Cathepsin D exon 2 polymorphism associated with ... - Semantic Scholar

Molecular Psychiatry (2003) 8, 14–18 & 2003 Nature Publishing Group All rights reserved 1359-4184/03 $25.00 www.nature.com/mp

REVIEW ARTICLE

Cathepsin D exon 2 polymorphism associated with general intelligence in a healthy older population A Payton1, F Holland2, P Diggle3, P Rabbitt4, M Horan5, Y Davidson5, L Gibbons5, J Worthington1, WER Ollier1 and N Pendleton5 1

Centre for Integrated Genomic Medical Research, Stopford Building, Manchester University, Oxford Road, Manchester M13 9PT, UK; 2GlaxoSmithKline, NFSP (S), Third Avenue, Harlow CM19 5AW, UK; 3Medical Statistics Unit, Department of Mathematics & Statistics, University of Lancaster, Lancaster LA1 4YF, UK; 4Age & Cognitive Performance Research Centre, University of Manchester, Zochonis Building, Oxford Road, Manchester M13 9PL, UK; 5Clinical Gerontology, University of Manchester, Clinical Sciences Building, Hope Hospital, Stott Lane, Salford, Greater Manchester M6 8HD, UK General intelligence is a heritable trait that is a risk factor for both the onset of dementia and the rate of cognitive decline in community-dwelling older persons. Previous studies screening for quantitative trait loci (QTLs) that influence general intelligence in healthy individuals have identified four loci, two of which are located within the genes insulin-like growth factor 2 receptor (IGF2R) and the Msx1 homeobox. Here, we report the finding of another QTL associated with general intelligence that is located within exon 2 of the cathepsin D (CTSD) gene. A group of 767 healthy adults with a follow-up period of over 15 years have been analyzed for cross-sectional and longitudinal trends in cognitive change using the Heim intelligence test score (AH4-1). We observed a significant association (P ¼ 0.01) between a functional C4T (Ala4Val) transition within exon 2 of the CTSD gene that increases the secretion of pro-CTSD from the cell, and the AH4-1 score at initial testing on entry to the longitudinal study. Interestingly, CTSD is transported by IGF2R from the trans Golgi network to the lysosome. Molecular Psychiatry (2003) 8, 14–18. doi:10.1038/sj.mp.4001239 Keywords: cathepsin D; intelligence; cognitive decline; polymorphism; association

Introduction General intelligence, also known as general cognitive ability, is a measure of an individual’s capacity to perform consistently across a variety of cognitive functions that include memory, spatial ability and processing speed. Both environmental and genetic factors influence the variation in cognitive ability observed between healthy individuals, with the contribution of each estimated at approximately 50%.1 General intelligence is an important risk factor for the incidence and prevalence of dementia and also the rate of cognitive decline in the nondemented elderly.2,3 Cognitive impairment affects approximately 10% of community-dwelling people aged 65 years and over, and accounts for 35% with disability in England and Wales.4 In an increasingly long-lived population, the identification and early treatment of individuals at risk from cognitive impairment will have profound social and economic implications.

Correspondence: A Payton, Centre for Integrated Genomic Medical Research, Stopford Building, Manchester University, Oxford Road, Manchester M13 9PT, UK. E-mail: [email protected] Received 26 March 2002; revised 21 May 2002; accepted 5 June 2002

A number of quantitative trait loci (QTLs) that are associated with cognitive deficit have already been reported. In particular, the presence of the APOE e4 allele confers a much greater risk of developing Alzheimer’s disease (AD), a condition that is characterized by irreversible and progressive memory dysfunction. We have recently reported that APOE genotype does not influence cognitive ability or its decline with age in healthy older individuals.5 The insulin-like growth factor 2 receptor (IGF2R) gene was the first reported QTL associated with general intelligence in healthy individuals. The gene codes for a mannose-6-phosphate receptor whose function includes the transport of phosphorylated lysosomal enzymes from the trans Golgi network to lysosomes. A microsatellite at the 30 end of the IGF2R gene contains a variable number of repeats that have been shown to be associated with children’s IQ.6 Other loci that have previously been associated with intelligence include a microsatellite polymorphism within the MSX1 gene located on chromosome 4 and two other microsatellite markers (D4S2943 and D4S1607) that are also located on chromosome 4.7 Cathepsin D (CTSD) is an active acid protease that is initially produced as a nonfunctional enzyme in the trans Golgi network. A post-translational modification that involves the addition of MP6 residues

Cathepsin D exon 2 polymorphism A Payton et al

enables pro-CTSD to bind to IGF2R and allows its transport to the lysosome where it is involved in intracellular protein breakdown. A functional polymorphism (C4T, Ala4Val) within exon 2 of the CTSD gene increases the secretion of pro-CTSD from the cell8 and has also been reported as a risk factor in AD, although this finding has been challenged.9–15 A role for CTSD in apoptosis has also been established,16,17 and its importance in brain development has recently been highlighted owing to its involvement in congenital ovine neuronal lipofuscinosis.18

Methods Group studied The individuals taking part in the research were all Caucasian volunteers from the University of Manchester Age and Cognitive Performance Research Centre, Manchester, UK. This longitudinal research program began in 1985 and comprises volunteers who can attend the research center with the only entry restriction being that they should be older than 49 years of age. Originally, this volunteer group comprised approximately 2500 individuals. The 767 individuals in this study are those who could attend for venesection in 1998. A large database exists of scores from these group on cognitive test batteries performed over the last 17 years. In addition to these data, information on age, gender, self-reported health, educational and socioeconomic attainment has been collected. Details of the recruitment, composition and selective attrition of the entire longitudinal study of elderly community residents have been described elsewhere.19 The group has been analyzed for crosssectional and longitudinal trends in cognitive change using the Heim intelligence test score (AH4-1).20 The volunteers performed the AH4-1 test in 1985, 1991 and 1997. The Heim AH4 test is a group general intelligence measure for use with a cross-section of the adult population. AH4-1 uses 65 pen-and-paper questions to access verbal and numerical skills. The test is performed within a 10-min time period and therefore has no ceiling effect. Results are expressed as number of correct responses. To determine the socioeconomic attainment of volunteers, we used the UK official social class classification system.21 Occupations were divided as follows: I, Professional; II, Managerial and Technical; III, Skilled ((N) nonmanual and (M) manual); IV, Partly skilled; V, Unskilled. University of Manchester ethics committee approval and individual written consent to the research was obtained. CTSD amplification The CTSD exon 2 C4T transition was amplified using primers (forward: 50 GTG ACA GGC AGG AGT TTG GT 30 ; reverse: 50 GGG CTA AGA CCT CAT ACT CAC G 30 ). PCRs were carried out in 96-well microtiter plates with a final reaction volume of 20 ml containing NH4 buffer (Bioline), 1.5 mM MgCl2, 0.1 mM dNTPs, 0.2 units Taq polymerase (Bioline, BioTaq), 10 pmol of

each primer, 50 ng of genomic DNA and 0.5 M Betaine. Cycle conditions were 35 cycles at 951C/ 40 s, 601C/30 s, 721C/30 s followed by 5 m at 721C using a PTC-225 Peltier Thermal Cycler (MJ Research). The 343 bp product was digested overnight with MwoI (New England Biolabs) and the fragments separated on a 2% agarose gel. The C allele digested into two distinctive 168 and 82 bp bands and the T allele formed a distinctive 250 bp band. To ensure genotyping accuracy and reproducibility, the first 95 samples were genotyped on two separate occasions, and on every plate of samples three of known genotype were included.

15

Statistical analysis A longitudinal analysis of the entire Newcastle and Manchester cohorts has been described in detail previously.19 In brief, random effects models were used to investigate factors predictive of decline in AH4-1 accounting for the fact that measurements taken on the same subject over time tend to be correlated. This type of model can be considered in two parts: a normal linear regression model for the average response over time for a subject with given values of all explanatory variables, and a model for the random variation about the mean response. For the second component, a set of latent variables or ‘random effects’ are postulated, which represent deviations of individual subjects from the population average for selected features. In the original analysis, the following explanatory variables were considered: age, sex, socioeconomic status, city of residence, year volunteer first participated in study and practice effect. The practice effect was modeled as a series of step increases between successive occasions on which the subject was tested; age was modeled as a quadratic trend, coding age in years over 49, the minimum age at entry across all subjects. The fitted population-average response was found to be quadratic in age with no evidence of interaction depending on social class, sex or practice. In the final model, the intercept of the quadratic depended additively on the subject’s sex, social class, year of entry to the study and city of residence with population-average step increases on second and subsequent testing. The random variation about the mean response was modeled by subject-specific intercept and slope parameters, which varied randomly from the average response curves according to a zero-mean bivariate normal distribution. The above model, excluding the term for city, was fitted to the subset of subjects who participated in the current study. The improvement in model fit on adding the C4T mutation/transition as a covariate, and its interaction with the quadratic age trend was assessed by the likelihood ratio test.

Results Blood samples were available for 776 patients. A total of nine patients lacking AH4-1 scores were excluded Molecular Psychiatry


16

Table 1 AH4-1 mean (SD) percent correct scores of volunteers when first tested at entry to study by gender, socioeconomic status, age of volunteer and presence or absence of C>T polymorphism Number

Mean

SD

Gender Female Male

536 231

56 59.3

16.0 16.1

Socioeconomic status C1 C2 C3(N) C3(M) C4 or C5 Missing

70 320 240 72 38 27

66.4 62.1 53.9 45.0 42.5 51.1

11.9 15.8 13.7 13.0 15.0 15.8

Age (years) 49–59 60–69 70–79 80 plus

232 424 104 7

63.1 55.2 50.8 49.0

15.1 15.9 14.4 20.8

CTSD No mutation Mutation

632 119

57.0 52.9

16.8 16.8

from the analyses, as were 16 individuals on whom the CTSD genotype was not available. Of the remaining 751 subjects, a total of 746 were still participating at the second visit and 499 remained at the third visit. Summary statistics for the group are shown in Table 1. The frequencies of the genotypes for CTSD were as follows: 632 (84.2%) were homozygous for the wildtype gene, 117 (15.6%) were heterozygous and 2 (0.3%) were homozygous for the C4T substitution at position 224. Therefore, 119 (15.8%) possessed at least one copy of the C4T substitution in the CTSD gene. Genotypes were in Hardy–Weinberg equilibrium. The results from fitting the selected random effects models for C4T substitution in the CTSD gene are shown in Table 2. The estimated mean AH4-1 score at entry onto the study was 3.3 percentage points lower for those subjects with the C4T substitution in the CTSD gene compared to those without after adjusting for several factors. This difference was statistically significant (P ¼ 0.01). Our sample size gives 80% power (5% significance level) to detect the observed effect size. There was no evidence that the rate of AH4-1 decline differed according to the presence or absence of the C4T substitution in the CTSD gene (P ¼ 0.78).

Discussion This is the first study to look for an association between CTSD and intelligence and is the first to describe a functional polymorphism being associated with cognitive ability. Our results showed an associaMolecular Psychiatry

tion between the CTSD C4T polymorphism and intelligence score at entry to the longitudinal study but no association between the polymorphism and the rate of cognitive decline over time. This would support the hypothesis that CTSD affects intelligence at an earlier stage, possibly during early brain development. Apoptosis is responsible for the death of between 50 and 80% of neurons during early brain development and is essential for the sculpting of efficient pathways. An increasing amount of evidence suggests that CTSD is an integral part of this process. Studies using rat neuronal cells have demonstrated that induction of apoptosis via neurotrophic factor deprivation increases CTSD levels and decreases levels of CTSB.22 Additionally, inhibiting CTSD reduces apoptosis in different cell types including neurons.23 The mechanism of CTSD action in apoptosis has been shown to occur before the activation of caspase-3-like proteins.17 Failure to express IGF2R or the presence of the T allele within the CTSD gene both result in the increased secretion of immature enzyme from the cell.8,24 This may decrease the amount of mature enzyme available for apoptosis during development and result in the survival of functionally inefficient neurons and hence give rise to a lower general intelligence. If the CTSD T allele is associated with general intelligence, then it would also be expected to be associated with dementia onset for which intelligence is an important risk factor.2,3 General intelligence as a risk factor for dementia onset has been explained using a concept called ‘brain reserve capacity’ (BRC).25 This concept implicates the presence of a threshold that needs to be exceeded before disease symptoms become evident. Individuals with a greater general intelligence will have a higher threshold and can therefore receive a larger brain insult before showing symptoms of cognitive impairment. A large well-characterized family-based study using the sibship disequilibrium test15 found no association between the CTSD T allele and AD, and concluded that there is no association between the C4T polymorphism and individuals with AD. This study did not look at the C4T polymorphism in relation to disease onset. The case–control studies looking at the association between the CTSD C4T polymorphism and AD have produced inconclusive results with two positive and four negative associations.9–14 However, three case–control studies that reported a negative association in Caucasians did observe a nonsignificant increase in T allele frequency in AD cases compared to controls.11–13 The fourth negative case–control study showed a nonsignificant increase in younger AD patients carrying the T allele.14 As intelligence is a risk factor for the onset of dementia we propose that the above age-matched case–control studies were detecting an association between intelligence and disease onset. The finding that Hispanics carrying a T allele had an earlier disease onset that was statistically significant supports this.11


17

Table 2 Random effect model parameter estimates for the mean AH4-1 percent correctresponse adjusting for CTSD Estimate

SE

t-value

P-value

95%

CI

(a) Mean AH4-1 percent correct response model Intercept 72.96 Age49 0.04 0.02 (Age49)2 Male vs female 0.41

2.81 0.13 0.003 1.14

25.99 0.31 8.28 0.36

o0.01 0.76 o0.01 0.72

67.47 0.21 0.03 1.82

78.44 0.28 0.02 2.65

Socioeconomic status C2 vs C1 C3(N) C3(M) C4/5 Missing

3.19 10.47 18.61 21.61 13.11

1.84 1.95 2.36 2.85 3.14

1.74 5.36 7.88 7.58 4.18

0.08 o0.01 o0.01 o0.01 o0.01

6.78 14.29 23.23 27.18 19.25

0.40 6.65 13.99 16.03 6.97

4.86 2.36

0.50 0.57

9.64 4.16

o0.01 o0.01

3.87 1.25

5.84 3.47

Entry year 1986 vs 1985 1987 vs 1985 1989 vs 1985 1991 vs 1985

2.22 1.16 1.13 1.34

1.40 1.42 1.44 2.17

1.58 0.82 0.78 0.62

0.11 0.41 0.44 0.54

0.53 3.93 1.70 5.58

4.96 1.61 3.95 2.90

CTSD Mutation vs none

3.28

1.34

2.44

0.01

5.91

0.65

12.54 0.24 0.59 5.15

15.59 0.48 0.21 5.73

Practice effect Visit Z2 Visit Z3

(b) Covariance and residual estimates SD (random intercept) 13.98 SD (random slope) 0.34 Corr (intercept,slope) 0.41 SD measurement error 5.43

The topic of genetics and intelligence has historically been a controversial one. The argument between nature vs nurture has been resolved by extensive studies involving families and twins, and has led the experts to conclude that each contributes approximately 50% to the variation observed between the cognitive ability of individuals. Concerns as to whether genetic screening for intelligence could be used to subgroup individuals at an early age must be weighed against the benefits of identifying and treating individuals who are at risk from developing cognitive deficit later in life and indeed any other disorder that the same genes may be associated with. For example, both the CTSD and IGF2R genes have also been implicated in cancer,26,27 emphasizing that while many diseases have multiple susceptibility genes it is also important to consider that many genes may be involved in multiple diseases. The relative small effects contributed by both CTSD and IGF2R also suggest that general intelligence will be a complex polygenic trait. The 3.3% effect size of CTSD C4T polymorphism toward the total variance of the AH4-1 score can be given perspective by comparing it to the effect size caused by ‘normal aging’, which is the biggest risk factor for cognitive decline in healthy individuals. Based on our statistical model with a negative quadratic age coefficient,

cognitive decline is predicted to accelerate with age. For example, using our reference group (female sex, first participated in study 1985 and social class C1), the average score at entry is estimated to be 73% points for a 49-year old falling to 53 points for an 80year old, that is, a difference of 20% points in mean scores between these two age groups. The previous finding that IGF2R6 is associated with intelligence and is also responsible for the transport of CTSD helps validate our result and suggests that the successful transport of CTSD is an important step for the optimal development of the human brain. Screening of both genes for new exonic and regulatory polymorphisms will be our next priority.

Acknowledgements This work was supported by the Wellcome Trust. Blood collection and DNA extraction was funded by Research into Ageing.

References 1 Bouchard Jr TJ, McGue M. Familial studies of intelligence: a review. Science 1981; 212: 1055–1059. Molecular Psychiatry


18 2 Schmand B, Smit JH, Geerlings MI, Lindeboom J. The effects of intelligence and education on the development of dementia. A test of the brain reserve hypothesis. Psychol Med 1997; 27: 1337–1344. 3 Evans DA, Beckett LA, Albert MS, Hebert LE, Scherr PA, Funkenstein HH et al. Level of education and change in cognitive function in a community population of older persons. Ann Epidemiol 1993; 3: 71–77. 4 MRC CFAS and RIS MRC CFAS. Profile of disability in elderly people: estimates from a longitudinal population study. B M J 1999; 318: 1108–1111. 5 Pendleton N, Payton A, van den Boogerd EH, Holland F, Diggle P, Rabbitt PM et al. Apolipoprotein E genotype does not predict decline in intelligence in healthy older adults. Neurosci Lett 2002; 324: 74–76. 6 Chorney MJ, Chorney K, Seese N, Owen MJ, Daniels J, McGuffin P et al. A quantitative trait locus associated with cognitive ability in children. Psychol Sci 1998; 9: 159–166. 7 Fisher PJ, Turic D, Williams NM, McGuffin P, Asherson P, Ball D et al. DNA pooling identifies QTLs on chromosome 4 for general cognitive ability in children. Hum Mol Genet 1999; 8: 915–922. 8 Touitou I, Capony F, Brouillet JP, Rochefort H. Missense polymorphism (C/T224) in the human cathepsin D pro-fragment determined by polymerase chain reactionFsingle strand conformational polymorphism analysis and possible consequences in cancer cells. Eur J Cancer 1994; 30A: 390–394. 9 Papassotiropoulos A, Bagli M, Feder O, Jessen F, Maier W, Rao ML et al. Genetic polymorphism of cathepsin D is strongly associated with the risk for developing sporadic Alzheimer’s disease. Neurosci Lett 1999; 262: 171–174. 10 Papassotiropoulos A, Bagli M, Kurz A, Kornhuber J, Forstl H, Maier W et al. A genetic variation of cathepsin D is a major risk factor for Alzheimer’s disease. Ann Neurol 2000; 47: 399–403. 11 Crawford FC, Freeman MJ, Schinka J, Abdullah LI, Richards D, Sevush S et al. The genetic association between cathepsin D and Alzheimers disease. Neurosci Lett 2000; 289: 61–65. 12 McIlroy SP, Dynan KB, McGleenon BM, Lawson JT, Passmore AP. Cathepsin D gene exon 2 polymorphism and sporadic Alzheimer’s disease. Neuro Sci Lett 1999; 273: 140–141. 13 Bhojak TJ, DeKosky ST, Ganguli M, Kamboh M. Genetic polymorphisms in the cathepsin D and interleukin-6 genes and the risk of Alzheimer’s disease. Neurosci Lett 2000; 288: 21–24. 14 Mateo I, Sanchez-Guerra M, Combarros O, Llorca J, Infante J, Gonzalez-Garcia J et al. Lack of association between cathepsin D and Alzheimer disease in a Spanish sample. Am J Med Gen 2002; 114: 31–33.

Molecular Psychiatry

15 Bertram L, Blacker D, Crystal A, Mullin K, Keeney D, Jones J et al. Candidate genes showing no evidence for association or linkage with Alzheimer’s disease using family-based methodologies. Exp Geron 2000; 35: 1353–1361. 16 Roberts LR, Adjei PN, Gores GJ. Cathepsins as effector proteases in hepatocytes apoptosis. Cell Biochem Biophys 1999; 30: 71–88. 17 Ollinger K. Inhibition of cathepsin D prevents free-radicalinduced apoptosis in rat cardiomyocytes. Arch Biochem Biophys 2000; 373: 346–351. 18 Tyynela J, Sohar I, Sleat S, Rosalie MG, Donnelly RJ, Baumann M et al. A mutation in the ovine cathepsin D gene causes a congenital lysosomal storage disease with profound neurodegeneration. EMBOJ 2000; 19: 2786–2792. 19 Rabbit P, Donlan C, Bent N, McInnes L, Abson V. The University of Manchester Age and Cognitive Performance Research Centre and North East Age Research Longitudinal Programmes, 1982 to 1997. Z Gerontol 1993; 26: 176–183. 20 Heim AW. The AH 4 Group of Intelligence Tests. NFER-Nelson: Windsor, 1971. 21 Standard Occupational Classification. Vol. 3, HMSO: London, 1991. 22 Shibata M, Kanamori S, Ishara K, Ohsawa Y, Konishi A, Kametaka S et al. Participation of cathepsins B and D in apoptosis of PC12 cells following serum deprivation. Biochem Biophys Res Commun 1998; 251: 199–203. 23 Ishara K, Ohsawa Y, Kanamori S, Shibata M, Waguri S, Sato N et al. Regulation of a novel pathway for cell death by lysosomal aspartic and cysteine proteinases. Neuroscience 1999; 91: 233– 249. 24 Lorenzo K, Ton P, Clark JL, Coulibaly S, Mach L. Invasive properties of murine squamous carcinoma cells: secretion of matrix-degrading cathepsins is attributable to a deficiency in the mannose 6-phosphate/insulin-like growth factor II receptor. Cancer Res 2000; 60: 4070–4076. 25 Satz P. Brain reserve capacity on symptom onset after brain injury: a formulation and review of evidence for threshold theory. Neuropsychology 1993; 3: 817–824. 26 Rochefort H, Garcia M, Glondu M, Laurent V, Liaudet E, Rey JM, et al. Cathepsin D in breast cancer: mechanisms and clinical applications, a 1999 overview. Clin Chim Acta 2000; 291: 157–170. 27 DaCosta SA, Schumaker LM, Ellis MJ. Mannose 6-phosphate/ insulin-like growth factor 2 receptor, a bona fide tumor suppressor gene or just a promising candidate? J Mammary Gland Biol Neoplasia 2000; 5: 85–94.