1. Introduction
Anderson–Fabry Disease (AFD) is an inherited sphingolipidosis due to a deficit of the lysosomal enzyme alpha-galactosidase (a-GAL A), which is responsible of the hydrolysis of terminal, non-reducing α-D-galactose residues in α-D-galactosides, catalyzing many catabolic processes, including cleavage of glycoproteins, glycolipids, and polysaccharides. The a-GAL A is encoded by the
GLA gene [
1], located in the X chromosome at the position Xq22.1 [
2]. AFD is an X-linked disease, affecting 1 in 40,000 males, but, unlike other X-linked diseases, the women carrying pathogenic
GLA variants may have the same systemic involvement as men and with the same severity. Another issue which makes it hard to precisely estimate the prevalence of AFD is that the degree of loss of a-GAL A function due to a heterogeneous effect of
GLA pathogenic variants may be responsible for two different clinical phenotypes: “classic” AFD and “late onset” AFD [
3]. The loss of a-GAL A function is responsible for the accumulation of the main substrate, globotriaosylceramide (Gb3), in the lysosomes. The main sites of Gb3 accumulation are the endothelial cells and the smooth muscle cells of the vessel’s walls, the heart muscle and endocardial cells, the glomerular nephrons and CNS cellular populations (neurons and glial cells), as well as the skin and the eye. The sites of accumulation predict systemic involvement in AFD, and several organs are directly assessable by biopsy for confirming and grading Gb3 accumulation (i.e., the heart and kidney). By contrast, CNS tissues do not undergo biopsy for such reasons, and the information about CNS involvement in AFD is mostly indirect and relies on clinical events (transient ischemic attacks, cerebral hemorrhages, thrombosis, and lacunar infarcts) and on neuroimaging findings. The CNS pathology in AFD has mostly been considered a secondary manifestation of endothelial dysfunction, as suggested by evidence of accumulation of the glycosphingolipid globotriaosylceramide (Gb3/GL-3) in endothelial cells as a consequence of alpha-GAL A enzyme deficiency. Smooth cells also have been demonstrated to contribute to the pathogenesis by releasing GL-3 and sphingosyne1 phosphate. However, neuropathological studies are still insufficient, limiting the accuracy in assessing the causal relationship between the neurological phenotype and pathogenic mechanisms linked to AFD. CNS involvement in AFD is not uncommon, being reported in 34% of patients within a multispecialty screening study [
3]. The prevalent involvement of the CNS refers to vascular damages of the parenchyma, especially the white matter, and the great vessels, e.g., vertebro-basilar dolichoectasia. Very little is known about pure neuronal and glial involvement in AFD. The therapies that have been available for a longer period of time, in particular enzyme replacement therapy (ERT), do not cross the blood–brain barrier, with an inevitable implication for the management of AFD-treated patients whose neurological involvement was not addressed. The more recently approved migalastat therapy has the potential to cross the blood–brain barrier, but its direct effects on the CNS have yet to be explored. Conversely, Parkinson disease (PD) is the most frequent neurodegenerative disease in the all-ages population [
4,
5], and its genetic determinants [
6,
7,
8,
9,
10] are currently increasingly expanding the definition of monogenic causes and the role of specific genetic variants as predisposing factors on which a “second hit”, either genetic or non-genetic, could act as a precipitating event for the development of the disease [
11,
12,
13,
14]. Pathogenic variants of the GBA gene, which encodes for the lysosomal hydrolase glucocerebrosidase (GCase) and represent the most common genetic risk factor for PD in the population, fall into the latter group. The role of lysosomal dysfunction, including specific enzymatic alterations such as that of GCase, has been implicated in PD pathogenesis even in the absence of the corresponding gene mutations [
15,
16,
17]. Recent reports suggested a possible association between AFD and PD. In particular, a higher prevalence of extrapyramidal signs and symptoms was observed in a subgroup of patients with AFD [
18]. This association deserves further study and has a biological and pathophysiological rationale that has not yet been fully explored.
2. Anderson–Fabry Disease, Brain and Neurodegeneration
As with many diseases affecting the CNS, the diagnostic gold standard for neurological involvement in AFD is neuropathology assessment. However, the availability of brain tissue for diagnostic purposes is critical, especially in diseases such as the AFD that do not require neuropathological analysis for diagnosis. There are isolated reports of non-autoptic brain tissue investigations in ADF patients, e.g., in patients who underwent brain biopsy or neurosurgical interventions for tumors (e.g., meningiomas). There are also very few published data on autopsy series [
19,
20,
21] of patients with AFD; such data were obtained mainly in a historical period prior to the introduction of ERT and without the aid of electron microscopy, which is usually used on biotic samples of other tissues (e.g., kidney, heart, and skin) for diagnostic purposes for AFD. Despite these limitations, the autopsy data provide us with some useful information to establish a solid basis for the pathophysiological hypotheses on the involvement of the CNS and its structures, from which the putative biological mechanisms of damage are better structured. One of the most interesting pieces of information coming from autoptic series [
19] is that the storage of glycosphingolipids is documented in muscle cells of large, intermediate, and small cerebral vessels and in neuronal tissue, with a distribution involving mainly components of the limbic system (basolateral nuclei of the amygdala and supraoptic and paraventricular nuclei of the hypothalamus) and the brainstem structures (substantia nigra, pontine reticular formation, dorsal efferent nucleus of vagus, salivary nuclei, nucleus ambiguous, and mesencephalic nucleus of the fifth cranial nerve) but also the spinal cord and peripheral nervous system.
Despite this documented involvement of the midbrain structures and in particular of the substantia nigra, an extrapyramidal clinical expression had never been described in patients with AFD until recently. The reasons for this are probably to be found not only in the multifaceted phenotype of AFD, but also in the substantial modification of natural history after the introduction of ERT. Before the introduction of ERT, the involvement of heart and kidney was the main factor influencing the reduced life expectancy of AFD patients with the classic form, and CNS involvement was partially due to systemic damage. Conversely, after the introduction of ERT, the improved control of heart and kidney involvement and the increased life expectancy have reduced the burden of CNS damage secondary to systemic disease. On the other hand, the increasingly frequent identification of late onset variants made it easier to identify milder phenotypes, even on the neurological side.
In favor of the hypothesis of a primary neurological involvement in AFD, several studies based on longitudinal neuroimaging and neuropsychological assessments suggested a possible neurodegenerative pathogenesis not secondary to cerebrovascular disease. The 8-year follow-up study published by Lelieveld et al. [
22] and focusing on neuropsychiatric symptoms and brain structural alterations in 14 AFD patients showed that, even without cognitive changes during the follow-up, a progressive decrease of hippocampal volume was recorded. Considering the not significant change of the white matter hyperintensities (WMHs) burden in the same interval, the loss of hippocampal tissue might be interpreted as pure neuronal involvement and therefore as a neurodegenerative phenotype of AFD patients, as suggested by the same research group at the baseline [
23]. The lack of correlation between neuroimaging parameters and neuropsychiatric parameters is interpreted by the authors as successful compensation of the progressive hippocampal volume loss. The MRI finding of hippocampal atrophy is concordant with the already reported autoptic studies [
24,
25,
26] demonstrating Gb3 storage in selective cortical and brainstem areas, including the neurons and ganglion cells of the hippocampus, in particular the presubiculum and the parahippocampal gyrus. It is interesting that in the case described by Okeda et al. [
26] the dementia syndrome of the patients was probably due to extensive cerebrovascular involvement and not to the mild Gb3 storage in the presubiculum and parahippocampal gyrus. The Gb3 storage might cause functional damage to the cells through oxidative stress and energy metabolism compromise up to cell death. Moreover, the lack of correlation between WMH volume and hippocampal atrophy in AFD is another issue in favor of the hypothesis of pure neuronal involvement in AFD [
23]. It seems that the hippocampal involvement could be considered a hallmark of neuronal involvement in AFD but without overt clinical manifestations outside of depression.
Neuroimaging studies [
27] suggested that advanced techniques can help to define the pathophysiology of CNS involvement in FD outside of WMHs. In particular, the loss of brain tissue volume (brain atrophy, differentiated into white matter—WM—and grey matter—GM—atrophy) is visually assessed in conventional morphologic MRI sequences, but this qualitative assessment has obvious limitations in reproducibility and accuracy both in general and in AFD [
28]. Advanced MRI techniques allow a quantitative and accurate assessment of brain volumes, and thus of atrophy. In AFD, as in other diseases, this evaluation is particularly relevant in absence of significant WMHs or cerebrovascular disease. Voxel-based morphometry (VBM) analysis has been applied to assess regional differences in GM volume in AFD patients, mostly failing to show a significant difference in comparison with healthy controls [
29,
30]. Conversely, atrophy in specific brain regions has been demonstrated in AFD, in particular in the thalami bilaterally and in the hippocampus [
23,
31], both with manual segmentation [
23] and with VBM [
31], suggesting direct neuronal involvement. An interesting finding is the significant reduction of the whole intracranial volume in AFD vs. healthy controls [
31] with preserved fractional brain tissue volumes (GM, WM, and cerebrospinal fluid). This finding too was interpreted as the consequence of abnormal neural development.
Another advanced MRI technique has been applied to assess the microstructural WM changes, i.e., diffusion tensor imaging (DTI) [
32]. The first studies focusing on cerebral diffusivity showed an elevated total brain parenchymal average diffusion constant in AFD vs. controls, and also in normal-appearing white matter [
33], because of increased brain interstitial water content due to microvascular changes. A similar finding was found in frontal, parietal, and temporal normal-appearing WM by using ROI-based DTI analysis [
34] and in the periventricular regions and the posterior portion of the thalamus by using voxel-based DTI study [
35]. In the same line are the findings of tract-based spatial statistics (TBSS) analyses [
36] and a combined TBSS–functional MRI (fMRI) study [
37], in which in AFD patients without extensive WMH loads, extensive areas of reduced fractional anisotropy were identified in different supratentorial and infratentorial WM regions. All these data supported the presence of microvascular injury in the territory supplied by the long perforating arteries in an early stage, i.e. before WM lesions are detectable on conventional images.
A different approach is the use of fMRI to study the presence of possible functional changes in AFD patients. Gavazzi et al. [
38] performed a motor task fMRI experiment (repetitive flexion–extension of the last four fingers of the right hand), finding an increased activation of additional cortical regions (cingulated motor area, secondary motor area, and primary sensorimotor cortex). Moreover, a resting-state fMRI (RS-fMRI) study [
39] demonstrated an alteration of the corticostriatal pathway in AFD, with reduced functional connectivity (FC) between motor cortices and the caudate and lenticular nuclei, bilaterally, and between the left motor cortex and cerebellar areas. In the same direction, another study [
37] demonstrated increased FC between the main hubs of the default mode network (DMN) (which is involved in the integration and coordination of sensorimotor and cognitive goal-directed activities) and different brain areas.
Finally, a promising technique is the qMRI, allowing a quantitative assessment of brain tissue relaxometry parameters and magnetic susceptibility; only isolated studies have applied it in AFD, but in one [
39] of those papers, the focus was possible iron accumulation in the striatonigral pathway. Quantitative susceptibility mapping (QSM) may help to evaluate the pathologic tissue changes and, in particular, iron accumulation [
40] in the striatonigral system due to different neurodegenerative disorders, including PD [
41] and atypical Parkinsonism [
42].
The application of advanced imaging techniques to study brain involvement in AFD suggested the presence of prodromal signs of neurodegeneration in AFD, with peculiar involvement of the motor system [
18] in the form of a subclinical extrapyramidal phenotype [
30] with motor slowing and postural instability more prevalent than rigidity and tremor. The above-reported neuroimaging findings supported the hypothesis of a neurodegenerative phenotype of AFD different from and not related to cerebrovascular involvement, which is better known and studied in the literature. This neurodegenerative phenotype was hypothesized also in the form of extrapyramidal dysfunction in general and mainly PD, as well as for other lysosomal storage disorders and in particular heterozygous GBA mutations.
This entry is adapted from the peer-reviewed paper 10.3390/biomedicines10123132