Animals and ethics statement

Eight-week-old male C57BL/6 mice were obtained from the Oswaldo Cruz Foundation’s Animal Breeding Center. The animals were kept at a controlled temperature (23 ± 1 °C) in a room under a 12-h light/dark cycle with ad libitum access to food and water. The Animal Welfare Committee of the Oswaldo Cruz Foundation (CEUA-FIOCRUZ) approved all animal procedures under license number L-012/2021.

Induction of metabolic syndrome and semaglutide treatment

We used a well-characterized experimental murine model of MetS. MetS was induced in a group of 10 animals by feeding them a high-fat diet (HFD) for 24 weeks, while a control group of 10 animals received a normalipid/regular fat diet (ND) (AIN93M) during the same period. Both chow diets were manipulated by Pragsoluções (Jau, São Paulo—Brazil).

The ND used was composed of corn starch, casein (14%), dextrinized starch, sucrose, soybean oil (4%), microcrystalline cellulose, mineral mix AIN 93M (sodium, iron, manganese, zinc, iodine, copper, selenium, cobalt and fluoride), mix Vit AIN 93 (vitamin A, vitamin B1, vitamin B2, vitamin B6, vitamin B12, choline bitartrate, vitamin D3, vitamin E, vitamin KE, niacin, biotin and folic acid), lysine, methionine, l-cystine, choline bitartrate and BHT (tertibutylhydroquinone). For the HFD, the same components as the control diet were used, but with changes in the amounts of casein (19%), increased porcine fat (32%), and sodium chloride (0.5%). In summary, the percentages of calories in each diet are as follows: carbohydrates: ND 75.81% vs. HFD 25.69%; lipids: ND 9.47% vs. HFD 60.19%; proteins: ND 14.73% vs. 14.12%.

For 24 weeks, C57BL/6 mice were first fed either a standard ND or a HFD. After this period, the mice on the HFD were randomly divided into two groups for an additional 4 weeks: one group received semaglutide treatment (HFD SEMA) and the other group received a saline solution (HFD SAL). The animals were administered a fixed daily dose of 0.2 mg/kg of semaglutide in 0.5 mL (Ozempic®, Novo Nordisk, Bagsværd, Denmark) subcutaneously on the back. The ND group also received the same semaglutide treatment (ND SEMA; n = 5) or saline solution (ND SAL; n = 5). The effects of semaglutide doses on mice with diabetes have been extensively studied to compare them to human doses. The dosage of semaglutide used in this study (0.2 mg/kg/day) is based on extrapolations from human dosages. In clinical settings, semaglutide is prescribed at doses of 0.5 mg to 1 mg per week for type 2 diabetes and obesity management, translating to approximately 0.01 to 0.02 mg/kg/day in humans. The higher dose used in mice accounts for differences in metabolism and body surface area scaling [15].

Biochemical analyses

Blood glucose was measured using an automatic glucometer (One Touch Ultra 2, Johnson & Johnson Medical SA, Argentina), and triglyceride levels were measured by Accutrend Plus (Roche). At the end of the experiments, blood samples were collected from all animals via intracardiac puncture, and serum was collected by centrifugation at 1500×g for 10 min at 4 °C, aliquoted, and stored at −80 °C for analysis. A biometric immunoassay, which uses fluorescent-dyed microspheres conjugated with a monoclonal antibody specific for a target protein such as insulin, leptin, or resistin, was conducted using a multiplex array reader from the Luminex™ Instrumentation System (Bio-Plex Workstation from Bio-Rad Laboratories). The concentration of the analyte was calculated using the software provided by the manufacturer (Bio-Plex Manager Software).

Histological analyses of brain and adipose tissues

To assess changes in adipose tissue, the epididymal fat contents of the animals were harvested immediately following the termination of the experiments. The tissues were fixed in 4% paraformaldehyde for 48 h at 4 °C before sectioning and H&E staining. Twelve-micron-thick adipose tissue sections were stained with hematoxylin and eosin (H&E) and imaged at 40X magnification using a Zeiss Primo Star light microscope (Oberkochen, Germany). The area of white adipocytes was determined using ImageJ software (NIH, Bethesda, MD, USA) by outlining white adipocytes with ellipses. For each animal, two to three images, each representing a distinct sample area, were obtained. The adipocyte size was quantified in 10–20 adipocytes per image and reported as the area of adipocyte size.

Twelve-micron-thick hippocampal sections stained with hematoxylin and eosin (H&E) were imaged using a Zeiss Primo Star light microscope at 40X magnification (Oberkochen, Germany). Analysis was conducted using ZenBlue software (Zeiss), through which the thickness of the hippocampus was measured. By counting the number of cells across a transverse line from the inner to the outer layers of the hippocampus, the cellular density in the hippocampus was calculated and is expressed as the ratio of cells per hippocampal thickness.

Brain intravital microscopy

The animals were randomly assigned to microcirculatory analysis at the end of the experiment. Mice were anesthetized with a combination of ketamine (100 mg/kg) and xylazine (10 mg/kg, i.p.) (Cristália, SP, Brazil). The tail vein was catheterized to allow injection of the fluorescent dye. To assess cerebral microcirculation, the mice were fixed in a stereotaxic apparatus, and a cranial window over the left parietal bone (1–5 mm lateral, between the coronal and lambdoid sutures) was created using a high-speed drill to expose the brain’s microvascular surface, as previously described [16]. The animals were placed under the light beam of a fluorescence microscope (Zeiss model AXIO SCOPE A1, Obercochen, Germany), and microcirculation images were acquired using Zen Blue software (Zeiss, Obercochen, Germany).

Leukocyte–endothelium interactions were assessed as previously described [17]. Rhodamine 6G (0.3 mg/kg) was injected into the tail vein to label circulating leukocytes. The interaction of leukocytes with the endothelial wall was evaluated by determining the number of adhered leukocytes that remained static along 100 µm of the venular wall for 30 s, and the number of rolling leukocytes that moved within the vessel at a slower speed than did the number of circulating erythrocytes showing slow contact with the inner venular walls. The values are expressed as the number of cells/min/100 μm. We determined these parameters in brain surface venules with diameters ranging from 50 to 100 µm, and images were acquired with a 10× ocular and 10× objective microscope (Zeiss—AXIO SCOPE A1, Obercochen, Germany), producing a final magnification of 100× on the monitor.

Immunohistochemical analysis and confocal microscopy

At the end of the intravital microscopy analyses, the animals were euthanized by an overdose of anesthetic drugs. Specifically, we used an overdose of pentobarbital sodium at a dose of 150 mg/kg, administered intraperitoneally. The mice were perfused with a saline solution through the heart, followed by 4% paraformaldehyde. Subsequently, the brains were collected and postfixed in paraformaldehyde for 48 h at 4 °C and sectioned to obtain 12-μm slices, which were permeabilized with 0.05% Triton X-100 (Vetec, Speyer am Rhein, Germany) solution in PBS for 30 min and subsequently incubated with blocking solution containing 5% bovine serum albumin (Sigma‒Aldrich), 2.5% normal goat serum (Thermo Fisher Scientific, Waltham, MA), and 0.02% Triton X-100 diluted in PBS for 1 h. Then, the tissues were incubated with primary antibodies and diluted in a blocking solution overnight at 4 °C. The sections were incubated with Biotinylated isolectin B4 (IB4) (Vector; 1:100) and streptavidin-Cy3 (1:400) to label cerebral microcirculation vessels. We used primary antibodies against astrocytes (anti-GFAP, Dako, 1:400 dilution), microglia (anti-Iba-1, Wako, 1:200 dilution), and ICAM-1 (Monoclonal Antibody for CD54, ICAM1, eBioKAT-1, eBioscience). Then, we used secondary antibodies and photographed the slides using a LSM 510 META confocal microscope (Zeiss). The BBB integrity was considered by measuring the parenchymal abundance of IgG using immunomicroscopy. Briefly, after blocking with 10% goat serum, 20 μm cryosections were incubated with goat anti-mouse IgG conjugated with Alexa488 (1:50, Life Technologies) for 20 h at 4 °C. The sections were counterstained with DAPI.

Structural capillary density analysis

To analyze the cerebral microvascular network, we used the AngioTool program (available in the public domain at https://ccrod.cancer.gov/confluence/display/ROB2/Downloads), a validated source for measuring vascular networks. We calculated the total length of IB4-labeled brain capillaries, which represented the brain capillary density. Lacunarity describes the distribution of spaces between vessels.

Analysis of microglial activation

To quantify morphological changes in Ionized calcium‑binding adaptor molecule 1 (Iba-1+) cells, consecutive Z-stack images were converted to a maximum intensity projection image by ImageJ software (NIH Bethesda, MD, USA). Using the Sholl analysis plugin, concentric circles were created centered on the soma, beginning at 4 μm radii and increasing by 2 μm with every circle. We determined the number of intersections made by microglial branching processes with each successive increasing circle, the maximum number of intersections for the cell (Nm), the critical value at which Nm occurred, and the maximum length at which a branch intersection was observed. We used the image tool “Sholl analysis” from the ImageJ Program to analyze microglial morphology. It was possible to analyze the number of processes of each microglia separately through concentric and consecutive circles every 2 µm around the cell and to compute the number of intersections of microglial processes with each circle, representing the degree of branching of the cell. The less branched the cell was, the more activated the cell was. At least 30 microglia per field were analyzed in 3 sections/4 slides per animal.

Analysis of vessel coverage by astrocytes

The degree of colocalization between the astrocyte marker GFAP and the blood vessel marker IB4 was analyzed using the ImageJ program. The tool "colocalization", which uses Mander's M1 coefficients, was used. These coefficients indicate the percentage of pixels in the green channel (GFAP+) intersecting with a red channel signal (IB4). In other words, the M1 coefficients show the fraction of intensity in each channel that coincides with some intensity in the other channel.

Cognition testsFear conditioning memory test

Mice were placed into a specific chamber individually. After 3 min of habituation, mice were subjected to an electric foot shock (0.75 mA, 3 s) simultaneously to a tone (2.5 kHz, 85 dB, 3 s) and returned to their home cages. The next day, mice were placed in the same chamber for 3 min, and the tone was applied without a foot-shock. A digital video camera recorded the behavior of the mice and was manually analyzed. Continuous immobility for 1 s was defined as freezing behavior, and the total freezing time (i.e., length of immobility) was measured. Throughout the experiment and analysis of results, the group of animals was not revealed to the researcher. The procedures followed the described by Zhang et al. [18] and Granja et al. [19].

Light/dark box test

The light/dark box test reliably predicts anxiolytic and anxiogenic-like effects in rodents and offers quick, easy performance without prior animal training or food/water deprivation. Transitions in this test indicate activity and exploration, while the time spent in each compartment reflects aversion (light) and attraction (dark). The apparatus consists of a wooden box (48 cm × 24 cm × 27 cm) divided into two compartments by a barrier with a doorway (10 cm × 10 cm). One compartment is black and covered, while the other is white and illuminated by a 60-W light bulb positioned 40 cm above the box. The test procedure involves transporting the mice to a darkened room, allowing them to acclimate in their home cages for 2 h, then placing them in the lit compartment to explore freely for 5 min. We recorded the time spent in the lit compartment, the number of entries into the lit compartment, and the latency to enter the dark compartment [20]. Throughout the experiment and analysis, the group of animals remained blinded to the researcher.