MATERIALS AND METHODS

Ethical approval

This study strictly followed the 3R principles of replacement, reduction, and refinement. All experimental procedures involving goldfish were reviewed and approved by the National Animal Care and Use Committee, Naresuan University, Phitsanulok, Thailand, under approval number NU-AQ660703. The experiment was also conducted in accordance with the guidelines of the Institute of Animals for Scientific Purpose Development under the National Research Council of Thailand’s Ethics of Animal Experimentation (License No. U1/00704/2558).

All fish handling, acclimatization, feeding, anesthesia, blood sampling, tissue collection, and euthanasia-related procedures were performed by trained personnel using methods designed to minimize stress, pain, and unnecessary harm. Fish were acclimatized before the feeding trial, maintained under controlled water-quality conditions, and monitored daily for health, behavior, feeding activity, and survival. Anesthesia was applied before color measurement, blood sampling, and terminal procedures, and fish welfare was prioritized throughout the 10-week experimental period. No treatment-related mortality, abnormal behavior, or external lesions were observed during the study.

Study period and location

The feeding trial was conducted over a 10-weeks period from May to July 2024 at the Laboratory of Fish Nutrition, Naresuan University, Phitsanulok, Thailand. Healthy goldfish (C. auratus) fingerlings were purchased as mixed-sex individuals from a fish village market in Baan Pong District, Ratchaburi Province, Thailand, and transported to the laboratory for acclimatization and experimental rearing.

Study design and rearing conditions

Healthy goldfish were initially screened based on active swimming behavior. Fish were acclimatized to laboratory conditions for 2 weeks in 500-L plastic tanks at a stocking density of 1.2 fish/L, with continuous water circulation maintained at a flow rate of 1200 L/h under a natural photoperiod of approximately 12 h light and 12 h dark. During acclimatization, fish were hand-fed a commercial diet containing 70 g/kg crude lipid, 400 g/kg crude protein, and 40 g/kg crude fiber three times daily to apparent satiation.

At the beginning of the feeding experiment, goldfish with an average body weight of 9.02 ± 0.01 g/fish were randomly collected, bulk-weighed, and allocated to each treatment at 20 fish per tank, in triplicate, into 15 glass tanks with a capacity of 150 L, each filled with 120 L of dechlorinated water. The experimental tanks were arranged using a completely randomized design to account for possible positional effects. Each tank was supplied with dechlorinated tap water and continuous aeration through an airstone to maintain adequate oxygenation. Approximately one-third of the water in each tank was siphoned daily after routine cleaning to remove accumulated waste and solid residues. Fish were manually fed the assigned experimental diets to apparent satiation three times daily at 08:30, 12:30, and 17:30 throughout the 10-week feeding trial. Water-quality parameters were monitored daily. Water temperature ranged from 26.4°C to 28.5°C, dissolved oxygen remained above 6.97 mg/L, and pH ranged from 7.61 to 7.71.

Gac fruit aril powder and oil preparation

Fresh ripe gac fruit (_M. _cochinchinensis) was collected from a home garden and farm at Chay Uncle Garden, Surin Province, Thailand. The fruit was washed twice with running tap water and air-dried. Subsequently, the peel and pulp were cut open, and the seeds were removed to collect the red arils. The collected arils were divided into two portions. The first portion was freeze-dried using a Christ Beta 2–8 LDplus freeze dryer (Martin Christ, Osterode am Harz, Germany) at −80°C for 48 h. The dried gac sample was ground into a fine powder, passed through a 250-µm mesh sieve, and stored in aluminum-plastic zip bags at −20°C until use for feed formulation.

The second portion of gac arils was air-dried at 50°C overnight. The dried arils were processed using a cold screw oil press (Euro Best Technology, Pathum Thani, Thailand). Subsequently, the gac aril extract was frozen at −80°C, concentrated using a Rotavapor R-210 rotary evaporator (Büchi Labortechnik AG, Flawil, Switzerland), and lyophilized using a Christ Beta 2–8 LDplus freeze-dryer (Martin Christ, Osterode am Harz, Germany) to obtain a stable product. The extraction yield of gac oil was approximately 28.7%. The obtained GO was stored in amber reagent glass bottles at −20°C until use.

Experimental diets

Five isonitrogenous (427.5 g/kg crude protein), isolipidic (91.9 g/kg crude lipid), and isocaloric (18.81 MJ/kg) diets were formulated with graded levels of GP and GO to distinguish the effects of ingredient form and bioactive concentration. A control diet was formulated without GP or GO supplementation, whereas the other four diets included GP at 10 and 30 g/kg and GO at 5 and 10 g/kg, designated GP10, GP30, GO5, and GO10, respectively (Table 1). The dietary inclusion levels of gac products were selected based on a previous study in laying hens [26]. To achieve isonitrogenous diets, wheat flour and fish oil levels were adjusted, while GP and GO levels were increased.

All powdered ingredients were mixed for 15 min using a C-B20G-A1 kitchen mixer (CKI Family, Nonthaburi, Thailand), then fish oil and lecithin were added and mixed for 5 min. Before pelletizing, water was added at 350 mL/kg diet, and the mixture was homogenized for 10 min. The diets were then pelletized into approximately 2-mm pellets using a meat mincer (CKI Family). The experimental diets were air-dried in a UL50 hot-air oven (Memmert GmbH + Co. KG, Schwabach, Germany) at 50°C overnight. Subsequently, the diets were placed in aluminum-plastic zip bags to prevent light exposure and stored at −20°C until use.

Color determination

At the end of the experiment, three fish from each tank (n = 9) were randomly collected to determine coloration in three body regions: the head, near the upper eye region; the abdominal region, at the lateral line on the dorsal section; and the caudal fin, at the dorsal region of the caudal fin rays. Fish were anesthetized before color measurement. Color was measured on the left side of each fish using a MiniScan EZ 4500L spectrophoto-meter (HunterLab, Reston, VA, USA), with one measurement recorded per body region. The instrument was calibrated using standard white and black calibration plates before analysis. The L* value represents lightness (100 = white, 0 = black), a* indicates redness/greenness, and b* indicates yellowness/blueness, according to the guidelines of the International Commission on Illumination [27].

Growth performance, feed utilization, and survival

At week 10, fish were starved for 24 h and anesthetized with 30 ppm clove oil solution prepared as a 1:9 mixture of clove oil and ethanol. The average body weight per fish was determined by weighing all fish in each tank collectively and dividing the value by the number of surviving fish. Growth performance and feed utilization parameters were calculated as follows: weight gain (g/fish) = final body weight − initial body weight; specific growth rate (SGR) (%/day) = [(ln final body weight − ln initial body weight)/days] × 100; Feed conversion ratio (FCR) = individual feed intake (g)/individual weight gain (g); protein efficiency ratio = wet weight gain (g)/protein intake (g); and protein productive value (PPV) (%) = protein gain (g)/protein intake (g) × 100. Survival (%) was calculated as 100 × [(final number of fish)/(initial number of fish)].

Items | Control | GP10 | GP30 | GO5 | GO10
Feed formulation (g/kg) |  |  |  |  | 
Fish meal | 390 | 390 | 390 | 390 | 390
Soybean meal | 290 | 290 | 290 | 290 | 290
Gluten meal | 80 | 80 | 80 | 80 | 80
Squid mealᵃ | 40 | 40 | 40 | 40 | 40
Rice bran | 35 | 35 | 35 | 35 | 35
Corn meal | 36 | 36 | 36 | 36 | 36
Wheat flour | 40 | 30 | 10 | 40 | 40
Gac powder | 0 | 10 | 30 | 0 | 0
Gac oil extract | 0 | 0 | 0 | 5 | 10
Fish oilᵇ | 45 | 45 | 45 | 40 | 35
Lysine | 8 | 8 | 8 | 8 | 8
Methionine | 7 | 7 | 7 | 7 | 7
Lecithin | 3 | 3 | 3 | 3 | 3
Vitamin C | 6 | 6 | 6 | 6 | 6
Vitamin premixᶜ | 10 | 10 | 10 | 10 | 10
Mineral premixᵈ | 10 | 10 | 10 | 10 | 10
Total | 1000 | 1000 | 1000 | 1000 | 1000
Proximate composition |  |  |  |  | 
Crude protein | 429.0 | 424.1 | 428.5 | 425.9 | 430.3
Crude fat | 91.1 | 91.2 | 92.3 | 93.1 | 92.1
Ash | 132.5 | 137.6 | 138.4 | 134.8 | 133.3
Dry matter | 953.2 | 946.8 | 944.3 | 954.2 | 955.4
Total carbohydrates | 300.6 | 293.9 | 285.1 | 300.4 | 299.7
β-carotene (mg/kg)ᵉ | 0 | 0.98 | 2.95 | 1.94 | 5.83
Lycopene (mg/kg)ᶠ | 0 | 0.65 | 1.97 | 4.01 | 12.03
Total carotenoids (g/kg) | 0.28 | 0.46 | 0.71 | 0.48 | 0.66
Gross energy (MJ/kg)ᵍ | 18.89 | 18.66 | 18.66 | 18.89 | 18.94

Somatic indices

Three fish were randomly collected from each tank and anesthetized with an overdose of clove oil solution. Each fish was individually weighed and measured for total length to determine condition factor using the following formula: (final body weight × body length³) × 100. Fish were dissected, and the weights of the viscera and liver were recorded to calculate viscerosomatic index (VSI) and hepatosomatic index (HIS) using the following formula: (organ weight/final body weight) × 100.

Chemical and phytochemical compositions

The proximate composition of whole-body samples was analyzed from the initial fish (10 fish; data not shown), fish at termination (two fish from each tank), and the experimental diets following the standard methods of the Association of Official Analytical Chemists [28]. Dry matter content was determined using the UL50 hot-air oven (Memmert, Schwabach, Germany) at 105°C until a constant weight was achieved. Crude protein content was determined by the Kjeldahl method (N × 6.25) with H₂SO₄ digestion using a Vapodest semi-automatic Kjeldahl system (Gerhardt GmbH & Co. KG, Königswinter, Germany; method 984.13). Crude lipid content was analyzed using the petroleum ether extraction method (method 920.85) with a classic Soxhlet apparatus (Gerhardt). Ash content was determined by incineration at 550°C for 8 h.

Carotenoids, including β-carotene, lycopene, lutein, zeaxanthin, and astaxanthin, were extracted based on a previously described method [29] with slight modifications. Approximately 0.1 g of sample was homogenized in a mortar using a solvent mixture of n-hexane, ethanol, and acetone at a ratio of 1.50:0.75:0.75 until complete extraction. The total carotenoids were then mixed with 5 mL of distilled water and centrifuged at 1,500 × g for 10 min at 25°C. The supernatant was analyzed as described previously [30] using a Waters Carotenoids C30 column (4.6 × 150 mm; Waters Corporation, Milford, MA, USA) with an Agilent 1260 high-performance liquid chromate-graphy system (Agilent Technologies Inc., Santa Clara, CA, USA). A gradient elution program was applied using methanol (A) and MTBE (B) as the mobile phases as follows: 0–8 min, 0% B; 8–14 min, 0%–22% B; 14–24 min, 22% B; 24–29 min, 22%–40% B; and 29–33 min, 40% B. The mobile phase was delivered at a flow rate of 1.0 mL/min, with the column temperature maintained at 30°C. The injection volume was 20 µL. Chromatographic peaks were identified by comparing retention times and spectra with authenticated standards. Detection wavelengths were set at 450 nm for β-carotene, 470 nm for lycopene, 445 nm for lutein, 450 nm for zeaxanthin, and 480 nm for quantification of free astaxanthin and its esterified forms.

Total flavonoid content was determined using a colorimetric method [31]. The absorbance of the reaction mixture was measured at 415 nm. Results were expressed as QE/g DW. All chemical reagents and assay kits were purchased from Sigma-Aldrich (St. Louis, MO, USA).

Total carotenoid content

Total carotenoid content in tissues and experimental diets was determined following the method of Torrissen and Naevdal [32] with slight modifications. Tissue samples, including fin, muscle, skin, and liver, were collected from two fish per tank (n = 6), with approximately 1 g collected from each tissue. Tissue samples were homogenized in darkness in a glass tube with 5 mL of cold acetone containing 1 g of anhydrous sodium sulfate (Na₂SO₄). The extraction solution was covered with aluminum foil to prevent light exposure and maintain sensitivity, then kept at 4°C overnight. The extraction procedure was repeated with an additional 5 mL of extraction solution, bringing the total volume to 10 mL, until complete pigment exhaustion was achieved. The solution was centrifuged at 3500 × g for 5 min at 4°C, and the supernatant was analyzed spectrophotometrically at 450 nm using a UV-1800 spectrophotometer (Shimadzu Corporation, Kyoto, Japan). Acetone was used as the blank reference, and measurements were performed using a 1-cm path-length cuvette. Total carotenoid content (µg/g) was determined using the following equation:

Total carotenoids (µg/g) = A × V × 10⁴ / (2500 × W)

where A is the absorbance at 450 nm, V is the extraction volume (mL), and W is the sample weight (g).

Total carotenoid content in serum was determined as described by Barbosa et al. [33]. Serum (50 µL; see the “Blood sampling, hematology, and biochemistry analysis” section) was homogenized with ethanol and hexane at a ratio of 4:1 (v/v) in a glass tube covered with aluminum foil to prevent light exposure and maintain sensitivity. The extraction solution was briefly mixed and centrifuged at 4500 × g for 10 min at 4°C. The supernatant was analyzed spectrophotometrically at 450 nm, using ethanol and hexane at a 4:1 (v/v) ratio as the reference.

Blood sampling, hematology, and biochemical analysis

Blood samples were collected from the caudal vein at the same time of day, between 08:00 and 10:00, to minimize diurnal variation. Samples were pooled (n = 3) and collected using a 26-G needle and 1-mL sterile syringe (Nipro Corporation, Osaka, Japan). Fish were fasted for 24 h before sampling. The fish were anesthetized with a 20 ppm clove oil solution, and blood samples were divided into two portions: one containing ethylenediaminetetraacetic acid (EDTA) as an anticoagulant for hematological analysis, and the other without EDTA for biochemical analysis. The second portion was allowed to clot on ice for 2 h and centrifuged at 2,000 × g for 15 min at 4°C. Serum from the upper layer was collected and transferred into a fresh tube for serum biochemical analysis.

Hematological parameters, including red blood cell (RBC) count (×10⁶ cells/µL) and white blood cell (WBC) count (×10⁴ cells/µL), were determined using a Neubauer hemocytometer following Rawling et al. [34]. Hematocrit (Hct, %) was determined using the microhematocrit method with a DM1424 hematocrit centrifuge (DLAB Scientific Co., Ltd., Beijing, China). Hemoglobin (Hb, g/dL) was analyzed using the colorimetric method with Drabkin’s assay kit (Sigma-Aldrich), and absorbance was measured at 540 nm.

Serum protein content was determined using a commercial colorimetric kit (Sigma-Aldrich). Serum albumin content was determined using the bromocresol green binding method [35]. Serum globulin, alanine transaminase (ALT), aspartate transaminase (AST), alkaline phosphatase (ALP), total cholesterol, triglycerides, high-density lipoprotein cholesterol (HDL-c), and low-density lipoprotein cholesterol (LDL-c) were determined using commercial assay kits (Sichuan Maker Biotechnology Co., Ltd., Chengdu, China) [36] and quantified using a Cobas C311 automated analyzer (Roche Diagnostics, Rotkreuz, Switzerland). Globulin content was calculated using the following formula: total protein content − albumin content.

Expression of immunity-related genes

Total RNA extraction and complementary DNA (cDNA) synthesis: Approximately 2 g of liver tissue was collected for liver morphology analysis, whereas another portion of liver tissue was preserved in RNAlater (Amnion Life Sciences, Cambridgeshire, UK) at −20°C for subsequent total RNA extraction. Total RNA was extracted from liver tissue ground in liquid nitrogen using a pestle and mortar, and the resulting fine powder was dissolved in 1 mL of QIAzol Lysis Reagent (Qiagen, Hilden, Germany). Samples were purified using the RNeasy Mini Kit (Qiagen) according to the manufacturer’s instructions. To avoid genomic DNA contamination, total RNA was incubated at 37°C for 15 min with DNase I (Thermo Fisher Scientific, Waltham, MA, USA). Total RNA quality was quantified using the 260:280 nm ratio with a Synergy H1 Multi-Mode Reader (BioTek Instruments Inc., Winooski, VT, USA). In addition, total RNA quality was assessed using 2% agarose gel electrophoresis. Total RNA was reverse-transcribed into single-strand cDNA using the RevertAid First-Strand cDNA Synthesis Kit (Thermo Fisher Scientific) according to the manufacturer’s protocol.

Real-time quantitative polymerase chain reaction (qPCR) analysis: Gene-specific primer sequences for tumor necrosis factor-alpha (TNF-α), interleukin-(IL_)__-1_β, IL-10, lysozyme, and heat shock protein-70 (HSP-70) used in this study were adopted from previously validated studies in ornamental goldfish [23, 24] (Table 2). These primer sets were also used to determine amplification efficiencies, which were 98%, 105%, 95%, 91%, and 99%, respectively, whereas the amplification efficiency of the β-actin gene was 99%, based on standard calibration curves generated from cycle threshold values of serially diluted samples (R² > 0.91).

qPCR was conducted using the PCRmax ECO48 real-time qPCR system (PCRmax Ltd., Staffordshire, UK) in a final volume of 20 µL containing 1 µL of diluted (100×) cDNA template, 0.4 µL (10 µM) each of forward and reverse primers, 10 µL of Maxima SYBR Green/ROX qPCR Master Mix (Thermo Fisher Scientific), and 8.2 µL of nuclease-free water. The real-time reverse transcription polymerase chain reaction (RT-PCR) program consisted of initial denaturation at 95°C for 10 min, followed by 40 cycles of denaturation at 95°C for 5 s and annealing/extension at 59°C for 40 s. All real-time RT-PCR samples were analyzed in triplicate, resulting in nine replicates in total. Gene-specific primers targeting TNF-α, IL-1β, IL-10, lysozyme, and HSP-70 in goldfish were used, with expression levels normalized to β-actin and quantified using the 2^−ΔΔCt method.

Gene | Primer | Sequence (5′–3′) | Reference
TNF-α | Forward | CATTCCTACGGATGGCATTTACTT | [23]
 | Reverse | CCTCAGGAATGTCAGTCTTGCAT | 
IL-1β | Forward | GATGCGCTGCTCAGCTTCT | [23]
 | Reverse | AGTGGGTGCTACATTAACCATACG | 
IL-10 | Forward | CAAGGAGCTCCGTTCTGCAT | [23]
 | Reverse | TCGAGTAATGGTGCCAAGTCATCA | 
Lysozyme | Forward | GTATCTTCAAGCGAGAGGGACT | [24]
 | Reverse | CCCTGTGGGTCTTATACTTACTC | 
HSP-70 | Forward | GGCAGAAGGTGACAAATGCA | [23]
 | Reverse | TGGGCTCGTTGATGTTCTCA | 
β-actin | Forward | GATGCGGAAACTGGAAAGGG | [23]
 | Reverse | ATGAGGGCAGAGTGGTAGACG | 

Liver morphology

Liver tissues from two fish per tank (n = 6) were fixed in cold, approximately 4°C, 10% neutral buffered formalin [10% (v/v) formalin containing 9 g/L NaCl and 12 g/L Na₂HPO₄, adjusted to pH 7.2–7.4]. The liver from each fish was trimmed into three pieces and processed for histological examination. After fixation, tissue samples were dehydrated in a graded ethanol series, embedded in Paraplast medium (Leica Microsystems, Nussloch, Germany), and cut into 5–6-µm sections using a Leica RM2235 rotary microtome (Leica Biosystems, Nussloch, Germany). Tissue sections were stained with hematoxylin for 3 min, rinsed, differentiated, and counterstained with eosin for 1 min before dehydration and mounting. Sections were examined and imaged using an Olympus BX40 light microscope (Olympus Corporation, Tokyo, Japan). Histopathological alterations were semi-quantitatively scored for lesion severity on a scale of 0–3, where 0 = normal, 1 = mild, 2 = moderate, and 3 = severe.

Statistical analysis

The SPSS statistical software package (SPSS Inc., Chicago, IL, USA) was used for statistical analyses. Before analysis, data normality and homogeneity of variance were assessed using the Shapiro–Wilk and Levene’s tests, respectively. The effects of dietary supplementation with GP and GO on all measured variables were evaluated using a one-way analysis of variance. Significant treatment effects (p < 0.05) were further analyzed using Tukey’s post hoc test to compare means among experimental groups. Results are presented as mean ± SD. The sample size was determined based on previous ornamental fish feeding studies with similar experimental designs and measured parameters [23], while also considering animal welfare and experimental feasibility.