Our exploration identified multiple circulating eCBs associated with cognitive function among community-dwelling, dementia-free older adults. Several eCBs are novel compounds with no previous information regarding their associations with cognitive health. Our study further underscores the importance of sex and the ApoEɛ4 as possible effect-modifiers, which should be taken into consideration in future studies of the eCBome and brain health.

The link between eCBs and cognitive function demonstrated in our study is supported by findings from multiple neuropathological, animal and cell culture studies. These studies highlighted several beneficial effects of eCBs including neuroprotection against excito-toxicity, inhibition of neuro-inflammation and neuro-degeneration, antioxidant activity and protection against Amyloid-Beta aggregation and toxicity [7, 23]. Yet, manipulations performed on the eCB system have shown contradictory effects on cognitive function, which are at least partly explained by excessive CB1 activation [24, 25].

Our findings suggest that increased circulating levels of the fatty acids LA, LnA and OA may be associated with poorer executive function. In general, fatty acids are essential players in central mechanisms underlying brain health including neurogenesis, neuronal inflammation, oxidative stress and neurotransmitter production [26]. Furthermore, fatty acid metabolism has been associated with AD pathology and cognitive function [27]. In line with our findings, a recent study showed that individuals with higher plasma levels of LA, the most abundant polyunsaturated n-6 fatty acid in the diet, had a faster cognitive decline [28]. This evidence raised the concern that excess levels of LA might adversely affect the brain due to its pro-inflammatory properties [29]. However, in contrast to our findings, other studies have shown that high circulating levels of LA may be related to lower risk of dementia and to improved cognitive function [30, 31]. Hence, it is postulated that LA may have neuroprotective effects, at least partly explained by its anti-oxidative properties, when not being used excessively [32]. Interestingly, we show that among carriers of the ApoEɛ4 genotype, high LA levels were related to better perceptual organization. Similarly, decreased stroke and mortality risk in relation to high plasma LA levels were previously observed only among FHS participants with, but not without, the ApoEɛ4 genotype [33]. These effect modifications by ApoEɛ4 can be explained by differential transport and metabolism of fatty acids based on ApoE ɛ4 status [34]. In addition to LA, high levels of LnA, a plant-derived n-3 fatty acid, have been associated in some prior literature with improved cardiometabolic health and anti-inflammatory effects [35, 36], while in our sample it was associated with poorer cognitive performance. However, most previous studies focused on LnA from diet rather than in blood, and its association with cognitive performance has been largely unexplored [36]. Lastly, OA, an n-9 fatty acid particularly abundant in olive oil is mostly known for its health benefits, including in the brain [37]. In addition, a recent study showed that high plasma levels of OA may be related to lower risk of mild cognitive impairment and AD [38-40]. The inconsistent findings comparing previous studies to the current research may stem from differences in study specious (i.e., cell cultures or animal models vs. humans), source of the fatty acids (e.g., diet, brain, CSF vs. blood), and differences in participants characteristics (e.g., age, disease/health status) [41]. Of note, the current associations of LnA and OA with executive function were found only in women. This finding is in accordance with previously noted interactions of fatty acids, particularly n-3 fatty acids, with sex with regard to cognition [40].

In general, fatty acid amides such PEA and AEA have consistently shown neuroprotective effects and a potential therapeutic effect for AD [42-45]. However, most research focused on the fatty acid ethanolamide class of fatty acid amides, while less is known on the association between fatty acid primary amides and brain health. Results from the current study does not support the evidence for neuroprotection of fatty acid ethanolamides, but rather suggest that high plasma levels of Ln-Am and L-Am, which are fatty acid primary amides, are associated with poorer verbal memory and attention. In agreement with these findings, a prior study that explored a panel of 883 plasma metabolites demonstrated a link of several fatty acid primary amides, including L-Am, with AD pathology and clinical diagnosis [46].

Another eCB class highlighted in our research is the N-acyl amino acids. N-acyl amino acids are eCBs that gained renewed scientific interest following the discovery of AEA, one of the most studied eCB compound, which binds directly to CB1 and CB2 receptors [47]. However, little is known to date on the molecular functions of the N-acyl amino acid family [47]. Specifically, our exploration showed that increased levels of N-acyl alanine (i.e., O-Ala and P-Ala) may be associated with poorer executive function. An in vitro exploration previously showed that compounds within the N-acyl alanine family may impair cell proliferation possibly through increased reactive oxygen species [48], a process that is altered in AD brains [49]. Our sex-stratified analyses identified additional N-acyl amino acids that are related to cognitive function, among men only. These findings are supported by a previous study from our group, demonstrating an association between eCBs from the N-acyl amino acid family with Glial fibrillary acidic protein (GFAP), a marker of neuronal injury, in men but not women [50].

Several biological explanations can be suggested for the sex interactions observed in our study. For example, the observation that the fatty acid LnA was associated with poorer executive function only among women may be explained by the decreased estrogen levels of the women in our sample, most of whom were at the postmenopausal phase. Indeed, estrogen may accelerate the transformation of LnA to Docosahexaenoic acid (DHA) and Eicosapentaenoic acid (EPA) [51, 52], which in turn, have shown associations with decreased AD risk [53, 54] and improved executive function [55, 56]. In addition, the fact that fatty acid amides were associated with poorer cognition only among women may be attributed to the strong influence of sex on fatty acid amide hydrolase (FAAH) [57, 58] which is responsible for the regulation of fatty acid amide levels [59]. Lastly, the observed effect modifications by ApoEɛ4 may be due to binding of ApoEɛ4 to sortilin receptor, which in turn disrupts sortilin’s ability to convert polyunsaturated fatty acid to eCB compounds [16].

Our study has several novelties and strengths. First, the study focused on many of the known eCB compounds, while others mostly addressed 3-5 specific compounds or single families [60]. Furthermore, the comprehensive cognitive assessment among community-dwelling dementia-free adults allows a valid generalizability to wide populations. Nevertheless, several limitations should be acknowledged. First, conclusions drawn from our study cannot infer a temporal relationship between eCBs levels and cognitive performance due to the study’s cross-sectional design. Second, the statistical power of our study is limited due to small sample sizes, particularly in the stratified analyses. Therefore, we consider our findings preliminary and encourage their validation in larger and diverse samples. Third, it is important to note the limited stability of eCB compounds. We acknowledged this limitation and therefore used the optimal laboratory practice (e.g., plasma frozen without delay, no repeated freeze/thaw cycles, blood draw in resting state after fasting). Fourth, little is known about the changes in eCBs levels over time in humans. Hence, changes in circulating eCB levels may have occurred in the time duration between the blood draw and the cognitive assessment. Fifth, it is unclear whether circulating eCBs levels reflect brain signaling. Yet, evidence shows that eCBs can cross the blood brain barrier [61]. Additionally, results showing that circulating eCB levels are linked with AD-related health conditions such as cardiovascular diseases, obesity and depression [24, 62] may support their role in cognition. Sixth, the external validity of our study is restricted, because it includes participants predominantly of high socioeconomic status and European ancestry and from one geographic region. Lastly, although main relevant covariates were considered, some residual confounding may remain. Particularly, lifestyle factors such as exercise and nutrition may confound the association between eCBs and cognitive since they have been associated with both cognitive function [63, 64] and eCBs levels [65, 66].

In conclusion, we demonstrate several significant associations between certain eCB molecules and cognitive function in dementia free middle- and old-age individuals who reside in the community. Overall, our findings support existing explorations in animal models and cell cultures, stressing the possible role of eCB signaling in cognitive performance. Yet, the variation in the direction of the eCB-cognition associations, together with the evidence of sex and ApoEɛ4 genotype highlight the complexity of the eCBome and the need for validations in external samples. A better understanding of the role of eCBs in cognitive health is important, since some eCB molecules may serve as blood-based biomarkers for cognitive impairment. Further, the levels of many eCB compounds including fatty acids, fatty acid amid and N-acyl amino acids may be modulated by interventions such as diet, hence may show therapeutic benefits for cognitive aging in older persons.