Neurocognition suggests more progressive impairments of TBI+PTSD and PTSD towards AD
Among subjects in the investigated clinical groups, PTSD subjects exhibited the worst cognitive performance in all assessments, followed by TBI+PTSD subjects, whereas cognitive scores in TBI subjects and healthy controls did not differ significantly (Fig.
1, Table
1). Yaffe et al. have shown that military personnel with PTSD are twice as likely to develop dementia as those without PTSD [
10]. TBI and PTSD are highly comorbid conditions in civilian life and among veterans [
7,
9], and both conditions are associated with an increased risk of developing dementia later in life [
5]. This link between TBI and PTSD may result from the physical injury and consequent cognitive impairments arising from TBI [
33], or may be due to persistent trauma-related memory [
34]. The ADAS-Cog, MMSE, ECog and CDR scores all showed greater memory and cognitive impairment in subjects with PTSD, recapitulating the findings of our earlier study in a larger group of ADNI-DOD subjects who had undergone amyloid PET imaging [
25]. In their review, Regehr and LeBlanc found that the degree of impairment of cognitive and working memory was correlated with the severity of PTSD [
35].
In the present study, >35% of the subjects with TBI and/or PTSD had some memory decline (CDR score ≥0.5), and most of these subjects were diagnosed with amnestic MCI, suggesting a progressive memory decline and raising the suspicion of early AD pathology. Indeed, these subjects were classified as Braak stages II–V, which is consistent with the range of Braak stages reported in MCI subjects [
22]. Furthermore, this also suggests that a history of TBI and/or PTSD might predict memory deficits occurring decades after the trauma.
Increased tau deposition might suggest typical AD progression in TBI+PTSD and PTSD as a possible link to AD
In the present study, elevated tau deposition (10–20%) was found in the cerebral cortex of TBI subjects compared with controls. Tau is a scaffolding protein binding axonal microtubules and other proteins, and TBI causes tau to abnormally phosphorylate, misfold and cleave, and thus to form NFTs [
36]. A post-mortem study of long-term (up to 49 years) survivors of a single TBI event showed exceptionally abundant NFTs in the cingulate gyrus, SFG and insular cortex, which led the authors to suggest a causal relationship between a single TBI event and the acquisition of AD-like neuropathological features [
15]. Tauopathy has also been reported in cohorts of individuals with a history of repetitive TBI leading to CTE and ultimately proceeding to AD [
16,
32], and in a group of players of American football with repeated concussion who showed high [
18F]AV1451 uptake in the cortical grey matter–white matter junction of multiple regions, which is considered pathognomonic for CTE [
16]. The relationship between TBI and tau deposition may be a consequence of the physical damage to the axonal cytoskeleton by shearing forces [
37] in conjunction with the nucleation of abnormal tau promoting the formation of NFTs [
38]. This biophysical model of tau pathogenesis was proposed by Ahmadzadeh et al., who suggested that tau-crosslinked microtubules are sufficiently flexible to accommodate mechanical strain in the brain when it arises slowly [
39,
40], but may fail if severe mechanical strain arises rapidly, and thus overwhelms the integrity of microtubules crosslinked by tau, causing tau dissociation and aggregation [
41,
42].
Another possible mechanism may be that damage to the blood–brain barrier (BBB) after TBI facilitates tau accumulation. In this scenario, TBI induces NFT formation particularly around small blood vessels of the cortex, typically in the depths of the sulci, and this may lead to CTE [
32]. Ramos-Cejudo et al. proposed that TBI first accelerates amyloid aggregation, leading to cerebrovascular injury and BBB damage, which then results in a deleterious feed-forward mechanism in which increased arterial stiffness favours further amyloid and tau deposition [
42]. PET and histopathological examination have shown that amyloid plaque density increased within a year of the occurrence of a TBI event [
43,
44]. On the other hand, Chen et al. found no evidence of provoked amyloid plaques in subjects who had suffered their TBI 3 years previously, despite ongoing elevation of the expression of the amyloid precursor protein in the white matter [
45].
Taken together, these studies imply that transient amyloid plaques may form rapidly after TBI, but are normally cleared in subsequent years. This acute or transient response to TBI might be an initiator of a more chronic increase in tau accumulation in a pathological cascade that eventually leads to a form of tauopathy. The TBI group included subjects showing an AD-typical profile of tau deposition, with regions of increased tau appearing during Braak stages I–IV, in addition to atypical-AD regions including the frontal and cingulate cortex (Fig.
2a). A sea change in the perception of the long-term consequences of TBI has been seen in recent years, suggesting that the risks of CTE, Lewy body disease and parkinsonism are higher than the risk of AD [
46‐
48]. However, we cannot currently establish if the increased tau in our TBI group was related to AD per se or to other tauopathies, mainly because of the absence of most of the cognitive impairments evident in the PTSD groups. Longitudinal tau PET studies in this or a similar cohort may better establish the relationship between TBI and AD-like pathology.
The PTSD group also showed elevated tau accumulation in the neocortex compared with controls. A single report has so far shown increased binding of the tau tracer [
18F]AV1451 in subjects with chronic PTSD from an Australian cohort of Vietnam War veterans [
19]. To elucidate the underlying mechanism by which PTSD induces tau accumulation, Miller et al. investigated the influence of the lipoxygenase genes
ALOX12 and
ALOX15 (enzymes involved in inflammatory responses) on the decreasing cerebrocortical thickness seen in subjects with PTSD, and found that
ALOX12 moderates the association between PTSD severity and thinning of the prefrontal cortex [
49]. The
ALOX12 pathway has been found to modulate tau metabolism [
50] and may be a mediator of inflammatory mechanisms in early AD [
51].
By examining the tau accumulation profiles in individual subjects, we were able to identify those with PTSD and TBI+PTSD who showed similar tau profiles to that in AD patients. Jack proposed that early accumulation of cortical amyloid might accelerate the progression and spread of tauopathy in AD [
52]. This author proposed that “primary age-related tauopathy” develops at some stage in life followed by increased amyloid deposition in certain neocortical areas that triggers (by an unidentified mechanism) accelerated tauopathy ultimately leading to severe cognitive deficits and AD [
20,
22‐
24]. In the current study, elevated tau binding on PET was positively correlated with amyloid positivity and cognitive impairment in the PTSD and TBI+PTSD groups, but this association was not present in the TBI or healthy control groups, suggesting a particular association with PTSD.
Although none of the participants in our cohorts met the clinical diagnosis of AD, the correlation analysis of amyloid and tau PET findings suggested a strong predisposition for tau accumulation to track amyloid deposition, especially in the TBI+PTSD group, thus suggesting a complex relationship between the two pathologies. However, further investigation is required to substantiate this association. In amyloid-negative subjects, tauopathy with Braak stages above zero might be primary age-related tauopathy [
37], and this also might explain the occasional finding of tau accumulation in our healthy control group. Alternately, our criterion for amyloid PET positivity of SUVr >1.1 in the whole cerebral cortex [
53] may have resulted in early amyloid changes being missed in some subjects.
We found significant correlations between ADAS-Cog, ECog total and CDR scores and tau accumulation in both the TBI+PTSD and PTSD groups, with the most compelling correlations in the TBI+PTSD group (Figs.
4 and
5, and Supplementary Fig.
1). In this group, the spatial pattern of positive correlations broadly matched the default mode network (DMN), that involves the precuneus, PCC and medial frontal cortex [
54]. Furthermore, tau accumulation in these same regions was positively correlated with total cortical amyloid deposition (Fig.
6). These regions of the DMN have previously been shown to contain amyloid deposits in patients with MCI [
55] and early AD [
54], suggesting that the DMN is the first functional network to be disrupted in AD [
55]. These various correlations between tau and cognitive impairments and amyloid may suggest that TBI+PTSD and PTSD subjects are at higher risk of conversion to AD, following the typical AD progression profile proposed by the NIA-AA framework [
20].
The data presented here imply that those veterans who developed PTSD following their TBI might be at the highest risk of progression to AD, while those with TBI only might be more at risk of developing other neuropathies [
46‐
48], a conjecture that could be investigated by longitudinal molecular imaging studies. Work by Li et al. showed that a self-reported history of TBI was associated with an onset of cognitive impairment in older adults 3–4 years earlier than in those without a history of TBI [
56], but these authors did not report interactions with PTSD.
The major limitation of this study was the small number of subjects in the TBI cohort (n = 10), which was insufficient to support strong conclusions. Further investigations are required to establish better links between TBI or PTSD with tau pathology and the risk of AD or other forms of dementia. In addition, there is a need for further investigation of the mechanisms triggering AD onset and progression. Future studies in a larger cohort may establish cut-off criteria for tau PET conforming to Braak staging.