Obesity differentially effects the somatosensory cortex and striatum of TgF344-AD rats

Lifestyle choices leading to obesity, hypertension and diabetes in mid-life contribute directly to the risk of late-life Alzheimer’s disease (AD). However, in late-life or in late-stage AD conditions, obesity reduces the risk of AD and disease progression. To examine the mechanisms underlying this paradox, TgF344-AD rats were fed a varied high-carbohydrate, high-fat (HCHF) diet to induce obesity from nine months of age representing early stages of AD to twelve months of age in which rats exhibit the full spectrum of AD symptomology. We hypothesized regions primarily composed of gray matter, such as the somatosensory cortex (SSC), would be differentially affected compared to regions primarily composed of white matter, such as the striatum. We found increased myelin and oligodendrocytes in the somatosensory cortex of rats fed the HCHF diet with an absence of neuronal loss. We observed decreased inflammation in the somatosensory cortex despite increased AD pathology. Compared to the somatosensory cortex, the striatum had fewer changes. Overall, our results suggest that the interaction between diet and AD progression affects myelination in a brain region specific manner such that regions with a lower density of white matter are preferentially affected. Our results offer a possible mechanistic explanation for the obesity paradox.


Diet
Rats fed CHOW were given standard chow and water ad libitum.Rats fed a HCHF diet were given the choice between standard chow and three HCHF food items ad libitum starting at 9 months of age.The HCHF rats were also given the choice between water and 12% sucrose solution to mimic soda consumption.The three HCHF items were rotated with three different HCHF items three times a week to maintain appetite and palatability.Food intake was tracked by weighing leftover food after the three HCHF items were rotated.Rats were weighed weekly.

Tissue collection
The nTg and TgAD rats were sacrificed under 5% isoflurane at 12 months of age.The rats were transcardially perfused with PBS-0.1% heparin followed by PBS-4% paraformaldehyde.Brains were extracted and post-fixed overnight in PBS-4% paraformaldehyde.The brains were cryopreserved in PBS-30% sucrose.Tissue was collected using a sliding microtome and 40 µm sections were collected from bregma + 4.00 to − 6.00.Four evenly spaced sections from bregma + 2.00 to − 1.00 were used for all pathological readout measures.
For the thioflavin-S stain, sections were washed in 1X PBS.Sections were incubated in 1% Thioflavin-S (Sigma-Aldrich #T1892) in distilled water for 8 min at room temperature.Sections were washed twice in 70% ethanol followed by three washes in 1X PBS.Sections were then processed for further immunofluorescent co-staining.

Immunohistochemistry for PLP1, Olig2, PDGFRα, Aβ, and tau
Olig2: All washes were in 1X PBS and were done between all antibody incubations, unless otherwise stated.Sections were incubated in 3% hydrogen peroxide in PBS for 30 min at room temperature.The blocking step and antibody dilutions were in 0.5% Triton X-100 and 0.5% BSA in PBS.The sections were blocked for 1 h at room temperature and then without washing, incubated with rabbit anti-Olig2 (abcam #ab109186; 1:200) overnight at room temperature.The sections were incubated with Vectastain ABC kit anti-rabbit secondary antibody (Vector Laboratories #PK-4001; 1:400) for 1.5 h at room temperature.After washing, the sections were incubated with ABC solution (Vector Laboratories #PK-4001; 1:400) diluted in PBS for 1 h at room temperature.Sections were developed using 3,3'-diaminobenzidine peroxidase substrate kit (Vector Laboratories #SK-4100).The sections were mounted, dehydrated in an ascending ethanol series (70%, 95%, 100%) and xylene, and then cover slipped using Cytoseal.
Aβ: The protocol is similar to Olig2, however, there was an additional antigen retrieval step prior to blocking.The sections underwent antigen retrieval incubated in 70% formic acid for 5 min at room temperature.ABC solution was diluted at 1:200.The primary antibody was mouse anti-human Aβ (Dako #M0872; 1:400) and the secondary antibody was Vectastain ABC kit anti-mouse secondary antibody (Vector Laboratories #PK-4002; 1:400).
Tau: All washes were in tris-buffered saline (TBS) and were conducted between all antibody incubations.0.05% Triton X-100 was added to TBS following primary incubation until 3,3'-diaminobenzidine development.Sections were incubated in 3% hydrogen peroxide in 0.25% Triton X-100 TBS for 30 min at room temperature.The sections were blocked in 5% skim milk in TBS for 1 h at room temperature and without washing, incubated with mouse IgG1 anti-PHF1 (a gift from Dr. Peter Davies; 1:1000) in 5% skim milk in TBS overnight at 4 ℃.The sections were incubated with goat anti-mouse IgG1-biotinXX antibody (Invitrogen #A10519; 1:200) in 20% superblock in TBS for 1.5 h at room temperature.The sections were incubated with ABC solution (Vector Laboratories #PK-4002; 1:200) diluted in 20% superblock in TBS for 1 h at room temperature.Sections were developed using 3,3'-diaminobenzidine peroxidase substrate kit (Vector Laboratories #SK-4100).The sections were mounted, dehydrated in an ascending ethanol series (70%, 95%, 100%) and xylene, and then cover slipped using Cytoseal.

Myelin staining
Myelin fibers were stained using Black Gold II (HistoChem #2BGII).Sections were washed in saline and then mounted on slides.After rehydrating in saline, the slides were incubated with 1 mL of 0.3% Black Gold II in saline on a heating block at 60 ℃ for 6 min.The slides were washed in distilled water and fixed in 1% sodium thiosulfate in saline.The slides were washed in distilled water then dehydrated in an ascending alcohol series (70%, 95%, 100%) and xylene, and then cover slipped with Cytoseal.

Imaging and statistical analyses
Whole brain images were taken using a Zeiss Observer.Z1.Fluorescent stains insets were taken on a Nikon A1 Confocal microscope.ImageJ was used to quantify the density of myelin and fluorescent labeling.Regions of interest were outlined to include either the somatosensory cortex or the striatum.These images were set to a background of 12-14 pixels per 1000 µm 2 .Area coverage was measured as a % area.To count the number of Olig2 + cells, objects between 25 pixels 2 and 300 pixels 2 were summed after thresholding.To count the number of PDGFRα + cells, an Ilastik program was trained to recognize cells and images were exported as pixel probabilities.After thresholding the pixel probabilities, objects greater than 25 pixels 2 were summed.To measure cerebral amyloid angiopathy, a semi-quantitative cerebral amyloid angiopathy score was used as previously established 28 .
In brief, a score of 1 was given to blood vessels with 0-20% thioflavin-S coverage, a score of 2 was given to vessels with 21-40% thioflavin-S coverage, up to a maximum score of 5. Tau was quantified by a manual count on ImageJ after adjusting brightness and contrast.Examiner was blind to sex, diet, and transgene throughout all analysis.Using Grubb's test, one TgAD HCHF rat was a significant outlier across several assays and was removed from the analysis.For activated microglia (1 nTg HCHF, 1 TgAD CHOW, 3 TgAD HCHF) and reactive astrocyte (1 nTg HCHF, 3 TgAD HCHF) analysis, some rats were removed from analysis due to a poor perfusion.GraphPad Prism 9.5.0 was used to perform all statistical analysis and to generate graphs.Two-way ANOVA with a post-hoc Tukey's HSD test was used for all analyses where transgene and diet were independent variables.Unpaired t-test was used where only diet was an independent variable.

Results
Sex differences were not observed in all readout measures examined within this study and thus data is presented in aggregate incorporating male and female data.

HCHF increased body weights of rats
To determine whether rats fed a HCHF diet became obese, their weights were tracked weekly over the study period from nine to twelve months of age.As expected, the mass of all rats increased over the study period.By the end of the study period, non-transgenic rats (nTg) on CHOW weighed 371 ± 43.2 g while nTg on HCHF diet weighed 464.2 ± 42.8 g (p = 0.005).TgAD on CHOW weighed 437.5 ± 40 g while TgAD on HCHF diet weighed 495.1 ± 36.7 g (p = 0.04).

Increased myelin and oligodendrocytes in the somatosensory cortex
Previous studies have found that obesity causes myelin damage and a loss of white matter integrity in both males and females of all age groups, with the most pronounced changes in people with body mass indices over thirty 29 .Thus, we investigated whether obesity impacted myelin in nTg and TgAD rats fed either a CHOW or a HCHF diet.We measured myelin density in the striatum and the somatosensory cortex using the myelin-specific Black-Gold II stain.Transgene (p = 0.39), diet (p = 0.59), and the transgene-diet interaction (p = 0.19) had no significant effects on myelin density in the striatum (Fig. 1a,b).However, diet (p = 0.02) and the transgene-diet interaction (p = 0.0002) had significant effects on myelin density in the somatosensory cortex, while transgene did not (p = 0.09) (Fig. 1c,d).nTg rats on HCHF or CHOW had no differences in myelin density (p = 0.77) while TgAD rats on HCHF diet had a 42.5 ± 7.94% increase in myelin density compared to TgAD on CHOW (p < 0.0001) (Fig. 1d).These data indicate that myelin density increases in response to the interaction between HCHF diet and AD transgene in the somatosensory cortex but not the striatum.
To further investigate the impact of HCHF and obesity on myelin, we measured the immunoreactivity of two myelin proteins, myelin-basic protein (MBP) and myelin proteolipid protein 1 (PLP1) in the striatum and the somatosensory cortex.MBP maintains the structure of myelin and is crucial to the myelination process 30 .Diet (p = 0.0004) and the transgene-diet interaction (p = 0.003) had significant effects on MBP immunoreactivity in the striatum while transgene did not (p = 0.58) (Fig. 1e,f).nTg on HCHF had no differences compared to nTg on www.nature.com/scientificreports/CHOW (p = 0.97) (Fig. 1e,f).Surprisingly, despite no change in myelin density in the striatum, TgAD on HCHF had increased MBP immunoreactivity compared to TgAD on CHOW in this region (33.2 ± 4.71%; p < 0.0001) (Fig. 1e,f).In the somatosensory cortex, diet (p < 0.0001) and the transgene-diet interaction (p = 0.0003) had significant impacts on MBP immunoreactivity, while transgene did not (p = 0.34) (Fig. 1g,h).nTg on HCHF had no changes in MBP immunoreactivity compared to nTg on CHOW (p = 0.36) in the somatosensory cortex (Fig. 1g,h).TgAD on HCHF had increased MBP immunoreactivity compared to TgAD on CHOW within the somatosensory cortex (88.1 ± 20.1%; p < 0.0001) (Fig. 1g,h).To verify the increased myelin in the somatosensory cortex, we measured the immunoreactivity of PLP1, which also stabilizes myelin structure and promotes oligodendrocyte development 31 .In agreement with the MBP immunoreactivity, transgene (p = 0.25), diet (p = 0.23), and the transgene-diet interaction (p = 0.15) had no significant impact on PLP1 in the striatum.(Fig. 1i,j).In the somatosensory cortex, the transgene-diet interaction was significant (p = 0.0004), while diet (p = 0.27) and transgene (p = 0.31) were not (Fig. 1k,l).Within the somatosensory cortex, nTg on HCHF diet had no changes in PLP1 immunoreactivity compared to nTg on CHOW (p = 0.29) in the somatosensory cortex (Fig. 1k,l).TgAD on HCHF had increased PLP1 immunoreactivity compared to TgAD on CHOW (20.6 ± 3.01%; p = 0.003) in the somatosensory cortex.Together, these results demonstrate an interaction between transgene and diet that increases myelin density and myelin-related proteins in the somatosensory cortex.
Oligodendrocytes are the myelin-producing cells of the central nervous system and, in AD, participate in defective myelination and undergo apoptosis 32 .Since we observed changes in myelination in the striatum and somatosensory cortex, we next quantified oligodendrocytes using the oligodendrocyte-specific marker, Olig2.Transgene (p = 0.64), diet (p = 0.78), and the transgene-diet interaction (p = 0.17) had no significant impacts on the number of Olig2-positive cells in the striatum (Fig. 2a,b).On the other hand, transgene (p < 0.0001), diet (p = 0.0001), and the transgene-diet interaction (p < 0.0001) were significant factors in Olig2 counts such that TgAD on HCHF diet had significantly more Olig2 + cells compared to TgAD on CHOW in the somatosensory cortex (19.3 ± 1.27%; p < 0.0001) while nTg on HCHF diet had no changes in Olig2 compared to nTg on CHOW (p = 0.85) (Fig. 2c,d).Overall, these results suggest an increase in oligodendrocytes in the somatosensory cortex due to a HCHF diet and AD progression.
AD pathology causes oligodendrocyte progenitor cells to undergo apoptosis, which compromises myelin integrity 32 .Since we observed an increase in oligodendrocytes, oligodendrocyte progenitor cells were quantified using the oligodendrocyte progenitor cell-specific marker, PDGFRα.In the striatum, diet (p = 0.04) had a significant effect on the number of PDGFRα + cells while transgene (p = 0.31) and the transgene-diet interaction (p = 0.33) did not (Fig. 2e,f).However, no significant differences were found between TgAD on HCHF compared to TgAD on CHOW (p = 0.86) or nTg on HCHF compared to nTg on CHOW (p = 0.20) (Fig. 2e,f).In the somatosensory cortex, diet (p = 0.0008) and the transgene-diet interaction (p = 0.0005) significantly impacted the number of PDGFRα + cells such that TgAD on HCHF diet had significantly more PDGFRα + cells than TgAD on CHOW (11.97 ± 0.66%; p < 0.0001) (Fig. 2g,h).However, in the somatosensory cortex transgene (p = 0.32) had no significant impact on PDGFRα + cells and there were no differences between nTg on HCHF compared to nTg on CHOW (p = 0.999) (Fig. 2g,h).Overall, these data demonstrate an increase in oligodendrocytes and oligodendrocyte progenitor cells in the somatosensory cortex that may have led to the increase in myelin density as a result of HCHF diet.

No neuronal loss, decreased inflammation in the somatosensory cortex
Oligodendrocytes and neurons interact at the axon to direct myelination.When this interaction is interrupted, neurodegeneration may occur 33 .AD is characterized by a loss of neurons 20,21,24 and thus we examined whether HCHF diet impacted neuronal loss using neuronal nuclei (NeuN) immunoreactivity.In the striatum, transgene (p = 0.06), diet (p = 0.55), and the transgene-diet interaction (p = 0.27) had no significant effects on NeuN immunoreactivity (Fig. 3a,b).Similarly in the somatosensory cortex, transgene (p = 0.07), diet (p = 0.33), and the transgene-diet interaction (p = 0.25) had no significant effects on NeuN immunoreactivity (Fig. 3c,d) This suggests that the HCHF diet does not contribute to neuronal loss in the striatum and the somatosensory cortex.
Microglia ensure proper oligodendrocyte and oligodendrocyte progenitor cell development, while contributing to normal myelination 34 .In contrast, activated microglia are a hallmark of AD progression and contribute to neuroinflammation, neuronal loss, and cognitive deficits 20 .Previously, obesity has been found to increase microglial activation 19 .We set out to determine whether HCHF diet and AD progression impacted microglia by measuring ionized calcium binding adaptor molecule 1 (Iba1) immunoreactivity.Increased Iba1 immunoreactivity results from an increase in activated microglia 20 .In the striatum, transgene (p = 0.003) significantly increased Iba1 immunoreactivity while diet (p = 0.005) significantly reduced Iba1 immunoreactivity (Fig. 4a,b).However, there was no transgene-diet interaction (p = 0.97) and so TgAD on HCHF compared to TgAD on CHOW (p = 0.14) and nTg on HCHF compared to nTg on CHOW (p = 0.21) had no significant changes in Iba1 immunoreactivity in the striatum (Fig. 4a,b).In the somatosensory cortex, transgene (p = 0.0006) significantly increased Iba1 immunoreactivity, diet (p < 0.0001) significantly reduced Iba1 immunoreactivity, and the transgene-diet interaction (p = 0.002) significantly impacted Iba1 immunoreactivity such that TgAD on HCHF diet had 24.0 ± 2.73% less Iba1 immunoreactivity compared to TgAD on CHOW in the somatosensory cortex (p < 0.0001) (Fig. 4c,d).However, nTg on HCHF had no significant changes in Iba1 immunoreactivity compared to nTg on CHOW (p = 0.82) in the somatosensory cortex (Fig. 4c,d).This suggests that the HCHF diet in an AD transgene specific manner decrease Iba1 immunoreactivity in the somatosensory cortex.
Astrocytes and oligodendrocytes also interact to induce myelination 35 .In AD, reactive astrocytes contribute to neuroinflammation and neurodegeneration 20 .We investigated the HCHF diet and AD progression on reactive astrocytes by measuring glial fibrillary acidic protein (GFAP) immunoreactivity.Increased GFAP immunoreactivity is a marker of reactive astrocytes 20  www.nature.com/scientificreports/immunoreactivity while diet (p = 0.43) and the transgene-diet interaction (p = 0.60) had no significant effect on GFAP immunoreactivity (Fig. 4e,f).In the striatum, TgAD on HCHF compared to TgAD on CHOW (p = 0.75) and nTg on HCHF compared to nTg on CHOW (p = 0.99) had no significant changes in GFAP immunoreactivity (Fig. 4e,f).Similarly, in the somatosensory cortex, transgene (p < 0.0001) significantly increased GFAP immunoreactivity while diet (p = 0.77) and the transgene-diet interaction (p = 0.76) had no significant effect (Fig. 4g,h).In the somatosensory cortex, TgAD on HCHF compared to TgAD on CHOW (p = 0.99) and nTg on HCHF compared to nTg on CHOW (p = 0.97) had no significant changes in GFAP immunoreactivity (Fig. 4g,h).Together, this suggests that HCHF diet has no effect on astrocyte activation as measured by GFAP immunoreactivity in either the somatosensory cortex or striatum.

Increased AD pathology in the somatosensory cortex
A key pathological hallmark of AD is the presence of Aβ plaques in the brain 20,21 .Aβ deposition is directly and indirectly linked to myelin dysfunction in AD, and a high-fat diet has been found to increase Aβ plaques 16,31 .Therefore, after observing increased myelin, we examined whether a HCHF diet has an impact on Aβ plaques utilizing the Ab-specific antibody, 6F3D.In the striatum, no significant difference was observed in Aβ plaques between TgAD on CHOW and TgAD on HCHF (p = 0.91) (Fig. 5a,c).Aβ plaques increased by 46.4 ± 19.2% in the somatosensory cortex in TgAD on HCHF compared to TgAD on CHOW (p = 0.04) (Fig. 5b,d).Cerebral amyloid angiopathy which involves Aβ deposition in the vasculature is present in up to 90% of AD patients 36 .Cerebral amyloid angiopathy impairs the brain's vasculature leading to brain atrophy, cognitive impairment, and ischemia 36 .By compromising the vasculature, cerebral amyloid angiopathy has been found to cause lesions in white matter 37 .We have previously found cerebral amyloid angiopathy in the penetrating arterioles and amyloid accumulation in the venules of the somatosensory cortex 21 .Thus, we set out to determine if cerebral amyloid angiopathy was impacted by diet.TgAD on HCHF had a 10.6 ± 1.28% increase in cerebral amyloid angiopathy in the somatosensory cortex compared to TgAD on CHOW (p = 0.04) (Fig. 5e,f).Aβ plaques and cerebral amyloid angiopathy are not detected in nTg rats and thus not quantified herein (data not shown).Together, this shows increased Aβ in the somatosensory cortex parenchyma and vasculature due to a HCHF diet.Intracellular tau inclusions contribute to neurodegeneration in AD and are frequently colocalized with Aβ plaques 38 .Thus, we set out to determine if tau inclusions increase in conjunction with the increase in Aβ plaques.We examined tau inclusions using the disease-specific antibody, paired helical filament-1 (PHF-1) immunoreactivity.We analyzed both plaque-associated and non-plaque associated tau inclusions (Fig. 6).In the striatum, plaque-associated (p = 0.33) and non-plaque associated inclusions (p = 0.49) remained unchanged between TgAD on CHOW and TgAD on HCHF diets (Fig. 6a,b).In the somatosensory cortex, there was a 48.5 ± 9.55% increase in the number of plaque-associated tau inclusions of TgAD on HCHF compared to TgAD on CHOW (p < 0.0001), with no significant differences in non-plaque associated tau inclusions (p = 0.30) (Fig. 6c,d).Overall, this shows that a HCHF diet increases AD pathology in the somatosensory cortex.

Discussion
Herein, we showed that a varied HCHF diet induced obesity in both TgAD and nTg rats from nine to twelve months of age.A HCHF diet in TgAD led to larger effect size outcomes in the somatosensory cortex compared to the striatum, confirming region-specific effects of a HCHF diet.We showed that HCHF diet-induced obesity increases myelin within the somatosensory cortex with a corresponding increase in oligodendrocytes and oligodendrocyte progenitor cells.We observed no loss in neurons.Further, there was decreased activated microglia, yet an increase in Aβ, cerebral amyloid angiopathy, and plaque-associated tau in the somatosensory cortex.These combined results suggests that diet effects on AD pathology are independent from other beneficial effects on myelin in the somatosensory cortex [39][40][41] .
Restoring myelin loss may be a promising method of reversing cognitive decline in AD 42 .Several studies have reported myelin loss in AD as a result of Aβ toxicity and tauopathy 43,44 .The loss in myelin exacerbated cognitive decline in human 45 .Further during normal aging in humans, myelin loss has been associated with age-related cognitive decline 46,47 .The increase in myelin that we report in the obese TgAD rats may provide one mechanism leading to the obesity paradox observed in aged people.Another study found that an increase in myelin can  rescue cognitive deficits in AD rodent models 41 .Chen et al. 41 utilized genetic knockouts and antagonists of the muscarinic receptor 1, both of which promoted oligodendrocyte progenitor cell differentiation.Oligodendrocyte progenitor cell differentiation led to an increase in oligodendrocytes, which enhanced myelin renewal in mice 41 .Increased myelin was found to rescue memory and improved neuronal function 41 .Increased MBP in both striatum and somatosensory cortex may also contribute to these benefits.AD patients have decreased MBP, which is consistent with the decreased levels of myelin 48 .Thus, by increasing and stabilizing myelin expression, neuronal function and cognitive deficits in AD may be attenuated.Behavioral experiments of our rats have similarly found that HCHF diet concomitant with AD transgenes improved spatial learning in male rats and executive function in female rats 49 .Notably, oligodendrocytes require significant amounts of energy to increase myelination 50 .One possible source of this energy is the HCHF diet consumed by the rats in our study, which is abundant in energy 51 .Another benefit of a HCHF diet is the diverse carbohydrate composition.Increasing carbohydrates, such as glucose and lactate, has been shown to increase oligodendrocyte development and myelination in rats 52 .
Since oligodendrocyte progenitor cells continue to divide and proliferate throughout adulthood, the observed increase in PDGFRα + cells may also contribute to the increase in oligodendrocytes 53 .Overall, our results suggest that a HCHF diet increases the number of oligodendrocytes, which increase myelination.
Our overall results show that a HCHF diet has a larger effect size on the somatosensory cortex compared to the striatum.This suggests that regions with less white matter may be more positively affected by a HCHF diet compared to regions dense in white matter.In human AD, both white and gray matter are impacted, however, damage to regions dominated by gray matter tend to be more severe 54,55 .One possible explanation for the observed increase in myelin in the somatosensory cortex is that gray matter contains axons that are scarcely myelinated.In contrast, white matter is heavily myelinated leaving little space for more myelin.This extra space in the gray matter may accommodate an increase in myelin.Functionally, the striatum acts as a relay between the thalamus and the cerebral cortex, thus, the dense myelin of the striatum is important for efficient communication 56 .However, the somatosensory cortex is involved in cognitive processes like motor and haptic memory 27 .Since myelination is important to preserve memories in the long-term, increased myelin may be the result of improved memory consolidation 57 .These functional differences may help to explain the larger effect size within the somatosensory cortex due to a HCHF diet compared to the striatum.
The decrease in microglial activation in the somatosensory cortex of TgAD HCHF treated rats may contribute to a beneficial oligodendrocyte-microglia interaction that increases myelin.Microglia promote survival, maturation and migration of oligodendrocyte progenitor cells that may have led to increased myelin 32 .This interaction is bidirectional, where oligodendrocytes in stressed states can induce activation of microglia.By reducing oligodendrocyte stress, neuroinflammation may also decrease 32 .Fewer activated microglia may also