In this study, we provide in vivo evidence of dopaminergic alterations occurring in A + T + N + AD cases and along the disease phases. The dopaminergic deficits were prominent in basal ganglia since the early stages, whereas limbic structures (i.e., amygdala and hippocampus) and cortical regions (i.e., lateral and ventral frontal cortex, anterior and middle cingulate cortex) showed a significant alteration only in the full-dementia phase. This result is also supported by the progressive disruption of the pattern of connectivity obtained using the caudate and thalamus as seeds, as the resulting AD-DEM network was characterized by a severe loss of connectivity between the seeds and limbic structures. Overall, the demonstrated dysfunction of the dopaminergic circuit in AD pathology is corroborated by post-mortem studies showing pathologic changes such as Aβ plaques, neuronal loss and reduction of density of dopamine receptors in brain regions belonging to the dopaminergic circuits [2, 6].
The study included a cohort of amyloid and tau-positive AD subjects both in MCI and dementia stages- classified according to the biological A + T + N + criteria [16]. Quantitative analyses based on parametric images allowed us to identify alterations of the dopaminergic pathways in the whole AD spectrum. These findings are consistent with results obtained with both neuropathological and neuroimaging evidence [6, 12]. Experimental data on AD mouse models found that dopaminergic pathology is related to Aβ accumulation in both basal ganglia and cortical regions, leading to the hypothesis of an Aβ-induced neuronal dysfunction [1, 23]. These findings are also consistent with other in vivo studies on dopaminergic alterations characterizing AD [12, 20]. Moreover, MCI-AD patients showed a higher laterality of dopaminergic alterations consistently with the asymmetry of Aβ plaques in early disease stages, becoming more symmetrical as the disease progresses [24]. Although basal ganglia dopaminergic alterations has been previously associated with motor symptoms severity[6], age-related changes in the striatum have long been considered an important predictor of memory decline [25, 26]. Consistently, the 30–40% of AD patients, mostly in later disease stages, also present with extrapyramidal signs like bradykinesia, tremor, and gait disturbances [5]. This may also suggest a possible overlap with alpha-synuclein pathology which has been recently described in AD patients[27] and definitively need to be verified in on-going studies. As for the extra-striatal dopaminergic targets, we found significant alterations in regions belonging to the mesolimbic pathway, namely the anterior and middle cingulate cortex, the amygdala, the hippocampus, and the ventral frontal cortex in AD-DEM stage. This is consistent with previous imaging findings showing a major vulnerability of cortico-limbic regions receiving projections from the ventral tegmental area [12]. Also, AD-DEM patients showed significant binding reduction in the right thalamus and the lateral frontal cortex. Neuropathological studies demonstrated that D2 receptor density is reduced within the striatum of post-mortem AD brains, possibly extending to frontal regions [7]. Interestingly, in our results both the ventral striatum (~ 80%) and the dorsal caudate (~ 80%) appeared more functionally impaired as compared to dorsal putamen (~ 30% of 123I-FP-CIT binding reduction) in the MCI-AD phase. Conversely, in Parkinson’s Disease – the most studied basal-ganglia disorder - the putamen nucleus appeares more severely depleted than the caudate nucleus[28].
This evidence suggests that the striatal vulnerability occurring in AD has a different topography to that observed in Parkinson’s Disease and Lewy bodies disorders [18]. Of note, the dopaminergic depletion since the MCI-AD phase of the ventral caudate – a key component of the ventral striatum associated with reward processing – is consistent with neuropsychiatric symptoms occurring early in patients with AD [29, 30]. In particular, apathy – which is the most frequent neuropsychiatric symptom – can be secondary to dopaminergic dysfunction [31]. In the same line of thinking, beyond motor functions, the dorsal caudate plays important roles in cognitive functions [32]: by integrating spatial information and motor planning, this brain region has been shown to be involved in spatial working memory and executive functions including deductive reasoning. Although episodic memory deficits during disease progression have been widely studied and are valuable biomarkers for the AD diagnosis, more recent research has investigated working memory and executive function decline during the mild cognitive impairment [33]. Thus, the evidence presented, here, showing ventral striatum and the dorsal caudate deficits since the MCI-AD phase are important tools for early diagnosis. This is perfectly in line with previous works pinpointing dopaminergic deficits as predictive hallmarks of the pathology and of early conversion from MCI to AD-DEM [12, 34, 35].
In the dementia phases, analyses additionally found a significant reductions in hippocampus and amygdala monoaminergic binding, as compared to controls. Previous post-mortem studies found that D1-like dopamine receptors were markedly decresed in hippocampal structures compared to control values in dentate gyrus, CA1, and CA3 subfields [10]. Dopamine is known to be a modulator of hippocampal synaptic plasticity and its binding to dopaminergic receptors in the dorsal hippocampus is a major determinant of memory encoding [36, 37]. Also, D2 dopamine receptors are described to be reduced in bilateral hippocampus and amygdala [7, 9].
Indeed, recent studies on mouse models of AD proved that dopaminergic neuron loss in the ventral tegmental area – a brain area with dopaminergic projections to both amygdala, and hippocampus – occurs since the early phases of the disease, and is associated with a dopamine reduction in the hippocampus and with deficits in synaptic plasticity, neuronal function, memory and reward processing [38–43]. Additionally, hippocampal dopaminergic alterations are associated in vivo with memory performance and picture naming tasks [9]. Overall, the involvement of limbic structures may be consistent with cognitive and behavioral symptoms occurring in patients with AD. In particular, apathy – which is the most frequent symptom associated with AD – can be secondary to dopaminergic dysfunction [31].
The voxel-wise direct comparison between MCI-AD and AD-DEM stages allowed us to describe those regions progressively impaired along the disease course, without any prior assumption. The dorsal portion of caudate nucleus and the thalamus resulted more impaired in AD-DEM patients as compared to MCI-AD. These regions are considered as seeds in the subsequent seed-based connectivity analysis, showing a progressive disruption of the dopaminergic network from prodromal to dementia stages. In CG the seed (i.e., dorsal caudate and thalamus) was positively associated with other striatal regions, together with medial frontal and middle temporal regions. Consistently, previous functional connectivity studies found that the dorsal parcellation of caudate is connected to medial frontal areas, posterior cingulate cortex, and middle temporal structures [44]. These interconnected regions support cognitive functions, such as working memory and fluid reasoning [44]. Thus, the progressive disruption of this covariance pattern may be consistent with cognitive symptoms described in AD [45].
We acknowledge some limitations of this study, mainly related to its cross-sectional design. The lack of correction for partial volume effects might represent a limit, also considering age or AD related structural changes. However, the use of a standardized atlas for ROI selection and the use of non-smoothed parametric images in ROI-based analyses may have in part reduced the partial volume effect. Limitations notwithstanding, this is one of the first studies describing in vivo dopaminergic alterations in a large cohort of AD subjects from prodromal to dementia stages. We provide a strong argue for dopaminergic vulnerability in AD, with a progressive disruption of the networks along the disease course. Future studies should consider careful longitudinal designs and the possible effect of dopaminergic disruption on clinical symptoms to further extend and confirm these results.