Beta-Sitosterol Beta-D-Glucoside (BSSG) causes gut inflammation in zebrafish and mice before brain cell damage begins

β-Sitosterol β-d-glucoside (BSSG) synthesis process and chemical characterization are extensively described in Additional File 1.

Zebrafish wild type (WT) lines used in this work, including those used for the generation of the stable mutant lines, derived from Tuebingen and Giotto strains matings. Embryos, larvae and adults, housed in the Zebrafish Facility of the University of Padova, were maintained according to standard procedures [12]. Embryos were obtained from natural mating of WT, mutant, or transgenic adult fish and raised at 28.5 °C in Petri dishes with Fish Water (50×: 25 g Instant Ocean, 39.25 g CaSO₄ and 5 g NaHCO₃ in 1L) maintained in a 12 h light:12 h dark (LD) cycle until 72 h post fertilization (hpf). At this life stage, larvae were screened for fluorescence, when necessary, and exposed to treatment or control vehicle.

WT C57BL/6 male mice, housed in the Animal Facility of the University of Padova, were fed for 15 weeks, following weaning (1 month after birth) and up to 6 months of age, with BSSG-enriched chow pellet or commercial food as a control. Mice food pellet was prepared by Mucedola Srl. The treatment paradigm, previously established in [11], consisted in feeding mice daily with 1 mg/day of BSSG for 5 days/week. At the end of the experimentation period, mice were sacrificed by cervical dislocation, weighed and dissected for organs collection. All husbandry and experimental protocols were in accordance with national and EU guidelines for the use of experimental animals and were approved by the Animal Care and Use Ethics Committee of the University of Padua and by the Italian Ministry of Health (Authorization n. 112/2015PR; Authorization n. 12/2023-PR; Authorization n. 690/2020-PR).

BSSG powder (MW = 576.86 g/mol) was dissolved in DMSO (Sigma-Aldrich), while β-sitosterol powder (β-Sito, MW = 414.71 g/mol; 85451, Sigma-Aldrich) was dissolved in ethanol (EtOH), both at a final concentration of 10 mM, and sonicated until complete solubilization. Ultrasonic bath temperature was kept at around 26–27 °C. Stock aliquots were stored at − 20 °C. All chemicals were administered to zebrafish larvae diluted to a final working concentration of 10 µM, while control larvae were exposed to the same volume of the vehicle, reaching a maximum concentration of 0.1% v/v.

For acute treatment, zebrafish larvae were incubated in Fish Water containing either BSSG, β-sitosterol or vehicle from 3 to 5 days post fertilization (dpf). Treatment was renewed every 24 h to avoid molecule deposits and changes in concentration.

Chronic 15-days-long BSSG treatment was performed on zebrafish larvae from 15 to 30 dpf. Briefly, larvae obtained from three independent spawns were raised in the Zebrafish Facility of the University of Padova according to standard procedures until 15 dpf. They were then divided into 3 treatment and 3 control groups and maintained during the experimentation period in glass beakers containing 300 mL of Fish Water at 28.5 °C under a 12:12 LD period and fed three times/day. Treatment was administered directly in the water to a final concentration of 10 µM and refreshed every 2 days.

The experimental BSSG-enriched feed was produced at the Department of Veterinary Sciences of the University of Pisa by incorporating the compound in a control feed, as described in [13]. Briefly, the control feed (a mixture of 50% TetraMin Flakes^® and 50% GEMMA Micro 300®) was first finely ground (< 100 µm) and then mixed with the BSSG powder to achieve a uniform dispersion. Distilled water was added to obtain a homogeneous dough, which was pelleted and dried at 40 °C in a drying chamber for 24 h. Finally, dry pellets were re-ground to restore feed particle size and kept at − 20 °C until use. Commercial feed without BSSG was processed using an identical procedure as a control. Male adult zebrafish at 4 months post fertilization (mpf), obtained from three independent spawns, were divided (N ≥ 4) into treatment and control groups, homogeneous in body weight. Fish were fed 5% of their body weight once per day, according to [14], to achieve a BSSG intake of approximately 15 µg/day/fish. The experiment lasted for 30 days.

Genotyping of mutant zebrafish was performed at early stages from genomic DNA extracted with the HotSHOT protocol [15] and amplified with locus-specific primers through PCR with 5× HOT FIREPol^® Blend Master Mix (04–25-00125, Solis BioDyne). Primers used for genotyping are listed in Additional file 2: Table 1.

Treated and control larvae (N = 10 per condition in each replicate) were anesthetized with Tricaine, mounted in 2% methylcellulose in H₂O on depression slides, and imaged under a Leica M165 FC microscope equipped with a Nikon DS-Fi2 digital camera. Quantification of the main morphological traits (standard length, eye area and area of the swimming bladder) was performed using Fiji-ImageJ software (Version 2.14.0/1.54p). The experiment was repeated for 3 times.

Heads and trunks of treated and control larvae (N = 30 in each condition) were separated with a sharp blade, pooled, snap frozen in liquid nitrogen, and stored at − 80 °C until use. Folch’s method [16] was used to extract lipids. Briefly, frozen pellets were resuspended with Milli-Q water and homogenized using a Pellet Pestle^® Motor (Kimble) for 30 s, alternating with 10 min incubation on ice (repeated for three times). After homogenization, a solvent solution composed of chloroform:methanol:water in a 8:4:3 volumetric ratio was added. The organic phase was collected, resuspended in the same solution and processed a second time to remove contaminants. Finally, samples were dried with a lyophilizer, flushed with nitrogen, sealed, and stored at − 80 °C. Subsequent LC–MS measurements were carried out to determine the internal concentration of BSSG in heads and trunks of treated larvae compared to controls. Lipid extracts were dissolved in 90:10 methanol/chloroform and 5 μL were injected into a Hewlett-Packard Model 1100 Series liquid chromatograph (Hewlett-Packard Development Company, CA) coupled to a photodiode array (PDA) detector (Agilent Technologies, Agilent 1100 Series) and to a Bruker Esquire-LC quadrupole ion-trap mass spectrometer (Bruker OptikGmbH) equipped with atmospheric pressure electrospray ion source. Quantitative LC–MS measurements were also performed in positive ion mode using a Triple Quadrupole (QQQ) mass spectrometer (Applied Biosystems, API 3000 QQQ) equipped with an electrospray ion source (ESI), and combined with a Shimadzu High Performance LC system (CBM-20 A, binary pump LC-20AB). Chromatographic analysis was carried out in both MS conditions at room temperature on a Kinetex-C8 100 × 4.6 mm, 2.6 μm column (Phenomenex, Italy). The eluent (1.0 mL/min) consisted of (A) methanol:water/10 mM ammonium acetate (70:30) and (B) methanol:isopropanol/10 mM ammonium acetate (90:10) using a linear gradient: 65%–100% B in 40 min, followed by isocratic B held for 10 min.

Zebrafish larvae treated with either BSSG, β-sito or the control vehicle were incubated with Neutral red solution (553-24-2, Sigma-Aldrich) at a final concentration of 4 µg/mL for 3 h in the dark, and then briefly rinsed in Fish Water. Larvae were anesthetized with Tricaine and imaged under a Leica M165 FC microscope with a Nikon DS-Fi2 digital camera. The length of the stained portion of the mid-intestine was measured with Fiji-ImageJ. Each treatment was performed four times with at least 10 larvae per condition.

Treated and control WT and gr^−/− zebrafish larvae were fixed overnight at 4 °C in 4% PFA (Sigma-Aldrich) in PBS. Staining with alcian blue (Alcian blue 8GX–A5268, Sigma-Aldrich Sigma) was performed as described in [17]. Number of goblet cells was manually scored starting from the junction between the intestinal bulb and the mid-intestine. The experiment was performed three times, with at least 10 larvae per condition.

Transgenic larvae were obtained crossing Tg(mpx:GFP)ⁱ¹¹⁴ [18], Tg(8xHs.NFκB:GFP, Luciferase)^hdb5 (hereafter Tg(NFkB:GFP)) [19] or Tg(7xSRE-HSV.Ul23:EGFP)^ia28 (hereafter Tg(Stat3:EGFP)) [20] x wild type zebrafish and cyp11c1^+/- x cyp11c1^+/- Tg(9xGCRE-HSV.Ul23:EGFP^ia20) (hereafter Tg(GRE:EGFP)) [21]. At 3 dpf, larvae were screened for specific fluorescence and only GFP-positive individuals were randomly assigned to treated and control groups. After treatment, larvae were anesthetized and mounted in 1% low melting agarose. The same intestinal region, starting from the junction between the intestinal bulb and the mid-intestine and spanning at least 4 somites, was imaged by z-stacks (5 µm step size) under a 20× (for Tg(mpx:GFP), Tg(NFkB:GFP) and cyp11c1;Tg(GRE:EGFP)) or 40× objective (for Tg(7xStat3:EGFP)) on a Nikon C2 confocal microscope, using NIS ELEMENTS software. Fluorescent neutrophils in Tg(mpx:GFP) larvae were manually counted by scrolling through the z-stacks, assisted by the Cell counter tool in Fiji-ImageJ to avoid repeated counting, as described in [22]. Only neutrophils located in the intestinal walls along 5 somites were considered.

To quantify fluorescence signal in Tg(NFkB:GFP), Tg(Stat3:EGFP) and cyp11c1;Tg(GRE:EGFP) larvae, z-stack images were converted to maximum intensity z-projections, and mean fluorescence was measured in the same region of interest (ROI) using Fiji-ImageJ software. After the image acquisition, cyp11c1;Tg(GRE:EGFP) larvae were genotyped to distinguish cyp11c1^+/+ and cyp11c1^−/−;Tg(GRE:EGFP) siblings.

Each experiment was performed three times, with at least 10 larvae per condition.

cyp11c1^+/+;Tg(GRE:EGFP) and cyp11c1^−/− Tg(GRE:EGFP) males of 4 mpf of age were randomly subdivided into treated (fed for 15 days with BSSG-enriched food) and control groups. At the end of the experimental period, fish were fasted 24 h prior to sacrifice to reduce the amount of food residues inside the intestine. After sacrifice, intestine was extracted and mounted in 1% low-melting agarose for imaging. Samples were imaged by z-stacks (7 µm step size) under a 20× objective on a Nikon C2 confocal microscope. To quantify intestine fluorescence, z-stack images were converted to maximum intensity z-projections, and the mean fluorescence was measured after setting a threshold to isolate the fluorescent region. Values were then normalized to the untreated cyp11c1^−/−;Tg(GRE:EGFP), considered as the background signal.

Control and BSSG-treated 5-dpf zebrafish larvae were incubated with Acridine Orange solution (A6014, Sigma-Aldrich) at a final concentration of 15 µg/mL for 10 min in the dark, and then extensively rinsed in Fish Water. Larvae were then anesthetized and mounted in 1% low melting agarose. The same intestinal region was imaged as z-stacks (5 µm step size) under a 20× objective on a Nikon C2 confocal microscope, using the NIS ELEMENTS software. To quantify fluorescence, z-stack images were converted to maximum intensity z-projections, and the mean fluorescence intensity was measured in the same ROI spanning 4 somites with Fiji-ImageJ software.

According to the protocol described in [8], control and BSSG-treated 5-dpf larvae were incubated with DCFH-DA (2’, 7’-Dichlorofluorescein diacetate, Sigma-Aldrich) at a final concentration of 1 mg/L overnight in the dark before the recording of peristaltic movements. Larvae were rinsed three times in Fish water, anesthetized for 1 min in Tricaine to limit its influence on muscular contraction, and mounted in 2% methylcellulose. Videos of 6–8 min were recorded under a Leica M165 FC fluorescence microscope. Each larva was analysed twice by scoring the number of peristaltic contractions, defined as partial or total invaginations of the intestinal bulb [23]. The average value was used to calculate the frequency of peristaltic movements over a 2 min period. The experiment was performed three times, with at least 10 larvae per condition.

According to the protocol described in [24] with slight modifications, gastrointestinal transit was evaluated as the progression of the food bolus along the digestive tract over 24 h after feeding. Treated and control 5-dpf larvae were fed commercial food in the morning (8:00 a.m.) and allowed to feed for 2 h. This was considered as “Time 0”. Larvae showing the intestinal bulb filled with a visible food bolus under brightfield were considered for the analysis and imaged under a Leica M165 FC microscope 4, 8 and 24 h post-feeding. Food progression was then evaluated according to the scoring system described in [24]. A total of N = 32 CTR and N = 24 treated larvae were examined.

The intestinal contractility of adult zebrafish was analysed ex vivo by measuring tension changes with the isolated organ bath technique, as previously described in mouse [25, 26] and fish preparations [27]. Experiments were performed on the full-length intestine of zebrafish fasted for 24 h prior to sacrifice to reduce the amount of food residues in the lumen. Once extracted, the intestine was maintained in Krebs solution (NaCl 118 mM, KCl 4.7 mM, CaCl₂∙2H₂O 2.5 mM, MgSO₄∙7H₂O 1.2 mM, K₂HPO₄ 1.2 mM, NaHCO₃ 25 mM, C₆H₁₂O₆ 11 mM). By creating two loops with silk thread, each intestine was mounted along its longitudinal axis in an organ bath containing 10 mL of oxygenated (95% O₂ + 5% CO₂) Krebs solution maintained at 28.5 °C. Changes in muscle tension were recorded by isometric force transducers (World Precision Instruments, Berlin, Germany) connected to a PowerLab 4/30 data acquisition system using LabChart 8 software (ADInstruments, Besozzo, VA, Italy). Intestinal preparations were stretched to an initial tension of 0.1 g and left to equilibrate for 45 min to allow the development of rhythmic spontaneous contractions [25, 26]. At the end of the equilibration period, samples were activated using 1 μM carbachol (CCh), a non-selective cholinergic receptor agonist, directly added to the Krebs solution. To evaluate smooth muscle contraction, each intestine was exposed to 40 mM KCl, a depolarizing agent that induces Ca²⁺ release and subsequent smooth muscle contraction. To assess the cholinergic receptor-mediated excitatory responses, samples were then exposed to increasing concentrations of CCh (0.001–100 μM) added cumulatively to the organ baths to generate concentration–effect curves. Neuronal-mediated contractions were evaluated using electrical field stimulation (EFS) at increasing frequencies with constant voltage (0–40 Hz; 1ms pulse duration; 10s pulse-trains, 80 V) via platinum electrodes connected to an S88 stimulator (Grass Instrument), inducing membrane depolarization and neurotransmitter release. Finally, to characterize the inhibitory neuromuscular responses, zebrafish intestines were exposed to 0.1 μM isoprenaline, a non-selective β-adrenergic receptor agonist. Concentration–response curves were subjected to a nonlinear regression analysis (fitted to a sigmoidal equation) to calculate maximal tension (Emax) values. Contractile responses were expressed as gram tension/gram dry tissue weight [25, 26]. N ≥ 4 animals were analysed.

Total RNA was extracted with TRIzol reagent (15596026, Thermo Fisher Scientific) from different tissue samples: pooled whole larvae at 5 dpf (at least N = 15 in each sample); pooled whole larvae at 30 dpf (N = 14 in each sample); single adult zebrafish intestine or brain; and single mouse small intestine samples approximately 1.5 cm in length. The quantity and quality of the isolated RNA were assessed using a NanoDrop 2000c (Thermo Fisher Scientific) and TapeStation System 4150 (Agilent). 1 µg (for zebrafish larvae and mouse tissue samples) or 500 ng (for adult zebrafish tissue samples) of total RNA were treated with RQ1 RNase-Free DNase (M6101, Promega) and used for cDNA synthesis, with the High-Capacity cDNA Reverse Transcription Kit (4368814, ThermoFisher). cDNA was diluted 1:10 (or 1:5 for mouse intestine samples) and amplified using 5× HOT FIREPol EvaGreen qPCR Mix Plus (08–36-00001, Solis BioDyne) with locus-specific primers. Genes expression was measured with SybrGreen method on a CFX384 Touch Real-Time PCR Detection System (BioRad). Primers’ melting curves and threshold cycles (Ct) were automatically generated. Relative gene expression levels were calculated using the comparative Ct method (2^–ΔΔCt) normalized to housekeeping genes (β-actin for adult zebrafish tissues, ef1α for larval zebrafish samples; Tbp for mouse intestine samples). Primers used for RT-qPCR analysis are listed in Additional file 2: Table 3.

Library construction from RNA extracted from 30 dpf zebrafish larvae was performed using the QuantSeq 3ʹ mRNASeq Library Prep Kit for Illumina (015UG009V0260, Lexogen) according to manufacturer’s protocol. Briefly, 500 ng of total RNA were retrotranscribed with oligod(T) primers and, after RNA degradation, the second cDNA strand was synthesized using random primers. Double-stranded DNA was amplified to incorporate sequencing adapter and library barcodes. For mouse samples, libraries were constructed starting from at least 500 ng of total RNA following ribosomal RNA depletion, using Illumina Ribo-Zero Plus rRNA Depletion Kit (Illumina). rRNA-depleted RNA was subsequently purified by ethanol precipitation and employed for library preparation. Briefly, after fragmentation, the first-strand cDNA was synthesized using random hexamer primers (NEBNext Ultra II RNA First Strand RNA Synthesis Module; NEB). Then, the second strand cDNA was synthesized, and dUTPs were replaced with dTTPs in the reaction buffer (NEBNext Ultra II Directional RNA Second Strand Synthesis Module; NEB). Double-strand DNA was purified using magnetic beads, end-repaired, and A-tailed to facilitate adapter ligation. After gel size selection (250–800 bp), USER enzyme digestion (NEB) was employed to eliminate UTP-containing second strand cDNA filaments. Prior to sequencing, libraries were quantified using the Qbit (dsDNA Quantitation, High Sensitivity; Thermofisher) and size distribution was assessed using the High Sensitivity DNA kit (Agilent Technologies). Zebrafish libraries were sequenced by the Sequencing facility of the Department of Biology, University of Padova, using a NextSeq 500 (Illumina) platform, employing a single read approach. Conversely, mouse libraries were sequenced by Novogene facility using Novoseq X plus (Illumina) sequencer employing a paired-end approach.

After adapter trimming with Cutadapt (version 4.7) [30], transcript expression was quantified using Salmon method (version 1.10.3) [31]. The resulting count matrix was imported into the R statistical environment, and the edgeR package (version 4.0.2) [32], was used for gene expression normalization and to identify differentially expressed genes (DEGs) between treated and control samples (FDR ≤ 0.05), reported in Additional file 4 and Additional file 5. Zebrafish and murine DEGs were used for Gene Ontology (GO) enrichment analysis via enrichplot package in R software, and bar charts were generated on ShinyGO [33]. Raw RNAseq data have been deposited in the SRA database.

After sacrifice, mouse small intestine was extracted and gently flushed with cold PBS. Small sections (0.5 cm long) of mouse small intestine were fixed overnight at 4 °C with Karnovsky fixative (2.5% glutaraldehyde, 2% paraformaldehyde in 0.1 M cacodylate buffer), washed with 0.1 M cacodylate buffer, post-fixed with osmium tetroxide for 2 h, and embedded in EMbed 812 (Electron Microscopy Sciences). Ultrathin sections were stained with uranyl acetate and lead citrate and examined using a Philips M400 microscope operating at 100 kV. The microvilli length was measured from the tip to the base with Fiji-ImageJ software. Analyses were repeated on multiple images derived from 3 individuals per condition.

After sacrifice, mice small intestine was extracted and gently flushed with cold PBS. Segments approximately 1.5 cm in length were fixed in Bouin fixative solution (30 mL saturated picric acid in ddH₂O, 10 mL formaldehyde, 2 mL glacial acetic acid for 42 mL) for 24 h and then rinsed in 70% ethanol. Samples were dehydrated in graded ethanol (80%−90%−100% in ddH2O) for 1 h at RT, cleared in xylene (Sigma-Aldrich), infiltrated with xylene: paraffin (1:1) and embedded in paraffin. Sections (7 µm) were cut using a Leica microtome. Tissues sections were deparaffinized, rehydrated to ddH₂O, incubated in 3% glacial acetic acid solution for 3 min, and stained with Alcian blue solution (1% Alcian blue 8GX – A5268, Sigma-Aldrich in 3% glacial acetic acid; pH 2.5) for 30 min. Sections were rapidly rinsed in 3% glacial acetic acid solution, washed in running tap water for 5 min, and counterstained with Eosin Y (Sigma-Aldrich, USA). Then were dehydrated in 100% ethanol and mounted for visualization. Alcian blue-stained goblet cells were counted along intestinal villi in 3 sections spaced 100 µm in 6 mice for each condition.

Mouse small intestines were paraffin-embedded and sectioned as previously described and mounted on Superfrost® Plus microscope slides (J1800AMNZ, Thermo Scientific). Sections were deparaffinized, rehydrated to ddH₂O with graded ethanol series (100%−90%−70% in ddH2O) for 5 min each and incubated for 15 min in a quenching solution (NH₄Cl 50 mM) to reduce autofluorescence. Antigen retrieval was performed by incubation in sub-boiling citrate buffer (citric acid 0.01 M, pH 6) for 10 min, followed by 10 min in TBS-1% Tween at RT. Slides were then blocked for 1 h in saturating solution (15% goat serum, 2% BSA, 0.25% gelatine, 0.2% glycine in PBS supplemented with 0.5% Triton X-100). Sections were incubated overnight at 4 °C in F4/80 monoclonal primary antibody (14–4801-82, Thermofisher Scientific) diluted 1:50. After 3 washes for 5 min in TBS-1% Tween, sections were incubated for 1 h at RT with goat anti-rat-Alexa Fluor 568 secondary antibody (A-11077, Thermofisher Scientific) at a final dilution of 1:200. To reduce autofluorescence, slides were treated with Sudan Black solution (0.1% in EtOH 70%) for 10 min, followed by extensive washing in TBS-1% Tween. Nuclei were stained with DAPI for 5 min and sections were mounted with SlowFade™ Diamond Antifade Mountant (S36967↗, Thermofisher Scientific). Images were acquired using a Zeiss AXIO Zoom.V16 fluorescence microscope equipped with an Axiocam 305 mono camera. Macrophages were manually counted in sections from 3 mice per condition.

To verify the ability of BSSG to bind steroid hormone receptors at a concentration of 10 µM, a radio-ligand binding assay was performed by Eurofins Panlabs Discovery Services using the NHR Binding Agonist Radioligand Assay. BSSG selectivity for human androgen receptor, estrogen receptor, glucocorticoid receptor, mineralocorticoid receptor and progesterone receptor were calculated as the percentage of inhibition for the binding of a radio-labeled ligand, specific for each receptor ([³H]-methyltrienolone, [³H]-estradiol, [³H]-dexamethasone, [³H]-aldosterone and [³H]-progesterone, respectively).

After BSSG treatment, 5-dpf larvae were fixed overnight in 4% PFA (Sigma-Aldrich) in PBS at 4 °C and dehydrated in 100% methanol. After rehydration with graded methanol series (75–50–25% methanol in PBS) and PBT (0.2% Triton X-100 in PBS) for 5 min each at RT, larvae were depigmented (2% KOH and 3% H₂O₂ in PBT for 5 min, followed by a 5 min wash in ddH₂O) and permeabilized (15 min at − 20 °C in ice-cold 100% acetone, followed by washes in ddH₂O and PBT). After a step in a blocking solution (1% BSA and 5% sheep serum in PBT) for 4 h at RT, larvae were incubated for 48 h at 4 °C with pan-neuronal anti-HuC/D (A-21272, ThermoFisher) or anti-Sox10 (GTX128374-S, GeneTex) primary antibody diluted 1:200 in the blocking solution. Larvae were then washed four times in PBT for 20 min each at RT, saturated in the same blocking solution for 4 h and incubated overnight in the dark at 4 °C with Streptavidin conjugate Alexa Fluor 555 secondary antibody (S32355, ThermoFisher) against HuC/D or goat anti-rabbit-Alexa Fluor 488 (A11034, Invitrogen) secondary antibody against Sox10 diluted 1:1000. After extensive washes in PBT, larvae were mounted in 1% low-melting agarose and imaged by z-stacks (3 µm step size) under a 20× objective at Nikon C2 confocal microscope. The same body region was imaged for all larvae. All HuC/D⁺ enteric neurons and Sox10⁺ neuronal progenitors visible on the ventral side of the intestine were manually scored by scrolling through the z-stacks, assisted by the Cell counter tool in Fiji-ImageJ to avoid repeated counting. N ≥ 10 animals/condition. The experiment was repeated three times.

FET test was performed according to the Organization for Economic Co-operation and Development (OECD, Paris, France) Guideline No. 236 (2013) [34] to define a non-lethal and non-toxic BSSG concentration for zebrafish treatment, in order to avoid teratogenic or developmental issues.

Briefly, 6 hpf embryos were transferred singularly into 24 well plates (1 embryo in 1 mL solution/well) and incubated with increasing concentrations of BSSG (2.5–5–10–20–40 µM) or the corresponding solvent control (DMSO at 0.1%–0.2%–0.4%). For each concentration, 20 embryos were individually incubated with BSSG, and the remaining 4 wells were used as internal negative controls (Fish Water). Negative and positive controls (1.5% EtOH) were also tested. The embryo medium was changed daily, and the developmental status of zebrafish embryos and larvae was monitored until 4 dpf. Survival and hatching rates, presence of cardiac edema and of swimming bladder were assessed based on the total number of surviving embryos.

Statistical analyses were performed using Graph Pad Prism V10.2.3. Data are expressed as mean ± SEM and statistical significance was calculated using unpaired Student's t-test for comparisons between two groups or one-way ANOVA followed by Tukey’s post-hoc test for multiple comparisons. For ex vivo gut contractility analyses, two-way ANOVA followed by Bonferroni post hoc test was used for multiple comparison. The differences between groups were considered significant at P < 0.05. Post hoc tests were run only if F achieved P < 0.05, and there was no significant variance inhomogeneity. For the gastrointestinal transit assay, differences in the distribution of food bolus across transit zones between control and treated larvae were assessed at each time point using Fisher’s exact test for 2 × 5 contingency tables based on absolute larval counts. p-values are indicated with the following symbols: *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001; ns, not significant.

Zebrafish larvae exposed to 10 µM BSSG presented normal development (Additional file 3: Figure 1 A, B) and mass spectrometry analysis of lipids extracted from larval heads and trunks confirmed the accumulation of this compound in the trunk region of treated animals (Additional file 3: Figure 1 C).

The presence of dark aggregates in the intestine of nearly all larvae exposed to BSSG prompted us to hypothesize that the gut may represent its first target (Fig. 1A). Consistently, in vivo analysis of acidified lysosomes evidenced a reduction in the number of lysosome-rich enterocytes (LREs) in treated larvae (Fig. 1B). To determine whether these intestinal effects were specifically attributable to BSSG rather than to possible metabolites, we performed the same assay on larvae treated with β-sitosterol, which shares the chemical structure of BSSG but lacks the glucose moiety. This molecule did not affect LREs (Additional file 3: Figure 2 A), suggesting that the detrimental effects only depend on BSSG consumption.

Mucus-secreting goblet cells along the mid-intestine, essential for protecting the gut against digestive enzymes and external insults, were reduced after exposure to BSSG (Fig. 1C). Conversely, the number of neutrophils infiltrating the mid-intestine was markedly increased in treated larvae (Fig. 1D), providing an indication of intestinal inflammation onset. This was further supported by increased fluorescence in the intestine of BSSG-treated transgenic Tg(NFκB:GFP) larvae, indicating the activation of NF-κB pathway (Fig. 1E). Consistently with the occurrence of an inflammatory phenotype, we also observed increased apoptosis in the mid-intestine (Fig. 1F).

Moreover, the reduction of Stat3-linked fluorescence specific of stem-like cells located at the base of the intestinal folds [20], which are analogous to crypt base columnar cells in mammals, may reflect a depletion of proliferating cells (Fig. 1G).

To further characterize the impact of BSSG on gut homeostasis, we analysed the expression of inflammation-related markers in whole larvae following acute exposure. We observed a trend towards increased expression of mmp9 (matrix metallopeptidase 9), along with significant upregulation of stat3 (signal transducer and activator of transcription 3) and pept1 (peptide transporter 1). Reduced expression of agr2 (anterior gradient 2), involved in mucus production by goblet cells, further supported the presence of a defective intestinal epithelial barrier (Fig. 1H).

We also observed reduced expression levels of the autophagy-related genes atg5 (autophagyrelated5) and lc3b (microtubule associated protein 1 light chain 3b), suggesting a possible impairment of autophagy (Additional file 3: Figure 2B).

Taken together, these results depict a scenario in which BSSG administration induces intestinal inflammation, disrupting essential cellular processes, potentially exacerbating gut dyshomeostasis.

To study the impact of BSSG on intestinal physiology, we firstly analysed peristaltic frequency. Treated larvae exhibited a lower number of gut contractile waves compared to controls (Fig. 2A). To corroborate this finding, we monitored larval gastrointestinal transit along the digestive tract, ideally divided into transit zones [24] (Fig. 2B, upper panel). Over time, differences in food bolus localization indicated that treated larvae displayed delayed gastrointestinal transit (***P < 0.001 with Fisher’s exact test 24 h after feeding) (Fig. 2B, lower panel).

Since peristaltic activity is regulated by the enteric nervous system (ENS), we evaluated the number of differentiated enteric neurons and neuronal precursors (labelled with HuC/D and Sox10, respectively), but found no differences between treated and control larvae (Additional file 3: Figure 2 C), suggesting that BSSG may affect the functionality of enteric innervation rather than its density.

Therefore, we investigated the involvement of enteric neurotransmission in gut dysmotility in adult zebrafish fed with a BSSG-enriched diet over a prolonged period, to mimic chronic exposure. For the first time, both receptor-mediated and non-receptor-mediated neuromuscular responses of isolated whole intestines were evaluated in zebrafish. Muscular contractile responses were analysed by exposing zebrafish intestine to KCl, a depolarizing agent that induces Ca²⁺ release and subsequent smooth muscle contraction. A significant increase in the KCl-induced contraction was observed in intestinal preparations of treated individuals (Fig. 2C). To further evaluate the excitatory cholinergic response, gut samples were exposed to increasing concentrations of carbachol (CCh), a non-selective cholinergic agonist. Cumulative concentration–response curves revealed a significant increase in CCh-mediated contraction in treated intestines (Fig. 2D), suggesting altered cholinergic neurotransmission following BSSG treatment. Then, to verify the impact on ENS neuronal activity, intestinal preparations were subjected to electrical field stimulation (EFS) at increasing frequencies with constant voltage, to induce neuronal depolarization and consequent neurotransmitters release. Treated individuals displayed an increased excitatory neuromuscular response, determining a significant upward shift of the frequency–response curve to EFS (Fig. 2E). Higher 10 Hz-EFS-mediated contraction confirmed alterations in excitatory cholinergic response. Conversely, muscular relaxation induced by isoprenaline, a non-selective β-adrenergic receptor agonist, was not affected by treatment (Fig. 2F).

Next, given the relevance of the microbiota in maintaining intestinal homeostasis and the well-established association between its alterations and intestinal inflammation in humans and animal models [35], we analysed microbiota composition in adult zebrafish following BSSG exposure. At the phylum level, Proteobacteria, Bacteroidota and Verrucomicrobiota increased, while Firmicutes and Actinobacteriota decreased (Fig. 2G), resulting in a reduced Firmicutes/Bacteroidetes ratio, a widely recognized indicator of gut health [36]. At the family level, we observed an increase in Barnesiellaceae, Aeromonadaceae and Rubritaleaceae, and a reduction in the relative abundance of Erysipelotrichaceae, Enterobacteriaceae and Weeksellaceae (Fig. 2H). Overall, although these results did not reach statistical significance, likely due to the limited sample size, they suggest that BSSG treatment may influence the gut microbial composition, potentially contributing to the onset of dysbiosis.

RNAseq analysis of chronically treated zebrafish larvae revealed 261 differentially expressed genes (168 upregulated and 93 downregulated) compared to controls. Upregulated genes were mainly involved in acute inflammatory response as well as in response to reactive oxygen species and defence response to bacterium (Fig. 3A). The upregulation of both mmp9 and pept1, in association with increased expression of saa (serum amyloid a) and s100a10a (s100 calcium binding protein a10 a) (Fig. 3B), corroborated previous findings in treated larvae (Fig. 1H), indicating the presence of gut inflammation.

BSSG treatment downregulated genes mainly associated with lateral line nerve development and oxygen transport (Fig. 3Aʹ). Decrease of oxygen binding resulted also among downregulated molecular functions (Additional file 3: Figure 3 A, right panel). Notably, the entire group of hemoglobin (hb) genes exhibited a significant reduction in treated individuals (Fig. 3B; Additional file 3: Figure 3B), a finding recently associated with the pathophysiology of different neurodegenerative diseases [37–39]. The decreased expression of mucin 5.3 (muc5.3) (Fig. 3B), secreted by goblet cells to neutralize digestive enzymes and pathogens, is consistent with the reduced goblet cells number observed in treated larvae (Fig. 1C).

To validate transcriptome results, we tested the expression of specific genes in the intestine of adult zebrafish confirming the upregulation of inflammation-related markers (mmp9, mmp13, il-4, il-13) and cellular stress markers (casp8, caspase 8; nupr1, nuclear protein 1) (Fig. 3C, upper panel).

Since RNAseq analysis evidenced the alteration of genes potentially involved in neurodegeneration, we also evaluated adult zebrafish brains. In contrast to intestinal findings, mmp9 expression in the brain was downregulated following treatment, potentially impacting its role in central nervous system (CNS) plasticity. We also observed the downmodulation of autophagy-related genes atg5, lc3b and p62 (Fig. 3C, lower panel).

To investigate whether BSSG dietary consumption could determine intestinal alterations also in the mouse model, WT mice were fed a BSSG-enriched diet for 15 weeks, followed by characterization of intestinal phenotype. We observed an increase in the number of resident and/or recruited macrophages in the lamina propria of the small intestine of treated mice (Additional file 3: Figure 4 A). We then evaluated goblet cell number and found a significant reduction in treated mice (Fig. 4A), evidencing also in this model hallmarks of intestinal inflammation following BSSG dietary uptake. Ultrastructural analysis revealed that BSSG-fed mice exhibited significantly shorter microvilli compared to controls (Fig. 4B), potentially implying impaired enteric absorption capacity. Consistently, treated mice weighed less than controls (Fig. 4C) despite similar food intake (Additional file 3: Figure 4B).

RNAseq analysis of gut tissue revealed 1835 differentially expressed genes between treated and untreated mice (982 upregulated and 853 downregulated in treated animals). Upregulated genes were primarily associated with regulation of immune system process and immune response (Fig. 4D, E; Additional file 3: Figure 4 C, left panel) and were related with actin cytoskeleton, side of membrane, and microvillus (Additional file 3: Figure 4D), thus confirming that BSSG directly impacts on genes involved in enterocytes’ brush border structure. Conversely, downregulated genes were involved in small molecules metabolic process, regulation of cell growth, heme/porphyrin-containing compound biosynthesis, and regulation of neuron projection development (Fig. 4Dʹ, E), and were associated with glutathione transferase activity and transmembrane transporter activity (Additional file 3: Figure 4 C, right panel). RT-qPCR analysis confirmed the activation of inflammatory and immune responses, evidencing upregulation of Nod1 (NOD-like receptor 1), Tlr-2, Tlr-4, Tlr-6 (Toll-like receptor-2, −4, −6), Nlrp3 (NLR family pyrin domain containing 3), Il-1β (Interleukin-1β) and Ifn-γ (Interferon-γ) and downregulation of Ikb-α (NF-kappa-B inhibitor alpha) and Reg3-γ (Regenerating islet-derived protein 3-γ). We also observed upregulation of Plp1 (Proteolipid protein 1), which is highly expressed in enteric glia (Fig. 4F).

Finally, analysis of fecal microbiota revealed some differences in bacterial composition between BSSG-fed mice and controls. Treated mice showed a higher relative abundance of potentially pathogenic bacterial families, including Bacteroidaceae, Helicobacteraceae and Prevotellaceae, typically associated with intestinal inflammation, and reduced anti-inflammatory taxa such as Lachnospiraceae (Additional file 3: Figure 4E). As observed in zebrafish, these data may reflect early alterations in gut microbiota in this pre-symptomatic ALS-PDC mouse model. However, no definitive conclusions can be drawn due to the lack of statistical significance, likely resulting from the limited sample size.

To investigate a possible mechanism of action of BSSG, we examined whether it could interact with known steroid hormones receptors, given its structural similarity to these molecules. A radioligand binding assay demonstrated that BSSG determined a 12.5% inhibition of the binding between glucocorticoid receptor (GR) and its radiolabelled specific ligand ([³H]-dexamethasone) and a 4.7% inhibition for the androgen receptor (AR) and its radiolabelled specific ligand ([³H]-methyltrienolone). No interference with estrogen receptor (ER), mineralcorticoid receptor (MR) or progesterone receptor (PR) was observed (Additional file 3: Figure 3 C).

To verify if BSSG effectively binds Gr in vivo, we exploited the cyp11c1 zebrafish mutant line recently generated in our laboratory. As other published cyp11c1 (cytochrome P450 family 1 subfamily C member 1) zebrafish mutant lines, homozygous cyp11c1^−/− cannot synthetize active glucocorticoids (GCs), while retaining a functional Gr [40, 41]. To visualize Gr activity, this cyp11c1 mutant line was crossed with transgenic Tg(GRE:EGFP) line, which expresses GFP after Gr activation [21] (Fig. 5A). Thus, cyp11c1^−/−;Tg(GRE:EGFP) animals, lacking endogenous GCs, allow to visually discriminate the effective binding of BSSG with the Gr in vivo. Following BSSG treatment, cyp11c1^−/−;Tg(GRE:GFP) larvae showed a significant increase in intestinal fluorescence, suggesting that BSSG can effectively modulate Gr activation (Fig. 5B). The same analysis performed in adult transgenic mutants exposed to BSSG-enriched diet also revealed a significant increase of intestinal fluorescence (Fig. 5C). The quantification of gut fluorescence in cyp11c1^+/+; Tg(GRE:GFP) larvae and adults showed, as expected, a higher baseline signal compared with cyp11c1^−/−;Tg(GRE:GFP), due to constitutive activation of Gr by endogenous glucocorticoids [21]. Accordingly, BSSG treatment in cyp11c1^+/+;Tg(GRE:GFP) larvae failed to induce any further detectable increase of intestinal fluorescence.

To confirm the role of the Gr in mediating the action of BSSG, we exploited the nr3c1^ia30/ia30 zebrafish mutant line (hereafter called gr^−/−), previously generated in our laboratory, in which the gr gene has been knocked out [17] (Fig. 6A).

Interestingly, no difference in the number of goblet cells was observed in BSSG-treated gr^−/− larvae compared to controls (Fig. 6B), in contrast to what observed in treated WT individuals (Fig. 1C). Similarly, no significant alterations were detected in the expression of mmp9, il-4 and il-13 in the intestine of adult gr^−/− zebrafish fed with BSSG-enriched diet. These findings suggest that BSSG likely interferes with the anti-inflammatory Gr activity. Conversely, the expression of casp8 and nupr1 was significantly upregulated, similarly to what was observed in WT animals, indicating that BSSG likely interferes with multiple pathways and cellular targets beyond Gr (Fig. 6C upper panel and Fig. 6C’).

Interestingly, also mRNA levels of mmp9, p62, atg5 and lc3b did not differ between brains of treated and control gr^−/− zebrafish (Fig. 6C lower panel and Fig. 6Cʹ).

Following in vivo evidence of a potential interaction between BSSG and Gr, we investigated whether BSSG could also impact on the fine-tuned regulation of gene expression commonly mediated by GCs/Gr signalling. We therefore evaluated the expression of Gr-target genes such as fkbp5 (FKBP prolyl isomerase 5) and foxo3b (forkhead box protein O 3b) in the intestine of WT adults fed with BSSG. We observed a significant reduction in fkbp5 expression, while foxo3b displayed a decreasing trend in treated individuals (Fig. 6D), further supporting our hypothesis.

To evaluate whether the impact of BSSG on intestinal neuromuscular function is modulated by Gr, we applied the ex vivo approach in gr^−/− adult zebrafish after administration of BSSG-enriched diet. We observed that gut samples from gr^−/− animals did not show any variation in muscle- and neuron-induced contractility when exposed to KCl, 1 µM CCh and 10 Hz EFS after BSSG treatment, maintaining responses comparable to untreated gr^−/− controls (Fig. 7A–C; for concentration–response curves see Additional file 3: Figure 2D, E). These data point out that defects in gut motility could be influenced by BSSG interaction with Gr and are abolished in gr^−/− mutant line.

Moreover, BSSG treatment appeared to induce fewer changes in the gut microbiota composition of gr^−/− animals compared with those reported for WT (Fig. 7D, E). At the phylum level, no evident alterations were detected, except for a slight increase in Firmicutes (Fig. 7D, right side). At the family level, we observed similar abundance of Barnesiellaceae, Erysipelotrichaceae and Rubritaleaceae, suggesting that BSSG exposure has a limited impact on these bacterial strains in gr^−/− individuals (Fig. 7E, right side).

It should be noted that the lack of the Gr is associated with baseline differences in gut microbiota composition. In particular, Fusobacteriota were more abundant in gr^−/− animals, whereas Firmicutes and Actinobacteriota were less represented, in opposition to what observed in WT. At the family level, Fusobacteriaceae, Comamonadaceae and Shewanellaceae were more abundant in gr^−/− than in WT. Nevertheless, our findings suggest that BSSG influences the gut microbial community differently in gr^−/− animals compared to WT, as also supported by 2D PCoA analysis and ASVs (Amplicon Sequence Variants) relative abundance calculated across all experimental groups (Additional file 3: Figure 5), where gr^−/− individuals appear more similar to each other regardless of BSSG treatment.

Not applicable.

Not applicable.

The authors declare no competing interests.

Transcriptional data generated during the current study have been deposited in SRA database (URL will be made available upon request). Other data supporting the findings of this study are available on reasonable request to the corresponding authors.