Abstract
Abstract
INTRODUCTION
Parkinson’s disease (PD) is the second most common neurodegenerative disorder, affecting 2% of the population over the age of 65 years [1]. Approximately 90% of PD cases occur sporadic of unknown cause and only 10% are inherited [2, 3]. The characteristic motor deficit of PD is due to a progressive loss of dopaminergic neurons in the substantia nigra pars compacta within the midbrain. The progressive degeneration of these nigrostriatal projection neurons results in deficient dopamine modulation of corticostriatal circuitry, leading to reduced spontaneous locomotion [4, 5]. The molecular events in this neurodegenerative cascade are crucial for the development of diagnostics and preventive therapies, but are still incompletely understood. Affected nerve cells display typical proteinaceous cytoplasmic inclusions, in the form of so called Lewy bodies (LB) and Lewy neurites (LN) [6–8]. The main ubiquitinated component of LB/LN is alpha-synuclein (SNCA) [9]. In some autosomal dominantly inherited cases of PD, this aggregation process was explained by SNCA missense mutations such as A53T or by SNCA gene duplication/triplication [10, 11]. In sporadic PD cases, genetic variants at the chromosomal locus of SNCA contribute to the disease risk, but a deficient removal of protein aggregates also seems to play a role [12, 13].
The physiological function of SNCA remains to be elucidated, but the protein is clearly concentrated in presynapses in association with vesicles, and there is evidence of its involvement with SNARE-complex assembly-dependent neurotransmitter release [14–19]. The synuclein protein family (comprising the 3 isoforms alpha, beta and gamma) was reported to share sequence and functional homology with the heterodimerizing 14-3-3 protein family and with small heat-shock proteins [20, 21]. There are seven known mammalian 14-3-3 isoforms (beta/β/YWHAB, gamma/γ/YWHAG, epsilon/ɛ/YWHAE, eta/η/YWHAH, zeta/ζ/YWHAZ, tau/τ= theta/θ/YWHAQ, and the selectively in T-cells as well as epithelial cells expressed sigma/σ). The species initially designated as α and δ are the phosphorylated forms of β and ζ) [22]. Protein interaction between SNCA and 14-3-3 within a 54-83 kDa complex was reported to promote apoptosis, possibly due to a reduction of cytoplasmic 14-3-3 leading to deficient binding and inactivation of pro-apoptotic proteins like BAD and BAX [23, 24]. The same authors showed 14-3-3 to co-precipitate with SNCA in the substantia nigra of PD patients in a process that turns the cells more susceptible to apoptosis [23]. Independent groups confirmed the presence of 14-3-3 proteins in Lewy bodies [25, 26], and documented immunoreactivity at least for epsilon-, zeta- and theta- isoform-specific antibodies [27]. SNCA and 14-3-3 proteins also co-aggregate in the glial cytoplasmic inclusion bodies of multiple system atrophy (MSA) [28–30]. SNCA mutations was observed to influence the expression levels of diverse 14-3-3 isoforms [14, 32]. Conversely, the abundance of 14-3-3 isoforms modulated the aggregation and the toxicity of alpha-synuclein in neuronal cultures [33, 34]. As direct protein interactors of alpha-synuclein, the different 14-3-3 isoforms might be among the first molecules to be affected by alpha-synuclein changes and to become relevant molecular read-outs at earliest stages of disease, while they are also progressively affected during the aggregation process until latest disease stages, and finally are holding therapeutic potential. However, their longitudinal assessment in vivo throughout the lifespan of synucleinopathy models is lacking.
It also remains unclear, which 14-3-3 isoforms are crucial in the midbrain neurodegeneration process of PD, given that cell cultures may have different expression profiles and that antibodies commercialized as isoform-specific may cross-react. Therefore, the aim of the present study study was to document for the first time the expression of all 14-3-3 isoforms in vivo for midbrain at the mRNA and protein levels by quantitative RT-PCR (qPCR) and immunoblots, along with data on their solubility during the progressive SNCA aggregation, from early adult to late ages of a PD mouse model. We employed inbred FVB/N animals with transgenic overexpression of human A53T-SNCA in dopaminergic nigrostriatal neurons driven by the neuron specific prion protein promoter (the homozygous PrPmtA line) [35]. This transgenic PrPmtA mouse line develops midbrain Lewy pathology upon crossbreeding with Pink1-deleted mice to double-mutant animals around the age 14 months [36]. These PrPmtA single mutants are also exceptional among the genetic rodent models of PD, because a pathological response to a dopaminergic stimulus with involuntary movements and increased post-synaptic sensitivity was demonstrable in these mice for the first time [37, 38]. In their striatal global proteome profile by 2D-gel based mass spectrometry, altered 14-3-3 epsilon levels were the earliest change by age 6 months that accompanies the dopaminergic signaling deficit of these mice [31]. In their striatum and midbrain global transcriptome profiles by oligonucleotide microarrays, a progressive dysregulation from age 6 months to age 18 months was detected for the 14-3-3 isoform epsilon, eta and zeta [39]. These mRNA data in our transgenic PD model were somewhat at odds with a previous report about PD patient brains on the basis of immunohistochemistry with 14-3-3 antibodies commercialized as isoform-specific, which claimed LB aggregation for the isoforms epsilon, zeta, gamma and theta [27], and with previous transgenic mouse and human neuronal culture studies emphasizing a role of 14-3-3 theta, epsilon and gamma in SNCA aggregation toxicity [33, 34]. Our longitudinal study now documents significant changes of various 14-3-3 isoforms with downregulated mRNA levels and altered protein solubility already at age 3 months, suggesting that 14-3-3 turnover is changed well before the advent of behavior or electrophysiology phenotypes. In addition we show that the levels of 14-3-3 eta mRNA correlate best with the progressionof pathology.
MATERIAL AND METHODS
Animal housing, welfare and breeding
Transgenic mice expressing the human A53T-SNCA cDNA under the control of the murine neuron specific prion protein promoter (the homozygous PrPmtA line) [35] as well as wild-type (WT) animals with the corresponding inbred FVB/N background were bred and aged in parallel at the FELASA-certified Central Animal Facility (ZFE) of the Frankfurt Goethe University Medical School. Both strains were housed in the same room in Type II L cages (365 × 207 × 140 mm, floor area 530 cm²; IVC-based) under 12/12 h light/dark cycles with water and food available ad libitum. Both mouse lines (PrPmtA and FVB/N) were derived from hemizygous littermates several generations ago, bred to homozygosity and then maintained by breeding between homozygous mice.
All experiments were performed in accordance with the European Communities Council Directive of 24 November 1986 (86/609/EEC) and the National Institute of Health Guide for the Care and Use of Laboratory Animals. The study reports data from a total of 43 PrPmtA and 42 WT mice.
Tissue dissection and preparation for qPCR and Western Blot
Mice were sacrificed using cervical dislocation. The brain was removed and midbrain was dissected as previously described [39]. The dissected tissue was immediately frozen in liquid nitrogen and stored at–80 °C until molecular analysis were performed.
Transcript level analysis using quantitative real-time reverse-transcriptase PCR (qPCR)
Tissue homogenization was performed with a pellet pestle motor (Kontes) and total midbrain RNA was isolated using TRIZOL reagent (Invitrogen) followed by digestion with DNase (amplification grade I; Invitrogen) according to manufacturer‘s instructions. Spectrophotometric quantification was used to determine the RNA concentration, and one microgram total RNA was reverse transcribed with SuperScript III (Invitrogen) utilizing oligo(dt)20 and random primers. The original amount of candidate gene transcript was quantified relatively to the constitutively expressed housekeeping gene TATA box binding protein (Tbp) via qPCR in a StepOnePlus Real-time PCR system (Applied Biosystems). The following gene expression TaqMan assays (Applied Biosystems) were used: Ywhab (Mm01247811_m1); Ywhae (Mm00494246_m1); Ywhah (Mm00834297); Ywhaz (Mm01722325_m1); Ywhag (Mm03047428_m1); Ywhaq (Mm01231061_g1); Tbp (Mm00446973_m1). Expression changes were calculated and plotted using the 2–ΔΔCT method [40], Microsoft Excel and GraphPad Prism5.
Protein level analysis using immmunoblotting
The protein levels of different 14-3-3 isoforms were studied in the midbrain of 3- and 18-month-old transgenic PrPmtA and wild type mice. In order to analyze the progressive insolubility and aggregation by differential detergents, proteins were separated in RIPA soluble and SDS soluble fractions. Homogenization of midbrain tissue was performed in RIPA buffer (50 mM TRIS/HCl pH 6.8, 150 mM NaCl, 1 mM EDTA, 1 mM EGTA, 5% Igepal CA-630, 0.5% Sodium Deoxycholate, 0.1% SDS) supplemented with protease inhibitor (Roche) using a pellet pestle motor (Kontes). Samples were rotated at 4°C for 15 min, followed by a centrifugation step at 16000 g for 20 min at 4°C. The RIPA soluble fraction was preserved in a new tube. Thereupon, 2x SDS-lysis buffer (137 mM TRIS/HCl pH 6.8, 4% SDS, 20% Glycerol) was added, the pellet was sonicated (3 cycles, 10 sec, 30–40% ) and centrifuged at 16000 g for 10 min at room temperature. The supernatant containing SDS soluble proteins was collected. Protein concentration was assigned with the BC assay protein quantification kit (Interchim) according to manufacturer’s instructions. Twenty microgram of protein were lysed with 2x loading-buffer (250 mM TRIS/HCl pH 6.9, 20% Glycerol, 4% SDS, 10% β-Mercaptoethanol, 0.005% Bromophenol blue) and incubated for 2-3 min at 90°C. Denaturated samples were then separated by SDS-PAGE (sodium dodecyl sulfate polyacrylamide gel electrophoresis) and blotted on a nitrocellulose membrane (Whatman) by wet blotting, followed by blocking for 1 h at room temperature in PBS-T containing 5% skim milk powder. Immunodetection was performed by incubating the membrane with the primary polyclonal 14-3-3 beta/alpha antibody (dilution 1:500; Cell Signaling), polyclonal 14-3-3 epsilon antibody (dilution 1:100; Santa Cruz), the polyclonal 14-3-3 eta antibody (dilution 1:100; Cell Signaling), the monoclonal 14-3-3 zeta/delta antibody (dilution 1:500; Cell Signaling), and the monoclonal β-Actin antibody (dilution 1:10000; Sigma Aldrich) in blocking buffer, at 4°C over night. The secondary antibody (ECL™ Anti-Rabbit IgG HRP linked for 14-3-3 detection and ECL™ Anti-Mouse IgG HRP linked for β-Actin detection; dilution 1:10000; GE Healthcare) was incubated for 1 h at room temperature in PBS-T and antibodies were visualized utilizing Signal West Pico Stable Peroxid Solution and SuperSignal West Pico luminol/Enhancer Solution (ThermoScientific, Germany) (ECL solution). Densitometric analysis was performed using the ImageJ software (1.40 g; National Institute of Health) and protein level changes were quantified using Microsoft Excel and GraphPad Prism5.
Statistical analysis
Statistical analysis of qPCR and quantitative immunoblots were performed using unpaired t-tests two-tailed via the Prism 5 software (GraphPad, La Jolla, CA, USA). Data are presented as mean ± SEM. Significant differences were highlighted with asterisks ( * p < 0.05; ** p < 0.01; *** p < 0.001).
RESULTS
qPCR analysis of 14-3-3 isoforms reveal early reduced and later elevated levels
In order to circumvent cross-reactivity issues of isoform-specific antibodies, we used commercial qPCR assays for an independent evaluation of 14-3-3 isoform expression at the mRNA level. In order to depict expression alterations that occur throughout life, adult mice were studied at four different ages to survey the complete lifespan of these animals. In midbrain samples from PrPmtA and WT mice aged 3 months, 6 months, 13 months and 18 months, the 14-3-3 levels for the beta, epsilon, eta, zeta, gamma and theta transcripts were quantified. Almost no changes were observed for gamma and theta mRNA in consistency with previous microarray profiling data at ages 6–18 months (data shown in Supplementary Figure 1), and none for epsilon with this technique in contrast to previous microarray findings [39] (Fig. 1B). At the early age of 3 months, novel significant reductions of the midbrain 14-3-3 mRNA levels for beta, eta and zeta isoforms were observed (Fig. 1A, C, D). For the beta isoform, also a novel significant increase was documented at age 18 months (Fig. 1A). For the eta and zeta isoforms, significant increases were documented at two different ages between 6 and 18 months (Fig. 1C, D), in agreement with previous microarray data [39]. The increase was almost linear over time in the case of the eta mRNA levels, from 0.77-fold at 3 months, via 0.99-fold at 6 months, to 1.16-fold at 13 months, and finally 1.37-fold at 18 months (Fig. 1C). Comparing the mRNA expression pattern of the analyzed 14-3-3 isoforms, it is remarkable that three of the isoforms show similar temporal dynamics of expression change, progressing from abnormally low to significantly elevated levels.
Quantitative immunoblots show several changes at 3 months, but only a reduction of soluble beta at 18 months
To determine whether the mRNA changes are mirrored by analogous anomalies of protein level in soluble and relatively insoluble tissue fractions,quantitative immunoblots were performed in midbrain proteins extracted sequentially with a RIPA buffer and then solubilizing the pellet again with high SDS concentrations. The two extreme age groups at 3 and 18 months show significant effects that illustrate a consistent overall trend (3 months in Fig. 2 and Supplementary Figure 2, 18 months in Fig. 3 and Supplementary Figure 3), while the age groups at 6 and 13 months revealed no significant alterations (6 months in Supplementary Figure 4 and 13 months in Supplementary Figure 5).
At age 3 months, the expected significant reductions of 14-3-3 beta/alpha (Fig. 2A) as well as epsilon (Fig. 2B), contrasting with an unexpected elevation of zeta, were observed in the RIPA fraction. Significantly increased presence in the relatively insoluble SDS fraction was noted for 14-3-3 epsilon and zeta at age 3 months (Fig. 2B, D). No significant alterations were detected in all analyses for other 14-3-3 isoforms (Fig. 2C and Supplementary Figure 2).
At age 18 months, again a significant reduction for 14-3-3 beta/alpha was observed in the soluble RIPA fraction (Fig. 3A). While all other isoforms and all studies of SDS fractions failed to exhibit significantly altered levels at old age (Fig. 3A?D and Supplementay Figure 3), a statistical trend towards reduction was notable for 14-3-3 eta within the SDS fraction (Fig. 3B) as well as for 14-3-3 gamma within the RIPA fraction (Supplementary Figure 3).
Comparing the protein level pattern of the 14-3-3 isoforms under study, it is remarkable that diverse effects at early age are consistent with previous observations, but get mostly compensated at later ages.
DISCUSSION
Novel insights were derived from this attempt to exploit a monogenic mouse model of PD to define the temporal evolution over 18 months of 14-3-3 isoform changes in midbrain, rather than previous studies of acute effects in a human neuronal cell line [33, 34] or very late effects in autopsy brain of polygenic PD cases [27]. Thus, the above findings represent the first longitudinal study in mouse, aiming to define molecular markers of prodromal disease and of disease progression in midbrain. Our previously published oligonucleotide microarray data [39] and the present qPCR data very consistently documented 14-3-3 eta and zeta mRNA dysregulations, with the new qPCR analysis of age 3 months demonstrating an initial reduction for both isoforms, which is then followed by progressively increased amounts as detected by both techniques. Unexpectedly, the systematic screen of 14-3-3 isoforms by qPCR uncovered a similar patternalso for 14-3-3 beta. Considering that 14-3-3 eta was found to be associated with SNCA in human late-stage Parkinsonian brains [41], that it is preferentially sequestrated by alpha-synuclein oligomers, undergoes a loss-of-function and acts to protect cells from SNCA aggregation and toxicity [42], and that it was observed to exhibit almost linearly increasing mRNA levels in the aging midbrain according to this study, the 14-3-3 eta mRNA seems particularly promising as an indicator of the progression of SNCA-triggered pathology at least in our mouse model.
This directly proportional correlation of 14-3-3 eta mRNA levels with synucleinopathy progression contrasts with the dynamics of 14-3-3 protein levels, where several isoforms show altered amounts at early ages and may thus reflect the initial disease burden or risk, but are later returning to homeostasis with exception of a constant reduction of soluble 14-3-3 beta. Although the antibody data concur with the mRNA assays that beta and zeta levels are reduced at age 3 months, both approaches are in conflict regarding epsilon versus eta. Furthermore, it is interesting to note that 14-3-3 epsilon and zeta protein are increased in insolubility (SDS fraction) at age 3 months on the one hand, and that elevated amounts of 14-3-3 beta/eta/zeta mRNA are needed to maintain normal protein levels at ages 6–18 months on the other hand. These consistent observations may be interpreted as consequencesof an insidiously progressive insolubility aggregation of SNCA with diverse 14-3-3 isoforms. If aggregated fibrillar SNCA/14-3-3 complexes are present in midbrain at age 18 months, they might not be solubilized even by high SDS, would not migrate in PAGE at the size of monomeric 14-3-3 and will perhaps not expose their specific epitopes for antibody recognition. Moreover, any aggregation process in the midbrain projection neurons might undergo trafficking of the resulting fibrils via anterograde axonal transport to striatal presynapses. In any of these scenarios, the aggregation process itself would remain undetectable for the quantitative immunoblotting approach of midbrain tissue in Fig. 3, but could leave traces such as the initial relative insolubility in the SDS fraction and the progressive upregulation of mRNA expression as were indeed observed. Overall, the 14-3-3 protein levels and solubility hold less promise as markers of disease progression in our PD mouse model, which does not progress to late-stages with neuronal death during its lifespan.
It is justified to question whether subtle 1.5-fold changes of abundance and solubility can be biologically relevant and are plausible mediators of PD pathology. Gene dosage changes with 1.5-fold increase are usually lethal in the case of chromosomal trisomies, result in manifest Alzheimer’s disease by the age of 40 years in practically all cases of trisomy 21, and famously underlie the autosomal PARK4 variant of PD with manifestation around age 50 years [43, 44]. Thus, it must be assumed that sporadic multifactorial PD with manifestation beyond 65 years is due to even more subtle effects that exert cumulative toxicity across decades. Given that the mice under study are modelling PD pathogenesis through the 1.5-fold overexpression of A53T-SNCA in brain tissue, we believe that it would be unreasonable to expect higher fold-changes in the interactor molecules.
What would the biological consequences of 14-3-3 changes be? First described in 1967, the 14-3-3 family was characterized as activators of tyrosine hydroxylase (thus modulating dopamine homeostasis) as well as inhibitors of PKCs [45–48]. Since then, 14-3-3 proteins have been shown to interact with more than 100 proteins and shield those targets against dephosphorylation [49]. Thus, they modulate neuronal differentiation, synaptic plasticity and behavioral plasticity [50]. Indeed, in the PD mouse model under study, we have previously documented such alterations. Our PrPmtA single mutant animals do not exhibit visible midbrain SNCA aggregates during their lifespan, as detectable by light or electron microscopy [35], but they do show a nigrostriatal synaptic dysfunction with progressive accumulation of dopamine [51], progressively diminished release of dopamine [52], dysregulated levels of the synaptic SNARE-complex component CPLX1 [39], and strong increases in the firing frequency of dopaminergic nigrostriatal neurons [53]. The downstream consequences of the documented neurotransmission deficiency are (I) transcript level alterations of the dopamine inducible genes Atf2, Cb1, Freq, Homer1 and Pde7b that reflect a reduced postsynaptic response in striatal tissue, (II) reduced spontaneous locomotor activity and (III) impaired synaptic plasticity that can be rescued by a phosphodiesterase inhibitor drug [51, 54]. In response to a dopaminergic challenge with apomorphine, the mice exhibit increased post-synaptic sensitivity, diverse alterations in phosphorylation cascades, involuntary movements and delayed recovery of normal behavior. These phenotypes can be demonstrated in the PrPmtA mice by the age 6 months onward. They could all be triggered or modified by the 14-3-3 changes now documented. It is important to note that the initial diverse reductions in 14-3-3 mRNA and soluble protein at age 3 months precede the manifestation of these electrophysiological and behavioral changes. The relevance of the 14-3-3 family is also underlined by recent observations that they are implicated in the pathogenesis of PD caused by LRRK2, Parkin and PINK1 mutations [41, 55–58].
Different 14-3-3 isoforms have been observed to change in response to synucleinopathy, depending on the model under study and the technical approach. In global transcriptome analyses via Affymetrix microarrays of single dopaminergic neurons from the substantia nigra pars compacta from late-stage patients with idiopathic PD, significant downregulations were observed for 14-3-3 beta (–5.2 fold), theta (–3.0-fold) and zeta (–1.5-fold) [59]. Conversely, upregulations of 14-3-3 beta and gamma as the main hub factors were defined in a systems biology approach that used a global transcriptome studies of substantia nigra tissue in late-stage PD and a genome-wide association study of single nucleotide polymorphisms to explore the protein interactions that are central for pathogenesis [60]. Trying to integrate these reports and our data, a plausible scenario emerges where midbrain tissue overall exhibits 14-3-3 transcript upregulations over time, while the specific dopaminergic neurons preferentially affected by cell death in PD undergo massive reductions of 14-3-3 mRNAs at late-stage. Global expression profiling of the midbrain in another transgenic mouse model also showed an initial downregulation appearing at the age of 9 months, specifically of the 14-3-3 theta in midbrain tissue with overexpression of wildtype alpha-synuclein under control of the PDGF-beta promoter [61]. Isoform differences between diverse genetic PD models may simply be due to transgene expression in different neuron populations of the midbrain, or to differential effects of wildtype alpha-synuclein versus A53T-alpha-synuclein.
Thus, the role of 14-3-3 isoforms in the neurodegenerative process of PD seems to be very different from their established position as biomarkers of several rapidly progressive nervous disorders [22]. The demonstration of elevated 14-3-3 levels in the CSF of prion disease victims is a consequence of many neuron deaths spilling this abundant cytosolic protein family into the extracellular fluid, and is therefore a late phenomenon after neuron loss. In PD, the joint human and murine data demonstrate a much earlier role in pathogenesis, with a 14-3-3 presence in the LB of remaining neurons, a progressive 14-3-3 transcript upregulation in nervous tissue before the onset of neuron loss, and a relative insolubility of 14-3-3 protein with concomitant reductions in the soluble fraction. Thus, our study of the temporal dynamics of 14-3-3 changes in the midbrain of a PD mouse model is consistent with a relevance of the 14-3-3 family early on, for nigrostriatal signaling adaptations, for the insidious aggregation process, and hopefully forneuroprotection.
CONFLICT OF INTEREST STATEMENT
The authors have no conflict of interest to report.
Footnotes
ACKNOWLEDGMENTS
We are grateful to our technical assistants Birgitt Meseck-Selchow and Mekhman Azizov. Our thanks go also to the animal care team at the ZFE Frankfurt (in particular the veterinarians Dr. A. Theisen and Dr. C. Tandi, and the caretakers E. Daut and B. Janton). This work was EU-funded through ERAnet Neuron consortium REPARK (BMBF 01EW1012) and by Goethe University Medical School. Nadine Brehm would also like to thank the International Max Planck Research School (IMPRS) for Neural Circuits in Frankfurt for a doctoral scholarship.