Cognitive dysfunction and increased phosphorylated tau are associated with reduced O‐GlcNAc signaling in an aging mouse model of metabolic syndrome

Metabolic syndrome (MetS), characterized by hyperglycemia, obesity, and hyperlipidemia, can increase the risk of developing late‐onset dementia. Recent studies in patients and mouse models suggest a putative link between hyperphosphorylated tau, a component of Alzheimer's disease‐related dementia (ADRD) pathology, and cerebral glucose hypometabolism. Impaired glucose metabolism reduces glucose flux through the hexosamine metabolic pathway triggering attenuated O‐linked N‐acetylglucosamine (O‐GlcNAc) protein modification. The goal of the current study was to investigate the link between cognitive function, tau pathology, and O‐GlcNAc signaling in an aging mouse model of MetS, agouti KKAy+/−. Male and female C57BL/6, non‐agouti KKAy−/−, and agouti KKAy+/− mice were aged 12–18 months on standard chow diet. Body weight, blood glucose, total cholesterol, and triglyceride were measured to confirm the MetS phenotype. Cognition, sensorimotor function, and emotional reactivity were assessed for each genotype followed by plasma and brain tissue collection for biochemical and molecular analyses. Body weight, blood glucose, total cholesterol, and triglyceride levels were significantly elevated in agouti KKAy+/− mice versus C57BL/6 controls and non‐agouti KKAy−/−. Behaviorally, agouti KKAy+/− revealed impairments in sensorimotor and cognitive function versus age‐matched C57BL/6 and non‐agouti KKAy−/− mice. Immunoblotting demonstrated increased phosphorylated tau accompanied with reduced O‐GlcNAc protein expression in hippocampal‐associated dorsal midbrain of female agouti KKAy+/− versus C57BL/6 control mice. Together, these data demonstrate that impaired cognitive function and AD‐related pathology are associated with reduced O‐GlcNAc signaling in aging MetS KKAy+/− mice. Overall, our study suggests that interaction of tau pathology with O‐GlcNAc signaling may contribute to MetS‐induced cognitive dysfunction in aging.


| INTRODUC TI ON
Metabolic syndrome (MetS) is characterized by hyperglycemia, obesity, and hyperlipidemia and significantly increases the risk for developing cardiovascular disease and diabetes. In the central nervous system (CNS), these conditions are associated with increased risk of dementia (Bahchevanov et al., 2021;Foret et al., 2021;Lai et al., 2020;Wooten et al., 2019;Yuan & Wang, 2017). Indeed, individuals with MetS are approximately 1.5 times more prone to develop late-onset dementia, and MetS is shown to accelerate the progression from mild cognitive impairment to dementia (Peila et al., 2002;Solfrizzi et al., 2011). Epidemiological studies indicate that metabolic anomalies including diabetes, insulin resistance, obesity, and dyslipidemia are key risk factors for dementia and up to 70% of individuals diagnosed with Type 2 diabetes are predisposed to cerebrovascular changes and mild cognitive impairment, with a substantial number developing dementia at a later stage (Akter et al., 2011;Biessels et al., 2014;Dash, 2013;Leibson et al., 1997;Ott et al., 1999;Sperling et al., 2011). Moreover, those with Type 2 diabetes for a minimum of 5 years have an increased risk for Alzheimer's disease-related dementia (ADRD), the most common age-related dementia, compared to those with Type 2 diabetes for less than 5 years (Leibson et al., 1997). However, how CNS dementiaassociated pathology and mechanisms are related to MetS remains unclear.
Neurofibrillary tangles composed of hyperphosphorylated tau and amyloid-beta (Aβ) plaques are the pathological hallmarks of ADRD (Sperling et al., 2011). Type 2 diabetes and ADRD share several related pathological features including impaired glucose metabolism, increased oxidative stress, defective insulin signaling, inflammation, and protein accumulation (Chen & Zhong, 2013;Zhao & Townsend, 2009). Postmortem studies show accumulation of hyperphosphorylated tau and Aβ plaques in the brain of diabetic patients (Valente et al., 2010). Insulin dysfunction can also result in overproduction and/or reduced clearance of hyperphosphorylated tau within the brain (Hoyer, 2000;Lee et al., 2008;Shonesy et al., 2012;Stranahan et al., 2008;Zhao & Townsend, 2009). Tau pathology, in particular, is correlated with cognitive dysfunction in ADRD (Braskie & Thompson, 2013;Lace et al., 2009;Maass et al., 2018). Collectively, evidence in patients and mouse models of ADRD highlight a putative link between cerebral glucose hypometabolism and abnormal tau accumulation (Hoyer, 2000;Lo et al., 2011;Mehla et al., 2014;Mosconi et al., 2008). Impaired glucose metabolism can also lead to reduced glucose flux through the hexosamine metabolic pathway and attenuated O-linked N-acetylglucosamine (O-GlcNAc) modification. Attachment of O-GlcNAc to nucleocytoplasmic proteins, a regulatory mechanism of intracellular glucose signaling, is an important post-translational protein modification known to affect multiple cellular processes including signal transduction, nuclear translocation, transcription, mitochondrial metabolism, and apoptotic pathways (Hart, 2014;Hart et al., 2007;Zachara et al., 2004). O-GlcNAcylation, mediated by the two key enzymes O-GlcNAc transferase (OGT) and O-GlcNAcase (OGA), can occur in a cell-and tissue-specific manner such that increased O-GlcNAc signaling may be beneficial for some cells while detrimental to others, in turn modulating various pathological conditions (Graham et al., 2014;Jacobsen & Iverfeldt, 2011;Yuzwa et al., 2012). O-GlcNAcylation of intracellular proteins is associated with numerous metabolic perturbations including hyperglycemia, insulin resistance, obesity, and hyperlipidemia (Ansari & Emerald, 2019;Dai et al., 2018;Zhao et al., 2016). In the central nervous system, studies suggest that decreased O-GlcNAcylation can be detrimental whereas increased O-GlcNAcylation may be protective, particularly in ADRD models and stroke (Huang et al., 2020;Lee et al., 2021;Pinho et al., 2018;Wang et al., 2021;Zhu et al., 2014).
Consistent with these findings, it is suggested that reduced O-GlcNAcylation in the CNS may contribute to hyperphosphorylation of tau and impaired cognitive function (Di Domenico et al., 2019;Gong et al., 2016;Liu et al., 2009;Park et al., 2020). KKAy +/− mice. Overall, our study suggests that interaction of tau pathology with O-GlcNAc signaling may contribute to MetS-induced cognitive dysfunction in aging.

K E Y W O R D S
aging, agouti mouse, cognitive dysfunction, metabolic syndrome, neurodegenerative disease, O-GlcNAc, sex-specific differences, tauopathy Significance Metabolic syndrome, characterized by hyperglycemia, hyperlipidemia, and obesity, is associated with an increased risk of late-onset dementia. However, how metabolic alterations alter memory and brain function is unclear. In the present study, we show increased phosphorylation of the microtubule protein tau in the brain and impaired recognition memory in an aging mouse model of metabolic syndrome, the KKAy mouse. It is further shown that O-GlcNAc signaling is reduced in the brain at the same time as tau pathology and memory deficits. This work increases our understanding of how metabolic syndrome may contribute to age-related dementia.
While glucose deprivation (short-term fasting), impaired cerebral glucose utilization and availability, and dysregulated energy metabolism (as in diabetes and obesity) may contribute to cognitive decline and dementia-related neuropathology, animal studies using diabetic and obese rodent models show conflicting results on the role of tau in cognitive dysfunction in these models (Abbondante et al., 2014;Jolivalt et al., 2008;Latina et al., 2021;Ramos-Rodriguez et al., 2013;Trujillo-Estrada et al., 2019;Wang et al., 2014). For instance, elevated tau pathology was observed in the Streptozotocin (STZ)-induced rodent model of Type 1 diabetes, while both tau-independent and taudependent mechanisms of cognitive impairment have been reported in mouse models of Type 2 diabetes (El Khoury et al., 2016;Gratuze et al., 2017;Kim et al., 2009;Trujillo-Estrada et al., 2019). In addition, a majority of the studies have focused on adult mice up to ~6 months of age, but few examined the pathological effects of longer term chronic metabolic dysregulation in older mice. Consequently, a mechanistic understanding of how aging interacts with metabolic anomalies such as diabetes, obesity, and dyslipidemia in the context of ADRD remains unclear. Thus, the goal of the present study was to investigate the behavioral phenotype, O-GlcNAc signaling, and tau phosphorylation in an aging mouse model of MetS, the agouti KKAy +/− mouse.

| Mice
Mouse breeding, experimental, and euthanasia procedures were conducted according to protocols annually approved by the Institutional Animal Care and Use Committee (IACUC) and performed in the Comparative Medicine Unit at Northeast Ohio Medical University.
Male agouti KKAy +/− mice (KK-Ay/TaJcl, CLEA, Japan) were cross bred with female non-agouti KKAy −/− (a/a stock # 002468, The Jackson Laboratory) to generate male and female non-agouti KKAy −/− and agouti KKAy +/− mice. Genotypes were confirmed by visual inspection of coat color with yellow coat indicative of positive Ay + (agouti) allele transmission and black coat reflecting negative Ay − allele transmission. C57BL/6 control mice (stock # 000664) were purchased from Jackson Laboratories (Bar Harbor, ME, USA). All mice were housed in a temperature-controlled environment (71°F-75°F) with 12:12 h light-dark cycle at 30%-70% humidity and food and water available ad libitum. For multiple-housed animals, a maximum of four mice were maintained in individual cages with physical enrichment provided for all singly housed animals. Cages were cleaned on a monthly basis for all singly housed mice while all other cages were changed weekly.

| Experimental design
Male and female agouti KKAy +/− and non-agouti KKAy −/− mice were weaned at 4 weeks of age, maintained on standard laboratory diet, and aged 12-18 months. All mice underwent periodic body weight and blood glucose monitoring conducted once a month up to 6 months; random blood glucose samples collected via lateral tail incision were used for glucose estimation using a One-touch glucometer. At 12-18 months of age, C57BL/6 control, non-agouti KKAy −/− , and agouti KKAy +/− mice were weighed again and subjected to a battery of behavioral tests described below. Following these tests, mice were euthanized with fatal plus per institutionally approved procedures, and blood and brain tissue samples were collected and utilized as described below.

| Plasma glucose and lipid measurement
After overnight fasting, blood was collected via cardiac puncture; plasma was isolated, aliquoted, and stored at −80°C for further analyses. Plasma glucose concentration was measured in duplicate in each mouse using a glucose colorimetric assay kit (Sigma), per manufacturer's instructions. Similarly, total cholesterol and total triglyceride concentrations were measured in duplicate in each mouse using Infinity Reagents (ThermoFisher), according to manufacturer's instructions.

Challenging beam traversal
The challenging beam traversal test was used to analyze motor performance and coordination. The test was performed as previously described (Fleming et al., 2004). The beam consisted of four sections connected linearly (25 cm each, 1 m total length) and was constructed from Plexiglas. Each section was a different width, the beam started at a width of 3.5 cm and gradually narrowed to .5 cm × 1 cm increments. Animals were trained to traverse the entire length of the beam starting at the widest section and ending at the narrowest section. The narrow end of the beam led directly into the animal's home cage. Before the day of the test, animals received 2 days of training consisting of five trials per animal. On the day of the test, a mesh grid (1 cm squares) of corresponding width was placed over the beam surface, leaving a 1 cm space between the grid and the beam surface. Animals were then recorded while traversing the gridcovered beam for a total of five trials. A rater, blind to genotype, viewed the videos in slow motion and scored for errors, number of steps made by each animal, and time taken to traverse the entire beam. An error was counted when a limb slipped through the grid during a forward movement and was visible between the mesh grid and the beam surface. Each limb slip was scored individually, and slips were only counted during forward movement. Average error per step, time to traverse, and number of steps were determined for all genotypes over the five trials.

Spontaneous activity
Spontaneous movements of the mice were measured in a small, transparent cylinder 15.5 cm high and 12.7 cm in diameter (Fleming et al., 2004). The cylinder was placed on a piece of glass with a mirror positioned at an angle beneath the cylinder to allow a clear view of movements along the floor and walls of the cylinder. Spontaneous movements were recorded for 3 min. Videotapes were viewed and rated in slow motion by an experimenter blind to mouse genotype.
The number of rears, forelimb and hindlimb steps, and time spent grooming were measured for each mouse.

Elevated plus maze
The elevated plus maze is a test of emotional reactivity and used to detect anxiety-related behavior in rodents. The maze consisted of a central platform (5 × 5 cm) from which two open arms and two enclosed arms (height = 15 cm) extended outward. The four arms form the shape of a plus sign and the maze has a height of 38 cm. The floor of each arm was lined with lightly textured contact paper, in order to increase grip and encourage the animals to explore the arms. Testing was conducted and recorded under red light. The test began when an animal was placed on the center platform and allowed to explore the maze for 5 min. If a mouse fell off the maze, it was replaced in the same location from which it fell. After testing, an experimenter blind to animal genotype measured the amount of time each animal spent in the open arms, closed arms, and center of the maze from the recordings. A mouse was considered to be in a section of the maze when all four limbs were in the section.

Object recognition
An object recognition test was used to measure attention and memory. The test was performed as previously described (Brown et al., 2010;Mittal et al., 2020). Prior to testing, the animals were habituated to the test bin once a day for 2 days, 15 min/day. On the test day, each animal received a sample trial and a test trial. The sample trial was conducted by placing the animal in the test bin with two identical objects at one end of the bin. The animal was then allowed to explore and investigate the objects for 10 min. After the sample trial, the animals were returned to their home cage for 60 min. For the test trial, one of the objects from the sample trial was replaced with a novel object, which differed in texture and shape but was of similar size. The animal was placed in the bin and allowed to explore and investigate the objects for 10 min. Each trial was recorded and later scored by an experimenter blind to genotype. The number of rears and the amount of time an animal spent investigating each object was measured.

| Immunoblotting
Following the behavioral tests, brains were harvested from each mouse and stored at −80°C for subsequent analyses. Frontal cortex and hippocampal-associated dorsal midbrain regions were dissected on ice and lysates were prepared using T-PER™ tissue protein extraction reagent (ThermoFisher) containing a protease and phosphatase inhibitor cocktail (Thermo Scientific). Protein content was determined using the BCA protein assay (BioRad). Equal quantities of protein lysates (20-40 μg) were resolved on 8%-10% SDS-PAGE gel and transferred to Immobilon-P Polyvinyl difluoride membrane.
Immunoblotting was then performed using the following antibodies (see Table 1

| Statistical analyses
Each dataset was assessed for normality and homoscedasticity.
Analysis of the MetS phenotype from 4 to 24 weeks was analyzed using a 2 × 7 mixed design ANOVA followed by Tukey's HSD, or Student's t-test. For body weights at 12-18 m, a randomized one-way ANOVA followed by Tukey's HSD was used to compare genotypes at the time of behavioral testing. Because the agouti KKAy +/− mice are obese and weigh significantly more than the other genotypes, this can influence performance in the behavioral tests. Therefore, analysis of covariance (ANCOVA) was used to partial out weight differences among genotypes on the challenging beam, spontaneous activity, and object recognition tests. The ANCOVA yields genotype means and error variance that are corrected for any influence in the tests due to weight differences. For the elevated plus maze, nonparametric analyses were used since animals may spend all or no time in a region of the maze. For tissue analysis, data were separated by sex and comparisons made between C57BL/6, non-agouti KKAy −/− , and agouti KKAy +/− mice. Either a randomized one-way ANOVA was used to compare genotypes followed by Tukey HSD post hoc test or if the data did not meet the assumptions for parametric statistics, then the Kruskal-Wallis nonparametric test was used. For normally distributed data with unequal variance, Welch ANOVA was applied followed by Dunnett's post hoc test. Results are expressed as Mean ± standard deviation (SD). Differences were considered statistically significant at p ≤ .05. All statistical analyses were performed using either GraphPad Prism version 9 or Matlab 2021A, Mathworks Inc.

| Agouti KKAy +/− mice exhibit increased body weight, plasma glucose, and lipid levels
Periodic monitoring of body weight and random blood glucose in younger (4-24 weeks) non-agouti KKAy −/− and agouti KKAy +/− mice combined with measurement of cholesterol and triglyceride levels confirmed the MetS phenotype in these animals (Figure 1a-d).
Specifically, male and female agouti KKAy +/− mice begin to weigh more than non-agouti KKAy −/− mice at approximately 7-10 weeks of age and that difference was maintained throughout aging Together, these results confirm the MetS phenotype in the agouti TA B L E 1 Antibodies used in the study.

Name
Immunogen Details  Figure S1a). Fasting plasma samples were utilized for glucose and lipid measurements.

| Emotional reactivity
The elevated plus maze data were analyzed using Kruskal-Wallis to compare multiple independent groups. In the maze, there were no significant differences between genotypes in the time spent in the closed, open, or center sections of the maze (Table 2). This indicates that neither KKAy genotypes develop an enhanced fear response.

| Object recognition
During the sample trial of the test, mice were presented with two novel identical objects and allowed to explore and investigate for 10 min. In this part of the test, ANCOVA revealed both non-agouti KKAy −/− and agouti KKAy +/− mice spent less time investigating the objects compared to C57BL/6 control mice (F (2, 33) = 14.61, p < .05; Figure 4). After the sample trial, mice were then placed back in the home cage for 1 h. For the test trial, mice were exposed to one object from the sample trial (familiar object) and a novel object. Here, ANCOVA showed that while both non-agouti KKAy −/− and agouti KKAy +/− mice investigated the novel object less than C57BL/6 controls (F (2, 33) = 17.46, p < .05), only the KKAy +/− mice displayed a significant reduction in the discrimination index which takes into account both novel and familiar investigation times (t (33) = 1.81, p < .05, one tail). These data indicate that both non-agouti KKAy −/− and agouti KKAy +/− mice display impaired attention, and agouti KKAy +/− develop more severe discrimination defects. While the study was not powered to directly compare males and females, when separated by sex, agouti KKAy +/− females in general showed more robust impairments in investigation time, a measure of attention ( Figure S2). and errors per step were measured over five trials and averaged for each mouse. Analysis of covariance was used to partial out weight differences, *p < .05 compared to C57BL/6 or non-agouti KKAy −/− mice. Figure 5). Together, our results suggest that MetS in older mice leads to sex-specific pathological changes in tau phosphorylation.

| Phosphorylated ERK and GSK3β
Next, to delineate a potential mechanism underlying increased tau phosphorylation, expression profiles of specific kinases (ERK, GSK3β) known to be involved in tau phosphorylation were measured in the brain. Phosphorylated ERK/total ERK (pERK/ tERK) and phosphorylated GSK3β/total-GSK3β (pGSK3β/tGSK3β) were measured in each genotype for each sex. In hippocampalassociated tissue in females, there was no difference in pERK/ tERK between the genotypes (Figure 6b: F (2, 16) = 1.293, p < .05).
In contrast, in male mice, pERK/tERK was significantly increased (a-d) Forelimb and hindlimb steps, rears, and time spent grooming were measured over 3 min for each mouse. Analysis of covariance was used to partial out weight differences; *p < .05, **p < .01 compared to C57BL/6 or non-agouti KKAy −/− mice.

| OGT and O-GlcNAc
We then examined the O-GlcNAc signaling pattern in brain lysates.

| DISCUSS ION
The World Health Organization reports that worldwide over 55 million people suffer from dementia and that almost 10 million new cases are diagnosed each year highlighting its profound impact on all populations (World Health Organization). In the aging population with co-existing metabolic anomalies, risk of dementia is even more amplified (Biessels & Reagan, 2015). To understand how MetS may contribute to cognitive dysfunction and decline, models that The present study utilized the well-established non-agouti KKAy −/− and agouti KKAy +/− mouse models of metabolic dysfunction and metabolic dysfunction with obesity, respectively (Chakraborty et al., 2009;Nakamura & Yamada, 1967;Suto et al., 1998;Tomino, 2012). In the current study, older KKAy −/− and KKAy +/− mice showed impairments in recognition memory and attention compared to C57BL/6 controls after weight was factored out using ANCOVA.
Of note, it was important to consider weight because only the obese agouti KKAy +/− mice displayed significant deficits in sensorimotor function which could negatively impact performance in other tests.
In addition, while the study was not powered to directly compare the sexes, when separated by sex, only females displayed significant differences in attention and memory dysfunction. Clinically, sex differences in cognitive dysfunction in Type 2 diabetes have also been reported. Indeed, a recent study showed that females with Type 2 diabetes had an increased risk for accelerated cognitive decline compared to diabetic males (Verhagen et al., 2022). Earlier studies also show behavioral anomalies in KKAy +/− mice including impairments in passive avoidance and spatial memory primarily using younger mice at 3-7 months of age (Huang et al., 2020;Li et al., 2018;Min et al., 2012;Sakata et al., 2010Sakata et al., , 2012Shi et al., 2020;Tsukuda et al., 2008). Furthermore, sex differences have been detected, with agouti KKAy +/− females displaying more severe cognitive impairments than agouti KKAy +/− males in these tests (Huang et al., 2020;Li et al., 2018;Min et al., 2012;Sakata et al., 2010Sakata et al., , 2012Tsukuda et al., 2008). However, weight was not factored in the analysis in these studies and KK mice without obesity (i.e., KKAy −/− mice) were not included. The findings in this study support previous work showing cognitive dysfunction in KKAy +/− mice and sex-specific effects of MetS on cognitive function. This underscores the importance of considering weight, sex, and age in preclinical and clinical studies on the relationship between metabolic syndrome and cognitive dysfunction.
Tau is a microtubule-associated protein involved in the cytoskeletal network and the stabilization of microtubules along axons.
Metabolic dysregulation involving hyperglycemia, obesity, insulin resistance, and dyslipidemia are risk factors for tauopathy, neurodegenerative diseases characterized by the accumulation of phosphorylated tau aggregates. Hyperphosphorylated tau in the form of tangles is one of the pathological hallmarks of ADRD and is correlated with cognitive dysfunction (Braskie & Thompson, 2013;Lace et al., 2009;Maass et al., 2018). In the present study, pTau (Ser202) and pTau (Thr231) were significantly increased in the frontal cortex and only pTau (Ser202) was increased in hippocampal-associated tis- Increased tau phosphorylation in the brain has been observed in the STZ model of Type 1 diabetes and the db/db mouse model of Type 2 diabetes supporting a putative link between ADRD pathology and MetS (Clodfelder-Miller et al., 2006;El Khoury et al., 2016;Jolivalt et al., 2008;Ke et al., 2009;Kim et al., 2009;Li et al., 2012;Planel et al., 2007). However, the mechanisms underlying increased phosphorylated tau in Type 1 and Type 2 diabetes may differ. For example, STZ-induced diabetes in humanized tau mice exacerbated cognitive impairments and genetic deletion of tau reduced cognitive and synaptic impairments induced by STZ (Abbondante et al., 2014;Trujillo-Estrada et al., 2019), while overexpression of humanized tau showed no effect in db/db mice (Trujillo-Estrada et al., 2019).
Additionally, an increase in cleaved tau in untreated db/db mice has been observed but not in STZ-injected mice (Kim et al., 2009). Taken together, the data in the current study indicate metabolic dysregulation and obesity lead to increased phosphorylated tau in multiple brain regions in older KKAy +/− mice.
Abnormal phosphorylation of tau and tau aggregation can promote neuronal microtubule destabilization and neuronal dysfunction (Barbier et al., 2019;Wang et al., 2013). Several different kinases including GSK3β, MAPK, cdk-5, PKA, and MARKs are known to regulate tau phosphorylation. Studies suggest that an imbalance between tau kinases and phosphatases plays a central role in the development of tau hyperphosphorylation (Martin et al., 2013;Stoothoff & Johnson, 2005). Disruption of microtubule assembly and tau aggregation is facilitated by increased phosphorylation in proline-rich regions of tau, with GSK3β recognized as a key prolinedirected tau kinase (Lauretti et al., 2020;Sayas & Avila, 2021). In vitro and in vivo studies using pharmacological GSK3β inhibitors and transgenic mouse lines show that increased GSK3β activity results in augmented tau phosphorylation, hippocampal degeneration, and learning deficits. Conversely, normalizing GSK3β activity is reported to reduce tau phosphorylation and neurodegeneration and reverse memory deficits (Ding et al., 2010;Engel et al., 2006;Giese, 2009;Gomez de Barreda et al., 2010;Onishi et al., 2011;Perez et al., 2002;Rodriguez-Matellan et al., 2020;Sereno et al., 2009). Similar findings linking tau and GSK3β have also been reported in the STZ model of Type 1 diabetes, but the link between tau and GSK3β is not as clear in the db/db model (Abbondante et al., 2014;Clodfelder-Miller et al., 2005;El Khoury et al., 2016;Jolivalt et al., 2008;Krishnankutty et al., 2017;Latina et al., 2021;Ponce-Lopez et al., 2017;Trujillo-Estrada et al., 2019). In the present study, despite augmented tau phosphorylation, female agouti mice did not reveal reduced p-GSK3β expression, indicative of GSK3β activation. Accordingly, these data support earlier findings suggesting a GSK3β-independent mechanism underlying tau hyperphosphorylation and cognitive impairment in aging female agouti genotypes.
Earlier work on the role of MAPK signaling in tau pathology suggests that increased ERK activation may also contribute to tau hyperphosphorylation and aggregation. While these findings have been confirmed in vitro, most in vivo studies show inconsistent results.
In the present study, tau hyperphosphorylation in female KKAy +/− was not accompanied by increased pERK expression, indicating an ERK-independent mechanism responsible for increased tau phosphorylation in these genotypes (Cai et al., 2011;Guise et al., 2001;Harris et al., 2004;Noel et al., 2015). Interestingly, ERK1/2 activation was profoundly attenuated in the frontal cortex of aging female agouti KKAy +/− mice compared with age-matched C57BL/6 controls.
Emerging literature suggests that alterations in ERK signaling in the prefrontal cortex may be linked to depression and cognitive dysfunction (Dwivedi et al., 2009;Hart & Balleine, 2016;Leem et al., 2014;Wang & Mao, 2019). Accordingly, the reduced ERK1/2 phosphorylation in the frontal cortex of female KKAy +/− genotypes may contribute to the observed impairments in attention and memory.
While substantial data supporting a reciprocal link between protein O-GlcNAcylation and phosphorylation abound in the current literature, the role of O-GlcNAc signaling in regulation of cognitive and neuronal function in the aging brain has been rather controversial, with O-GlcNAc levels reported to be both increased and decreased in the ADRD brain (Forster et al., 2014;Griffith & Schmitz, 1995;Hwang & Rhim, 2019;Taylor et al., 2014;Wang et al., 2021). Overall, the present results suggest a potential protective role of O-GlcNAc signaling on ADRD-related pathology and cognitive deficit induced by metabolic disorders during aging.
In conclusion, the current study provides compelling evidence for an interaction of aging with MetS in the context of tau pathology.
Importantly, the results suggest an association between abnormal tau phosphorylation, cognitive impairment, and reduced O-GlcNAc signaling patterns in aging MetS mice, that is specific to the female genotype.

DECLARATION OF TRANSPARENCY
The authors, reviewers and editors affirm that in accordance to the policies set by the Journal of Neuroscience Research, this manuscript presents an accurate and transparent account of the study being reported and that all critical details describing the methods and results are present.

ACK N OWLED G M ENTS
We would like to thank Sophia Scott for her assistance with manuscript preparation and literature review. We would like to acknowledge support from The Neurodegenerative Disease and Aging Research Focus Area at NEOMED and NIH ES031124-01 to SMF; NEOMED Research Funds and NIH R56HL141409-01 to PR.

CO N FLI C T O F I NTER E S T S TATEM ENT
The authors disclose no conflict of interest.

PEER R E V I E W
The peer review history for this article is available at https:// www.webof scien ce.com/api/gatew ay/wos/peer-revie w/10.1002/ jnr.25196.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are available from the corresponding author upon reasonable request. Two-way ANOVA to account for sex and genotype, Tukey's HSD post hoc; *p ≤ .05, ***p < .0005, ****p < .0001 compared to C57BL/6 mice.

FIGURE S7
Uncropped full blots for female frontal cortex GSK and male hippocampus ERK immunoblotting.