OUP user menu

The ΔK280 Mutation in MAP tau Favors Exon 10 Skipping In Vivo

John C. van Swieten MD, PhD, Iraad F. Bronner MSc, Asma Azmani MD, PhD, Lies-Anne Severijnen MD, PhD, Wouter Kamphorst MD, PhD, Rivka Ravid PhD, Patrizia Rizzu PhD, Rob Willemsen PhD, Peter Heutink PhD
DOI: http://dx.doi.org/10.1097/nen.0b013e31802c39a4 17-25 First published online: 1 January 2007

Abstract

Tau mutations in frontotemporal dementia and parkinsonism linked to chromosome 17 (FTDP-17) are associated with changes in alternative splicing of exon 10. The ΔK280 mutation in exon 10 is exceptional because in vitro observations suggest a dramatic effect on microtubule binding, enhanced self-aggregation, as well as a decrease of the 4R/3R ratio by the ablation of an exon splicing enhancer element. Using immunohistochemistry, Western blotting, and electron microscopy on brain material with the ΔK280 mutation, we investigated which of these effects is most dominant in vivo. The brain showed abundant Pick bodies in several brain regions, which stained positive with 3-repeat-specific but not with 4-repeat-specific tau antibodies. Western blots of sarkosyl-insoluble tau showed exclusively three repeat (3R0N and 3R1N) tau in most regions, although some 4R1N could be detected in the frontal cortex. In addition, the sarkosyl-soluble tau fraction showed a significantly higher amount of 3-repeat tau. Because quantitative analysis of 4R and 3R mRNA transcripts showed a 4R/3R ratio of only 0.3, association between increased transcription and protein expression was observed. These observations confirm the postulated hypothesis that the ΔK280 mutation abolishes a splice enhancer element, which overrules the decreased microtubule binding and enhanced self-aggregation.

Key Words
  • 3-repeat tauopathy
  • Pick disease
  • Tau mutation

Introduction

Mutations in the tau gene are associated with familial frontotemporal dementia and parkinsonism linked to chromosome 17 (FTDP-17). FTDP-17 is usually characterized by neuronal and glial tau pathology, including neurofibrillary tangles, pretangles, coiled bodies, and tufted astrocytes (1-3). The more than 40 different identified mutations can be divided into 2 classes. The first class of mutations give rise to amino acid changes that affect the ability of tau to promote microtubule assembly in vitro (2, 4-7). The second class of mutations affect alternative splicing of exon 10, leading to changes in relative levels of mRNA encoding for tau protein with 3 amino-acid repeat regions (3R) versus isoforms with 4-repeat regions (4R) (2,8). The alternative splicing of MAP tau is regulated by a complex combination of intronic and exonic sequences and normally gives a ratio of approximately 1:1 for 3R:4R (9-11). Most of the "splice" mutations lead to an excess of 4R tau isoforms and are associated with the presence of both neuronal and glial tau aggregates consisting predominantly of 4R tau isoforms.

The ΔK280 mutation, resulting from a deletion of the nucleotides AAG in exon 10, was detected in a single Dutch patient with FTD as previously reported (12). In vitro experiments suggest that the lack of a lysine residue in 4R tau isoforms resulting from this mutation dramatically decreases both microtubule binding and strongly increases the induction of self-aggregation of tau into filaments (12-15). Strikingly, in vitro splicing assays showed that the ΔK280 mutation also affected the alternative splicing of exon 10 by abolishing an exon splicing enhancer element, resulting in a decrease of the 4R:3R ratio (8). This reduction of the 4R:3R ratio is unique because all other "splice" mutations have the opposite effect. In this case, the dramatic effect of the ΔK280 mutation on microtubule binding of 4R tau and self-aggregation would be overruled by the postulated excess of 3R and we would expect that tau aggregates should predominantly be composed of 3R tau analogous to Pick disease (8). To clarify which of these mechanisms is most important for the disease-causing process in this mutation, it is essential to investigate these effects in available brain material from a patient with this unique mutation.

Recently, formalin-fixed and fresh-frozen brain material from the only known patient with the ΔK280 mutation has become available and opens perspectives to address these issues by means of immunohistochemistry with isoform-specific antibodies, semiquantitative Western blotting, and quantitative mRNA analysis.

Materials and Methods

Clinical History

The patient with the ΔK280 mutation presented with restlessness, agitation, and emotional bluntness at the age of 53. She also developed loss of initiative and roaming behavior without losing her way. Family history revealed that her father had died from Parkinson disease and dementia at the age of 71, and her grandfather had also had dementia. The patient became impulsive and distractible.

Neuropsychologic testing showed impairment in executive functions, reduced mental flexibility, and verbal memory dysfunction. Neurologic examination revealed normal muscle, no rigidity or tremor, no signs of motor neuron disease, and no focal neurologic deficits. Magnetic resonance imaging of the brain showed asymmetric left-sided frontotemporal atrophy. Her speech gradually reduced to stereotyped phrases, which finally developed into mutism. She was admitted to a nursing home 4 years after onset and her condition gradually worsened. Neurologic reexamination 8 years after onset showed no extrapyramidal rigidity or postural instability. The patient died from sudden heart arrest at the age of 63 years.

Brain Pathology and Staining

Brain autopsy was carried out within 4 hours after death according to the Legal and Ethical Code of Conduct of the Netherlands Brain Bank. Tissue blocks were taken from the frontal, temporal, parietal, and occipital cortex, hippocampus, striatum, thalamus, substantia nigra, pons, medulla, and cerebellum and frozen at -80°C. Half of the brain was fixed in 10% buffered formalin solution for 4 weeks. Eight-micron paraffin-embedded sections of the same brain regions underwent routine staining with hematoxylin and eosin, Bodian, methenamine silver, Gallyas, and Congo red staining.

Immunohistochemistry was performed using phosphorylation-dependent antibodies: AT8 (1:40; Innogenetics, Ghent, Belgium), AT180 (1:100; Innogenetics), PHF1 (1:100; donated by P. Davies, Albert Einstein College of Medicine, New York, NY), antibodies directed against ubiquitin (1:500; Dako, Glostrup, Denmark), β-amyloid (1:100; Dako), and α-synuclein (1:1; Zymed Laboratories, San Francisco, CA). To expose antibody-specific epitopes, tissue sections were pretreated at 80°C for 30 minutes in 0.1 M sodium citrate buffer at pH 7.7 before incubation with ubiquitin antibody (16), whereas a 99% formic acid solution was used for pretreatment for α-synuclein (5 minutes) and β-amyloid (20 seconds). Before incubation with the anti-tau phosphorylation-specific Ser262 antibody (1:100; 577814; Calbiochem, San Diego, CA), tissue sections were pretreated by microwave. Tau antibodies specific against 3-repeat (RD3 [1:6000], Upstate, Charlottesville, VA) and 4-repeat tau isoforms (RD4 [1:100]; Upstate) were used to investigate the tau isoform composition of tau-positive inclusions (17). Sections from the ΔK280 brain as well as an Alzheimer disease (AD), progressive supranuclear palsy, (PSP), and Pick controls were pretreated by pressure-cooking in 0.1 M sodium citrate buffer (pH 6) for 5 minutes followed by incubation with RD3 (1:6000) and RD4 (1:100) at 4°C overnight (17). The Histostain-Plus broad-spectrum kit DAB (Zymed) was used as a detection system.

Tau Extraction, Immunoblotting, and Electron Microscopy

Sarkosyl-soluble and -insoluble tau proteins were extracted from the inferior frontal gyrus, medial temporal gyrus, hippocampus, caudate nucleus, and mesencephalon as described (18). Half of the sarkosyl-insoluble tau pellet was stored at −80°C until loaded on a 12% SDS-polyacrylamide gel (PAGE, see subsequently). The other half was incubated in 25 mM Tris-HCl pH 7.5 containing 4 M guanidine hydrochloride for 1 hour. Next, the guanidine hydrochloride was washed away and the sample concentrated using Microcon centrifugal filter devices (Millipore, Bedford, MA). Approximately one tenth of sarkosyl-soluble tau extracted from 100 mg frozen brain was also concentrated using Microcon centrifugal filter devices (Millipore). Subsequently, both the sarkosyl-insoluble and sarkosyl-soluble tau samples were redissolved in 50 mM Tris-HCl pH 7.5 and incubated with 0.6 units of bacterial alkaline phosphatase (Sigma Benelux, Aalsmeer, The Netherlands) overnight. All samples were run on a 10% 1:37.5 Bis-/acrylamide SDS-PAGE. Immunoblotting using the phosphorylation-independent anti-tau antibody H-7 (1:2000) (19) was performed as described (16, 20). Recombinant tau was a generous gift from Dr. M. Goedert, MRC Laboratory of Molecular Biology, Cambridge, United Kingdom.

Western blots containing sarkosyl-soluble tau homogenates from different regions of the ΔK280 patient's brain, an AD brain, and a control brain were exposed to Fuji Super RX film (Fuji Photo Film Co., Tokyo, Japan) using ECL (Amersham Biosciences, Buckinghamshire, UK) as a detection agent. As a result of the nonlinear nature of film, 2 separate Western blots containing the same samples were exposed to the same film and scanned using a HP Scanjet 3570c scanner (Hewlett-Packard Nederland B.V., Amstelveen, The Netherlands). Obtained TIFF files were imported into ImageJ (21, 22) and lane profile plots were generated. Next, peak areas were measured and 3R/4R soluble tau ratios were determined. Statistical significance between samples and controls was calculated using a one-way analysis of variance, including a Tukey honest significant difference (HSD) post hoc test to determine differences between regions and control.

Electron Microscopy

Sarkosyl-insoluble tau extracted from freshly frozen medial temporal gyrus with abundant and representative tau pathology was processed for electron microscopy as described (18, 23). A small droplet of resuspended sarkosyl-insoluble pellet was placed on a formvar-coated copper grid (200-mesh) for 2 minutes. Unlabeled filaments were prepared by negative staining with 2% uranyl acetate. For antibody labeling, the grid with adsorbed specimen was placed on a drop of phosphate-buffered saline (PBS)+ (BSA 0.5%, glycin 0.15%) and blocked for 15 minutes. Subsequently, the grid was placed on a solution of the (AT8 1:100; Innogenetics) primary antibody for 30 minutes at room temperature followed by 3 washing steps in PBS+. Antigen-antibody complexes were visualized with a secondary antibody conjugated to gold 10 nm (Aurion, Wageningen, The Netherlands) for 30 minutes at room temperature. The grid was washed with PBS+ and aquadest and subsequently stained with 2% uranyl acetate and allowed to air dry.

Tau Haplotyping, cDNA Synthesis, and 4R/3R Quantification

Tau haplotype was determined on genomic DNA as described previously (24). RNA was extracted from snap-frozen medial temporal lobe using RNA-Bee (IsoTex Diagnostics, Friendswood, TX) according to standard protocols with the modification that the volume of RNA-Bee per tissue weight was doubled. RNA was purified using QIAgen RNeasy Kit (Qiagen Benelux, Venlo, The Netherlands) according to manufacturer protocol. cDNA was made according to standard protocols using Superscript III (Invitrogen, Breda, The Netherlands) at 50°C. Polymerase chain reaction (PCR) was performed between exon 9 and 11 and exon 9 and 13 in triplicate as described previously (2) except that FAM was used as fluorochrome on the 5′ end of the forward primer. For every experiment, the optimal amount of cycles (i.e. 27-30) was determined on agarose gel. PCR products were confirmed by sequencing. Quantitative fluorescent signals were detected on an ABI 3730 and analyzed using Genemapper 3.7 (Applied Biosystems, Foster City, CA).

Results

Pathology

The brain (weight 870 g) showed severe atrophy of the frontal and temporal cortex extending into the inferior parietal lobe (no macroscopic pictures). The ventricles were enlarged and the white matter of the frontal and temporal lobes was reduced. The caudate nucleus, the amygdala, and hippocampus were very atrophic, whereas thalamus, putamen, pallidum, and substantia nigra had a normal appearance. Microscopically, the frontal and the anterior pole of the temporal cortex showed subtotal loss of neurons with severe gliosis (Fig. 1), including the underlying white matter. There was severe neuronal loss in the granular layer of the dentate gyrus, Ammon's horn 1-4, subiculum, parahippocampal gyrus, and in the inferior parietal cortex. The caudate nucleus, claustrum, medial part of the thalamus, subthalamic nucleus, and insular cortex also showed severe gliosis and neuronal loss. Mesencephalon, pons, locus coeruleus, medulla, and cerebellum were normal. Bodian staining showed argyrophilic round inclusions resembling Pick bodies in the granular cells of the dentate gyrus, in the remaining neurons of CA 1-4, and the frontal and temporal cortex. Some neurons of the substantia nigra, caudate nucleus, and amygdala contained round argyrophilic inclusions as well. Pick bodies did not stain with Gallyas silver stain.

FIGURE 1.

Hematoxylin & eosin (H&E) staining of the frontal cortex. (A) Severe neuronal loss in layer III and spongiform changes in layer II of the frontal cortex and (B) severe gliosis of the frontal cortex (H&E, original magnification: 200×). (C) Neuronal loss in layer II and III of the temporal cortex (H&E, original magnification: 200×). (D) Mild neuronal loss in layer II and III of the inferior parietal cortex (H&E, original magnification: 200×). Scale bar = 100 μm.

Immunohistochemical staining with antibody AT8 showed numerous Pick bodies in the dentate gyrus (Fig. 2A) and in neurons of CA1-4 subiculum and entorhinal cortex. In addition, many Pick bodies were present in neurons of layer II and III and V and VI of the frontal cortex (Fig. 2B) and temporal and cingular cortices. Numerous small neurons in the caudate nucleus also contained tau-positive inclusions (Fig. 2G). Abundant tau-positive astrocytes and a few oligodendroglial cells were present in the cortical regions (Fig. 2B), whereas the white matter showed very few glial inclusions. The amygdala showed many tau-positive neurons, some of them with Pick bodies and others with more granular staining. The substantia nigra exhibited abundant tau-positive neuropil threads, some tau-positive neurons, and a few tau-positive astrocytes. Eccentric Pick bodies or neurons with more granular staining were visible in the pontine nuclei.

FIGURE 2.

Immunohistochemistry with tau antibodies. (A) Numerous tau-positive Pick bodies in the granular cells of the dentate gyrus. (B) The frontal cortex showed tau-positive Pick bodies and also tau-positive astrocytes (AT8 tau, original magnification: 400×). Positive staining of Pick bodies with a 3-repeat tau specific antibody in the dentate gyrus (C) and in the frontal cortex (D) (RD3, original magnification: 400×). (E, F) The same regions showed negative staining with a 4-repeat tau specific antibody (RD4, original magnification: 400×). (G) Tau-positive Pick bodies were also present in the caudate nucleus (AT8, original magnification: 400×). (H) RD4 antibody gave positive staining of globoid tangles, coiled bodies, and tufted astrocytes with a 4-repeat specific antibody in the striatum of a control progressive supranuclear palsy brain. (B, H) Black arrows point to tufted astrocytes. (H) White arrows point to oligodendroglial tau-positive inclusions. Scale bar = (A-H) 100 μm.

Immunohistochemistry with the antibody PHF gave a similar pattern of staining, although less intense. The staining with AT180 was negative but positive for the AD and Pick samples that were used as positive controls. The staining with the Ser262 antibody was negative for Pick bodies, although a few of them and some astrocytic dendritic processes gave some punctate staining.

Staining with the 3-repeat specific antibody, RD3, revealed many Pick bodies in the granular cells of the dentate gyrus (Fig. 2C) and in the pyramidal cells of CA2 and 3, but did not stain the AT8-positive astrocytic ramifications. Abundant RD3-positive neurons were also present in the substantia nigra, caudate nucleus, and pontine nuclei. In contrast, RD4 antibody did not stain Pick bodies and also failed to stain the tau-positive astrocytic ramifications in the ΔK280 brain (Fig. 2F), whereas NFTs in AD and PSP were RD4-positive (Fig. 2H). Pick bodies showed weak staining with ubiquitin antibody and negative staining with α-synuclein. Staining with β-amyloid was negative as well.

Characterization of Frozen Brain Material

Western blots of the sarkosyl-insoluble fraction of tau from frontal and temporal cortex, hippocampus, and caudate nucleus showed a 60- and a 64-kDa band (Fig. 3A). After dephosphorylation, sarkosyl-insoluble tau from the temporal cortex and hippocampus consisted of 3-repeat (3R0N and 3R1N) tau exclusively, whereas the frontal cortex showed 3R0N and 3R1N and a small amount 4R1N (Fig. 3B). These results are surprising if we consider the available in vitro data on the enhanced self-aggregation properties of 4R tau with a ΔK280 mutation but they are consistent with an effect of the mutation on the alternative splicing of exon 10.

FIGURE 3.

Western blots of sarkosyl-insoluble tau protein. (A) Hyperphosphorylated sarkosyl-insoluble tau from the ΔK280 patient's brain was run in two gels: gel 1: Alzheimer-diseased brain (AD) and temporal cortex (T); and gel 2: AD, caudate (Cau), frontal cortex (F), and hippocampus (H). All regions showed only 60- and 64-kDa bands. (B) Dephosphorylated sarkosyl-insoluble tau from the ΔK280 patient's brain. Pathogenic tau is hyperphosphorylated. Because of this posttranslational modification, the behavior on SDS-PAGE changed. To determine the precise isoform contents of the pathogenic tau, hyperphosphorylated tau was dephosphorylated. Temporal (T), hippocampus (H), frontal cortex (F), and recombinant tau (Rec). All regions contained mainly 3R0N and 3R1N sarkosyl-insoluble tau. In the temporal lobe, some 4R01 is present, whereas in the frontal lobe, some 4R0N and 4R1N is seen. (C) Sarkosyl-soluble fraction from Alzheimer control (AD), unaffected control (Con), caudate nucleus (Cau), cerebellum (Ce), frontal cortex (F), substantia nigra (SN), and hippocampus (H). All regions of the ΔK280 patient's brain showed a relative predominance of 3R1N over 4R1N and 4R0N, especially compared with both Alzheimer and unaffected controls. Names are according to tau isoforms. Sizes are in kDa. (D) Graph of 4R/3R signal ratios on Western blot. Unaffected control (Con), Alzheimer control, and brain regions from the ΔK280 patient (AD), caudate putamen (Cau), cerebellum (Ce), frontal cortex (F), substantia nigra (SN), and hippocampus (H). Signal intensities were quantified from 2 separate blots using ImageJ and ratios were calculated. *, p < 0.01; **, p < 0.001 (one-way analysis of variance) when compared with both Alzheimer and unaffected control. Error bars depict 95% confidence intervals. Higher 3R/4R signal ratios in frontal lobe and hippocampus could not be confirmed as a result of low signal intensities, the limited amount of brain material, and nonlinearity of film.

Western blots of the sarkosyl-soluble fraction from several regions demonstrated all 6 isoforms, although the soluble tau fraction showed a higher expression of 3R1N in all regions compared with the unaffected or AD controls (Fig. 3C).

When 3R and 4R tau expression in those regions was compared (twice) with the expression in unaffected or AD controls using ImageJ (21, 22), the 3R/4R tau ratio was significantly higher (1.5-4 times excess of 3R to 4R-soluble tau) in cerebellum (p = 0.01), caudate putamen (p < 0.0001), and substantia nigra (p < 0.001) in the ΔK280 patient, whereas in unaffected or AD control material, this ratio was approximately one (0.9-1.0) (Fig. 3D). Significance was determined using a one-way analysis of variance including a Tukey HSD post hoc test. When multiple blots were compared, expression of 3R1N was also higher in the frontal lobe. However, as a result of the limited amount of brain material, together with nonlinear properties of signal strength using film, this higher ratio of 3R tau could not be established in ΔK280 frontal lobe or possibly also in hippocampus material (Fig. 3D).

In conclusion, the combined results of the insoluble and soluble fractions strongly suggest that less 4R tau is produced, which would be consistent with the predicted effect of the ΔK280 mutation on the alternative splicing of exon 10.

Sarkosyl-Insoluble tau Filaments

Electron microscopy of sarkosyl-insoluble tau material extracted from the temporal cortex showed a low number of twisted and straight filaments that were positively labeled with antibody AT-8 (Fig. 4). The majority of filaments were narrowly twisted with a diameter of 7 to 8 nm and a half-periodicity of 95 nm, similar to twisted ribbons.

FIGURE 4.

Electron microscopy of ΔK280 tau filaments. Electron microscopic examination of sarkosyl-insoluble tau extracted from freshly frozen frontal lobe of the patient's brain. Sarkosyl-insoluble tau was fixed and mounted on a grid. Several filaments were measured and average half-periodicity and diameter were calculated. (A) Paired helical filaments with a diameter of 7 to 8 nm and half-periodicity of 95 nm labeled with gold conjugated AT8. (B) Straight filaments labeled with gold conjugated AT8. In both micrographs, the scale bar depicts 50 nm. White arrows point to crossovers of filaments.

Quantitative Analysis of 4R/3R RNA Isoform Ratio In Vivo

The mRNA expression of 3R and 4R tau from this patient was compared with that from 4 controls and a patient with dementia-disinhibition-parkinsonism-amyotrophy complex (DDPAC) carrying a +14 splice site mutation for whom we have previously shown an increased 4R/3R ratio (2). All individuals were carrying an H1/H1 haplotype; therefore, all mutant alleles were expressed in an H1 background. Two separate reverse transcriptase-polymerase chain reactions were done, one reaction amplifying exons 9 to 11 and a separate reaction essentially amplifying exons 9 to 13 (2). All experiments were conducted in triplicate. Both reactions generated 2 products, one corresponding to the 3R mRNA isoform and the other to the 4R isoform. In controls, the ratio between 4R and 3R was on average 0.7 for reactions amplifying exons 9 to 13 and 0.8 for reactions including exon 9 to 11 (Fig. 5). As previously described, the patient carrying the +14 splice site mutation showed an increased 4R/3R ratio of 3.0 for reactions, including exons 9 to 13, and 3.3 for reactions, including exon 9 to 11 (Fig. 5) (2). The ΔK280 patient showed a 4R/3R ratio of only 0.3 for both reactions (Fig. 5). These findings indicate that although 4R expression is largely decreased, the change in 3R/4R ratio with a factor of approximately 2.3 is smaller than that observed in the DDPAC patient (a factor of approximately 4.3). In conclusion, the ΔK280 mutation has a significant effect on the alternative splicing of exon 10; however, the effect is less dramatic than previously suggested by in vitro splicing assays (8).

FIGURE 5.

Ratio between 4- and 3-repeat tau. All experiments were done in triplicate. Two separate reverse transcriptase-polymerase chain reactions were done. The black bars depict the ratio of amplification reactions for exon 9 to 13 and the gray bars depict the ratio of reactions amplifying exons 9 to 11. Error bars depict the standard error. Delta K280 is the patient with the ΔK280 mutation and the DDPAC patient carries a + 14 splice site mutation. DDPAC indicates dementia-disinhibition-parkinsonism-amyotrophy complex.

Discussion

The present study shows that the ΔK280 mutation in the tau gene is associated with the presence of abundant Pick bodies consisting predominantly of 3-repeat tau isoforms as was shown with immunohistochemistry using a 3-repeat specific antibody and Western blotting of dephosphorylated sarkosyl-insoluble tau.

This neuropathologic picture of Pick bodies in the absence of glial tangles or coiled bodies clearly distinguishes this patient from patients with mutations, resulting in the overexpression of 4R mRNA transcripts characterized by abundant neurofibrillary and glial tangles. It is also in contrast to the observations of the E10 + 19 mutation with decreased E10 inclusion in splicing assays in which tau pathology is lacking in the brain (25). The 4R-specific antibody stained the threads and tufts in the L266V mutation with Pick bodies (26,27), whereas the AT8-positive astrocytic ramifications in the present mutation failed to stain with both RD3 and RD4 antibody. The fact that we could not draw any conclusions about the exact (3R or 4R) composition within these tau-positive astrocytes could be explained by the low amount of abnormal tau present within astrocytes. Apparently, the affinity of RD3 and RD4 for abnormal tau (compared with AT8) was below the threshold for visualizing these lesions using these antibodies.

The relative overexpression of 3R transcripts confirms the hypothesis that the ΔK280 mutation disrupts an exon splicing enhancer that plays a role in a complex network of exonic and intronic splicing regulatory sequences (11). However, the mutation only results in an approximately 2-fold decrease in 4R mRNA transcripts. This is consistent with the observation that the relative excess of 3R over 4R tau in the sarkosyl-soluble fraction is lower than a ratio of 75:25, although as a result of the nonlinear properties of film, an exact 3R/4R tau ratio could not be established with this method.

Our results only partly confirm the results of in vitro splicing assays that suggest an almost exclusive production of 3R mRNA transcripts (8). However, one should keep in mind that these in vitro assays are artificial and can only partly validate the more complex mechanisms that occur in brain. Therefore, it is very likely that small amounts of 4R tau isoforms containing the ΔK280 mutation are still produced. The nearly complete absence of 4R tau in neuronal and astrocytic inclusions and sarkosyl-insoluble tau fraction of the ΔK280 brain is intriguing in the light of its effects on microtubule binding and self-aggregation. In vitro results have shown that this mutation in the VQIVYNK motif of the second repeat has the largest effect on microtubule binding compared with other mutations (13). Additionally, mutant ΔK280 tau leads to remarkable rapid and extensive aggregation of tau into filaments by increasing the β-structure propensity within the repeat region in vitro (13). Therefore, one might hypothesize that the small amount of mutant 4R tau would act as a seed for tau aggregation and that sarkosyl-insoluble tau would contain a significant amount of 4R mutant tau. However, only in the frontal cortex a small amount of sarkosyl-insoluble 4R1N tau could be detected on Western blots (Fig. 3B). Furthermore, it is unknown whether this is normal or mutant 4R tau. We have attempted to generate specific peptide-based polyclonal antibodies against the mutant ΔK280 tau but have failed to obtain a good antigenic response (data not shown).

Alternatively, the increased amount of unbound 3R tau protein might act as a seed and during the further process, some 4R normal or mutant protein is trapped in the aggregates. This mechanism is essentially similar to the effect of the splice mutations leading to excess of 4R tau protein isoforms.

The molecular mechanism of neuronal cell death and selective aggregation of 4R isoforms in intronic mutations with overexpression of 4R tau is poorly understood (28-31). Also, in AD, a shift in the ratio of 3R tau to 4R tau mRNA has been found in individual neurons of the human cholinergic basal forebrain (32). Analogous to this, it is also unknown how the relative overexpression of 3R tau in the ΔK280 mutation would lead to neuronal cell loss and accumulation of exclusively 3R tau (33). The 3R tau isoforms have a significantly weaker ability to stabilize microtubules. Furthermore, 3R isoforms show a reduced rate and extent of microtubule growth when compared with 4R tau isoforms. Therefore, an increase in 3R tau protein isoforms would lead to increased microtubule shortening and decreased microtubule stabilization (33). This is likely to be explained by specific sequences within and outside the first 2-repeat regions and their interrepeat in 3-repeat and 4-repeat isoforms, which are responsible for isoform-specific conformational changes and lead to isoform-specific interactions of tau with microtubules (34). Additionally, there is evidence for 2 distinct microtubule-bound tau populations in which tau might be reversibly and weakly bound to the surface of microtubules or might be stronger and irreversibly incorporated into microtubule network (35-37). The difference in rate of microtubule assembly and the isoform-specific interactions may explain why a change in the ratio of tau isoforms will have deleterious changes in the regulation of microtubule dynamics in neurons leading to cell death (38).

The morphology of filaments with a crossover spacing of 95 nm in the ΔK280 mutation shows a typical paired helical structure, as expected (13). They differ from those found in other E10 splicing mutations in which there is an accumulation of 4-repeat tau isoforms (3, 39). This is not surprising because the conformational characteristics of filaments are highly determined by the seventh residue of the second and third repeat region (40).

In conclusion, the immunohistochemical and biochemical features of the ΔK280 mutation confirm the hypothesis of decreased splicing of exon 10 and relative increase in 3R tau protein isoforms. Second, the present observations prove the etiologic significance of an excess of 3R tau to the disease process. Although the pathophysiological significance of mutant 4R tau could not be ruled out, tau deposits do not contain mutant 4R tau at significant levels. However, because mutant 4R tau isoforms appear to be produced, its contribution to the disease process remains unclear.

Acknowledgments

The authors thank Dr. M. Goedert for providing the recombinant tau, Dr. P. Seubert for donating the antibody 12E8, Dr. P. Davies for providing the antibodies PHF1 and MC1, Dr. Z. Bochdanovits for statistical analysis, and T. de Vries Lentsch for photography and artwork.

Footnotes

  • This study was supported by the Hersenstichting (project 13F05 (2). 14) and by the Centre for Medical Systems Biology (CMSB), a centre of excellence approved by the Netherlands Genomics Initiative/Netherlands Organisation for Scientific Research (NWO); and the Center for Biomedical Genetics.

References

  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.
  22. 22.
  23. 23.
  24. 24.
  25. 25.
  26. 26.
  27. 27.
  28. 28.
  29. 29.
  30. 30.
  31. 31.
  32. 32.
  33. 33.
  34. 34.
  35. 35.
  36. 36.
  37. 37.
  38. 38.
  39. 39.
  40. 40.
View Abstract