In this paper, we characterize two kindred cases, mother and son, that were affected by a classic form of AD and by an aggressive form of dementia, firstly classified as probable AD and subsequently as an AD/LBD form. Patients have been deeply characterized from a neuropathological, genetic, and transcriptomic point of view with the aim to highlight the possible reasons that triggered the son to develop such a different clinical dementia phenotype despite the common genetic background shared with the mother. Clinical diagnosis of both patients was confirmed by a thorough neuropathological examination. High AD pathology with widespread deposition of Aβ aggregates and diffuse TAU pathology were found in the mother’s brain, while a severe LTS was found in the son’s brain, accompanied by intermediate levels of Aβ aggregates and TAU pathology. Moreover, both mother and son show moderate SVD with cerebral amyloid angiopathy.
Concerning their genetics, previous investigations showed that the two cases carried the ApoE ε4 allele, which is known to increase the risk of developing AD of 3-4-fold [25, 26]. Indeed, both showed a severe cortical amyloid load with amyloid angiopathy [27]. However, through NGS we found no pathogenic variants in the causative genes known to be involved in AD and LBD. Through the hereditary hypotheses, we found common variants that may have triggered dementia in both individuals and variants present only in the son that may be responsible for the parkinsonian spectrum and the LTS. Among the common variants present in both cases, the only one classified as pathogenic by the ACMG guidelines was a disruptive in-frame insertion (c.12948_12950dupAAG) in the RYR2 gene, which codifies for the Ryanodine Receptor Type 2 (RyR-2), a Ca2 + channel expressed either in the cardiac muscle sarcoplasmic reticulum and in the endoplasmic reticulum (ER) of Purkinje cells in the cerebellum, of other neurons in the cerebral cortex, and the dentate gyrus of the HiC [28, 29]. Perturbed ER Ca2 + homeostasis is recognized as a central player in AD [30, 31] and, in axon terminals; indeed, ER Ca2 + release is involved in vesicle fusion and neurotransmitter release [32]. This evidence suggests that the mutation found in the two subjects may have altered the RyR-2 channel function in neurons, with consequences on the Ca2 + homeostasis during the aging trajectory of the brain, contributing to the synaptic loss and to the development of dementia. Among the “not in common” (son-specific) variants, we found a missense variant in the USP24 gene (PARK10), which encodes for a deubiquitinating enzyme, involved in protein turnover and degradation. It is well known that impairments of the ubiquitin proteasome system lead to a non-effective clearance of α-synuclein, with consequent accumulation throughout time and LBs formation [33]. The variant in the USP24 gene may have impaired the deubiquitinating function, playing a role in the LTS of the son [34]. Moreover, USP24 has been previously associated with late-onset PD [23, 35]. Although this association with PD remains to be elucidated our results suggest that this variant may have influenced the pathology and the clinical phenotype of the son.
Through RNA-seq of four selected brain areas (PL, BG, HiC and SN) of the two cases and the control case, we identified the most dysregulated brain areas of the two subjects, the SN in the son and the PL in the mother. Specifically, the SN of the son was found to be the brain area with the highest number of DEGs compared to all the other brain areas analyzed either in the subject and in the mother. Downregulated genes in this area were related to the synaptic signaling, mainly to the dopaminergic and the GABAergic synapses. Among the retrieved DEGs, TH, SLC6A3 and SLC18A2 were the genes with the largest differences in expression between the patient and the non-demented control, and all three resulted highly downregulated. TH codifies for the Tyrosine hydroxylase, which is involved in the conversion of tyrosine to dopamine in the dopaminergic neurons; SLC18A2 is involved in the loading of dopamine in the presynaptic vesicle for its release, and SLC6A3 codifies for the dopamine transporter (DAT) which is involved in the reuptake of dopamine in the presynaptic terminals. Downregulation of these three genes in the SN of the son may reflect the loss of dopaminergic neurons in the SN pars compacta that characterizes Parkinson disease showing that transcriptomic analysis strictly correlates with neuropathological examination and may become a useful supportive tool in flanking challenging diagnoses. On the opposite side, among the most upregulated pathways in the son’s SN, we found the “Synaptic vesicle cycle” and the “Ca2 + signaling” pathways. While both pathways have been already reported in association with PD[36–38], no studies linked these pathways to LBD yet. These data confirm that PD and LBD probably share the same pathogenetic pathways and represent two sides of the same coin. Transcriptomic analysis demonstrates a clear-cut difference between LDB in the son and AD in the mother in spite of the common clinical phenotype (dementia) and genetic background.
Unlikely the son, the transcriptome analysis of the mother revealed that the PL was the most dysregulated brain area, according with the neuropathology which showed abundance of TAU deposition precisely in the PL. The parietal lobe is gaining even more attention in the development of AD [39–41]. Particularly, among the dysregulated pathways, we found the “steroid biosynthesis” and the “glycerolipid metabolism” ones. Several studies established a close relationship between alteration in steroidogenesis and fatty acid biosynthesis and neurodegeneration [42–44]. Moreover, pathways related to inflammation have been found dysregulated as expected in a late-stage AD brain, when inflammation contributes to neurodegeneration [45, 46].
The other two brain areas, the HiC and the BG resulted less dysregulated compared to the SN and the PL. Particularly, mother and son resulted similarly dysregulated within the BG according to their pathology, which was quite similar in this area and characterized by moderate amyloid deposition. It is noteworthy that in both cases the BG were not affected by specific proteinopathies (TAU/synuclein) but only by diffuse amyloid plaques, with little impact on the transcriptome. Conversely, the HiC was early affected by the disease and was an area with severe degeneration in both subjects, with double proteinopathy (TAU/synuclein) in the son. Nonetheless, this area appeared slightly dysregulated. Our data let us hypothesize that the amount of transcriptional dysregulation may be related to the level of damage accumulated across time by the different brain areas. The HiC which was the first and the most damaged brain area in both mother and son, with the majority of dying or critically damaged cells, become less prone to respond to tissue damage through transcriptional dysregulation, reducing the transcriptional machinery to work at basal levels due to cell dysfunction or, possibly, with the aim to preserve at most cell survival. Brain areas with modest or late involvement of pathological changes, as in the case of the BG of the two cases, commence to overcome tissue damage, resulting in a low and similar level of dysregulation. Lastly, areas affected by an evolving active pathology result highly and specifically dysregulated, as in the case of the SN and the PL in this study. They are still transcriptionally active and result in a peculiar transcriptional dysregulation depending upon the underlying pathogenesis and influencing the type of pathology. This work shows that the alterations of the transcriptome might follow the pathology according to a temporal and topographical trajectory where the less affected areas and, paradoxically, the most affected ones appear to be the least dysregulated, while areas with greater dysregulation are those in which the pathology is more active.