Partial Serotonin Transporter Deficiency Modulates Plasma Metabolome, Arginine-Nitric Oxide Pathway and Emotional Behavior in Mice Exposed to Western Diet

Abstract

Background/Objectives: Reduced serotonin transporter (SERT) function is associated with increased vulnerability to emotional and metabolic dysregulation, particularly in elderly women. Most preclinical studies relied on young male rodents with complete Sert deficiency; the Western diet (WD) acerbates these abnormalities. However, complete Sert loss does not fully reflect the human condition of partial SERT dysfunction. Here, we examined the effects of WD in aged female Sert^+/− mice on metabolic, biochemical, molecular, and behavioral outcomes. Methods: Wild-type (WT) and Sert^+/− mice were fed WD or a control diet. Emotionality, cognition, glucose tolerance (GT), plasma ¹HNMR spectroscopy metabolome and biochemical parameters were studied. Gene expression analyses of nitric oxide (NO)-related markers were performed in the hypothalamus, dorsal raphe, and liver. Results: WD-exposed WT mice showed impaired GT and reduced plasma lactate and branched-chain amino acid levels; metabolome changes were more pronounced in mutants, while GT was unchanged. Naïve Sert^+/− mice exhibited lower lactate and alanine levels compared with WT controls. WD increased leptin and cholesterol levels in both genotypes, whereas triglyceride concentrations were reduced in Sert^+/− mice. Both WD and Sert deficiency increased Nos expression, while arginase expression was differentially regulated by genotype and diet. Malondialdehyde levels were elevated in the prefrontal cortex of Sert^+/− mice regardless diet. WD also impaired object recognition memory and induced anxiety- and depression-like behaviors, with more pronounced effects in Sert^+/− mice, except marble test behavior. Conclusions: Partial Sert deficiency aggravates some but not all WD-induced metabolic alterations, enhances oxidative stress, dysregulates arginine–NO signaling, and modifies behavior, highlighting the translational relevance of Sert^+/− mice for modeling SERT dysfunction.

1. Introduction

Evidence indicates that reduced function of the serotonin transporter (SERT), which regulates serotonin reuptake and thereby influences both central and peripheral serotonergic signaling [1], is associated with increased vulnerability to mood and anxiety disorders and metabolic dysregulation [2,3]. In humans, the most common functional variation involving a polymorphic promoter region in the SLC6A4 gene (5-HTTLPR), the “short” (S) allele, moderates the environmental effects on these conditions [2,4,5,6,7]. This polymorphism is associated with weight gain and altered responses to high-calorie nutritional interventions [8,9]. Neuroimaging studies have further suggested a relationship between decreased SERT binding, insulin resistance, obesity and serotoninergic regulation [10,11,12].

Environmental factors, such as the consumption of a Western-style diet (WD), which is high in saturated fats, cholesterol, and sugars, are major contributors to metabolic syndrome, non-alcoholic fatty liver disease (NAFLD), and behavioral alterations [13]. The global adoption of WD-like nutrition has increased sharply, resulting in an increased incidence of obesity, diabetes, and metabolic syndrome [14]. At the molecular and cellular levels, systematic WD intake induces systemic inflammation, oxidative stress, and mitochondrial dysfunction [15], as well as reduced SERT expression and activity [16,17]. These mechanisms are considered potential targets for therapeutic and preventive interventions, the need for which is rapidly increasing due to the growing epidemiology of metabolic disorders and obesity.

The use of animal models is indispensable in the investigation of new therapeutic strategies for metabolic syndrome and related complications, particularly in the context of the epidemiologically spread S-SLC6A4 polymorphism. However, most preclinical studies have relied on young male rodents with complete SERT deficiency, whereas complete Sert loss does not fully reflect the human condition. Here, we examined the effects of WD on metabolic, biochemical, molecular, and behavioral outcomes in aged female Sert^+/− mice. Therefore, we used a WD paradigm based on a 3-week exposure to experimental mouse chow with high sugar, saturated fat, and cholesterol content provided by Research Diets, a well-established dietary paradigm that reliably recapitulates the key features of human metabolic syndrome and NAFLD in female mice [18,19,20,21]. In particular, this model has been extensively validated in our laboratory and has been shown to recapitulate abdominal obesity, insulin resistance, hepatic steatosis, and associated neuroinflammation and aberrant behaviors.

WD exposure is known to exacerbate metabolic abnormalities in the presence of genetic susceptibility, such as Sert deficiency, providing mechanistic insight into these gene × diet interactions [22,23,24,25,26,27,28,29,30,31]. For instance, aged female Sert^−/− mice exposed to a 3-week WD showed exacerbated glucose intolerance and suppressed hypothalamic and hepatic expression of the mitochondrial regulators Ppargc1a and Ppargc1b. Housing on WD also increased the expression of the inflammation marker, Toll-like receptor 4 (Tlr4), in the prefrontal cortex and dorsal raphe of Sert^−/− mice, but not in Sert^+/− mice [23]. Sert^−/− female mice also displayed hepatosteatosis and overexpression of insulin receptor isoforms A and B in the hippocampus and hypothalamus [26,27]. Likewise, male Sert^−/− mice exposed to WD for 19 weeks showed impaired glucose and insulin tolerance, liver hypertrophy, increased visceral adiposity, adipose tissue inflammation, and hepatic steatosis accompanied by reduced insulin receptor substrate 1 (Irs1), Glut2, and Glut4 gene expression [30,31].

Metabolic abnormalities in naïve and WD-fed Sert^−/− mice are accompanied by aberrant behaviors, such as reduced novelty exploration, impulsivity, disrupted hippocampal-dependent behavior, and helplessness [23,26,32]. A recent metabolomic study using nuclear magnetic resonance (NMR) spectroscopy revealed elevated plasma leptin and lipid levels and decreased glucose, lactate, and amino acid levels in naïve and WD-challenged Sert^−/− mice compared to wild-type WD-fed animals. However, metabolomic responses to WD in Sert^+/− rodents have not yet been reported.

While constitutive Sert knockout mice provide valuable insights into serotonergic mechanisms associated with metabolic disturbances, the inherited lack of Sert triggers developmental compensation and neurochemical adaptations that may substantially distort regulatory processes from partial transporter reductions in S-SLC6A4 careers [33]. Hence, partial Sert deficiency models may better approximate human gene × environment interactions, which warrants the importance of studies using Sert^+/− rodents.

As female sex is a major biological determinant of vulnerability to diet-induced metabolic and behavioral disturbances [18], particularly during aging, we chose to use female mice in the current study. Previous studies have demonstrated that aged female rodents display increased susceptibility to WD-induced adiposity, glucose intolerance, and inflammatory responses [34,35]. This pattern is consistent with epidemiological data showing a higher BMI, faster progression to obesity, and increased prevalence of type 2 diabetes in women exposed to obesogenic environments [36,37].

Age is a critical factor in the compromise of both metabolic homeostasis and serotonergic function [38,39]. Age-related declines in SERT expression, serotonin synthesis, transport, and receptor sensitivity have been evidenced by positron emission tomography (PET) [40,41]. This is concurrent with lowered mitochondrial efficiency, oxidative capacity, and nitric oxide (NO) bioavailability, further eroding metabolic flexibility and cellular resilience, and contributing to an increased risk of insulin resistance [42,43,44,45,46,47,48]. In addition, as estrogens protect against type 2 diabetes by enhancing insulin sensitivity and β-cell function, this protective effect diminishes with aging due to menopause-related estrogen decline, thereby elevating the risk of diabetes [47]. Rodent studies have also demonstrated that aging exacerbates diet-induced metabolic dysfunction [43].

Given the escalating global burden of metabolic syndrome and NAFLD, particularly in the older population, which affects over 1 billion individuals worldwide and contributes to 4 million deaths annually, it is critically important to address the mechanisms of these conditions in preclinical models using aged animals [49,50,51]. These risks are especially high in older female populations, which face a 2–3-fold higher incidence of NAFLD and insulin resistance than young cohorts of women [36,52]. In addition, the percentage of S-SLC6A4 polymorphism carriers is higher in women than in men [7,8]. However, the currently available therapies for insulin resistance remain largely insufficient [53], while the socioeconomic costs resulting from these morbidities are projected to reach $2.2 trillion annually by 2030 [54]. As such, current preclinical research in this field is of high medicinal and social relevance [23,25,27,55,56]. In the present study, we aimed to fill the gaps in this pre-clinical research by addressing the metabolic and physiological responses of aged female Sert^+/− mice to WD and comparing the outcomes with the biochemical, molecular, and behavioral changes in Sert^−/− animals reported previously [22,23,25,55,56,57].

Therefore, in the present study, plasma samples were analyzed using NMR spectroscopy, followed by metabolite profiling and measurements of cholesterol, leptin, triglycerides, and total protein levels, which are known to be altered in the WD model [23]. We also performed gene expression analyses of nitric oxide (NO)-related markers in the hypothalamus, dorsal raphe, and liver. Because nitric oxide (NO) signaling links metabolic stress with serotonergic function, we additionally analyzed the expression of iNos, nNos, eNos, arginase (Arg), Arg1, and Arg2 in the brain and liver. The NO pathway is a key downstream target of both WD exposure and reduced SERT function and is implicated in anxiety- and depression-related phenotypes via serotonergic modulation [58,59,60]. Arginase regulates arginine availability for NO synthesis. WD-induced obesity increases hepatic Arg1 expression, reduces L-arginine levels, and contributes to insulin resistance and impaired NO signaling [61,62]. Elevated extracellular serotonin inhibits nNOS [63] and shifts arginine metabolism toward arginase-dependent pathways, lowering NO bioavailability and promoting endothelial dysfunction and inflammation [64]. Consistently, WD-fed Sert^−/− mice exhibit elevated hepatic Arg2 expression [27]. The increased risk of insulin resistance and type 2 diabetes in S-SLC6A4 carriers is mediated by serotonin-driven arginase upregulation and arginine depletion [65,66]. WD exposure may amplify these effects by inducing oxidative stress and hypercholesterolemia, which impairs NO signaling through arginase induction and eNOS uncoupling [67].

In addition, we assessed the malondialdehyde (MDA) concentration, a marker of oxidative stress that is functionally related to NO transmission in the prefrontal cortex [68]. WD-exposed WT and Sert^+/− mice were also studied for anxiety-like, exploratory, and depression-like behaviors in novel cages, elevated O-mazes, step-down anxiety, and forced swim models, as well as for object recognition memory and hippocampal-dependent performance in the marble test. These outcomes enabled the assessment of the translational value of Sert^+/− mice in modeling physiological responses to WD in aged female S-SLC6A4 carriers.

2. Materials and Methods

2.1. Animals

Experiments were conducted using 12-month-old female mice that were either heterozygous Sert^+/− or wild-type littermates, all derived from heterozygous mutants at the 10th generation (F10) of backcrossing with C57BL/6J mice. All genotypes were confirmed by PCR. Mice were housed three to four per cage under a reversed 12 h light–dark cycle (lights on at 21:00 h), with ad libitum access to food and water, and maintained under controlled laboratory conditions (22 ± 1 °C, 55% humidity). To minimize potential environmental influences, behavioral testing was conducted during the dark phase of the animals’ light cycle (after 9:00 h), and other potential confounding factors were controlled, as described elsewhere [69]. The animals were monitored every morning and evening throughout the experimental period. All experimental procedures were conducted in accordance with the ARRIVE guidelines and the European Communities Council Directive for the Care and Use of Laboratory Animals (2010/63/EU) and were approved by the local ethics committee of C. Bernard University (Ethical Committee of C. Bernard University on animal care and welfare), approval date 1 July 2017, approval code CBU08RC2017) and were compliant with ARRIVE guidelines (http://www.nc3rs.org.uk/arrive-guidelines, 2 January 2022).

2.2. Study Flow and Dietary Challenge

Mice were fed either a standard laboratory diet (control diet, CD; groups WT/CD, n = 7 and Sert^+/−/CD, n = 7) or a Western diet (WD; groups WT/WD, n = 7 and Sert^+/−/WD, n = 7) for three weeks, as described previously [23,70]. A total of 28 mice were used in this study. The groups were randomized according to body weight; statistical power was evaluated as described previously [23,70]; experimenter was blind for animals’ ID in a course of the behavioral evaluation and physiological manipulations. No specific inclusion or exclusion criteria were defined; this study had no human endpoints. Potential confounding factors were systematically controlled as described elsewhere [69], and all mice were given a 10-day acclimatization period prior to the experiment. The CD provided 3.8 kcal/g and contained 4.3% fat (1.3% saturated fat) (D18071801, Research Diet Inc., New Brunswick, NJ, USA), whereas the WD provided 4.6 kcal/g and contained 0.2% cholesterol and 21.3% fat (10.5% saturated fat) (D11012302, Research Diet Inc., New Brunswick, NJ, USA). The nutrient composition and ingredient details are presented in Supplementary Table S1 (see Supplementary File). Mice were exposed to CD or WD diets for 21 days. On days 20–23, all mice were subjected to the novel cage, O-maze, marble, object recognition, forced swim, step-down anxiety, and glucose tolerance tests (Figure 1A; see below). On the day 24, the animals were euthanized, and their brains and livers were dissected for gene expression analysis (see below). Blood samples were collected for metabolomic and biochemical analysis (see below).

2.3. Behavioral Tests

2.3.1. Novel Cage

A novel cage test was conducted to evaluate exploratory behavior in a novel environment. Mice were placed in a standard plastic cage (21 × 27 × 14 cm) containing a small amount of fresh bedding under red light. The number of exploratory rearings was manually recorded every minute for 5 min, as described previously [71].

2.3.2. Step-Down Anxiety Test

The apparatus (Evolocus LLC, Tarrytown, NY, USA; Technosmart, Rome, Italy) consisted of a transparent plastic chamber (25 cm × 25 cm× 48 cm) with a stainless-steel grid floor. A small wooden platform (7 cm× 7 cm× 1.5 cm) was positioned on the grid floor, and the illumination level was set at 25 lx. Each mouse was placed on the platform and enclosed within a transparent cylinder for 30 s to prevent immediate movement. After the cylinder was removed, the latency to step down from the platform with all four paws was recorded as the step-down latency [69].

2.3.3. Marble Test

All experimental groups were evaluated for pellet displacement in the marble test, following previously described methods [69,72]. The innate tendency of mice to remove small objects, such as stones or food pellets, from a tube placed in their cage is species-specific and depends on the integrity of the hippocampal formation [72]. In this test, a paper tube (4 cm internal diameter, 10 cm length) filled with 20 food pellets was placed in the home cage of the experimental animals, and the latency to displace the first pellet as well as the latency to empty the tube were recorded.

2.3.4. Object Recognition Learning

Mice were assessed for novel object exploration and recognition in a two-day test, as described previously [24]. The apparatus consisted of a plastic cage (21 cm × 27 cm× 14 cm) with opaque walls and two objects (“brush” and “flower,” 7 cm × 4 cm ×3 cm) placed symmetrically 2 cm from the walls in opposite corners. The objects used in the novel object recognition test were either disposable and unique to each mouse (paper flowers) or washable and reusable (plastic brushes), ensuring the exclusion of olfactory contamination from previous subjects. This method has been validated extensively in previous studies [24,73]. Illumination was maintained at 5 lx. On day 1, two identical objects were presented, and each mouse was placed in a cage equidistant from both objects and allowed to explore freely for 15 min. On day 2, one object was replaced with a novel object, and the mice were again allowed to explore for 15 min. Object exploration was defined as the mouse directing its nose toward the object at a distance of less than 2 cm, and was scored offline. The percentage of time spent exploring the novel object relative to the total exploration time was compared to a 50% chance level of random exploration and used as a measure of object recognition memory. The latency to explore the old and new objects was also recorded. All behaviors were manually scored offline.

2.3.5. Elevated O-Maze

The maze consisted of a black circular runway (5.5 cm wide and 46 cm in diameter) elevated 20 cm above the floor. The illumination level was maintained at 5 lx during the experiment. Two opposite sections of the maze were enclosed with 10 cm-high walls. Each mouse was placed in one of the enclosed compartments at the beginning of the test. The latency to enter the anxiety-related open sections of the maze was recorded as previously described [55].

2.3.6. Forced Swim Test

The test was conducted as previously described [74]. Mice were placed in a transparent plastic cylinder (diameter 17 cm, height 20 cm) filled with water (23 °C, depth 13 cm) under low illumination (25 lx) for 6 min to assess depression-like behavior. Floating behavior, defined as the absence of directed movements of the head and body, was analyzed offline using an automated video-tracking system (Viewpoint, Lyon, France).

2.3.7. Glucose Tolerance Test

A glucose tolerance test was performed as described previously [26]. The mice were fasted overnight for 18 h, starting at 16:00. Following the fasting period, a glucose solution (2 g/kg, 1.8 g/L) was administered by oral gavage, and blood samples were collected from the tail vein. Sampling was performed immediately before glucose administration and at 5, 15, 30, and 60 min thereafter. Blood glucose levels were measured using a OneTouch UltraEasy glucometer with corresponding test strips (LifeScan OneTouch, Dubai, United Arab Emirates). The area under the curve (AUC) for the 0–30- and 0–60 min intervals following glucose administration was calculated.

2.4. Culling and Tissue Samples Collection

The mice were terminally anesthetized using isoflurane. Blood collection was performed transcardially, blood was stored in heparinized vials prior to centrifugation (1500 rcf, 15 min, 4 °C); plasma was removed and immediately stored at −80 °C until use [75]. Following blood collection, each mouse brain was perfused with saline and dissected as described previously [26,27]. Brain regions were isolated according to Paxinos and Franklin’s Mouse Brain in Stereotaxic Coordinates, including the prefrontal cortex, hypothalamus, hippocampus, and dorsal raphe region. The livers were also dissected. The collected tissue samples were stored at −80 °C until further use in the PCR assay (see below).

2.5. RNA Extraction, cDNA Synthesis, and Real-Time Polymerase Chain Reaction

Total RNA from prefrontal cortex, dorsal raphe, striatum, hippocampus and liver samples was extracted using QIAzol^® Lysis Reagent (Qiagen Sciences Inc., Germantown, MD, USA). Each tissue sample was placed in 1 mL of QIAzol and homogenized with a TissueRuptor (Qiagen Sciences Inc., Germantown, MD, USA) for two 30 s cycles at half speed, with the samples kept on ice for one minute between cycles. The homogenates were centrifuged at 12,000× g for 15 min at 4 °C to remove cellular debris. Chloroform was then added, and the samples were centrifuged again under the same conditions to separate the phases of the samples. The aqueous phase was collected, and RNA was precipitated with ethanol. The resulting RNA pellet was washed and purified using an RNeasy Mini Kit (Qiagen, Hilden, Germany). RNA concentrations were determined using a NanoDrop spectrophotometer (Thermo Fisher Scientific, Waltham, MA, USA). For cDNA synthesis, 1 µg of total RNA was reverse-transcribed using random primers and the QuantiTect Reverse Transcription Kit (Qiagen, Hilden, Germany), according to the manufacturer’s protocol. The reaction was carried out in an Eppendorf Mastercycler^® Gradient (Eppendorf SE, Hamburg, Germany) under the following conditions: 68 °C for 5 min, 42 °C for 60 min, and 70 °C for 10 min.

Gene expression analysis was performed using real-time polymerase chain reaction (RT-PCR) with SYBR Green Master Mix (Applied Biosystems, Foster City, CA, USA). Specific primers for the target genes were designed, with Gapdh (Glyceraldehyde-3-phosphate dehydrogenase) serving as the housekeeping reference gene. Primer sequences are listed in Table S2 (Supplementary File). Each RT-PCR reaction (10 µL total volume) contained 5 µL SYBR Green Master Mix, 3 µL RNase-free water, 1 µL primer mix (20 pmol/µL), and 1 µL of cDNA. The thermal cycling program consisted of an initial denaturation step at 95 °C for 5 min, followed by 40 cycles at 95 °C for 20 s, 60 °C for 30 s, and 68 °C for 30 s. All primers were obtained from Sigma-Aldrich (St. Louis, MO, USA). Each sample was analyzed in triplicate using an ABI Prism 7900 HT SDS system (Applied Biosystems, Foster City, CA, USA). Relative gene expression levels were normalized to Gapdh and calculated as fold changes compared to the WT mice fed CD, as described previously [76].

2.6. Nuclear Magnetic Resonance (NMR) Spectroscopy and Metabolome Assay

NMR spectroscopy, data processing, and analysis were performed as described previously [26]. Briefly, 150 µL of plasma was mixed with 400 µL of 75 mM sodium phosphate D₂O buffer (pH 7.4). Carr–Purcell–Meiboom–Gill (CPMG) spectra were acquired using a 700 MHz Bruker AVII spectrometer (Bruker Daltonics GmbH & Co., Bremen, Germany). Spectra were manually phased, baseline-corrected with a third-degree polynomial, and referenced to the CH₃-lactate doublet at δ = 1.33 ppm using TopSpin 4.0.5 (Bruker Daltonics GmbH & Co., Bremen, Germany). The processed spectra were then exported to ACD/Labs Spectrus Processor Academic Edition 12.01 (Advanced Chemistry Development, Inc., Toronto, ON, Canada), segmented into 0.02 ppm bins, and the integrals of each bin were calculated. The integrals were normalized to the total spectral area and Pareto-scaled prior to statistical analysis. The acquired spectra were manually phased, baseline-corrected, and referenced to the lactate-CH₃ doublet resonance at δ = 1.33 ppm using TopSpin 4.0 (Bruker, Germany). The processed spectra were exported to the ACD/Labs Spectrus Processor Academic Edition 12.01 (Advanced Chemistry Development, Inc., Toronto, ON, Canada) for further analysis. Each resonance signal, excluding the water region, was manually bucketed, and the integral of each bin was normalized to the total spectral integral for every sample.

2.7. Orthogonal Partial Least Squares Discriminant Analysis (OPLS-DA)

Data were analyzed using orthogonal partial least squares discriminant analysis (OPLS-DA) to assess differences between two predefined experimental groups. OPLS-DA is a supervised multivariate method that models the relationship between measured variables and class membership while separating variation that is predictive of group differences from variation that is orthogonal (unrelated) to group separation.

To evaluate the robustness and predictive performance of the model, the dataset was repeatedly divided into training and test sets. For each iteration, the model was built using the training data and then applied to the test data, which were classified in a blinded manner. Model performance was assessed by comparing the accuracy of classification against that expected by chance, allowing evaluation of whether group discrimination was significantly better than random assignment.

The contribution of individual variables to group separation was assessed using Variable Importance in Projection (VIP) scores. VIP values reflect the relative importance of each variable in the OPLS-DA model, with higher scores indicating a greater contribution to class discrimination.

2.8. Blood Biochemical Analysis

2.8.1. Leptin Concentration

Leptin concentrations were determined using a commercially available Mouse Leptin (OB) ELISA Kit (Sigma-Aldrich, St. Louis, MO, USA). Optical density was measured at 450 nm using a Wallac 1420 VICTOR plate reader (PerkinElmer, Waltham, MA, USA). All samples were analyzed in duplicate, and the procedures were performed according to the manufacturer’s instructions, as described previously [23].

2.8.2. Cholesterol Concentration

Cholesterol concentration was determined using a commercially available Mouse Total Cholesterol ELISA Kit (Abcam, Cambridge, UK). Optical densities were measured at 450 nm using a Wallac 1420 VICTOR plate reader (PerkinElmer, Waltham, MA, USA). All samples were analyzed in duplicate, and the procedures were performed according to the manufacturer’s instructions, as described elsewhere [23].

2.8.3. Triglyceride Concentration

Triglyceride concentrations were determined using a commercially available Triglyceride Assay Fluorometric Kit (Abcam, Cambridge, UK). Optical density was measured at 570 nm using a Wallac 1420 VICTOR plate reader (PerkinElmer, Waltham, MA, USA). All samples were analyzed in duplicate, and the procedures were performed according to the manufacturer’s instructions and as described elsewhere [23].

2.8.4. Total Protein Concentration

Total protein concentrations were quantified using the Pierce BCA Protein Assay Kit (Thermo Fisher Scientific, Waltham, MA, USA). Optical density was measured at 562 nm using a Wallac 1420 VICTOR plate reader (PerkinElmer, Waltham, MA, USA). All samples were analyzed in duplicate, and the procedures were performed according to the manufacturer’s instructions and as described elsewhere [23].

2.9. Statistical Analysis

Data were analyzed using GraphPad Prism (version 8.0.2; GraphPad Software, San Diego, CA, USA). The normality of the data distribution was assessed using the Shapiro–Wilk test. For comparisons among the four groups, two-way analysis of variance (ANOVA) followed by Tukey’s post hoc test was used, which applies the necessary correction for multiple comparisons to minimize the false-positive results. A one-sample t-test was used to compare the data with chance levels in the object recognition test. No data points were excluded from statistical analysis. Statistical significance was set at p < 0.05. Data are presented as mean ± SEM, with group sizes specified in the figure legends.

3. Results

3.1. Western Diet and Partial SERT Deficiency Affected Glucose Tolerance, Emotionality and Learning

A significant diet effect was observed for the AUC of 30 min of the glucose tolerance test in absolute values and in percent to basal glucose level (F = 8.6, p = 0.0097 and F = 6.91, p = 0.0181, respectively, two-way ANOVA) without significant effects of genotype (F = 0.08, p = 0.77 and F = 0.94, p = 0.34) or their interaction (F = 2.56, p = 0.128 and F = 2.71, p = 0.12). Both these measures were significantly higher in WD-fed control animals than in WT CD group (p = 0.0055 and p = 0.0081, respectively; Tukey’s test; Figure 1B,C). Similarly, a significant diet effect was revealed for the AUC of 60 min of glucose tolerance test in absolute values and in percent to basal glucose level (F = 11.92, p = 0.0033 and F = 8.51, p = 0.0101, respectively) without significant genotype effect (F = 0.06, p = 0.805 and F = 0.809, p = 0.381) but with a strong trend for their interaction (F = 3.53, p = 0.078 and F = 3.57, p = 0.077, two-way ANOVA). The both parameters were significantly elevated in the WT WD group compared to those in the WT CD group (p = 0.0017 and p = 0.0037, respectively, Tukey’s test; Figure 1D,E).

Two-way ANOVA revealed a strong trend for genotype effect (F = 3.96, p = 0.06), no diet effect (F = 0.52, p = 0.48), and no genotype × diet interaction (F = 2.1, p = 0.17) for the number of rears during the last two minutes of the novel cage test. This measure was significantly reduced in WD-fed Sert^+/− mice compared to that in WD-fed WT mice on the same diet (p = 0.02; Tukey’s test; Figure 2A). No other significant changes were observed in this assay. A significant genotype effect (F = 11.63, p = 0.003), no diet effect (F = 0.15, p = 0.7), and no genotype × diet interaction (F = 0.02, p = 0.88) were revealed for the duration of floating for the last two minutes of the forced swim test. This parameter was increased in Sert^+/− mice compared to the control group, both on CD (p = 0.02; Tukey’s test) and WD (p = 0.03; Figure 2B). Two-way ANOVA indicated no significant genotype effect (F = 2.38, p = 0.14), a significant diet effect (F = 6.1, p = 0.02), and a trend for genotype × diet interaction (F = 3.7, p = 0.07) for the latency to descend in the step-down test, which was significantly increased in WD-fed Sert^+/− mice (p = 0.005; Tukey’s test; Figure 2C).

A significant genotype effect (F = 9.72, p = 0.006), a significant diet effect (F = 9.24, p = 0.007), and a trend for genotype × diet interaction (F = 3.78, p = 0.07) were demonstrated for the latency to explore the non-anxiogenic area in the object recognition test. This measure was significantly increased in the WD-fed Sert^+/− group compared to Sert^+/− mice on a control diet (p = 0.002; Tukey’s test) and the WT-WD group (p = 0.003; Figure 2D). Simultaneously, two-way ANOVA revealed no significant effect of genotype (F = 1.97, p = 0.18), diet (F = 0.75, p = 0.4), or their interaction (F = 1, p = 0.25; Figure 2E) on the latency to explore the anxiogenic area. Similarly, no significant effects on novel object preference were observed (genotype: F = 1.6, p = 0.22; diet: F = 2.88, p = 0.11; interaction: F = 0.38, p = 0.55). However, only the WD CD group demonstrated a significant difference in this measure from the chance level (p = 0.0002, one-sample t-test; Figure 2F).

For the number of exits and duration of exits in the elevated O-maze, two-way ANOVA revealed a significant genotype effect (F = 4.65, p = 0.04 and F = 4.69, p = 0.044), but no diet effect (F = 0.99, p = 0.33 and F = 0.42, p = 0.52) and no genotype × diet interaction (F = 0.26, p = 0.62 and F = 0.45, p = 0.51, respectively). A strong trend for a decrease in both these measures was revealed for the WD-fed Sert^+/− group compared with Sert^+/− mice on CD (p = 0.075 and p = 0.059, respectively, Tukey’s test; Figure 2G,H). In the food pellet displacement test, no significant effects of genotype, diet, or their interaction were observed on the latency to displace the first pellet (F = 0.32, p = 0.57; F = 0.67, p = 0.42; and F = 0.67, p = 0.43, respectively, two-way ANOVA) and on the latency to empty the tube (F = 0.15, p = 0.704; F = 0.677, p = 0.39; and F = 0.3, p = 0.59, respectively; Figure 2I,J).

3.2. Western Diet and Partial SERT Deficiency Altered Metabolome Parameters

Distinct metabolite-associated peaks and spectral regions were observed in the ¹H NMR plasma spectra (Figure S1, see Supplementary File). Following spectral binning into 0.02 ppm intervals, integrals were calculated for each bin, reflecting relative metabolite concentrations. Our results revealed high accuracy of OPLS-DA analysis (Figure 3A–H, see also Table S1, Supplementary File). Multivariate analysis revealed clear differences between the experimental groups. Several spectral bins contributed strongly to group separation, indicating altered metabolite profiles. These discriminating bins were subsequently assigned to their corresponding metabolites, identifying specific metabolic changes associated with group differences (Figure 3I).

Statistical analysis of lactate and alanine levels in the blood of experimental groups showed significant main effects of genotype (lactate level: F = 12.05, p = 0.002; alanine level: F = 8.29, p = 0.008) and diet (lactate level: F = 8.78, p = 0.0068; alanine level: F = 33.13, p < 0.0001). A significant genotype × diet interaction was specific to lactate concentrations (F = 4.81, p = 0.0381), whereas this interaction was non-significant for alanine levels (F = 0.02, p = 0.878). Both Sert^+/− CD and WT-WD mice showed significantly lower lactate levels than WT-CD mice (p = 0.0007 and p = 0.0017, respectively; Figure 4A). Both WD-fed WT and WD-fed Sert^+/− mice had significantly lower alanine levels than the respective CD-fed controls (p = 0.0005 and p = 0.0004, Tukey’s test). It was also significantly lower in CD-fed Sert^+/− animals than in WT CD group (p = 0.0492; Figure 4B). For blood glucose, significant main effects of genotype (F = 7.65, p = 0.01) and diet (F = 6.46, p = 0.02) were observed without significant interaction (F = 0.398, p = 0.53). Glucose levels were significantly increased in WT WD mice compared to those in the WT CD group (p = 0.04; Tukey’s test), but Sert^+/− WD mice had significantly lower glucose levels than their WT-WD counterparts (p = 0.02; Figure 4C).

For isoleucine and valine levels, two-way ANOVA confirmed significant diet (isoleucine: F = 44.95, p < 0.0001; valine level: F = 33.68, p < 0.0001) and genotype effects (valine level: F = 8.46, p = 0.008; isoleucine level: F = 5.6, p = 0.06), with no significant interactions (isoleucine level: F = 0.003, p = 0.96; valine level: F = 3.35, p = 0.08). These parameters were substantially reduced by WD in both genotypes (p < 0.0001 for both; Figure 4D,E). For unsaturated lipids concentrations, HDL and = CH-CH2-CH= levels no significant effect of genotype (F = 1.1, p = 0.304; F = 0.35, p = 0.56, and F = 1.26, p = 0.27, two-way ANOVA), diet (F = 1.15, p = 0.29; F = 0.26, p = 0.61, and F = 0.53, p = 0.47) and their interaction (F = 0.16, p = 0.69; F = 1.9, p = 0.17, and F = 0.13, p = 0.72, respectively; Figure 4F,H,I) were found. For VLDL levels, a significant diet effect was observed (F = 5.26, p = 0.0309, two-way ANOVA) without a significant effect of diet or genotype × diet interaction (F = 0.81, p = 0.38 and F = 0.04, p = 0.82, respectively; Figure 4G).

A heatmap showing the relative relationship of metabolites in experimental groups of CD-fed, WD-fed WT and Sert^+/− mice is presented on Figure 5.

3.3. Effects of Western Diet and Sert^+/− Genotype on the Expression of Arginases and Nitric Oxide Synthases

Arg1 expression in the dorsal raphe was significantly affected by the genotype × diet interaction (F = 9.58, p = 0.006), whereas the main effects of genotype (F = 3.1, p = 0.086) and diet (F = 0.023, p = 0.88) were not significant. Post hoc analysis demonstrated reduced Arg1 expression in the dorsal raphe of WT WD mice compared to that in WT CD controls (p = 0.048, Tukey’s test, Figure 6A). The WD-fed Sert^+/− group showed significantly elevated Arg1 expression compared to both Sert^+/− CD and WT WD groups (p = 0.0303 and p = 0.029, respectively; Figure 6B). At the same time, two-way ANOVA showed no significant effects for Arg1 expression in the hypothalamus (genotype: F = 2.49, p = 0.13; diet: F = 2.81, p = 0.11; interaction: F = 2.74, p = 0.11; Figure 6A) or liver (genotype: F = 2.17, p = 0.16; diet: F = 0.6, p = 0.16; interaction: F = 0.11, p = 0.75; Figure 6C). For Arg2 expression in the hypothalamus and liver, two-way ANOVA indicated a significant diet effect (F = 4.6, p = 0.04 and F = 4.4, p = 0.048, respectively), with no significant genotype effect (F = 0.47, p = 0.5 and F = 1.86, p = 0.18) or interaction (F = 1.29, p = 0.27 and F = 2.26, p = 0.15, two-way ANOVA). WD-fed Sert^+/− mice exhibited significantly upregulated Arg2 expression in the hypothalamus and liver compared to Sert^+/− CD group (p = 0.024 and p = 0.0097, respectively; Figure 6D,F). A trend for increased hepatic Arg2 expression in WD-fed Sert^+/− mice compared to WD-fed WT mice was also observed (p = 0.0523). Two-way ANOVA revealed no significant main effects on Arg2 expression in the dorsal raphe, except for a strong trend for genotype × diet interaction (genotype: F = 0.001, p = 0.97; diet: F = 0.26, p = 0.61; interaction: F = 3.85, p = 0.06; Figure 6E).

Two-way ANOVA identified a significant diet effect (F = 4.99, p = 0.038) and genotype × diet interaction (F = 6.84, p = 0.017) for hepatic iNos expression, with no significant genotype effect (F = 1.35, p = 0.26). Sert^+/− WD-fed mice showed significantly increased iNos expression in the liver compared to Sert^+/− CD (p = 0.001, Tukey’s test) and WT WD groups (p = 0.016; Figure 6I). Two-way ANOVA showed no significant main effects for iNos expression in the hypothalamus (genotype: F = 2.07, p = 0.16; diet: F = 0.27, p = 0.61; interaction: F = 0.62, p = 0.44; Figure 6G) or dorsal raphe (genotype: F = 0.05, p = 0.82; diet: F = 0.05, p = 0.83; interaction: F = 0.27, p = 0.61; Figure 6H). For hypothalamic eNos expression, two-way ANOVA revealed a significant genotype × diet interaction (F = 6.24, p = 0.02), with no main effects of genotype (F = 2.89, p = 0.10) or diet (F = 0.0001, p = 0.99). The Sert^+/− CD group exhibited significantly elevated eNos expression compared to the WT CD group (p = 0.0085, Tukey’s test) and a trend for its increase when compared to the Sert^+/− WD-fed group (p = 0.077; Figure 6J).

Two-way ANOVA indicated no significant main effects for eNos expression in the dorsal raphe (genotype: F = 0.39, p = 0.54; diet: F = 0.64, p = 0.43; interaction: F = 2.98, p = 0.098; Figure 6K) or liver (genotype: F = 0.35, p = 0.56; diet: F = 0.5, p = 0.49; interaction: F = 1.95, p = 0.18; Figure 6L), except for a strong trend for genotype × diet interaction for this measure in the dorsal raphe. Similarly, no significant main effects were found for nNos expression in the hypothalamus (genotype: F = 0.19, p = 0.67; diet: F = 0.25, p = 0.62; interaction: F = 0.03, p = 0.87; Figure 6M) or dorsal raphe (genotype: F = 0.77, p = 0.39; diet: F = 1.17, p = 0.29; interaction: F = 1.37, p = 0.25; Figure 6N). Finally, two-way ANOVA identified a significant genotype effect on MDA concentration in the prefrontal cortex (F = 40.84, p < 0.0001), with no significant diet effect (F = 0.75, p = 0.79) or their interaction (F = 0.61, p = 0.44). This parameter was significantly higher in both CD-fed and WD-fed Sert^+/− compared to the respective WT groups (p = 0.0005 and p = 0.0034, respectively; Figure 6O).

3.4. Effects of Western Diet and Partial SERT Deficiency on the Blood Biochemical Parameters

A significant effect of diet (F = 18.49, p = 0.001) but no effect of genotype (F = 0.18, p = 0.67) or their interaction (F = 0.92, p = 0.35, two-way ANOVA) was observed on leptin blood concentration. It was significantly higher in WD-fed control and Sert^+/− mice than in the respective controls (p = 0.0422 and p = 0.0022, respectively, Tukey’s test; Figure 7A). Similarly, a significant diet effect (F = 182.1, p < 0.0001) but no effect of genotype (F = 0.19, p = 0.66) or genotype × diet interaction (F = 0.0013, p = 0.971, two-way ANOVA) were revealed for cholesterol blood levels, which were significantly increased in WD-fed control and Sert^+/− mice compared to CD-fed controls (both p < 0.0001, Tukey’s test; Figure 7B). No significant effect of genotype, diet or their interaction were shown for triglycerides (F = 1.98, p = 0.18; F = 2.12, p = 0.17 and F = 0.19, p = 0.66, respectively; Figure 7C) and for total protein level (F = 0.11, p = 0.73; F = 1.45, p = 0.25 and F = 0.009, p = 0.93, respectively; Figure 7D).

4. Discussion

Our study showed that WD exposure compromised glucose tolerance and caused dyslipidemia in WT mice, whereas Sert^+/− mice did not show these abnormalities. However, the WD-induced decrease in circulating lactate and branched-chain amino acids (BCAA) was further exacerbated in Sert^+/− mice compared to that in WT mice. Naïve Sert^+/− mice had lower baseline lactate and alanine levels than WT controls did. The metabolome blood parameters of unsaturated lipid levels, VLDL, HDL, and = CH-CH2-CH= remained unaltered in mutants with partial Sert deficiency, which was different from previously reported data on WD-challenged Sert^−/− mice. The findings reported here demonstrate the high accuracy of OPLS-DA multivariate analysis applied to metabolome data, which revealed clear differences between the experimental groups of WT and Sert^+/− mice exposed to WD or control diets.

Biochemical analysis has shown that WD also elevates leptin and cholesterol levels in both genotypes. Exposure to WD and partial Sert deficiency upregulated Nos expression in several brain regions and the liver of the experimental groups, and the expression of arginase genes was altered by both genotype and WD. Notably, prefrontal cortex MDA concentrations markedly increased in Sert^+/− mice, regardless of dietary conditions. In the novel object recognition model, WD-fed Sert^+/− mice displayed increased latency to approach objects, shortened time spent in the open arms of the elevated O-maze, and prolonged latency to step down in the step-down anxiety test, suggesting elevated anxiety-like behavior similar to the behavioral profile of Sert^−/− mice fed WD. Meanwhile, in the marble test for hippocampus-dependent performance, WD-fed Sert^+/− mice showed unaltered behavior, unlike the previously reported deficit in Sert^−/− mice. Comparative analysis of the metabolic, physiological, behavioral, and phenotypes of Sert^+/− and Sert^−/− mice fed with WD suggests that while many WD-induced changes overlapped between the two genotypes, there were substantial differences in their responses (

Table 1. Comparative behavioral and metabolic phenotypes of Sert^+/− and Sert^−/− mice fed with WD.

Parameter	WD-Sert+/−	WD-Sert−/−
Glucose tolerance	↑	↑
Depressive-like behavior	↑	↑
Anxiety-like behavior	↑	↑
Hippocampal-dependent performance	—	↑
Leptin level	↑	↑
Lactate level	↓	↓
Alanine level	↓	↓
Glucose level	↓	↓
Isoleucin level	↓	↓
Valine level	↓	↓
Unsaturated lipids level	—	↑
VLDL level	—	↑
HDL level	—	↑
=CH-CH2-CH= level	—	↑

) [23,26,27].

Available evidence suggests that physiological differences between rodents with partial and complete Sert deficiency may arise from gene dose-dependent increases in extracellular serotonin and threshold-limited compensatory mechanisms in serotonin neurotransmission. Extracellular serotonin increases approximately threefold in the brains of Sert^+/− mice and up to 9–13 fold in Sert^−/− mutants [77,78], with excessive serotonin impairing insulin sensitivity antioxidant defense and nitric oxide-dependent pathways [29,79,80,81,82]. Despite elevated serotonin levels, Sert^+/− animals largely resemble wild-type controls, whereas Sert^−/− mice display marked physiological alterations, reflecting divergent receptor adaptations and limited non-receptor compensation [23,27,83,84]. Only Sert^−/− animals exhibit persistent serotonergic dysregulation in the brain [78,85]; similar maladaptive serotonergic mechanisms may occur in peripheral organs, such as the gut, liver, and pancreas, potentially extending to carriers of the S-SLC6A4 polymorphism. Together with genotype-specific microbiome alterations [25], these factors likely underlie the distinct physiological responses observed between Sert^+/− and Sert^−/− animals.

The present study demonstrated decreased lactate levels and compromised amino acid metabolism in naïve and WD-challenged Sert^+/− mice, which have also been found in Sert^−/− mice [26]. Disruptions in amino acid and energy metabolism may be a key point of convergence between serotonergic vulnerability and dietary stress [86,87]. Elevated circulating BCAAs levels are considered early markers of insulin resistance. Under prolonged metabolic overload and mitochondrial dysfunction, BCAA levels may decline, reflecting impaired oxidative catabolism and reduced metabolic flexibility [86,87]. This shift might be further exacerbated by serotonergic signaling, as Sert deficiency has been associated with reduced plasma BCAAs concentrations following WD exposure, with more pronounced deficits in aged Sert^−/− mice than in WT littermates [26]. A similar pattern was observed for lactate levels. Chronic WD intake was previously shown to disrupt lactate metabolism, the lowered concentration of which results in compromised glycolysis and astrocyte–neuron metabolic coupling, suggesting a lowered glycolytic flux and impaired mitochondrial integration [88].

Other metabolomic studies have further documented WD-induced alterations in circulating organic and amino acids, including BCAAs and alanine [89]. For example, in male C57BL/6 mice, 12-week WD consumption resulted in reduced hippocampal lactate levels that correlated with the downregulation of glial glutamate transporters GLT-1 and GLAST, suggesting impaired astrocyte–neuron metabolic coupling [88]. Previous studies have shown that SERT deficiency results in mitochondrial dysfunction, impaired glycolysis, dysregulated amino acid metabolism, reduced BCAAs levels (e.g., valine and isoleucine), and decreased alanine concentrations [26,90,91,92].

Notably, both lactate and BCAAs are critical for maintaining NAD(p)H balance [93,94], contributing to mitochondrial oxidative metabolism and neurotransmitter synthesis and turnover [95,96]. Their depletion may further impair serotonergic transmission by limiting tryptophan availability for 5-HT biosynthesis [96,97], thereby exacerbating the consequences of SERT deficiency. These effects might explain the previously reported mitochondrial dysfunction and compromised redox homeostasis in the neural and peripheral tissues of Sert-deficient mutants [90,91]. A demonstrated here increase in MDA levels in the prefrontal cortex, indicative of lipid peroxidation, is in line with these findings.

The current study showed elevated blood glucose concentrations in the WD-fed WT group, where there was no such increase in Sert^+/− mice, unlike the previously reported WD-induced augmentation of glucose concentration in Sert^−/− mice [23]. This pattern aligns with the established association between reduced SERT function, enhanced peripheral tissue glucose uptake, and shifts in β-cell compensatory function [29]. Earlier, increased glucose uptake alongside altered insulin secretion dynamics in Sert^−/− mice was associated with disrupted β-cell function in Sert^+/− and Sert^−/− male mice [29].

In our experiments, WD also elevated circulating blood leptin and cholesterol levels in mice of both genotypes, consistent with the literature [18,98,99]. In previous studies on Sert^−/− mice leptin levels were increased in knockout animals regardless of WD diet intervention [26]. However, Sert^+/− mice did not display such changes unless challenged with WD, indicating a genotype × diet interaction effect. It has been shown that unsaturated lipids, including HDL-associated –CH₃ groups and olefinic = CH–CH₂–CH= moieties, were significantly increased in Sert^−/− mice [26].

Interestingly, the expression of Arg1 and iNos exhibited significant regional and tissue specificity, with most group differences observed in the liver. These regional differences might be explained by the fact that Arg1 is highly expressed in hepatocytes, where it plays a central role in regulating systemic L-arginine availability and, consequently, NO bioavailability. iNos expression in the liver is also highly responsive to diet-induced metabolic shifts, being upregulated in Kupffer cells and hepatocytes, contributing to the key mechanisms of metabolic dysfunction. Increased hepatic iNos activity has been linked to NAFLD progression, insulin resistance, and inflammatory signaling cascades that exacerbate metabolic dysfunction. We also found that WD-exposed Sert^+/− mice showed hepatic upregulation of Arg1 and iNos, in contrast to the minimal changes observed in the hypothalamus and dorsal raphe, which is in line with previous evidence [61,100,101,102]. In addition, arginine availability can be further decreased by the depletion of lactate and BCAAs in the blood of WD-challenged mice [101,103]. Thus, the findings reported here indicate that the arginine–NO pathway is altered in both the brain and liver, consistent with previous studies [61,66,104,105,106,107,108].

Finally, WD-fed Sert^+/− animals showed numerous changes in their behavioral parameters, such as reduced exploration in the novel cage test, increased floating duration in the forced swim test, indicating enhanced depression-like behavior, and impaired recognition memory in the novel object recognition test, keeping with previous literature on rodent models of WD-like diets [83,84] and clinical findings [109,110]. The elevated anxiety of WD-fed Sert^+/− mice in object exploration situations is consistent with evidence of heightened anxiety- and depression-like behaviors in Sert-deficient rodents and S-SLC6A4 carriers [26,32,111,112,113,114,115,116].

A comparison of our findings with reports on Sert^−/− mice suggests a gene-dose effect of Sert deficit on various in vivo parameters, indirectly supported by recent in vitro results. For example, Chaji et al. (2021) demonstrated gene-dose–dependent reductions in dendritic spine density in primary cortical neurons derived from Sert-deficient rat brains, which may underlie altered synaptic plasticity and emotionality [117,118]. Earlier experiments on HEK-293 cells expressing rat SERT, which reduced transporter activity, suggested a relationship between cholesterol and compromised SERT function [119]. Martí et al. (2017) [120] showed higher mitochondrial vulnerability to toxins in serotonergic neurons with lower SERT activity.

Several limitations of our study should also be acknowledged, such as the limited translational value of the relatively short dietary intervention used in mice in comparison to clinical situations, substantial species-specific differences between humans and mice in metabolic regulation and overall metabolic rate, limited group sizes, the use of just one sex and one age in a study design, and lack of direct comparison of experimental groups with animals completely lacking Sert.

5. Conclusions

Thus, partial loss of Sert interacts with WD in aged female mice, producing metabolic, molecular, and behavioral effects that differ in important ways from those previously reported in rodents completely lacking Sert. Because the control and WD were matched for micro- and macronutrients, except for high sugar, cholesterol, and saturated fat content, the observed effects can be attributed specifically to these latter elements of the WD. Hence, our results emphasize the translational value of partial SERT deficiency models for understanding how genetic vulnerability and diet interact to promote metabolic dysfunction in at-risk populations, particularly in the context of female sex and aging. Future studies should focus on the mechanisms underlying this vulnerability and the contributing to developmental neurochemical adaptations. In this context, the use of mutants with inducible Sert deficits, both partial and complete, is likely to uncover downstream pathways in the CNS and peripheral organs, such as the gut, pancreas, and liver. This will hopefully help identify new targets for the personalized prevention and treatment of diet-induced metabolic disorders in the future.