Yasmine Fathy

Difficulties in planning and solving problems, as a result of executive deficits, end up disrupting PD patients’ social and functional lives [60]. There are various pathophysiological models to explain executive deficits in PD, one is damage to the mesocortical dopaminergic network. Dopaminergic fibers emerging from the ventral tegmental area project profusely to the neocortex, particularly the prefrontal, insular, and cingulate cortical regions, modulating D2 receptors within those regions. The insula is thought to play a major role in this network by recruiting and controlling other networks involved in cognition, a function possibly disrupted in PD due to damage of mesocortical dopaminergic projections [60]. Furthermore, recent studies suggest the anterior insula atrophies early and could function as a diagnostic biomarker in prodromal DLB [61].

Our meta-analysis showed that insular atrophy in FTD was associated with speech deficits, fear conditioning, deficits in empathy, recognition of emotions, and memory enhancement through emotion. A large cluster of left insular atrophy was associated with deficits in behavior such as disgust, abnormal eating behavior, and compliance with social norms. This suggests that insular atrophy and the vulnerability of VENs in both the insula and cingulate cortex are involved in the clinical deficits seen in FTD. We additionally showed in voxel-based morphometry (VBM) studies that insular atrophy, particularly of the right insula, is associated with cognitive impairment, lack of awareness, psychosis, delusions, and hallucinations in AD. Other studies have shown that the right insula is closely associated with several neuropsychiatric deficits including irritability, hallucinations, delusions, and agitation in AD [62].

It’s all about networking

Axons play a crucial role in connecting different brain regions with each other, allowing transfer of information. Myelinated axons particularly allow fast impulse propagation through saltatory conduction, necessary to yield complex functions such as motor, sensory, and cognition. Through the anisotropic diffusion of water molecules in nerve fibers, driven by alignment of fibers within axonal cytoskeleton [63], white matter tracts can be identified and reconstructed through diffusion tensor imaging (DTI) [64]. By providing information on water diffusion and fiber orientation, measured through mean diffusivity (MD) and fractional anisotropy (FA), DTI provides quantitative indices for the detection of white matter microstructural abnormalities [65, 66]. In PD, DTI shows abnormal FA and MD in multiple brain regions even prior to atrophy [67], indicating it could be used as a biomarker for disease progression. In chapter 5, we discussed findings from several VBM studies in PD-MCI, in which insular atrophy correlated with executive function deficits and cognitive impairment [68]. Moreover, the salience network, consisting of cingulate and insular cortices as core regions, as well as right insula volume correlate with the severity of depression in PD [69]. Overall, studies focusing on the substrates of non-motor deficits in PD have found that structural, metabolic, and pathological changes within the insular cortex could potentially play a profound role in cognitive, behavioral, and somatosensory disturbances in PD [70]. Considering the role of both the anterior insula and the cingulate cortex in the salience network, as well as the involvement of both the insula and cingulate in the pathological staging of PD, we then explored the microstructural white matter changes of the anterior insula in relation with the cingulate cortex in PD. In chapter 6, we showed that tracts between the dorsal anterior insula and anterior cingulate had lower FA and higher MD values associated with lower semantic fluency, working memory, and executive functions. These values were also associated with higher anxiety scores [71]. This indicates that pathology and axonal degeneration in the anterior insula leads to a specific loss of connectivity between the anterior insula and cingulate cortices. The anterior cingulate cortex has anatomical and functional characteristics very similar to the anterior insula. Both regions have affective and cognitive functions and together they form the salience network, which plays an important role in cognitive control [72]. Lower FA and higher MD values may reflect a number of microstructural tissue changes such as demyelination, inflammation, edema, or necrosis [73]. Considering the neuropathological findings present in PD and the axonal and myelin neurodegenerative features in the disease, as described in chapter 4, further disconnection between the anterior insula and cingulate cortices, and to a lesser degree disconnection from other regions within the salience network, would lead to several non-motor deficits, most importantly an impairment of cognitive control. As seen in chapter 7, a disconnection between the anterior insula and cingulate cortex is also associated with executive deficits [74], commonly present in early-stage PD. Executive functions include three core functions, namely, inhibition and interference control (such as self-control and selective attention), working memory, and cognitive flexibility. These are all essential functions for reasoning, problem solving, and planning [75].

In the late 90s, by understanding how complex systems work, through interactions among its constituent elements, the concept of network neuroscience gained increasing attention allowing us to decipher the interactions between large-scale brain networks that underlie complex functions such as cognition [76]. Understanding the human brain connectome has shed much light on how the brain is an intricate collection of networks with clusters and hubs integrating information from specialized modules, the latter being target regions for degeneration [77]. The insular cortex is known to function as a structural brain hub integrating information from various other brain regions [78]. Although there are a few studies on the functional connectivity of the insula, none break the region down into its separate sub-regions in PD. In chapter 7, we describe significantly reduced cognitive performance in the PD patient cohort compared to age-matched controls using an extensive cognitive battery (Cambridge cognition- CAMCOG), assessing seven different cognitive domains [74]. Reduced cognitive performance in this PD cohort was associated with functional connectivity between the dorsal anterior insula and the default mode network; while in controls connectivity between the insula and the fronto-parietal network was associated with cognitive performance. Across all brain regions, only dorsal anterior insula and anterior cingulate showed reduced connectivity, which again was associated with cognitive performance in PD only. These results indicate that damage of the dorsal anterior insula connectivity with the anterior cingulate may contribute to poor cognition in PD. Since both regions play a role in the salience network where they control and switch between large-scale networks such as the default mode network (DMN) and fronto-parietal network (FPN) [79], we assumed that damage of the insula-cingulate functional network leads to disrupted control of these networks. To assess this further, we computed the degree of centrality of each of those nodes and found increased betweenness centrality of both the DMN and FPN in PD which in turn was associated with reduced insular-cingulate connectivity and cognitive performance. If we add up all the information on insula-cingulate structural and functional similarities, interactions, shared networks along with our findings on their disconnection, structurally and functionally, and pathological involvement, we have a simulated model of degeneration that not only impacts the integrative functions of both regions but also of large-scale networks playing important roles in cognition (Figure 1).

Strengths and limitations

This thesis includes a combination of post-mortem human brain studies and MRI studies as well as a systematic review and meta-analysis, all aimed at elucidating the role of the insular cortex and its sub-regions in neurodegeneration, particularly in PD and DLB. By diving into the insular sub-regions at both the micro and macro-scales, we were able to identify characteristic cellular, structural, and functional features of each sub-region, their degenerative features, and role in clinical deficits. This thesis provides novel insights in the extent of insular cortex involvement in clinical features of PD and DLB. Despite providing novel results on the insular role in PD and DLB, the studies included in this thesis have limitations. In the neuropathological and neuroanatomical studies, due to the scarcity of post-mortem tissues, the sample sizes were small. Furthermore, the studies were rather descriptive in nature and mainly described the morphological features of cellular and axonal pathology and neurodegeneration in the insular sub-regions. Furthermore, MRI studies contained small sample sizes, particularly the control group, which prevented us from generalizing our results. Future studies recruiting larger cohorts and using expanded quantitative analyses on insular cell loss, molecular mechanisms underlying myelin abnormalities, and the relationship with axonal transport could shed much light on the extent of neuronal and axonal damage in the insula in different stages of disease. Moreover, studies combining post-mortem quantitative MRI and pathological features, as in our NABCA pipeline [80, 81], would elucidate the relationship between local insular pathology and network characteristics as well as changes in PD and DLB.

Figure 1. Selective neuropathology and degeneration in the anterior insular and cingulate cortex, impact on brain networks, and role in cognitive impairment in PD and DLB. The insula, sub-divided into 3 sub-regions, anterior agranular (vAI; red) and dysgranular (dAI; green) and posterior granular (yellow) is heavily affected by pathological aggregates and shows abnormalities in neurons (α-synuclein aggregates in VENs), axons (demyelination and axon-myelin detachment, top right), and glia (synucleinopathy; top left) in PD and DLB. Both AIC and ACC contain VENs, show same sub-divisions, and similar pathological staging. Together they form the salience network which is disrupted resulting in large-scale network dysfunctions and disrupted cognitive control as well as emotional processing (triple network). ACC: anterior cingulate cortex; AIC: anterior insular cortex; dAI: dysgranular anterior insula; DLB: dementia with Lewy bodies; LBs: Lewy bodies; PD: Parkinson’s disease; VENs: von Economo neurons.

Future perspective

After identifying the underlying neuropathological, cellular, behavioral and network abnormalities in the insula in PD and DLB, one question remains: how can this information benefit patients? Considering the complexity of neurodegenerative diseases and the challenges associated with proper disease management in these patient groups, there are three main components to consider: 1) protein aggregates, 2) diagnostic and prognostic tools, 3) treatments (neuropsychiatric and cognitive). First, understanding PD and DLB, as well as many other neurodegenerative diseases, requires an understanding of the associated neuropathological protein aggregates and their effects on neuronal/non-neuronal dysfunction. Given the evidence provided herein of the existence of multiple neuropathological protein aggregates within the anterior insula in PD(D) and DLB, it is important to understand the consequences of their combined presence. So far, mechanistic studies trying to understand the effects of LBs and LNs have yielded controversial and contradicting results [82, 83]. What happens when we study the combined effect of α-synuclein, amyloid-β, p-tau, and AGD on axonal transport deficits and cellular (dys)function and death? Future studies should investigate the combined effect of different pathologies on cellular and axonal dysfunctions and the mechanisms through which different protein aggregates may stimulate each other and enhance the overall process of degeneration. Second, despite having several imaging, CSF, and other diagnostic tools, the prediction of progression to dementia, a pre-determined fate, remains difficult to measure. Since the stages of hippocampal/cortical involvement (Braak stage 4-5) in PD are thought to be associated with cognitive impairment, the insular cortex, including insular atrophy and reduced dopaminergic receptor binding, as a prognostic marker for cognitive and emotional deficits remains underestimated yet may be promising. Future studies should thoroughly investigate the role of insular and cingulate abnormalities, such as atrophy, glucose metabolism, dopaminergic receptor binding, tau and amyloid binding using PET scans in PD and DLB, including their potential to serve as prognostic biomarkers in terms of progression to dementia. Third, cognitive and neuropsychiatric deficits pose a great burden on patients and are important to truly understand. The insular cortex, along with the cingulate, are two regions largely playing a role in such deficits and may be studied further to try to grasp the neurochemical abnormalities underlying these deficits. Translational studies assessing the intersection of neuroanatomical, neuropathological, and connectome abnormalities within the insular cortex and associated clinical deficits in these patient groups would transform our understanding of the insula’s contribution to disease.

Conclusions

Throughout this work, there are a few exceptionally illuminating pieces of information worth considering: 1) Studying the insular cortex as a set of sub-regions is important as each sub-region harbors its own cellular, structural, and functional features as well as variable contribution to disease, 2) Differential characteristics of the insular sub-regions from cellular components to myelination and even type and gradient of pathological aggregates are defining features of the insula and possibly other similar cortical brain regions, 3) The insula supports complex intellectual functions through its interaction with many other brain regions and is a hub for pathological deposits, 4) Degeneration of the insula contributes to extensive deficits through loss of internal specialized functions, disruption of the salience network and other related large-scale networks playing a role in cognition, and loss of integrative functions, and 5) the anterior cingulate and anterior insula together represent influential brain regions sharing unique cellular and cytoarchitectonic features, special circuitries and networks, and a similar spatio-temporal sequence of disease-specific pathology. Altogether, neurodegenerative changes in the anterior insula are severe in PD/PDD, and particularly in DLB where different pathologies coincide and damage extends to neurons, axons, and the cytoskeleton. At the connectome and brain network level, anterior insula - anterior cingulate connectivity is disrupted and the insula loses its central position in the brain network, including its control of other related networks involved in cognitive control. Overall, the evidence presented herein raises an important question: if proteinopathies target essential brain hubs, shouldn’t disease management strategies target these regions too?