24(S)-Saringosterol Prevents Cognitive Decline in a Mouse Model for Alzheimer’s Disease

We recently found that dietary supplementation with the seaweed Sargassum fusiforme, containing the preferential LXRβ-agonist 24(S)-saringosterol, prevented memory decline and reduced amyloid-β (Aβ) deposition in an Alzheimer’s disease (AD) mouse model without inducing hepatic steatosis. Here, we examined the effects of 24(S)-saringosterol as a food additive on cognition and neuropathology in AD mice. Six-month-old male APPswePS1ΔE9 mice and wildtype C57BL/6J littermates received 24(S)-saringosterol (0.5 mg/25 g body weight/day) (APPswePS1ΔE9 n = 20; C57BL/6J n = 19) or vehicle (APPswePS1ΔE9 n = 17; C57BL/6J n = 19) for 10 weeks. Cognition was assessed using object recognition and object location tasks. Sterols were analyzed by gas chromatography/mass spectrometry, Aβ and inflammatory markers by immunohistochemistry, and gene expression by quantitative real-time PCR. Hepatic lipids were quantified after Oil-Red-O staining. Administration of 24(S)-saringosterol prevented cognitive decline in APPswePS1ΔE9 mice without affecting the Aβ plaque load. Moreover, 24(S)-saringosterol prevented the increase in the inflammatory marker Iba1 in the cortex of APPswePS1ΔE9 mice (p < 0.001). Furthermore, 24(S)-saringosterol did not affect the expression of lipid metabolism-related LXR-response genes in the hippocampus nor the hepatic neutral lipid content. Thus, administration of 24(S)-saringosterol prevented cognitive decline in APPswePS1ΔE9 mice independent of effects on Aβ load and without adverse effects on liver fat content. The anti-inflammatory effects of 24(S)-saringosterol may contribute to the prevention of cognitive decline.


Introduction
Alzheimer's disease (AD) is the most prevalent form of dementia in the elderly. This neurodegenerative disorder is characterized by a progressive cognitive decline, accumulation of amyloid-β (Aβ), formation of neurofibrillary tangles, neuroinflammation, and loss

24(S)-Saringosterol Does not Affect the Aβ Plaque Load in the Cortex and the Hippocampus
The Aβ plaque load in the cortex (p = 0.963) and in the hippocampus (p = 0.450) of APPswePS1ΔE9 mice (Figure 3a,b) were not affected by 24(S)-saringosterol administration. There were also no differences in concentrations of insoluble Aβ40 and Aβ42 or soluble extracellular, intracellular, or membrane-associated Aβ40 and Aβ42 (p > 0.05) (Figure 3dg). The percentage of surface coverage of the Aβ staining was determined in the total cortical (a) and hippocampal area (b) of APPswePS1ΔE9 mice after immunohistochemical staining (cortex: n = 6 per group; hippocampus: n = 5 and 6 per group, respectively). Photos of Aβ-stained cortical and hippocampal areas representative for the experimental groups are shown (c). Soluble (extracellular (d), intracellular (e), and membrane-associated (f)) Aβ and insoluble Aβ (g) in the cortex of APPswePS1ΔE9 mice administered with vehicle or 24(S)-saringosterol was quantified using ELISA (n = 16 and 20 per group, respectively). Bars represent mean ± SEM. The percentage of surface coverage of the Aβ staining was determined in the total cortical (a) and hippocampal area (b) of APPswePS1∆E9 mice after immunohistochemical staining (cortex: n = 6 per group; hippocampus: n = 5 and 6 per group, respectively). Photos of Aβ-stained cortical and hippocampal areas representative for the experimental groups are shown (c). Soluble (extracellular (d), intracellular (e), and membrane-associated (f)) Aβ and insoluble Aβ (g) in the cortex of APPswePS1∆E9 mice administered with vehicle or 24(S)-saringosterol was quantified using ELISA (n = 16 and 20 per group, respectively). Bars represent mean ± SEM.

24(S)-Saringosterol Affects the Expression of LXR Target Genes In Vitro, But Not In Vivo
We assessed the effect of 24(S)-saringosterol on the expression of lipid metabolismrelated LXR target genes in CCF-STTG1 cells ( Figure 5) and in the hippocampus of APPswePS1∆E9 mice (Supplementary Figure S1). Incubation of CCF-STTG1 glial cells with 24(S)-saringosterol increased the expression of ABCA1 (Figure 5a), ABCG1 (Figure 5b), and APOE ( Figure 5c) in a dose-dependent manner. The expression of ABCA1 and ABCG1 was increased to a comparable extend by the positive control T0901317, while the expression of APOE was increased to a lesser extent by 24(S)-saringosterol than by T0901317. However, no effect of 24(S)-saringosterol administration could be detected on the expression of Abca1, Abcg1, Apoe, Scd1, or Srebf1 in the hippocampus of WT or APPswePS1∆E9 mice (p > 0.05) (Supplementary Figure S1). in WT mice (F(1, 15) = 7.871, p < 0.05) (Figure 4c). The difference in microglia cell count in the cortex of APPswePS1ΔE9 mice and WT mice on the vehicle treatment (p < 0.05) disappeared upon 24(S)-saringosterol administration (p = 0.718). There were no differences in CD68 levels in the cortex of WT and APPswePS1ΔE9 mice (F(1, 17) = 0.969, p = 0.339), and no effects of 24(S)-saringosterol treatment (F(1, 17) = 0.109, p = 0.746) (Figure 4d). . Administration of 24(S)-saringosterol reduces the microglia marker Iba1 and microglial density, but not CD68 in APPswePS1ΔE9 mice. Coronal sections of the brain of WT and APPswePS1ΔE9 (AD) mice were stained for Iba1 (a-c,e) and CD68 (d) by immunohistochemistry, and the percentage of surface coverage of the staining in the total cortical and hippocampal area was determined (a,b,d). Photos of the Iba1-stained cortex representative for the experimental groups are shown (e). Iba1 and CD68 levels are presented as the percentage of surface coverage, the microglia cell count as the number of Iba1-positive stained cell bodies per 100 inch 2 cortex. Bars represent mean ± SEM (n = 5-6, 3, 3-6, and 5 per group, respectively). * p ≤ 0.05, ** p ≤ 0.01.

24(S)-Saringosterol Affects the Expression of LXR Target Genes In Vitro, but not In Vivo
We assessed the effect of 24(S)-saringosterol on the expression of lipid metabolismrelated LXR target genes in CCF-STTG1 cells ( Figure 5) and in the hippocampus of APPswePS1ΔE9 mice (Supplementary Figure S1). Incubation of CCF-STTG1 glial cells with 24(S)-saringosterol increased the expression of ABCA1 (Figure 5a), ABCG1 (Figure 5b), and APOE ( Figure 5c) in a dose-dependent manner. The expression of ABCA1 and ABCG1 was increased to a comparable extend by the positive control T0901317, while the expression of APOE was increased to a lesser extent by 24(S)-saringosterol than by T0901317. However, no effect of 24(S)-saringosterol administration could be detected on the expression of Abca1, Abcg1, Apoe, Scd1, or Srebf1 in the hippocampus of WT or APPswePS1ΔE9 mice (p > 0.05) (Supplementary Figure S1). . Administration of 24(S)-saringosterol reduces the microglia marker Iba1 and microglial density, but not CD68 in APPswePS1∆E9 mice. Coronal sections of the brain of WT and APPswePS1∆E9 (AD) mice were stained for Iba1 (a-c,e) and CD68 (d) by immunohistochemistry, and the percentage of surface coverage of the staining in the total cortical and hippocampal area was determined (a,b,d). Photos of the Iba1-stained cortex representative for the experimental groups are shown (e). Iba1 and CD68 levels are presented as the percentage of surface coverage, the microglia cell count as the number of Iba1-positive stained cell bodies per 100 inch 2 cortex. Bars represent mean ± SEM (n = 5-6, 3, 3-6, and 5 per group, respectively). * p ≤ 0.05, ** p ≤ 0.01.

Discussion
In this study, we examined whether purified 24(S)-saringosterol can preserve cognition and prevent the development of neuropathology in a mouse model for AD. Our data show that 10 weeks of administration with the semi-synthetic preferential LXRβ activating phytosterol 24(S)-saringosterol prevented cognitive decline in APPswePS1ΔE9 mice,

Discussion
In this study, we examined whether purified 24(S)-saringosterol can preserve cognition and prevent the development of neuropathology in a mouse model for AD. Our data show that 10 weeks of administration with the semi-synthetic preferential LXRβ activating phytosterol 24(S)-saringosterol prevented cognitive decline in APPswePS1∆E9 mice, without affecting the Aβ plaque load. Concentrations of 24(S)-saringosterol in the brain were significantly increased upon 24(S)-saringosterol administration, and microglial activation in the brain of APPswePS1∆E9 mice was found to be reduced.
The prevention of the cognitive decline in APPswePS1∆E9 mice upon 24(S)-saringosterol administration is in accordance with our previously reported data demonstrating neuroprotective effects of T0901317 and 24(S)-saringosterol-containing Sargassum fusiforme or its lipid extract [14,21]. Although 24(S)-saringosterol reduces neuronal Aβ 42 release and promotes microglial Aβ clearance in vitro [21], no effect of 24(S)-saringosterol administration on Aβ plaque load could be detected. This observation is in line with our data showing no effect of T0901317 on Aβ plaque load despite effects on cognition [14]. However, the mice treated with T0901713 were much older than the mice in the present study and therefore Aβ deposition could not be prevented. Because Sargassum fusiforme, either in crude form or as a lipid extract, did decrease the Aβ plaque load in APPswePS1∆E9 mice [21], constituents other than saringosterol are likely to reduce the deposition of Aβ. Aβ plaque lowering effects have been reported for several constituents besides 24(S)-saringosterol contained by Sargassum fusiforme including the phytosterols β-sitosterol [37,38] and stigmasterol [39,40], the carotenoid fucoxanthin [28,31,32], and oligosaccharide sodium oligomannate [24]. These findings suggest that LXR activation by 24(S)-saringosterol can prevent the decline in cognition independently of its effect on Aβ deposition.
Although 24(S)-saringosterol concentrations were significantly increased in serum and in the brain of the mice after its administration, this did not affect the concentrations of cholesterol, its precursors, or metabolites. Phytosterol concentrations in serum were reduced, possibly as a result of competition for incorporation in micelles and the subsequent intestinal absorption or an inhibitory effect of 24(S)-saringosterol on intestinal sterol absorption. An alternative explanation is an enhancing effect of 24(S)-saringosterol on sterol excretion via LXR activation and upregulation of Abcg5/8 in the liver and the intestine [41,42]. The expression of lipid metabolism-related LXR target genes in the brain remained unaffected despite the increased concentrations of 24(S)-saringosterol. However, in cultured CCF-STTG1 cells, we observed that 24(S)-saringosterol administration did increase the expression of LXR target genes ABCA1, ABCG1, and APOE. Chen et al. (2014) obtained similar results after saringosterol administration to HEK293T, HepG2, THP-1 monocytes, and RAW264.7 cells [36]. The absence of an effect in vivo may be caused by the chronic nature of the 24(S)-saringosterol administration leading to a new balance in gene expression levels or the difference in 24(S)-saringosterol concentration in vivo (9 × 10 −4 mM in serum) compared to in vitro (1.25-7.5 mM). Previously, we observed a limited increase in expression of LXR target genes in the brains of WT and APPswePS1∆E9 mice upon administration of Sargassum fusiforme containing a similar concentration of 24(S)-saringosterol [21]. Sargassum fusiforme contains additional LXR agonists, including fucosterol, that may contribute to the gene expression profile resulting from Sargassum fusiforme supplementation. Therefore, it remains to be established via which biological pathways, involving LXR activation or not, 24(S)-saringosterol exerts its neuroprotective effects in AD mice.
Activation of LXR is known to transrepress inflammatory pathways in immune cells in the central nervous system through SUMOylation [43]. Because chronic neuroinflammationattributed to excessive activation of microglia-exacerbates AD pathologies [10], activating LXR might alleviate AD symptoms by dampening this process. This is supported by our findings that 24(S)-saringosterol administration did restore the expression of Iba1 (a marker for microglial activation and inflammation [44]) and the number of microglia in the AD mice to levels similar to those in WT mice. On the other hand, the absence of differences in the expression of CD68 suggests no involvement of phagocytic activity of microglia [45]. Increased Iba1 expression, and thus microglial activation, in brain tissue of AD patients has been reported [46]. An increased ApoE production may contribute to the reduction in neuroinflammation [21,47]. However, an upregulated ApoE production was induced by 24(S)-saringosterol only in vitro and not in vivo [21]. Therefore, the exact immunomodulatory effects of 24(S)-saringosterol in AD remain to be elucidated.
Our observation that administration of 24(S)-saringosterol as a preferential LXRβ activator does not induce hepatic steatosis is in line with the assumption that triglyceride accumulation in the liver is predominantly driven by LXRα activation [19]. Further research should elaborate on the potential adverse effects of long-term 24(S)-saringosterol administration. Our data therefore support the potential application of pure 24(S)-saringosterol, Sargassum fusiforme, either crude or as an extract, in the prevention and retardation of AD-related symptoms.
An advantage to the use of crude seaweed or a seaweed extract over pure 24(S)saringosterol could be the presence of constituents with potential additional or synergistic effects. Seaweed constituents other than 24(S)-saringosterol have been reported to display beneficial effects in AD models [24]. Several phytosterols, including fucosterol, sitosterol, stigmasterol, and brassicasterol, have been reported to exert anti-inflammatory effects via LXR activation [24,48,49]. However, we could not confirm these results in an LXR assay using physiological sterol concentrations [21]. Evidence indicates that sitosterol can augment the polarization of macrophages towards an anti-inflammatory phenotype via transrepression of toll-like receptor activation [48]. Moreover, the formation of Aβ can be reduced by sitosterol [38], stigmasterol [39,40], and fucosterol [28]. Fucosterol also alleviates Aβ-induced ER stress in primary rat hippocampal neurons and prevents soluble Aβ 42 exposure-induced cognitive decline in aging rats [29]. The carotenoid fucoxanthin may contribute to the prevention of AD-related symptoms by its anti-oxidative properties [34] and by preventing the formation of Aβ peptides [28,[31][32][33] or Aβ neurotoxicity [31]. Furthermore, seaweed-derived polyphenols such as phloroglucinol have been reported to prevent cognitive impairment in AD mice, possibly via anti-oxidative effects [49], and fucoidans were found to ameliorate cognitive impairments in models for neurodegeneration via anti-oxidative or anti-inflammatory mechanisms [26]. The oligosaccharide sodium oli-gomannate, which has demonstrated cognitive improvement in a phase 3 clinical trial in China, was found to reduce neuroinflammation by remodeling the gut microbiome [25]. Furthermore, seaweeds may have a beneficial effect on cognition by being a low glycemic food, tending to release glucose slowly and steadily [50]. Because a steady supply of glucose to the brain is required for optimal cognitive performance and an impaired glucose metabolism has been linked to AD, blood glucose control is crucial [50][51][52]. Therefore, dietary supplementation with seaweed combines the beneficial properties of multiple constituents that may act synergistically in the prevention and retardation of AD-related symptoms.
In conclusion, we showed that semi-synthetic purified 24(S)-saringosterol, which is bioavailable in the central nervous system, prevents cognitive decline in a well-established AD mouse model, despite having no effect on Aβ plaque load, or any detectable effects on lipid or cholesterol homeostasis. Our data point to immunomodulatory effects of 24(S)-saringosterol contributing to its neuroprotective properties, but detailed mechanisms remain to be elucidated. Moreover, 24(S)-saringosterol can be regarded as a promising agent for the prevention of deterioration of AD-related symptoms.

Route of 24(S)-Saringosterol Administration
Prior to this experiment, two routes of 24(S)-saringosterol administration were compared: oral gavage and subcutaneous injection. Six C57BL6/J mice received an oral gavage and six C57BL6/J mice received a subcutaneous injection containing 0.5 mg 24(S)saringosterol-semi-synthesized from kelp-derived fucosterol-(purity of 98.2%; COM-FiON B.V., Leimuiden, The Netherlands) per 25 g body weight twice-daily for three consecutive days, whereafter sterol concentrations in serum, brain, and other tissues were determined by gas chromatography/mass spectrometry (GC/MS) as described previously [21,53] (Supplementary Table S1). Oral gavage and subcutaneous injection led to saringosterol concentrations of 40.9 ± 12.7 and 28.0 ± 3.3 µg/dL in serum and to 6.43 ± 1.40 and 5.51 ± 1.85 ng saringosterol per mg dry weight cerebellum. Because the bioavailability of saringosterol was higher upon oral gavage, we administered saringosterol via this route.

Animals and Diet
Male APPswePS1∆E9 (AD) and wildtype C57BL6/J (WT) littermate mice were obtained by backcrossing male APPswePS1∆E9 mice (The Jackson Laboratory, Bar Harbor, ME, USA) with female C57BL6/J mice (Envigo, Horst, The Netherlands). The animals were housed in a conventional animal facility at Hasselt University. Two weeks prior to the start of the behavioral experiments, 5-month old mice were housed individually. Mice were fed ad libitum and kept in an inversed 12/12 h light/dark cycle with behavioral experiments performed during the dark phase of the cycle. The cognitive performances were scored blindly. The body weight of the mice was monitored twice a week. Two series of animal experiments were conducted. The animal procedures were approved by the ethical committee for the animal experiments of Hasselt University and performed in accordance with institutional guidelines (protocol ID201849).

Cognitive Testing
Prior to the baseline assessment, the mice were habituated to the arena and to the four different objects used for cognitive testing, as previously described [21]. At baseline, a functional memory of both WT and APPswePS1∆E9 mice was confirmed with the OLT. After one resting day, the experiment was initiated. The ORT and the OLT were conducted after the treatment period of 10 weeks by a researcher that was blinded to the experimental groups. The same objects were used for the ORT and OLT and objects were selected following a randomized scheme (Supplementary Tables S2 and S3).
The ORT was conducted as described previously [54]. During the first trial (T1), the animal was exposed to two similar objects for 4 min after which it was placed back in its home cage. After a 1 h inter-trial interval (ITI), a second trial (T2) was performed during which the animal was exposed for 4 min to one familiar object from T1 and one novel object. The times spent exploring each object during T1 and T2 were recorded manually. Biting or sitting on the object was not considered exploratory behavior. As a measure of object memory, the discrimination index (D2) ((exploration time for novel object)-(exploration time for familiar object)/(total exploration time in T2)) in T2 was calculated.
The OLT was conducted as a modified form of the ORT described elsewhere [54]. During the first trial (T1), the animal was exposed to two similar objects placed symmetrically in the arena center for 4 min, after which it was placed back in its home cage. After a delay interval of 4 h, a second trial (T2) was performed during which the animal was exposed for 4 min to the two objects from T1 of which one was displaced. The times the mice spent exploring each object during T1 and T2 were recorded manually. As a measure of spatial memory, the discrimination index (D2) ((exploration time for displaced object) -(exploration time for stationary object)/(total exploration time in T2)) in T2 was calculated.

Tissue Sample Preparation
After the post-treatment memory assessment, mice were sacrificed and tissues were isolated for further analyses. Mice were anesthetized by intraperitoneal injection of Dolethal (Vetoquinol, Aartselaar, Belgium) (200 mg per kg body weight) followed by transcardial perfusion with Heparin-phosphate-buffered saline (PBS). Blood was collected via cardiac puncture and was centrifuged for 10 min at 200 g to separate serum, which was stored at −80 • C until use. Brains were isolated and divided into the forebrain (above bregma 0), the cerebellum, and the remaining two hemispheres. The cerebellum was snap-frozen and stored at −80 • C until sterol profiling. The left hemisphere was fixed in formalin and embedded in paraffin for immunohistochemistry. The cortex from the right hemisphere was snap-frozen and cryopreserved for Aβ ELISA analyses. Half of the liver was directly snap-frozen for mRNA expression analyses, the other half was stored at −80 • C in O.C.T. embedding compound (Sakura Finetek USA, Inc., Torrance, CA, USA) for Oil Red O staining.

Determination of Lipid Profile
Sterol profiles in serum and in the cerebellum were determined by GC/MS as described previously [21,53]. In short, prior to sterol analysis, the brain tissue samples (cerebellum) were spun in a speed vacuum dryer to relate individual sterol concentrations to dry weight. The sterols were extracted from the dried tissues by placing them in a 5-mL mixture of chloroform-methanol. Subsequently, 1 mL of the brain sterol extracts was evaporated to dryness. Furthermore, 1 mL of distilled water was added to the samples. To extract the neutral sterols, 3 mL of cyclohexane was added twice. The combined cyclohexane phases were again evaporated to dryness under a stream of nitrogen at 63 • C, and the sterols were dissolved in 100 µL n-decane. After transfer to gas-chromatography (GC)-vials, the sterols were converted to trimethylsilyl ethers (TMSis) and incubated at 60 • C for 1 h [7]. Levels of cholesterol were determined in a gas-chromatograph-flame ionization detector (GC-FID) with 50 µL 5α-cholestane-solution (1 mg/mL 5α-cholestane in cyclohexane) as an internal standard. Levels of plant sterols (campesterol, sitosterol), cholesterol precursors (lanosterol, lathosterol, and desmosterol), and cholesterol metabolites (24S-OH-cholesterol and cholestanol) were determined using gas chromatography-mass spectrometry (GC-MS) using epicoprostanol as an internal standard.
Triglyceride and cholesterol concentrations in serum were determined with enzymatic reagent kits according to the manufacturer's instructions (DiaSys Diagnostic Systems, Holzheim, Germany).

ELISA-Quantification of Aβ
For quantification of Aβ using ELISA, the cortex of the right hemisphere of the brains of APPswePS1∆E9 mice was homogenized in TBS/0.1% Triton X-100 containing 2% complete protease inhibitor cocktail (Roche Diagnostics Ltd., Mannheim, Germany) (pH 7.2) and centrifuged (21,000× g, 10 min). The supernatant containing the extracellular soluble Aβ was obtained and stored at −80 • C until use, the pellet was sonicated in TBS containing 2% complete protease inhibitor cocktail (Roche Diagnostics Ltd.) and centrifuged (21,000× g, 10 min). The supernatant containing intracellular soluble Aβ was obtained and stored at −80 • C until use, the pellet was sonicated in 2% sodium dodecyl sulfate (SDS) in distilled water and centrifuged (21,000× g, 10 min). The supernatant containing membrane-associated soluble Aβ was obtained and stored at −80 • C until use, the pellet was sonicated in 70% formic acid in distilled water and centrifuged (44,000× g, 10 min). The supernatant containing insoluble Aβ was obtained and stored at −80 • C until use. In the obtained samples, Aβ 40 and Aβ 42 levels were quantified using an Aβ 40 and Aβ 42 ELISA (Invitrogen, Carlsbad, CA, USA) and related to the total protein content (in the extracellular soluble Aβ fraction) determined with a BCA protein assay kit (Thermo Fisher Scientific, Waltham, MA, USA) according to the manufacturer's instructions.

RNA Isolation and RT-Q-PCR
Hippocampus and liver tissue were homogenized using the BioSpec Mini-Beadbeater (Biospec Products, Bartlesville, OK, USA) and CCF-STTG1 cells were washed with cold PBS. RNA was prepared using Trizol (Invitrogen, Carlsbad, CA, USA) and RNA was reverse transcribed to cDNA using the QuantiTect Reverse Transcription Kit (Qiagen), according to the manufacturer's instructions. Quantitative PCR (qPCR) was conducted, as previously described, on a CFX384 Thermal Cycler (Bio-Rad Laboratories) using SYBR Green PCR Select Master Mix (Applied Biosystems, Warrington, UK) [55]. Relative quantification of gene expression was accomplished by using the comparative Ct method. Data were normalized to the most stable reference genes (Actb, B2m, Hprt1, and Sdha (hippocampus) or Actb (liver)), which were analyzed and selected with geNorm 3.5 and StepOnePlus. Expression levels are indicated by fold change values and compared to the WT mice on the vehicle control. Details of the primers used are shown in Table 1.

Hepatic Neutral Lipid Quantification
Tissue Tek-embedded livers were cut with a cryostat CM3050S (Leica, Wetzlar, Germany) to obtain 14 µm sections which were mounted on SuperFrost Plus adhesion slides (Thermo Fisher Scientific, Waltham, MA, USA), air-dried overnight, and stored at room temperature until use. For hepatic neutral lipid staining, liver sections were fixed in 4% neutral buffered formalin, washed with tap water, and rinsed with 60% isopropanol. Hepatic lipids were stained with Oil Red O (Polysiences Inc., Warrington, FL, USA) for 15 min. Next, the liver sections were rinsed with 60% isopropanol, lightly stained with hematoxylin, and covered with a coverslip. Digital images of the sections were obtained using a Leica DMLB microscope (Leica Microsystems, Rijswijk, The Netherlands) equipped with software from the Leica Applications Suite (Leica Microsystems, Rijswijk, The Netherlands).

Statistical Analyses
All statistical analyses were performed using IBM SPSS Statistics 25. The Shapiro-Wilk normality test was used to test normal distribution. Unless stated otherwise, normally distributed data are presented as mean ± SD and analyzed using two-way ANOVA (with treatment and genotype as independent variables) and the Tukey post hoc test. Notnormally distributed data are presented as median (25th-75th percentile) and analyzed using the Mann-Whitney U test. The OLT and ORT discrimination index D2 and the fold change values of CCF-STTG1 cells (compared to the DMEM-F12 medium control) were analyzed using a one-sample T-test. Animals that did not reach the minimum of 4 s of exploration in T1 or T2 were excluded from further analyses. Extreme values were excluded using Dixon's principles of exclusion of extreme values [56,57]. Significance are denoted as follows: * p ≤ 0.05, ** p ≤ 0.01, and *** p ≤ 0.001.