MAPT H1 Haplotype is Associated with Late-Onset Alzheimer’s Disease Risk in APOE ɛ 4 Noncarriers: Results from the Dementia Genetics Spanish Consortium

INTRODUCTION

Tau protein plays an essential role in the central nervous system by promoting microtubule assembly and stability in neuronal cells. Neurofibrillary tangles composed of truncated and hyperphosphorylated tau proteins are one of the hallmarks of Alzheimer’s disease (AD) pathology [1]. Neurofibrillary tangles are also present in a substantial subgroup of frontotemporal dementia patients (FTD), and in other FTD-spectrum tauopathies, such as progressive supranuclear palsy (PSP) and corticobasal degeneration (CBD). Tau deposits also colocalize with alpha-synuclein in Lewy bodies of Parkinson’s disease (PD) patients [1–4].

Tau protein is encoded by the MAPT gene (MAPT: OMIM: *157140), located at chromosome 17q21-22. There are two common MAPT extended haplotypes in Caucasians resulting from an ancestral inversion: H1 and H2. The H1 haplotype has been linked with sporadic and familial neurodegenerative disorders like PSP [5–8], CBD [9], FTD [10], PD [11–13], and inconsistently with AD [14]. In fact, the last AlzGene meta-analysis including case-control data showed no significant association between MAPT H1 haplotype and AD [15] and, so far, available genome-wide association study found no MAPT risk variants in AD subjects [16], and only very recently the IGAP consortium has found a significant association with AD near MAPT in subjects not carrying APOE ɛ4 [17].

Mutations in MAPT have been identified in familial FTD syndromes [18–24]; however, the role of rare genetic MAPT variants in sporadic neurodegenerative diseases is not well established. More recently, a rare variation in MAPT exon 7 (p.A152T) has been linked to both sporadic FTD and AD risk [25–27]; however, to date, p.A152T association has not been replicated in large independent populations.

In the present study, we assessed the risk effect of the rare variant MAPT p.A152T and the common extended MAPT H1/H2 haplotypes in a large series of participants with sporadic and genetic neurodegenerative disorders from Spain.

MATERIALS AND METHODS

RESULTS

No deviation from Hardy Weinberg equilibrium (Pearson’s Chi-Square) was found in controls for both studied variants (p = 0.78 for MAPT p.A152T and p = 0.86 for MAPT H1/H2).

DISCUSSION

In our first analysis, we tested whether the MAPT p.A152T rare genetic variant was associated with risk for various neurodegenerative diseases (AD, FTD, PSP, and PD). We found that MAPT p.A152T occurs more frequently in Spanish patients with neurodegenerative disease compared with the study by Coppola et al. [25], whose cohort was primarily comprised of the US population (AD:0.97% versus 0.69% ; FTD: 1.42% versus 0.89% and PD: 077% versus 0.48% respectively). Because the frequency of MAPT p.A152T was also significantly higher in our healthy controls than the healthy control cohort of Coppola et al. (0.71% versus 0.30% , respectively) [25], the association between AD risk and MAPT p.A152T was not significant in our population. Although our OR for AD risk associating with p.A152T occurred in same direction as in the previous study [25], our OR was considerably lower and thus did not reach statistical significance (OR = 1.4; 95% CI = 0.9–2.1 versus OR = 2.3; 95% CI = 1.3–4.2, respectively). Similarly, the OR we obtained for p.A152T in FTD risk trended toward significance, but was also lower than the OR for FTD risk in the previous study (OR = 2.0; 95% CI = 0.9–4.3 versus OR = 3.0; 95% CI = 1.6–5.6, respectively) [25].

Several factors may explain the lack of replication of previous results. Rare genetic variant frequencies can differ across populations, and MAPT p.A152T appears to occur more frequently in the general Spanish population than in the US. Another consideration is that the real ORs for diseases associated with the variant may be lower than the ORs in the discovery cohort due to the “winner’s course” effect, a common phenomenon observed in pioneer genetic epidemiological studies [35]. Another potential influence on the differencebetween our results and those of Coppola et al. [25] is the mean age at which controls were deemed healthy; for example, p.A152T carriers in one cohort may have been classified as controls at a younger age, prior to disease onset. Since the controls of Coppola et al. [25] were significantly younger (50 ± 16 years) than those analyzed in our study (64.1 ± 14.8 years), it is less likely that misclassification of our p.A152T carriers as controls who might manifest future degenerative disease could explain the higher p.A152T allelic frequency observed in our cohort.

A surprising finding of the present study was the co-segregation of MAPT p.A152T in 70% of carriers of the GRN mutation IVS6-1G>A (g.1872G>A) unique to the Basque Country [32]. This is a splicing mutation located at chromosome 17 (base pair position 139486) that causes truncated GRN protein due to mRNA degradation [33]. The fact that the MAPT p.A152T variant co-occurred with the GRN mutation only in families in a limited geographical region suggests that these individuals share the same haplotype, most likely from a common ancestor. However, MAPT p.A152T variant in patients carrying the GRN IVS6-1G>A mutation did not influence age at onset. Future studies are necessary to probe the influence of the co-ocurrence of MAPT p.A152T and GRN IVS6-1G>A on the FTD clinical or neuropathological phenotype.

Our last finding was that MAPT H1 haplotype is overrepresented in patients with AD, PD, and PSP compared to controls. Although the association of MAPT with PD and PSP risk is well documented [5–8, 11–13], its association with AD is much more controversial. To date, genome-wide association studies and case-control data meta-analyses such as AlzGene have not been able to link MAPT genetic variants to AD [15] despite numerous experimental evidence of the involvement of tau protein in AD pathogenesis [36]. In our study, we found a very significant overrepresentation of the MAPT H1 haplotype in patients with AD compared to controls. The mildly increased risk for AD conferred by the H1 haplotype emerged only in our subgroup of noncarriers for APOE ɛ4, especially in the oldest subjects. This is in line with a recent publication re-analyzing the IGAP consortium data in APOE ɛ4 carriers and non-carriers. That study reported genome-wide significant association with many SNPs across a region on chromosome 17 including MAPT and with the H1 haplotype, however, the association was accounted for by SNPs located between two genes (KANSL1 and LRRC37A) adjacent to MAPT [17].

Our results are consistent with the hypothesis that AD pathology could develop through different causal pathways with several genetic factors likely to be involved, APOE ɛ4 being the strongest one. APOE ɛ4 lowers the threshold for AD susceptibility, which 1) associates with an earlier age of disease onset, and 2) may decrease the number and magnitude of etiological factors that would be necessary to start the disease’s pathological mechanisms. However, in the absence of APOE ɛ4, the participation of an ensemble of alternative etiological factors, and for longer periods of time, might be necessary to elicit the disease. For instance, if MAPT H1 haplotype confers a modest risk for AD independent of APOE ɛ4, we may be able to detect this association only in elderly individuals not carrying APOE ɛ4; otherwise, APOE ɛ4’s effect on AD risk might mask the ability to detect the effect of MAPT H1 on AD risk. The association between the MAPT haplotype and AD is consistent with studies suggesting that H1/H1 status is associated with an increased rate of conversion from mild cognitive impairment to AD [37]. Our results are also in line with a recent publication studying MAPT haplotypes in a large sample of late onset AD from the US that found that H2 haplotype carriers were protected from AD and had lower MAPT levels in brain [38]. An alternative explanation to our results could be that within the APOE ɛ4 non-carriers group the number of subjects withdementia due to pure tauopathies (PSP, CBD, or FTD-tau) misdiagnosed as AD might be higher than those among the APOE ɛ4 carriers group. We suggest that in large population samples this phenomena is likely to occur to a certain degree, but we consider unlikely that these disorders with a low prevalence are contributing significantly to our results. Additionally, the fact by which most patients included in our study come from specialized memory units from academic hospitals increases the likelihood of a correct AD diagnosis.

In summary, we did not find a significant association between the rare variant MAPT p.A152T and AD risk, although our findings trended toward significance for p.A152T being associated with FTD risk. Despite our large sample size, our results should be interpreted with caution, as our study may be underpowered to detect the effect of such an infrequent genetic variant if the real OR is lower in our population than found in previous studies. Our finding that MAPT p.A152T and the progranulin IVS6-1G>A mutation cosegregates in families from the Basque region raises interesting questions about the influence of multiple risk genetic variants coinciding in neurodegenerative diseases; future studies will address these questions [39]. Finally, we found a robust statistical association between MAPT H1 extended haplotype and risk of late-onset AD in APOE ɛ4 noncarriers. Our results, in a large sample of Spanish population, represent strong evidence supporting a link between common MAPT genetic variants and AD. The modest risk effect conferred by MAPT H1 haplotype and the fact that it is restricted to APOE ɛ4 negative subjects might contribute to clarify controversial results inprevious studies.

DEGESCO project members

Center for Applied Medical Research, University of Navarra (CIMA): Sara Ortega-Cubero, Oswaldo Lorenzo-Betancor, Elena Lorenzo, Maria A. Pastor, Pau Pastor

Research Center and Memory Clinic of Fundació ACE: Isabel Hernández, Asunción Lafuente, Maitée Rosende-Roca, Ana Mauleón, Liliana Vargas, Octavio Rodríguez, Carla Abdelnour, Montserrat Alegret, Ana Espinosa, Gemma Ortega, Ángela Sanabria, Marta Ibarria, Susana Diego, Pilar Canyabate, Mariola Moreno, Mar Buendía, Ana Pancho, María Eugenia Palacio, Susana Ruiz, Natalia Tantinya, Marina Tarragona, América Morera, Marina Guitart, Oscar Sotolongo Grau, Elvira Martín, Victòria Fernández, Agustín Ruiz, Lluís Tárraga, Mercè Boada

Hospital Clí nic: Raquel Sánchez-Valle, José Luis Molinuevo, Albert Lladó, Anna Antonell, Lorena Rami

Gen ética Molecular-HUCA: Eliecer Coto, Victoria Alvarez

Biodonostia: Fermín Moreno, Begoña Indakoetxea, Myriam Barandiarán, Ana Gorostidi, Adolfo López de Munain

FCIEN: Alberto Rábano, Miguel Calero

HUMV: José Luis Vázquez-Higuera, Ana Pozueta, Andrea González Suarez, Eloy Rodriguez-Rodriguez, Ignacio Mateo, José Berciano, Pascual Sánchez-Juan, Onofre Combarros

Centro de Biología Molecular Severo Ochoa: María J. Bullido, Ana Frank-García, Isabel Sastre

Doce de Octubre: Eva Carro

BIOMICs: Marian M de Pancorbo, Xabier Elcoroaristizabal

Hospital de la Santa Creu i Sant Pau-Neurology Department-Memory Unit: Alberto Lleó, Rafael Blesa, Juan Fortea, Olivia Belbin, Daniel Alcolea, María Carmona-Iragui, Mª Belén Sánchez-Saudinós, Isabel Sala, Sofía Anton-Aguirre, Estrella Morenas, Roser Ribosa, Martí Colom-Cadena, Laura Cervera, Laia Muñoz, Oriol Dols-Icardo, Jordi Clarimón