From 1997 to 2007: A decade journey through the H1 haplotype on 17q21 chromosome

1. The genomic structure of the H1 haplotype 

Conrad et al. [9] were the first to demonstrate that a genetic variation at MAPT could be associated with PSP disease. One specific allele, designated A0 which consists of 11TG dinucleotides and is located in the intron downstream of MAPT exon 9 was present more frequently in PSP patients than in the control group and approximately 95.5% of the PSP patients studied were found homozygous for allele A0 [9]. This overrepresentation of the allele A0 and genotype A0/A0 in PSP patients was also confirmed by other studies [10], [11], while similar results were found in CBD patients, as well [12], [13].

Once the association with the repeat polymorphism was found the question was whether the repeat itself or another variant in MAPT, in linkage disequilibrium with the repeat polymorphism, was responsible for the increased risk for PSP. After sequencing MAPT, except for the dinucleotide polymorphism, eight more SNPs were detected. All of them were in LD and defined two MAPT haplotypes covering ∼62kb of the gene, namely H1 and H2. No recombination events were found between those two haplotypes, suggesting that their establishment was an ancient event and that, since then, recombination was either suppressed or selected against. In addition, the more common H1 haplotype and the H1/H1 genotype were found to be overrepresented in PSP patients compared to controls. The sequencing analysis also revealed a deletion of 238bp in intron 9 of MAPT, always inherited as part of the less common MAPT H2 haplotype, whereas the wild-type sequence is present on the more common MAPT H1 haplotype. This insertion/deletion (ins/del) polymorphism is now routinely used for defining the two MAPT haplotypes [14].

The initially defined ∼62kb H1 haplotype, soon extended its outer edges to MAPT promoter region [15], [16], [17] and has now expanded to a region of ∼1.8Mb [18], [19], [20]. Pastor et al. [21] found a significant association between the A0/A0 genotype and Parkinson's disease (PD), thus broadening the range of MAPT possible clinical significance. Further research and the observed overrepresention of the MAPT H1 haplotype not only in PD patients, but in patients with frontotemporal dementia (FTD) and primary progressive aphasia, as well, added weight to the hypothesis that tau is a major regulator protein in the pathogenesis of many neurodegenerative diseases [22], [23], [24]. In addition, the extended ∼1.8Mb H1 haplotype encompasses not only MAPT, but the NSF (N-ethylmaleimide sensitive factor), the IMP5 (a presenilin homologue), the CRHR1 (corticotrophin releasing hormone receptor) and the WNT3 (wingless-type MMTV integration site family, member 3) genes, as well. The identification of so many genes harboured by the H1 haplotype complicated the issue of the genetic background of neurological disorders such as PSP and CBD and the existence of specific sub-haplotype blocks determining the genetic pathogenesis of various such diseases, remained to be determined.

Additional single nucleotide polymorphisms (SNPs) defining MAPT H1 sub-haplotypes were examined. In particular, with the aid of the HapMap Project [25], specific “tag SNPs” were selected, in order to define a minimum number of SNPs that are in LD and can differentiate between all MAPT H1 sub-haplotypes [20]. Rademakers et al. based the “tag SNPs” selection on the complete resequencing of the MAPT genomic region in 24 individuals [26]. Moreover, the architecture of the H1 haplotype was separated in smaller haplotype blocks, in an attempt to delineate crucial-regions, with potential pathogenic variants [18], [20], [26]. The H1 haplotype was also examined in different ethnicities. Interestingly, the LD region had the same pattern in many different populations and retained the same “margin” genes, verifying the early suggestion of the ancient formation of the H1 haplotype [27]. Comparably, the H1 haplotype displays a considerable variability, whereas the H2 haplotype is found conserved and manifests a protective role in both PSP and CBD [28]. The conservation of H2 haplotype in European populations could be attributed to “founder effect” and “positive selection” phenomena [29]. It is also of note that, according to a recent study, approximately 3 million years ago, a region of 900kb was inverted in the H2 chromosomes compared to the H1 chromosomes [29]. It has been suggested that the presence of the low-copy repeats (LCRs) in the MAPT genomic region have led to this inversion as a result of non-homologous recombination [30].

The MAPT inversion was demonstrated both on mechanically stretched chromosomes of tau-negative FTD patients, and in control individuals [31]. Whereas the presence of the MAPT inversion in FTD patients is not an argument for a genetic association of the MAPT haplotype in FTD, the fact that about half of the FTD patients have tau depositions in their brain is in favor of studying MAPT haplotypes in FTD [19], [32], [33]. Until now, the genetic association of the MAPT haplotype with FTD remains equivocal as the initial report of Verpillat et al. [23] has not been replicated by others reports [32], [33]. Moreover, no association was found between the MAPT H1 haplotype and Pick's disease [34]. On the other hand, recent studies indicate that part of the H1 haplotype, defined by the promoter polymorphism rs242557, (H1c) is a risk factor for developing AD [35], [36]. Likewise, while a large amount of data support a significant association of the H1 haplotype with PD, other studies show conflicting results [36], [37], [38], [39], [40], [41]. In fact, the role of the H1 haplotype among different ethnicities seems to differ and worldwide PD association studies highlight an ethnically dependent genetic impact of the H1 haplotype on PD pathology [42]. Moreover, different genetic variants can characterize specific ethnic groups and contribute to PD pathogenesis. For instance, a recent study in a Japanese population indicates that polymorphisms in MAPT promoter may influence the age of PD onset [43].

2. Functional effects of H1 haplotype 

Lately, a lot of attention has been directed to the functional significance of the H1 haplotype, particularly focusing on tau expression and the splicing events relevant to the MAPT exon 10. Up to now, it has been found that the H1 and H2 haplotypes have different transcriptional activity. In fact, the H1 haplotype, and particularly the H1c, has been documented to show significantly greater expression than the H2 haplotype [26], [36], [44], [45]. Evenmore, the H1 haplotype appears to express up to 43% more exon 10+ 4R MAPT transcripts than H2 [36], [45], [46]. The inclusion of exon 3 in MAPT transcripts may also contribute to protecting H2 carriers from neurodegeneration [47]. The exact pathogenic mechanism by which the H1 haplotype can cause an increased risk for some tauopathies is unknown. However, there are some specific points that should be taken into consideration. High levels of tau protein encoded by the H1 haplotype may be the basis for the risk factor associated to MAPT haplotype [48]. An imbalance in the ratio of 3R and 4R tau isoforms has been documented to affect the normal production of the tau protein and has been associated with many neurodegenerative diseases. Genetic polymorphisms can also influence the modification of tau expression and splicing. The fact that the H1 haplotype can be divided into different sub-haplotypes and the major potentials provided by the HapMap Project can pave the way of the identification of disease associated variants. Actually, as recently reported, the sub-haplotype H1c, defined by the promoter polymorphism rs242557, has been shown to be a more specific risk factor for PSP and AD [18], [35] and has been proposed to be involved in tau expression modulation [26]. The identification of additional genetic factors is also expected to provide the scientific community with more conclusive data. In addition, mechanisms that can lead to the aggregation of the encoded tau protein, such as the insufficient clearance of tau protein by the ubiquitin–proteasome degradation system might be of vital importance. Abnormally deposited tau protein could also be a toxic intermediate of oxidative damage procedures or even a product of post-translational modifications such as phosphorylation, nitration, glycation, glycosylation, ubiquitination, etc. In fact, phosphorylated tau has been found to have reduced capacity of binding to microtubules and tau hyperphosphorylation leads to the formation of pathological tau filaments [48]. Interacting phenomena between potential pathogenic genes should be also examined with concern. For instance, a-synuclein is a protein that co-localizes with tau in pathologic inclusions that characterize many neurological diseases. These proteins act synergistically and promote each others' polymerization into fibrils. The combination of risk genotypes of both a-synuclein protein and MAPT has been recently shown to approximately double the risk for development of PD [49], supporting the hypothesis that tau and a-synuclein might be implicated in a common PD pathogenic pathway. However, data obtained from statistical analysis cannot alone be efficient to interpret complicated and multifactor biological phenomena. GSK3B is a kinase that promotes tau phosphorylation. Functional polymorphisms in the GSK3B gene have also been shown to interact with MAPT H1 haplotype [50]. GSK3B–MAPT gene interaction remains to be replicated by future studies, as this synergistic effect was not observed in the study of Goris et al. [49]. At last, it should also be born in mind that different populations may be associated with different genetic variations and may show different tau expressive motifs or effects of the MAPT haplotypes on splicing events. Moreover, the role of tau and its precise defective mechanism involved in many neurodegenerative diseases may be different, as there is a lot of molecular heterogeneity in neurological groups with similar clinical symptomatology.

3. Conclusions 

Despite a decade of intense genetic research we are only beginning to decipher the H1 haplotype and its role in many neurodegenerative diseases. MAPT, a gene part of the H1 haplotype, has undoubtedly a prominent role in a number of neurological diseases such as PSP, CBD, PD, FTDP-17 and AD; however, the biological mechanisms that underline these diseases remain largely unknown. The gene complexity that characterizes the H1 haplotype, the intricacies of tau expression, splicing events and ethnic variations are some of the hindering difficulties faced. A lot of work remains to be done in order to elucidate the complex genetic structure of the H1 haplotype on chromosome 17, its functional effects and its role in a possible common pathway in many neuropathologic diseases.