A high fat, sugar, and salt Western diet induces motor‐muscular and sensory dysfunctions and neurodegeneration in mice during aging: Ameliorative action of metformin

Abstract Aims To explore the novel linkage between a Western diet combining high saturated fat, sugar, and salt (HFSS) and neurological dysfunctions during aging as well as Metformin intervention, we assessed cerebral cortex abnormalities associated with sensory and motor dysfunctions and cellular and molecular insights in brains using HFSS‐fed mice during aging. We also explored the effect of Metformin treatment on these mice. Methods C57BL/6 mice were fed with HFSS and treated with metformin from 20 to 22 months of age, resembling human aging from 56 to 68 years of age (an entry phase of the aged portion of lifespan). Results The motor and sensory cortexes in mice during aging after HFSS diet showed: (A) decreased motor‐muscular and sensory functions; (B) reduced inflammation‐resolving Arg‐1+ microglia; (C) increased inflammatory iNOs+ microglia and TNFα levels; (D) enhanced abundance of amyloid‐β peptide and of phosphorylated Tau. Metformin attenuated these changes. Conclusion A HFSS‐combined diet caused motor‐muscular and sensory dysfunctions, neuroinflammation, and neurodegeneration, whereas metformin counteracted these effects. Our findings show neuroinflammatory consequences of a HFSS diet in aging. Metformin curbs the HFSS‐related neuroinflammation eliciting neuroprotection.

here the effects of a HFSS diet on motor-muscular and sensory behaviors, as well as on neuronal loss by Western blotting and by immunohistological measurement of the levels of the M1-like inflammatory and M2-like inflammation-resolving phenotypes of microglia cells in the cerebral cortex. [32][33][34][35][36] We also explored the effects of HFSS on the accumulation of phosphorylated tau (pTau) and beta-amyloid (Aβ), which are the biomarkers and postulated causes of AD pathogenesis. 4,16,18,33,[37][38][39][40][41] In addition, we determined whether metformin treatment would ameliorate any deleterious effects of HFSS diet.
Peripheral sensory behavior is controlled by the sensory cortex, while the motor cortex is critical for controlling muscle reflexes and strength. 42, 43 We used a novel approach that incorporated consumption of HFSS by C57BL/6 mice from 20 to 22 months of age.
This age approximates human aging from 56 to 68 years of age, [44][45][46][47] an apparently critical entry phase of the aged portion of the human lifespan that is likely to be important in the neurodegeneration associated with the combination of HFSS and aging. The results of this study provide mechanistic insights into microglial phenotypes and suggest the intervention using metformin to curb the onset of neurodegenerative diseases.

| Animals
The Louisiana State University Health Science Center (LSUHSC) IACUC committee approved all the animal procedures. C57BL/6 (B6) mice were maintained at LSUHSC at a controlled temperature of 25 ± 2°C and 50%-65% humidity with a fixed 12:12 h light-dark cycle. The animal data report has followed the ARRIVE guidelines. 48

| HFSS Western diet and metformin treatment
Twenty-month-old male B6 mice were randomly separated into salt water drink (1% NaCl for 7 weeks, followed by 4% salt for 4 days and then 3% salt for a week). The ND consisted of 10% kcal fat, 73% kcal carbohydrate (corn starch, 82.2% plus maltodextrin-10%, 17.8%, by weight), 17% kcal protein, and 0.25% salt, and tap-water drink contained <0.02% salt. Our design of the HFSS diet was based on previous studies on Western diets, although all reported studies focused on the single or double high feature of fat, simple sugar, and salt. 4,18,[37][38][39]49,50 All the mice were separated into two groups and fed either the ND or HFSS diet for two months. During the second month of the diet, one set of mice from the HFSS group was treated with metformin (Sigma, 100 mg/kg body weight, i.p., every two days for a month) and designated as the HFSS + Met group.

| Loaded grid test
This test is based on the tendency of a mouse, when suspended by the tail, to grasp a grid with its forelimbs. 51 Here, weights of specific mass are hung from the grid, and the total time the mouse held a given weight was used as an assessment of motor-muscular strength.
Mice were habituated for 60 min before the tests. Prior to the test, the body weight of the mouse was recorded to allow normalization of the body weight. The mouse was held by its tail and allowed to grasp the weighted grid. Trials were provided with different weight measures. A resting period of 5 min was provided for each mouse between each series of pulls to allow recovery and to avoid habit formation.

| Hot water tail flick test
This test evaluates the animal's sensitivity to a hot temperature. 52 The mice were habituated to the experimental setting for 60 min. A 500 ml glass beaker filled with 450 ml distilled water was warmed on a hot plate to the specified temperatures (ie, 42 and 55°C). The water temperature was monitored with a glass thermometer and maintained throughout the experiment. Approximately 5 min before the test, the mouse was removed from the cage and allowed to crawl into a precut 50 ml tube. The tail was marked with waterproof ink 3 cm from the distal tip. The tube holding the mouse was held in a horizontal position to allow the distal 3 cm of the tail to be submerged in the preheated water bath. The tail submersion time was limited to 30 s. The flicking movement of the tail out of the water was recorded as the tail flick time and recorded accordingly. Each mouse underwent four trials, with a 10 min interval between each trial.

| Tissue collection and processing for immunohistochemistry
After behavioral assessments, mice were anesthetized with ketamine and xylazine and then sacrificed through decapitation. One cerebral hemisphere of each mouse was preserved in 4% paraformaldehyde and then cryoprotected in phosphate-buffered sucrose gradients (10%, 20%, and 30%) at 4°C until the tissue settled at the bottom. After processing, coronal sections of the motor and sensory cortex were sequentially cut at a thickness of 15 µm using a cryostat (Shandon Cryotome SME). The sections were collected serially on Superfrost slides (Fisher Scientific) and stored at −20°C until used for immunohistological studies. 53

| Cell quantification and analysis
Cell quantification was done on sections stained with the standard immunofluorescence method. [53][54][55][56][57][58][59] The cell populations per microscope field were estimated in the motor and sensory cortexes. Equal areas were assessed in every section by applying an 1162.5 µm frame encompassing the anatomical areas of interest to ensure that counts were representative of the analyzed areas. Microimages of the motor and sensory cortex regions located between bregma 0.74 mm and 0.02 mm, respectively, 60 were photographed with a Zeiss LSM710 confocal microscope. Only tissue sections corresponding to those coordinates were included in the quantification to ensure that the regions of interest were equivalent among animals and experiments. Co-localization and cell counts were analyzed on 20× and 40x magnification images. Two proteins (one stained green and the other stained red) were considered as co-localized at the section sites that displayed yellow from the superimposition of the red and green light. Nuclei were stained blue with DAPI. Cell population percentages and levels of specific proteins (labeling intensity integrated) were analyzed on the digital images using the NIH Image J software and specific plugins (http://rsb.info.nih.gov/ij/downl oad. html). 54,55 Iba-1 + microglia that were Arg-1 + or iNOs + were counted under the confocal microscope. The measurement settings for every sample were kept the same throughout the quantification.

| Western blotting analysis
The other hemisphere of the brain was snap-frozen, homogenized, and lysed in buffer containing protease and phosphatase inhibitors (ThermoFisher) and analyzed by Western blotting, as described. 61,62 In brief, the lysates were centrifuged at 14,000 rpm for 25 min at

| Statistical analysis
Data were analyzed by Kruskal-Wallis (nonparametric) test followed by Dunn test for the assessment of significant differences between groups in case data that do not have normal distribution. 63,64 The pvalues were adjusted by Holm method. Data were reported as dots and lines (medians) in figures or as medians in text. The correlations between specific microglia abundancies and Aβ or pTau levels in the motor or sensory cortex were analyzed via Spearman's nonparametric method. 63,64 RStudio (Version 1.4) was used. A p-value ≤0.05 was considered statistically significant.

| HFSS decreases motor-muscle grip strength
This strength measured by loaded grid test was used to assess the motor-muscle functions 51 that was hypothesized to be affected by HFSS. The motor-muscle grip strength was lower in HFSS-fed than in age-matched ND-fed mice ( Figure 1A) when lifting a 30 g (p = 0.0465) weight. No significant difference was observed for lifting a 10, 20, or 40 g weight (data not shown). Met treatment did not show effect on the muscle strength in HFSS-fed mice ( Figure 1A).

| HFSS reduces heat sensitivity
We tested the prediction that HFSS attenuates the sensory function by measuring the heat sensitivity of mice. The tail flick test with 55°C water revealed a significantly reduced heat sensitivity of the HFSSfed mice compared with age-matched ND-fed mice (p = 0.0230), as the HFSS-fed mice took a longer time to respond to the heated water ( Figure 1B). Met treatment improved the heat sensitivity behavior in HFSS-fed mice compared with the untreated HFSS-fed mice (p = 0.0201), indicating a protective effect of the drug. No differences were observed with the 42°C water test (data not shown).

| Metformin enhances cerebral neuronal density in the aging mice fed with HFSS diet
The neuroprotective action of Met was expected to counteract the neurodegenerative potential of HFSS, thus was explored. The rela-
These data indicated that the relative abundancy of inflammationresolving, reparative Arg-1 + microglia was reduced in the motor and sensory cortexes by HFSS, but the Met treatment curbed this decline.
Another protein biomarker that is used to characterize microglia is inducible nitric oxide synthase (iNOs), which is viewed as a marker for the M1-like (classical activation/mediator of inflammatory F I G U R E 1 Chronic HFSS consumption and Met on the motormuscular grip strength and sensory behavior of mice during aging. A, HFSS-fed mice had a significant reduction in their motormuscular grip strength (lifting a 30 gram weight) compared with ND-fed age-matched mice. B, tail flick test. HFSS consumption reduced the mouse heat sensitivity at 55°C compared with the ND control, as mice took more time to respond to the hot temperature. Following Met treatment, the HFSS-fed mice showed improvement in their sensitivity. Data are displayed as dots and lines (medians). p ≤ 0.05 is significant (n = 4 mice) response) phenotype. 33 The HFSS-fed mice showed an increase in iNOs-colabeled microglial cells ( Figure 4A

| Aβ deposition along the Col-IV-labeled blood vessels in motor and sensory cortex
β-amyloid (Aβ) deposition along neurovascular units is another characteristic of AD, in addition to the excessive expression of pTau. [66][67][68] The co-localization of some Aβ (green) and Col-VI + blood vessels (red) has been revealed as yellow regions in the brain cortexes ( Figure 6 F I G U R E 3 HFSS diet decreased Arg-1-colabeled Iba-1 + microglia in the motor and sensory cortexes, whereas Met reversed this trend. A, Representative microimages of Arg-1-colabeled Iba-1 + microglia in the motor and sensory cortexes in brain cryosections of aging B6 mice fed with ND or HFSS and treated with or without Met. B, Abundancy, M2-like Arg-1 + Iba-1 + microglia /total Iba-1 + microglia (Arg-1 + Iba-1 + /Iba-1 + microglia), % in the microscope field of the motor and sensory cortexes. Arrows annotate microglial cells with reprehensive morphology. The magnified images at the most right column of "A" show the morphology and expression of Arg-1 (yellow) of Iba-1 + microglia. Data are presented as dots and lines (medians). p ≤ 0.05 is significant (n = 4 mice). Scale bar =50 µm (p = 0.1413) group ( Figure 6B lower-left). No difference was observed in the Col-IV + intensity sum per microscopic field between groups ( Figure 6A, quantitative data not shown).

| DISCUSS ION
This study demonstrates that HFSS fosters neurological dysfunctions in the aged mice on cerebral cortexes in motor-muscular strength and sensory behavioral functions, and neurodegenerative and microglial changes during aging. Our focus on HFSS has arisen from findings that Western diets having a single or double high feature of fat, sugar, and salt promulgate some neurodegenerationrelated characteristics. 4,18,[37][38][39]49,50 Some of these effects are more or less visible in rodents fed the diet for 2-8 months, depending on the exact diet, animal strain, and age. 4,[69][70][71] However, the real-world Western diets are usually high in all three components (fat, sugar, and salt) and are not well studied. This report has initially addressed the knowledge gap on the potential consequences of a Western diet combining HFSS on motor and sensory cortexes and related neurological outcomes. It also has provided a unique protocol of 2 months feeding, a relatively short duration compared to other similar studies in the field. Aging impairs the above functions and relevant cells and molecules, 72-75 thereby contributing to the outcomes observed in these old mice as they age from 20 to 22 months of age. Notably, this HFSS diet did not result in a significant difference in body weights ( Figure S2) or blood glucose compared to the ND diet, suggesting that obesity or diabetes could not account for the neurodegenerative effects of HFSS. Our observations of 20-22 months-old B6 mice paralleled observation of a previous report showing that 10 weeks of a high-fat diet did not cause a significant change in the body weights of 5xFAD mice, but accelerated the pathogenesis of Alzheimer's-like disease. 37 We observed marked deficiencies in motor-muscular grip strength and in heat sensitivities in HFSS-fed compared with ND-fed aged mice. Although our behavior studies are preliminary because of the small sample size in each group, the results are suggestive of significant effects that warrant future studies. The alterations in motormuscular grip strength and sensory abilities have been explained by neurotransmission modifications and cerebral atrophy in the motor and sensory cortexes. 73,[76][77][78] These changes are likely to affect the cortico-cortical and corticospinal connectivity.
We also observed lower densities of neurons marked by NeuN in the hemisphere of HFSS-fed mice than ND-fed mice, although the difference was not significant (p = 0.1302). Previous studies have shown neuronal death following the consumption of a high-fat diet. 70,79 By contrast, another recent study reported that consumption of a high salt diet does not cause any neuronal death. 4 Our study used HFSS, the combination of high fat, sugar, and salt common in Western diets, with the assumption that these components together would have a collective effect that could reduce motor-muscular grip strength and sensory behavior in aged mice. These data suggest that future studies should quantitatively delineate the effects of HFSS on neuronal distribution in motor and sensory cortexes using larger sample sizes.
Our results showed that Met attenuated the above detrimental effects of HFSS diet. Our Met treatment offset some HFSSinduced defects of cortex-related behavior and neural cells in aging mice (Figures 1-6), supporting a neuroprotective effect of Met during HFSS consumption and aging-related neurodegeneration.
The beneficial actions of Met are known to extend beyond diabetes management, since Met is also used in the treatment of various neurodegenerative disorders, 80-84 motor dysfunctions, 85 aging, 29 and inflammatory signaling. [86][87][88] Previous studies reported more microglia with a reactive morphology (large cell bodies) from the cerebral cortexes of aged mice fed ND than in young adult mice fed ND. 89,90 The HFSS diet resulted in even more microglia with the reactive morphology in the aged B6 mice compared with mice fed ND, consistent with increased neuroinflammation. 65 This observation is also in accordance with a new genome-wide transcriptomic study that indicated heightened chronic inflammation in aged microglia. A significant increase was also noted for the inflammatory iNOs + microglia and a decrease in the anti-inflammatory Arg-1 + microglia in the cortexes, as well as an elevation of TNFα levels in brains of the HFSS-fed aged mice compared with the ND-fed aged mice (Figures 3,4, Figure S1). A high-fat diet increases oxidative stress in both serum and tissues throughout the body, 7,11,12 consequently increases the influx of oxidative stress radicals across the blood-brain barrier into the brain. 8,71 Excessive oxidative radicals and inflammatory mediators, including TNFα, 8 are therefore likely to play a role in neurodegeneration, in the inflammatory activation of microglia, and in the increased expression of iNOs observed in cortexes of HFSS-fed aged mice, as observed in the present study (Figure 4). In cortexes of the HFSS-fed aged mice, Met treatment not only reduced the levels of inflammatory TNFα and iNOs + -microglia, but also increased the level of anti-inflammatory Arg-1 + microglia, demonstrating an immunomodulating action of Met on microglia cells ( Figure S1, Figures 3,4). Moreover, aging impaired the regenerative functions of microglia, 89,90 which is likely to F I G U R E 4 HFSS diet increased the level of iNOs-colabeled Iba-1 + microglia in the motor and sensory cortexes, whereas Met reversed this effect. A, Representative microimages of iNOs-colabeled Iba-1 + microglia in motor and sensory cortexes in brain cryosections of mice during aging after treatment with ND or HFSS and/or Met. B, Abundancy, M1-like iNOs + Iba-1 + microglia /total Iba-1 + microglia (iNOs + Iba-1 + /Iba-1 + microglia), % in the microscope field of the motor and sensory cortexes. Arrows indicate microglial cells with representative morphology. The magnified images at the farthest right column of "A" show the morphology and expression of iNOs (yellow) of microglia. Data are presented as dots and lines (medians). p ≤ 0.05 is significant (n = 4 mice). Scale bar =50 µm interplay with HFSS in neurodegeneration and invites future study to address the novel interaction.
HFSS augmented Aβ clustering along blood vessels as well as Aβ and pTau accumulation in the cortexes (Figures 5 and 6). This is consistent with the HFSS-induced reduction of Arg-1 + microglia and surge of inflammatory iNOs + microglia (Figures 3,4) because Inflammatory phenotype of microglia has lower ability in phagocytosis to clean or reduce excessive Aβ and pTau compared with F I G U R E 5 HFSS diet augmented pTau abundance in the motor and sensory cortexes of mice during aging, whereas some of Met treatment insignificantly curbed this HFSS effect. A, Microimages of pTau and NeuN in motor and sensory cortexes in brain cryosections. B, left: intensity sum of pTau + green pixels per microscope field representing pTau levels in the brain cortexes; right: abundancy of pTau + NeuN + / NeuN + cells, % per microscope field in the cortexes of B6 mice. The magnified images at the farthest right column of "A" show the typical expression and localization of pTau and NeuN. Data are presented as dots and lines (medians). p ≤ 0.05 is significant (n = 4 mice). Scale bar =50 µm F I G U R E 6 HFSS consumption induced Aβ accumulation in the vicinity of Col-IV + blood vessels in both motor and sensory cortexes of mice during aging, whereas Met treatment tended to curb this induction. A, Typical microimages of Aβ and Col-IV in the cortexes in the brain cryosections of B6 mice. B, Intensity sum of Aβ + pixels per microscopic field representing the densities of Aβ deposition. Solid arrows: blood vessels; hollow arrows: Aβ deposits. Data are presented as dots and lines (medians). p ≤ 0.05 is significant (n = 4 mice). Scale bar =50 µm inflammation-resolving Arg-1 + microglia. [32][33][34][35][36] In our results from ND, HFSS, and HFSS + Met groups, there is a negative correlation between the M2-like Arg-1 + Iba-1 + microglial abundancy and the level of Aβ or pTau or pTau + NeuN + cell abundancy in brain cortexes, whereas the correlation between M1-like iNOs + Iba-1 + microglia abundancy and the level of Aβ or pTau or pTau + NeuN + cell abundancy is positive. Moreover, Aβ and pTau can promote inflammation, thus switch microglia to M1 type and foster neuroinflammation, neurodegeneration, and incidence of AD. 32,33 Aβ deposition in the vicinity of a blood vessel wall can lead to vascular rupture and intracerebral hemorrhage. 40,41 Our results are consistent with the knowledge that M2 microglia are more effective than M1 microglia in the phagocytosis of extracellular Aβ and Tau/pTau, and that extracellular Aβ and pTau are pro-inflammatory, consequently may promote the switch from M2 to M1 microglia. [32][33][34][35][36]91 The microimages acquired for co-staining of NeuN and pTau (pSer199/202) revealed that HFSS also increased the pTau level in cortexes and in NeuN + neuronal somas (pTau + NeuN + ) of these cortexes. 32,36,[66][67][68]92 These findings indicate that long-term consumption of a combined HFSS diet might exacerbate AD susceptibility and progression in the aged brain. Moreover, Met tended to curb the accumulation of Aβ along blood vessels and pTau, suggesting neuroprotection by Met against AD-like changes and the aging process. 21,[93][94][95] Further experiments are warranted to determine whether Met is more effective with longer treatment or at higher doses. Whether Met prevents or promotes the removal of Aβ and pTau accumulation also warrants future investigation. This exploratory study tested male mice only. Numerous studies suggest the existence of sex differences in brain metabolism, 96 microglial activation and neuroinflammation, 97,98 and vulnerability to brain injury, 99,100 which are involved in the AD neurodegeneration.
Thus, this study needs to be also conducted on female mice in the future to test the prediction of sex differences on the effects of HFSS diet and Met and involved mechanisms.

| CON CLUS IONS
Our data showed that aging male B6 mice fed a HFSS diet for 2 months developed motor-muscular and sensory defects and displayed an inflammatory switch of microglial phenotypes and TNFα production. In addition, it demonstrates an increased burden of Aβ and pTau in the motor and sensory cortexes. This study is the first exploration, to the best of our knowledge, of the connection between a Western diet combining all components of HFSS and the early events leading to impaired motor and sensory functions in the aged. The HFSS-induced pathogenesis disclosed here provides an effective window into mechanisms involved, while targeting potential new therapeutic avenues. The ameliorating neuroprotection of Met in HFSS-induced pathogenesis is promising and requires future investigation. [101][102][103][104] Additional exploration of genetically traceable motor and sensory circuits will further augment our understanding of HFSS-induced defects at both the circuit and network levels. Relative to the lifelong dietary style of some humans, our chronic HFSS treatment could be still too short; therefore, a longer HFSS duration deserves future investigation.
HFSS, aging, and Met could also affect vascular structures and blood pressure 19,105 ; thus, this possibility should be studied. We recommend a more explorative paradigm in this area that examines various aspects, such as longer time for the pathogenesis, different Met dosages, the comparison of HFSS versus a Western diet with single or double high levels of its three components at different ages to identify the most critical factors leading to neurological defects, and the effects of Met at different life stages to advance the therapeutic approaches to curb pathogenesis. We anticipate that the findings of this study could lead to improved tools for effective intervention and a better understanding of chronic neural diseases, including AD.

CO N FLI C T S O F I NTE R E S T
The authors declare that they have no conflict of interest.

AUTH O R CO NTR I B UTI O N S
SH involved in conceiving the entire project, conceptualization, experimental design, data curation, validation, writing, supervision, and funding acquisition. AN and YL involved in data curation, formal analysis, validation, and writing. HP: experimental design, formal analysis, data validation, and writing. QAVD involved in data curation and validation. NBP and CAV involved in data curation. NGB involved in conceptualization, data interpretation, validation, writing, reviewing and editing, and resources.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data of this report are available from the corresponding author upon reasonable request.