Unveiling the mechanism of abdominal and transcranial ultrasound stimulation against DSS-induced colitis based on proteomic analysis

Abstract

Inflammatory bowel disease (IBD), particularly ulcerative colitis (UC), is increasingly recognized for its systemic effects, including neuroinflammation and cognitive deficits mediated through the gut-brain axis. This study investigates the potential mechanisms for treating UC with low-intensity pulsed ultrasound (LIPUS). A murine model of UC was established using 3% dextran sulfate sodium (DSS) in C57BL/6J mice. Disease progression was monitored via the Disease Activity Index (DAI). Histopathological evaluations were conducted using Hematoxylin and Eosin (H&E) staining. To elucidate molecular alterations, hippocampal tissues underwent quantitative proteomic analysis employing high-throughput liquid chromatography-tandem mass spectrometry (LC-MS/MS). Differentially expressed proteins (DEPs) were identified and analyzed to understand the impact of both abdominal and transcranial LIPUS treatments. Both abdominal and transcranial LIPUS treatments were found to alleviate symptoms of colitis. Proteomic analysis of hippocampal tissues identified five DEPs—REPS1, MYG1, KRT13, SRSF10, and CDC42BPG—whose expression levels were modulated by LIPUS interventions. Notably, REPS1 and MYG1, which were downregulated in UC conditions, showed increased expression following LIPUS treatment. KEGG pathway enrichment analysis revealed that these DEPs are primarily involved in the Ras/MAPK signaling pathways. The modulation of these pathways by LIPUS suggests a mechanism by which it exerts anti-inflammatory effects, potentially restoring metabolic balance and reducing inflammation in both the gut and brain. These findings highlight the role of the gut-brain axis in mediating the beneficial effects of LIPUS and suggest its potential as a non-invasive therapeutic strategy for UC and associated neuroinflammatory conditions.

Supplementary Information

The online version contains supplementary material available at 10.1186/s13036-025-00619-4.

Introduction

Inflammatory bowel disease (IBD), encompassing ulcerative colitis (UC) and Crohn’s disease (CD), is a chronic inflammatory disorder of the gastrointestinal tract that not only affects the digestive system but also contributes to systemic complications, including neuroinflammation and cognitive impairment [1]. The gut-brain axis is a critical mediator of these interactions, connecting intestinal inflammation with systemic manifestations via immune, neural, and molecular signaling pathways [2–4]. Pro-inflammatory cytokines, such as TNF-α and IL-6, released in response to intestinal inflammation, can traverse the blood-brain barrier, exacerbating neuroinflammation and perpetuating a detrimental feedback loop [5, 6]. Elucidating the molecular mechanisms governing this intricate interplay is crucial for developing targeted therapeutic strategies.

Proteomic analysis has become a valuable approach for detecting molecular alterations in dextran sulfate sodium (DSS)-induced colitis, a widely utilized model of IBD. An acute colitis mouse model was established using DSS. This compound does not cross the blood-brain barrier, making it an ideal system for investigating the directional inflammatory response from the gut to the brain [7]. Previous studies have highlighted critical pathways and proteins involved in disease progression, with a focus on differential protein expression. For instance, Wong et al. highlighted the role of exosomes in macrophage activation, whereas Qi et al. investigated immune and oxidative stress pathways in relation to conventional therapeutic strategies [8, 9]. Although these advancements have enhanced our understanding, most studies have predominantly concentrated on validating proteomic findings, providing limited insight into the underlying mechanisms of disease progression and therapeutic responses.

Several studies have highlighted the significance of low-intensity pulsed ultrasound (LIPUS) as a noninvasive brain stimulation modality with anti-inflammatory properties for preventing neuroinflammation [10, 11]. Furthermore, abdominal LIPUS treatment has been shown to alleviate colitis severity by activating the splenic nerve and modulating the cholinergic anti-inflammatory pathway [12]. Our research group has demonstrated a novel, non-invasive therapeutic approach for LPS-induced neuroinflammation by applying LIPUS [10, 13]. Recently, further studies have explored that LIPUS, delivered either transcranially or abdominally, effectively mitigates inflammation in both the gut and brain [14, 15]. This attenuation of inflammation may alleviate inflammation-induced behavioral changes and associated disorders, potentially through gut-brain communication mechanisms. However, the underlying mechanisms by which transcranial and abdominal ultrasound stimulation simultaneously alleviate UC and neuroinflammation remain to be elucidated.

This study employed quantitative proteomic techniques to assess protein expression in the hippocampal tissues of UC mice. Differentially expressed proteins (DEPs) were characterized for their biological functions, and the modulatory effects of both abdominal and transcranial LIPUS were evaluated. The aim was to identify UC-relevant protein targets and to elucidate further the mechanisms underlying the therapeutic effects of these LIPUS modalities on UC. Focusing on Ras/MAPK signaling pathways, our findings lay the groundwork for innovative strategies to alleviate IBD and its systemic complications. These results offer valuable insights into regulating inflammation in both the gut and the brain, paving the way for new therapeutic interventions.

Results

LIPUS stimulation alleviated the symptoms of colitis and neuroinflammation

An acute colitis model was established in mice using DSS to investigate the therapeutic potential of LIPUS stimulation in UC. DSS administration induced a progressive loss of body weight, starting on day 5, which reached a maximum reduction of approximately 14.9% by day 7, compared with the Sham group (p < 0.001; Fig. 1A). However, neither abdominal nor transcranial LIPUS treatment significantly attenuated the body weight loss. In contrast, the DAI score increased sharply in the DSS group after day 3 and peaked on day 7 (p < 0.001 vs. Sham; Fig. 1B). Notably, both LIPUS interventions significantly decreased the DAI from day 6 onward, with lower scores maintained through day 7 (both p < 0.01 vs. DSS; Fig. 1B). Additionally, the colon length of DSS-treated mice was significantly shorter than that of the Sham group (p < 0.001; Fig. 1C). However, no significant difference in colon length was observed between the DSS-only group and the two LIPUS-treated groups. The DSS-only group exhibited a significantly higher spleen weight-to-body weight ratio compared to the Sham group (p < 0.001; Fig. 1D). In contrast, both LIPUS treatments led to a significant reduction in this ratio (both p < 0.01; Fig. 1D). Furthermore, H&E staining revealed that the DSS group exhibited disrupted glandular architecture, a decrease in goblet cells, and extensive infiltration of inflammatory cells. In contrast, both LIPUS treatments significantly alleviated these histological abnormalities (Fig. 1E) and led to a significant reduction in histopathological scores (both p < 0.001; Fig. 1F). DSS exposure led to a marked upregulation of the proinflammatory cytokines TNF-α, IL-1β, and IL-6 in the colon (all p < 0.01; Fig. 1G–I). In contrast, administration of either LIPUS intervention significantly suppressed the elevated levels of these cytokines compared with the DSS-only group (all p < 0.05; Fig. 1G–I). A similar inflammatory response was observed in the hippocampal tissues of mice, where DSS significantly enhanced the expression of TNF-α, IL-1β, and IL-6 (all p < 0.05; Fig. 1J–L). Both LIPUS treatments effectively attenuated these increases (all p < 0.05; Fig. 1J–L).

Fig. 1.

LIPUS attenuated colon damage and neuroinflammation in DSS-induced acute colitis. (A) Changes in body weight in different groups. (B) DAI scores for each group. (C) Measurement of colonic length. (D) LIPUS mitigated the elevation of the spleen weight-to-body weight ratio in DSS-induced acute colitis. (E) Representative hematoxylin and eosin (H&E) stained images and (F) histological scoring of tissue damage. (G) TNF-α, (H) IL-1β, and (I) IL-6 mRNA in the colons were quantified with qRT-PCR. (J) TNF-α, (K) IL-1β, and (L) IL-6 mRNA in the hippocampus were quantified with qRT-PCR. In (A) and (B), *, ^#, and ^† denote a significant difference between the DSS group and the Sham group, DSS + LIPUS_gut group, and DSS + LIPUS_brain group, respectively. In (C), (D), and (F)-(L), * and ^# denote significant differences from the Sham group and the DSS group, respectively (*^,#, p < 0.05;**^,##,††, p < 0.01; ***^,###, p < 0.001; (A)-(D) n = 8; (F)-(L) n = 6). DAI: disease activity index; DSS: dextran sulfate sodium; LIPUS: low-intensity pulsed ultrasound

Quantification of differentially expressed proteins

The volcano plot in Fig. 2A-C illustrates the DEPs between the groups. Quantitative analysis revealed that, compared to the Sham group, DSS-induced UC mice exhibited 28 DEPs, with 17 upregulated and 11 downregulated. In the LIPUS_gut group, 28 DEPs were identified compared to the DSS group, comprising 10 upregulated and 18 downregulated proteins. Meanwhile, the LIPUS_brain group displayed 25 DEPs relative to the DSS model group, with six proteins showing upregulation and 19 demonstrating downregulation (Table 1).

Fig. 2.

Volcano plot and Venn diagram of differentially expressed proteins (DEPs) between groups. (A) Volcano plot of DEPs in Sham and DSS group. (B) Volcano plot of DEPs between the DSS model and LIPUS_gut group. (C) Volcano plot of DEPs between the DSS model and LIPUS_brain group. (D) Venn diagram of DEPs in DSS/Sham and DSS + LIPUS_gut/DSS group. (E) Venn diagram of DEPs in the DSS/Sham group that DSS + LIPUS_brain/DSS group. (F) Venn diagram of DEPs in DSS/Sham, DSS + LIPUS_gut/DSS group, and DSS + LIPUS_brain/DSS group. The x-axis displays protein expression values as log2-transformed fold changes, while the y-axis represents the adjusted p-values transformed as -log10. DSS: dextran sulfate sodium; LIPUS: low-intensity pulsed ultrasound

Table 1.

List of differential protein numbers

Compare group	Up-regulated	Down-regulated	All-regulated
DSS/Sham	17	11	28
DSS + LIPUS_gut/DSS	10	18	28
DSS + LIPUS_brain/DSS	6	19	25

An overlap analysis of the DEPs among the Sham group, LIPUS_gut group, and LIPUS_brain group was conducted (Fig. 2D-F). Comparison between the DSS/Sham and DSS + LIPUS_gut/DSS groups revealed 11 DEPs shared among the Sham, DSS, and DSS + LIPUS_gut groups. Seven proteins [REPS1, MYG1, PTGR2, DAZAP1, TXNDC5, EIF3K, RGS7BP] were downregulated in UC. Still, they showed upregulation following DSS + LIPUS_gut treatment, while four proteins [KRT13, SRSF10, CDC42BPG, SLC25A51] exhibited the opposite trend, being upregulated in UC but downregulated after DSS + LIPUS_gut treatment (Fig. 2D; Table 2). Similarly, a comparison between the DSS/Sham and DSS + LIPUS_brain/DSS groups identified 11 DEPs common to the Sham, DSS, and LIPUS_brain groups. Among these, four proteins (REPS1, MYG1, NDUFS4, REM2) were downregulated in UC but upregulated following LIPUS_brain treatment, whereas seven proteins (TSFN, KRT13, SRSF10, CDC42BPG, ACTBL2, MAPK8IP3, NT5C2) were upregulated in UC but downregulated after LIPUS_brain intervention (Fig. 2E; Table 3). Moreover, comparison among the DSS/Sham, DSS + LIPUS_gut/DSS, and DSS + LIPUS_brain/DSS groups revealed five DEPs shared among the Sham, DSS, DSS + LIPUS_gut, and DSS + LIPUS_brain groups. Two proteins [REPS1, MYG1] were downregulated in UC. Still, they showed upregulation following both LIPUS treatments, while three proteins [KRT13, SRSF10, CDC42BPG] exhibited the opposite trend, being upregulated in UC but downregulated after both LIPUS treatments (Fig. 2F; Tables 2 and 3).

Table 2.

Differentially expressed proteins in dextran sulfate sodium-induced ulcerative colitis model group mice regulated by LIPUS_gut

Symbol	Protein ID	Protein Name	DSS/Sham		DSS + LIPUS_gut/DSS
			Normalized ratio	Q value	Normalized ratio	Q value
REPS1	O54916	RalBP1-associated Eps domain-containing protein 1	0.01	< 0.001	95.97	< 0.001
MYG1	Q9JK81	MYG1 exonuclease	0.01	< 0.001	95.97	< 0.001
PTGR2	Q8VDQ1	Prostaglandin reductase 2	0.01	< 0.001	95.97	< 0.001
DAZAP1	Q9JII5	DAZ-associated protein 1	0.01	< 0.001	95.97	< 0.001
TXNDC5	Q91W90	Thioredoxin domain-containing protein 5	0.01	< 0.001	95.97	< 0.001
EIF3K	Q9DBZ5	Eukaryotic translation initiation factor 3 subunit K	0.01	< 0.001	95.97	< 0.001
RGS7BP	Q8BQP9	Regulator of G-protein signaling 7-binding protein	0.27	< 0.001	3.43	< 0.001
KRT13	P08730	Keratin, type I cytoskeletal 13	101.73	< 0.001	0.01	< 0.001
SRSF10	Q9R0U0	Serine/arginine-rich splicing factor 10	101.73	< 0.001	0.01	< 0.001
CDC42BPG	Q80UW5	Serine/threonine-protein kinase MRCK gamma	101.73	< 0.001	0.01	< 0.001
SLC25A51	Q5HZI9	Mitochondrial nicotinamide adenine dinucleotide transporter SLC25A51	101.73	< 0.001	0.01	< 0.001

Table 3.

Differentially expressed proteins in dextran sulfate sodium-induced ulcerative colitis model group mice regulated by LIPUS_brain

Symbol	Protein ID	Protein Name	DSS/Sham		DSS + LIPUS_brain/DSS
			Normalized ratio	Q value	Normalized ratio	Q value
REPS1	O54916	RalBP1-associated Eps domain-containing protein 1	0.01	< 0.001	94.88	< 0.001
MYG1	Q9JK81	MYG1 exonuclease	0.01	< 0.001	94.88	< 0.001
NDUFS4	Q9CXZ1	NADH: Ubiquinone Oxidoreductase Subunit S4	0.01	< 0.001	94.88	< 0.001
REM2	Q8VEL9	GTP-binding protein REM 2	0.36	< 0.001	2.78	< 0.001
TSFM	Q9CZR8	Elongation factor Ts, mitochondrial	2.66	< 0.001	0.32	< 0.001
KRT13	P08730	Keratin, type I cytoskeletal 13	101.73	< 0.001	0.01	< 0.001
SRSF10	Q9R0U0	Serine/arginine-rich splicing factor 10	101.73	< 0.001	0.01	< 0.001
CDC42BPG	Q80UW5	Serine/threonine-protein kinase MRCK gamma	101.73	< 0.001	0.01	< 0.001
ACTBL2	Q8BFZ3	Beta-actin-like protein 2	101.73	< 0.001	0.01	< 0.001
MAPK8IP3	Q9ESN9	C-Jun-amino-terminal kinase-interacting protein 3	101.73	< 0.001	0.01	< 0.001
NT5C2	Q3V1L4	Cytosolic purine 5’-nucleotidase	101.73	< 0.001	0.01	< 0.001

The effect of LIPUS stimulation on the alteration of hierarchical clustering and rank plot

Hierarchical clustering analysis using heat maps enables the identification of proteins that exhibit similar expression patterns across different experimental conditions. A heatmap was generated to visualize protein expression clusters among DSS, DSS + LIPUS_gut, and DSS + LIPUS_brain groups using the top 10 and bottom 10 proteins with a mean of log2fold change (Fig. 3A). Compared to the DSS/Sham group, the overall protein expression profiles of the DSS + LIPUS_gut/DSS and DSS + LIPUS_brain/DSS groups were more similar. Besides, a rank plot visualizes the relationship between fold change across samples and the corresponding ranked protein expression values. In both the DSS + LIPUS_gut/DSS (Fig. 3B) and DSS + LIPUS_brain/DSS (Fig. 3C) groups, REPS1 and MYG1 exhibited the highest fold change rankings among the differentially expressed proteins, compared to all other proteins. KRT13, SRSF10, and CDC42BPG were the three proteins with the lowest fold change in both groups. We further validated the expression of REPS1 and MYG1, which were identified as LIPUS-responsive proteins in the proteomic analysis. Western blot analysis revealed that, compared with the Sham group, the hippocampal levels of REPS1 and MYG1 were significantly reduced in UC mice (all p < 0.01; Fig. 3D, E). Notably, both abdominal and transcranial LIPUS treatments markedly restored the expression of these proteins (all p < 0.05; Fig. 3D, E). These findings confirm that the western blot results are consistent with the proteomic data, supporting the reliability of the identified LIPUS-regulated targets.

Fig. 3.

Heatmap clustering, protein rank, and western blot analysis. (A) Heatmap of different proteins between the DSS group and the Sham group, DSS + LIPUS_gut group, and DSS + LIPUS_brain group. The x-axis denotes various experimental groups, while the y-axis corresponds to different proteins. Expression values are normalized and converted to a log2 scale. In the heat map, proteins with higher expression levels are represented in red, whereas those with lower levels are shown in blue. (B) and (C) Rank plot depicts the ranked protein expression values and the fold change between the DSS group and the DSS + LIPUS_gut group and DSS + LIPUS_brain group, respectively. The protein with the highest fold change is assigned rank 1 on the left side of the x-axis, while the protein with the lowest fold change is positioned at the final rank on the right side of the x-axis. (D) and (E) Western blot analysis of REPS1 and MYG1 protein expression levels in the Sham and DSS groups, with showing the effect of LIPUS_gut treatment and LIPUS_brain treatment. * and ^# denote significant differences from the Sham group and the DSS group, respectively (^#, p < 0.05; **^,##, p < 0.01; ***^,###, p < 0.001; n = 7). DSS: dextran sulfate sodium; LIPUS: low-intensity pulsed ultrasound

Gene ontology enrichment analysis

The GO provides a structured framework for classifying gene functions, facilitating the interpretation of gene roles in the context of DEPs. Figure 4 depicts the 10 most significantly enriched GO terms across the three main categories: BP, CC, and MF. Compared to the DSS/Sham group (Fig. 4A), the GO functional annotations of DEPs in the DSS + LIPUS_gut/DSS (Fig. 4B) and DSS + LIPUS_brain/DSS (Fig. 4C) groups exhibited similar patterns. BP was the most favorable enrichment component. DEPs were enriched in the negative regulation of granulocyte differentiation, observational learning, positive regulation of myeloid cell apoptotic process, and regulation of Schwann cell proliferation. CC analysis revealed enrichment in the Cul4A-RING E3 ubiquitin ligase complex and mitotic spindle midzone, while MF analysis identified enrichment in phosphatidylethanolamine binding and muscle alpha-actinin binding. Notably, the only shared GO term among all three groups was observational learning, which was enriched within the BP category.

Fig. 4.

GO classification of all differentially expressed proteins (DEPs). (A) Biological function annotation of DEPs between the Sham and DSS group. (B) Biological function annotation of DEPs between the DSS model and DSS + LIPUS_gut group. (C) Biological function annotation of DEPs between the DSS model and DSS + LIPUS_brain group. BP: Biological process; CC: Cellular component; MF: Molecular function; GO: Gene Ontology; DSS: dextran sulfate sodium; LIPUS: low-intensity pulsed ultrasound

KEGG pathways enrichment and IPA analysis

KEGG pathway enrichment analysis was conducted to investigate the biological pathways associated with DEPs among the experimental groups. In the comparison between the Sham and DSS groups, the most significantly enriched pathways included the estrogen signaling pathway, apelin signaling pathway, and Parkinson’s disease pathway (Fig. 5A; Table 4). For the DSS + LIPUS_gut versus DSS comparison, DEPs were primarily associated with the Ras and MAPK signaling pathways (Fig. 5B; Table 5). Similarly, in the DSS + LIPUS_brain versus DSS comparison, the enriched pathways included Ras, MAPK, and pyruvate metabolism (Fig. 5C; Table 6). The initial proteomic analysis indicated that LIPUS_gut and LIPUS_brain treatment modalities share a common feature of significantly enriching the Ras and MAPK signaling pathways.

Fig. 5.

The top 10 pathways identified by KEGG. (A) KEGG pathway enrichment of DEPs between the Sham and DSS groups. (B) KEGG pathway enrichment of DEPs between the DSS model and DSS + LIPUS_gut group. (C) KEGG pathway enrichment of DEPs between the DSS model and DSS + LIPUS_brain group. The x-axis represents GeneRatio, indicating the proportion of differential genes for each KEGG pathway relative to the total number of differential genes. The y-axis displays the top 10 most significant KEGG pathways, selected based on the adjusted p-value. KEGG: Kyoto Encyclopedia of Genes and Genomes; DSS: dextran sulfate sodium; LIPUS: low-intensity pulsed ultrasound

Table 4.

DSS/sham KEGG pathway

Pathway name	Protein-ID	Number	Q value
Estrogen signaling pathway	Q9D6P8, P08730, P70313	3	0.043
Apelin signaling pathway	Q9D6P8, P70313, Q5EG47	3	0.043
Parkinson disease	Q9D6P8, Q9CXZ1, Q922F4	3	0.135

Table 5.

DSS + LIPUS_gut/DSS KEGG pathway

Pathway name	Protein-ID	Number	Q value
Ras signaling pathway	Q04690	1	0.114
MAPK signaling pathway	Q04690	1	0.119

Table 6.

DSS + LIPUS_brain/DSS KEGG pathway

Pathway name	Protein-ID	Number	Q value
Ras signaling pathway	Q04690, Q9JIW9	2	0.051
MAPK signaling pathway	Q9ESN9, Q04690	2	0.078
Pyruvate metabolism	Q99KE1	1	0.329

IPA was utilized to explore the functional networks associated with the overexpressed DEPs between groups. The most prominent networks, encompassing a substantial number of upregulated protein biomarkers in both LIPUS-treated UC models, were found to be related to REPS1 and MYG1, as illustrated in Fig. 6. Within these networks, proteins exhibiting increased expression in UC are highlighted in red, while those with decreased expression are shown in green.

Fig. 6.

Interaction network of differentially expressed proteins (DEPs) between groups. (A) The protein-protein interaction (PPI) DEP network of DEPs between the Sham and DSS group. (B) PPI network of DEPs between the DSS model and DSS + LIPUS_gut group. (C) PPI network of DEPs between the DSS model and DSS + LIPUS_brain group. Colors indicate changes in protein expression, with red indicating up-regulation and green indicating down-regulation

Discussion

IBD is increasingly linked to neuroinflammation and behavioral changes in patients [16, 17]. Abdominal and transcranial LIPUS have demonstrated significant efficacy in concurrently alleviating gut and brain inflammation. Over the past several years, the application of proteomic techniques to UC research has expanded, offering valuable insights into its pathogenesis and contributing to improvements in clinical diagnosis and therapeutic interventions. In this study, proteomic analysis was conducted to identify DEPs in the hippocampus of DSS-induced UC mice, focusing on those DEPs modulated in response to both abdominal and transcranial LIPUS treatments.

The DSS-induced colitis model is commonly employed in research [18], as it induces intestinal inflammation, activates microglia, and triggers neuroinflammation [19, 20]. Additionally, the interaction between the brain and gut inflammation can further worsen the damage to the intestine [21]. Previous studies by our group have demonstrated that LIPUS significantly attenuates DSS-induced colitis and neuroinflammation by downregulating the expression of TNF-α, IL-1β, and IL-6, and by restoring intestinal tight-junction proteins (occludin and ZO-1), as well as suppressing hippocampal inflammatory responses [14, 15]. Treatment with either abdominal or transcranial LIPUS was found to ameliorate behavioral impairments, protect against gut barrier dysfunction, and reduce the concentrations of serum lipopolysaccharide (LPS) and lipopolysaccharide-binding protein (LBP). Compared to transcranial LIPUS, abdominal LIPUS further induced alterations in the gut microbiota composition and fecal metabolomic profiles in DSS-treated mice. These findings highlight a potential role for the gut-brain axis in mediating the beneficial effects of both forms of LIPUS on behavioral impairments and colonic inflammation induced by DSS.

In the present study, our results showed that abdominal and transcranial LIPUS mitigated the severity of colitis symptoms caused by DSS administration (Fig. 1). Furthermore, proteomic analysis of the DSS/Sham, DSS + LIPUS_gut/DSS, and DSS + LIPUS_brain/DSS groups revealed five DEPs that were shared across all three groups. Notably, two proteins, REPS1 and MYG1, were found to be downregulated in UC but showed increased expression following both LIPUS treatments (Fig. 2F; Tables 2 and 3). Previous research has suggested that REPS1 may serve as a potential biomarker and therapeutic target for both Alzheimer’s disease (AD) and vascular dementia (VD). Moreover, REPS1 was linked to suppressing the Ras signaling pathway and enhancing pyruvate metabolism [22]. Myg1 has been proposed as a potential gene implicated in memory function and the pathophysiology of Alzheimer’s disease, with its expression decreased in specific brain regions affected by AD [23, 24]. The expression of Myg1 is often suppressed under conditions of cellular stress, such as cytotoxicity, nutrient starvation, and cellular degeneration [25]. The decline in Myg1 expression observed under conditions of starvation or cellular degeneration suggests that mitochondrially localized Myg1 may play a role in regulating metabolic processes. Additionally, KRT13 was found to be significantly associated with the severity of cognitive impairment, exhibiting a positive correlation with measures of cognitive dysfunction. SRSF10, a gene that regulates alternative splicing, has been widely associated with the pathogenesis of AD [26, 27]. Our findings should be considered preliminary, as they are derived solely from bioinformatic analyses. Further experimental validation is necessary to determine whether these protein expressions are altered in the brains of individuals with UC.

We further explored the potential functions of DEPs by KEGG functional enrichment analyses. KEGG analysis revealed that the majority of the DEPs were primarily involved in Ras and MAPK signaling pathways in the LIPUS-treated DSS mouse model (Fig. 5). The Ras/MAPK and NF-κB signaling pathways function as key mediators in regulating cellular responses to external stimuli, including inflammation, proliferation, and stress. Although these pathways were initially considered independent, accumulating evidence reveals that Ras/MAPK activation can influence NF-κB signaling at multiple levels. For example, components of the MAPK cascade, such as ERK and JNK, have been implicated in the activation of the IKK complex, leading to phosphorylation and degradation of IκBα, thereby promoting NF-κB nuclear entry [28]. Furthermore, constitutive Ras activity, often observed in cancer, has been shown to sustain NF-κB-driven transcription of genes involved in cell survival and inflammation, such as Bcl-xL and COX-2 [29, 30]. In addition, transcriptional cooperation between AP-1 (a downstream target of MAPK) and NF-κB enhances the expression of pro-inflammatory genes [31]. These interactions highlight a coordinated regulatory mechanism integrating mitogenic and inflammatory signals, underscoring their relevance in oncogenesis and chronic inflammatory conditions.

LIPUS exerts therapeutic effects by modulating the Ras/MAPK signaling pathway, affecting pyruvate metabolism. Bioinformatic analysis indicated that LIPUS may exert anti-inflammatory effects by modulating the Ras/MAPK signaling pathway through the regulation of DEPs. While these results suggest a potential mechanism, they remain inferential. Future studies will include functional validation, such as knockdown or inhibition of key proteins and assessment of Ras/MAPK activation, to directly confirm the regulatory effects of LIPUS. The potential involvement of pyruvate metabolism discussed in this study is hypothesis-generating and based on indirect proteomic evidence rather than direct metabolic assessment. While these findings suggest that LIPUS may influence energy metabolism and mitochondrial function, metabolomic profiling and mitochondrial respiration assays will be required in future studies to confirm these proposed mechanisms. By downregulating pyruvate kinase M2 and pyruvate dehydrogenase kinases, LIPUS promotes mitochondrial pyruvate oxidation over lactate production, helping to restore metabolic balance and reduce inflammation [32]. In colitis, this reduces intestinal inflammation and strengthens the epithelial barrier [14] while influencing gut microbiota and metabolites that affect the brain via the gut-brain axis. Additionally, LIPUS can directly reduce neuroinflammation by inhibiting microglial activation and restoring brain energy metabolism, possibly through the same MAPK-metabolic pathway [33, 34]. These coordinated actions suggest that LIPUS provides bidirectional benefits, treating gut and brain inflammation through shared molecular and metabolic pathways.

Conclusion

In summary, this study employed proteomic analysis to reveal that both abdominal and transcranial LIPUS may exert anti-inflammatory effects in DSS-induced UC mice by modulating the Ras and MAPK signaling pathways through the regulation of proteins such as REPS1, MYG1, KRT13, SRSF10, and CDC42BPG. These findings suggest a potential mechanism by which LIPUS alleviates inflammation in both the gut and brain. Further research into the specific roles of these proteins in UC pathogenesis could provide deeper insights into the therapeutic mechanisms of LIPUS for UC and its associated neurological manifestations.

Materials and methods

Establishment of the colitis model and experimental protocol

Eight-week-old male C57BL/6J mice (22–25 g) were obtained from BioLASCO Taiwan Co., Ltd. The animals were maintained under a 12-hour light/dark cycle with free access to food and water. All procedures were strictly followed by the guidelines approved by the Animal Care and Use Committee of National Yang Ming Chiao Tung University. Colitis was induced by providing a 3% (wt/vol) dextran sulfate sodium (DSS, molecular weight 36–50 kDa) solution in the drinking water for 7 days. In contrast, the sham group received standard water (Fig. 7A). After DSS administration, the mice were randomly divided into Sham, DSS, DSS + LIPUS_gut, and DSS + LIPUS_brain. In both the DSS + LIPUS_gut and DSS + LIPUS_brain groups, low-intensity pulsed ultrasound (LIPUS) was applied at an intensity of 0.5 W/cm² from day 4 to day 7, targeting the gut and brain, respectively. The LIPUS treatment protocol involved anesthetic induction using isoflurane combined with oxygen; for consistency, the sham and DSS groups were also subjected to isoflurane anesthesia during this period. On day 7 following DSS induction, the mice were euthanized, and hippocampal tissues were harvested for subsequent analysis.

Fig. 7.

Study design and LIPUS system configuration. (A) Experimental timeline. Mice received 3% DSS in drinking water for 7 days, while LIPUS treatment (red dashed line) was applied to the brain or abdomen daily from day 4 to day 7. (B) Illustration of the LIPUS setup targeting the brain and abdomen in DSS-treated mice. DSS: dextran sulfate sodium; LIPUS: low-intensity pulsed ultrasound

Ultrasound stimulation system

The ultrasound system was configured using a therapeutic ultrasound generator (ME740, Mettler Electronics, Anaheim, CA) with a 1-MHz planar transducer (ME7413, Mettler Electronics) with a 4.4‐cm² effective radiating area. The device was operated with a burst duration of 2 ms, a duty cycle of 20%, and a repetition frequency of 100 Hz, with the spatial average intensity at the transducer head determined to be 0.5 W/cm² via a radiation force balance (Precision Acoustics, Dorset, UK) in degassed water. For transcranial LIPUS stimulation, the transducer was mounted onto a detachable aluminum cone (10 mm tip diameter). A stereotaxic apparatus (Stoelting, Wood Dale, IL, USA) was employed to precisely focus the ultrasound on a target located 2.0 mm posterior to the bregma (Fig. 7B). Conversely, for abdominal LIPUS treatment, mice were shaved, and ultrasound transmission gel (Pharmaceutical Innovations, Newark, NJ, USA) was applied between the plane transducer and the abdomen to optimize acoustic coupling. The treatment protocol encompassed the entire abdominal region—from the diaphragm to the groin—with each sonication lasting 5 min and interspersed by 5-minute intervals, resulting in a cumulative daily sonication time of 15 min.

Protein extraction and LC-MS/MS-based proteomic analysis

Proteomic profiling was conducted following previously established protocols. Frozen mouse tissues were lysed using a buffer composed of phosphate-buffered saline (PBS) supplemented with 0.1% Triton X-100, PhosSTOP™ phosphatase inhibitor, and cOmplete™ protease inhibitor cocktail (Roche Diagnostics Corporation, Indianapolis, IN, USA). Zirconium oxide beads (1.0 mm) were added at a volumetric ratio of 1:2:1 (tissue/lysis buffer/beads) to facilitate homogenization. Tissue disruption was performed at maximum speed for 30 s using a homogenizer (Next Advance Bullet Blender, BBX24B, Troy, NY, USA), followed by a 60-second cooling period on ice. This cycle was repeated at least three times until complete homogenization was achieved. The resulting homogenate was centrifuged at 15,000 rpm for 15 min at 4 °C, after which the supernatant was collected for protein quantification using the Pierce™ BCA Protein Assay Kit or stored at − 80 °C for future analysis. Protein digestion was carried out using the SMART Digest kit, following the manufacturer’s protocol. In brief, tissue lysates were combined with Digestion Buffer at a 1:3 ratio, and immobilized trypsin beads were introduced. The digestion reaction was conducted at 72 °C with continuous shaking for 4 h. After centrifugation, the supernatant was collected, and dithiothreitol (final concentration: 1 mM) was added, followed by incubation at 53 °C for 30 min. Iodoacetamide (final concentration: 5 mM) was introduced, and the reaction was left in the dark at room temperature for 30 min. Post-alkylation, the samples were purified using solid-phase extraction to remove unwanted contaminants. Before mass spectrometry (MS) analysis, the lyophilized peptides were reconstituted in 0.1% formic acid and separated using a nanoflow liquid chromatography system (Vanquish neo). The ionization process was performed using a nano-spray source, and peptide analysis was conducted with an Orbitrap Fusion Lumos tandem mass spectrometer operating in automated data-dependent acquisition mode. The full scan of positively charged ions was performed within an m/z range of 375–1500, with ion selection based on signal intensity within a three-second window. High-energy collision dissociation (HCD) was applied to fragment precursor ions with charge states ranging from 2 + to 7+, and MS/MS spectra were acquired for downstream data interpretation. The raw MS and MS/MS datasets were collected for further computational analysis.

Mass spectrometry data analysis

The data analysis was conducted using Proteome Discoverer software (version 2.3, Thermo Fisher Scientific). Peptide identification was performed with the Mascot search engine (Matrix Science, London, UK; version 2.5), using the SwissProt database for spectral matching. Mass tolerances were set to 10 ppm for precursor ions and 0.5 Da for CID-generated fragment ions, allowing up to two missed cleavage sites from trypsin digestion. Oxidation of methionine and N-terminal acetylation were considered variable modifications, while carbamidomethylation of cysteine was treated as a fixed modification. Peptide-spectrum matches (PSMs) were filtered based on high confidence and ranked 1 peptide in the Mascot search results, ensuring an overall false discovery rate below 1%. Proteins were considered identified if at least two unique peptides were detected. Quantification was performed by summing the peptide peak areas, as determined by the Minora algorithm. Differential protein expression was assessed using the log2-transformed protein abundance ratios, with thresholds defined as the mean ± 2 × standard deviation (SD) under the assumption of a normal distribution.

Bioinformatic analysis

To investigate the functional roles of identified proteins, annotation was carried out using the Gene Ontology (GO) database, covering three main categories: biological process (BP), cellular component (CC), and molecular function (MF). Canonical pathway analysis was performed using the 2022 version of Ingenuity Pathway Analysis (IPA) software (QIAGEN Inc., Hilden, Germany), accessed on October 3, 2022. The accession numbers and expression fold changes of differentially regulated proteins were uploaded to the IPA platform for this analysis. The software then categorized the proteins based on biological functions and conducted pathway enrichment analysis. The statistical significance of pathway enrichment was evaluated using Fisher’s exact test, with the p-value of overlap serving as a measure of relevance. This approach enabled the identification and characterization of biological pathways and functions associated with the DEPs within the experimental framework.

Western blot analysis

Four hours after the final ultrasound treatment, mice were euthanized, and hippocampal tissues were collected for protein analysis. Total proteins were extracted using T-PER reagent supplemented with Halt™ Protease Inhibitor Cocktail (Pierce Biotechnology, Rockford, IL, USA). Lysates were centrifuged, and protein concentrations in the supernatants were determined using the Bio-Rad Protein Assay Reagent (Bio-Rad, CA, USA). Equal amounts of protein (30 µg) were separated on 12% SDS–polyacrylamide gels and transferred to Immun-Blot^® PVDF membranes (Bio-Rad, CA, USA). Membranes were blocked for 1 h with blocking buffer (Hycell, Taipei, Taiwan) and incubated overnight at 4 °C with rabbit primary antibodies against REPS1 (CPA5489, Cohesion Biosciences, 1:1000), MYG1 (CQA3944, Cohesion Biosciences, 1:1000), and GAPDH (GTX100118, GeneTex, 1:5000). After washing with PBST, membranes were incubated for 1 h at room temperature with HRP-conjugated secondary antibodies. Protein signals were visualized using Western Lightning ECL Pro reagent (Bio-Rad, CA, USA) and imaged with an ImageQuant™ LAS 4000 system (GE Healthcare Life Sciences, PA, USA). Band intensities were quantified using ImageJ software.

Statistical analysis

For all data, means ± standard deviation (SD) were presented. Statistical analyses were conducted using one-way ANOVA, followed by Tukey’s post-hoc test. Bonferroni correction was applied for multiple comparisons to establish significant differences among various groups, with a significance threshold set at p < 0.05. Proteomics data were analyzed using IBM SPSS 22 statistical software and MedCalc software (Version 20.211). The statistical significance for the proteomics data was determined through a one-way analysis of variance (ANOVA), followed by the Duncan post hoc multiple range test. A significance level of p < 0.05 was set to identify significant differences among groups in the proteomic analyses.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

Feng-Yi Yang, Meng-Ting Wu, Yi-Ju Pan, Wei-Shen Su, Zih-Yun Pan, Yi-Tang Lin, Chung-Fu Sun, and Yu-Chen Lin conducted experiments and analyzed data. Feng-Yi Yang conceived the study. All authors were involved in the writing and reviewing of the manuscript. The authors read and approved the final manuscript.

Funding

This study was supported by grants from the National Science and Technology Council of Taiwan (no. NSTC 114-2314-B-A49-045- and NSTC 114-2221-E-A49-073-MY3), the Cheng Hsin General Hospital Foundation (no. CY11409 and CY11322), and the Far Eastern Memorial Hospital National Yang Ming Chiao Tung University Joint Research Program (no. 114DN20 and 113DN22).

Data availability

No datasets were generated or analysed during the current study. Data are available from the authors upon request.

Declarations

Ethics and consent to participate

Not applicable.

Consent to publish

Not applicable.

Competing interests

The authors declare no competing interests.