Circadian rhythms regulate many physiological and behavioral processes, including the formation of memory (Davies et al., 1973; Eckel-Mahan et al., 2008). Aging notably impairs both circadian rhythms (Renfrew et al., 1987; Weitzman et al., 1982; Witting et al., 1990) and memory (Craik and McDowd, 1987; Youngjohn and Crook, 1993), though the molecular mechanisms underlying these impairments are not well understood, nor is the relationship between them fully characterized (e.g., if age-related memory impairments are caused by age-related circadian deficits).

In mammals, circadian rhythms are generated chiefly by the body's central pacemaker, the suprachiasmatic nucleus (SCN), which is found at the base of the hypothalamus and synchronizes satellite oscillators throughout the body. Circadian oscillators, including the SCN, contain molecular clocks largely driven by a transcription-translation feedback loop (TTFL) with four critical components: genes Circadian Locomotor Output Cycles Kaput (Clock) and Brain and Muscle ARNT-like 1 (Bmal1) and gene families Period (Per) and Cryptochrome (Cry). In brief, CLOCK and BMAL heterodimerize and promote transcription of Cry and Per genes. After translation in the cytoplasm, CRY and PER proteins enter the nucleus and inhibit their own transcription, creating a negative feedback loop that lasts approximately 24 hours. The molecular clock within the SCN entrains other oscillators located throughout the rest of the brain and body, however it is worth noting that many of these peripheral oscillators cycle in the absence of SCN input (Yamazaki et al., 2000; Yoo et al., 2004), although many report that the hippocampus does not (Abe et al., 2002; Phan et al., 2011; but see also Chaudhury et al., 2005). Most of the existing research on clock genes focuses on their activity within the SCN, but these genes are also present across both the nervous system (Abe et al., 2002) and other tissues (Plautz et al., 1997). The function of clock genes outside of the TTFL is not yet fully characterized.

Recent reports have suggested that clock genes might operate in satellite brain regions (i.e., outside the SCN) to control local functions, including learning and memory (Snider et al., 2016; Woodruff et al., 2018; reviewed by Smies et al., 2022) and other behavior (McClung et al., 2005; Mukherjee et al., 2010; Spencer et al., 2013) in a brain region–dependent manner. Previous work has demonstrated that one such clock gene, Period1 (Per1) modulates the phosphorylation of cAMP response element binding protein (CREB) in the dorsal hippocampus (DH; Rawashdeh et al., 2016), suggesting a possible role for Per1 in hippocampal memory, given CREB's known role as a memory modulator (Yin et al., 1994, 1995). We recently demonstrated that bidirectional, local manipulations of Per1 in the DH (Bellfy et al., 2022; Kwapis et al., 2018) and retrosplenial cortex (RSC; Urban et al., 2021) are sufficient to affect spatial memory. Although we reported that Per1 expression in the DH is impaired by aging (Kwapis et al., 2018), no one has yet investigated how Per1 expression in memory-relevant neocortical regions (e.g., the RSC) changes as a result of age and whether these changes are linked to memory performance.

The RSC (Brodmann areas 29 and 30) is a cortical brain structure integral to spatial memory and critically affected by age. This region—located immediately posterior to the corpus callosum in primates and directly dorsal to the DH in rodents—is densely connected with the DH and both the prefrontal and cingulate cortices (Shibata et al., 2004; Wang et al., 2016). Notably, expression of the immediate early gene (IEG) Fos is induced in the RSC following spatial learning (Vann et al., 2000), and lesions of the rodent RSC impair performance in both the Morris water maze (Harker and Whishaw, 2002) and radial arm maze (Vann and Aggleton, 2004). Further studies have suggested that RSC lesions specifically impair the ability to bind complex stimuli together, rather than the act of navigation itself (Nelson et al., 2015). Additionally, the RSC has been demonstrated to play an important role in both the retrieval (Corcoran et al., 2011) and formation (Kwapis et al., 2015; Urban et al., 2021) of contextual fear memory. Both the anterior RSC (aRSC) and the posterior RSC (pRSC) have been implicated in spatial memory, although these subregions may each support different types of information (see Discussion). Interestingly, both contextual fear memory (Moyer Jr. and Brown, 2006) and object location memory (Wimmer et al., 2012) are impaired as a result of aging, but RSC-independent modalities like delay fear conditioning (Trask et al., 2021a) and semantic memory (Wiggs et al., 1998) are resistant to the effects of aging (Craik and Grady, 2002; Moyer Jr. and Brown, 2006; reviewed by Trask and Fournier, 2022). Thus, the RSC is tightly linked to age-related memory deficits, and further investigation may elucidate the link between these memory deficits and age-related disruptions in circadian rhythms.

Here, we report that dysregulation of Per1 within the RSC contributes to age-related memory deficits. We find that although learning induces Per1 expression in the RSC of aging mice, this induction is smaller than in young mice. Additionally, this Per1 induction fluctuates across the day-night cycle in both young and aging animals, although aging induces minor disruptions of this pattern. Furthermore, local downregulation of RSC Per1 in young mice impairs the formation of spatial memory, while local upregulation of Per1 in the RSC of aging mice is sufficient to rescue memory formation. Finally, although learning induces gene expression in the SCN of young but not aging mice, locally downregulating Per1 in the SCN of young mice has no effect on spatial memory. Together, these data indicate that Per1 functions within the RSC to modulate memory, whereas Per1 in the SCN is dispensable for normal memory formation in young mice. Age-related repression of Per1 in the RSC may therefore contribute to age-related spatial memory deficits. Although we did not directly test the role of Per1 as a circadian regulator, these results further elucidate the relationships between clock genes, aging, and learning.