Can the entorhinal cortex help distinguish healthy aging brains from pathological aging brains?

Not all brain aging is equal

The aging brain progresses through two distinct processes: one is healthy aging, the other is pathological aging which causes brain dysfunction such as dementia, of which there are 5 forms: Alzheimer's disease (AD), dementia with Lewy Bodies, vascular dementia, frontotemporal dementia, and Parkinson’s disease dementia. Other forms of dementia are “mixed dementia” dementia associated with Huntington’s disease and Creutzfeld-Jakob disease. Alzheimer’s disease (AD) is the most common form of age-related dementia, its occurrence rises from about 10% in persons in their 60 s to > 50% in individuals aged 80 years and more.

A major challenge for clinicians is to distinguish between a healthily aging brain from one that is undergoing pathological aging. Senile plaques (composed of insoluble amyloid β or Aβ) neurofibrillary changes, and brain atrophy are still considered to be the pathological hallmarks of AD [48]. However, persons who do not display clinical signs of dementia of the AD type may nevertheless display some of these pathological features [1], [37], [44]. While Aβ production increases with aging [47], Aβ deposition and decline of brain function is not correlated [22], [23]). In contrast, the load of neurofibrillary tangles (NFT, comprised of abnormally hyperphosphorylated tau protein) correlates well with the extent of brain atrophy and functional impairments [22], [23], [39], [43]; however, dementia associated with increased NFT and brain atrophy does not necessarily warrant a diagnosis of AD[39], [15]. Despite this, it is interesting to note that NFT that emerge during aging are first observable in entorhinal cortex (EC), the integrity of which is crucial for memory formation. Over time, NFT spread from the EC to the limbic cortex and neocortex, regions associated with emotions and executive functions, respectively (Fig. 1). Thus, it seems likely that this sequential formation of NFT at least partly underpins the clinical progression of AD [26], [54].

Fig. 1.

Development of NFTs and neural connection. NFTs spread from the entorhinal cortex to the limbic and neocortex regions, and the entorhinal cortex and limbic region, and the limbic region and neocortex region have neural connections. NFTs and associated neuronal cell death make the connections between those sites and the sites to which they are connected vulnerable, resulting in progressive loss of brain function.

Connections of the entorhinal cortex and graph-theory based understanding of multi-level damage to neurocircuits responsible for optimal brain function

Interestingly, although NFT in the EC have so far not been seen to be accompanied by signs of dementia, NFT in the limbic areas and neocortex usually associated with mild cognitive impairment or even overt dementia [5], [30]. That NFT in the EC apparently do not cause loss of function may reflect compensatory changes in other brain areas such as prefrontal area and primary sensory cortices. We conjecture that the appearance of functional deficits in association with the buildup of NFT in the limbic and neocortex may be indicative of insufficient compensatory plasticity.

Neurons of EC are mostly innervated by the fibers carrying sensory signals via the perirhinal, prefrontal and retrosplenial cortex, olfactory structures, and the subiculum; in turn, they send fibers to the hippocampal dentate gyrus and CA3 layer, and the prefrontal and retrosplenial cortex [29], [51], [41]. Thus, the EC serves as major hub that connects the cortex with the hippocampal formation, a sub-cortical structure (Fig. 2). These anatomical arrangements align with graph theory which predicts that neural networks are hierarchically organized. Thus, lower networks (e.g. primary sensory and motor neurons), comprising clusters of neurons (hubs), couple to higher (e.g. fronto-cortical) networks through inter-connected hubs that receive inputs from clusters of neurons (micro-networks) [52]. Given this organizational plan, it follows that, if age-related damage occurs at a lower-level hub (e.g. EC), brain function can only be preserved, at least partially, through the implementation of compensatory mechanisms within higher-level hubs [2]. These corrective measures may be however subject to overload, ultimately resulting in widespread damage to higher-level hub(s) in a domino-like manner (Fig. 3).

Fig. 2.

Navigation and neural network. In navigation, visual information is transmitted to the entorhinal cortex and hippocampus, activating grid cells and place cells. This information is transmitted through retrosplenal cortex to Parietal cortex and Precuneus (overlapping with default mode network) for egocentric coding. Landmark location and attention code allocentric navigation in the posterior cingulate cortex via the frontal lobe (overrapping with central executive control network). When visual information is reduced, the entorhinal cortex is overloaded to maintain the activity of the retorosplenal cortex and prefrontal cortex. As a result, NFTs are formed in the entorhinal cortex. When it happens, the prefrontal cortex and retrosplenial cortex, which are the networks above the entorhinal cortex, become overactive.

Fig. 3.

Network Collapse by Graph Theory. According to graph theory, each neural circuit form a hub structures (round). When one hub fails (stop mark), the hub above it is overloaded (star mark) to maintain its function. However, when the number of hubs that fail increases due to overwork, it causes a loss of function.

Egocentric navigation – An example of the role of the EC as a lower-level integrative hub

Spatial navigation is the process by which an individual determines and maintains the optimal trajectory from one location to another. It depends on multiple cognitive and perceptual processes that generally decline at older ages [21], [36]. Two – not necessarily mutually exclusive – strategies are used in spatial navigation: self-centered egocentric navigation, and landmark-dependent allocentric navigation [34]. Allocentric navigation uses landmarks encoded by place cells in the hippocampus [42], and in egocentric navigation, information about location is mainly provided from grid cells localized in the EC that respond to every regular hexagonal grid point in space, localize in entorhinal cortex [24], [19]. Grid cell-like activity has been observed in functional brain imaging of the human brain [13], [33]. Such activity correlates with path integration [53], a function that is impaired in subjects with mild cognitive impairment (MCI) and which is negatively correlated with EC volume [28].

Whereas young adults use both, the allocentric and egocentric strategies for navigation, older subjects predominantly navigate using an egocentric strategy [46], [20], [9]. Increased utilization of egocentric navigation, which may be accounted for by age-related deficits in vision and hearing, is thought to be a compensatory mechanism [9] in response to dysfunctional (sensory) neural networks. In young, healthy adults, interactions between the parietal lobe which is essential for egocentric navigation and the medial temporal lobe which is required for allocentric navigation are reconciled by the retrosplenial cortex (RSC) [8], [59], [12], [3], as shown by connectome analysis [57], [13], [10]. Accordingly, we postulate that the EC, which locates to a lower-level hub as compared to the RSC (Fig. 2), becomes overloaded in aged subjects resulting in a shift to almost exclusive dependence upon an egocentric navigation strategy.

Appearance of NFT in the EC herald pathological aging of the brain

Functions other than navigation may be impaired by the increase in EC activity in response to deficient input from the primary sensory cortex. We propose this compensatory mechanism to trigger the formation of NFT in the EC since hyper-activation of neurons stimulates the translation of tau in dendrites, followed by the hyperphosphorylation and accumulation of tau [31], [35], [32]. Since the build-up of insoluble tau-containing NFT is irreversible, hippocampal and fronto-cortical neurons become hyper-excited in an attempt to maintain brain function. However, the latter facilitates a vicious cycle whereby the sequential accumulation of NFT in the hippocampus and frontal cortex results in a gradual decline of cognitive functions. This hypothesis is not necessarily incompatible with suggestion that tau seeds “shed” from the EC are spread through the neural circuitry [11], [38]. Importantly, the available evidence indicates that the appearance of NFT in the EC mark the start of a pathological process that eventually leads to cognitive dysfunction typical of AD and other tau-related dementias.

Where do we go from here?

The search for treatments for AD domain has been largely dominated by the Aβ [25]. However, to date, none of therapeutic trials targeting Aβ have proven successful in terms of inhibiting the progression of dementia [18], [14], [58], [56], [7], [16], [27]. It seems that the accumulation of Aβ marks a point of no return; therefore, therapeutic developments must aim at events that precede Aβ deposition. Here, we may be guided by revisiting early studies on the progression of AD neuropathology.

The idea that the EC may be the location at which pathological aging commences is, in fact, not new! It is rather an extension of classical work showing that NFT appear in the EC during the third decade of life, well before Aβ deposition at mid-life [6]. Thus, it is plausible that increases in Aβ are secondary to compensatory mechanisms in the EC that inadvertently trigger NFT formation and the vicious cycle referred to in the previous section where lower-level hubs are disturbed before higher-level hubs. Such a scenario could underpin the functional deficits seen after cerebral ischemia, for example, although these are usually attributed to increases in the levels of Aβ [60], [17].

The temporal sequence of neuropathological events indicates that AD therapies should consider the Braaks’ description of the trans-EC stage of the disease, because it is too late for intiating therapy based on neuropsychological assessments form the basis of clinical staging of AD [45], [4], [40]. In order to allow earlier detection of pathology, at the Braak “trans-EC stage”, the conventional tests would require refinements, including improved sensitivity, especially in light of the aforementioned functional compensatory mechanisms as NFT begin to disrupt EC activity.

If the assumption that brain aging, at least with respect to AD, begins with NFT formation in the EC, evaluation of path integration performance (see above) in clinical research is likely to help early, non-invasive, detection of disease. The results of such research will, in turn, help verify the view that NFT in the EC mark the earliest-detectable signs of AD pathology. If the hypothesis is proven correct, research should then focus on finding ways to delay or prevent NFT formation in the EC and the propagation of aberrant signals to higher-level hubs responsible for memory formation and recall as well as executive functions. Preclinical studies have brought insight into the mechanisms responsible for NFT formation [55] and have also begun to identify means of preventing the generation of insoluble tau fibrils [50], [49].

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.