Ethics statement and mice

Mice were treated in accordance with the National Research Council's Guide for the Care and Use of Laboratory Animals, the Guidelines for Proper Conduct of Animal Experiments published by the Science Council of Japan. The protocol was approved by the Committee on the Ethics of Animal Experiments of the Kyoto Sangyo University (Approval Nos. 2017–08, 2018–08, 2019–07, 2020–10, 2021–07, 2022–07, 2023–09, 2024–12). St3gal4-KO mice (C57BL/6-St3gal4 < tm1.1Bsi >, RBRC02286), in which the Cre-loxP system deleted exon 3 of St3gal4, were generated previously (Srimontri et al. 2014). The experiments used mice of the 17th–22nd generations backcrossed with C57Bl/6 J mice. Genotyping of St3gal4-KO and littermate wild-type (WT) mice was determined using tail genome-PCR, as previously described (Srimontri et al. 2014).

In daily care, four mice were bred in open-top plastic cages (225 mm width × 338 mm length × 140 mm height) covered with stainless steel wire grid lids after weaning. Mice were housed in a temperature-controlled room (22 °C–24 °C) with a 12-h light/dark cycle (lights on at 8:00 am) with environmental enrichment (Clea House (S), CLEA Japan, Inc., Tokyo, Japan) and were allowed free access to food (MF 12-mm diameter pellet; Oriental Yeast Co., Ltd., Tokyo, Japan) and water. All mice were fed normal chow with a caloric content of 3.6 kcal/g. The floor bedding (paper chip, 5 mm diameter; CREA Japan, Inc.) was changed twice weekly.

Mice were individually housed in metabolic cages (KN-645, Natsume Seisakusho, Tokyo, Japan) to measure daily feed intake (g), water intake (g), urine weight (g), and number of feces for 11 days. The mean of the values measured for 7 days, from days 3 to 9, following habituation to the metabolic cage, was used for each analysis.

Tissue sample collection

Non-fasting whole blood with heparin was collected from the hearts of 13–49-week-old mice of both sexes following euthanasia with isoflurane (Pfizer, New York, NY, USA) inhalation. Plasma treated with heparin (sodium salt, Mochida Pharmaceutical Co., Ltd., Tokyo, Japan) was collected following blood centrifugation for 20 min at 800 × g and promptly separated from the clot or packed cells, then transferred to an Eppendorf tube and immediately frozen at -80 °C until analysis. Plasma levels of ALP, protein, total cholesterol (T-Cho), high-density lipoprotein-cholesterol (HDL-C), low-density lipoprotein-cholesterol (LDL-C), very low-density lipoprotein-cholesterol (VLDL-C), TGs, glycerol, insulin, and free amino acids in non-fasting plasma were determined.

Analysis of plasma samples using absorbance spectrophotometer

ALP activity (units/μL) was detected at 405 nm using the LabAssay ALP kit (FUJIFILM Wako Pure Chemical Corporation, Osaka, Japan). Plasma TGs and glycerol concentrations (mg/dL) were measured at 540 nm using the Serum Triglyceride Determination Kit (TR0100, Sigma-Aldrich, St. Louis, MO, USA). Plasma T-Cho (mg/dL) was measured at 600 nm using the Cholesterol E test kit (999–02601, FUJIFILM Wako Pure Chemical Corporation). Plasma protein concentration was determined at 526 nm using the Pierce bicinchoninic acid protein assay kit (Thermo Scientific, Tokyo, Japan). The plasma assays were performed using an absorbance spectrophotometer (Synergy HTX multi-mode reader and Gen5.20 data analysis software, BioTek Instruments, Winooski, VT, USA).

The plasma lipoproteins were fractionated via commercial agarose gel electrophoresis, and the cholesterol concentrations were determined by staining with a methanol solution of Fat Red 7B (Cholesterol fraction kit in Helena Laboratories, Texas, USA) (Sato et al. 2006). The stained bands were analyzed using densitometry with a 525-nm filter (Quick Scan, Helena Laboratories, Saitama, Japan) and the relative concentration of cholesterol contained in lipoprotein fractions α (HDL-C), preβ (VLDL), and β (LDL) were determined relative to the total plasma cholesterol concentration (mg/dL) (Helena Laboratories).

Intraperitoneal glucose tolerance test (ip-GTT) and the insulin level

For glucose tolerance testing, the mice were fasted for 16 h before obtaining blood samples from the tail vein at baseline (0 min). Subsequently, 2 g/kg glucose in a 50% solution was injected intraperitoneally. Glucose concentrations in the blood were measured at 15, 30, 60, 90, and 120 min using a glucose test meter (Glutest Neo Super, GT-1820, Arkray, Inc., Kyoto, Japan). Plasma insulin levels were determined using an EIA kit (LBIS U-type, FUJIFILM Wako Pure Chemical Corporation, Osaka, Japan).

Derivatization of amino acids and gas chromatography-mass spectrometry (GC–MS) analysis

EZ:faast™ was used to extract and analyze free amino acids from plasma complex matrices (Phenomenex, Torrance, CA, USA) (Badawy et al. 2008). Based on solid-phase extraction (SPE), 50 μL of mouse plasma was mixed with 20 nmol norvaline as internal standard, pipetted into a glass vial, and passed through an SPE tip. Free amino acids were bound to the sorbent material by cation exchange and washed with 2-propanol/water. Amino compounds were eluted with an alkaline eluting medium consisting of sodium hydroxide in water, 2-propanol, and an organic base. The eluate was derivatized by propyl chloroformate for GC-MS analysis.

Free amino acids were identified using a GC–MS (QP-2010 Ultra) equipped with a PAL autosampler (AOC-6000). This analysis used a Zebron ZB-AAA column (10 m × 0.25 mm inner diameter, 0.25 μm film thickness, Phenomenex) for detection. Helium was used as the carrier gas at a pressure of 15 kPa. The temperature of the injector was set at 250 °C. In the GC–MS conditions, starting at 110 °C, the oven temperature was increased to 320 °C (30 °C/min) and finally held at 320 °C for 1.5 min. The interface and ion source temperatures were set at 280 °C and 200 °C, respectively. The ion trap mass spectrometer was operated at an ionization energy of 70 eV. The sample (2 µL) was injected in split mode (1:10, v/v). Full scans were recorded over a mass range of 45–450 m/z and a scan velocity of 0.15 scans/sec. Arginine among proteinogenic amino acids was not detected here.

ALP enzyme histology

Brains were collected from mice following euthanasia with isoflurane and frozen on dry ice. Sectioning of the frozen brain block was performed in the coronal plane (-0.2 mm to 3.5 mm from bregma) at 14 μm on a sliding microtome (CM1950, Leica Biosystems Nussloch GmbH, Nußloch, Germany) at -17 °C.

The brain sections were dried, fixed with 4% formaldehyde (Fisher Scientific, Pittsburgh, PA, USA) in 0.1 M phosphate buffer, pH 7.4 for 20 min on ice, and washed twice for 5 min at room temperature in 0.1 M pH 7.4 phosphate buffer, once for 5 min in 0.1 M pH 9.5 Tris-HCl, and once for 5 min in AP buffer (0.1 M pH 9.5 Tris-HCl, 0.1 M NaCl, and 50 mM MgCl2). The brain specimens were treated with BCIP/NBT substrate (11383221001/11383213001, Sigma-Aldrich, Tokyo, Japan) in AP buffer for 10 min at room temperature. The specimens were incubated for 5 min in 0.01 M pH 7.5 Tris–HCl containing 1 mM EDTA to terminate the ALP enzymatic activity. After washing for 5 min in 0.1 M phosphate buffer, the specimens were sealed with a cover glass using Permount (Fisher Scientific, Pittsburgh, PA).

Subsequently, slides were observed using an upright fluorescence microscope (ECLIPSE 80i, Nikon Corporation, Tokyo, Japan), and images were obtained using a camera (DS-Qi1Mc, Nikon) and imaging software (NIS Elements ver 4.10, Nikon). To allow analysis of the BCIP/NBT signal intensity, images, including the septum in a section cut at 1.0 mm anterior from bregma and the thalamus in sections cut from 0.2–2.7 mm posterior from bregma, were processed by binary treatment using image analysis software (WinROOF version 3.3, Mitani Corporation, Osaka, Japan). The volumes (1.5 mm2 * intensity) within the fields were calculated as the ALP activity.

Fear conditioning test

The auditory fear conditioning test was performed as described previously (Srimontri et al. 2014; Kato 2015). An auditory fear conditioning system was used: scrambled foot shocks were delivered as unconditioned stimuli (US; 0.5 mA, 1 s; O’hara and Co., Ltd., Tokyo, Japan), and the auditory conditioning stimuli (CS; white noise, 70 dB, 10 s) were delivered from a speaker on the right side of the chamber in an acoustic isolation box (50 dB background noise level, 220 lx from a light-emitting diode). On day 1, each mouse was placed in the clear square-shaped arena with a metal grid floor and allowed to explore for 60 s, then received the CS for 10 s and the US during the final 1 s of the presence of CS. The CS and US were delivered automatically with a 20-s interval twice. Freezing times of more than 2 s were measured over 180 s. Twenty-four hours later, the contextual fear conditioning test was performed in the same box used on day 1 without the CS or US. Forty-eight hours later on day 3, each mouse was placed in the cued test box, a novel solid gray square-shaped arena (50 lx), allowed to explore for 60 s, received the CS for 30 s, and then allowed to explore for 90 s. Freezing times of more than 2 s were measured over 180 s. All data were collected and analyzed with Time FZ1. Ten days after the fear conditioning test, mouse plasma was collected to measure plasma ALP levels.

Statistical analysis

Raw data were first analyzed using Excel (Microsoft) and GraphPad Prism version 8 (Prism 8, GraphPad Software Inc., CA, USA). Values are indicated as mean ± standard deviation. In group comparisons, values were compared between two groups using the Mann–Whitney U-test and four groups using the Kruskal–Wallis test. P-values of ≤ 0.05 were considered statistically significant. Statistically significant effects identified using the Kruskal–Wallis test were further evaluated using multiple comparisons by controlling the false discovery rate (Benjamini et al. 2006). The relations of all pairs of variables with body weight and concentrations of plasma compounds were analyzed using Pearson’s correlation coefficients (r) with the associated two-tailed p-value using Prism 8 software. Power analyses were performed using G*Power 3.1 software for Mann–Whitney U-test comparisons with p < 0.05.