Discussion
Our previous studies showed a striking acceleration of synucleinopathy in a transgenic mouse model (M83) expressing the A53T mutated human αS protein, when mice were intra-cerebrally inoculated with brain extracts from old, sick M83 mice [
14]. We have now confirmed these initial observations in experiments using hemi-brain extracts from these mice that showed accelerated disease after experimental challenge, thus representing a “second passage” of the disease, as is routinely performed to propagate prion strains in experimental models. In these experiments, similar incubation periods of the disease were observed after stereotactic inoculation of the brain extracts into either the cerebellum or hippocampus, although this latter brain region is acknowledged to be largely spared during normal aging in this mouse line [
4]. All but one mouse inoculated with hemi-brain extracts in the cerebellum developed the disease to a terminal stage well before the age of 8 months, i.e. the age of disease onset in uninoculated M83 mice [
10]. Survival, in these second passage experiments, was significantly longer when an extract prepared from the cerebral cortex from a sick mouse was used, with 5/9 mice surviving more than 8 months, although these mice developed the disease more rapidly than uninoculated M83 mice, which is consistent with the much lower levels of αS
D in this brain region (Additional file
1: Figure S1A). In the cerebral cortex, αS
D is however clearly detected in our study by immunohistochemistry, although to a lesser extent compared to more caudal regions or the spinal cord, consistent with other similar studies [
4]. In the second passage experiment performed by non stereotactic inoculation (experiment 4), survival was significantly shorter with 1% whole brain homogenate (from an inoculated mouse) compared to that previously observed at first passage (from an old uninoculated mouse) (experiment 1) [
5], although insoluble pSer129 αS
D levels in the inocula were comparable (Additional file
1: Figure S1A); in this experiment, a 20 μl volume at 1% of inoculum was inoculated instead of 2 μl at 10% in the stereotactic experiments which showed a similar survival to that observed at first passage. The precise mechanisms and molecular species that contribute to the acceleration of the disease in such experiments are far from being understood yet. A recent study showed induction of αS pathology by intra-cerebral inoculation of a non amyloidogenic form of recombinants αS, raising the question of the contribution of neuroinflammation in such experiments [
18]. In addition, the presence of αS oligomers has been described in all examined brain regions of M83 mice, sharing similar basic biochemical properties but showing subtle conformational differences, those found in inclusion-bearing brain regions significantly accelerating αS aggregation
in vitro and causing primary cortical neuron degeneration [
19]. The contribution of these oligomers to the development of neuronal dysfunction appears to be independent of their absolute quantities and basic biochemical properties but is dictated by the composition and conformation of the intermediates as well as unrecognized brain-region-specific intrinsic factors. Such molecular species could differ between first and second passage experiments, as brain lesions appeared more important after inoculation than during normal aging [
4].
We then used these experiments to obtain a detailed characterization of αS aggregation in the brain. Western blot analyses specifically revealed pSer129 αS in the insoluble fractions prepared by ultracentrifugation in the presence of sarkosyl, most abundantly in the mesencephalon, brain stem and spinal cord. This disease-associated αS protein (αS
D) was typically undetected or only barely detectable by this method in the more frontal brain regions, such as the olfactory bulb, cerebral cortex, hippocampus or striatum. The typical 4 band pattern, representing monomeric or oligomeric ubiquitinated or not αS forms [
4,
12,
15], was similarly recognized by clone 42 antibody and, to a much lesser extent, by LB509 antibody. Similar αS
D distribution was observed by immunohistochemistry using an antibody against pSer129 αS, which however also revealed specific immunolabeling of neurons in more frontal parts of the brain such as the cerebral cortex. In addition, when stereotaxically inoculated into the hippocampus, αS
D was also detected in this brain region, more intensely in the inoculated side of the brain than in the controlateral hippocampus. More intense αS
D labeling was also observed in the cerebellum of mice that had been stereotactically injected in this brain region. These data are consistent with previous reports that αS aggregation was increased and/or earlier around the sites of injection of brain homogenates from sick mice [
4,
12]. The robust detection of αS
D in caudal brain regions that do not share direct innervations with the injection site, similar to that observed during normal aging of M83 mice, further supports the hypothesis of αS
D trans-synaptic spreading as a possible mode of propagation of αS pathology [
4], this transfer being possibly more or less efficient depending on the molecular species (monomers, oligomers or fibril) [
20]. In a recent study involving unilateral inoculation of αS preformed fibrils (PFFs) in the hippocampus of transgenic human P301S mutant tau mice, tau inclusions were reported not only in all parts of the hippocampus, including regions that were more rostral and caudal to the injection site, but also in the contralateral hippocampus, and even in the locus coeruleus, a brainstem structure distant from the injection site [
21]. Whereas this tau aggregation was triggered to a greater extent by one of two aggregated forms of αS, both failed to induce substantial αS pathology, in contrast with previous studies in wild-type mice [
11]. Interestingly, some recent studies of the genesis of infectious prions from recombinant prion proteins (PrP) suggested a new mechanism, designated “deformed templating”, and postulated a required switch from the protein folding pattern of recombinant PrP fibrils to that of the disease-associated PrP to explain the long silent stage before any disease can occur [
22]. However, our study in the M83 transgenic mouse model overexpressing A53T mutated human αS, still confirmed an acceleration of the pathological process after intra-cerebral inoculation of fibrillar recombinant human A53T αS, similar to that observed following the inoculation of brain extracts from sick M83 mice. Recent findings obtained following the analysis of presymptomatic versus symptomatic M83 mice suggested that the formation of αS inclusions could be a relatively rapid and probably synchronized process. Indeed, in mice with sparse or moderate levels of αS inclusions, the distribution of aggregates within affected areas of the neuroaxis was diffuse and without clustering of inclusions, whereas none of the presymptomatic mice exhibited abundant αS inclusions [
17].
The disease in M83 mice was further characterized by setting up an ELISA test that has been shown to specifically detect αS
D directly from mouse brain homogenates. Indeed, ELISA analyses of brain homogenates, prepared from whole brains without any concentration step, readily enabled sick mice to be distinguished from healthy M83 mice, although some variability in αS
D levels was apparent between different mice from the same experimental group. In a second passage experiment with different inocula concentrations, mice inoculated with 10 mg of brain tissue equivalents had significantly lower αS
D levels than those inoculated with only 2 or 0.4 mg. The biological significance of this variability, if any, remains unexplained, but the ELISA results were tightly correlated with the Western blot analyses of insoluble pSer129 αS. Although such experiments should still be repeated with larger number of animals, it could suggest subtle differences in αS biochemical properties that could be more difficult to extract when higher amounts of aggregated αS have been inoculated. The possible variable effects of different quantities of aggregated αS inoculated on the features of αS aggregates ultimately found in the brain of recipients have been considered in a recent study, emphasizing that different lesions have been observed in different studies involving inoculations of different quantities of aggregated recombinant αS [
23]. Alternatively, considering that mice inoculated with the higher αS
D concentrations have also received higher amounts of non-αS brain components in our experiments, it should be noticed that it has been suggested that, beside “prion-like” protein self templating, neuroinflammation could be an important event in such experiments, thus possibly modifying the course of the disease according to the inoculation conditions [
18,
23], giving shorter survival despite lower αS
D levels in the brain of M83 mice. In agreement with Western blot analyses, immunoreactivity was specifically identified in the mesencephalon, brain stem and spinal cord by ELISA, but was not detected in the olfactory bulb, cerebral cortex or striatum. Similarly, it was not detected in the hippocampus, except in mice stereotaxically inoculated in this brain region. Although the analytical sensitivity of the ELISA test on brain homogenates appeared to be ~ 20× more sensitive than our previously described Western blot method on brain extracts prepared by ultracentrifugation in the presence of sarkozyl [
5], it remained lower at this stage than that of immunohistochemical detection, which also allows the detection of αS
D in individual cells in the frontal brain regions. This is consistent with our observation of accelerated disease following intra-cerebral inoculation of a brain extract prepared from the cerebral cortex from a sick M83 mouse, although these mice showed a longer survival than those inoculated with a spinal cord homogenate. These differences in sensitivity were also apparent in ELISA analyses of mice 8 or 12 weeks after their inoculation with brain extracts from sick M83, which were positive at 12 weeks, but negative at 8 weeks. In contrast, data previously obtained from similar experiments in this same M83 model using immunohistochemistry, showed that αS
D could be detected as early as 30 days after experimental challenge [
4]. All together, our data show clear ELISA differentiation between sick and healthy mice, and between brain regions differently affected by the pathological process, thereby demonstrating that the ELISA approach specifically recognizes disease-associated αS.
Importantly, similar results were obtained by this ELISA test, not only with a monoclonal antibody specifically recognizing pSer129 αS
D, but also with several other antibodies. It is however remarkable that we failed to detect any immunoreactivity by ELISA analysis of sick M83 mice using the clone 42 antibody against a central region of αS (91–96) [
24,
25]. This suggests that the recognized epitope could be cryptic under our ELISA conditions while it is exposed in samples denatured for Western blot detection. This agrees with our recent structural investigations which revealed that αS residues 91–93 or 91–97 are involved in beta-sheet structures within aggregated αS [
26]. Only very slight immunoreactivity was observed with an antibody against the N-terminal end (2–14) of the protein, Syn-514, that is reported to recognize conformational variants of αS enhanced by the presence of the double mutation E46K/A53T [
15,
27]. However much higher immunoreactivity was detected with several antibodies directed against the C-terminal part (115–140) of αS: these included LB509, produced against Lewy bodies (amino acid 115–122) and specifically recognizing the human protein, and 8A5 reported to be produced against 129–140 amino-acid and reacting with αS species in Lewy bodies [
28]. Interestingly, high immunoreactivity was also observed with the 4D6 antibody, that was recently described to specifically recognize the 124–134 region of the non phosphorylated form of αS at serine residue 129 [
29]. This indicates that non phosphorylated αS could also be significantly involved in the aggregation process in this model, even though recent studies have suggested that this phosphorylation occurs in the brain after Lewy bodies formation and more importantly at their periphery [
30,
31]. Neuronal inclusions not reacting with pSer129 antibodies by immunohistochemistry have already been reported with the M83 model [
15], as well as a dramatic increase of total αS in the brain between the ages of 2 and 6 months [
32]. In human PD patients, a recent study in which αS levels were assessed by ELISA in the plasma up to 20 years after the initial symptoms, showed a steady increase of total αS as the disease progressed, whereas the pSer129 αS levels remained relatively constant, suggesting that disease progression could be monitored by measuring non phosphorylated αS [
30].
Finally, we were able to use the ELISA approach to demonstrate immunoreactivity with an antibody specifically recognizing mouse αS (amino-acids 103–110) from brain homogenates that also reacted with the antibody against pSer 129 αS or with the LB509 antibody that only recognizes human αS. This immunoreactivity with the anti-mouse antibody, although much smaller than that observed with the two other antibodies, was quite consistent. This result is in line with previous observations in similar experiments involving M83 mice, where it was concluded that murine αS might be recruited during human αS aggregation in M83 mice [
4]. More recently, aggregation of murine αS was also reported following intra-cerebral inoculation in wild-type mice, not only of fibrillar recombinant mouse αS [
11], but also of fibrillar recombinant human αS or brain extracts from human patients with dementia with Lewy bodies [
12].
Authors’ contributions
DB participated in the design of the study, carried out the biochemical experiments, analysed the data, and drafted the paper. JV carried out and analyzed the biochemical experiments. SB carried out and analysed the stereotactic experiments and immunohistochemical analyses. EM performed the statistical analyses. DG participated in the stereotactic experiments and Latifa Lakhdar supervised the animal experiments. LB and RM produced and characterized the fibrillar recombinant protein. TB conceived the study, analyzed the data and drafted the paper. All authors read and approved the final manuscript.