Multi-omics reveals that dietary arachidonic acid reshapes hepatopancreatic lipid remodeling and modulates steroidogenesis-related pathways in mud crab

Abstract

Arachidonic acid (ARA) is a bioactive polyunsaturated fatty acid involved in lipid metabolism, antioxidant defense, and steroidogenesis in aquatic animals. Previous studies have demonstrated that dietary ARA regulates lipid metabolism and steroid biosynthesis in the mud crab Scylla paramamosain. It was hypothesized that dietary ARA promotes steroidogenesis by reshaping hepatopancreatic lipid composition. To test this hypothesis, the parallel transcriptomic and lipidomic analyses with pathway-level comparative interpretation were performed to characterize the hepatopancreatic responses of mud crabs fed experimental diets containing different concentrations of ARA for 22 weeks. Transcriptomic analysis revealed that an appropriate dietary ARA concentration significantly regulated lipid metabolic pathways, as indicated by the down-regulation of acs, enhanced antioxidant capacity through the up-regulation of sod, sod2, and sod3. Transcriptomic analysis indicated molecular responses associated with ovarian steroidogenesis-related pathways, including the up-regulation of pla2 and ARTISt. Lipidomic analysis further demonstrated that dietary ARA markedly remodeled hepatopancreatic lipid composition, particularly glycerophospholipids and sterol-related lipids. In conclusion, these findings supported our hypothesis that dietary ARA regulated steroidogenesis by reshaping hepatopancreatic lipid composition. This study provides mechanistic insights that may facilitate the development of nutritionally optimized broodstock diets to improve reproductive performance in mud crab.

Highlights

• •This study provides the first integrated transcriptomic–lipidomic analysis of dietary ARA regulation in mud crab.
• •Dietary ARA reshaped hepatopancreatic lipid composition and fatty acid positional distribution in TG, PC, and PE.
• •ARA modulated lipid transport and metabolism through PPAR signaling–associated pathways.
• •Enhanced antioxidant capacity was associated with longevity-regulating signaling pathways under optimal ARA levels.
• •ARA-induced remodeling of cholesterol esters and glycerolipids was linked to ovarian steroidogenesis.

1. Introduction

Arachidonic acid (ARA; 20:4n-6) is a biologically active long-chain polyunsaturated fatty acid that participates in lipid metabolism, immune regulation, and reproductive processes in aquatic animals. Although ARA is generally present at lower concentrations than n-3 long-chain polyunsaturated fatty acids in aquatic animal tissues, it serves as a precursor of eicosanoids and other signaling molecules involved in physiological regulation (Grosser et al., 2002; Ljubobratović et al., 2020; Patiño et al., 2003). In aquatic species with limited capacity for endogenous ARA biosynthesis, dietary ARA is considered an essential fatty acid (Trushenski & Rombenso, 2020). Our previous studies demonstrated that appropriate dietary ARA supplementation enhanced growth performance, antioxidant capacity, and health status in juvenile mud crabs, whereas excessive ARA intake suppressed lipogenesis (Fang et al., 2022). Furthermore, optimal dietary Dietary ARA supplementation promoted ARA deposition and metabolism, elevated sex steroid hormone concentrations, and accelerated ovarian development in female mud crabs (Fang et al., 2023). However, how dietary ARA links hepatopancreatic lipid metabolism with steroidogenesis remains unclear.

Recent advances in multi-omics technologies have provided powerful approaches for investigating nutrient-regulated metabolic responses in aquatic animals. Parallel transcriptomic and metabolomic analyses have been applied to characterize physiological regulation under environmental stress in the Chinese mitten crab (Eriocheir sinensis) (Mo et al., 2022). However, multi-omics investigations focusing on dietary ARA responses in aquatic animals remain limited. To date, only a few studies have combined transcriptomic and lipidomic approaches to examine ARA-associated metabolic regulation, such as in female Chinese sturgeon (Acipenser sinensis) (Wu et al., 2022). Compared with transcriptomics or lipidomics alone, their integration enables the identification of metabolic pathways supported by both gene expression changes and lipid alterations, thereby providing a more comprehensive view of nutrient-induced metabolic remodeling. Nevertheless, these associations should be interpreted cautiously because transcriptomic and lipidomic changes do not necessarily indicate direct causality, and regulatory layers such as post-translational modification and enzymatic activity are not fully captured. Therefore, parallel multi-omics approaches are particularly valuable for generating testable hypotheses regarding nutrient regulation but require complementary functional validation.

The mud crab (Scylla paramamosain) is an economically important mariculture species in China owing to its rapid growth, short reproductive cycle, environmental adaptability, and high nutritional value (Wang et al., 2022). However, the development of nutritionally optimized formulated diets for mud crab remains constrained by limited knowledge of precise nutrient requirements, particularly regarding lipid nutrition (Dong et al., 2017). Although previous studies have demonstrated the physiological importance of dietary ARA in crustaceans, the molecular mechanisms linking dietary ARA to hepatopancreatic lipid remodeling and steroidogenesis remain poorly understood. In particular, transcriptomic and lipidomic evidence explaining these responses is still lacking. Therefore, we hypothesized that dietary ARA modulates hepatopancreatic lipid remodeling associated with ovarian steroidogenesis-related pathways. Specifically, whether hepatopancreatic lipid remodeling mediates the relationship between dietary ARA and steroidogenesis has not been determined. To address this knowledge gap, the primary outcome of this study was to determine how dietary ARA modulates hepatopancreatic transcriptomic and lipidomic profiles, particularly pathways associated with lipid metabolism and steroid metabolism. Secondary outcomes included the validation of representative differentially expressed genes (DEGs) and characterization of lipid class remodeling. This study aimed to investigate the transcriptomic and lipidomic responses associated with dietary ARA supplementation in the hepatopancreas of mud crab, with particular emphasis on lipid metabolism and ovarian steroidogenesis-related pathways. This study provides new insights into how dietary ARA modulates hepatopancreatic lipid metabolism and steroid metabolism-associated transcriptional responses in mud crab.

2. Materials and methods

2.1. Ethics statement

All experimental procedures were approved by the Institutional Animal Care and Use Committee of Ningbo University (Protocol No. 139111111; SYXK20190005), and were conducted in accordance with the Guide for the Care and Use of Laboratory Animals and relevant institutional guidelines. The investigators were not blinded to dietary treatments during animal husbandry and sample collection due to the nature of the dietary intervention experiment. However, blinding was applied during subsequent laboratory analyses and data processing. Specifically, personnel involved in LC-MS/MS data acquisition, lipid identification and signal intensity, transcriptomic data analysis, and initial statistical processing were unaware of sample group assignments until the completion of the corresponding analytical workflows. No protocol deviations occurred during the experiment. All procedures, including dietary administration, animal management (mortality and molting), sample collection, and analytical procedures, were performed according to the approved experimental protocol.

2.2. Diet preparation, feeding trial and sampling

Based on our previous nutritional studies investigating the physiological responses of mud crab to dietary ARA supplementation, six isonitrogenous and isolipidic experimental diets were formulated with graded ARA concentrations (0%, 0.25%, 0.50%, 1.00%, 1.50%, and 2.00%) (Table S1) (Fang et al., 2022, Fang et al., 2023). In the following text, we will use H1, H2, H3, H4, H5 and H6 to represent these six groups. The analyzed ARA contents of the six diets were 0.04% (H1), 0.25% (H2), 0.46% (H3), 0.89% (H4), 1.25% (H5), and 1.57% (H6), respectively. Notably, the omics results focus mostly on H1, H3 and H6. These three groups were selected because they represent the key nutritional gradients of interest (lowest, intermediate, and highest concentrations) based on our previous studies (Fang et al., 2022, Fang et al., 2023). The H1, H3, and H6 groups were selected a priori based on the study hypothesis that biologically meaningful transcriptomic and lipidomic responses would be most apparent across the dietary ARA gradient. H1 and H6 represented the lowest and highest dietary ARA concentrations, respectively, whereas H3 represented the intermediate (approximately optimal) dietary ARA concentration identified in the feeding trial. These three groups were therefore chosen to capture low, optimal, and high dietary ARA conditions while maintaining a feasible sample size for the resource-intensive transcriptomic and lipidomic analyses. These dietary concentrations were selected to cover a broad biological range, including low, optimal, and high ARA supplementation concentrations, enabling evaluation of the effects of dietary ARA on lipid metabolism and related physiological processes. An ARA-enriched oil (Changsha Kenan Biotechnology Co., Ltd., Changsha, China) was incorporated into the basal diet to achieve the target ARA concentrations. The fatty acid compositions of the experimental diets are shown in Table S2. Diet preparation was performed according to our previous study with minor modifications (Fang et al., 2023).

The feeding trial was conducted for 22 weeks at the pilot-scale aquaculture facility of Ningbo University (Meishan Campus) using an intelligent recirculating aquaculture system (Shanghai Haisheng Biotech Co., Ltd., Shanghai, China). A total of 144 juvenile crabs were maintained in 36 compartmentalized 100-L glass tanks connected to a common recirculating aquaculture system. Each glass tank was divided into four separate compartments using 5-mm perforated partitions, with one crab housed in each compartment to prevent cannibalism and behavioral interference. After a 7-day acclimation period, apparently healthy intermolt female crabs of similar body weight (approximately 10 g) were selected. Individuals showing signs of disease, recent molting, limb loss, or abnormal behavior during acclimation were excluded before experimental allocation. The remaining crabs were randomly allocated to six dietary treatments using a computer-generated random number sequence. Each dietary treatment consisted of three biological replicates, with each replicate comprising two glass tanks (eight crabs per replicate). All glass tanks shared the same recirculating water system to maintain stable environmental conditions throughout the experiment. The experiment comprised six dietary treatments, including the control diet, with three biological replicates assigned to each treatment. Each biological replicate consisted of two 100-L tanks, and each tank contained eight crabs. Thus, each dietary treatment included six tanks in total, corresponding to three replicate tank-pairs. The two tanks within each replicate tank-pair received the same dietary treatment and were considered as one biological replicate. For each tank, the measurements of the eight crabs were first averaged to obtain a tank-level value. The two tank-level values within each replicate tank-pair were subsequently averaged to generate one replicate-level mean for statistical analysis. Accordingly, three replicate-level means (n = 3) were used for each dietary treatment, and individual crabs or individual tanks were not treated as independent observations. Dietary treatments were assigned to replicate tank-pairs rather than to individual tanks, such that both tanks within a replicate tank-pair received the same diet. All experimental tanks were connected to a common recirculating aquaculture system (RAS). Therefore, the biological replicates were not maintained in completely independent water-circulation systems, and potential exchange of waterborne factors among tanks could not be completely excluded. This feature of the experimental design was taken into consideration when interpreting the experimental results. Three biological replicates were included for each dietary treatment according to the predefined experimental design.

During the experimental period, crabs were fed once daily at 18:00. Three biological replicates were used for each dietary treatment. The number of replicates was practically determined for this exploratory multi-omics study based on the predefined experimental design and logistical and resource constraints. No formal a priori power analysis was performed to determine the sample size. Experimental diets were provided to all crabs within each replicate according to the assigned treatment. Uneaten feed and feces were removed daily, and approximately one-third of the seawater was renewed every morning to maintain water quality. The feeding ration was adjusted daily according to actual feed consumption. Molting events, molting time, and mortality were recorded throughout the experiment. Water quality parameters were maintained as stable as possible, with salinity at 23.0 ppt, temperature at 26.0 °C, pH at 7.0, dissolved oxygen at approximately 7.0 mg/L, and ammonia nitrogen below 0.05 mg/L. These parameters were monitored using commercial test kits (Huzhou Haqi Environmental Protection Technology Co., Ltd., Huzhou, China).

At the end of the 22-week feeding trial, each individual crab was housed in an independent tank; therefore, individual crabs were considered independent experimental units for growth and survival analyses. For molecular analyses, three individuals per dietary treatment were randomly selected as independent biological replicates for tissue collection. This sample size is consistent with or exceeds that reported in comparable omics studies in decapod crustaceans, where n = 3 per group is common. We acknowledge the limited statistical power arising from this small sample size. The hepatopancreas was immediately dissected after sampling. This tissue was selected because it functions as the major metabolic organ in crustaceans and is considered functionally analogous to the vertebrate liver and adipose tissue (Štrus et al., 2019).

For transcriptomic analysis, hepatopancreas samples were preserved in RNA stabilization solution (Takara Bio, Kusatsu, Japan) at 4 °C overnight and subsequently stored at −80 °C until eukaryotic reference transcriptome sequencing and RNA-seq validation. Parallel hepatopancreas samples from the same individuals were collected, immediately frozen in liquid nitrogen, and stored at −80 °C for relative quantitative lipidomic analysis. All samples were coded according to dietary treatment and biological replicate to ensure accurate sample identification during subsequent analyses.

2.3. Transcriptome analysis

2.3.1. RNA extraction

Three independent biological replicates per dietary group were used for transcriptomic analysis. Each RNA sample was sequenced once, and no technical sequencing replicates were performed. Total RNA was extracted from hepatopancreas tissues using TRIzol reagent (Thermo Fisher Scientific, Waltham, MA, USA) according to the manufacturer's instructions. Briefly, approximately 50.0 mg of frozen hepatopancreas tissue was homogenized in a 2.0 mL centrifuge tube containing stainless steel beads using a mechanical tissue grinder (Ningbo Scientz Biotechnology Co., Ltd., Ningbo, China) (55 Hz, 30 s, twice). Subsequently, 1.0 mL of TRIzol reagent was added, and samples were incubated at room temperature for 10 min after thorough mixing. Chloroform (200 μL) was then added, followed by vigorous shaking and incubation for 2–3 min at room temperature. After centrifugation at 13,700 ×g for 10 min at 4 °C, the aqueous phase was carefully transferred to a new tube. RNA was precipitated using an equal volume of isopropanol and incubated at −20 °C for 10 min. The RNA pellet was washed twice with 75% ethanol, air-dried briefly, and dissolved in DEPC-treated water. RNA concentration and purity were determined using a NanoDrop 2000 spectrophotometer based on A260/A280 and A260/A230 ratios. RNA integrity was further evaluated using an Agilent 4150 Bioanalyzer (Agilent Technologies, Santa Clara, CA, USA). Qualified RNA samples were stored at −80 °C until library preparation and transcriptomic sequencing.

2.3.2. Library construction and sequencing

RNA samples with acceptable quality based on concentration, purity, and integrity assessment were selected for transcriptome library construction. Each library was prepared from an independent biological replicate, and three libraries were constructed for each dietary treatment. Paired-end mRNA libraries were generated using the ABclonal mRNA-seq Library Preparation Kit (ABclonal Technology, Wuhan, China) according to the manufacturer's instructions. Briefly, mRNA was enriched from total RNA using oligo (dT) magnetic beads, followed by fragmentation, first- and second-strand cDNA synthesis, end repair, adaptor ligation, and PCR amplification. The amplified libraries were purified, and library quality was evaluated using an Agilent 4150 Bioanalyzer (Agilent Technologies, Santa Clara, CA, USA). Qualified libraries were sequenced on the MGISEQ-T7 platform (MGI Tech Co., Ltd., Shenzhen, China) to generate 150 bp paired-end reads (PE150).

2.3.3. Library quality assessment and sequencing

Prior to sequencing, the quality of each transcriptome library was evaluated based on insert size distribution and effective library concentration. Only qualified libraries were selected for sequencing. Libraries derived from independent biological replicates were pooled according to their effective concentrations and the expected sequencing depth. High-throughput sequencing was subsequently performed on the MGISEQ-T7 platform to generate PE150. The raw sequencing data were further subjected to quality filtering before downstream transcriptomic analysis.

2.3.4. Bioinformatics analysis

All bioinformatics analyses were performed by Shanghai Applied Protein Technology Co., Ltd. (Shanghai, China). Raw sequencing data generated from the MGISEQ-T7 platform were subjected to quality control before downstream analyses. Raw reads in FASTQ format were processed using custom Perl scripts to remove adaptor sequences, low-quality reads, and reads containing more than 5% ambiguous nucleotides (N), resulting in high-quality clean reads.

Clean reads were aligned to the S. paramamosain reference genome using HISAT2 v2.2.1. Gene-level read counts were generated using FeatureCounts v2.0.3. Differential expression analysis was performed using DESeq2 v1.40.2 based on raw read counts, whereas gene expression levels were normalized as fragments per kilobase of transcript per million mapped reads (FPKM) for expression visualization and comparison.

DEGs were identified using the Benjamini–Hochberg procedure to control the false discovery rate (FDR). Genes with an adjusted P value (FDR) < 0.05 and an absolute log₂ fold change (|log₂FC|) > 1 were considered significantly differentially expressed. These criteria were applied consistently throughout the transcriptomic analysis. Functional annotation and enrichment analyses of Gene Ontology (GO) terms and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways were performed using the clusterProfiler v4.6.0 package in R. Pathway enrichment results were first filtered by FDR-adjusted p < 0.05. From this filtered list, pathways were prioritized for discussion only if they met all of the following criteria: the pathway contained at least three annotated genes from the input list; the pathway was annotated in the relevant Arthropoda or crustacean KEGG reference database, where available, or had clear ortholog-based support for at least half of the mapped genes; the pathway was biologically plausible in the context of the tissue and species studied, based on existing literature for crustacean or invertebrate systems.

2.3.5. Real-time quantitative PCR (RT-qPCR) validation of DEGs

To validate the reliability of the transcriptomic data, six representative DEGs involved in lipid metabolism, antioxidant defense, and steroidogenesis were selected for RT-qPCR analysis. RT-qPCR was used as an orthogonal analytical method to measure the expression of selected transcripts in the same biological RNA samples, with the explicit acknowledgment that this does not constitute independent biological replication. Gene-specific primers are listed in Table S3. β-actin, which exhibited stable expression across all dietary treatments, was used as the internal reference gene, and the H1 group served as the calibrator. It was used as the reference gene based on its common use in crustacean gene expression studies. RT-qPCR was performed according to the method described by Wang et al. (2022). RT-qPCR data were analyzed using the 2^-ΔΔCt method. The stability of β-actin expression among dietary treatments was evaluated by comparing its Ct values across groups, and no significant differences were detected (P > 0.05), supporting its suitability as the reference gene. Relative expression levels were compared between treatment groups using Student's t-test / one-way ANOVA with post-hoc test, as appropriate, with P < 0.05 considered statistically significant. All RT-qPCR reactions were performed in technical triplicate, and data are presented as mean ± SD from three biological replicates. Gene-specific primers were designed using Primer Premier 5.0 (PREMIER Biosoft International, Palo Alto, CA, USA) based on the corresponding S. paramamosain cDNA sequences available in the NCBI database. Primer amplification efficiencies were determined using six serial dilutions of cDNA, and all primer pairs exhibited amplification efficiencies ranging from 93% to 102%. PCR amplification was carried out on a Roche real-time PCR system (Roche Diagnostics, Basel, Switzerland) using a 20.0 μL reaction mixture containing 2.0 μL cDNA template, 1.0 μL of each primer, 10.0 μL 2× SYBR Green I Master Mix (Vazyme Biotech, Nanjing, China), and nuclease-free water. The amplification program consisted of an initial denaturation at 95 °C for 2 min, followed by 45 cycles of 95 °C for 10 s, 58 °C for 10 s, and 72 °C for 20 s.

2.4. Lipidomics analysis

2.4.1. Sample pretreatment and lipid extraction

Equal aliquots of hepatopancreas extracts from all samples were pooled to generate quality control (QC) samples. QC samples were injected at regular intervals (every three samples) throughout the LC-MS/MS sequence to monitor instrumental stability and injection reproducibility. Because the pooled QC samples were prepared from already extracted samples, they primarily assessed instrumental and injection repeatability rather than extraction-to-extraction variability. QC-based quality assessment, including principal component analysis (PCA) and coefficient of variation (CV) analysis, was performed during data processing. Lipid features with a CV >30% among QC samples were excluded from subsequent analyses. Missing values in the non-targeted lipidomics dataset were imputed using the K-nearest-neighbor (KNN) method.

Total lipids were extracted from hepatopancreas tissues using the methyl tert-butyl ether (MTBE) method according to Giavalisco et al. (2011), with minor modifications. Briefly, hepatopancreas samples were thawed at 4 °C, and approximately 30.0 mg of tissue was homogenized with 200.0 μL ultrapure water. Subsequently, 240.0 μL pre-cooled methanol (HPLC grade; Aladdin Biochemical Technology Co., Ltd., Shanghai, China) and 800.0 μL MTBE (HPLC grade; Sigma-Aldrich, St. Louis, MO, USA) were added. Samples were vortexed thoroughly, sonicated for 20 min, and incubated at room temperature for 30 min. After centrifugation at 14,000 ×g for 15 min at 10 °C, the upper organic phase was collected and evaporated to dryness under a gentle stream of nitrogen. The dried lipid extracts were reconstituted in 200.0 μL isopropanol/acetonitrile solution (9:1, v/v; LC-MS grade; Thermo Fisher Scientific, Waltham, MA, USA), followed by centrifugation at 14,000 ×g for 15 min at 10 °C. The resulting supernatants were transferred for LC-MS/MS analysis.

2.4.2. LC-MS analysis

Three biological replicates from each dietary treatment were subjected to LC-MS/MS analysis. Each biological sample was analyzed once, while pooled quality control (QC) samples were repeatedly injected at regular intervals throughout the analytical sequence to monitor technical reproducibility and instrument stability. Lipid separation was performed using a Nexera LC-30 A ultra-performance liquid chromatography (UPLC) system (Shimadzu Corporation, Kyoto, Japan) equipped with a Waters ACQUITY UPLC CSH C18 column (Waters Corporation, Milford, MA, USA). The column temperature was maintained at 45 °C, and the autosampler temperature was set at 10 °C. The injection volume was 3.0 μL, with a flow rate of 300 μL/min. The mobile phase consisted of solvent A (10 mM ammonium formate in acetonitrile/water, v/v) and solvent B (10 mM ammonium formate in acetonitrile/isopropanol, v/v). Gradient elution was performed as follows: 30% solvent B from 0 to 2 min, increased linearly from 30% to 100% solvent B over 2–25 min, and maintained at 100% solvent B from 25 to 35 min for column washing.

Mass spectrometric analysis was conducted using a Q Exactive series mass spectrometer (Thermo Fisher Scientific, Waltham, MA, USA) equipped with an electrospray ionization (ESI) source operated in both positive and negative ion modes. The ESI parameters were set as follows: sheath gas flow rate, 45 arb; auxiliary gas flow rate, 15 arb; heater temperature, 300 °C; capillary temperature, 350 °C; spray voltage, 3 kV; nebulizer gas flow rate, 1 arb; and S-Lens RF level, 50%. Full-scan MS data were acquired over an m/z range of 200–1800, with mass resolutions of 70,000 for MS1 and 17,500 for MS2 at m/z 200.

2.4.3. Data processing and statistical analysis

The biological replicate (consisting of two glass tanks and eight crabs) was considered the experimental unit for statistical analyses. Mean values of the eight crabs within each replicate were used for statistical analyses (n = 3 per dietary treatment). Integration of transcriptomic and lipidomic datasets was performed at the pathway level. DEGs and differentially abundant molecular species were independently identified and subsequently interpreted together by mapping them onto common KEGG pathways related to lipid metabolism and steroidogenesis-associated processes. The raw RNA sequencing data generated in this study have been deposited in the NCBI Sequence Read Archive (SRA) with accession number PRJCA070490. Other datasets generated or analyzed during the current study are available from the corresponding author upon reasonable request. Raw LC-MS/MS data were processed using LipidSearch 5.0 (Thermo Fisher Scientific, Waltham, MA, USA) for peak detection, alignment, and lipid identification. Molecular species were annotated based on accurate mass and MS/MS fragmentation patterns using the LipidSearch database and LIPID MAPS database. During data processing, a retention time tolerance of 0.1 min was applied for peak alignment, with mass tolerances of ±5 ppm for MS¹ and ± 10 ppm for MS². Adducts considered included [M + H]⁺, [M + NH₄]⁺, and [M + Na]⁺ in positive ion mode, and [M-H]⁻, [M + HCOO]⁻, and [M + CH₃COO]⁻ in negative ion mode. LipidSearch automatically groups multiple adducts (e.g., [M + H]⁺, [M + NH₄]⁺, [M + Na]⁺) of the same molecular species into a single identification. We used the software's adduct grouping function and confirmed that each reported lipid feature represents a unique molecular species, not multiple adducts of the same species. We monitored for common in-source fragments (e.g., loss of water, loss of headgroup for phospholipids) and confirmed that LipidSearch accounts for these. Any potential redundant features arising from in-source fragmentation were manually reviewed and removed if identified. The software's internal grading (A–D) based on the number and intensity of matched fragment ions relative to the base peak, and only lipids with Grade A, B, or C were retained for quantitative analysis. Grade D identifications were excluded. In detail, Grade A: All fatty acyl chains and the class-specific head group fully identified with high-confidence fragment ion matching; Grade B: All fatty acyl chains and the class-specific head group identified but with lower fragment ion matching scores; Grade C: Partial identification; class and some but not all acyl chains resolved; Grade D: Low-confidence identifications, excluded from the dataset. Molecular species were confirmed by matching to authentic standards.

The vast majority of lipid features were quantified based on relative signal intensity (peak area) without external calibration curves or absolute quantification. No internal standards were used for signal normalization, and peak areas were not adjusted according to exact tissue mass. No QC-based signal-drift correction was applied to the peak-area data. Unless otherwise stated, lipid abundances are therefore reported as peak areas reflecting relative signal intensity rather than absolute molar concentrations. Accordingly, the terms “relative abundance” and “peak area” are used throughout the manuscript, whereas terminology implying absolute quantification, such as “concentration” or “pmol/mg,” is avoided. Where applicable, Pareto scaling was used as a variable-scaling procedure for subsequent multivariate analysis and was not considered a signal-normalization method.

We have adopted the Metabolomics Standards Initiative (MSI) confidence levels to classify all reported lipid identifications. Level 1 (Confirmed identification): Lipid identified by matching both MS/MS spectrum and retention time to an authentic standard analyzed under identical conditions; Level 2 (Putative annotation): Lipid identified by MS/MS spectral matching to a spectral library (LipidSearch database) without authentic standard confirmation. The majority of our reported lipids fall into this category; Level 3 (Putative compound class): Lipid assigned to a class or subclass based on characteristic fragment ions or neutral losses, without full acyl-chain resolution; Level 4 (Unknown): Features with detectable signals but no structural assignment. These were excluded from differential analysis. Notably, the analytical approach used in the present study did not establish the stereospecific positions of fatty acyl chains on the glycerol backbone; therefore, all descriptions of fatty acid composition refer only to the putative acyl-chain composition and do not imply positional assignment.

Before statistical analysis, lipid abundance data were preprocessed using Pareto scaling. Differential lipid abundance analysis was performed using MetaboAnalyst 5.0. Multivariate analyses, including principal component analysis (PCA). For univariate analysis, Student's t-test and fold-change analysis were performed, and multiple testing correction was conducted using the Benjamini–Hochberg false discovery rate (FDR) method. Lipids with variable importance in projection (VIP) >1 and FDR-adjusted P < 0.05 were considered significantly altered. No batch correction was applied because all samples were analyzed within the same analytical batch. Hierarchical clustering analysis, heat maps, and volcano plots were generated using Cloud Platform 3.0 (Shanghai Applied Protein Technology Co., Ltd., China). Conventional statistical analyses and graphical visualization were performed using GraphPad Prism version 10.0 (GraphPad Software, San Diego, CA, USA), unless otherwise specified.

3. Results

3.1. Transcriptomic profiling of hepatopancreas

3.1.1. Overview of RNA sequencing data

High-throughput transcriptome sequencing was performed to investigate the hepatopancreatic responses of mud crabs fed diets containing different concentrations of ARA. Raw sequencing signals were converted into sequenced reads and base-called using the MGI base-calling software.

3.1.2. Quality assessment of sequencing data

Sequencing quality was evaluated based on Q-phred scores and base error rates, with the correspondence between Q-phred values and error probabilities shown in Table S4. Detailed quality metrics for each sample are summarized in Table S5. The sequencing error rate distribution is presented in Fig. S1. GC content analysis was conducted to assess base composition bias, and the distributions of A, T, G, and C bases are shown in Fig. S2. After removing adapter sequences, low-quality reads, and reads containing more than 5% ambiguous bases (N), high-quality clean reads were obtained for subsequent analyses. The composition of raw and clean reads is illustrated in Fig. S3.

3.1.3. Mapping to the reference genome

Clean reads were aligned to the reference genome of S. paramamosain by HISAT2, which is based on an improved Burrows-Wheeler Transform algorithm. The alignment statistics are summarized in Table S6, and the mapping distribution of reads across the reference genome is shown in Fig. S4, indicating a high mapping efficiency suitable for downstream analyses.

3.1.4. Gene expression profiling

3.1.4.1. Quantification of gene expression levels

Gene expression levels were quantified by featureCounts tool and normalized as fragments per kilobase of transcript per million mapped reads (FPKM). The distribution of genes across different expression level intervals is shown in Fig. S5 and summarized in Table S7.

3.1.4.2. Global distribution of gene expression

Boxplots of FPKM values were generated to visualize the overall expression patterns and dispersion among samples (Fig. S6). The results indicated comparable expression distributions across all samples, suggesting good consistency among biological replicates.

3.1.4.3. Assessment of sample reproducibility

To evaluate sample reproducibility, correlation analysis and PCA were conducted. The correlation heatmap (Fig. S7) demonstrated strong correlations among the three biological replicates within each group. PCA further revealed clear clustering patterns among the H1, H3, and H6 groups (Fig. S8), with H1 and H6 exhibiting relatively similar global expression profiles.

3.1.4.4. DEGs analysis

Six representative DEGs annotated to the KEGG lipid metabolism, antioxidant defense, and ovarian steroidogenesis pathways were selected for qPCR analysis to examine whether their expression patterns were consistent with RNA-seq results. Differential expression analysis was performed using the DESeq2 package, with thresholds set at |log₂(fold change)| > 1 and adjusted P-value (padj) < 0.05. Comparative analyses identified 1883 DEGs in H3 vs H1, including 1220 upregulated and 663 downregulated genes. In H6 vs H1, 1414 DEGs were detected (793 upregulated and 621 downregulated), while 1612 DEGs were identified in H6 vs H3, comprising 570 upregulated and 1042 downregulated genes (Fig. S9).

Hierarchical clustering heatmaps of H3 vs H1 (Fig. 1 A) and H6 vs H1 (Fig. 1 B) revealed distinct transcriptional differences among groups. Volcano plots of DEGs in H3 vs H1 (Fig. 2 A) and H6 vs H1 (Fig. 2 B) illustrated the overall distribution of DEGs. Additionally, Venn analysis (Fig. S10) identified 44 common DEGs shared among the three pairwise comparisons, suggesting their potential involvement in ARA-mediated regulatory processes.

Fig. 1.

Hierarchical clustering heatmaps of differentially expressed genes in H3 vs H1 (A) and H6 vs H1 (B). Red and purple indicate higher and lower relative expression levels, respectively. Rows represent individual samples, and columns represent genes. The dendrograms indicate hierarchical clustering of samples and genes based on their expression profiles, with shorter branch distances representing greater similarity. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Fig. 2.

Volcano plots of differentially expressed genes in H3 vs H1 (A) and H6 vs H1 (B). Each dot represents an individual gene. Gray dots indicate genes without significant differential expression, whereas red and blue dots indicate significantly up-regulated and down-regulated genes, respectively. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

3.1.4.5. Functional annotation and pathway enrichment analysis

KEGG pathway enrichment analysis was conducted using Fisher's exact test to identify biological pathways significantly affected by dietary Dietary ARA supplementation. It was showed that the up-regulated DEGs obtained in H3 vs H1 comparison group were mainly mapped into caffeine metabolism, longevity regulating pathway, phagosome, etc. (Fig. 3 A). However, there were some enriched among downregulated genes pathways in this comparison group, including longevity regulating pathway, selenocompound metabolism, butanoate metabolism, etc. (Fig. 3 C). Additionally, DEGs in H6 vs H1 comparison group were mapped into some significantly enriched among upregulated genes pathways, mainly including Toll and Imd signaling pathway, MAPK signaling pathway, hematopoietic cell lineage, etc. (Fig. 3 B). ‌While, this comparison group had some significantly enriched among downregulated genes pathways, mainly including pentose phosphate pathway, glycosaminoglycan degradation, thermogenesis, etc. (Fig. 3 D).

Fig. 3.

KEGG pathway enrichment bubble plots for H3 vs H1 (A, C) and H6 vs H1 (B, D). Each bubble represents an enriched KEGG pathway. Bubble size indicates the number of differentially expressed genes mapped to the pathway, and the colour gradient represents the significance of pathway enrichment based on the P value.

To further elucidate key regulatory genes, DEGs from H3 vs H1 were mapped onto representative KEGG pathways. Notably, the fatty acid transport-related gene acs was significantly enriched among downregulated genes in PPAR signaling pathway (Fig. S11). In longevity regulating pathway-multiple species signaling pathway, antioxidant-related genes sod, sod2, and sod3 were significantly up-regulated (Fig. S12). Additionally, pla2 and artist were markedly up-regulated in the steroidogenesis-related pathways (Fig. S13).

To validate the transcriptomic results, six DEGs (acs, sod, sod2, sod3, pla2, and artist) were selected for RT-qPCR analysis. The expression patterns obtained by RT-qPCR were consistent with the RNA-seq data (Fig. S14), confirming the reliability of transcriptome sequencing results. RT-qPCR results should be interpreted with caution because reference-gene stability was not formally assessed and because the RT-qPCR analysis was performed on the same biological samples used for transcriptomic profiling, rather than on independent biological replicates. We recommend that future studies include a systematic evaluation of candidate reference genes and independent biological replication.

3.2. Lipidomic profiling of hepatopancreas

3.2.1. Data quality assessment

The reliability and stability of lipidomic dataset were evaluated by six QC strategies to assess experimental repeatability, instrument stability, and overall data quality. Base peak chromatograms of QC samples acquired in both positive and negative ionization modes (Fig. S15 and Fig. S16) exhibited highly consistent retention times and peak intensities, indicating good analytical repeatability. Pearson correlation analysis further showed that correlation coefficients among QC samples exceeded 0.9 (Fig. S17), confirming high reproducibility. PCA based on all detected ion features demonstrated tight clustering of samples within each group and clear separation among different dietary treatments (Fig. S18). Hotelling's T² test indicated that all QC samples were within 99% confidence interval (Fig. S19), with no outliers detected. Multivariate control charts revealed that all QC measurements remained within ±3 standard deviations (Fig. S20), suggesting stable instrument performance. Additionally, more than 80% of detected ion peaks exhibited relative standard deviation (RSD) values ≤30% (Fig. S21), confirming the robustness of lipidomic data for downstream analyses.

3.2.2. Lipid subclasses and lipid coverage

A total of 5699 lipid features were annotated as molecular species across 44 lipid subclasses based on MS/MS spectral matching in both positive and negative ionization modes (Table S8; Fig. 4). These annotations represent putatively characterized molecular species based on the LipidSearch database and MS/MS information. The major annotated lipid subclasses included triacylglycerols (TG; 1273 features, 22.3%), phosphatidylcholines (PC; 777 features, 13.6%), phosphatidylethanolamines (PE; 512 features, 9.0%), diacylglycerols (DG; 500 features, 8.8%), Hex1Cer (429 features, 7.5%), and ceramides (Cer; 421 features, 7.4%).

Fig. 4.

Number of annotated lipid subclasses and lipid features identified in the hepatopancreas of mud crab. TG, triglyceride; PC, phosphatidylcholine; PE, phosphatidylethanolamine; DG, diglyceride; Hex1Cer, hexosyl ceramide; Cer, ceramide; PS, phosphatidylserine; CL, cardiolipin; PG, phosphatidylglycerol; PI, phosphatidylinositol; Hex2Cer, dihexosyl ceramide; LPC, lysophosphatidylcholine; SM, sphingomyelin; PIP, phosphatidylinositol-4-phosphate; CerP, ceramide phosphate; Hex3Cer, trihexosyl ceramide; LPE, lysophosphatidylethanolamine; ST, sulfatide (galactosyl ceramide sulfate); AcCa, acylcarnitine; PA, phosphatidic acid; MG, monoglyceride; WE, wax ester; FA, fatty acid; phSM, phytosphingosine-containing sphingomyelin; OAHFA, (O-acyl)-1-hydroxy fatty acid; SPH, sphingosine base; LPS, lysophosphatidylserine; LPI, lysophosphatidylinositol; CerG2GNAc1, N-acetylhexosyl ceramide; ChE, cholesterol ester; LPG, lysophosphatidylglycerol; CerG3GNAc1, dihexosyl N-acetylhexosyl ceramide; SPHP, sphingosine phosphate; GM3, monosialotrihexosyl ceramide; PIP2, phosphatidylinositol-4,5-bisphosphate; LPA, lysophosphatidic acid; PIP3, phosphatidylinositol-3,4,5-trisphosphate; ZyE, zymosterol ester; Co, coenzyme Q; Hex1SPH, hexosyl sphingosine; LSM, lysosphingomyelin; CerG3GNAc2, trihexosyl di-N-acetylhexosyl ceramide; GM1, monosialotetrahexosyl ceramide; StE, stigmasterol ester.

3.2.3. Lipid composition and distribution among dietary groups

FC analysis (Fig. 5) revealed pronounced alterations in lipid composition of the hepatopancreas in response to dietary Dietary ARA supplementation. In H3 vs H1 comparison, the relative abundance of PC, PE, LPC, PIP, PIP2, LPA, ZyE, WE, OAHFA, Hex1Cer, Cer, SM, CerP, phSM, CerG2GNAc1, SPHP, LSM, and coenzyme Q (Co) were significantly reduced in H3. In H6 vs H1 comparison, significantly lower relative abundance of DG, PC, PE, LPC, PIP, LPE, LPS, LPG, LPA, ChE, ZyE, StE, WE, OAHFA, Hex1Cer, Cer, SM, CerP, phSM, SPHP, CerG3GNAc2, and Co were observed in H6.

Fig. 5.

Effects of dietary arachidonic acid (ARA) levels on the relative abundance of lipid classes in the hepatopancreas of mud crabs fed the H1, H3, and H6 diets. (A) Glycerolipids; (B) glycerophospholipids; (C) sterol lipids; (D) fatty acyls; (E) sphingolipids; and (F) prenol lipids.

3.2.4. Univariate analysis of lipid metabolites

Volcano plot-based univariate analysis was initially performed using a nominal P < 0.05 and fold-change (FC) > 1.5 or < 0.67 as preliminary screening criteria. In the H3 vs H1 comparison (Fig. 6A, C), 452 molecular species showed nominally significant abundance changes, including 314 increased and 138 decreased species. In the H6 vs H1 comparison (Fig. 6B, D), 1539 molecular species showed nominally significant changes, including 1034 increased and 505 decreased species. These nominally significant features were subsequently subjected to Benjamini–Hochberg false discovery rate (FDR) correction, after which 81 and 314 molecular species remained significant at FDR-adjusted P < 0.05 in the H3 vs H1 and H6 vs H1 comparisons, respectively. Multivariate analysis was then paralleled with the univariate results, and robust differential lipid candidates were identified using VIP > 1, FDR-adjusted P < 0.05, and the predefined FC thresholds. The resulting differential molecular species were used for downstream pathway enrichment and biological interpretation.

Fig. 6.

Venn diagrams and volcano plots of differential lipid features in the hepatopancreas of mud crabs fed diets containing different levels of ARA. (A, C) H3 vs H1; (B, D) H6 vs H1. In the volcano plots, ▲, ▾ and ● indicate increased, decreased, and non-significantly changed lipid features, respectively. In the univariate statistical analysis, lipid features with FC > 1.5 or FC < 0.67 and FDR-adjusted P < 0.05 were considered significantly different. The labeled lipid molecules represent the top 10 up-regulated and top 10 down-regulated lipid features among all significantly altered features.

3.2.5. Multivariate statistical analysis of lipid metabolites

PCA models constructed for H3 vs H1 and H6 vs H1 explained 62.4% and 74.5% of total variance, respectively (Fig. S22; Table S9), indicating good model stability.

3.2.6. Differentially abundant molecular species

Differentially abundant molecular species were identified using the predefined criteria of FDR-adjusted P < 0.05, VIP > 1, and FC > 1.5 or < 0.67. In the H3 vs H1 comparison (Fig. 7A, C), 81 molecular species met these criteria, including 64 increased and 17 decreased species. In the H6 vs H1 comparison (Fig. 7B, D), 314 molecular species were identified, including 234 increased and 80 decreased species. In H3 vs H1, individual molecular species belonging to the PE, PC, TG, LPC, Hex1Cer, PS, DG, LPE, CL, FA, LPI, LPS, and SM classes showed increased relative abundance in H3, whereas PG species showed higher relative abundance in H1. In H6 vs H1, individual molecular species assigned to TG, PI, PG, DG, PS, Hex1Cer, AcCa, sterol (ST), Cer, CL, FA, and SPH classes exhibited increased relative abundance in H6, whereas LPC and Co species showed higher abundance in H1. It should be noted that the results presented in this section represent changes in individual annotated molecular species rather than total lipid class abundance. Therefore, increases in selected PC or PE species do not necessarily contradict the class-level reductions described in Section 3.2.3, where abundance changes were calculated based on summed signal intensities of all annotated species within each lipid class.

Fig. 7.

Venn diagrams of differential lipid features and their fold-change distributions in the hepatopancreas of mud crabs. (A, C) H3 vs H1; (B, D) H6 vs H1. In panels C and D, ▲, ▾ and ● indicate increased, decreased, and non-significantly changed lipid features, respectively.

3.2.7. Composition of major differential lipid classes

Hierarchical clustering heat maps of TG, PC, and PE species revealed clear separation between dietary groups, with distinct clustering of H1, H3, and H6 samples (Fig. S23). These results indicate that dietary ARA supplementation significantly altered the molecular composition of major lipid classes in the hepatopancreas.

3.3. Comparative interpretation of transcriptomic and lipidomic responses

Transcriptomic analysis identified significant enrichment of pathways associated with lipid metabolism and steroid metabolism-related pathways, whereas lipidomic profiling revealed alterations in glycerophospholipid- and sterol-related molecular species. Although analyzed independently, both datasets consistently highlighted biological processes associated with cholesterol metabolism, arachidonic acid metabolism, and steroidogenesis. These complementary findings suggest that dietary ARA elicited molecular responses at both the transcriptional and lipid metabolic levels. Nevertheless, the present analyses do not establish direct regulatory relationships between DEGs and individual molecular species, and such relationships will require formal multi-omics integration analyses in future studies.

4. Discussion

In the present study, transcriptomic and lipidomic analyses were employed to elucidate the molecular responsess underlying the hepatopancreatic response of S. paramamosain to dietary ARA-associated changes. The results revealed that dietary ARA supplementation was associated with alterations in lipid metabolism, antioxidant capacity, and steroidogenesis through coordinated modulation of gene expression and lipid composition. Transcriptome profiling demonstrated extensive differential gene expression in response to dietary ARA supplementation. In the comparison between H3 and H1, 1883 DEGs were identified in pathways associated with metabolism, genes annotated to the KEGG ovarian steroidogenesis pathway, MAPK signaling, TNF signaling, and longevity regulation. Similar enrichment patterns were reported in female Chinese sturgeon fed ARA-enriched diets, where DEGs were mainly concentrated in cellular processes, developmental regulation, and reproductive pathways (Wu et al., 2022). These findings suggest that ARA-mediated regulation of steroid synthesis is closely linked to enhanced ChE metabolism and transcriptional activation of steroidogenic pathways. Notably, our experiment involved dietary supplementation with ARA, a polyunsaturated fatty acid with well-established roles as a precursor for eicosanoid signaling and as a structural component of membrane phospholipids. We therefore gave particular attention to pathways with known functional connections to fatty acid metabolism, lipid signaling, and membrane remodeling.

Lipid metabolic regulation by dietary ARA was further supported by the down-regulation of acs, a key gene involved in fatty acid activation and lipid metabolism within the PPAR signaling pathway. ACS, together with fatty acid transport proteins (FATP), participates in the regulation of fatty acid uptake, activation, and intracellular lipid handling (Martin et al., 1997). In the present study, the reduced expression of acs may indicate an altered transcriptional regulation of fatty acid metabolic processes in the hepatopancreas. Although decreased acs expression could potentially be associated with reduced fatty acid esterification and intracellular lipid retention, these possibilities require further validation through measurements of ACS activity, fatty acid flux, and lipid metabolic intermediates. Whether this transcriptional downregulation results in reduced ACS enzymatic activity and subsequent changes in fatty acid partitioning requires further experimental validation. This molecular response is consistent with our previous findings that appropriate dietary Dietary ARA supplementation exert lipid-lowering effects in mud crab hepatopancreas (Fang et al., 2022). In parallel, dietary ARA markedly enhanced antioxidant capacity, as evidenced by the significant up-regulation of sod, sod2, and sod3. Superoxide dismutases constitute a crucial antioxidant enzyme family responsible for scavenging superoxide radicals and maintaining cellular redox homeostasis (Zelko et al., 2002). We guess that dietary ARA strengthens the antioxidant defense system of the hepatopancreas, which aligns with previous observations of improved antioxidant status in mud crabs receiving optimal ARA supplementation (Fang et al., 2022).

Steroidogenic regulation emerged as another key response to dietary ARA. The significant up-regulation of pla2 and ARTISt within the genes annotated to the KEGG ovarian steroidogenesis pathway were observed and then we put forward a hypothesis that it indicates enhanced liberation and utilization of ARA for steroid hormone synthesis. ARA is predominantly esterified at the sn-2 position of membrane phospholipids, and cellular stimulation triggers its release via phospholipase A2 (PLA2) (Harizi et al., 2008; Martin et al., 2016). The released free ARA serves as a substrate for eicosanoid biosynthesis through the COX, LOX, and CYP450 pathways. Moreover, ARTISt (ARA-related thioesterase involved in steroidogenesis) has been identified in mammalian steroidogenic tissues as an acyl-CoA thioesterase that can hydrolyze arachidonoyl-CoA to release free arachidonic acid and has been implicated in the regulation of steroidogenesis (Finkielstein et al., 1998). However, its precise biochemical function in S. paramamosain, including whether it catalyzes the hydrolysis of arachidonoyl-CoA to release free ARA or contributes to steroid metabolism, has not been experimentally established and requires dedicated functional investigation.

Non-targeted lipidomics further revealed that dietary ARA significantly altered the lipid composition of the hepatopancreas. TG, PC, PE, and DG constituted the major lipid subclasses, with TG accounting for the largest proportion. TG serves as the primary neutral lipid for energy storage and contains long-chain polyunsaturated fatty acids that influence lipid metabolism, membrane properties, and immune function (Li et al., 2020). PC and PE are major membrane phospholipids that regulate membrane structure, fluidity, and signal transduction (Wellner et al., 2013), while DG functions as both a metabolic intermediate and a bioactive lipid involved in lipid signaling (Miklos et al., 2011). The observed reduction in PC, PE, and DG contents in response to increasing dietary ARA supplementation indicates a substantial remodeling of hepatopancreatic lipid architecture, consistent with previous reports of diet-dependent fatty acid enrichment in crustacean tissues (Araújo et al., 2020; Xu et al., 2022). Both univariate and multivariate lipidomic analyses demonstrated that higher dietary ARA supplementation induced a greater number of DLMs. Alterations in specific molecular species, including TG, PE, cardiolipin (CL), phosphatidylglycerol (PG), lysophosphatidylcholine (LPC), and coenzyme Q-related lipids, further suggest that dietary ARA influences mitochondrial function, membrane dynamics, and lipid-mediated signaling pathways. In particular, CL, a mitochondria-specific phospholipid involved in maintaining mitochondrial structure and regulating apoptosis-related signaling (Gonzalvez & Gottlieb, 2007), was significantly elevated at higher dietary ARA supplementation. Therefore, we put forward a hypothesis that higher dietary ARA supplementation enhanced mitochondrial membrane remodeling. Collectively, the transcriptomic and lipidomic datasets consistently implicated biological processes related to cholesterol metabolism, arachidonic acid metabolism, and steroidogenesis. Although these complementary datasets converged on several biological pathways, the present study did not include formal multi-omics integration analyses, such as gene–lipid correlation, network analysis, or multivariate integration models. Therefore, the proposed relationships between transcriptional responses and lipid remodeling should be regarded as biological associations rather than direct mechanistic evidence. Nevertheless, the concordance of the two independently analyzed datasets provides complementary evidence that dietary ARA influences hepatopancreatic lipid metabolism and supports the hypothesis that lipid remodeling may contribute to steroidogenic processes.

Several limitations should be acknowledged in the present study. First, all experimental tanks shared a common recirculating aquaculture system, which may have reduced the environmental independence among replicate units because waterborne factors could potentially be exchanged among tanks. In addition, only three biological replicates were used per dietary treatment, which may have limited the statistical power to detect small or moderate treatment effects. Therefore, the present multi-omics findings should be considered exploratory and hypothesis-generating and warrant validation in future studies using a larger number of independent biological replicates and independent recirculating systems. Second, although transcriptomic and lipidomic analyses revealed coordinated alterations in lipid metabolism and steroidogenesis-associated pathways following dietary ARA supplementation, direct measurements of steroid hormones, steroid intermediates, and metabolic flux were not performed. Therefore, the potential role of ARA-induced lipid remodeling in regulating steroidogenesis remains a mechanistic hypothesis rather than a demonstrated causal relationship. In addition, the multi-omics interpretation was primarily based on pathway-level integration, and further validation using targeted metabolomics, hormone quantification, enzyme activity assays, and functional manipulation approaches will be required to establish causal mechanisms. Third, the present multi-omics analyses were conducted using representative dietary treatments rather than all six dietary ARA levels. Although physiological parameters exhibited ARA concentration-associated responses across the six dietary treatments, the transcriptomic and lipidomic results should be interpreted as ARA-associated molecular alterations rather than direct evidence of a continuous dose-dependent regulatory pattern. Comprehensive multi-omics profiling across all dietary levels combined with formal trend analyses will be necessary to further characterize the dose-response characteristics of ARA-mediated metabolic regulation. Finally, the current dataset reflects steady-state changes in transcript and lipid abundance and cannot distinguish whether these alterations result from changes in biosynthesis, remodeling, or degradation processes.

The results of present study demonstrate that dietary ARA supplementation modulates hepatopancreatic metabolism through coordinated regulation of cholesterol metabolism, ARA-related pathways, and steroidogenic processes. Several limitations of the present study should be acknowledged when interpreting these findings. The sample size for multi-omics analyses was relatively limited (n = 3 per group), reflecting the practical challenges associated with maintaining individually housed crabs under controlled dietary conditions for a 22-week feeding trial, as well as the cost and analytical capacity constraints of multi-omics profiling. The limited sample size may reduce statistical power for detecting moderate or small effect sizes and increase the possibility of false-negative results, indicating that the identified DEGs and molecular species likely represent the most pronounced molecular responses to dietary ARA. Therefore, these classification results should be interpreted cautiously and require validation in larger, independently replicated datasets. Although physiological parameters exhibited ARA concentration-associated responses across the six dietary treatments, transcriptomic and lipidomic analyses were conducted using only three representative dietary groups. Therefore, the multi-omics findings were interpreted as ARA-associated molecular alterations rather than direct evidence of a continuous dose-dependent regulatory pattern. Future studies integrating multi-omics profiling across all dietary ARA levels and applying formal trend analyses will be necessary to further elucidate the dose-response characteristics of ARA-mediated metabolic regulation. Finally, all animals were sampled at a single time point, preventing evaluation of the temporal dynamics underlying molecular adaptation to dietary ARA supplementation. Despite these limitations, this study provides a valuable exploratory dataset identifying candidate genes and molecular species associated with dietary ARA intake and establishes a foundation for future hypothesis-driven investigations using larger cohorts and comprehensive dose-response experimental designs. Although transcriptomic and lipidomic analyses provided valuable insights into potential metabolic pathways affected by dietary ARA, further studies involving enzyme activity assays, isotope tracing, and targeted metabolite quantification are required to confirm the proposed regulatory mechanis.

5. Conclusion

In the present study, a approach of the parallel transcriptomic and lipidomic analyses with pathway-level comparative interpretation was applied to investigate the regulatory responses of hepatopancreatic metabolism to dietary ARA supplementation in mud crab. The analysis revealed that dietary ARA reshaped hepatopancreatic lipid profiles and modulated the expression of genes involved in lipid metabolism, antioxidant-related responses, and steroid metabolism-associated pathways. In particular, coordinated alterations in ChE profiles and lipid metabolism-related gene expression suggest that dietary ARA may influence lipid remodeling processes potentially linked to steroid metabolism. Several limitations should be acknowledged, as direct measurements of free cholesterol, steroid intermediates, steroid hormones, and relevant enzyme activities were not performed. Therefore, the potential relationship between ARA-induced lipid remodeling and steroid metabolism should be interpreted as a hypothesis requiring further validation rather than a demonstrated mechanism. Overall, this study provides new insights into the molecular responses associated with dietary ARA-mediated lipid remodeling and metabolic regulation in mud crab and highlights potential pathways for future nutritional and functional investigations.

CRediT authorship contribution statement

Fang Fang: Writing – original draft, Investigation, Data curation. Ye Yuan: Writing – review & editing. Min Jin: Funding acquisition. Xuexi Wang: Writing – review & editing. Tingting Zhu: Methodology. Qicun Zhou: Project administration, Conceptualization.

Ethics declaration

Animal subject

This study was conducted in accordance with the ARRIVE (Animal Research: Reporting of In Vivo Experiments) guidelines.

This study was approved by the Institutional Animal Care and Use Committee of Ningbo University (Approval No. SYXK20190005).

Declaration of competing interest

The authors state that there are no known financial interests or personal relationships that could potentially influence the research presented in this paper.

Appendix A. Supplementary data

Supplementary material 1

Supplementary material 2

Data availability

Data will be made available on request.