Acute inflammatory events attenuate high-sucrose diet-induced neurodegenerative processes in reproductively normal female wild-type mice

It is known that diabetic and chronic inflammatory conditions can increase the risk of Alzheimer’s disease (AD)-like neurodegeneration in isolation. As certain elements of the diabetic/pre-diabetic state may sensitize the brain to inflammatory insult (i.e. excess glucocorticoid activity), there is reason to believe that obesogenic and inflammatory factors may accelerate neurodegeneration in a synergistic manner. Also, given that most AD research utilizes male animal models despite increased prevalence of AD among women, we sought to characterize elements of the established (in males) high-sucrose model of neurodegeneration, for the first time, in reproductively normal (pre-menopausal) female mice. A high-sucrose diet (20% of the drinking water) was combined with systemic intraperitoneal lipopolysaccharide (LPS) injections (0.1 mg/kg; 1x/month over 3 months) over seven months in reproductively normal female wild-type mice (C57Bl/6; n=10/group). Although a deleterious effect was hypothesized, low-dose LPS proved to protect against high sucrose diet-induced pathologies in female wild-type mice. Results from our high-sucrose group confirmed that a high-sucrose diet is a mild model of neurodegeneration in wild-type females, as evidenced by exaggerated glucocorticoid expression, spatial learning deficits, irregularities within the insulin pathway, and increased β-amyloid production and Tau phosphorylation. While LPS had little to no effect in isolation, it exerted a protective influence when added to animals sustained on a high-sucrose diet. Corticosterone homeostasis, and levels of amyloid-β (Aβ) and pTau were rescued following addition of LPS. The work presented supports a high-sucrose diet as a model of mild neurodegeneration in female mice and highlights a protective role for transient inflammation against dietary-insult that may be sex dependent.

anti-inflammatory and anti-β-amyloidogenic(4,14), pre-menopausal females may demonstrate a heightened resistance to 118 neurodegeneration relative to their male and/or post-menopausal female counterparts.

121
Experimental mice and study groups 122 Three-month-old female C57/Bl6 (Charles River, Canada) mice were randomized into four groups of ten animals constituting a 123 2x2 design. The control group was provided normal drinking water and given three intraperitoneal (IP) saline injections delivered  After seven months, mice were weighed and sacrificed for harvesting of liver and brain tissue. One-half of the brain was flash 162 frozen in isopentane and ground into powder for long-term storage at -80° C.

Statistical analysis
215 Data are expressed as mean ± SEM and compared using two-way ANOVA to assess treatment effects and potential interactions 216 between a high-sucrose diet and repeated lipopolysaccharide challenge. Fisher's LSD post-hoc tests were performed to assess the 217 significance of differences between group means relative to control. Prism V 8.1.2 (GraphPad Software, Inc. SD, CA) was used to 218 analyze two-way ANOVA data. As both the small sample size used (n=10) and the lengthy duration of the study (7 months) likely 219 contributed to increased variation within groups, Cohen's D measure of effect size was also used to assess the magnitude of the 220 difference between group means(37-39). Cohen's D values > 0.5 represent a medium effect size while values > 0.8 denote a large 221 effect. Effect size pertains to the standard deviations by which two group means differ; i.e. a D of 0.5 suggests that two groups' 222 means differ by half a standard deviation(37). It has been proposed that small effect sizes (x < 0.5) are likely to be trivial, regardless 223 of significance, whereas medium to large effects (x > 0.5) warrant further exploration, even in cases where statistical significance is 224 not reached(37,40,41). A p-value of 0.05 is not necessarily more reliable than one of 0.08, as the difference is likely the result of 225 minute differences in sample size (i.e. n of 9 vs. n of 10)(40). It can be argued that the size of an effect is more important than its 226 likelihood of being attributable to type I error, so long as the associated p-value is reasonable. For this reason, p-values are provided 227 along with the magnitude of effect (Cohen's D). Seemingly meaningful differences between groups should not be disregarded simply 228 because an alpha greater than 0.05 was calculated(39,41-43). P/D LPS , P/D hS , and P/D hSL pertain to the differences between the control 229 mean and LPS, hS, and hSL group means, respectively. As aberrant glucocorticoid and ceramide activity is associated with dysregulated brain insulin signaling(12,13,22,23), we 258 examined baseline total and phospho concentrations of insulin pathway-associated second messenger protein analytes in hemibrain 259 homogenates to probe for any irregularities. It should be noted that the second messengers explored are not exclusive to the insulin 260 pathway. It is therefore possible that any differences observed might not be solely attributable to insulin dysfunction.

261
Total insulin-receptor substrate 1 (IRS1) protein levels were increased after seven months in all mice on a 20% sucrose diet 262 (sucrose effect P < 0.001; P/D hS = 0.007/1.57, P/D hSL = 0.056/0.924; Fig. 2A (Fig. 3B). Although the anti-inflammatory IL-10 and inflammatory TNF-α, IL-1β, IL-5 and IL-6 showed a 289 few differences between groups (Table 1), when data were normalized against IL-10 (no changes observed across all groups) 290 variation was improved (Fig. 3C-F). Interestingly, although all inflammatory cytokines showed a sucrose effect except for IL-6, only Cytokines were normalized to the anti-inflammatory cytokine IL-10 to help reduce variation within groups. All values are normalized 298 to control (fold change for y-axis). Two-way ANOVA with Fisher's LSD post-hoc test. *P < 0.05 vs control. # Cohen's D > 0.5. 299 n = 9-10 for all groups. 300 301 Table 1. Hemibrain cytokine expression (pg/ml). Open box means Cohen's D > 0.5. Grey box indicates P < 0.05 via two-way 302 ANOVA Fisher's LSD. N = 9-10 for all groups.  Fig. 5A, left) that did not coincide with increased fecal corticosterone levels. Interestingly, 332 although LPS or hS treatment had no effect on long-term spatial recall alone, when combined we observed an improvement in these 333 reproductively normal female mice (Fig. 5A, right).

352
Given that trends in AD incidence coincide with similar trends in obesity, diabetes, and chronic inflammation -all known risk 353 factors for AD -the present study combined a high-sucrose drinking water regimen with repeated monthly intraperitoneal 354 lipopolysaccharide injections to accelerate AD-related pathology in estrous-female wild-type mice. Female mice were used as the 355 bulk of AD literature concerns male animal models despite nearly two-thirds of all AD patients being women(2). We demonstrate 356 that high sugar consumption promotes upregulation of early-stage AD-related processes after just 6-7 months, as evidenced by 357 elevated fecal corticosterone (Fig. 1C), increased liver sphingomyelinase activity (Fig. 1D), brain insulin pathway dysregulation ( Fig.   358 2), increased β-amyloidogenesis (Fig. 4A), an altered brain inflammation state ( Fig. 3 and Table 1), and worsened spatial learning in 359 the Barnes maze (Fig. 5). Interestingly, the addition of LPS prevented/lessened many of these effects; fecal corticosterone ( Fig. 1C) 360 was normal, Tau phosphorylation (Fig. 4B) was suppressed, and spatial acquisition deficits were curtailed.

386
We previously demonstrated that a 20% sucrose diet increases Tau phosphorylation after just 4-months in male mice(9). Similar 387 -though lesser -observations were noted in the present work, confirming for the first time the high-sucrose diet as a model of mild 388 neurodegeneration in female wild-type mice. Both total Aβ 42 and the ratio of Aβ 42 to Aβ 40 were increased by the high sugar diet.

389
Furthermore, high-sucrose animals displayed increased nitrate + nitrite expression, suggesting enhanced NO production (perhaps 390 through upregulated iNOS). Exaggerated NO has been linked to neuroinflammation and nitrosative activity, both of which are known 391 to enhance the processing of APP to Aβ(33,34). Increased β-amyloidogenesis and NO expression were associated with worsened 392 spatial learning performance, highlighting a potential decline in cognition. Given the data presented, it seems possible that a high-sugar diet upregulated neurodegenerative processes (i.e. β-amyloidogenesis, nitrosative activity, etc.) through glucocorticoid(13)-394 and hepatic ceramide-mediated mechanisms (16,19,19) to influence spatial memory.

395
It should be noted that the mild phenotype observed may be related to the sex of the animals. Estrogen has been shown to exert 396 neuroprotection in a variety of models (cell culture as well as animal) and can diminish many pathological processes associated with 397 AD, such as β-amyloidopathy, glucocorticoid over-expression, mitochondrial dysfunction, and oxidative stress(4,5). Also, the mild 398 elevation in Aβ 42 coupled with impaired spatial memory acquisition but intact long-term recall could represent an early time-point 399 in late-onset AD pathogenesis(59-62) (Fig. 6) processes suggestive of an early late-onset AD-like phenotype. Early AD is characterized, in part, by impairments in short-term 408 memory formation. As the disease progresses, long-term recall begins to diminish. This coincides with increased Aβ and pTau 409 burden. Our mice demonstrate a mild though present Aβ and pTau phenotype that coincides with worsened spatial acquisition 410 (delayed learning of a visual escape location in the Barnes maze) and spared long-term recall. This may suggest that our mice display 411 an AD-like phenotype representative of an early time-point within late-onset pathogenesis. 412 413 The modest effects of hS treatment on Tau phosphorylation and β-amyloidogenesis could also have been due to the use of 414 hemibrain homogenates, which may have diluted region-specific differences. In addition, it has been proposed that while diabetes 415 alone may be insufficient to generate a robust AD phenotype, it could serve as a cofactor in the etiology and progression of the which raises the possibility that the lack of LPS-induced neuroinflammation observed in this study could be attributable to estrogen 450 interference. Furthermore, gonadectomy leads to an increase in mortality in female rats exposed to LPS that is attenuated by estrogen 451 replacement therapy(75). Taking potential estrogen interference into account, it is possible that male and/or post-menopausal females 452 would demonstrate more robust neuroinflammatory phenotypes in response to the dose of LPS administered in this study.

453
If transient low-dose LPS-induced inflammation did protect against select elements of the deleterious effects of a high-sucrose 454 diet by 're-setting' the corticosterone response in these reproductively normal female mice, it would reinforce the notion that high 455 sugar diet-induced pathology is mediated primarily by glucocorticoids, and that transient inflammatory events may be beneficial in 456 the long-term management of chronic stress.