Authors: Zhiyuan Chen, Qingxuan Li, Ruimin Sheng, Jiacheng Zhang, Jianlin Guo, Peng Tan, Shengjun Bao, Yujie Liu, Youqin Kong, Hongfeng Bai, Zhili Ding
Categories: Original Research Article, Fishmeal replacement, Growth performance, Hermetia illucens, Intestinal health, Macrobrachium rosenbergii, Phosphorus emission
Source: Animal Nutrition
Authors: Zhiyuan Chen, Qingxuan Li, Ruimin Sheng, Jiacheng Zhang, Jianlin Guo, Peng Tan, Shengjun Bao, Yujie Liu, Youqin Kong, Hongfeng Bai, Zhili Ding
In recent years, the use of insect meal in feed has been regarded as an important alternative to fishmeal. Defatted black soldier fly larvae meal (BSFLM), derived from Hermetia illucens, is a promising fishmeal alternative in aquafeeds for various aquatic species. However, its potential as a protein source in diets for the giant freshwater prawn (Macrobrachium rosenbergii) remains underexplored. This study provides a systematic evaluation of BSFLM as a fishmeal substitute in M. rosenbergii diets, assessing growth performance, nutrient digestibility, hepatopancreatic biochemistry, intestinal microbiota composition, and phosphorus discharge. Five isonitrogenous experimental diets were formulated, with BSFLM replacing 0 (control, BSFLM0), 10% (BSFLM10), 20% (BSFLM20), 40% (BSFLM40), or 80% (BSFLM80) of dietary fishmeal. A total of 900 juvenile prawns (initial weight 0.25 ± 0.03 g, 40-d-old) were fed experimental diets. They were reared in 300-L tanks (60 prawns/tank, 3 replicates/group) and fed to satiation twice daily for 8 weeks. Each tank was treated as an independent experimental unit for statistical analysis. Growth performance (survival, weight gain, and feed conversion ratio) did not differ significantly among groups (P > 0.05). BSFLM40 and BSFLM80 groups showed significantly improved phosphorus retention efficiency (14.57% to 19.24%) and deposition rate (13.69% to 18.16%), alongside reduced phosphorus waste (P < 0.05). Intestinal digestive enzyme activities (trypsin, amylase, and lipase) were significantly upregulated at 40% to 80% BSFLM inclusion (P < 0.05). BSFLM-fed prawns exhibited enhanced antioxidant capacity (total superoxide dismutase activity was significantly elevated, P < 0.001) with stable hepatopancreatic malondialdehyde levels (P = 0.343). Transcriptomic analysis revealed differential expression of nutrient metabolism (involving in the insulin signaling pathway and mineral absorption pathway) and immune response genes (associating with phagosome activity and antigen processing/presentation) in BSFLM20 vs. control. BSFLM inclusion altered intestinal BSFLM10 enriched Firmicutes abundance, while BSFLM40 increased Agromyces abundance (P < 0.05). BSFLM-fed prawns displayed enhanced intestinal morphology (villus height and muscularis thickness increased, P < 0.001). In summary, BSFLM can replace up to 80% of fishmeal in M. rosenbergii diets without compromising growth. Moderate replacement (10% to 20%) enhanced antioxidant capacity and optimized microbiota; higher levels (40% to 80%) improved digestive function, reduced phosphorus discharge, and stimulated chitinolytic bacteria, confirming BSFLM as a nutritionally balanced and environmentally sustainable fishmeal alternative in M. rosenbergii aquaculture.
With the relentless expansion of the global population, food security concerns have escalated into a pressing worldwide challenge (Ntiamoah et al., 2023). Within this context, aquaculture has emerged as a pivotal solution for meeting escalating protein demands, undergoing remarkable development in recent decades (Afewerki et al., 2023). Notably, aquaculture now dominates as the principal source of aquatic products for human nutrition (Yue and Shen, 2022). However, intensive aquaculture systems face substantial economic pressures, with feed costs constituting more than half of total production expenditures (Ye et al., 2020), thereby creating unprecedented demand for protein sources. Fishmeal, serving as the primary protein source in aquafeeds (Zhao et al., 2021), confronts production limitations due to global fishery resource constraints, with output having reached a plateau (Gokulakrishnan et al., 2023; Kaiser et al., 2022). This supply-demand imbalance (Liu et al., 2021) has driven international fishmeal prices to historic high, progressively undermining the sustainable development of aquaculture (Majluf et al., 2024).
To mitigate the critical dependency on fishmeal, the scientific community has intensified efforts to develop innovative non-grain protein sources as sustainable alternatives. Traditional grain-based substitutes such as soybean and corn have been widely incorporated into aquafeeds (Hang et al., 2022; Mugwanya et al., 2023). However, these staple crops predominantly serve as fundamental components in human nutrition and livestock production, creating competitive pressure on aquatic feed systems. Against the backdrop of escalating global food security challenges (Naheed, 2023), non-grain protein research has emerged as an innovative developmental pathway in aquaculture nutrition (Chen et al., 2024; Su et al., 2023). Particular noteworthy, insect proteins exhibit exceptional potential as fishmeal alternatives due to their minimal land requirements, rapid growth cycles, high feed conversion efficiency, and reduced environmental impact (Cámara-Ruiz et al., 2023). For instance, Zheng et al. (2023) demonstrated that replacing 30% of fishmeal with yellow mealworm (Tenebrio molitor) improved growth performance in Pacific white shrimp (Litopenaeus vannamei). Notably, the well-balanced amino acid profile of insect proteins demonstrates superior nutritional compatibility with crustacean physiological demands compared to other alternative sources (Köhler et al., 2019).
Among insect proteins, black soldier fly (Hermetia illucens) larvae meal (BSFLM) has garnered substantial research interest as a sustainable fishmeal alternative, distinguished by its high protein content (35% in dry matter, increasing to 60% post-defatting) and nutritionally balanced profile (Gasco et al., 2021; Nogales-Mérida et al., 2019). This saprophagous insect demonstrates remarkable bioconversion efficiency, completing rapid life cycles while valorizing organic wastes including food residues and livestock manure into high-value feed ingredients (Kim et al., 2021). Nutritional analyses reveal BSFLM's amino acid composition parallels fishmeal, while additionally providing beneficial fatty acids, essential minerals, and trace elements that enhances aquatic growth performance (Mousavi et al., 2020). Recent studies have demonstrated successful fishmeal substitution with BSFLM in diverse 200 g/kg inclusion in koi carp (Cyprinus carpio) diets maintained growth while upregulating immune-related genes (Linh et al., 2024), with comparable efficacy observed in climbing perch (Anabas testudineus) (Mapanao et al., 2023), European seabass (Dicentrarchus labrax) (Magalhães et al., 2017), and largemouth bass (Micropterus salmoides) (Fischer et al., 2022). Of particular significance is BSFLM's inherent antimicrobial capacity derived from its detritivorous nature. The insect biomass contains bioactive compounds including antimicrobial peptides, proteases, and lauric acid (a medium-chain saturated fatty acid representing 40% to 55% of its lipid profile) (Choi et al., 2012; Li et al., 2021; Weththasinghe et al., 2022). Lauric acid exhibits dual nutritional-pharmacological its rapid β-oxidation provides immediate energy, while membrane-disrupting properties inhibit pathogenic bacteria and viruses through lipid envelope destabilization (Dayrit, 2015).
A critical consideration in aquaculture nutrition involves the high phosphorus content of fishmeal (2.68%) (Parveen et al., 2000) coupled with the low phosphorus utilization efficiency in shrimp species (Lemos et al., 2021) necessitating careful management of aquaculture effluent. Shrimp farming wastewater, particularly its elevated phosphorus levels, can induce eutrophication and stimulate algal blooms, which rapidly deplete dissolved oxygen and lead to secondary pollution through algal decomposition (Cheng et al., 2018). In contrast, BSFLM exhibits a substantially reduced phosphorus content (0.9%) (Alfiko et al., 2022). Although systematic modeling data remain unavailable, this characteristic suggests potential reductions in nitrogen and phosphorus discharge from aquaculture systems, thereby mitigating environmental impacts.
Macrobrachium rosenbergii, a commercially valuable freshwater prawn species, has demonstrated remarkable production growth in global aquaculture systems, with consecutive record-breaking yields significantly contributing to sectoral expansion (González-Figueroa et al., 2024). Nevertheless, its heavy reliance on fishmeal in formulated feeds (D'Abramo and Sheen, 1994) urgently necessitates the development of low-fishmeal diets to ensure industry sustainability. Recent research advances have advanced on various non-grain protein substitutes, including yeast hydrolysate (Kong et al., 2025), algal-based proteins (Sukri et al., 2016), black soldier fly (Zarantoniello et al., 2023), and agro-industrial byproducts (Boateng et al., 2023). Although a preliminary study has explored the fishmeal substitution with BSFLM on this prawn, its BSFLM origin is full fat and the replacement level is only 3% and 20% (Zarantoniello et al., 2023). Indeed, this current comprehensive, multi-angle study investigating fishmeal replacement with BSFLM in M. rosenbergii nutrition represents a significant scientific advancement. This research provides a full evaluation of defatted BSFLM in giant freshwater prawn diets, integrating transcriptomics, hepatopancreatic-intestinal morphology, gut microbiota analysis, digestibility assessments, and phosphorus emission quantification. This integrated strategy not only reduces fishmeal dependence and production costs but also establishes optimal substitution ratios through environmental impact assessment, offering actionable strategies for sustainable aquaculture practices.
All experiments on animals were approved by the Committee on the Ethics of Animal Experiments at Huzhou University and the Care and Use of Laboratory Animals in China. Approval 20240701; approval 12 July 2024.
Five isonitrogenous and isoenergetic semi-purified diets were formulated using chicken meal, fish meal, soybean meal, and defatted BSFLM as protein sources, with fish oil and soybean oil as lipid sources and wheat flour as the carbohydrate source. All feeds were prepared on the same day using the same batch of raw materials. All protein ingredients underwent prior analysis of crude protein (CP), crude lipid, and amino acid composition (Table S1 and S2) to determine inclusion ratios. The experimental diet formulations are presented in Table 1. The control group (BSFLM0) received a basal diet without BSFLM, while four experimental groups (BSFLM10, BSFLM20, BSFLM40, and BSFLM80) replaced 10%, 20%, 40%, and 80% of fish meal with BSFLM, respectively. Methionine (99% purity) and lysine (70% purity) were supplemented to balance methionine and lysine levels across all diets. Both methionine and lysine used in this study were purchased from Zhejiang Hengxing Feed Co., Ltd. (Jiaxing, Zhejiang, China).Table 1Ingredients and proximate compositions of the experimental diets (%, air-dry basis).Table 1ItemsGroups2BSFLM0BSFLM10BSFLM20BSFLM40BSFLM80IngredientsFish meal36.0032.4028.8021.607.20Chicken meal19.0019.0019.0019.0019.00Soybean meal15.0015.0015.0015.0015.00BSFLM4.408.8017.7035.40Wheat flour16.0016.0016.0016.0016.00Fish oil/soybean oil (1:1)3.102.702.301.500.00Ca(H2PO4)21.501.501.501.501.50Sodium carboxymethylcellulose2.002.002.002.002.00Lysine (70% purity)0.000.050.100.190.37Methionine (99% purity)0.000.030.060.110.23Microcrystalline cellulose4.403.923.442.400.30Other ingredients13.003.003.003.003.00Total100.00100.00100.00100.00100.00Proximate compositionMoisture6.646.886.767.157.67CP44.3044.1844.1244.2144.52Crude lipid9.069.079.078.898.76Organic matter86.2386.2986.9487.2887.64Phosphorus2.172.052.011.841.63Lysine2.862.862.872.872.87Methionine0.960.960.960.950.96Gross energy, kJ/g16.9716.8916.8216.6216.38BSFLM = black soldier fly larvae meal; CP = crude protein.1Other ingredients contained the following components in equal proportions, each accounting for 0.5%: Y2O3 (yttrium oxide), mineral premix, vitamin premix, choline chloride, cholesterol, and soybean lecithin. Vitamin premix contained the following (per kg): vitamin A 6.4 g, vitamin C 151.52 g, vitamin E 71 g, vitamin D3 2 g, vitamin B1 3.2 g, vitamin B2 9 g, vitamin B6 4 g, vitamin B12 0.08 g, nicotinic acid 16 g, folic acid 1 g, inositol 64 g, biotin 0.2 g, calcium pantothenate 14 g, zeolite powder 777.02 g; mineral premix contained the following (per kg): CuSO4·5H2O 1.95 g, ZnSO4·7H2O 30.91 g, MnSO4·H2O 5.23 g, FeSO4·7H2O 24.83 g, Ca(IO3)2 0.46 g, Na2SeO3 0.09 g, CoCl2·6H2O 0.32 g, MgSO4·7H2O 244.9 g.2BSFLM0 (control), BSFLM10, BSFLM20, BSFLM40, and BSFLM80 (80%), were the basal diet with 0, 10%, 20%, 40%, or 80% fishmeal replaced by BSFLM, respectively.
BSFLM was sourced from Zhejiang Kunwei Agricultural Technology Co., Ltd. (Hangzhou, Zhejiang, China), and other protein ingredients (fishmeal, soybean meal, and chicken meal) were procured from Zhejiang Hengxing Feed Co., Ltd. (Jiaxing, Zhejiang, China). The defatting process of BSFLM employs the following first, larvae undergo homogenization and dehydration to form a concentrate. Subsequently, 3 to 5 L of defatting solvent (n-hexane, Shanghai Macklin Biochemical Co., Ltd., Shanghai, China) is added per kg of concentrate. This extraction procedure is repeated 3 to 5 times to remove crude fat. Final separation of oil and solid matter is achieved by centrifugation. All solid components were ground through a 60-mesh sieve, then homogenized stepwise according to the formulation. Fish oil, soybean oil, and water were added to achieve complete homogenization. The mixture was pelleted via a 1.5-mm die using an F-26 twin-screw extruder (South China University of Technology, Guangzhou, Guangdong, China). Pellets were dried in a 40 °C air oven, packaged, labeled, and stored at −20 °C until use.
Healthy M. rosenbergii juveniles were obtained from Zhejiang Zhongyi Aquatic Seedling Technology Co., Ltd. (Huzhou, Zhejiang, China) and acclimated for one week in cement tanks under controlled conditions using a commercially available acclimation diet. Following this period, 900 prawn (initial body weight 0.25 ± 0.03 g, initial age 40 d post-hatching) were randomly allocated into 15 high-density polyethylene tanks. The tanks used as experimental groups were randomly allocated in a glass greenhouse with good lighting and ventilation. Each tank had the same specifications, with a bottom area of approximately 0.5 square meters and a total volume of 300 L. Five dietary treatments were tested, with three replicate tanks per treatment, each stocked with 60 prawns, the bottom of the tanks was kept clean throughout the experiment, and algae growth was prevented.
Prawns were fed to satiation twice daily (08:00 and 00) using the experimental diets. To mitigate aggression and cannibalism during molting, shelters composed of nylon netting and bundled polyvinyl chloride pipes were provided in each tank. Water quality was maintained by replacing 30% of the tank volume daily while monitoring physicochemical temperature (25-28 °C), dissolved oxygen (>6.5 mg/L). Nitrogenous waste was maintained at low levels (ammonia < 0.1 mg/L, nitrite < 0.1 mg/L) through regular water exchange. Water quality physicochemical parameters, including temperature, dissolved oxygen, and pH, were measured using a YSI Pro Plus portable multi-parameter meter (YSI Inc., Yellow Springs, OH, USA). Ammonia nitrogen and nitrite concentrations were determined via colorimetric qualitative measurement using rapid detection kit (LOT: 20240520, Nanhua Qianmu Biotechnology Co., Ltd., Zhengzhou, Henan, China). If concentrations exceeded the standard, the daily water exchange volume was increased until values complied. The experiment was conducted from early July to early September 2024 in Huzhou, China (approximately 30° N, 120° E), within a glass greenhouse under natural light for 8 weeks. Fecal collection commenced in the third week. Residual feed was siphoned 1 h post-feeding; feces were collected from the tank bottom using a siphon and 150-μm mesh sieve 3 h later. Samples were oven-dried at 40 °C, stored at −20 °C, and kept separately by tank for subsequent digestibility analysis.
M. Rosenbergii were fasted for 24 h prior to sampling. Prawns from each tank were enumerated and weighed to evaluate growth performance. Following anesthesia in an ice-water bath (4 °C), approximately 10 individuals per tank were aseptically dissected to collect hemolymph and intestinal tissues. Intestinal samples were divided into two one preserved in eppendorf tubes for gut microbiota analysis and the other homogenized in pre-chilled phosphate buffer (pH 7.4) for digestive enzyme assays. Hepatopancreas tissues from 10 prawns per tank were pooled into four sterile centrifuge tubes.
For morphological analysis, hepatopancreas and intestinal segments from two intact prawns per tank were fixed in 4% paraformaldehyde (Beijing Labgic Technology Co., Ltd., Beijing, China). Muscle tissues from ten prawns per tank were collected for amino acid profiling. The hepatopancreas and intestines of 10 prawns were also aliquoted into cryovials separately. All tissues (hepatopancreas, intestine, and muscle) were flash-frozen in liquid nitrogen immediately after collection and stored at −80 °C for subsequent enzymatic, genomic, and transcriptomic analyses. Hemolymph samples were refrigerated overnight at 4 °C, centrifuged at 3500 × g (4 °C, 10 min), and supernatants stored at −80 °C. Following sampling, all subsequent data analyses were performed at the tank level, with replicate samples derived from the same tank subjected to averaging to yield tank-specific values, which served as the fundamental experimental units for statistical comparisons.
Feed and body composition analyses (Tao et al., 2025) were conducted following Chinese National Standard protocols. Dry matter content (China National Standard, 2006b;ISO 1999) was determined by oven-drying samples at 105 °C to constant weight. Crude lipid (China National Standard, 2006a;ISO 1999) was quantified via Soxhlet extraction using anhydrous ethyl ether. Ash content (China National Standard, 2007;ISO 2002) was measured after incineration in a muffle furnace (MF-36-12P, Tianjin Taisite Instrument Co., Ltd., Tianjin, China) at 550 °C for 6 h. The determination of gross energy (GE) in feed followed the national standard (China National Standard, 2024;), and a bomb calorimeter (Model C2000, IKA Group, Staufen, Germany) was used for the analysis.
Organic matter (%) = 100 - Ash content (%).
Crude protein was analyzed by the Kjeldahl method (China National Standard, 2018b) using an automated nitrogen analyzer (KjelFlex 840, Büchi Laboratory Equipment Trading [Shanghai] Ltd. Shanghai, China) with a conversion factor of 6.25, as described previously (Lynch and Barbano, 1999). All analyses were performed in triplicate, with instrument calibration verified using certified reference materials (bovine serum albumin for protein).
Amino acid separation and detection (China National Standard 2016a) were performed using a Hitachi L-8900 automated amino acid analyzer (Hitachi Ltd., Tokyo, Japan) (Qu et al., 2020). Diet and muscle samples were hydrolyzed in 6 mol/L hydrochloric acid (HCl) at 110 °C for 24 h to ensure complete amino acid hydrolysis. After filtration, centrifugation, and nitrogen blow-drying, the residues were dissolved in 0.02 mol/L HCl solution. The samples were then analyzed using the amino acid analyzer (Hitachi L-8900).
Phosphorus content was quantified following Chinese National Standard (2018c) using dry digestion. Briefly, samples were incinerated at 550 °C for 6 h to obtain ash residues. The ash was digested with 6 mol/L HCl and trace HNO3, boiled for 10 min to solubilize phosphorus, then diluted to volume in volumetric flasks. Phosphorus concentrations were determined spectrophotometrically at 420 nm using a vanadomolybdate colorimetric assay, with reference to a calibrated standard curve.
Phosphorus deposition rate (PDR, %) = (Final weight × Final phosphorus content – Initial weight × Initial phosphorus content)/(Feed intake × Feed phosphorus content) × 100; Phosphorus retention efficiency (PRE, %) = 100 × (Final weight × Final phosphorus content − Initial weight × Initial phosphorus content + Weight of die prawns × Phosphorus content of die prawns)/(Feed intake × Feed phosphorus content); Phosphorus waste (PW, g/kg) = 1000 × (Feed intake × Feed phosphorus content) × (1 − PRE)/(Final weight − Initial weight + Weight of die prawns).
Growth parameters were calculated as
Weight gain (%) = 100 × (Final body weight - Initial body weight)/Initial body weight;
Survival rate (%) = 100 × Final prawn number/Initial prawn number;
Feed conversion ratio = Feed intake/Weight gain;
Average daily feed intake (ADFI, g) = Total feed intake/(Trial days × Number of animals).
Diets were supplemented with 0.5% yttrium oxide (Y2O3) as an inert marker. Apparent digestibility (AD) was calculated as ADdry=1−(Di/Fi),ADprotein=1−[(F/D)×(Di/Fi)],where D, protein of diet (%); F, protein of feces (%); Di, digestion indicator of diet (%); Fi, digestion indicator of feces (%).
The determination of yttrium (Y) content was conducted following the procedures outlined in the GB/T 35871-2018 (China National Standard, 2018a). Briefly, an accurately weighed sample was digested with 5 mL of nitric acid and 1 mL of perchloric acid. The mixture was placed on a hot plate and heated at 130 to 150 °C until the majority of the brown fumes disappeared. The temperature was then increased to 180 °C to continue digestion. If the digestion solution turned brown-black, a small amount of nitric acid was supplemented until white fumes appeared and the solution became colorless or slightly yellow and transparent. After cooling, the digestion solution was transferred into a 50 mL volumetric flask. The beaker was rinsed multiple times with distilled water, and the rinsing solutions were combined into the flask, which was then filled to volume. A reagent blank was prepared simultaneously. The final filtrate was analyzed using an inductively coupled plasma atomic emission spectrometer (ICP-AES) and an inductively coupled plasma mass spectrometer (ICP-MS). The main instruments used included a iCAP 7400 dual-view ICP-AES and a iCAP RQ ICP-MS (Thermo Fisher Scientific Inc., Waltham, MA, USA).
For the determination of antioxidant enzymes in the hepatopancreas and digestive enzyme activities in the intestine, and tissue samples were thawed on ice and homogenized in phosphate-buffered saline (PBS) using a tissue homogenizer (High-speed tissue grinder, Model KZ-II, Jingxin Industrial Development Co., Ltd., Shanghai, China). The homogenates were then centrifuged at 1500 × g for 20 min at 4 °C, and the supernatants were collected for subsequent analyses.
The levels of malondialdehyde (MDA; LOT: 20240904) and the activities of total superoxide dismutase (T-SOD; LOT: 20240827) in the hepatopancreas were determined using commercial assay kits (Nanjing Jiancheng Bioengineering Institute, Nanjing, China). Similarly, the activities of amylase (LOT: 20240904), trypsin and lipase (LOT: 20240907) in the intestine were measured using the corresponding assay kits from the same supplier . The enzyme activities of hepatopancreas and intestine are both expressed as mg prot (tissue homogenate protein), i.e., the enzyme activity units per mg of protein, following the manufacturer's protocols strictly.
Fresh intestinal and hepatopancreas tissue samples were immersed in 4% paraformaldehyde solution (pH 7.4) and fixed at 4 °C for 24 to 48 h, with the fixative volume maintained at 10 to 15 times the tissue volume. After fixation, the samples were washed to remove residual fixative, dehydrated through a graded ethanol series, embedded in paraffin, and sectioned using a rotary microtome. The sections were subsequently stained with hematoxylin and eosin and examined with CaseViewer software (3DHistech Ltd., Budapest, Hungary).
For intestinal microbiota analysis, gut samples (including intestinal contents) were collected from five treatment groups, each with three biological replicates of M. rosenbergii. Samples were sent to Majorbio Bio-Pharm Technology Co., Ltd. (Shanghai, China) for microbial community profiling. Species annotation was conducted using the Bayesian classifier with the Silva 138/16S_bacteria database, and sequence denoising was performed using the DADA2 algorithm.
Microbial community results were analyzed at the phylum and genus levels, with operational taxonomic units (OTUs) serving as the minimum classification units. In microbial ecological and phylogenetic studies, OTUs are commonly used to represent distinct taxa (e.g., strains, species, or genera) to simplify classification. OTUs were clustered at 97% sequence similarity. Community richness was evaluated using the abundance-based coverage estimator (ACE) and Chao indices, community diversity was assessed using the Shannon and Simpson indices, and community coverage was estimated by the coverage indices.
Based on phosphorus utilization and discharge results, the BSFLM0 and BSFLM20 groups were selected for transcriptomic analysis. Three hepatopancreas samples from each group were collected and sent to Majorbio Bio-Pharm Technology Co., Ltd. (Shanghai, China) for analysis. Next-generation high-throughput sequencing technology was employed, and the data were assembled using Trinity. The assembly quality was assessed using BUSCO. All transcripts obtained from this transcriptome sequencing were then aligned with six major databases (NR, non-redundant protein sequence database; Swiss-Prot, Swiss-Prot protein knowledge base; Pfam, Pfam protein families database; EggNOG, evolutionary genealogy of Non-supervised Orthologous Groups; GO, gene ontology; KEGG, Kyoto Encyclopedia of Genes and Genomes) to retrieve annotation information. The results of the annotations were statistically analyzed. Expression levels of the transcripts were quantified using RSEM software (RNA-Seq by Expectation-Maximization, https://github.com/deweylab/RSEM.git). Subsequently, DESeq2 was used to analyze the differential expression of transcripts across samples, and gene regulatory mechanisms were explored by integrating sequence functional information.
Total RNA was extracted using the RN28 kit (Aidlab Biotechnologies Co., Ltd., Beijing, China). The concentration and purity of the extracted RNA were determined using a Thermo Scientific NanoDrop2000 spectrophotometer (Thermo Fisher Scientific Inc., Waltham, MA, USA), ensuring an OD260/280 ratio between 1.8 and 2.0. Subsequently, complementary DNA (cDNA) was synthesized from RNA using the PC70 reverse transcription kit (LOT: 352239AH, Aidlab Biotechnologies Co., Ltd., Beijing, China). Quantitative real-time PCR (qRT-PCR) was performed using the pc5902 SYBR Green kit (LOT: 352041AX, Aidlab Biotechnologies Co., Ltd., Beijing, China) on a CFX Opus 384 Real-Time PCR system (Bio-Rad Laboratories Inc., Hercules, CA, USA). Each qRT-PCR reaction had a final volume of 20 μL. Primers were designed using Primer6 software. The qRT-PCR was performed using a two-step cycling protocol, which included an initial denaturation step at 94 °C for 2 to 3 min (for template denaturation and enzyme activation), followed by 40 amplification cycles consisting of denaturation at 94 °C for 5 to 10 s and combined annealing-extension at 60 °C for 30 to 34 s, and a final dissociation stage where the temperature was from 60 °C to 95 °C at a rate of 0.5 °C/min with real-time fluorescence monitoring to verify product specificity. The relative mRNA expression levels of target genes were calculated using the 2^−ΔΔCt^ method, with β-actin used as the internal reference gene. The sequences of all primers are listed in Table S3.
All data analyzed in this study, were presented as the mean values of each tank, with statistical comparisons performed among different tanks. All data were subjected to one-way analysis of variance (ANOVA) using SPSS 25.0 statistical software (International Business Machines Corp., Armonk, NY, USA). as shown Xij=μ+αi+eij,where Xij is the dependent variable, μ is the overall mean, αi represents the treatment effect, and eij is the random error of the sample.
The linear and quadratic effects of BSFLM increase were analyzed by CONTRAST program in SPSS 25.0. When significant differences among groups were detected (P < 0.05), Duncan's method was performed to identify differences between treatments. Results are expressed as mean. The inter-group standard error of the mean (SEM), linear and quadratic curve trend analysis were shown in the tables.
As shown in Table 2, the replacement of fishmeal with BSFLM did not significantly affect the growth performance of M. rosenbergii. No significant differences were observed among all treatment groups in survival rate, weight gain, and feed conversion ratio (P > 0.05). Similarly, ADFI remained comparable across all groups after 8-week trial (P = 0.306).Table 2Growth performance and feed efficiency of Macrobrachium rosenbergii fed different levels of defatted black soldier fly larvae meal (BSFLM) for 8 weeks.Table 2ItemsGroups1SEMP-valueBSFLM0BSFLM10BSFLM20BSFLM40BSFLM80ANOVALinearQuadraticFinal weight, g3.193.003.412.893.110.0770.2500.6210.798Weight gain,%1176.361100.361265.681055.191143.6030.6300.2500.6210.798Survival rate,%75.4275.8371.6779.1778.331.2460.3720.2380.508Feed conversion ratio1.351.431.341.431.320.0340.7950.6290.687ADFI, g/prawn0.0690.0670.0730.0650.0660.0010.3060.2830.559Total feed intake, g/tank173.73170.80173.84173.03171.940.3820.063––ADFI= average daily feed intake; SEM= standard error of the mean.1BSFLM0 (control), BSFLM10, BSFLM20, BSFLM40, and BSFLM80 (80%), were the basal diet with 0, 10%, 20%, 40%, or 80% fishmeal replaced by BSFLM, respectively, n = 3.
Body composition analysis revealed that CP content increased with higher BSFLM inclusion levels (Table 3). The BSFLM80 group exhibited significantly higher CP levels than other groups (P < 0.001). Furthermore, the CP content in body composition showed a significant linear effect with increasing levels of BSFLM supplementation (P < 0.001). The crude lipid content in the BSFLM10 group was significantly lower than that in the control group (P = 0.014). Ash content was significantly reduced in the control group (BSFLM0) relative to BSFLM-fed groups (P < 0.001). Phosphorus content showed no significant differences among all experimental groups (P = 0.437).Table 3Whole-body proximate composition of Macrobrachium rosenbergii fed different levels of defatted black soldier fly larvae meal (BSFLM) for 8 weeks (% wet weight).Table 3ItemGroups1SEMP-valueBSFLM0BSFLM10BSFLM20BSFLM40BSFLM80ANOVALinearQuadraticCP15.16^e^15.49^d^16.31^c^16.70^b^17.53^a^0.229<0.001<0.001<0.001Crude lipid2.15^ab^1.91^c^2.27^a^2.01^bc^2.03^bc^0.0410.0140.5320.823Ash5.65^c^5.72^b^5.86^a^5.90^a^5.75^b^0.026<0.0010.241<0.001Phosphorus0.410.410.400.420.400.0040.4370.7020.707Moisture69.4769.9169.5370.2969.710.1560.476––SEM= standard error of the mean; CP= crude protein.Within a row, means without a common superscript letter differ at P < 0.05, n = 3.1BSFLM0 (control), BSFLM10, BSFLM20, BSFLM40, and BSFLM80 (80%), were the basal diet with 0, 10%, 20%, 40%, or 80% fishmeal replaced by BSFLM, respectively.
Dietary BSFLM inclusion significantly influenced the muscle amino acid profiles (Table 4). Proline levels increased in all BSFLM-fed groups compared to the control group (P < 0.001), with proline content being particularly elevated in BSFLM20, BSFLM40, and BSFLM80 groups relative to BSFLM10 (P < 0.001), while serine levels were reduced in BSFLM20 compared to BSFLM0, BSFLM10, and BSFLM80 (P = 0.025).Table 4Muscle amino acid compositions of Macrobrachium rosenbergii fed different levels of defatted black soldier fly larvae meal (BSFLM) for 8 weeks (%).Table 4ItemsGroups1SEMP-valueBSFLM0BSFLM10BSFLM20BSFLM40BSFLM80ANOVALinearQuadraticAspartate and asparagine2.232.212.002.002.050.0360.0640.0960.029Threonine0.750.750.720.720.740.0070.5200.6980.303Serine0.81^a^0.82^a^0.76^b^0.78^ab^0.80^a^0.0070.0250.6070.100Glutamate and glutamate0.650.780.720.800.850.0280.1370.0230.063Glycine1.171.191.161.181.100.0160.5280.1150.233Alanine1.101.101.071.071.080.0060.4020.2080.204Valine0.980.990.950.910.970.0150.5780.6840.297Methionine0.610.640.600.610.620.0060.5180.9470.916Isoleucine1.051.020.920.920.950.0230.2150.1720.076Leucine1.631.661.601.601.620.0140.5550.6740.588Tyrosine0.640.740.730.720.730.0130.0540.1960.107Phenylalanine0.790.770.730.720.740.0110.1590.1370.030Lysine1.611.621.561.541.570.0160.4280.2780.206Histidine0.460.460.430.430.450.0060.2680.6970.106Arginine1.881.961.841.841.880.0190.2410.4840.495Proline0.65^c^0.86^b^1.17^a^1.22^a^1.27^a^0.068<0.001<0.001<0.001SEM= standard error of the mean.Within a row, means without a common superscript letter differ at P < 0.05, n = 3.1BSFLM0 (control), BSFLM10, BSFLM20, BSFLM40, and BSFLM80 (80%), were the basal diet with 0, 10%, 20%, 40%, or 80% fishmeal replaced by BSFLM, respectively.
Phosphorus retention efficiency and PDR improved significantly with increasing BSFLM inclusion (Table 5). The BSFLM40 and BSFLM80 groups showed higher PDR values than other groups (P < 0.001), and PRE was significantly elevated in these groups compared to BSFLM0 and BSFLM10 (P < 0.001). Phosphorus waste was significantly lower in BSFLM20, BSFLM40, and BSFLM80 groups (P < 0.001). Both PRE and PDR, along with PW, exhibited significant linear and quadratic effects as BSFLM inclusion levels increased (P < 0.05).Table 5Phosphorus utilization and discharge of Macrobrachium rosenbergii fed different levels of defatted black soldier fly larvae meal (BSFLM) for 8 weeks.Table 5ItemsGroups1SEMP-valueBSFLM0BSFLM10BSFLM20BSFLM40BSFLM80ANOVALinearQuadraticPDR,%13.69^c^13.83^c^14.49^c^16.20^b^18.16^a^0.322<0.001<0.001<0.001PRE,%14.57^b^14.68^b^15.93^b^19.24^a^18.87^a^0.372<0.001<0.001<0.001PW, g/kg14.46^b^15.27^a^13.17^c^13.65^c^13.47^c^0.161<0.0010.0060.005PDR= phosphorus deposition rate; PRE= phosphorus retention efficiency; PW= phosphorus waste (P per prawn weight gain) g/kg; SEM= standard error of the mean.Within a row, means without a common superscript letter differ at P < 0.05, n = 3.1BSFLM0 (control), BSFLM10, BSFLM20, BSFLM40, and BSFLM80 (80%), were the basal diet with 0, 10%, 20%, 40%, or 80% fishmeal replaced by BSFLM, respectively.
Intestinal trypsin activity increased significantly in all BSFLM-fed groups except BSFLM10 compared to the control (P = 0.004; Table 6). Amylase and lipase activities were also significantly higher in BSFLM40 and BSFLM80 groups (P < 0.05).Table 6Intestinal digestive enzyme activities of Macrobrachium rosenbergii fed different levels of defatted black soldier fly larvae meal (BSFLM) for 8 weeks.Table 6ItemsGroups1SEMP-valueBSFLM0BSFLM10BSFLM20BSFLM40BSFLM80ANOVALinearQuadraticTrypsin, U/mg prot207.99^c^276.56^bc^374.85^a^300.86^ab^369.55^a^18.1320.0040.0150.022Amylase, U/mg prot2.85^c^3.26^b^2.89^c^3.70^a^3.39^ab^0.0950.0020.0390.025Lipase, U/g prot4.10^b^4.87^b^4.18^b^7.51^a^8.27^a^0.516<0.001<0.001<0.001SEM= standard error of the mean; prot = protein.Within a row, means without a common superscript letter differ at P < 0.05, n = 3.1BSFLM0 (control), BSFLM10, BSFLM20, BSFLM40, and BSFLM80 (80%), were the basal diet with 0, 10%, 20%, 40%, or 80% fishmeal replaced by BSFLM, respectively.
No significant differences were observed in MDA content among groups (P = 0.343; Table 7). However, T-SOD activity was significantly elevated in the BSFLM-fed groups (P < 0.001).Table 7Enzyme activities involved in antioxidant capacity in hepatopancreas of Macrobrachium rosenbergii fed test diet.Table 7ItemsGroups1SEMP-valueBSFLM0BSFLM10BSFLM20BSFLM40BSFLM80ANOVALinearQuadraticMDA, nmol/mg prot38.9333.3539.0935.8736.510.9910.3430.7730.908T-SOD, U/mg prot36.67^b^43.56^a^41.96^a^46.23^a^45.49^a^0.763<0.0010.001<0.001MDA= malondialdehyde; T-SOD= total superoxide dismutase; SEM= standard error of the mean; prot = protein.Within a row, means without a common superscript letter differ at P < 0.05, n = 3.1BSFLM0 (control), BSFLM10, BSFLM20, BSFLM40, and BSFLM80 (80%), were the basal diet with 0, 10%, 20%, 40%, or 80% fishmeal replaced by BSFLM, respectively.
Dry matter and protein apparent digestibility (ADdry and ADprotein) did not differ significantly (P > 0.05) among groups (Table 8).Table 8The apparent digestibility (AD) of Macrobrachium rosenbergii fed different levels of defatted black soldier fly larvae meal (BSFLM) for 8 weeks (%).Table 8ItemsGroups1SEMP-valueBSFLM0BSFLM10BSFLM20BSFLM40BSFLM80ANOVALinearQuadraticADdry72.6972.6873.0972.4371.890.3230.8700.3250.581ADprotein85.1385.0185.3084.6384.900.1580.7840.4970.722SEM= standard error of the mean;ADdry = the apparent digestibility of dry matter; ADprotein = the apparent digestibility of protein.1BSFLM0 (control), BSFLM10, BSFLM20, BSFLM40, and BSFLM80 (80%), were the basal diet with 0, 10%, 20%, 40%, or 80% fishmeal replaced by BSFLM, respectively (n = 3).
Transcriptome sequencing of hepatopancreas samples yielded high-quality data (Q20 > 98.52%, Q30 > 95.58%), with 31,761 unigenes and 41,003 transcripts assembled (average N50 = 2801 bp). Differential expression analysis identified 1141 upregulated and 694 downregulated genes in BSFLM20 vs. BSFLM0 (Fig. 1A). KEGG enrichment revealed significant involvement of differential expression genes (DEGs) in tight junctions, phagosomes, antigen processing/presentation, insulin signaling pathway, FoxO signaling pathway, mineral absorption, AGE-RAGE signaling pathway, PI3K-Akt signaling pathway, and cGMP-PKG signaling pathway (Fig. 1B).Fig. 1Analysis of differentially expressed genes in the hepatopancreas of Macrobrachium rosenbergii fed different levels of defatted black soldier fly larvae meal (BSFLM) for 8 weeks. (A) Scatter plot of expression level differences of BSFLM20 vs. BSFLM0 (A). (B) KEGG enrichment analysis of DEGs in the hepatopancreas of Macrobrachium rosenbergii of BSFLM20 vs. BSFLM0. BSFLM0 (control) and BSFLM20 were the basal diet with 0 and 20% fishmeal replaced by BSFLM, respectively. KEGG = Kyoto Encyclopedia of Genes and Genomes; DEG = differential expressed genes.Fig. 1
Gene ontology enrichment showed downregulated DEGs were associated with macromolecule and nitrogen compound metabolic processes (8 processes; Fig. 2A), while upregulated DEGs participated in gluconeogenesis and hexose biosynthesis (13 processes; Fig. 2B). Quantitative RT-PCR validation confirmed the transcriptomic results (Fig. 3), and among all 10 DEGs tested, the trend of log2FC was consistent with the transcriptomic results.Fig. 2Enrichment analysis of DEGs in the hepatopancreas of Macrobrachium rosenbergii fed different levels of defatted black soldier fly larvae meal (BSFLM) for 8 weeks. (A) GO enrichment analysis of downregulated DEGs of BSFLM20 vs. BSFLM0. (B) GO enrichment analysis of upregulated DEGs of BSFLM20 vs. BSFLM0. BSFLM0 (control) and BSFLM20 were the basal diet with 0 and 20% fishmeal replaced by BSFLM, respectively. GO = gene ontology; DEG = differential expressed genes.Fig. 2Fig. 3Validation of transcriptome sequencing data in BSFLM20 vs. BSFLM0 by qRT-PCR. BSFLM0 (control) and BSFLM20 were the basal diet with 0 and 20% fishmeal replaced by BSFLM, respectively. BSFLM = defatted black soldier fly larvae meal; RNA-Seq = RNA-Sequening.Fig. 3
Comprehensive analysis of intestinal microbiota was performed on samples collected from 15 M. rosenbergii individuals distributed across five treament groups. High-throughput sequencing generated high-quality 16S rRNA gene reads, with no significant difference (P > 0.05) observed in the number of reads among samples (Table S4). The sequencing coverage exceeded 99.9% for all groups. Alpha diversity metrics, including Simpson and Shannon indices calculated from OTUs, revealed no significant variations (P > 0.05) among treatment groups (Table S4).
Principal coordinates analysis (PCoA) clearly demonstrated distinct clustering patterns of microbial communities across different dietary treatments (Fig. 4A), indicating substantial modifications in microbiota composition induced by BSFLM supplementation (P = 0.001). At the phylum level, the intestinal microbiota was predominantly composed of Proteobacteria, Bacteroidota, Firmicutes, Actinobacteriota, and Patescibacteria (Fig. 4B). Detailed taxonomic analysis at the genus level identified five dominant bacterial Cloacibacterium, Lactococcus, Acinetobacter, Paracoccus, and Agromyces (Fig. 4C).Fig. 4Analysis of intestinal microbiota community structure in Macrobrachium rosenbergii fed different levels of defatted black soldier fly larvae meal (BSFLM) for 8 weeks. (A) Principal coordinates analysis (PCoA) among BSFLM0, BSFLM10,BSFLM20, BSFLM40 and BSFLM80 groups showing clustering by diet. Histogram of intestinal microbiota distribution at the phylum (B) and genus (C) level in Macrobrachium rosenbergii fed different levels of defatted black soldier fly larvae meal for 8 weeks. For each treatment, n = 3, where each sample is a mixture of hindgut contents collected from each tank. BSFLM0 (control), BSFLM10, BSFLM20, BSFLM40, and BSFLM80 (80%), were the basal diet with 0, 10%, 20%, 40%, or 80% fishmeal replaced by BSFLM, respectively. ASV = amplicon sequence variant.Fig. 4
Statistical analysis of microbial composition revealed significant dietary effects at both phylum and genus levels (Table 9). The BSFLM10 group showed a marked increase in Firmicutes abundance compared to other groups (P < 0.001), while the control group (BSFLM0) maintained significantly higher levels of Patescibacteria (P = 0.031). At the genus level, Acinetobacter was particularly abundant in the BSFLM20 group (P < 0.001), whereas Agromyces showed significant elevation in BSFLM40 relative to all groups except BSFLM80 (P = 0.031).Table 9Main bacterial phyla and genera in intestinal microflora of Macrobrachium rosenbergii fed different levels of defatted black soldier fly larvae meal (BSFLM) for 8 weeks.Table 9ItemsGroups1SEMP-valueBSFLM0BSFLM10BSFLM20BSFLM40BSFLM80ANOVALinearQuadraticPhylum levelProteobacteria29.1536.2344.5431.9340.252.8140.4710.4910.754Bacteroidota35.186.1931.9118.4119.943.2800.0520.5140.671Firmicutes14.93^bc^42.64^a^6.22^c^19.90^b^11.93^bc^3.679<0.0010.2920.586Actinobacteriota6.1910.276.8921.6618.931.9170.0650.0040.006Patescibacteria9.73^a^1.50^c^5.68^b^2.35^bc^4.67^bc^0.8860.0310.3490.087Genus levelCloacibacterium30.70^a^5.72^b^30.11^a^15.73^ab^19.50^ab^3.1220.0250.6600.756Lactococcus7.8127.824.5418.2410.092.9480.0520.7350.800Acinetobacter5.83^b^1.88^b^25.84^a^3.52^b^1.24^b^2.531<0.0010.3140.276Paracoccus6.417.003.0910.602.661.0370.0650.4180.304Agromyces3.44^b^5.30^b^3.54^b^10.25^a^6.83^ab^0.8520.0310.0990.059SEM= standard error of the mean.Within a row, means without a common superscript letter differ at P < 0.05, n = 3.1BSFLM0 (control), BSFLM10, BSFLM20, BSFLM40, and BSFLM80 (80%), were the basal diet with 0, 10%, 20%, 40%, or 80% fishmeal replaced by BSFLM, respectively.
Histological examination revealed well-preserved hepatopancreas architecture across all treatment groups (Fig. 5). The organ displayed characteristic pseudostratified epithelium with tightly connected cells lining the hepatopancreatic tubules. Comparative analysis indicated that BSFLM80-fed prawns contained a greater number of blister cells (B-cells) compared to other dietary groups (Fig. 5E).Fig. 5The hepatopancreas morphology of Macrobrachium rosenbergii fed different levels of defatted black soldier fly larvae meal (BSFLM) for 8 weeks. (A), (B), (C), (D), and (E) represent BSFLM0, BSFLM10, BSFLM20, BSFLM40 and BSFLM80 group, respectively. BSFLM0 (control), BSFLM10, BSFLM20, BSFLM40, and BSFLM80 (80%), were the basal diet with 0, 10%, 20%, 40%, or 80% fishmeal replaced by BSFLM, respectively. Blister cells (B-cells, indicated by arrows in panel E) appear as large vacuolated cells. Scale bar = 100 μm, the magnification is 200×.Fig. 5
Intestinal morphology assessment demonstrated significant dietary effects on gut structure (Fig. 6, Table 10). BSFLM supplementation resulted in substantial increases in both villus height and muscularis thickness compared to the control diet (P < 0.001). Among BSFLM-fed groups, the BSFLM20 treatment elicited the most pronounced morphological improvements, showing significantly greater villus height and muscularis thickness than other groups (P < 0.001). However, higher inclusion levels (BSFLM40 and BSFLM80) resulted in reduced villus height and muscularis thickness relative to BSFLM10 and BSFLM20 groups (P < 0.05).Fig. 6The intestinal morphology of Macrobrachium rosenbergii fed different levels of defatted black soldier fly larvae meal (BSFLM) for 8 weeks. (A), (B), (C), (D), and (E) represent BSFLM0, BSFLM10, BSFLM20, BSFLM40 and BSFLM80 group, respectively. Groups: BSFLM0 (control), BSFLM10, BSFLM20, BSFLM40, and BSFLM80 (80%), were the basal diet with 0, 10%, 20%, 40%, or 80% fishmeal replaced by BSFLM, respectively. Blue arrows mark muscularis thickness; Red arrows mark villous height. Scale bar = 50 μm, the magnification is 400×.Fig. 6Table 10Histological parameters of intestine of Macrobrachium rosenbergii fed different levels of defatted black soldier fly larvae meal (BSFLM) for 8 weeks (μm).Table 10ItemsGroups1SEMP-valueBSFLM0BSFLM10BSFLM20BSFLM40BSFLM80ANOVALinearQuadraticVillous height52.03^d^87.55^b^104.61^a^69.09^c^66.56^c^2.835<0.0010.404<0.001Muscularis thickness13.45^d^22.94^b^27.94^a^18.21^c^17.50^c^0.763<0.0010.439<0.001SEM= standard error of the mean.Within a row, means without a common superscript letter differ at P < 0.05, n = 3.1BSFLM0 (control), BSFLM10, BSFLM20, BSFLM40, and BSFLM80 (80%), were the basal diet with 0, 10%, 20%, 40%, or 80% fishmeal replaced by BSFLM, respectively.
Compared with traditional plant protein alternatives, insect meal offers the advantage of reducing the negative impacts of anti-nutritional factors such as phytic acid (Silva and Bracarense, 2016), thereby potentially improving feed palatability and providing a more balanced amino acid profile (Robinson et al., 2024). In the present study, an 8-week feeding trial with various replacement levels of BSFLM to fish meal revealed no significant differences in growth rate, survival rate, or feed conversion ratio of M. rosenbergii. Furthermore, the ADFI showed no significant variation among treatment groups. These findings demonstrate the considerable potential of BSFLM as a fishmeal substitute in M. rosenbergii aquaculture. Specifically, a short-term result indicates that a defatted BSFLM can replace up to 80% of fishmeal (in a diet containing 36% fishmeal) without compromising prawn yield. This aligns with the findings in Pacific white shrimp, where fishmeal replacement level of 65% or 75% with BSFLM showed no adverse effects on growth performance of (Chang et al., 2025; Wang et al., 2021).
In contrast, many fish species exhibited reduced growth performance and feed utilization at high BSFLM inclusion rates (Hua, 2021), potentially due to low chitin-related palatability issues and impaired nutrient digestibility (Rust et al., 2002). Notably, BSFLM contain approximately 3.6% chitin (Wang et al., 2020). Interestingly, M. rosenbergii appears uniquely adapted to chitin, with of 5% purified chitin actually enhancing growth in juveniles (Kumar et al., 2006). This may be attributed to endogenous chitinase enzymes that facilitate chitin digestion and participate in immune responses (Niu et al., 2018), suggesting minimal digestive impacts from BSFLM inclusion.
The apparent digestibility of dry matter, reflecting overall nutrient absorption efficiency (Terrazas-Fierro et al., 2010), showed no significant differences among groups, indicating that high BSFLM replacement does not impair gross nutrient utilization in M. rosenbergii. Digestive enzyme analyses revealed that BSFLM inclusion enhanced intestinal lipase activity of M. rosenbergii, particularly in BSFLM40 and BSFLM80 groups. This may reflect improved utilization of BSFLM lipids, which are rich in lauric acid and other saturated fatty acids that enhance prawn metabolism (Ewald et al., 2020). Trypsin activity was also significantly elevated in all BSFLM groups except BSFLM10, demonstrating effective adaption to BSFLM protein. Similarly, the intestinal amylase activity increased significantly in the BSFLM40 and BSFLM80 groups, possibly due to residual enzymatic activity from BSF larvae (Kim et al., 2011). These findings collectively position BSFLM not only as a viable fishmeal alternative but also as a functional feed ingredient that enhances digestive performance of M. rosenbergii.
Notably, in high-proportion BSFLM replacement groups, the reduction of fish oil and soybean oil to maintain constant lipid levels enabled investigation into M. rosenbergii's adaptation to BSFLM-derived fatty acids. Despite BSFLM introducing elevated saturated fatty acids (e.g., lauric acid, C12:0) and reducing highly unsaturated fatty acids content, prawns maintained unaffected body composition and growth performance, indicating significant adaptive capacity. This aligns with their inherent fatty acid metabolism M. rosenbergii exhibits no preferential utilization of n-3 vs. n-6 polyunsaturated fatty acids when nutritional requirements are met (D'Abramo and Sheen, 1993), facilitating adaptation to BSFLM-induced fatty acid profile alterations. Crucially, their capacity to elongate and desaturate C18 polyunsaturated fatty acids into highly unsaturated fatty acids (Kamarudin and Roustaian, 2002) explains why BSFLM—despite its low content of highly unsaturated fatty acids—did not impede they effectively compensate for highly unsaturated fatty acids deficits. These results substantiate these mechanisms, confirming their metabolic flexibility and supporting the feasibility of high-proportion BSFLM substitution for fishmeal.
After the 8 weeks feeding trial, the CP content of prawns exhibited an increasing trend following BSFLM replacement, suggesting high protein bioavailability in BSFLM. This observation is supported by the elevated intestinal trypsin activity detected in this study. However, previous reports found increased body CP content within certain BSFLM replacement ranges (Cummins et al., 2017). The observed variations may be attributed to either the defatting process of BSFLM or species-specific digestive characteristics.
Interestingly, the CP content of prawns remained stable across increasing BSFLM inclusion levels except BSFLM10. BSFLM is known to contain high-quality lipid sources, particularly lauric acid (Choi et al., 2012), a short-chain saturated fatty acid that undergoes rapid oxidation for energy production. This suggests efficient utilization of BSFLM lipids, which can effectively replace traditional lipid sources such as fish oil. Kroeckel et al. (2012) proposed that the defatting process might integrate lipids into the chitin structure, thereby affecting lipid availability.
Notably, an increase in crude ash content was observed in the body composition of prawns following BSFLM inclusion. According to Rødde et al. (2008), the prawn exoskeleton primarily consists of chitin, mineral compounds (predominantly calcium carbonate), and proteins. The elevated ash content may therefore reflect enhanced digestion of chitin and improved absorption of minerals and proteins in the exoskeletal structure of M. rosenbergii.
Regarding phosphorus metabolism, BSFLM contains significantly less phosphorus than fishmeal (Alfiko et al., 2022), resulting in decreased dietary phosphorus levels with increasing BSFLM replacement in the experimental diets. However, BSFLM supplementation improved phosphorus retention efficiency and deposition rate while reducing phosphorus discharge in M. rosenbergii, suggesting potential environmental benefits. This is likely primarily due to the decrease in phosphorus levels in the feed. In the present study, the PRE of M. rosenbergii ranged from 14.57% to 18.87%, indicating relatively low phosphorus utilization efficiency. This finding is consistent with observations in other shrimp species. Da Silva et al. (2013) reported only 35% dietary phosphorus absorption in L. vannamei, while Su et al. (2009) found that merely 20.69% of input phosphorus was utilized by Chinese white shrimp (Fenneropenaeus chinensis), with over half (52.04%) accumulating in pond sediments. Notably, fishmeal contains substantial quantities of indigestible phosphorus forms such as bone phosphates. Research on Atlantic salmon (Nordrum et al., 1997) demonstrates that phosphorus utilization efficiency from fish bone meal is significantly lower than from inorganic phosphorus sources. Consequently, the absence of significant differences in body phosphorus content in the present study is not unexpected. According to Sahu et al. (2013), in large-scale ponds simulating actual M. rosenbergii production environments with monitored phosphorus inputs, only 10.09% to 10.97% of total phosphorus was recovered at harvest. This range aligns with the phosphorus digestibility rate of the fishmeal group observed in the present study (13.69%), yet remains substantially lower than all replacement groups. These findings suggest that BSFLM replacement feeds with reduced phosphorus content may enhance phosphorus utilization efficiency and mitigate phosphorus waste emissions, thus systematically reducing the risk of water eutrophication in the aquaculture process.
The amino acid profiles in muscle were observed to be significantly influenced by the inclusion of BSFLM, particularly proline. Although BSFLM contains lower concentrations of proline compared to the fish meal, their levels in muscle tissue increased following fish meal replacement with BSFLM. This intriguing observation suggests that BSFLM inclusion may stimulate proline-related metabolic pathways. Meanwhile, proline has been shown to modulate innate lymphoid cell type 3 metabolism and cytokine production, thereby supporting intestinal barrier integrity and immune homeostasis (Hepworth, 2023). Thus, the elevated levels of proline may confer health benefits to prawns. Notably, proline is a major structural component of collagen and plays a critical role in connective tissue formation. Therefore, the enhancement of the immune-antioxidant system may be associated with changes in proline.
Dietary BSFLM supplementation significantly influenced the antioxidant capacity of M. rosenbergii. Results demonstrated a notable increase in T-SOD activity in BSFLM-fed groups compared to the control, while MDA levels in hepatopancreas tissue remained stable (Table 7). As a terminal product of lipid peroxidation, MDA serves as a reliable biomarker for evaluating oxidative stress status (Ito et al., 2019). The stable MDA levels indicate that dietary changes in this experiment did not induce lipid peroxidation in M. rosenbergii. The enhanced T-SOD activity, which plays a crucial role in neutralizing reactive oxygen species (Moreno et al., 2005), suggests that BSFLM inclusion effectively bolstered the prawns' antioxidant defense system. These findings align with previous observations in Pacific white shrimp (Shin et al., 2020). This may be attributed to the immunostimulatory effects of compounds like chitin and lauric acid. Notably, chitin has been demonstrated to enhance the immune and antioxidant systems in M. rosenbergii (Kumar et al., 2015).
Histopathological examination revealed well-preserved hepatopancreas architecture across all treatment groups, indicating that BSFLM supplementation did not induce adverse morphological changes (Fig. 5). Of particular interest, prawns fed the BSFLM80 diet exhibited a marked increase in B-cell population. B-cells are recognized to process digestive remnants within hepatopancreatic tubules (Vogt, 1993), and are proposed to synthesize and recycle fat emulsifiers (Vogt, 2019). This functionality not only facilitates digestion but may also constitute an adaptive mechanism against nutritional stressors or elevated metabolic demands. Given the absence of necrosis or inflammation in histological sections, it is postulated that the B-cell proliferation likely represents both an adaptation to BSFLM and enhancement of digestive capacity. This adaptation may contribute to the elevated digestive enzyme activity observed in BSFLM-fed prawns, as documented in preceding sections.
In this study, the composition and abundance of intestinal microbiota in M. rosenbergii were significantly altered with changes in dietary BSFLM levels. The intestinal microbiota composition showed distinct phylum-level patterns, with Proteobacteria, Bacteroidota, Firmicutes, Actinobacteriota, and Patescibacteria constituting the dominant populations, consistent with previous reports in this species (Xu et al., 2023). Notably, the BSFLM10 group exhibited the highest relative abundance of Firmicutes, a phylum known for its role in short chain fatty acid production (Ghosh and Pramanik, 2021). Short chain fatty acids serve as crucial energy substrates for intestinal epithelial cells and play pivotal roles in maintaining gut barrier function and immune homeostasis, suggesting that moderate BSFLM inclusion (10%) may confer intestinal health benefits. Notably, high BSFLM inclusion levels (40% to 80%) induced the enrichment of Actinobacteriota compared to the control group. This phylum, known for its ecological role in organic matter decomposition (Iwasaki et al., 2020), contains numerous species capable of chitinolytic activity. The observed microbial shift likely represents an adaptive response to the chitin content in BSFLM, as these bacteria can facilitate chitin breakdown while potentially enhancing host digestion and immune function (Costantini et al., 2017).
Microbial analysis at the genus level revealed distinct probiotic effects at lower inclusion levels. The BSFLM10 group showed enrichment of Lactococcus, a well-documented probiotic fermentative bacterial genus, it can fully exert its function in regulating the intestinal health of M. rosenbergii by producing beneficial secondary metabolites and improves growth performance and gut health in aquatic species (Wang et al., 2024; Zhu et al., 2021). Concurrently, a marked reduction in the abundance of Cloacibacterium, a genus associated with nitrogen cycling (Emmanuel et al., 2025) and potential disease transmission (Lin et al., 2023), was observed. Through genomic analysis, Lund et al. (2018) elucidated the symbiotic relationship of Agromyces with host organisms in intestinal environments and its heterotrophic metabolic characteristics. The genome encodes multiple carbohydrate transporters, demonstrating dependence on host-derived intestinal polysaccharides and amino acids, while harboring relatively complete biosynthetic and salvage pathways for both amino acids and nucleotides. Consequently, it is hypothesized that Agromyces may perform analogous digestive functions in the intestinal tract of M. rosenbergii. The observed abundance increase likely represents an adaptive response to elevated chitin levels or altered amino acid profiles following increased BSFLM inclusion, suggesting the presence of growth-promoting factors for Agromyces in BSFLM. These coordinated changes suggest that optimal BSFLM inclusion (10%) promotes beneficial microbial communities while suppressing potentially pathogenic taxa. Most importantly, there were no significant differences in opportunistic pathogenic genera (such as Vibrio spp.) among the BSFLM-treated groups, indicating that BSFLM supplementation did not disrupt the stability of the intestinal microbiota in a harmful manner.
Intestinal structural integrity primarily encompasses the tightness of epithelial cells and villus height, both of which are closely associated with crustacean health and nutrient absorption (Liao et al., 2022). In this study, the inclusion of black soldier fly larvae meal, particularly at the 20% replacement level, significantly increased intestinal villus height and muscularis thickness, suggesting enhanced intestinal absorptive capacity and structural integrity. This finding aligns with previous research in piglets, where BSFLM supplementation improved intestinal morphology under viral challenge by increasing ileal villus height and the villus height-to-crypt depth ratio (Yu et al., 2024). Similarly, complete replacement of fishmeal with BSFLM in Atlantic salmon showed no negative effects on intestinal health (Li et al., 2020).
Collectively, these findings demonstrate that low-level substitution (10% to 20%) of BSFLM optimizes intestinal microbiota structure by enriching probiotic taxa and morphology, while higher substitution levels (40% to 80%) maintain prawn safety by selectively enhancing chitinolytic bacterial populations.
Transcriptome sequencing has been extensively applied in aquatic animal research to elucidate molecular mechanisms underlying growth regulation, immune responses, and evolutionary adaptation. In the present investigation, KEGG enrichment analysis of M. rosenbergii fed fishmeal substituted with BSFLM revealed significant enrichment of differentially expressed genes in four metabolic the insulin signaling pathway, FoxO signaling pathway, mineral absorption pathway, and advanced glycation end product-receptor for advanced glycation end product (AGE-RAGE) signaling pathway associated with diabetic complications. These findings collectively demonstrate that BSFLM substitution substantially modulates carbohydrate metabolism, mineral homeostasis, and lipid metabolic processes in prawns. Furthermore, transcriptomic profiling identified additional enriched pathways including phosphatidylinositol 3 kinase-protein kinase B (PI3K-Akt) signaling, cyclic GMP monophosphate-protein kinase G (cGMP-PKG) signaling, tight junction regulation, phagosome activity, and antigen processing/presentation, suggesting multifaceted impacts of BSFLM on both physiological adaptation and innate immunity.
The observed alterations in carbohydrate metabolism were particularly pronounced, potentially attributable to chitosan content in insect-derived proteins like BSFLM (Lagat et al., 2021), which may modify gut microbiota composition and digestive enzyme functionality, thereby regulating glucose transporter expression. Concurrently, the enrichment of the mineral absorption pathway implies enhanced bioavailability of essential minerals including calcium, phosphorus, and iron. The activation of AGE-RAGE signaling likely results from advanced glycation end-products (AGEs) formed during BSFLM processing, particularly through hot-air drying defatting procedures (Teodorowicz et al., 2018). Given that AGE-RAGE interactions initiate inflammatory cascades via receptor binding (Medeiros et al., 2014), optimization of processing methodologies could mitigate associated oxidative stress. In lipid metabolism, the activation of phospholipase D and sphingolipid signaling pathways indicates BSFLM-induced modifications in membrane lipid dynamics. This phenomenon may be driven by BSFLM-derived fatty acids such as lauric acid and palmitic acid, which are known regulators of lipid metabolism gene expression and cellular membrane architecture (Barker et al., 2004). The changes in these digestion and absorption-related pathways may precisely account for the significant increase in intestinal digestive enzyme activity in the high-proportion replacement groups.
The significant enrichment of the PI3K-Akt pathway—a central regulator of cellular growth and survival (Franke et al., 2003)—was marked by upregulation of PI3K and 4EBP, suggesting BSFLM promotes muscle protein synthesis through mammalian target of rapamycin signaling activation (Saxton and Sabatini, 2017). This may be related to the increase in CP content in body composition. Parallel upregulation of the cGMP-PKG pathway, which governs physiological processes ranging from smooth muscle relaxation to gene expression (Gambaryan, 2022; Kim et al., 2021), implies systemic adaptation to dietary BSFLM.
Moreover, the enrichment of pathways such as tight junctions, phagosomes, and antigen processing and presentation may indicate complex effects of BSFLM on the innate immune system of M. rosenbergii. BSFLM are rich in antimicrobial peptides, and some of these bioactive molecules may influence immune performance through modulation of heat shock protein families. This may reflect the reasons behind the changes in the immune-antioxidant system, which in turn lead to the enhanced hepatopancreatic antioxidant capacity in the high-proportion replacement groups. In this study, Hsp90 was significantly downregulated. Given that Hsp90 is associated with responses to chronic stress (Prodromou, 2016), its downregulation may reflect the prawn's good adaptability to BSFLM diets.
This study provides compelling evidence that defatted BSFLM serves as a nutritionally complete and functionally effective alternative to fishmeal in M. rosenbergii diets. Findings demonstrate that BSFLM substitution at varying inclusion levels (10% to 80%) maintained comparable growth performance to conventional fishmeal-based diets, with no significant differences observed in key production parameters including survival rate, weight gain, and feed conversion efficiency. Notably, the study revealed distinct beneficial effects associated with different substitution levels. Low-level inclusion (10%) induced modulation of gut microbiota, characterized by increased abundance of Firmicutes and probiotic Lactococcus species, which correlated with enhanced intestinal morphology and improved antioxidant capacity. These changes suggest BSFLM's potential to promote gut health and immune function at optimal inclusion levels. Higher substitution rates (40% to 80%) demonstrated reduced phosphorus discharge, while simultaneously enriching chitin-digesting microbial populations and enhancing digestive enzyme activity. These findings position BSFLM as a multifunctional feed ingredient capable of addressing both production and environmental sustainability challenges in prawn aquaculture due to reduced fishmeal use and lower phosphorus waste. However, the 8-week experimental duration highlights the need for further investigation into the long-term effects of high BSFLM inclusion in commercial-scale operations to fully evaluate its practical application potential in the aquaculture industry.
Zhiyuan Chen: Writing – original draft, Data curation. Qingxuan Li: Validation, Supervision, Project administration. Ruimin Sheng: Project administration, Methodology. Jiacheng Zhang: Visualization, Data curation. Jianlin Guo: Software, Investigation, Formal analysis. Peng Tan: Supervision, Resources, Methodology. Shengjun Bao: Resources. Yujie Liu: Resources, Project administration. Youqin Kong: Resources, Data curation. Hongfeng Bai: Resources, Data curation. Zhili Ding: Writing – review & editing, Funding acquisition.
We declare that we have no financial and personal relationships with other people or organizations that can inappropriately influence our work, and there is no professional or other personal interest of any nature or kind in any product, service and/or company that could be construed as influencing the content of this paper. The authors declare the following financial interests/personal relationships which may be considered as potential competing Shengjun Bao is currently employed by Zhejiang Kunwei Agricultural Technology Co., Ltd. (Hangzhou, Zhejiang, China).