Note: Descriptions are shown in the official language in which they were submitted.
CA 02718392 2010-09-13
WO 2009/112882 PCT/IB2008/002161
BIOMARKERS FOR ALZHEIMER'S DISEASE
The present invention pertains to the domain of Alzheimer's disease
(AD), and provides new genetic markers for this disease, distributed on 17
genes
initially selected by transcriptomic analysis because of their over- or under-
expression
in the cerebral tissue of AD cases compared to normal individuals. Some of
these
markers can also be used for predicting an increased risk of developing an
early-onset
form of Alzheimer's disease.
AD is already a major problem of public health in our developed
countries. About 5% of the people aged 65 or above are affected with AD and
the
prevalence rises steeply to 19% after age 75 and to 47% after age 85.
In the brain tissue, two anatomo-pathological damages characterize
this pathology:
- an abnormal accumulation of extracellular deposits, mainly
constituted by a peptide of 40 to 42 amino acids called amyloid peptide (or
A13);
- a neuronal intracellular aggregation of modified proteins form the
cytoskeleton under paired helical filaments (PHF).
It is not very probable that a single causal factor is responsible for the
development of these two brain damages. AD appears, indeed, as a
multifactorial and
complex pathology. The appearance and the evolution of this pathology, as
those of
most of the multifactorial diseases, are conditioned by the interaction of
various genetic
susceptibility factors and environmental risk factors.
At the clinical level, although difficult to determine, the age of
appearance of the first symptoms allows to distinguish the relatively rare
early-onset
forms, affecting the patient before the age of 65 years, and the frequent late-
onset forms
arising after 65 years.
Most of the premature forms of AD are hereditary and their
transmission follows a mendelian autosomic dominant mode. Although these cases
represent less than 2 % of all the forms of AD, the discovery of genes
involved in these
familial forms allowed understanding better some pathophysiological mechanisms
(Campion, Dumanchin et al. 1999).
To date, pathogenic mutations were discovered in the amyloid
precursor gene (APP) (protein which metabolism produces the A13 peptide), the
presenilin 1 (PSI) and presenilin 2 (PS2) genes. These mutations were
associated with
a specific increase in the secretion of the A131-42 peptide, the most toxic
form of all
amyloid peptides (Cruts and Van Broeckhoven 1998).
These modifications in the APP metabolism led to propose a
pathophysiological hypothesis based on the concept of the amyloid cascade.
CA 02718392 2010-09-13
WO 2009/112882 PCT/IB2008/002161
2
Although the bibliographical data let think of the existence of 4 major
genes, only the s4 allele of the apolipoprotein E gene (APOE) is recognized as
a genetic
determinant of AD. The expression of this allele is associated with an
increase from 3
to 5 times of the risk of developing AD (Farrer, Cupples et al. 1997).
Encountered difficulties for characterizing new susceptibility genes in
AD are partially related to methodological problems. Indeed, these approaches
get
organized essentially around two axes: genome-wild linkage or linkage
disequilibrium
studies (LD) on familial (early-onset) AD and case-control studies in late-
onset AD
based on a bibliographical gene selection.
Linkage disequilibrium (LD) studies are very well adapted when
looking for pathogenic mutations responsible for early-onset AD forms and
presenting a
monogenic segregation (Kamboh 2004). However, this approach seems less
appropriate
when there are several implied genetic factors, for which impact differs from
one family
to the other, especially because of the implication of environmental factors.
Such LD
studies are made by genomic screening. This technique consists in studying the
segregation of genetic markers regularly spaced out along the genome through
several
generations. This screening allows defining loci of interest susceptible to
contain a gene
implicated in the pathology.
However, the current studies based on this methodology and using
populations suffering from early-onset AD forms lead to define chromosomal
regions
of interest sometimes extending over more than 60 centimorgan (Roberts,
MacLean et
al. 1999). These regions contain hundreds of genes and the systematic search
for
mutations or for pathogenic polymorphisms is consequently a very heavy task.
The use of case-control studies can also seem very effective to bring
to light interactions between studied factors (genetic or environmental)
and/or restricted
effects on the risk of developing the disease (Bertram, McQueen et al. 2007).
However,
this approach presents numerous problems as for the selection of the studied
genes, for
the quality and the size of studied populations and for the functionality of
the
incriminated polymorphisms. This latter point is indeed crucial because,
further to the
complete sequencing of the human genome, a considerable number of
polymorphisms
are now available in databases for this kind of studies.
Hence, despite the intrinsic qualities of each method, the obtained
results could not facilitate the identification of new genetic markers of AD.
In this context, the inventors developed filters in order to decrease the
number of genes and thus polymorphisms to be studied. Their original approach,
described in more details in the experimental part which follows, is based on
the
hypothesis that new genetic markers of AD could be identified by analysing the
CA 02718392 2010-09-13
WO 2009/112882 PCT/IB2008/002161
3
differential expression of genes located in chromosomal regions of interest
(Bensemain,
Hot et al. 2007).
By performing such an analysis on two sub-populations of individuals
of Caucasian origin (a French cohort of 1370 individuals and an American
cohort of
1700 individuals), the inventors identified single nucleotide polymorphisms
(or Tag-
SNPs), located in 17 different genes, which are associated with the risk of
developing
Alzheimer's disease. These genes are indicated in Table 1 below, and the Tag-
SNPs are
detailed in Table 2.
Number of Association
Gene Abbreviation Ref. Seq Chr. associated with age at
SNPs onset
Myosin X MYO10 NM-012334 5 24 4
Integrin alpha 2 ITGA2 NM 002203 5 3
precursor -
Butyrophilin, subfamily BTN3A2 NM 007047 6 2
3, member A2 -
Tenascin XB isoform 1 TNXB NM 019105 6 4
Advanced glycosylation AGER NM 001136 6 4
end product-specific -
Major
histocompatibility HLA-DRA NM019111 6 5
complex, class II, DR
PHF 1 PHF 1 NM 002636 6 3
Cyclin D3 CCDN3 NM_001760 6 3
Interleukin 33 /
Nuclear factor from IL-33/NF- NM 033439 9 1
high endothelial HEV -
venules
FLJ20375 or FLJ20375 AK000382 9 9
KIAA 1797
Beta-1,4- B4GALT1 NM 001497 9 5
galactosyltransferase 1 -
Ciliary neurotrophic CNTFR NM 001842 9 7 2
factor receptor -
Carbonic anhydrase IX CA9 NM-00 1216 9 1
precursor
KIAA1462 KIAA1462 NM020848 10 3 5
Calcium/calmodulin-
dependent protein CAMK2G NM_001222 10 7
kinase II gamma
Angiopoietin 4 ANGPTL4 NM_009641 20 6 5
Spermine oxidase SMOX NM 175840 20 2
Table 1: list of genes having at least one Tag-SNP associated with the risk of
i o developing Alzheimer's disease.
CA 02718392 2010-09-13
WO 2009/112882 PCT/IB2008/002161
4
12+22 22 Co-
SNP Chr. Gene 1/2 versus 11 versus dominant
12+11 model
rs153202 5 myosin X C / T 0,005 ns 0,01
rs27430 5 myosin X T / C 0,008 0,005 0,002
rs2401987 5 myosin X A / G 0,05 ns ns
rs26315 5 myosin X A / T 0,05 ns ns
rs6881621 5 myosin X C / T ns 0,04 ns
rs1445946 5 myosin X A / G 0,009 ns 0,05
rs13356962 5 myosin X A / G ns 0,01 ns
rs250339 5 myosin X A / G 0,03 ns 0,04
rs 10051929 5 myosin X A/C ns ns 0,03
rs12520877 5 myosin X A / G ns ns 0,04
rs4702173 5 myosin X C / T ns 0,04 ns
rs17651023 5 myosin X C / T 0,05 ns ns
rs6898592 5 myosin X C / T 0,01 ns ns
rs17707609 5 myosin X C / G ns 0,007 ns
rs17651165 5 myosin X C / T 0,007 ns ns
rs31505 5 myosin X A / G 0,02 ns 0,05
rs17614059 5 myosin X C / G 0,01 ns 0,02
rs 17651266 5 myosin X C / T 0.05 ns 0,05
rs40985 5 myosin X A / G 0,007 ns 0,007
rs716555 5 myosin X C / T Ns 0,02 0,03
rs253315 5 myosin X A/C 0,009 0,02 0,002
rs2560852 5 myosin X C / G 0,002 ns 0,001
rs7710976 5 myosin X C / G 0,008 ns ns
rs153709 5 myosin X A/C ns 0,04 ns
rs716436 5 ITGA2 C / T ns 0,009 ns
rs4865756 5 ITGA2 A / G 0,02 ns 0,03
rs11741738 5 ITGA2 A / G ns 0,02 ns
rs9467731 6 BTN3A2 C / T ns 0,01 ns
rs3757142 6 BTN3A2 A / G ns 0,009 ns
rs3130285 6 TNXB C / T 0,003 ns 0,009
rs3130286 6 TNXB C / T 0,006 ns 0,01
rs185819 6 TNXB T / C ns 0,03 ns
rs3130287 6 TNXB C / T ns 0,002 0,004
rs3134943 6 AGER A / G ns 0,005 0,007
rs8365 6 AGER C / G 0,03 ns ns
rs3134940 6 AGER A / G 0,04 ns ns
rs1800684 6 AGER A / T 0,008 ns 0,01
rs3129876 6 HLA-DRA A / G ns 0,008 0,009
rs3129881 6 HLA-DRA C / T 0,03 ns 0,05
rs9268658 6 HLA-DRA G / A 0.05 0,01 0,008
rs8084 6 HLA-DRA A / C ns 0,009 ns
CA 02718392 2010-09-13
WO 2009/112882 PCT/IB2008/002161
rs7192 6 HLA-DRA G / T 0,004 ns 0,03
rs3116713 6 PHF1 A / G ns ns 0,04
rs442745 6 PHF 1 A / G ns 0,007 0,02
rs3106196 6 PHF 1 G / T ns 0,05 0,03
rs3218114 6 cyclin D3 A / G ns 0,003 0,002
rs9529 6 cyclin D3 A / G 0.05 ns ns
rs3218086 6 cyclin D3 A / G ns 0,03 0,02
rs7044343 9 IL-33 T / C 0.00003 ns 0.00001
rs10757145 9 KIAA1797 A / G ns 0,05 ns
rs1332322 9 KIAA1797 A/C ns ns 0,04
rs4978111 9 KIAA1797 A / G 0,02 ns 0,01
rs7860490 9 KIAA1797 G / T ns 0,04 0,05
rs7033259 9 KIAA1797 A / T 0,05 ns ns
rs6475473 9 KIAA1797 A/C 0,002 0,02 0,002
rs6475474 9 KIAA1797 C / T ns 0,02 0,01
rs 1081143 8 9 KIAA 1797 C / G ns ns 0,04
rs10964779 9 KIAA1797 A / G 0,03 ns ns
rs3780491 9 B4GALT1 C / T ns ns 0,03
rs10511909 9 B4GALT1 C / G 0,01 ns 0,008
rs2774272 9 B4GALT1 A / G ns 0.05 0,02
rs10758194 9 B4GALT1 G / T ns 0,04 0,02
rs1328898 9 B4GALT1 C / T 0,05 ns 0,02
rs2274592 9 CNTFR C / T 0,04 ns 0,03
rs2381164 9 CNTFR C / G 0,008 0,02 0,002
rs10972149 9 CNTFR A / G 0,0008 ns 0,04
rs 10814123 9 CNTFR C / T ns 0,001 0,05
rs10758268 9 CNTFR G / T ns 0,0003 0,001
rs12551429 9 CNTFR C / G ns 0,05 ns
rs3763613 9 CNTFR A/C ns 0,03 ns
rs3829078 9 CA9 A / G 0,01 ns 0,01
rs6481654 10 KIAA1462 A / G 0,02 ns ns
rs9988732 10 KIAA1462 A / G ns 0,003 0,01
rs2488024 10 KIAA1462 G / A 0,006 ns 0,02
rs11000780 10 CAMK2G A / G ns 0,05 ns
rs 10824037 10 CAMK2G A / G ns 0,05 ns
rs10458656 10 CAMK2G A / G 0,05 ns ns
rs2633310 10 CAMK2G G / T 0,02 ns 0,02
rs2675662 10 CAMK2G C / T ns 0,02 0,02
rs2459446 10 CAMK2G C / T 0,01 ns 0,02
rs2688614 10 CAMK2G C / T ns 0,007 0,005
rs6055551 20 ANGPTL4 C / T ns 0,04 ns
rs3787566 20 ANGPTL4 C / T ns 0,007 ns
rs4816050 20 ANGPTL4 C / T ns 0,002 ns
rs13040505 20 ANGPTL4 C / G 0,02 ns 0,01
CA 02718392 2010-09-13
WO 2009/112882 PCT/IB2008/002161
6
rs730169 20 ANGPTL4 C / G ns 0,03 0,04
rs574628 20 ANGPTL4 A / G 0,04 ns ns
rs1741296 20 SMOX C / T ns 0,004 0,03
rs6076623 20 SMOX G / T 0,01 ns 0,04
Table 2: list of the SNPs showing a significant association with the risk of
developing
AD based on statistical analyses with recessive, dominant or co-dominant
models. P
was adjusted on the age, gender kind and the genetic status of APOE gene. The
allele
indicated in bold is the more frequent one.
The inventors also identified 16 SNPs which are associated with an
increased risk, for an individual, of developing an earlier-onset form of
Alzheimer's
disease. This means that people having, for at least one of these SNPs, an
unfavourable
form of said polymorphism, statistically develop Alheimer's disease at least
one year
earlier than other people developing the disease. These SNPs are located in 4
different
genes, as disclosed in Table 3 below.
Gene ANGPTL4 MYO 10 CNTFR KIAA1462
SNPs rs1124740 rs17651266 rs6476454 rs1063205
rs6118095 rs4702179 rs3763615 rs2478839
rs4813852 rs2562343 rs7071535
rs4816051 rs152342 rs2014144
rs6118137 rs9337951
Table 3 : SNPs associated with the age at onset.
Accordingly, the invention pertains to the use of at least one single
nucleotide polymorphism (SNP) selected in the group consisting of rs 153202,
rs27430,
rs2401987, rs26315, rs6881621, rs1445946, rs13356962, rs250339, rsl0051929,
rs 12520877, rs4702173, rs 17651023, rs6898592, rs 17707609, rs 17651165,
rs31505,
rs17614059, rsl7651266, rs40985, rs716555, rs253315, rs2560852, rs7710976,
rsl53709, rs716436, rs4865756, rs11741738, rs9467731, rs3757142, rs3130285,
rs3130286, rsl85819, rs3130287, rs3134943, rs8365, rs3134940, rs1800684,
rs3129876, rs3129881, rs9268658, rs8084, rs7192, rs3116713, rs442745,
rs3106196,
rs3218114, rs9529, rs3218086, rs7044343, rs10757145, rs1332322, rs4978111,
rs7860490, rs7033259, rs6475473, rs6475474, rs10811438, rs10964779, rs3780491,
rs10511909, rs2774272, rs10758194, rs1328898, rs2274592, rs2381164,
rsl0972149,
rs10814123, rs10758268, rs12551429, rs3763613, rs3829078, rs6481654,
rs9988732,
rs2488024, rsll000780, rs10824037, rs10458656, rs2633310, rs2675662,
rs2459446,
rs2688614, rs6055551, rs3787566, rs4816050, rs13040505, rs730169, rs574628,
rsl741296, rs6076623, rs1124740, rs6118095, rs4813852, rs4816051, rs6118137,
rs4702179, rs2562343, rsl52342, rs6476454, rs3763615, rs1063205, rs2478839,
rs7071535, rs2014144 and rs9337951, as a genetic marker for determining the
genetic
predisposition of an individual to Alzheimer's disease. In a preferred
embodiment of the
CA 02718392 2010-09-13
WO 2009/112882 PCT/IB2008/002161
7
invention, a group of 2 to 5, 6 to 10, 11 to 20 or even more than 20 SNPs
selected in the
above list is used as a genetic predisposition marker.
The present invention also pertains to the use of at least one SNP
selected in the group consisting of rs1124740, rs6118095, rs4813852,
rs4816051,
rs6118137, rs17651266, rs4702179, rs2562343, rs152342, rs6476454, rs3763615,
rs1063205, rs2478839, rs7071535, rs2014144 and rs9337951, as a genetic marker
for
determining the genetic predisposition of an individual to develop an early-
onset form
of Alzheimer's disease. Groups of 2, 3, 4, 5, 6 to 10 or even more SNPs
selected in the
above list can also be used to determine the predisposition of an individual
to develop
an early-onset form of Alzheimer's disease.
According to a particular embodiment, the present invention relates to
a method for in vitro predicting an increased risk, for an individual, of
developing
Alzheimer's disease, comprising a step of performing a genotyping assay of at
least one
gene selected in the group consisting of myosin X, integrin alpha 2 precursor,
NM 007047, tenascin XB isoform 1, advanced glycosylation end product-specific,
major histocompatibility complex, class II, DR, NM_002636, cyclin D3,
Interleukin-33,
KIAA1797, NM_001497, ciliary neurotrophic factor receptor, carbonic anhydrase
IX
precursor, KIAA1462, calcium/calmodulin-dependent protein kinase II gamma,
angiopoietin 4 and SMOX, in a biological sample from said individual.
An example of suitable sample for the SNP genotyping is blood. The
SNP genotyping assay can be performed using any SNP analysis method known in
the
art. For example, hybridization methods can be used, with technologies such as
macro-
or micro-array, or with technologies based on probes which are specific for
the
changing base of the SNPs of interest (e.g., Tagman probes, Molecular
beacons,
Scorpion probes etc.).Other methods, such as restriction fragment length
polymorphism
(RFLP) methods, can also be used, as well as chromatography (dHPLC), capillary
electrophoresis sequencing, mass spectrometry, or specific PCR. Various
methods, such
as SNPplexTM, Amplifluor SNPs Genotyping System and single base extension
(SBE)
followed by tag-array hybridization and allele-specific primer extension, are
commercially available for performing SNP genotyping assays.
In particular, the above method can be performed by genotyping at
least one SNP selected in the group consisting of rs153202, rs27430,
rs2401987,
rs26315, rs6881621, rs1445946, rs13356962, rs250339, rs10051929, rs12520877,
rs4702173, rs17651023, rs6898592, rs17707609, rs17651165, rs31505, rs17614059,
rs17651266, rs40985, rs716555, rs253315, rs2560852, rs7710976, rs153709,
rs716436,
rs4865756, rs11741738, rs9467731, rs3757142, rs3130285, rs3130286, rs185819,
rs3130287, rs3134943, rs8365, rs3134940, rs1800684, rs3129876, rs3129881,
CA 02718392 2010-09-13
WO 2009/112882 PCT/IB2008/002161
8
rs9268658, rs8084, rs7192, rs3116713, rs442745, rs3106196, rs3218114, rs9529,
rs3218086, rs7044343, rs10757145, rs1332322, rs4978111, rs7860490, rs7033259,
rs6475473, rs6475474, rs10811438, rs10964779, rs3780491, rs10511909,
rs2774272,
rs10758194, rs1328898, rs2274592, rs2381164, rsl0972149, rs10814123,
rs10758268,
rs12551429, rs3763613, rs3829078, rs6481654, rs9988732, rs2488024, rs11000780,
rs10824037, rs10458656, rs2633310, rs2675662, rs2459446, rs2688614, rs6055551,
rs3787566, rs4816050, rs13040505, rs730169, rs574628, rs1741296 and rs6076623.
Of
course, the skilled artisan can chose to genotype a determined group of 2, 3,
4, 5, 6-10,
11-20 or more than 20 SNPs. When performing this method, the skilled artisan
can also
attribute a weight to each of the genotyped SNPs, in order to obtain a score
integrating
the results of all SNPs genotypes. Of course, the skilled artisan can also
combine the
information obtained by SNP genotyping as described above with other
information
such as, for example, information concerning the e4 allele in said individual.
The present invention also pertains to a method for in vitro predicting
an increased risk, for an individual, of developing an earlier-onset form of
Alzheimer's
disease, comprising a step of performing a genotyping assay of at least one
gene
selected in the group consisting of myosin X, ciliary neurotrophic factor
receptor,
KIAA1462 and angiopoietin 4. in particular, this method can be performed by
genotyping one, 2, 3, 4, 6 to 10 or even more SNP(s) selected in the group
consisting of
rs1124740, rs6118095, rs4813852, rs4816051, rs6118137, rsl7651266, rs4702179,
rs2562343, rsl52342, rs6476454, rs3763615, rs1063205, rs2478839, rs7071535,
rs2014144 and rs9337951.
By "earlier-onset form of Alzheimer's disease" is meant that an
individual having a marker as mentioned above will statistically develop
Alzheimer's
disease at least one year earlier than a person not having this marker.
In a preferred embodiment, the above methods are performed with a
sample from a Caucasian individual.
Another aspect of the present invention is a kit for performing a
method as above described, comprising means for genotyping at least one SNP
selected
in the group consisting of rs153202, rs27430, rs2401987, rs26315, rs6881621,
rs1445946, rs13356962, rs250339, rs10051929, rsl2520877, rs4702173,
rsl7651023,
rs6898592, rs17707609, rs17651165, rs31505, rs17614059, rs17651266, rs40985,
rs716555, rs253315, rs2560852, rs7710976, rs153709, rs716436, rs4865756,
rs11741738, rs9467731, rs3757142, rs3130285, rs3130286, rs185819, rs3130287,
rs3134943, rs8365, rs3134940, rs1800684, rs3129876, rs3129881, rs9268658,
rs8084,
rs7192, rs3116713, rs442745, rs3106196, rs3218114, rs9529, rs3218086,
rs7044343,
rs10757145, rs1332322, rs4978111, rs7860490, rs7033259, rs6475473, rs6475474,
CA 02718392 2010-09-13
WO 2009/112882 PCT/IB2008/002161
9
rs10811438, rs10964779, rs3780491, rs10511909, rs2774272, rs10758194,
rs1328898,
rs2274592, rs2381164, rs10972149, rs10814123, rs10758268, rs12551429,
rs3763613,
rs3829078, rs6481654, rs9988732, rs2488024, rs11000780, rs10824037,
rs10458656,
rs2633310, rs2675662, rs2459446, rs2688614, rs6055551, rs3787566, rs4816050,
rs13040505, rs730169, rs574628, rs1741296, rs6076623, rs1124740, rs6118095,
rs4813852, rs4816051, rs6118137, rs4702179, rs2562343, rs152342, rs6476454,
rs3763615, rs1063205, rs2478839, rs7071535, rs2014144 and rs9337951, and a
notice
of use mentioning the relevance of said SNP(s) in the prognosis of Alzheimer's
disease.
In a preferred embodiment of said kit, the kit comprises means for
genotyping at least two, and more preferably at least 5 SNPs selected in the
above list.
Non-limitative examples of genotyping means which can be included
in a kit according to the invention are oligonucleotides and restriction
enzymes specific
for said SNP(s). Oligonucleotide(s) included in the kit can be bound to a
solid support,
for example in a micro- or macro-array. Alternatively or complementarily,
oligonucleotide(s) included in the kit can be labelled (for example TagMan
probes,
molecular beacons, Scorpion probes, etc.). Reagents for performing HPLC, RFLP,
capillary electrophoresis, specific PCR (such as SNPplexTM or Amplifluor SNP
genotyping), mass spectrometry etc., can also be included in the kit.
Other characteristics of the invention will also become apparent in the
course of the description which follows of the biological assays which have
been
performed in the framework of the invention and which provide it with the
required
experimental support, without limiting its scope.
Examples
Materials and methods
Populations
Both analyzed sub-populations were obtained by drawing lots from
two case-control studies, a French one (n=1370) and an American one (n=1700).
The
characteristics of these two sub-populations are described in table 4 below.
French case- American case-control
control sub-population
AD controls AD cases controls
n 307 238 200 200
Mean age' 74.1 73.1 8.5 77.2 6.1 74.3 5.8
Mean age at 68.6 - 72.0 -
% of men 35 43 42 38
Table 4
CA 02718392 2010-09-13
WO 2009/112882 PCT/IB2008/002161
All samples were from individuals of Caucasian origin. The clinical
diagnoses were made according to the criteria NINCDS / ADRDA and DSM-III-R.
Controls were recruited in the same geographical region as individuals
affected by AD
(in the North of France or in the East-region of Pennsylvania) and did not
present
5 criteria of dementia at the time of the recruitment.
Chip-Genotyping
Using HapMap database (http://www.hapmap.org), a set of 1156
polymorphisms was selected on 82 genes. These polymorphisms are considered as
being Tag-SNPs, allowing to define the most complete genetic information
(frequency
10 of the rare allele > 10 % and r2 < 0.8).
The experiments were realized by a service provider, DNAvision
company (Belgium), as described by the supplier (Affymetrix company) by using
a
GeneChip oven 640 hybridization station and a GeneChip 3000 7G 4C scanner.
Among all analyzed SNPs, only twenty one SNPs (1.8 %) could not
be genotyped due to technological problems. All other SNPs were genotyped with
a rate
of success greater than 90 %. Finally, seventy SNPs (6 %) were not in Hardy-
Weinberg
equilibrium (p < 0.05), the majority of them weakly differing from this
equilibrium
(60 %, with a probability included between 0.01 and 0.05). Besides, an
important
proportion of these SNPs were located inside or near a potential CNV (copy
Number
Variation).
Statistical Analyses
The software SAS 8.02 was used (SAS institute, Cary, NC). The
association of the 1156 SNPs with the risk of developing AD was estimated by
logistic
regression and adjustments on the age, the sex, the presence or the absence of
allele e4
and the recruitment centre. Three models were systematically tested:
recessive, co-
dominant and dominant.
The model 12+22 versus 11 corresponds to a dominant model (that is
at least one copy of an allele is sufficient to induce a modification of the
risk).
The model 22 versus 11+12 corresponds to a recessive model (that is
two copies of one allele are necessary to induce a modification of the risk)
Co-predominant model assume that the more the number of copy is
important, the more the risk increase.
Results and conclusions
Genes Selection by transcriptomic studies
Previous linkage-disequilibrium LD studies allowed determining
chromosomal regions susceptible to contain one or several genetic markers of
AD.
These regions extend sometimes over more than 95 cM.
CA 02718392 2010-09-13
WO 2009/112882 PCT/IB2008/002161
11
From these data was developed within the laboratory a biochip
allowing the study of the expression of 2741 ORFs ("Open Reading Frame")
contained
in these chromosome regions of interest (Lambert, Testa et al. 2003). The
level of
expression of these ORFs was compared in cerebral tissues resulting from 12
individuals affected by AD and from 12 controls (selected on the quality of
extracted
total ARN).
107 genes, differentially expressed in AD brains compared to control
one, were consequently identified (table 5) (Bensemain, Hot et al. 2007).
Locus Distance in Selected ORFs Differentially expressed
cm ORFs
Chr. 1 50 393 13
Chr.5 50 174 6
Chr.6 40 535 24
Chr. 9 55 230 12
Chr. 10 95 415 15
Chr. 12 40 306 11
Chr. 20 50 239 9
Chr. 21 58 267 6
Chr. X 25 182 11
Table 5 : number of differentially expressed genes in the various chromosomal
regions of
interest.
Association of identified genes with the risk of developing AD
To accelerate the analysis of the genetic association of these genes
with the risk of developing AD, a genotyping chip with an average flow
(Affymetrix)
was thus developed. From 82 genes, 1156 Tag-SNPs were selected on the Hapmap
site,
as mentioned above.
These Tag-SNPs were analyzed on two sub-populations, French
(n=545) and American (n=400), obtained by drawing lots from more important
populations. Seventy Tag-SNPs were not in Hardy-Weinberg equilibrium in the
control
combined population (French and American), most of them weakly differing from
this
balance (60 % between 0.01 < p < 0.05).
Among remainders Tag-SNPs and according to the three statistical
models that were used (recessive, co-dominant and dominant), 89 of these Tag-
SNPs
located in 17 different genes presented at least one polymorphism associated
with the
risk of developing AD in the combined population and this in at least one of
the three
used models (p < 0.05). Besides, 4 of these genes also present Tag-SNPs
presenting the
characteristic to modulate the age of appearance of the pathology. (cf table 6
below and
table 2 above).
CA 02718392 2010-09-13
WO 2009/112882 PCT/IB2008/002161
12
Chromosomal Differential Number Number Association
Gene Ref. Seq Chr. locations expression b of of with age at
studied associated
SNPs SNPs onset
16,715,016- +201% 103
MYO10 NM_012334 5 16,989,385 24 4
52,320,913- + 215% 26
ITGA2 NM_002203 5 52,426,365 3
26,473,377-
BTN3A2 NM 007047 6 26,486,525 - 45% 13 2
32,116,911- + 100% 31
TNXB NM 019105 6 4
32,185,131
32,256,724-
_
AGER NM 001136 6 32,260,001 - 53% 7 4
32,515,625-
-
HLA-DRA NM 019111 6 32,520,799 - 57% 10 5
33,486,934-
_
PHFI NM 002636 6 33,492,187 - 47% 4 3
42,010,650-
-
CCDN3 NM 001760 6 42,017,530 + 157% 9 3
IL-33/NF- 6,231,678-
- 66/0 4
HEV -033439 9 6,247,982 1
20,939,639-
FLJ20375 AK000382 9 20,985,947 - 62% 55 9
33,100,642-
B4GALT1 NM 001497 9 33,157,231 -43% 9 5
34,541,431- +132 % 16
CNTFR NM_oo 1842 9 34,579,722 7 2
35,663,915-
_
CA9 NM001216 9 35,671,152 -50% 11 1
30,343,390- + 262% 18 KIAA 1462 NM-020848 10 30,376,806 3 5
75,242,265-
_
CAMK2G NM 001222 10 75,304,344 -50% 10 7
801,604-
ANGPTL4 NM-009641 20 844,605 - 65% 42 6 5
4,103,675-
_
SMOX NM 175840 20 4,116,352 - 42% 15 2
Table 6
The fact that SNPs different from those associated to the risk of
developing AD, are linked to different age of AD onset result from the
combination of
numerous factors:
= The statistical tools that have been used: logistical decline
(qualitative / qualitative) versus analysis of the variance (qualitative /
quantitative)
CA 02718392 2010-09-13
WO 2009/112882 PCT/IB2008/002161
13
= The statistical models that have been used (recessive,
predominant or co-predominant)
= The frequency of studied SNPs.
= The variability in the measurement of the age of onset and its
heterogeneity between the different individuals
= The linkage disequilibrium more or less important between
different SNPs
Conclusion
This integrated approach, based on the study of genes i) located in the
chromosomal regions of interest defined by genomic screening and ii)
differentially
expressed in AD brains compared to control brains, allowed the identification
of 17
susceptibility genes for AD. The analysis of separated and/or combined genetic
variants
of the genes which have been identified as associated with the risk of
developing AD
according to the present study will allow developing tools for the diagnosis
and
prognosis of AD.
CA 02718392 2010-09-13
WO 2009/112882 PCT/IB2008/002161
14
REFERENCES
Bensemain, F., D. Hot, et al. (2007). "Evidence for induction of the
ornithine transcarbamylase expression in Alzheimer's disease." Mol Psychiatry.
Bertram, L., M. B. McQueen, et al. (2007). "Systematic meta-analyses
of Alzheimer disease genetic association studies: the A1zGene database." Nat
Genet
39(1): 17-23.
Campion, D., C. Dumanchin, et al. (1999). "Early-onset autosomal
dominant Alzheimer disease: prevalence, genetic heterogeneity, and mutation
1o spectrum." Am J Hum Genet 65(3): 664-70.
Cruts, M. and C. Van Broeckhoven (1998). "Molecular genetics of
Alzheimer's disease." Ann Med 30(6): 560-5.
Farrer, L. A., L. A. Cupples, et al. (1997). "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
278(16): 1349-56.
Kamboh, M. I. (2004). "Molecular genetics of late-onset Alzheimer's
disease." Ann Hum Genet 68(Pt 4): 381-404.
Lambert, J. C., E. Testa, et al. (2003). "Relevance and limitations of
public databases for microarray design: a critical approach to gene
predictions."
Pharmacogenomics J 3(4): 235-41.
Roberts, S. B., C. J. MacLean, et al. (1999). "Replication of linkage
studies of complex traits: an examination of variation in location estimates."
Am J Hum
Genet 65(3): 876-84.
