Protein Coaggregation in Caribbean Atypical Parkinsonism: The Contribution of Annonacin

ABSTRACT

Aims

There is an unexpectedly high proportion of atypical forms of degenerative parkinsonism in the French Caribbean islands. Residents of these islands are thought to be susceptible to Caribbean atypical parkinsonism (CAP) owing to their consumption of Annonaceae plant products containing the mitochondrial toxin annonacin. Here, we aimed to better correlate the clinical diagnosis of CAP with the misfolded protein pathology observed in affected individuals and to further investigate how annonacin could contribute to CAP pathogenesis.

Methods

We conducted postmortem histopathological analysis of brain samples from eight patients; more specifically, we assessed the distribution and burden of α‐synuclein (αS) and tau pathologies. Additionally, we studied how annonacin influences αS and tau aggregation using biophysical assays, with the corresponding recombinant human proteins serving as substrates.

Results

CAP was associated with heterogeneous clinical and histopathological features. Tau/αS copathology with a predominance of either αS or tau aggregates was observed in the majority (5/8) of patients. Tau and αS aggregates were sometimes colocalised in the same brain regions or cells. In biophysical assays, we showed that annonacin leads to an increase in αS aggregation and the formation of αS fibrils that could cross‐seed tau aggregation.

Conclusions

Annonacin may contribute to degenerative CAP by modulating the production of tau and αS pathogenic protein assemblies.

Summary

• Caribbean atypical parkinsonism (CAP) is suspected to be linked to the consumption of Annonaceae plant products containing the mitochondrial toxin, annonacin.

• Histopathological analysis of the brain samples revealed a tau/α‐synuclein (αS) copathology in most CAP patients.

• Biophysical assays demonstrated that annonacin can promote αS aggregation and that newly formed αS fibrils can cross‐seed tau aggregation.

• Proaggregation effects of annonacin may represent a new pathogenic mechanism that contributes to CAP.

Caribbean atypical Parkinsonism (CAP) is believed to be associated with the consumption of Annonaceae plant products containing the mitochondrial toxin, annonacin. The analysis of brain samples from CAP patients revealed a co‐pathology involving tau and α‐synuclein (αS) in the majority of cases. Tau and αS aggregates sometimes co‐localised in the same brain regions or cells. Biophysical assays demonstrated that annonacin can promote αS aggregation and that newly formed αS fibrils can cross‐seed tau aggregation. Pro‐aggregation effects of annonacin may represent a new pathogenic mechanism contributing to CAP.

1. Introduction

There is an unexpectedly higher proportion of atypical forms of degenerative parkinsonism in the French Caribbean islands than in Europe or North America, likely reflecting exposure to environmental agents [1, 2, 3].

Caribbean atypical parkinsonism (CAP) patients develop levodopa‐resistant motor symptoms with dementia and, in approximately half of the patients, supranuclear gaze palsy [4, 5, 6, 7]. Based on the neuropathological description of a restricted number of CAP patients with supranuclear ophthalmoplegia, the disease was initially defined as a tauopathy [7]. However, a significant proportion of patients have manifestations suggestive of synucleinopathies, such as dysautonomia, REM‐sleep behaviour disorder (RBD) and hyposmia [5, 6, 8]. Furthermore, CAP patients can present with myoclonus, with electrophysiological and clinical characteristics that are very similar to those observed in patients with dementia with Lewy bodies (DLB) or Parkinson's disease (PD) with dementia [9]. This finding suggested that CAP might sometimes be associated with mixed tau and α‐synuclein (αS) pathology [6].

Fruits, juices and herbal teas from the Annonaceae plant family are widely consumed in the Caribbean for their taste and as traditional medicines. However, experimental and epidemiological evidence suggests that the consumption of Annonaceae plant products could worsen disease severity and cognitive deficits in patients with CAP and PD [1, 5, 10]. For instance, in vitro and in vivo studies have reported that Annonaceous acetogenins, especially annonacin, one of the most abundant acetogenins, can damage different populations of neurons in the brain, including dopaminergic neurons and cortical and subcortical neurons, by promoting mitochondrial dysfunction [1, 11, 12]. Annonacin can also induce the redistribution of abnormally phosphorylated tau from neurites to cell bodies in striatal neuronal cultures [13]. In line with this observation annonacin was reported to aggravate tau pathology in mutant R406W human tau transgenic mice and oral consumption of Annona muricata juice containing annonacin was found to trigger and enhance tau phosphorylation in both wild‐type mice and transgenic mice expressing the human microtubule‐associated protein tau [14, 15]. Conversely, annonacin did not promote the formation of αS pathology in cellular or animal models [12, 13].

We hypothesised that exposure to an environmental toxin, such as annonacin, might promote tau/αS copathology and influence clinical presentation in Caribbean patients with degenerative parkinsonism. To investigate this hypothesis, we further characterised brain samples from patients with Caribbean degenerative parkinsonism, focusing on the possible coexistence of tau and αS pathology, and we conducted biophysical assays to determine whether annonacin could have an impact on αS and tau aggregation.

2. Material and Methods

2.1. Clinical Information and Diagnosis

We included eight patients with atypical forms of Caribbean degenerative parkinsonism. The clinical data for seven of these patients have been previously published [5, 6, 7]. Qualitative data are presented as counts and frequencies, whereas quantitative data are expressed as median and interquartile range (1st quartile—3rd quartile). For three patients, partial neuropathological data were published 20 years ago [7] (Patient 5 = Case 1; Patient 7 = Case 2; Patient 8 = Case 3; Table 1). All had degenerative parkinsonism but did not fulfil the criteria for clinically probable PD [19]. All of them were consumers of Annonaceae plant products. In addition, we compared the αS distribution and characteristics of Patients 1–3 with six advanced non‐Caribbean PD cases obtained from the French National Brain Bank Network‐Neuro‐CEB (Table S1), originating from mainland France, who had no history of annonacin exposure and had diffuse neocortical Lewy pathology and similar extent of ad neuropathological change.

TABLE 1.

Clinical and neuropathological characteristics of patients presenting with Caribbean atypical parkinsonism.

General features				Initial symptoms	Clinical features					Immunohistochemistry			Neuropath diagnosis
	Age at death	Sex	Duration (years)		Parkinsonian features	Supranuclear palsy	Cognitive impairment	Hallucinations	Others	αS Lewy pathology	Tau	Aβ Thal phase
P1&	80	m	8	Bradykinesia	B, R, T	No	Yes	V	RBD, OH CA, E	Neocortical	AD‐type c (Braak II)	1	LP, neocortical ad II/1; AA, 2
P2#	81	m	10	Bradykinesia	B, R, T a	No	No b		F, RBD	Neocortical	AD‐type c (Braak III)	2	LP, neocortical ad III/2; AA,1
P3#	72	m	18	Right tremor	B, R, T	No	Yes	V‐A	RBD, M, LD, BBK	Neocortical	AD‐type c (Braak IV)	3	LP, neocortical ad IV/3; AA, 1
P4#	59	m	4	Right‐hand tremor	Axial, B, R, T	Yes	Yes	V‐A	F, RBD, E, M, EOD	Brainstem (rare, dmX, retic form, LC)	PSP‐type d	0	PSP LP, brainstem
P5£	57	m	4	Bradykinesia	Axial, B, R	Yes	Yes	V	EOD	Brainstem (rare, dmX)	CBD‐type e	0	CBD f LP, brainstem
P6	66	m	2	Bilateral hand tremor	B, R, T	No	Yes		F, OH, CA	No	PSP‐type d	0	PSP
P7£	68	m	2	Postural instability, MCI	B, R	Yes	Yes	Occasional	BBK	No	PSP‐type d	0	PSP
P8£	71	f	3.5	Falls, MCI	Axial, B, R	Yes	Yes	V‐A	M, LD, EOD, BBK	No	PSP‐type d	0	PSP

Note: Patients from published cohorts (^&[5]; ^#[6]; ^£[7]; Patient 5 = Case 1; Patient 7 = Case 2; Patient 8 = Case 3).

AA: amyloid angiopathy, Type 1 (capillary AA) or 2 (no capillary AA).

ad Braak/Thal [16, 17]: Alzheimer disease Braak stage/Thal phase. Braak stages: Stages I and II, transentorhinal; Stages III and IV, limbic; and Stages V and VI, isocortical. Thal phases: 0, no amyloid; 1, isocortical; 2, allocortical; 3, basal ganglia; 4, brainstem; 5, cerebellum [17].

LP: Lewy pathology; 3 categories: brainstem/limbic/neocortical [18].

Abbreviations: αS: α synuclein; A: auditory hallucinations; ad: Alzheimer disease; Aβ: Aβ protein; B: bradykinesia; BBK: Babinski sign; CA: cerebellar ataxia; CAP: Caribbean atypical parkinsonism; CBD: corticobasal degeneration; dmX: dorsal motor nucleus of the vagal nerve; E: epileptic seizures; EOD: eyelid opening apraxia and blepharospasm; F: early falls (first 2 years); f: female; LBD: Lewy body disease; LC: locus cœruleus; LD: limb dystonia; m: male; M: myoclonus; MCI: Mild Cognitive Impairment; OH: orthostatic hypotension; P: patient; PSP: progressive supranuclear palsy; R: rigidity; RBD: REM‐sleep behaviour disorder; Retic form: Reticular formation; T: tremor; V: visual hallucinations.

^a Low response to L‐DOPA (22%; acute test 300‐mg L‐DOPA) and subjective improvement with chronic treatment lower than 30%.

^b MMSE 27/30; FAB 16/18.

^c AD‐type: Alzheimer disease–type tau misfolded protein aggregates are observed only in neuronal soma and processes (neurofibrillary tangles, neuropil threads, dystrophic neurites).

^d PSP‐type: progressive supranuclear palsy–type tau misfolded protein aggregates involve neurons (globose tangles and diffuse granular cytoplasmic immunoreactivity) and glial cells (astrocytes: tufted astrocytes, oligodendrocytes: coiled bodies).

^e CBD‐type: corticobasal degeneration–type tau misfolded protein aggregates involve neurons (globose tangles and diffuse granular cytoplasmic immunoreactivity) and glial cells (astrocytes: astrocytic plaques; oligodendrocytes: coiled bodies).

^f Numerous astrocytic plaques in the affected neocortex; abundant neuropil tau pathology involving the white matter of the centrum semiovale, the internal capsule and other white matter tracts with numerous oligodendrocytic coiled bodies. Neuronal tau inclusions were also abundant in the cortex, striatum, and brainstem and were found as either intracytoplasmic diffuse granular deposits or globose neurofibrillary tangles.

2.2. Brain Samples and Neuropathology

Formalin‐fixed paraffin‐embedded tissue sections of representative brain areas were stained with haematoxylin and eosin. In addition, immunohistochemistry with mouse monoclonal anti‐αS (clone 5G4; Millipore, 1/4000), anti‐phospho‐tau (pS202, pT205) (clone AT8; Thermo Fisher, 1/500), anti‐tau‐RD3 (3‐repeat isoforms, clone 8E6/C11; Millipore, 1/100), anti‐tau‐RD4 (4‐repeat isoforms, clone 1E1/A6; Millipore, 1/50), anti‐Aβ (clone 6F/3D; Agilent, 1/200) and rabbit polyclonal anti‐TDP‐43 (Protein Tech, 1/2000) antibodies was performed on selected brain regions with a Ventana BenchMark stainer (Roche). The antibodies were revealed with a diaminobenzidine (DAB)–based detection kit (Ventana Medical System Inc., Roche) before conducting neuropathological assessment on an Olympus CX41 bright‐field microscope. The distribution of tau and Aβ deposits was evaluated according to the Braak neurofibrillary tangle (NFT) staging and Thal phases, respectively [16, 17]. Lewy pathology was classified into three stages based on regional extent: brainstem predominant, limbic or neocortical [18]. Misfolded protein pathology burden was assessed on immunostained slides using a semiquantitative scoring scheme based on a five‐point scale: absent (0), rare (R), mild (1), moderate (2) or severe (3).

Double‐labelling immunofluorescence with an anti‐phospho‐tau antibody (clone AT8; Thermo Fisher; 1/200) and a phosphorylation‐dependent anti‐αS antibody (pSer129) (clone EP1536Y; rabbit monoclonal; Abcam ab51253, 1/200) was also performed. Primary antibodies were revealed with a secondary antibody coupled with Alexa Fluor 488 or 555, respectively. Slide imaging was carried out using a confocal Nikon AX R inverted microscope equipped with a 60× oil immersion objective (CFI Plan Apochromat Lambda D 60× Oil; NA 1.42; W. D 0.15) and a bright‐field detector (Nikon France). Specifically, slides were positioned upside‐down (i.e., with the glass coverslip facing down), and images were captured every 0.175 µm in the Z direction. Fluorescent images were obtained using Galvano scanner mode, and illustrations are presented as maximal Z‐projections.

2.3. Preparation of Monomeric Human αS

The purification of recombinant human αS and the analysis of its expression were performed as previously described [20, 21]. SDS–PAGE was used to assess the purity of the proteins. Monomeric αS (αS_m) stock solutions were prepared in 20‐mM HEPES and 150‐mM NaCl at pH 7.4. Before the measurements, the protein solutions were filtered and centrifuged at 12,000 × g for 30 min at 4°C. The protein concentration was determined by measuring the absorbance at 280 nm using the extinction coefficient ε275 = 5600 cm⁻¹·M⁻¹.

2.4. Preparation of Monomeric Full‐Length Tau (2N4R)

The purification of recombinant human tau and the analysis of its expression were performed as previously described by Medina and colleagues [22]. Briefly, fresh monomeric tau (tau_m) stock solutions were prepared in 40‐mM Tris–HCl and 20‐mM MgCl₂ (pH 7.4) supplemented with 0.05% dithiothreitol. The protein solutions were centrifuged at 12,000 × g for 40 min at 4°C and then filtered with a 0.22‐μm syringe filter (Miltenyi‐GV. Millipore). The protein concentration was determined by measuring the absorbance at 280 nm using the extinction coefficient ε = 7700 cm⁻¹·M⁻¹.

2.5. Protein Aggregation Assays

αS and tau aggregation assays were adapted from previous protocols [21, 22] using a total reaction volume of 500 μL. Briefly, the aggregation kinetics of αS were monitored in samples containing monomeric αS (70 μM) and thioflavin‐T (ThT; 10 μM), which served as a fluorescence reporter for amyloid fibril formation. αS was assembled at 37°C in the presence or absence of annonacin using constant orbital agitation at 600 rpm (Thermomixer Comfort, Eppendorf). The aggregation kinetics of tau were monitored by incubating 22 μM of monomeric 2N4R tau with 10‐μM ThT using the same conditions of orbital agitation as for αS. Tau aggregation was monitored in the presence or absence of a 50‐μL aliquot of αS preformed fibrils generated without (αS_f) or with (αS:A_f) 10‐μM annonacin, respectively, to reach a final concentration (monomeric equivalent) of 7‐μM preformed fibrils in the reaction mixture. We also evaluated the impact of 10‐μM annonacin on tau aggregation in the absence of αS_f. Finally, heparin (0.2 mg/mL) was used as a reference cofactor to seed tau aggregation in the presence or absence of 10‐μM annonacin. The excitation and recording of ThT emission were performed at 450 and 485 nm, respectively, using an ISS PC1 spectrofluorometer (Champaign, IL).

2.6. Acquisition of Transmission Electron Microscopy (TEM) Images

The protocol for obtaining TEM images was adapted from a previous study [21]. Samples (50 μL) of αS solution were adsorbed onto glow‐discharged 200 mesh Formvar carbon film‐coated copper grids (Electron Microscopy Sciences) and stained with UranyLess (Electron Microscopy Sciences). Excess liquid was removed, and the grids were allowed to air dry. Images were captured using a Hitachi 7700 TEM system (Hitachi, Tokyo, Japan).

2.7. Annonacin Purification

Annonacin was purified from the leaves and twigs of Annona montana as described previously [23].

3. Results

3.1. Clinical and Histopathological Characterisation

The clinical and neuropathological features of the patients are described in Table 1. Of the eight selected patients, seven were men, and one was a woman. The median age at death was 69.5 years (IQR 64.3–74.0). The median duration of the disease was 4 years (IQR 3.1–8.5). All bar one had dementia. Five of them had a mixed pathology with either αS (n = 3, the oldest in the series) or tau (n = 2) predominant pathology. Of the remaining patients, three had pure primary tauopathies.

Histological examination of the three patients (P1–P3) with αS predominant pathology also revealed ad‐type tau and Aβ neuropathological changes. Lewy bodies were observed throughout the brain, including the neocortex, which, together with the clinical presentation, was suggestive of DLB [18]. In these three patients, the neuropathological changes associated with ad were either mild (tau Braak Stage II or III; Thal Aβ Phase 1 or 2 in two patients) or intermediate, with tau Braak Stage IV and Thal Aβ Phase 3 in the third patient [24]. Comparing these three Caribbean patients (median age: 80 years; IQR: 76–80.5) with the six non‐Caribbean PD patients (median age: 85 years; IQR: 83.3–86), we found no notable difference in the aggregate appearance, burden or regional distribution of αS pathology (Tables S1 and S2).

Patients 4 and 5 had predominant non‐ad–type tau pathology with additional rare αS‐positive Lewy pathology. The αS pathology involved the dorsal motor nucleus (dmX) of the vagal nerve, the reticular formation and the locus cœruleus (LC) in Patient 4 but was limited to the dmX in Patient 5. Neuropathological analysis of Patient 4 revealed characteristic features of progressive supranuclear palsy (PSP) with a brainstem predominant involvement, including scant tufted astrocytes in the frontal cortex and striatum, as well as neuronal globose neurofibrillary tangles in the substantia nigra, subthalamic nucleus and globus pallidus [25]. This patient presented with early falls and supranuclear gaze palsy, along with clinical symptoms such as hallucinations, which are considered red flags that reduce diagnostic certainty for PSP, according to the MDS‐PSP criteria [26], as well as with myoclonus and RBD, which are relatively uncommon in PSP [27, 28]. Patient 5, also with predominant tau pathology, was clinically diagnosed with PSP‐Richardson syndrome (RS) [26], characterised by early postural instability and supranuclear oculomotor palsy, although the presence of visual hallucinations, which are atypical for PSP‐RS, raised diagnostic uncertainty [28]. Neuropathological findings, including diffuse granular neuronal tau‐positive cytoplasmic aggregates, astrocytic plaques in the neocortex and striatum, abundant thread‐like lesions in both grey and white matter, marked neuronal loss of the substantia nigra with few globose tangles and preservation of the globus pallidus and subthalamic nucleus, pointed towards a definitive diagnosis of corticobasal degeneration (CBD), according to neuropathologic criteria from Dickson and colleagues [29] (Table 1).

The final three patients who had pure neurodegenerative tauopathies and no αS pathology exhibited clinical heterogeneity that could be explained by the differences in the distribution of the tau pathology seen in these cases. P6 had a histological diagnosis of PSP with a predominant brainstem distribution. Numerous globose neurofibrillary tangles were observed in the pontine base, substantia nigra and medulla oblongata, along with abundant glial pathology, mainly composed of coiled bodies. Tufted astrocytes were scant in the neocortex and striatum. In contrast, P7 and P8 showed a more classic distribution, with numerous tufted astrocytes in the neocortex and striatum, prominent glial tau aggregates in globus pallidus and globose tangles in the subthalamic nucleus and brainstem, including the substantia nigra and pontine base. Patients 7 and 8 showed a phenotype resembling PSP‐RS but also experienced hallucinations (P7 and P8), which are considered red flags that reduce diagnostic certainty for PSP [26] and/or myoclonus (P8), a clinical manifestation that is rare in PSP [27]. Patient 6 presented with degenerative parkinsonism, dysautonomia and cerebellar ataxia, but without oculomotor dysfunction.

3.2. Localisation of Misfolded Protein Pathology Burden in Brain Samples

The localisation and pathological burden in the brain tissue samples are summarised in Tables 2 and S3. Among the five patients with mixed tau/αS copathology, the brain regions predominantly affected were the locus coeruleus and the raphe nuclei in the pons, and the reticular formation and the dmX, in the medulla oblongata (100%), that is, brain areas involved in oculomotor and axial symptoms, such as postural instability and falls. In the patients with ad‐type tau pathology (P1–P3), the basal forebrain cholinergic nucleus of Meynert, the hippocampus, the amygdala and the frontotemporal and entorhinal cortices, that is, neuronal circuits linked to memory processes, were also affected. The presence of tau and αS inclusions in the brains of CAP patients is illustrated in Figure 1. The morphology, distribution and intensity of αS pathology in these regions were not different from those observed in the six non‐Caribbean PD cases with similar ad neuropathological changes.

TABLE 2.

Localisation and burden of αS Lewy pathology, AD‐, PSP‐ and CBD‐type Tau pathology, and Aβ misfolded protein aggregates in various brain areas of seven cases of Caribbean atypical parkinsonism.

	Patient 1			Patient 2			Patient 3			Patient 4			Patient 5			Patient 6			Patient 8
	αS Lewy pathology	Tau AD‐type	Aβ	αS Lewy pathology	Tau AD‐type	Aβ	αS Lewy pathology	Tau AD‐type	Aβ	αS Lewy pathology	Tau PSP‐type	Aβ	αS Lewy pathology	Tau CBD‐type	Aβ	αS Lewy pathology	Tau PSP‐type	Aβ	αS Lewy pathology	Tau PSP‐type	Aβ
FC	1	0	1	2	0	1	2	1	1	0	1	0	0	3	0	0	1	0	0	3	0
TC	3	R	0	3	R	1	3	3	3	0	R	0	0	3	0	0	0	0	0	2	0
PC	1	0	0	1	0	1	1	0	1	0	0	0	0	3	0	0	R	0	0	2	0
OC	R	0	0	0	0	1	0	0	1	0	0	0	0	1	0	0	0	0	0	1	0
Amyg	0	1	0	3	1	—	3	3	—	0	1	—	0	3	—	0	R	—	0	2	—
Hippoc	3	1	0	3	2	1	0	3	3	0	2	0	0	3	0	0	R	0	0	1	0
Entorh	1	2	0	3	3	1	3	3	3	0	2	—	0	2	—	—	1	0	0	1	0
Meynert	3	R	—	3	1	—	3	3	—	0	2	—	0	3	—	0	1	—	0	3	—
Caudate	3	0	0	1	0	1	3	1	3	0	2	0	0	3	0	—	1	0	0	3	0
Pallidum	1	0	—	R	0	0	2	0	2	0	1	0	0	3	—	0	2	0	—	3	—
Thal + STN	2	0	—	1	R	0	2	1	1	0	3	—	0	3	—	0	2	—	—	3	—
SN	3	R	—	3	R	0	3	1	0	0	3	0	0	3	0	0	2	0	0	3	0
Pons:
LC	3	R	0	2	1	0	2	2	0	R	3	0	0	3	0	0	—	0	0	3	0
Pont tegm	2	R	0	1	0	0	1	R	0	0	3	0	0	3	0	0	2	0	0	3	0
Pont base	0	0	0	0	0	0	0	0	0	0	1	0	0	3	0	0	3	0	0	3	0
DRC	2	0	0	—	—	—	2	2	0	0	2	0	0	3	0	0	—	0	0	—	0
Med oblon:
dmX	—	—	—	3	1	—	3	—	—	R	1	—	R	2	—	0	R	0	0	3	—
Retic form	—	—	—	3	1	—	3	1	—	R	3	—	0	3	—	0	3	0	0	3	—
Inf olive	—	—	—	0	0	—	0	0	—	0	2	—	0	3	—	0	3	0	0	2	—
Cerebellum	—	0	0	—	0	0	—	0	0	—	2	0	—	2	0	—	2	0	—	2	0

Note: Burden of pathology. 0: absent, R: rare, 1: mild, 2: moderate, 3: severe, −: not available.

.

Data for Patient 7 were excluded from this table due to the unavailability of slides required for immunostaining.

Abbreviations: αS Lewy pathology: Lewy type synuclein pathological aggregates (Lewy bodies, Lewy neurites); AD‐type: Alzheimer disease–type tau misfolded protein aggregates (in neurons: neurofibrillary tangles, neuropil threads, neuritic plaques); Amyg: amygdala; Aβ: Aβ pathological aggregates (plaques); Caudate: caudate nucleus; CBD‐type: corticobasal degeneration–type tau misfolded protein aggregates in neurons (diffuse granular cytoplasmic, globose tangles) and glia (astrocytic plaques, coiled bodies); Cerebellum: cerebellum; dmX: dorsal motor nucleus of the vagal nerve; DRC: dorsal raphe nucleus; Entorh: entorhinal cortex; FC: frontal cortex; Hippoc: hippocampus; Inf olive: inferior olive; LC: locus coeruleus; Med oblon: medulla oblongata; Meynert: nucleus basalis of Meynert; OC: occipital cortex; Pallidum: nucleus pallidus; PC: parietal cortex; Pont base: pontine base; Pont tegm: Pontine tegmentum; PSP‐type: progressive supraneuronal palsy–type tau misfolded protein aggregates in neurons (globose tangles, diffuse granular intracytoplamic) and glia (tufted astrocytes, coiled bodies); Retic form: Reticular formation; SN: substantia nigra; TAU: tau pathological aggregates; TC: temporal cortex; Thal + STN: thalamus + subthalamic nucleus.

FIGURE 1.

The majority of Caribbean atypical parkinsonism patients have a mixed p‐tau/αS pathology. Tau and αS pathology was observed in five out of eight patients in different regions of the brain through the use of anti‐tau (clone AT8) (p‐tau) and anti‐αS (clone 5G4) antibodies that detect hyperphosphorylated (Ser202, Thr205) tau and aggregated αS, respectively. In some of these regions, concomitant αS and p‐tau pathological aggregates were observed. P1 showed marked pathological αS and p‐tau aggregates in the CA2 region of the hippocampus (a, b). αS and p‐tau aggregation was moderate in P2 in the same region as in P1, and a few intraneuronal neurofibrillary tangles were seen (inset) (hippocampus; CA2) (c, d). In P3, there was a moderate amount of pathological αS and tau aggregates in the reticular formation of the medulla oblongata (e, f). In contrast, in P4, there were rare Lewy neurites (g: arrow and arrowhead [inset]) in the dorsal nucleus of the vagus nerve (dmX) in the medulla oblongata associated with a low amount of p‐tau pathological aggregates (h). P5 (i, j) was primarily a tauopathy in which we observed a moderate number of p‐tau deposits in the dmX within the medulla oblongata, with only rare αS deposits (i, arrow and arrowhead [inset]). In P6, only p‐tau aggregates were observed in the reticular formation of the medulla oblongata, whereas αS immunostaining was negative (k, l). Finally, in P8, a marked amount of pathological p‐tau was detected in the locus coeruleus, with no αS aggregates (m, n). Scale bars: 50 μm.

We further investigated whether in CAP patients, the two misfolded proteins were colocalised at the cellular level. Among the patients with the highest burden of αS‐positive Lewy pathology (P1, P2 and P3), we analysed the regions where tau and αS were most abundant (P1 and P2: hippocampus; P3: nucleus basalis of Meynert). We observed a colocalisation of these two proteins in neurites and, to a lesser degree, in neuronal somas within the CA2 region of the hippocampus in P1 and P2 and the nucleus basalis of Meynert in P3 (Figure 2). Two of the three patients without mixed pathology (P7 and P8) had extensive PSP‐type tau pathology throughout the brain after 3 years of disease evolution. The third patient (P6) showed moderate to severe PSP‐type tau pathology, predominantly located in the brainstem and, to a lesser degree, in the basal ganglia. Concomitant TDP‐43 pathology was also observed in the substantia nigra of P5. In this case, TDP‐43 deposition may explain why the PSP‐RS clinical presentation was associated with underlying CBD pathology [30]. No TDP‐43 pathology was detected in the limbic system or in any other brain regions in the remaining patients.

FIGURE 2.

Codetection of p‐tau and p‐αS aggregates in the brains of CAP patients. Coexistence of p‐αS and p‐tau aggregates in a large pyramidal cell soma from the CA2 hippocampal region in Patient 1 (P1). p‐αS and p‐tau aggregates are present concomitantly in neuropil (a, first column). Coexistence of p‐αS and p‐tau aggregates in cell somas but less often in neuritic processes in the CA2 hippocampal region of Patient 2 (P2) (a, second column). p‐tau aggregates are comparatively more abundant in patient P2 than in P1. p‐αS and p‐tau aggregates coexist in cell somas but not in neurites in the nucleus basalis of Meynert from Patient 3 (P3) (a, third column). Note the presence of small p‐tau puncta in the cell body with a large αS deposit. A higher magnification of selected areas is presented above (b). Large and thin arrows point to cell bodies and neuropil containing both p‐αS and p‐tau aggregates, respectively in (a, b). Arrowheads point to cells comprising a single type of aggregate. Scale bar: 20 µm (a) and 10 µm (b).

3.3. αS Aggregation Is Increased by the Plant Toxin Annonacin

In vitro, we first tested the ability of annonacin to exacerbate αS aggregation using a ThT fluorescence assay. We found that annonacin at a concentration of 10 μM (but not at lower concentrations) led to an approximately threefold increase in αS aggregation, as indicated by the increase in the fluorescent signal of the reporter at the endpoint of the assay (72 h) (Figure 3a). When the aggregation kinetics of αS were monitored in the presence or absence of 10 μM annonacin (Figure 3b), we observed that the acetogenin did not affect the lag phase for nucleation, suggesting that it enhanced the elongation phase of amyloid fibril formation. Note that annonacin did not increase the intensity of the ThT fluorescent signal in the absence of αS_m, suggesting that the increase in the intensity of the ThT signal truly reflects αS amyloid aggregation. Turbidity measurements of the αS samples confirmed that the elongation process improved in the presence of 10 μM annonacin, whereas the lag phase remained unchanged (not shown).

FIGURE 3.

The impact of the mitochondrial Complex I inhibitor annonacin on αS and tau aggregation. αS aggregation measured by ThT fluorescence after 72 h of incubation of αS_m (70 μM) with (1–10 μM) or without annonacin (−). Data in % of maximal ThT fluorescence levels are presented as mean values ± SEM (n = 3). ***p < 0.001 versus αS. One‐way ANOVA followed by Dunnett's test (a). Fibrillization kinetics (0–72 h) of αS measured by ThT fluorescence in the presence or absence of αS_m (70 μM) mixed (red line and squares) or not mixed (black line and circles) with annonacin (a; 10 μM). Variations in ThT fluorescence levels are shown for samples containing annonacin only (green line and triangles). Data expressed in % of maximal ThT fluorescence at the endpoint stage are presented as mean values ± SEM (n = 3) (b). TEM images of αS samples incubated at 37°C for 72 h under continuous orbital agitation in the absence (upper panel) or presence (lower panel) of annonacin (10 μM). The right panel shows enlarged images depicted by the dotted line squares. Scale bars: 500 nm (left) and 50 nm (right) (c). Measurements of the length (in μm) and diameter (in nm) of fibrils formed in the presence (αS:A_f) or absence (αS_f) of annonacin. Data are presented as mean values ± SEM (n = 16). ***p < 0.001 versus αS. Student's t‐test (d, e). Tau fibrillization kinetics monitored by ThT fluorescence in conditions where tau_m (22 μM) is mixed with 7 μM (monomer equivalent) fibrillar species of αS formed in the presence (αS:A_f; red line and circles) or absence (αS_f; black line and squares) of annonacin (10 μM). Variations in ThT fluorescence levels are also shown when annonacin (10 μM) was added in the presence (inverted green triangle and line) or absence (blue triangle and line) of tau_m (f). Tau fibrillization kinetics monitored by ThT fluorescence in conditions where tau_m (22 μM) was mixed (red square and line) or not (green circle and line) with 0.2 mg/mL heparin to promote aggregation. ThT fluorescence levels are estimated comparatively under conditions where test samples contain tau_m with heparin and annonacin (black triangle and line) or heparin alone (inverted yellow triangle and line) (g). In (f) and (g), data expressed in % of maximal ThT fluorescence signal at the endpoint stage are presented as mean values ± SEM (n = 3).

TEM image analysis revealed that the noncanonical fibrils formed in the presence of 10 μM annonacin (αS:A_f) were morphologically different (Figure 3c). TEM image analysis revealed that the noncanonical fibrils were approximately three times longer than the canonical αS_f (Figure 3d). However, the diameter of these fibrils was not modified in the presence of the acetogenin annonacin (Figure 3e), suggesting that the same fibrillary amyloid structure was conserved.

3.4. αS Fibrils Formed in the Presence of Annonacin Have the Potential to Cross‐Seed Tau Aggregation

We subsequently investigated whether the newly formed αS:A_f could cross‐seed tau aggregation, a process that usually requires cofactors such as anionic lipids or polyanionic biomolecules like heparin (Figure 3f) [31, 32, 33]. We found that the αS:A_f were indeed capable of seeding tau aggregation in the absence of heparin (Figure 3f). In agreement with previous reports [34, 35], we found that αS_f generated without annonacin were also capable of seeding tau aggregation, although with much lower efficacy than αS:A_f. Notably, annonacin alone failed to seed the aggregation of tau_m (Figure 3f). The data presented in Figure 3g depict the aggregation kinetics of tau in the presence of the reference cofactor heparin (0.2 mg/mL). Annonacin did not further enhance the proaggregation effect of heparin.

4. Discussion

We found that five atypical degenerative CAP patients had a mixed tau/αS copathology, whereas three patients had pure degenerative PSP‐type tauopathies. Patients with a mixed misfolded protein pathology exhibited either predominant αS‐positive Lewy pathology with mild to moderate ad‐type tau pathology or primary PSP‐ or CBD‐type tauopathy with rare Lewy pathology in the brainstem. We demonstrated biochemically that exposure to the mitochondrial Complex I inhibitor annonacin via the consumption of Annonaceae plant products can aggravate the purified monomeric recombinant human αS aggregation process and indirectly seed purified monomeric recombinant human tau aggregation. Our results provide further evidence that Annonaceae plant products could contribute to CAP pathogenesis.

As with any study, the present work has limitations: (i) We examined brain samples from only eight patients, and (ii) the relevance of using synthetic protein strains and biophysical assays to model pathological aggregation in a human neurodegenerative disease is questionable. Nonetheless, our study may help to explain how Annonaceae plant products contribute to degenerative Caribbean parkinsonism.

CAP has been previously described as a tauopathy [7]. Here, we found that tau and αS pathology coexist in five of the eight patients with CAP. Brain tissue samples from three of them exhibited a predominant αS pathology associated with more discrete ad‐type tau pathology. Misfolded αS protein aggregates were found in the same regions affected by tau pathology, that is, the medial temporal lobe and brainstem. However, colocalisation in the same cellular structures was only observed in the CA2 region of the hippocampus and the nucleus basalis of Meynert. In two other patients, with a tau/αS copathology, tau abnormalities were prominent with rare αS deposits exclusively located in the brainstem. In these two cases, PSP‐ and CDB‐type tau pathology, characterised by mixed neuronal and glial involvement, with tufted astrocytes in the former and astrocytic plaques in the latter, was distributed throughout the brain. αS deposition was restricted to the dmX, reticular formation and/or locus coeruleus, without evidence of colocalisation within the same cell structures, likely due to the scarcity of αS aggregates in these regions. The remaining three patients were classified as having pure tauopathies of the PSP‐type without evidence of αS misfolded protein pathology.

In the first subgroup of patients (the three oldest individuals), all exhibited, in addition to αS pathology, low‐to‐moderate ad‐related neuropathological changes, such as NFTs and Aβ deposits, which are also frequently reported in Lewy body disorders, particularly in regions involved in cognitive processing [18]. However, their clinical presentation was uncommon for DLB, as none of them exhibited fluctuating cognition or variations in attention or alertness. In addition, two of the three patients presented with atypical features for DLB, that is, cerebellar ataxia in one (P1) and the absence of dementia and hallucinations in the other (P2). ad‐related neuropathological changes, including NFTs and Aβ deposits, are also known to increase with ageing, even in cognitively unimpaired individuals [36, 37]. Therefore, in this specific subgroup, these changes may reflect either coincidental age‐related processes or represent an integral component of the underlying proteinopathy. The clinical presentation was heterogeneous among the five other patients with pure or predominant tauopathy.

In the three patients with a predominant αS‐positive Lewy pathology associated with more discrete tau lesions in different regions of the brain, colocalization of αS and tau was observed in the same neuronal soma or neuritic processes.

Several clinical signs that may commonly occur in certain synucleinopathies but are not exclusive to them were observed in CAP patients regardless of their neuropathological profile. Hallucinations were noted in 6/8 patients, and four of these patients had predominant or pure tauopathies. RBD was reported in 4/8 patients, one of whom had a predominant tauopathy. Myoclonus was observed in 3/8 patients, including two with predominant or pure tauopathies, and dysautonomia was present in 2/8 patients, one of whom had pure tauopathy. Unexpectedly, in patients with predominant tauopathy presenting with myoclonus and/or hallucinations, Lewy bodies (LBs) were absent from cortical networks commonly associated with these symptoms [9, 38, 39]. Misfolded protein pathology burden involved mainly the brainstem, with moderate to severe tau and/or αS deposits in the substantia nigra, pons and nucleus basalis of Meynert in six of the seven patients. Moderate to severe misfolded protein pathology was also observed in the medial temporal lobe and the neocortex in 4/7 patients. The involvement of cortical regions in CAP patients is reminiscent of what we have previously observed. Using multimodal MRI, we showed that, compared with PSP‐RS patients, the predominance of diffusion changes in the temporal and occipital lobes in CAP patients was linked to cognitive changes and hallucinations [40]. Overall, our results suggest that the occurrence of multiple atypical clinical features within the same patient, along with the heterogeneity of the observed clinical manifestations, may reflect a widespread distribution and high burden of misfolded protein pathology more than the nature of aggregates. Copathology might also contribute to disease heterogeneity and severity in CAP patients.

Using a biophysical approach, we demonstrated that a concentration of 10 µM annonacin enhances monomeric recombinant αS aggregation and modifies the morphology of αS fibrils. It is unlikely that this concentration is reached in the brain through the consumption of Annonaceae plant products [41]. Nonetheless, much lower concentrations of annonacin may still be sufficient to replicate this effect in a pathological context, as long as a ratio of 7:1 between αS monomers and annonacin is achieved in the brain, similar to what we observed in the thioflavin‐T assay. Annonacin, which is a fatty acid‐derived product, might exert this type of effect because αS possesses intrinsic lipid‐binding properties, which are important for its physiological function [42, 43, 44]. Therefore, annonacin might operate as a decoy lipid structure, promoting the aggregation of αS. Notably, experimental evidence demonstrates that short‐term or subchronic administration of annonacin through Alzet minipumps in rodents leads to neurodegenerative changes in the brain without significant formation of LB‐like aggregates [12, 14]. Thus, annonacin may enhance rather than induce the synucleinopathy process in CAP patients.

In contrast to αS, which can form amyloid‐type aggregates in vitro by simple agitation without the need for a cofactor, tau fibrillization necessitates the presence of anionic biomolecules such as heparin [31], RNA [45] or arachidonic acid [46]. We found that, unlike heparin, which was used as a reference cofactor in the present study, annonacin does not promote the aggregation of human recombinant 2N4R tau. This finding is surprising considering that annonacin shares a fatty acid‐like structure with the tau aggregation inducer arachidonic acid, but contrary to arachidonic acid, annonacin has a neutral characteristic [47]. This suggests that in pure tauopathy forms of CAP, annonacin might help to induce pathological tau phosphorylation, possibly through its ability to promote mitochondrial dysfunction [13, 14, 15], a process that could subsequently drive tau aggregation [48, 49]. It is worth noting that the likelihood of tau nucleation involving other aggregation‐prone proteins, such as αS and TDP‐43 [34, 35, 50], with annonacin acting as a cofactor for their aggregation, is virtually excluded in these pure tauopathy forms.

However, this scenario becomes plausible in CAP patients with prominent tau/αS copathology. Indeed, we demonstrated that the intrinsic capacity of αS_f to cross‐seed tau aggregation [34, 35] is strongly enhanced in the presence of noncanonical fibrils of αS formed in the presence of annonacin. Therefore, annonacin may contribute to the tau/αS copathology in CAP patients by two complementary mechanisms: (i) the enhancement of αS aggregation and (ii) the promotion of tau cross‐seeding. In this context, it can be assumed that brain annonacin concentrations would remain below the levels necessary to directly induce tau pathology. Overall, our findings support the view that regular consumption of Annonaceae plant products containing annonacin might contribute to the development of a tau/αS copathology in some CAP patients. Note, however, that previous reports [51, 52, 53] and our present observations indicate that a copathology involving tau and αS also exists in non‐Caribbean PD patients. Notably, these patients had no history of consuming Annonaceae plant products. This raises the question of whether they might have been exposed to another environmental factor contributing to their tau/αS copathology, similar to what may occur in CAP patients.

Together, our clinical and histopathological data demonstrate that the CAP consists of a heterogeneous group of patients in whom neurodegenerative parkinsonism is associated with either pure tau or mixed tau/αS pathologies. CAP characteristics might reveal exposure to an environmental modifier that leads to susceptibility to a copathology rather than a distinct clinicopathological entity. Our results further support the notion that the consumption of Annonaceae plant products is a risk factor for CAP [4, 5]. More broadly, we speculate that exposure to annonacin and the subsequent increase in αS aggregation and cross‐seeding of tau aggregation by newly formed αS fibrils might also influence the pathological and clinical course of any form of parkinsonism, including PD. This finding is consistent with previous observations showing the impact of Annonaceae consumption on the overall severity of any type of degenerative parkinsonism in the Caribbean, especially regarding cognitive deficits [10].

Author Contributions

F.G.L., S.B., P.P.M. and A.L. contributed to the study conception and design, the acquisition and interpretation of data and the drafting of the manuscript; F.G.L., S.B., C.L.A., B.S., L.M. and P.P.M. contributed to data collection; J.R.H. provided purified annonacin; C.D.S. and D.A. were in charge of confocal image acquisition and processing; E.R. and H.C. contributed to the drafting of the manuscript; E.R., C.L.A., R.C. and R.V.R. contributed to the interpretation of the data; A.L. provided general supervision of the research group and acquired funding. A.L. was involved in critically revising the manuscript together with S.B. and P.P.M. All the authors have read and approved the final manuscript.

Ethics Statement

The study was approved by the French Sud Ouest and Outre Mer III Ethics Committee (2011/86). Autopsies were authorised according to current French regulations. The use of samples for research was approved by the next of kin.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Table S1 Comparison of demographics and neuropathological characteristics of patients presenting with non‐Caribbean Parkinson’s disease (PD) (absence of annonacin exposure) and Caribbean atypical parkinsonism (CAP, P1–P3).

Table S2 Localisation and burden of α‐synuclein Lewy pathology, ad‐type tau, Aβ and TDP‐43 pathological aggregates in representative brain regions of non‐Caribbean Parkinson’s disease (PD) cases (absence of annonacin exposure).

Table S3 Localisation and burden of TDP‐43, 3R‐tau and 4R‐tau isoforms in representative brain regions of seven cases of Caribbean atypical parkinsonism (CAP).

González‐Lizárraga F., Boluda S., Hidalgo J., et al., “Protein Coaggregation in Caribbean Atypical Parkinsonism: The Contribution of Annonacin,” Neuropathology and Applied Neurobiology 51, no. 4 (2025): e70026, 10.1111/nan.70026.

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.