Introduction

Intra- or extracellular accumulation of abnormal conformers of physiological proteins like β-amyloid derived from the amyloid precursor protein, the microtubule-associated protein tau, α-synuclein, prion protein, or proteins linked to trinucleotide repeat disorders like huntingtin characterizes neurodegenerative diseases [22]. Immunohistochemical detection of α-synuclein, phosphorylated tau, abnormal prion protein, and β-amyloid has become the histological technique of choice for the diagnosis [6]. However, as previous studies pointed out, variability across antibodies and antigen unmasking methods may account for the discrepancies between research studies and affect diagnostic accuracy (e.g. GFAP, ubiquitin, prion protein, α-synuclein) [5, 8, 12, 13]. This is of outmost importance when searching immunohistochemically for novel proteins, like TPPP/p25, related to pathological protein aggregates.

Tubulin polymerization promoting protein TPPP/p25 or p25α was originally co-purified with a tau kinase preparation from bovine brain [24]. In normal rat and human brain it was localized predominantly in oligodendrocytes [14, 18, 23, 25]. This protein must be distinguished from the extensively characterized p25 (hence introduction of the term TPPP/p25), which is a truncated form of p35 that deregulates cyclin-dependent kinase activity by causing prolonged activation and mislocalization of the kinase [20]. TPPP/p25 promotes the polymerization of tubulin into double-walled tubules and polymorphic aggregates [9]. It inhibits mitotic spindle assembly without affecting other cellular events like centrosome replication and segregation, nucleation of microtubules by the centrosomes, and nuclear growth [26]. In addition, it inhibits glycogen synthase kinase 3 that phosphorylates tau protein [16]. Purified recombinant human TPPP/p25 strongly stimulates the aggregation of α-synuclein in vitro [15].

Recently we demonstrated that TPPP/p25 is enriched in filamentous α-synuclein bearing Lewy-bodies of Parkinson’s (PD) and diffuse Lewy-body diseases (DLBD), as well as in glial cytoplasmic inclusions (Papp-Lantos bodies) of multiple system atrophy (MSA) [14]. Furthermore, it was reported in α-synuclein immunonegative neuronal inclusions in MSA [2, 10]. Previous studies on human diseased brain tissue agreed that TPPP/p25 is a marker of pathological α-synuclein immunoreactive structures [2, 14, 15], although the amount of immunolabelled neuronal inclusions of PD and DLBD was different [14, 15]. In our present study we evaluate antigen unmasking methods and antibodies raised against a small fragment and the full length of TPPP/p25. We extend the number of investigated neurodegenerative disorders to further specify the possible pathogenic pro-aggregatory role of TPPP/p25.

Materials and methods

We included three neuropathologically characterized cases each of Alzheimer’s disease (AD), DLBD, PD, MSA, progressive supranuclear palsy (PSP), corticobasal degeneration (CBD, only one case), Pick’s disease (PiD), motor neuron disease (MND, five cases), argyrophilic grain disease (AgD), frontotemporal lobar degeneration with ubiquitin-only immunoreactive neuronal changes (FTLD-U), Huntington’s disease (HD), and control brains (Co, five cases). Diagnoses of controls comprise extraneural carcinoma, HIV infection without affecting CNS, and alcoholic encephalopathy. Data of cases and examined anatomical regions are summarized in Table 1.

Table 1 Data of examined cases and anatomical regions
Full size table

Production of TPPP/p25 antibodies

Two different antibodies were used to immunodetect TPPP/p25 in human brain tissue.

Antibody TPPP/p25-A

Immunogen (peptide of 184–200 aa, from sequence of hp25) was synthesized by standard Fmoc strategy. Extra Cys amino acid was added at the first N-terminal position for aiming conjugation. After conjugation of the unblocked peptide to KHL, rats were immunized by standard protocol, with a 4-week interval of boosting injections. After the third boosting, blood samples of the animals were withdrawn and tested against rechp25 by ELISA, and by western blotting on human cortex extract. This antibody was described already in detail [14].

Antibody TPPP/p25-B

This antibody is raised against full-length human recombinant TPPP/p25. His-tagged rechp25 was purified by IMAC chromatography from E. coli lysate and was used as immunogen. The purity of the protein was controlled on SDS-PAGE. Rats were immunized against the protein according to the same protocol as detailed above, but without conjugation. After the third boosting, blood samples of the animals were withdrawn and tested against rechp25 by ELISA, and by western blotting on human cortex extract. Specificity of this antibody was described earlier [19]. We also tested it in Western blots against human cortex extracts (data not shown). For immunohistochemistry we omitted the primary antibody or we incubated the peptide with the TPPP-B antibody before immunostaining and we did not observe any immunoreactivity as detailed below (data not shown).

Pilot study of pre-treatments

Using adjacent sections of pons from MSA and mesencephalon from PD cases, we tested both anti-TPPP/p25 antibodies applying the following epitope retrieval methods: (A) 10 min autoclaving (100°C) in 0.01 M citrate buffer (pH 6.0); (B) pre-treatment (A) followed by 1 min formic acid (88%) treatment; and (C) 0.03% proteinase K treatment for 10 min.

Immunohistochemistry

We immunostained adjacent sections of different anatomical regions (Table 1) using two different TPPP/p25 antibodies (A and B, both rat polyclonal, 1:200, over night incubation) and two different α-synuclein antibodies (mouse monoclonal, 1:10.000, clone 4D6, immunogen: human α-synuclein, Signet, Dedham, MA, USA; rabbit polyclonal, 1:200, immunogen C-terminus aa 111–132, Sigma, St. Louis, MO, USA) using the same epitope retrieval including 10 min 0.01 M citrate buffer (pH 6.0) followed by 1 min 88% formic acid. Additionally, in AD, PSP, CBD, PiD, AgD we applied anti-phospho-tau (mouse monoclonal, clone AT8, 1:200 Pierce Biotechnology, Rockford, IL, USA), in FTLD-U and MND anti-ubiquitin (rabbit polyclonal, 1:200, Dako, Glostrup, Denmark), and in HD anti-huntingtin antibody (mouse monoclonal, 1:100, Chemicon, Temecula, CA, USA). As a secondary system we used the Envision kit (Dako, Glostrup, Denmark). In addition, for double immunolabelling we used GFAP (mouse monoclonal, 1:50, Chemicon, Temecula, CA, USA) and S100 (rabbit polyclonal, 1:100, Dako, Glostrup, Denmark) antibodies.

Double immunolabelling and laser confocal microscopy

Double immunofluorescent labelling was evaluated by a ZEISS LSM 510 confocal laser microscope. The fluorescent-labelled secondary antibody for TPPP/p25-A, and C was Alexa Fluor 633 goat anti-rabbit IgG (1:200, Molecular Probes Inc. Eugene, OR, USA), for ubiquitin and S100 was Alexa Fluor 488 goat anti-mouse IgG (1:200, Molecular Probes), and for AT8, GFAP, huntingtin, and alpha-synuclein Alexa Fluor 488 goat anti-mouse IgG (1:200, Molecular Probes). Primary antibodies were incubated over night. For TPPP/p25-A we applied rabbit anti-rat (1:200, Dako) as intermediary incubation step. We applied Sytox orange (1:8000, Molecular Probes) for 15 min to detect nuclear staining using Helium/Neon 546 nm laser. We used Argon 488 nm and Helium/Neon 633 nm lasers to elicit immunofluorescent staining. In order to control specific immunofluroescence we omitted the primary antibodies and used intermediary and/or secondary antibodies without detecting any fluorescence.

Quantitative evaluation

To evaluate the amount of TPPP/p25-A, B, and α-synuclein immunopositive pathological structures, we counted intra- and extracellular globular bodies in the same 1.0 × 1.0 cm framed temporal cortical area in adjacent sections in three DLBD cases using objective ×40. Similarly, we counted neuronal and glial cytoplasmic TPPP/p25-A, B, and α-synuclein immunoreactive profiles in the same 1.0 × 1.0 cm framed area in adjacent sections containing the pontine base in three MSA cases using objective ×40.

Results

Comparison of epitope unmasking methods

In our hands 10 min autoclaving (100°C) in 0.01 M citrate buffer (pH 6.0) followed by 1 min formic acid (88%) treatment (pre-treatment B) was the most efficient method for the immunodetection of pathological TPPP/p25 immunoreactive structures (Fig. 1a–c). Immunolabelling of oligodendrocytes was detectable with all three pre-treatments. Thus, to compare immunoreactivity for TPPP/p25-A and B we used pre-treatment B.

Fig. 1
figure 1

Testing of different epitope retrieval methods (a–c) and TPPP/p25 antibodies (d–r) in adjacent sections of pons of a MSA case (a–c, g–r) and mesencephalon of a PD case (d–f). Epitope retrieval methods include (a) 10 min autoclaving (100°C) in citrate buffer (pH 6.0), (b) the previous pre-treatment followed by 1 min formic acid (88%) treatment, and (c) 0.03% proteinase K treatment for 10 min. All photos represent antibody TPPP/p25-A and shows the same area with neurites and neuronal inclusions. Using the pre-treatment method shown in (b), TPPP/p25-A (e, h, k, n, q) gives better result when compared to immunoreactivity observed by using polyclonal anti-(-synuclein antibody (d, g, j, m, p), and TPPP/p25-B (f, i, l, o, r). Bar indicates 10 (m in a–c, 5 (m in d–i, m–o, and 30 (m in j–l and p–r

Full size image

Comparison of antibodies

Results, including present and previous observations, are summarized in Tables 2 and 3.

Table 2 Summary of immunoreactivity using two different anti-TPPP/p25 and two distinct anti-α-synuclein antibodies (see also [14])
Full size table
Table 3 Summary of immunoreactivity observed using anti-TPPP/p25-A, B and anti-α-synuclein antibodies
Full size table

In summary, TPPP/p25 immunoreactivity was observed in S100 immunopositive oligodendroglia-like cells and not in GFAP immunopositive astroglial cells (Fig. 2a, b) in all cases. TPPP/p25-A was less efficient in detecting oligodendroglial cells than antibody B. Lewy-bodies and Lewy-neurites in PD and DLBD cases (Fig. 1d–f), neuromelanin, and lipofuscin-like dots in all cases were immunostained predominantly with antibody TPPP/p25-A. It must be noted that anti-TPPP/p25-A immunolabelled practically all morphological types of pathological α-synuclein immunoreactive structures (Fig. 3). In contrast, TPPP/p25-B infrequently immunolabelled the rim of brainstem-type classical Lewy-bodies. This was never seen in cortical-type Lewy-bodies.

Fig. 2
figure 2

Double immunolabelling laser confocal microscopical images of S100 (a), GFAP (b), monoclonal anti-α-synuclein antibody (c, d), and anti-huntingtin (e) and anti-TPPP/p25-A (a–e). Sytox indicates the staining of nuclei. Bar indicates 10 μm in a–c; 5 μm. Note that TPPP/p25 co-localizes with S100 (a) and not with GFAP (b). There is a high degree overlap of TPPP/p25 and α-synuclein in glial cytoplasmic inclusions in MSA (c). In neuronal cytoplasmic inclusions, co-localization is present in contrast to nuclear pathological structures (arrowhead, d). In Huntington’s disease, TPPP/p25 is present only in some extracellular neuritic profiles and not in intranuclear huntingtin immunoreactive inclusions (e). Here, additional granular TPPP/p25-A immunoreactivity is seen in astroglia like cells

Full size image
Fig. 3
figure 3

Immunostaining for TPPP/p25 in various pathological α-synuclein immunopositive neuronal structures, comprising (a) early dot-like accumulation of TPPP/p25 immunoreactivity (indicated by the white asterisk) beside neuromelanin granules, (b) classical Lewy-bodies, and (c) Lewy-neurites in a representative case of Parkinson’s disease. Similarly, in diffuse Lewy-body disease, different morphological variants of cortical Lewy-bodies (d, e) and α-synuclein immunopositive Lewy-neurites in the CA2/3 plexus (f) are strongly immunostained with anti-TPPP/p25. All photographs represent sections immunostained using anti-TPPP/p25-A antibody. Bars represent 10 μm in a–e and 25 μm in f

Full size image

Both TPPP/p25 antibodies immunostained some granules in granulovacuolar degeneration, seen mostly in AD and PiD cases. In addition, glial and neuronal cytoplasmic inclusions and α-synuclein immunopositive threads in MSA cases were immunolabelled by both anti-TPPP/p25 antibodies (Figs. 1g–r, 2c, d). We observed immunoreactivity for neuronal nuclear inclusions in MSA cases very rarely with anti-TPPP/p25-A antibody, and these intranuclear inclusions were undetectable with anti-TPPP/p25-B antibody. Interestingly, huntingtin immunopositive extracellular globular immunodeposits, but not intranuclear inclusions, were occasionally co-immunostained with anti-TPPP/p25-A (Fig. 2e). This was not observed with two anti-α-synuclein and anti-TPPP/p25-B antibodies. In HD cases, in severely degenerated areas we noted granular TPPP/p25 immunoreactivity in astroglia-like cells not immunolabelled with anti-huntingtin (Fig. 2e).

Quantification of TPPP/p25 and α-synuclein immunoreactivity

The median value of intra- and extracellular globular bodies in DLBD immunolabelled with the monoclonal and polyclonal anti-α-synuclein antibodies was 127 and 128, and with anti-TPPP/p25-A antibody it was 103 (≈80% of α-synuclein immunopositive structures). The median value for glial cytoplasmic inclusions in MSA counted in sections immunostained with the monoclonal anti-α-synuclein antibody was 24, with polyclonal anti-α-synuclein antibody was 26, with anti-TPPP/p25-A was 20.6 (≈80%), and with TPPP/p25-B it was 14 (≈54%). The median value for neuronal cytoplasmic inclusions in MSA counted in sections immunostained with the monoclonal anti-α-synuclein antibody was 18, with the polyclonal anti-α-synuclein antibody was 23, and with the anti-TPPP/p25-A and B antibodies was 10 (≈50%) and none, respectively.

Discussion

In the present study we show that immunostaining for TPPP/p25 in various pathological structures in human brain depends on the antibody and pre-treatment used. This is similar for other antibodies, such as α-synuclein [5], and supports the need for harmonization of laboratory methodology in the evaluation of neurodegenerative disorders [1].

We demonstrate that one antibody, raised against aa 184–200 segment of TPPP/p25 is better in immunolabelling the majority of α-synuclein immunopositive neuronal and glial pathological profiles detectable in PD, DLBD, and MSA, while another, raised against full-length human recombinant TPPP/p25 is more suitable to immunodetect normal oligodendrocytes, rendering the latter antibody more convenient to evaluate disease states affecting oligodendrocytes (e.g. brain tumours [21], myelination) (Table 3). Since pre-treatment with formic acid, which we selected in our study, is well known to enhance the immunodetection of aggregated proteins and those showing amyloid features [11], it is likely that TPPP/p25 is also in an altered, aggregated state. The process of pathological TPPP/p25 accumulation in oligodendrocytes comprises the expansion [15] of the regularly seen perinuclear cytoplasmic immunoreactivity and further its deposition in typical glial cytoplasmic inclusions (Papp-Lantos bodies) of MSA brains. Interestingly, α-synuclein may be observed in oligodendrocytes of normal brains [17], thus, the α-synuclein pro-aggregatory role of TPPP/p25 may lead to widespread accumulation in the form of cytoplasmic inclusions. Although neuronal expression of TPPP/p25 is considered as less physiological or at least lower than in oligodendroglia, an aggregation enhancing effect of TPPP/p25 may lead to the appearance of neuronal cytoplasmic inclusions in MSA and Lewy-body disorders.

It can be speculated that the segment between aa 184–200 of TPPP/p25 is more exposed in protein aggregates. The distinct immunolabelling patterns observed by anti-TPPP/P25-A and anti-TPPP/p25-B suggest different conformational structures for this protein within oligodendroglial or neuronal cytoplasm (soluble or attached to microtubular network), and within protein aggregates in the inclusions. Analogously, processing of α-synuclein in glial cytoplasmic inclusions and Lewy-bodies is different [3]. This is based on the observation that in MSA brains SDS-insoluble α-synuclein is not found in contrast to an elevated level of SDS-soluble and even TBS-soluble fraction of α-synuclein [3]. SDS-insoluble α-synuclein depicts less aggregated/membrane-associated forms of α-synuclein. Thus, our TPPP/p25-A antibody more likely associates both with the SDS-soluble and insoluble α-synuclein, contrasting our TPPP/p25-B antibody that detects predominantly epitopes of SDS or TBS soluble α-synuclein-associated forms of TPPP/p25. Nonetheless, this merits further immunohistochemical and biochemical studies using monoclonal antibodies against different epitopes of TPPP/p25. However, there is an additional issue that is also related to the distinct feature of the two antisera. The segment aa 184–200 of TPPP/p25 is localized at the C-terminal part of the protein that includes partly the predicted tubulin binding domain (J. Ovádi, unpublished data). Theoretically, this segment may be masked by microtubules or other interacting components under physiological conditions. During the pathological process that involves alterations in the microtubular network, this epitope might be exposed. An alternative explanation could be that the polyclonal antibody raised against the whole protein does not recognize the segment aa 184–200 of TPPP/p25. Interestingly, we found co-localization of TPPP/p25-A and huntingtin in some extracellular globular structures in HD cases (Fig. 2e). α-Synuclein immunoreactivity of huntingtin polyglutamine aggregates in HD was reported using another anti-α-synuclein antibody [4]. The anti-α-synuclein antibodies used in our study did not show similar immunoreactivity. This emphasizes the importance of optimized and comparable laboratory methodology. In addition, it suggests that TPPP/p25 may be either recruited to huntingtin aggregates together with α-synuclein, or theoretically may enhance extranuclear aggregation of huntingtin. While there is a lack of immunodetection of TPPP/p25 in normal astrocytes, the detection of granular TPPP/p25 immunoreactivity in reactive astrocytes may be indicative of a low level of expression in these cells related to the microtubule system that degenerates and accumulates, or less likely astrocytes might phagocytose breakdown products of other cells, e.g. oligodendroglia.

The detection of TPPP/p25 immunoreactivity in cells bearing lipofuscin and granulovacuolar degeneration may be due to the degeneration or breakdown of the microtubule system of the neuronal cytoskeleton, which may allow access of antibodies to epitopes. This suggests that neurons also possibly harbour low levels of TPPP/p25, although not detected with these immunohistochemical procedures. Since α-synuclein may be cross-linked to substantia nigra neuromelanin in PD [7] and subcellular proteomics revealed TPPP/p25 as a component of neuromelanin granules [27], it is not surprising to find TPPP/p25 immunoreactivity in neuromelanin granules of substantia nigra neurons.

In summary, we demonstrate that TPPP/p25 is consistently associated with pathological α-synuclein profiles. We identify aa 184–200 as a more exposed segment of TPPP/p25 in pathological protein aggregates related predominantly to altered α-synuclein or less frequently huntingtin, or in some circumstances associated with early stages of tangle formation composed of hyperphosphorylated tau [14]. This suggests a possible role of TPPP/p25 in protein aggregation during some neurodegenerative processes.