boue A. Chromsome analysis of two consecutive abortuses in each of 43
women. Hum Genet. 1973; 1 5:1 12-114.

Alberman E, Elliot M, Creasy M, Dhadial R. Previous reproductive history in
mothers presenting with spontatneous abortions. Br J Obstet Gynaecol.
1975;82:366-373.

Lauritsen J. Aetiology of spontaneous abortion: a cytogenetic and epidemiologic
study of 288 abortuses and their parents. Acta Obstet Gynecol Scand.
1976;(Supp1.) 52: 1-29.

Kajii T, Ferrier A. Cytogenetics of aborters and abortuses. Am J Obstet Gynecol.
1978;131:33-38.

Warburton D, Kline J, Stein Z, Hutzler M, Chin A, Hassold T. Does the karyotype
of a spontaneous abortion predict the karyotype of a sunsequent abortion?
Evidence from 273 women with two karyotyped Spontaneous abortions. Am J
Hum Genet. 1987;41:465-483.

Stene J. detection of higher recurrence risk for age-dependent chromosome
abnormalities, with an appication to trisomy G (Down's syndrome). hum Hered.
1970;20:112-122.

Hassold T, Sherman S, Hunt P. The origin of trisomy in humans. Prog Clin Biol
Res. 1995;393:1-12.

Fisher J, Harvey J, Lindenbaum R, Boyd P, Jacobs P. Molecular studies of
trisomy 18. Am J Hum Genet. 1993;52:1139-1144.

Hassold T. Cytogenetic and molecular studies of trisomy 13. J of Med Genet.
1987;24:725-732.

Hassold T, Pettay D, Freeman S, Grantham M, Takaesu N. Molecular studies of
non-disjunction in trisomy 16. J Med Genet. 1991;28:159-162.

Fisher JM, Harvey JF, Morton NE, Jacobs PA. Trisomy 18: studies of the parent

and cell division of origin and the effect of aberrant recombination on
nondisjunction. Am J Hum Genet. l995;56(3):669-75.

71

69.

70.

71.

72.

73.

74.

75.

76.

77.

78.

79.

80.

Jacobs P, Hassold T. Proceedings of the Seventh International Congress, Berlin
1986. Berlin: Springer-Verlag; 1987.

Risch N, Stein Z, Kline J, Warburton D. The relationship between maternal age
and chromosome size in autosomal trisomy. Am J Hum Genet. 1986;39:68-78.

Menzel HJ, Kladetzky RG, Assmann G. Apolipoprotein E polymorphism and
coronary artery disease. Arteriosclerosis. 1983;3(4):310-5.

Zekraoui L, Lagarde J, Raisonnier A, Gerald N, Aouizerate A, Lucotte G. High
frequency of the apoliprotein E 4 allele in African pygrnies and most of the
Aﬁican populations in sub-Saharan Aﬁica. Hum Biol. 1997;69:575-81.

Ehnholrn C, Lukka M, Kuusi T, Nikkila E, Utermann G. Apolipoprotein E
polymorphism in the Finnish population: gene frequencies and relation to
lipoprotein concentrations. J Lipid Res. 1986;27(3):227-35.

Lehtimaki T, Moilanen T, Viikari J, et al. Apolipoprotein E phenotypes in Finnish
youths: a cross-sectional and 6- year follow-up study. J Lipid Res.
1990;31(3):487-95.

Eggertsen G, Tegelrnan R, Ericsson S, Angelin B, Berglund L. Apolipoprotein E
polymorphism in a healthy Swedish population: variation of allele frequency with
age and relation to serrun lipid concentrations. Clin Chem. 1993;39(10):2125-9.

Evans AE, Zhang W, Moreel JF, et al. Polymorphisms of the apolipoprotein B
and E genes and their relationship to plasma lipid variables in healthy Chinese
men. Hum Genet. 1993;92(2):191-7.

Schachter F, Faure-Delanef L, Guenot F, et al. Genetic associations with human
longevity at the APOE and ACE loci. Nat Genet. 1994;6(1):29-32.

Louhija J, Miettinen HE, Kontula K, Tikkanen MJ, Miettinen TA, Tilvis RS.
Aging and genetic variation of plasma apolipoproteins. Relative loss of the
apolipoprotein E4 phenotype in centenarians. Arterioscler Thromb.
1994;14(7):1084-9.

Kervinen K, Savolainen MJ, Salokannel J, et al. Apolipoprotein E and B
polymorphisms-longevity factors assessed in nonagenarians. Atherosclerosis.
1994;105(1):89-95.

Cauley JA, Eichner JE, Kamboh Nﬂ, Ferrell RE, Kuller LH. Apo B allele

frequencies in younger (age 42-50) vs older (age 65-90) women. Genet
Epidemiol. 1993;10(1):27-34.

72

81.

82.

83.

84.

85.

86.

87.

88.

89.

90.

91.

92.

93.

Jian-Gang Z, Yong-Xing M, Chuan-Fu W, et al. Apolipoprotein E and longevity
among Han Chinese population. Mech Ageing Dev. 1998;104(2):159-67.

Hirose N, Homma S, Arai Y, et al. [Tokyo Centenarian Study. 4. Apolipoprotein
E phenotype in Japanese centenarians living in the Tokyo Metropolitan area].
Nippon Ronen Igakkai Zasshi. 1997;34(4):267-72.

Bader G, Zuliani G, Kostner GM, F ellin R. Apolipoprotein E polymorphism is not
associated with longevity or disability in a sample of Italian octo- and
nonagenarians. Gerontology. 1998;44(5):293-9.

Thieszen SL, Hixson JE, Nagengast DJ, Wilson JE, McManus BM. Lipid
phenotypes, apolipoprotein genotypes and cardiovascular risk in nonagenarians.
Atherosclerosis. l990;83(2-3): 1 37-46.

Brattstrom L, Zhang Y, Hurtig M, et al. A common methylenetetrahydrofolate
reductase gene mutation and longevity. Atherosclerosis. 1998;141(2):315-9.

Bladbjerg EM, Andersen-Ranberg K, de Maat MP, et al. Longevity is
independent of common variations in genes associated with cardiovascular risk.
Thromb Haemost. 1999;82(3):1100-5.

Wilson PW, Schaefer EJ, Larson MG, Ordovas JM. Apolipoprotein E alleles and
risk of coronary disease. A meta-analysis. Arterioscler Thromb Vasc Biol.
1996 ,16(10). 1250-5.

Kuusisto J, Mykkanen L, Kervinen K, Kesaniemi YA, Laakso M. Apolipoprotein
E4 phenotype is not an important risk factor for coronary heart disease or stroke
in elderly subjects. Arterioscler Thromb Vasc Biol. 1995;15(9):1280-6.

Teasdale GM, Nicoll JA, Murray G, Fiddes M. Association of apolipoprotein E
polymorphism with outcome after head injury. Lancet. 1997;350(9084):1069-71.

Friedman G, F room P, Sazbon L, et a1. Apolipoprotein E-epsilon4 genotype
predicts a poor outcome in survivors of traumatic brain injury. Neurology.
1999;52(2):244-8.

Bertomeu A, Ros E, Zambon D, et al. Apolipoprotein E polymorphism and
gallstones. Gastroenterology. 1996;11 1(6): 1603-10.

Niemi M, Kervinen K, Rantala A, et al. The role of apolipoprotein E and glucose
intolerance in gallstone disease in middle aged subjects. Gut. 1999;44(4):557-62.

Cauley JA, Zmuda JM, Yaffe K, et al. Apolipoprotein E polymorphism: A new

genetic marker of hip ﬁacture risk-The Study of Osteoporotic Fractures. J Bone
Miner Res. 1999;14(7):1175-81.

73

94.

95.

96.

97.

98.

99.

100.

101.

102.

103.

104.

105.

Jahn CE, Oette K, Esser A, von Bergmann K, Leiss 0. Increased prevalence of
apolipoprotein E2 in patients with retinitis pigmentosa. Ophthalmic Res.
1987;19(5):285-8.

Huq L, McLachlan T, Hammer HM, et al. An increased incidence of
apolipoprotein E2/E2 and E4/E4 in retinitis pigrnentosa. Lipids. 1993;28(11):995-
8.

Rocca W, a1. e. Ann Neurol. 1991;19:415-424.

Farrer LA, Cupples LA, Haines JL, et al. Effects of age, sex, and ethnicity on the
association between apolipoprotein E genotype and Alzheimer disease. A meta-
analysis. APOE and Alzheimer Disease Meta Analysis Consortium. Jama.
1997;278(16):1349-56.

Hyman BT, West HL, Rebeck GW, et al. Quantitative analysis of senile plaques
in Alzheimer disease: observation of log-normal size distribution and molecular
epidemiology of differences associated with apolipoprotein E genotype and
trisomy 21 (Down syndrome). Proc Natl Acad Sci U S A. 1995;92(8):3586-90.

Craﬁ S, Teri L, Edland SD, et al. Accelerated decline in apolipoprotein E-
epsilon4 homozygotes with Alzheimer's disease. Neurology. 1998;51(1):149-53.

Malamud N. Neuropathology of organic brain syndromes associated with aging.
In: Gaitz C, ed. Advances in Behavioural Biology. New York: Plenum Press;
1 972.

Schhupf N, Kapell D, Lee J, Ottman R, Mayeux R. Increased risk of Alzheimer's
disease in mothers of adults with Down's syndrome. Lancet. 1994;344:353-356.

Yatham LN, McHale PA, Kinsella A. Down's syndrome and its association with
Alzheimer's disease. Acta Psychiatr Scand. 1988;77(1):38-41.

Van Duijn C, Clayton D, Chandra V, al. e. Familial aggregation of Alzheimer's
disease and related disorders: a collaborative reanalysis of case-contorl studies.
Int J Epidemiol. 1991 ;20(Suppl 2):S 1 3-20.

Heston LL, Mastri AR, Anderson VE, White J. Dementia of the Alzheimer type.
Clinical genetics, natural history, and associated conditions. Arch Gen Psychiatry.
1981;38(10):1085-90.

Heyman A, Wilkinson WE, Hurwitz BJ, et al. Alzheimer's disease: genetic
aspects and associated clinical disorders. Ann Neurol. 1983;14(5):507-15.

74

106.

107.

108.

109.

110.

111.

112.

113.

Mendez MF, Underwood KL, Zander BA, Mastri AR, Sung JH, Frey WH, 2nd.
Risk factors in Alzheimer's disease: a clinicopathologic study. Neurology.
1992;42(4):770-5.

Sadovnick AD, Yee IM, Hirst C. The rate of the Down syndrome among
offspring of women with Alzheimer disease. Psychiatr Genet. 1994;4(2):87-9.

Berr C, Okra-Podrabinek N, F eteanu D, et al. Derrnatoglyphic patterns in
dementia of the Alzheimer type: a case- control study. J Epidemiol Community
Health. 1992;46(5):512-6.

Seltzer B, Sherwin 1. Fingerprint pattern differences in early- and late-onset
primary degenerative dementia. Arch Neural. 1986;43(7):665-8.

Avramopoulos D, Mikkelesen M, Vassilopoulos D, Grigoriadou M, Petersen M.
Apolipoprotein E allele distribution in parents of Down's syndrome children.
Lancet. 1996;347:862-865.

Potter H, Ma J, al. e. Beyond beta-protein: new steps in the pathogenic pathway to
Alzheimer's Disease. In: Iqbal 1, ed. Recent Advances in Alzheimer's Disase and
Related Disorders. Chinchester: John Wiley & Sons; 1995.

Potter H, Geller LN. Alzheimer's disease, Down's syndrome, and chromosome
segregation. Lancet. 1996;348(9019):66.

Schimmel MS, Eidehnan AI, Zadka P, Kombluth E, Hammerrnan C. Increased

parity and risk of trisomy 21: review of 37,1 10 live births. ij.
1997;314(7082):720-1.

75