Whole genome analysis in a consanguineous family with early onset Alzheimer’s disease

Abstract

Early-onset Alzheimer’s disease (EOAD) is a clinically and genetically heterogeneous condition in which the typical features appear significantly earlier in life (before 65 years). Mutations in three genes (PSEN1, PSEN2, and APP) have been identified in autosomal dominant forms of EOAD. However, in about 50% of Mendelian cases and in most of the sporadic EOAD patients, no mutations have been found. We present clinical characteristics of an Israeli family comprising two affected siblings with EOAD born to neurologically healthy parents who were first cousins (both parents died after 90 years old). Sequence analysis of PSEN1, PSEN2, APP, TAU, PGRN, and PRNP failed to reveal any mutations in the affected siblings. Because the disease in this family is consistent with an autosomal recessive mode of inheritance we identified all homozygous regions identical by descent (IBD) in both siblings, by high-density SNP genotyping. We provide here the first catalog of autozygosity in EOAD and suggest that the regions identified are excellent candidate loci for a recessive genetic lesion causing this disease.

1. Introduction

Only in 1–2% of all Alzheimer’s disease (AD) cases, the disease appears before the fifth decade of life (<60–65 years of age) and is known as early-onset AD (EOAD), In the majority, but not all, EOAD patients, the disease aggregates within families, and ~10% of these families show an autosomal dominant pattern of inheritance. To date, three genes have been unequivocally related to these familial forms of EOAD: the presenilin 1 gene (PSEN1), accounting for 15–50% of the cases, the amyloid precursor protein (APP), accounting for 5% of cases, and the presenilin 2 (PSEN2), mutations of which have been identified in less than 1% of EOAD (for review see Rademakers et al., 2003). However, despite the great effort that has been made to find other genetic causes, no other Mendelian gene has been found in the last 11 years.

The ε4 allele of the Apolipoprotein E (APOE) gene has been related to both familial and sporadic EOAD. Several studies have shown that the risk associated with APOE-ε4 is dose dependent, and correlates with the age at onset. That is, individuals with two APOE-ε4 alleles have an earlier age of onset than those without APOE-ε4. In total, the proportion of patients with dementia that is attributable to APOE-ε4 has been estimated to be 7–20% (Saunders et al., 1993; Strittmatter et al., 1993; van Duijn et al., 1994).

In summary, although genetic causes for EOAD have been found, a substantial proportion of sporadic and familial EOAD cases are of unknown genetic etiology (Cruts et al., 1998; Campion et al., 1999). Surprisingly, there has been no formal assessment of the possibility that a proportion of cases have a simple recessive aetiology. Although there have been two studies of isolated populations which have had a high incidence of disease, the primary analyses have been using dominant or additive modes of inheritance (Farrer et al., 2003; Liu et al., 2007). We describe a consanguineous family in which two siblings, sons of a first-cousin, neurologically healthy parents, suffered from EOAD. We catalog the homozygous genomic regions identical by descent (IBD) in both patients.

2. Patients and methods

2.1. Patients

We analyzed a Jewish Israeli family originating from Morocco (Fig. 1). The family was composed of seven siblings born to first-degree consanguineous parents who died after the age of 90 years without any sign of cognitive impairment. Two sibs suffered from EOAD and agreed to participate in the genetic study. The others refused to give any blood sample. Written informed consent was obtained from both patients.

Fig. 1.

Pedigree of the Israeli family. Squares represent men, circles represent women: diamonds are used in the present generation to obscure identity. Black filling indicates early-onset Alzheimer’s disease. Actual ages (and age of death of one of the siblings) are shown underneath. Ages at onset are indicated underneath.

2.2. Genetic analyses

After informed consent was obtained, blood sample was collected and DNA extraction was performed by standard methods. Mutations in PSEN1 (exons 3–12). PSEN2 (exons 3–12), APP (exons 16 and 17), TAU, PGRN, and PRNP were discarded by direct sequencing.

Genome wide SNP typing using the Illumina HumanHap240S Genotyping was performed for the two siblings. A total of 241,848 SNPs (236,852 located in autosomes) were genotyped for each sample. The experiments were carried out as per the manufacturer’s instructions. Genotype success rate greater than 95% was obtained for both assays. Visualization of homozygosity tracks and gene copy variation was performed through Genome Viewer tool within Beadstudio v2.2.22 (Illumina Inc., San Diego, CA). B allele frequency and log R ratio were analyzed. Briefly, B allele gives an estimate of the proportion of times an individual allele is called A or B; thus an individual homozygous for the B allele would have a score close to 1, an individual homozygous for the A allele a score close to 0, and a score of 0.5 would indicate a heterozygous genotype. Value of the log R ratio is the log (base 2) ratio of the observed normalized R value for the individual SNP divided by the expected normalized R value for the SNPs theta value. The ratio of observed R to expected R for any SNP yields an indirect measure of the binding efficiency of detected alleles for each polymorphism and therefore, gives an estimate of genomic copy number. An R above 1 is indicative of an increase in copy number (duplication or triplication), and values below 1 suggest a deletion.

3. Results

3.1. Family description and clinical characteristics

The family consisted of seven siblings from a first-cousin marriage (Fig. 1). At the time of the clinical study the mother was 90 years old with mild gait difficulties but otherwise healthy, with no cognitive loss. We are aware that she died after few years. The father died at age 90 due to cerebral stroke. The proband is a 64-year-old man who was born in Morocco and migrated to Israel in 1956. At the age of 58 he presented with a 9-month history of insidious, progressive impairment of memory first noticed by the family. Early symptoms were repeated questioning, forgetting names and loss of interest in his family. Other features included an inability to plan or order tasks, for which he was fired from his work in a tire factory. During the neurological examination, no involuntary movements, pyramidal, extrapyramidal, or cerebellar signs were found. Primitive frontal release signs were absent. Two months after the initial visit the Mini-Mental State Examination score was 9/30. Cognitive examination revealed diffuse cognitive impairment, most notably of short and long-term memory, verbal and visual recall, and visuospatial abilities. Calculation, attention tasks, and information processing speed were markedly impaired. In addition, he had moderate dressing apraxia and right–left disorientation. Other neuropsychological examinations (i.e., Wechsler test, word naming and general knowledge) were difficult to perform due to advanced dementia. According to the NINCDS-ADRDA criteria the clinical examination and neuropsychological profile was consistent with EOAD (McKhann et al., 1984). The diagnostic workup included brain MRI that showed generalized, symmetrical cerebral atrophy, and an EEG with diffuse mild slowness at the range of 5–7 Hz. The lumbar puncture showed normal protein and glucose levels and PCR from the CSF was negative for tuberculosis. One of his siblings started with clinical symptoms at the age of 58, when he was laid-off from work due to cognitive decline. One year later, he was institutionalized in a psychiatric institution with a clinical diagnosis of EOAD. He died at the age of 70 and did not present any extrapyramidal sign during the course of disease. The actual ages of the other sibs are 46, 55, 57, 64, and 66. All of them are neurologically healthy. APOE genotyping revealed that both siblings carried an APOE-ε3ε4 genotype.

3.2. Whole genome genotyping

We looked at both log R ratio and B allele frequency plots across the whole genome in both siblings with EOAD, This high-throughput genotyping technology allowed a rapid production of high-quality, ultra-dense genotypes and provided direct visualization of extended tracks of homozygosity (Gibbs and Singleton, 2006). Information of all the homozygozity genomic regions greater than 1 MB that are shared by both siblings is presented in Table 1. Also, a diagram showing the homozygous regions of the entire genome of both siblings is depicted in Fig. 2. There were no regions of homozygosity IBD greater than 1 MB on chromosomes 7, 12, 15, 18, 19, 20, 21, and 22. Also, regions such as 14q24.3 (PSEN1), 1q31-q42 (PSEN2), 21q21 (APP), 17q21.1 (MAPT), 10q23-q25 (IDE), 3q21-q27 (MME), lp36.1 (ECE), or 1 1q23.2-24.2 (SORL1) did not show shared homozygosity. The total length of genomic homozygosity that was shared by both siblings was of ~250 MB, which represents ~8.32% of the genome, with the longest track of homozygosity in chromosome 11 (with 59.1 MB). This is considerably longer than would be expected (~2%) and suggests, unsurprisingly, that there has been additional consanguinity in previous generations. There were no duplications or triplications in either of the genomes (data not shown). Three homozygous deletions were found in both siblings on chromosomes 2p22.3, 4q26, and Xp 11.1, with lengths of 7 kb, 6 kb, and 700 base pairs, respectively (Fig. 2). However, all have been previously reported as polymorphic copy number variations of the human genome (Iafrate et al., 2004; McCarroll et al., 2006: Korbel et al., 2007). and are described in the Database of Genomic Variants (http://projects.tcag.ca/variation/). Only deletion in chromosome 2 is located within a gene (RASGRP3), comprising the exons 12 and 13 of this locus. The others are placed in intergenic regions.

Table 1.

Lengths and positions of homozygosity tracks shared by both affected siblings

Chrom.	From (SNP ID)	Position	To (SNP ID)	Position	Total homozygosity length (MB)	Genes in Alzene found in these regions
1	rs1390484	64,603,980	rs6681765	76,549,653	11.95
1	rs2994809	120,153,500	rs661678	145,445,019	25.29	PRKAB2
1	rs2584315	220,273,510	rs851170	221,416,553	1.14
2	rs13423995	89,769	rs3910069	7,472,223	7.38
2	rs10189659	193,944,304	rs2123738	195,479,306	1.54
3	rs1517520	16,482,329	rs17006228	19,761,798	3.28	RFTN1
3	rs9863064	61,590,243	rs704267	71,784,695	10.19
3	rs1546304	80,618,614	rs893518	81,898,253	1.28
3	rs12630403	103,152,590	rs1107889	116,721,177	13.57
4	rs7672923	110,173,872	rs4975181	129,466,845	19.29	CASP6; FABP2
5	rs6870997	34,587,622	rs7711836	55,344,388	20.76	PRKAA1; HMGCS1; NDUFS4
5	rs17237118	67,441,983	rs7732832	90,645,624	23.20	HMGCR; CRHBP
6	rs9460326	11,382,125	rs13202453	21,091,421	9.71
8	rs10503942	33,300,800	rs16882274	34,447,332	1.15
8	rs6470874	132,053,392	rs7017860	135,220,187	3.17
9	rs10814410	36,587	rs1378324	3,834,391	3.80	VLDLR
10	rs10823828	73,129,862	rs607483	78,895,276	5.77	PSAP; CHST3; PPP3CB; SEC24C; NDST2; CAMK2G; PLAU; VCL; AP3M1; MYST4; KCNMA1
10	rs12267798	106,125,107	rs4918515	112,181,982	6.06	SORCS3; SORCS1; hCG2039140
11	rs1503502	19,215,941	rs611962	78,319,725	59.10	BDNF; CAT; MAPK81P1; TCN1; CNTF; CHRM1; GSTP1; GAL; FADD; 1NPPL1; UCP2; GAB2
13	rs12430069	59,745,163	rs7999738	61,243,897	1.50
13	rs1316712	84,627,554	rs2148065	94,432,314	9.80
14	rs11845833	73,772,002	rs17103130	74,844,099	1.07	NPC2; DLST; FOS
16	rs4081759	65,505,284	rs10431924	67,396,803	1.89	HSD11B2
17	rs16968682	34,640,425	rs16939964	40,779,673	6.14	GRB7; PNMT; PYY

Fig. 2.

Data from HumanHap240S beadchips. For each autosome, plots depicting genotypes for all SNPs are indicated for both siblings. The plot shows the B allele frequency across the chromosome. Genotypes for BB have an allele frequency of 1; heterozygous (AB) genotypes have an allele frequency of 0.5; and AA genotypes have an allele frequency of 0. Regions of shared homozygosity (>1 MB) between siblings are indicated in gray boxes. Homozygous deletions in chromosomes 2p22.3 (locus number 0532 in the Database of Genomic Variants) and 4q26 (locus number 9064), are indicated with an arrow. Regions with significant associations in the study by Farrer et al. (2003) are indicated with black squares underneath the specific chromosomes.

4. Discussion

As far as we know, this is the first study describing a family with a segregation pattern of EOAD suggesting a recessive mode of inheritance. In 2003, an association study performed in late onset AD patients from the Wadi Ara community, with high levels of inbreeding, reported significant associations in chromosomes 2, 9, 10, and 12 (Fig. 2) (Farrer et al., 2003). Interestingly, allelic association in chromosome 9 was due to significant overrepresentation of homozygous genotypes in patients compared to controls. However, due to the nature of the analysis (this was a case-control study in a defined population), the lack of pedigree data, and the excess of comparisons performed, caution should be taken in interpreting this data as recessive causality. Furthermore, the chromosome 9 region reported in the study by Farrer and cols. (from 9p22.2 to 9p21.2) is more than 14 MB upstream from the IBD region shared by both siblings in this same chromosome (Fig. 2). This would indicate that if recessive causality could underlay chromosome 9 association, heterogeneity of recessive causes could exist in AD. It is important to note that the other regions reported by Farrer and cols. do not locate within any of the shared homozygosity tracks presented by the two siblings analyzed in the present study.

We have used a high-throughput genotyping technology to accurately define the length of homozygosity that is IBD in both siblings with the disease. Since the HumanHap240S beadchip average spacing between SNPs (MAF ≥ 0.05) is 5.5 kb for the CEU population, genomic rearrangements equal or greater than this length could have been detected. Although we are aware that this is an approximate approach and that this is not a powerful sample to limit the loci of interest, our aim was to stress that an autosomal recessive mode of inheritance is a possible cause of dementia and more importantly to provide an initial catalog of candidate loci for us and others to follow-up on. In contrast to autosomal dominant cases, estimates of the incidence of sporadic EOAD have not been accurately described in the literature. However, patients with EOAD and no family history of disease are relatively common. We suggest that recessive segregation of disease could be playing a role in, at least, some forms of early onset familial AD, such as the present family, and could also account for a proportion of early onset “sporadic” cases. In the present study, it is unlikely that the presence of one single APOE-ε4 allele could lower the age of onset to 58 years, as is the case of the two patients analyzed.

We have narrowed the homozygosity region shared by both siblings to ~250 MB, thus discarding ~92% of the genome to contain any hypothetical recessive gene (at least in the present family). In Table 1, we outline this region and note all the loci from Alzgene (http://www.alzgene.org/, Bertram et al., 2007). Unfortunately, since other family members refused to participate, genotypic data on the cognitively healthy siblings was not possible. This information would have been of enormous value in reducing the length of shared regions and, therefore, to better define the locus of interest. A desirable next step to reduce candidate region for a recessive AD gene, would be additional high-throughput genotyping studies in small families with possible recessive patterns of inheritance, or sporadic EOAD patients. In summary, the present data contributes as a significant start for subsequent studies in which large scale genotyping and sequencing will be performed in both familial and sporadic EOAD patients not carrying mutations in any of the known Mendelian genes linked to the disease.