Significant Changes in the Tau A0 and A3 Alleles in Progressive Supranuclear Palsy and Improved Genotyping by Silver Detection

Background  Progressive supranuclear palsy (PSP) is characterized by intraneuronal inclusions of neurofibrillary tangles formed by aggregated tau protein. A significant association between the tau gene A0/A0 genotype and PSP recently has been reported.

Objectives  To determine if a significant association between the tau gene A0/A0 genotype and PSP could be found in an independent population with a genetic background different from that in which the initial association was reported, and to standardize a nonradioactive method for tau gene genotyping.

Setting  Hospital and university research laboratories.

Subjects and Methods  To facilitate genotyping of the tau gene, we standardized the conditions for silver-based detection of the tau gene dinucleotide polymorphism. Thirty patients from Spain clinically diagnosed as having probable PSP were included in the study and compared with different control groups.

Results  A highly significant overrepresentation of the A0/A0 genotype (P<.001) and a decrease in the frequency of the A0/A3 genotype were found in the Spanish patients with PSP compared with the control group. A method based on silver detection was standardized for the genotyping of the tau gene.

Conclusions  The detection of a significant association between the tau gene A0/A0 genotype and PSP in 2 independent populations rules out genetic stratification as an explanation for the association and indicates that the presence of the tau A0/A0 genotype is a risk factor for developing PSP independent of genetic background. Alternatively, the results could be interpreted as a protective effect of the A3 allele.

PROGRESSIVE supranuclear palsy (PSP) is a neurological degenerative disorder clinically characterized by supranuclear ophthalmoplegia, parkinsonism, prominent early postural instability and frequent falls, along with secondary symptoms including neck dystonia and frontal lobe–type dementia.^1-3 Progressive supranuclear palsy is considered to be a tau pathologic disorder. Histopathologically, PSP is characterized by the presence of subcortical neurofibrillary tangles (NFTs).^4-7 Other neurological disorders with tau pathologic characteristics include Alzheimer disease (AD),^4,8-10 parkinsonism-dementia complex of Guam, and Pick disease.¹⁰ However, a differential characteristic of PSP is the absence of amyloid deposits and neuritic plaques.^4-7

Evidence for a possible genetic origin or predisposition for PSP has been presented.^11,12 Based on tau pathologic characteristics and the knowledge that the ϵ4 allele of the apolipoprotein E (APOE) gene is a risk factor for AD, several studies involving patients with PSP with the APOE genotype have searched for a correlation between PSP and the ϵ4 allele of the APOE gene, although no association has been found.^13,14 Notably, it has been recently reported that the A0 allele of a polymorphic marker (dinucleotide repeat) of the tau gene is significantly overrepresented in PSP.¹⁵

We initiated the present study to determine if a similar association could be detected in an independent population with a different genetic background. Additionally, we aimed to standardize the method for tau gene genotyping based on silver detection, thus avoiding the use of radioactive isotopes in the experimental procedure.

We included 30 patients in the study with the clinical diagnosis of PSP. Following the diagnostic criteria of the National Institute of Neurological Disorders and Stroke/Society for PSP Inc,¹⁶ all 30 patients were diagnosed as having probable PSP, while following the diagnostic criteria of our Movement Disorder Clinic,¹⁷ 29 participants were diagnosed as having probable PSP and 1 participant was diagnosed as having possible PSP. The mean±SD age of the patients with PSP was 70.7±7.5 years, and the mean±SD age at onset of the disease was 65±7.5 years. Several control groups were included in the study. The young subjects control group consisted of 108 random subjects (mean age, 24 years) who were attending the genetic service because of pathologic conditions unrelated to PSP or AD. Two other control groups in the study included 44 patients with AD (mean age, 71.2 years) and their corresponding spouses (control group 2; mean age, 71.5 years). The selection of these subjects is described elsewhere.^8,9 Patients and controls came from 3 main hospitals in Spain: Hospital Clínic Provincial, Barcelona, Hospital La Fe, Valencia, and Hospital de Bellvitge, Hospitalet de Llobregat.

DNA was isolated from blood samples and the tau dinucleotide repeat was amplified (forward primer, 5′-GCCTCGCAAATTGCTGGGAT-3′; reverse primer, 5′-AGGTGACTGGGTAGAGACAGAGC-3′) with the conditions described using [α-phosphorus P 32]dCTP.¹⁵ In addition, parallel duplicate polymerase chain reactions were run, but no [α-phosphorus P 32]dCTP was included in those reactions. Radioactive samples were run exactly as described¹⁵ in 6% polyacrylamide gels containing 8-mol/L urea, dried, exposed overnight to autoradiographic film, and developed to read the genotypes. Nonradioactive samples were run on 40-cm-long 6% polyacrylamide gels containing 8-mol/L urea as described, except that thick-wedged (1 mm at the bottom) spacers were used to increase the strength of the gels in the area of interest. The samples (50 µL of the polymerase chain reactions) were supplemented with 10 µL of loading buffer (0.05% xylene cyanol, 0.05% bromphenol blue, 20-mmol/L EDTA, and 98% formamide) and denatured, and the volume was reduced by heating (30 minutes at 98°C) and run at 1500 V for 3 hours until the bromphenol blue dye was 10 cm from the bottom of the gel. The bottom 15 cm of the gel was cut with a scalpel and silver stained as described¹⁸ with the following modifications: the gel was fixed for 10 minutes in 10% ethanol, incubated in 1% HNO₃ for 3 minutes, washed once in water, incubated in 0.2% NO₃Ag for 20 minutes, washed twice in water, and developed in 2.96% CO₃Na in 0.02% formaldehyde (2 washes in the developing solution; the first wash was discarded as soon as a black precipitate appeared in the wash solution). Once the microsatellite bands became visible (after approximately 10 minutes) the developing process was stopped in a 10% acetic acid. Subsequently, the gels were soaked in 10% glycerol for 5 minutes, transferred to a piece of Whatmann paper (Whatmann International LTD, Maidstone, England), covered with plastic wrap, and dried under a vacuum. After the genotypes were read, the results were analyzed with the SPSS 6.1 statistical package (SPSS Inc, Chicago, Ill).

Comparisons of the genotypes determined by the radioactive procedure with the genotypes determined with nonradioactive silver detection coincided in all cases. The nonradioactive silver detection method yielded slightly sharper bands and allowed for the evaluation of the genotype immediately (30 minutes) after the gel was run. However, the major advantage of using silver detection was that it obviated the need for special radiation protection requirements and licenses for the laboratory and personnel, thus making the procedure more widely accessible.

After analyzing the tau genotypes, we found that the frequency of A0/A0 homozygous individuals was 38% in the young subjects control group, 52% in the spouses of patients with AD control group, and 54.5% in the group of patients with AD (Table 1). No significant differences were found among these 3 groups. In contrast, a highly significant increase in the frequency of A0/A0 individuals (86.7%; P<.001) was found in the PSP group compared with the young subjects control group or the spouses of patients with AD control group (Table 1). The frequency of the A0/A3 genotype was also significantly lower in the PSP group compared with the young subjects control group (P =.001) and with the spouses of patients with AD control group (P =.02). Consistently, the A3 allelic frequency was also lower in the PSP group compared with the young subjects control group (P<.001) and with the spouses of patients with AD control group (P =.01).

We found a highly significant overrepresentation of the homozygous A0/A0 tau gene dinucleotide polymorphism in patients with PSP compared with control groups. The finding of this association between the A0 homozygous state of the tau gene dinucleotide polymorphism and PSP in a population independent of that described by Conrad et al¹⁵ makes it unlikely that the association is due to genetic stratification of this particular loci in the populations studied. Thus, the association between PSP and the genotype A0/A0 is likely due to either linkage disequilibrium with another genomic change in the tau gene or another gene close by, or the possibility that the tau gene dinucleotide polymorphism itself, perhaps through altered tau gene expression, is responsible for increasing the risk of developing PSP.

Consistent with the previous finding,¹⁵ the allele with the highest frequency in our population was the A0 allele, followed by the A3 allele and the A1 allele (Table 1). However, the frequency of the A0 allele in our population (63.9%) was significantly lower (P<.001) than the corresponding frequencies previously reported,¹⁵ and the frequency of the A3 allele (31.9%) was significantly higher (P<.001) (Table 1). These differences are likely due to the different genetic backgrounds of the respective populations. The data provided herein (Table 1) together with those previously reported¹⁵ indicate that the genotype A0/A0 is a risk factor for PSP regardless of the genetic background of the respective populations (odds ratios from our data and the previous data¹⁵ were 10.62 and 15.60, respectively). A similar association has been reported between the ϵ4 allele of the APOE gene and AD.^8,9,19 Consistent with the above findings, the frequency of A0/A0 homozygous individuals in the control groups of our population was also lower than that in the population described by Conrad et al.¹⁵ The frequency of A0/A0 homozygous individuals with PSP was also slightly lower in our population (Table 1). However, the significance of the association between the A0 allele and PSP indicated by our results was higher (P =.00019) than that previously described (P =.001).

Another possible interpretation of our results could be that the differences among the PSP and control groups are due to a protective effect of the A3 allele rather than a deleterious effect of the A0 allele. The A0 and A3 alleles are the most frequent alleles in our population. Therefore, as expected, any significant increase in the A0 allele frequency was accompanied by a decrease in the A3 frequency (P<.001 when the A3 allele frequency in the PSP group was compared with the young subjects control group, and P =.01 when the PSP group was compared with the control group of spouses of those with AD). Consistently, there was also a significant decrease in the A0/A3 genotype frequency in our PSP group compared with our young control group (P =.001) or with the spouses of patients with AD control group (P =.02). It is now possible to search for the pathogenic mechanism or the genomic change near or within the tau gene polymorphism responsible for its association with PSP or a possible protective effect.

Accepted for publication January 5, 1998.

This study was supported by grant FIS96/0658 (Dr Oliva) from the Fondo de Investigaciones Sanitarias, Spain, and by grants 1997SRG00169 (Dr Oliva) and 1996SRG00064 (Dr Tolosa) from the Comissionat per a Universitats i Recerca de la Generalitat de Catalunya, Spain.

Reprints: Rafael Oliva, MD, PhD, Genetics Service, Hospital Clínic i Provincial, Villarroel 170, 08036 Barcelona, Spain (e-mail: oliva@medicina.ub.es).