In the present study, we systematically characterized changes in the eCB system during the aging process in mice. Our findings demonstrate that both major eCBs, 2-AG and AEA, were affected with age in a brain region-specific manner. The 2-AG was most prominently reduced in the Hip and moderately in the CPu, whereas AEA was mostly decreased in the CPu as well as in the mPFC+Cg. Consistent with the changes in 2-AG, its synthesizing enzyme DAGLα was also down-regulated in the Hip. Conversely, the protein levels along with its ligand-binding density of its cognate receptor CB1 were prominently increased in the Hip. This increase may reflect a compensatory mechanism.

These data are consistent with previous results showing that the tone of the eCB system declines during aging. For example, our lab previously showed a prominent decline of 2-AG and its synthesizing enzyme DAGLα in the Hip at old age using two different mouse strains [9]. Although that study provided the first evidence of these changes at the time, the conclusions were drawn from a direct comparison of young (2 month) and old (12–15 month) mice. Furthermore, the alterations in the eCB tone during the aging process were not determined in that study. In the current study, we used five different age groups, ranging from 2 months to 18 months in order to characterize the influence of aging on the eCB system. We followed the age-categorization of Jackson Laboratory and other studies (https://www.jax.org/news-and-insights/jax-blog/2017/november/when-are-mice-considered-old# (accessed on 5 December 2021) [18,19]); therefore, 2-month-old mice were termed as adolescent, 6-month-old as adult, 9- and 12-month-old as middle aged and 18-month-old as old. The major finding of the current study is that the eCB tone is not linearly decreased with age, rather it is most affected during the middle age, and thereafter, it revived modestly.

Middle age is critical for several molecular and behavioral changes. Deep proteomic profiling of plasma proteome from young individuals to nonagenarians revealed undulating changes that indicate three waves of aging across the lifespan. They identified three distinct peaks during the fourth, seventh and eighth decades of life, indicating the number of proteins that are differentially expressed at that particular time window. Interestingly, among the most strongly altered proteins during the middle and old age, there are proteins that are associated with cognitive functions. These findings are an indication that the middle age is crucial for robust molecular changes [20]. Further, several rodent studies also reported prominent behavioral alterations during the middle age. For example, 2–3-, 4–5-, 6–7- and 8–12-month-old mice were tested for changes in motor function, locomotor activity, anxiety-like behavior, social behavior, depression-like behavior, and learning and memory functions with respect to age. Changes were observed in behaviors between individual age groups tested against the youngest group. Interestingly, for most behaviors, 8–12-month-old mice exhibited more pronounced changes than the other age groups compared to the youngest group [21]. A more recent and similar study also reported age-related changes in spatial learning and memory in C57BL/6J mice [22]. However, this study tested the learning and memory functions at 2, 4, 6, 9 and 12 months of age and reported a gradual decline in these performances only from 6 months of age [22]. In another study, deficits were observed in the Hip-dependent contextual fear conditioning task in middle-aged mice. Afterhyperpolarization is a mechanism generally engaged during the learning process. Although basal neuronal excitability seems to be comparable in both young and middle-aged mice, learning-related modulation of the post-burst afterhyperpolarization was impaired in CA1 neurons in middle-aged mice [23]. Emotional behaviors were also tested in rats at 3 (young) and 12 months (middle-age) of age. It was shown that middle-aged rats were more anxious, depressed, and displayed aversive memory impairments [24].

Importantly, age affects brain on an area-specific manner: reduction in white-matter and cell loss is reported in the cortex, hippocampus, putamen, thalamus or striatum, whereas other brain regions remain largely preserved during physiological aging [25]. Thus, the cerebellum, which is a major brain area with an exceptionally high CB1 receptor expression [26], was not investigated in the present study.

CB1 mediates two types of signaling cascades in the hippocampus based on its location. CB1 on the Schaffer commissural projections mediates phosphorylation of Munc18-1 through ERK1/2 signaling to depress the transmitter release [27]. Most CB1 agonists are capable of activating this signaling cascade. However, 2-AG mediated CB1 signaling on the lateral perforant path terminals entails β1-integrin to effect synaptic potentiation by stably enhancing the transmitter release. This signaling cascade is explicitly engaged during the lateral perforant path-dependent learning [28,29]. Interestingly, perforant path projections of layer II entorhinal cortex neurons deteriorate during middle age, accompanied by prominently impaired long-term potentiation (LTP) in the medial and lateral perforant paths in both mice and rats. Accordingly, middle-aged mice also performed poorly in the lateral perforant path-dependent episodic memory task. Importantly, disrupted eCB signaling due to changes in 2-AG levels were attributed to the deterioration of later perforant path LTP during the middle age. Although data for basal 2-AG levels among the age groups were lacking, it was shown that increasing 2-AG levels by inducing DAGLα expression using physostigmine or by blocking the breakdown of 2-AG using JZL184 restored the lateral perforant path LTP [30]. In line with the above findings, we also show here that 2-AG levels were prominently affected during the early middle-age in both Hip and CPu. The findings from Daglα knock out (KO) mice have clearly shown that a loss of 2-AG adversely affects the emotional state in mice, resulting in enhanced anxiety, stress, fear responses, sociability deficits, and impaired exploratory drive [3,31]. Moreover, the fact that middle-aged C57BL/6J mice also exhibit the above stated behavioral deficits while 2-AG signaling is most affected is quite fitting together.

Here, we show that CB1 protein levels were first slightly increasing in the 2–9 month age interval in most sub-regions of the Hip and dropped significantly at 12 months in all the sub-regions of the Hip. Further, we also observed a CB1 protein decrease in CPu (at 12 months), Acb (at 9, 12 and 18 months), and Cg (at 6, 9 and 12 months). These undulating changes in CB1 protein levels, specifically during the middle age in the sub-regions of the Hip, may constitute a compensation for the 2-AG loss. Although, the molecular mechanisms underlying this phenomenon are not clear, the compensation seems to be specific to the Hip and not for other regions tested. In agreement with this, early studies also reported a decrease in Cb1 mRNA levels in aged rats (>24 months) against young (3 months) rats. It was shown that Cb1 mRNA levels were specifically decreased in CPu, and to a small extent, in the sub-regions of Hip (CA1, CA2, CA3, and DG) [32,33]. Age-related changes in CB1 protein levels were also compared between 4- and 24-month-old rats using Western blotting. Interestingly, the study reported an increase in CB1 protein levels in entorhinal and temporal cortices, and a decrease in the postrhinal cortex, whereas there were no changes in the Hip or its sub-regions [34]. The above studies unfortunately lack data from other age groups. In human muscle, the CB1 protein levels seem to increase with age when compared between young (20–27 years) and old (≥65 years) individuals, and CB1 expression seems to increase with resistance exercise in older individuals [35]. Mice subjected to voluntary exercise for eight days exhibited increased eCB signaling in the Hip, accompanied by an increased progenitor cell proliferation within the DG. Eight days of free access to running wheels also increased the agonist binding site density of CB1 receptors within the Hip [36]. Aerobic exercise with moderate intensity induced antinociception in rats that was mediated through CB1. Exercise also induced the expression and activation of CB1 in rat brains (periaqueductal gray matter) assessed using Western blotting and immunofluorescence. In addition, exercise also elevated the plasma levels of both AEA and 2-AG [37]. This is an indication that a healthy lifestyle by means of regular exercise could delay or postpone the middle-age crisis in the eCB system.

Aging has been reported to alter binding of CB1 in the brain. In aged rats (>24 months), [³H]-WIN 55,212-2 specific binding of CB1 was shown to be decreased in most basal ganglia regions (entopeduncular nucleus, SN and CPu) other than globus pallidus [32]. The same group also analyzed CB1 receptor binding of [³H]-WIN 55,212-2 in the sub-regions of Hip and showed similar binding levels in both young and aged rats, except for the DG, where the binding was increased in aged rats. Further, CB1 binding was significantly decreased in the Hy, whereas similar binding levels were detected in the limbic structures (nucleus accumbens, septum nuclei, and basolateral amygdaloid nucleus) of young and aged rats [33]. In line with the above findings, [³H] CP55940 specific binding of CB1 was significantly decreased in the SN at the age of 18 months. Furthermore, CB1 binding in the dorsal Hip (CA1, CA2, CA3 and DG) and the ventral Hip was increased specifically at 9 months, but in LEnt, CB1 binding increased at 9, 12, and 18 months. However, unlike CB1 protein levels in the sub-regions of the dorsal Hip at 12 months, no prominent decrease in CB1 binding was detected at 12 months and later. The discrepancy between CB1 protein levels after 9 months of age as measured by immunofluorescence, and CB1 binding as measured by receptor autoradiography, may be due to the different experimental techniques. Immunofluorescence measures epitope of a receptor, whereas receptor binding measures the interaction between the receptor binding site and the ligand. Changes in binding levels reflect not only altered receptor numbers, but also changes in affinity (Kd). Indeed, CB1 is known to undergo positive allosteric modulation by endogenous anti-inflammatory lipids, such as lipoxin A4 [38,39]. As such, changes in CB1 binding detected in 9-month-old mice may reflect aging-dependent changes in the levels of endogenous allosteric modulators of CB1 which inadvertently would impact CB1 affinity and, hence, binding levels. Nonetheless, age-related changes in the receptor number (Bmax), but not in Kd were also reported for other receptor systems in the brain, such as dopamine D2/D3 receptors [40]. Hence, the changes in binding in our study most likely represent changes in receptor number rather than changes in affinity.

The density and signaling properties of CB1 vary based on the cell type in the Hip of adult mice. GABAergic neurons express higher amounts of CB1 receptors compared to glutamatergic neurons, but these CB1 receptors are less efficiently coupled to G-protein signaling than their glutamatergic counter parts [41]. Although we have not determined this in the current study, it is necessary to understand how cell-type specific CB1, its ligand-binding affinities, and downstream signaling are affected with age. As such, future work is warranted, focusing on the consequences of aging on CB1 function and signaling with the use of a combination of several techniques, including CB1 agonist [³⁵S]GTPγS binding, intracellular ⁺⁺Ca imaging, signaling, and electrophysiology.

In conclusion, our observations indicate that the eCB system is most affected during the middle age in a brain region-specific manner. The 2-AG levels were mostly affected in the Hip and moderately in CPu. While AEA levels were not altered in Hip, they were mainly affected in CPu as well as in mPFC+Cg. Although CB1 protein levels prominently decreased at 12 and 18 months of age in most sub-regions of the Hip, CB1 binding was unaltered at this age in these regions. Taken together, the middle-age crisis in the eCB signaling corresponds well with the onset of neuroinflammatory glial activity and cognitive deficits in mice. We now hypothesize that late middle-age is the time period when a therapy based on the activation of the cannabinoid system has the highest efficacy to prevent cognitive aging and pathologies related to brain aging.