Introduction

The fish farming industry plays a significant role in meeting the growing global demand for seafood (FAO 2022). The expansion of fish farming is heavily dependent on the quality of feed produced. Traditionally, fishmeal and fish oil have been crucial components in aquafeeds, providing protein, and energy, respectively. However, fishmeal and fish oil cannot meet the need of aquafeeds for various reasons such as limited availability, impact on marine ecosystems, and rising costs. Consequently, the dependence of fishmeal and fish oil for the aquafeed industry is a global debate. Moreover, the sustainable growth of aquaculture is directly dependent on high-quality feed formulated according to the specific requirements of each species (Tacon and Metian, 2015; FAO 2022). Protein source is of great importance in the feed of all carnivorous farmed fish species, such as rainbow trout (Oncorhynchus mykiss). As one of the most widely cultured fish species, rainbow trout farming holds great economic and environmental importance (Tacon and Metian, 2008; FAO 2019). There is a growing interest in exploring alternative protein sources to enhance production efficiency sustainably for this species and reduce reliance on traditional feed ingredients (Alam et al. 2018; Carlberg et al. 2018; Biswas et al. 2020; Ruiz et al. 2023). However, the quality of protein sources utilized in aquafeeds also significantly affects the quality of fish flesh (Zhang et al. 2022; Yao et al. 2022; Xu et al. 2022). Vegetable and animal protein sources have been extensively researched in carnivorous fish feeds with high economic value such as rainbow trout (Aksnes et al. 2006; Asadi et al. 2021; Richard et al. 2021). However, advancements in biochemical technology have made it feasible to incorporate novel protein sources into aquafeeds. Recently, feather meal has emerged as a notable animal protein source in the formulation of aquafeeds (Poppi et al. 2011; Campos et al. 2017; Psofakis et al. 2020; Fornari et al. 2023).

Hydrolyzed feather meal (HFM) is a rich protein source containing between 74 and 91% crude protein (CP) (Bureau et al. 2000; Grazziotin et al. 2008; Campos et al. 2017; Davies et al. 2018), rich in cystine (4–5% of CP), but deficient in lysine (2% of CP) and methionine (1% of CP) (Baker et al. 1981; Klemesrud et al. 2000). The low lysine and methionine and high cystine content of HFM, compared to fishmeal (NRC 2011), may cause an imbalance and reduce the dietary energy available for fish growth (Klemesrud et al. 1997; Wang and Parsons 1998). Feather meal is produced through the hydrolysis of poultry feathers, resulting in a protein-rich feed ingredient that could potentially replace traditional fishmeal sources in rainbow trout diets. Moreover, the nutrient availability of HFM can vary according to the rendering process applied to the feathers (Lee et al. 2022; Buse et al. 2022; Asadnezhad et al. 2023). However, the utilization of feather meal in aquafeeds requires a comprehensive understanding of its effects on growth performance and physiology in rainbow trout (Bureau et al. 2000; Poppi et al. 2011; Glencross 2011).

Research on FM and HFM utilization in a limited number of species shows the substitution levels for fishmeal can be implemented without significant adverse effects on growth performance and digestibility (Davies et al. 2018; Cao et al. 2020; Psofakis et al. 2020; Damir and Yıldız, 2022). Studies conducted in species with carnivorous feeding habits, such as Japanese flounder (Paralichthys olivaceus), European seabass (Dicentrarchus labrax), gilthead sea bream (Sparus aurata), and turbot (Scophthalmus maximus) have shown that HFM replacement levels vary depending on the species. Several studies (Pfeffer et al. 1994; Steffens 1994; Pfeuti et al. 2019a, b) have examined various parameters in rainbow trout, incorporating feather meal alongside diverse animal protein sources (blood meal and poultry by-product meal), aside from fishmeal. However, there are only two studies that focused on the growth performance of juvenile rainbow trout with the use of both feather meal and fishmeal into their diets, which reported that up to 30% of fishmeal can be replaced with HFM in their feeds (Bureau et al. 2000; Glencross 2011).

More recently, there are some scientific opinions regarding the inclusion of probiotics in feeds to enhance the digestibility of protein sources utilized as substitutes for fishmeal in diets designed for various cultured fish species. However, the number of studies on the subject is very limited (Wang et al. 2008; He et al. 2022; Del Valle et al. 2023). Previous studies have reported the beneficial effects associated with dietary inclusion of probiotics on fish, such as enriching the digestive system flora, increasing growth, feed utilization, digestibility, disease resistance, innate immune responses, intestinal microbiota, and gut health (González-Félix et al. 2018; Puvanendran et al. 2021; Hany et al. 2022). Recently, probiotics have been added to diets along with alternative feed ingredients to enhance their effective utilization by fish (Balcazar et al. 2007; Guidoli et al. 2018; Silva et al. 2021). However, research on the use of probiotics in fish feeds with alternative ingredients is quite limited (Nadanasabesan et al. 2022; Zan et al. 2023; Abdel-Latif et al. 2023). Only one study has reported increased feed utilization through the application of probiotics in Totoaba (Totoaba macdonaldi) diets with feather meal (González-Félix et al. 2018). However, no research has been found in rainbow trout feeding on the subject. In our study, we aimed to increase the digestibility of feather meal and increase its replacement levels by juvenile rainbow trout by adding probiotics to diets.

The growth performance of rainbow trout is a crucial indicator of feed efficiency and overall productivity in aquaculture systems. Evaluating the impact of feather meal incorporation on growth performance parameters such as weight gain, FCR, and specific growth rate is essential to determine its suitability as a dietary protein source (Campos et al. 2017; Davies et al. 2018; Cao et al. 2020; Psofakis et al. 2020). Additionally, the examination of gene expression patterns in rainbow trout can provide valuable insights into the molecular mechanisms underlying the physiological responses to feather meal-based diets (Psofakis et al. 2020; Damir and Yıldız, 2022).

Hence, the present study aimed to investigate the growth performance, feed utilization, digestibility, intestinal histology, and gene expression of juvenile rainbow trout fed diets in which HFM replaced fishmeal at the levels of 30, 35, 40, and 45% with the addition of essential amino acids (lysine and methionine) and 0.2% probiotics (Lactobacillus plantarum and Bacillus subtilis). In addition, the effects of the diets were investigated on the proximate and whole-body amino acid composition of fish.

Materials and methods

Experimental diets

Five iso-nitrogenous (approximately 50% crude protein), iso-energetic (approximately 16 kJ/g), and iso-lipidic (approximately 22% crude lipid) diets (Table 1) (2–4-mm diameter) were prepared at the Sapanca Inland Fisheries Production Research and Application Unit of the Faculty of Aquatic Sciences, Istanbul University, Türkiye. The control diet used in the study consisted of fishmeal, corn gluten, wheat gluten, and wheat bran as protein sources. Hydrolyzed feather meal (HFM) supplemented with probiotics (Lactobacillus plantarum and Bacillus subtilis) (2 g/kg diet) and some essential amino acids (lysine and methionine) replaced 30% (HFM30), 35% (HFM35), 40% (HFM40), and 45% (HFM45) of dietary fishmeal in the remaining diets (Table 2). After the feed ingredients were mixed homogeneously, approximately 30% water was added and turned into dough. This dough was passed through a meat grinder and turned into pellets and dried in a cold storage environment with the aid of a fan. Prior to 2 weeks of feces collection, 1% titanium dioxide (TiO2) was added in diets as an indicator to determine digestibility. Diets were kept in plastic bags at − 20 °C until they were used.

Table 1 Feed formulation and proximate compositions of diets
Full size table
Table 2 Amino acid compositions of the protein sources and dietary treatments used in the study (g per 100 g of crude protein)
Full size table

Experiment conditions

The feeding trial was conducted with juvenile rainbow trout, of average initial weight 29.65 ± 0.39 g, and length 14 ± 0.15 cm, from Sapanca Inland Fisheries Production Research and Application Unit. In the study, a total of 600 fish were randomly distributed into 15 cylindrical tanks (40 fish per tank) with capacity of 1000 L, with 3 replicates for each experimental group. Experimental fish were acclimated to experimental conditions 2 weeks prior to the study. During this time, they were fed with commercial rainbow trout feed. The water parameters were maintained at a temperature of 13.9 ± 0.3 °C, dissolved oxygen of 7.7 ± 0.1 mg/L, and pH of 7.8 throughout the experiment. During the experiment, 12 h light–12 h dark photoperiod regime was utilized. Filtered water from a well was used in the experimental tanks. The daily feed requirements for the fish were calculated based on their live weights, ranging from 2 to 3% of their body weight, taking into account water temperature and fish size. The fish were manually fed twice per day (10 a.m. and 5 p.m.) for 75 days, during the feeding trials.

Sampling

At the beginning and end of the experiment, all fish in the tanks were weighed for weight calculations; five fish randomly selected from each experimental tank were weighed to determine their total weight, liver weight, and weight of internal organs, in order to calculate the viscerosomatic and hepatosomatic index values of the fish. After individual measurements of size and weight for all fish, five fish were randomly selected from each tank and stored at − 80 °C freezer for proximate composition and amino acid analysis. For intestinal histology analysis, intestinal tissues of five fish from each tank were taken and preserved in 10% formaldehyde solution. For gene expression analysis, stomach tissues of five fish from each tank were taken and stored at − 80 °C freezer. A lethal dose of anesthetic substance (2.5 g Benzaquin + 10 mL acetone + 10 L water) was added to water in a container to perform these measurements. Subsequently, the fish collected from the experimental tanks were placed into this container to induce euthanasia. Additionally, fish feces were collected for the purpose of calculating amino acid metabolism of the fish. The fish were fed with TiO2 supplemented feeds and their feces were collected 2 weeks before the end of the experiment. The fecal matter at the bottom of each tank was collected by siphoning and dried using a plankton net, then stored in suitable containers and stored at − 80 °C prior to analyses.

Growth performance

Growth performance of juvenile rainbow trout for the duration of feeding trial was evaluated with the parameters listed below:

  • Weight gain (WG) = final body weight − initial body weight

  • Specific growth rate (SGR, %) = 100 × [ln(final body weight) − ln(initial body weight)] × (days).−1

  • Feed conversion ratio (FCR) = feed consumed (g) × wet weight gain.−1 (g)

  • Protein efficiency ratio (PER) = weight gain (g) × protein intake.−1 (g)

  • Hepatosomatic index (HSΙ, %) = (weight of liver / weight of fish) × 100

  • Viscerosomatic index (VSI, %) = (weight of viscera / weight of fish) × 100

  • Nitrogen free extract (NFE, %) = % dry matter − (% crude protein + % crude lipid + % ash + % fiber)

  • Metabolizable energy (ME) = [3.9 × (% crude protein) + 8.0 × (% crude lipid) + 1.6 × (% nitrogen free extract)] × 10/1000 × 4.186

  • Gross energy (GE) = [5.65 × (% crude protein) + 9.45 × (% crude lipid) + 4.10 × (% nitrogen free extract)] × 10/1000 × 4.186

Proximate composition and amino acid analysis

The proximate composition of the diets and whole fish samples (for protein, lipid, ash, and moisture) and also fiber contents of the diets were analyzed by the methodology of AOAC (2006). Crude protein (N × 6.25) was analyzed using a semi-automatic Kjeldahl technique (Gerhardt Vapodest, 45 s). Lipid was measured by ether extraction using Velp Scientifica Ser, 148 (Gerhardt, Germany). Moisture content was measured by drying samples in an oven (Nüve FN 400, TS 6073) at 105 °C for 12 h. Ash was measured by incinerating samples at 550 °C for about 12 h in a Muffle Furnace (Gerhardt, Germany). Fiber was determined by using sulfuric acid and then sodium hydroxide, 12.5% (w/w) for half an hour each, with the final residue washed with 5% HCl and water, filtered, dried, and weighed. Amino acid was extracted from the whole body of fish, feed, and feces samples. Analyses were performed using a high-sensitivity, high-speed triple quadrupole mass spectrometer method for the liquid chromatography mass spectrometer system (LCMS/8050). Ten milliliter of petroleum ether was added to 1 g sample, and the oil was vortexed for 2 min. Then, the samples were centrifuged for 3 min, and the upper petroleum ether phase was removed. Twenty-five milliliter of 6N HCl solution was added to the sample, then poured into a 250-mL Schott bottle. Then, Schott bottles were placed into an oven preheated to 110 °C. The bottles were left open for an hour in the oven. After 1 h, the bottles were closed and sealed to continue hydrolysis for 23 h. At the end of 24 h, samples were filtered using a 45-µm filter and transferred to a 15-mL falcon tube. The samples were diluted 1500 times before placing them on an LCMS-MS device to finish the AA analysis. All AA analyses were done in triplicate.

Histological analysis

For the histological analysis, ≈1.5-cm-long tissue segments were taken from the midpoint of the intestine from three fish per tank and flushed with a saline solution. These segments were fixed in 10% formalin at room temperature for 72 h. Samples were fixed in paraffin wax after routine histological procedures. Sections with thickness of 4 µm were taken on the glass slide using a microtome and then stained with periodic acid-Schiff. Images were taken with a computer‐supported imaging system (ImageJ, U.S. National Institutes of Health, Bethesda, MD) connected to a light microscope (Olympus Optical). Villus length (VL), villus width (VW), crypt depth (CD), villus length to crypt depth ratio (VL:CD), and villus surface area were measured. Villus width was measured in the middle of each villus. Villus length was measured from the top of the villus to the point where the villus connects with the crypts, and the crypt depth was measured from its base up to the region of transition between the crypt and villus. The surface area of the villus was measured using the following formula (Solis de los Santos et al. 2005; Ceylan et al. 2023):

$$(2\mathrm\pi)\:\times\:(\mathrm{villus} \ \mathrm{width}/2)\:\times\:(\mathrm{villus} \ \mathrm{length})/10^6$$

Digestibility analysis

Fish, feed, and feces samples were crushed with mortar (0.25-mm screen) for digestibility analysis. Titanium concentration of diets and ileal digesta was determined by a modified protocol outlined by Short et al. (1996). For this, 0.5 g of each sample was ashed for 12 h at 580 °C. After titrating ashed samples with 10 mL of sulfuric acid (7.4 M), samples were boiled at 200 °C for 2 h until fully dissolved. The solutions were titrated with 20 mL of 30% hydrogen peroxide and brought to 100 mL using distilled water after cooling. The absorbance was determined by a spectrophotometer (UV‐1200, Shimadzu) at 410 nm after keeping for 48 h at room temperature. The determined values were used for the calculation of the utilization of various nutrients. The apparent digestibility coefficients (ADC) of crude protein, amino acid, and crude fat were calculated using the following equation (Cho et al. 1982; Krontveit et al. 2014):

$$\mathrm{ADC}\;(\%)=100-\;\left[100\times(\%\;\mathrm{indicator}\;\mathrm{in}\;\mathrm{feed}/\%\;\mathrm{indicator}\;\mathrm{in}\;\mathrm{feces})\times(\%\;\mathrm{nutrient}\;\mathrm{in}\;\mathrm{feces}/\%\;\mathrm{nutrient}\;\mathrm{in}\;\mathrm{feed})\right]$$

Real-time PCR analysis

The expression of the growth-related genes was determined by real-time PCR analysis. Approximately 10 mg of fish stomach tissues from each experimental group was used to extract the total RNA using the kit (Qiagen, USA). Nanodrop 1000 (Thermo Fisher Scientific, Inc USA, LightCycler® 480 System, Roche) was used to specify the purity of RNA. cDNA was synthesized from 2 µg of RNA (20 µL reaction volume) using a high-capacity RNA to cDNA reverse transcriptase kit (PE Applied Biosystems). The primer sequences used in the research are presented in Table 3. Each primer set was used at 200 nM concentration from a volume of 14 µL using Power SYBR Green PCR Master Mix (PE Applied Biosystems). The glyceraldehyde-3-phosphate dehydrogenase (GAPDH) (GenBank accession no. NM001124209.1) of rainbow trout was used as a housekeeping gene. Gene expression primers were validated using the BLAST sequence alignment algorithm, and specific PCR products were produced to determine the absence of primer-dimerization that could affect the analysis. The house‐keeping gene GADPH was utilized as an internal reference to normalize the expression of the other genes according to Livak and Schmittgen (2001).

Table 3 Forward and reverse primer list for RT-qPCR analysis
Full size table

Statistical analysis

Results were expressed as means ± SD. All the statistical analyses were made using SPSS statistical software (SPSS 23). The normal distribution of data was assessed using Shapiro–Wilk tests. Percentage data were arcsine-transformed prior to analysis. After confirmation of normality and homogeneity of variance, data were subjected to one-way ANOVA, and subsequent comparison of means was performed using Tukey’s (P < 0.05) multiple range test.

Results

Growth performance and proximate composition

After 75 days of feeding, survival rates were 100% in all treatments. The growth performance and proximate composition results are given in Table 4. Relationship between HFM replacement level in the diets and final weight of fish are shown in Fig. 1. The growth performance of the fish fed the control diet was higher than the other experimental groups (P < 0.05). The growth performance of the HFM30, HFM35 and HFM40 groups was similar between them (P > 0.05) and lower than the control group (P < 0.05). However, the FCR values of the control group were higher than the other experimental groups (P < 0.05). The growth performance of fish fed with the HFM45 diet was found at the lowest level (P < 0.05). HSI values were similar in the control and HFM30 groups (P > 0.05) and higher than the other groups (P < 0.05). VSI values were highest in HFM30 and lowest in HFM35 and HFM45 groups (P < 0.05). Whole body crude protein was highest in fish fed HFM40 (P < 0.05), whereas no significant differences (P > 0.05) were observed among the other groups. There were no significant differences on dry matter and crude lipid values for treatment groups (P > 0.05). The ash levels of fish fed HFM45 diet were the lowest compared to the other dietary groups (P < 0.05).

Table 4 Growth performance and proximate composition of juvenile rainbow trout groups fed with experimental diets
Full size table
Fig. 1
figure 1

Relationship between HFM replacement level (% of FM) in the diets and final weight (g) of juvenile rainbow trout. a–c Letters denote the statistical significance of fish final weight (P < 0.05)

Full size image

Whole-body amino acid profile

Whole body EAA and NEAA content of rainbow trout are shown in Table 5. Among the trial groups, EAAs of whole body, histidine, and lysine were found lowest in the HFM45 group (P < 0.05). Lysine and methionine levels were highest in control, and lowest in the HFM45 (P < 0.05). Phenylalanine was found higher in HFM45 than control and other experimental groups (P < 0.05). At the end of the 75-day feeding experiment, the total EAAs was highest in fish fed with the control diet and was lowest in HFM45 (P < 0.05). The total EAAs of HFM30, HFM35, and HFM40 were higher than initial (P < 0.05).

Table 5 Whole-body amino acid compositions of juvenile rainbow trout fed with experimental diets (g/100 g protein)1
Full size table

Feed analysis and apparent digestibility

Nutrient and amino acid apparent digestibility analyses are shown in Table 6. The relationship between HFM replacement level in the diets and protein digestibility of fish is shown in Fig. 2. Crude protein digestibility in control, HFM30, HFM35, and HFM40 groups was similar. There was no significant difference in energy digestibility in all experimental groups (P > 0.05). There were no significant differences between the levels of EAA digestibility in the control and other experimental groups (P > 0.05). However, isoleucine digestibility was significantly lowest in the HFM45 group (P < 0.05). Moreover, there is no significant difference between the EAA digestibility in all experimental groups (P > 0.05); however, it was observed that the EAA digestibility in the HFM45 group was lower than other experimental groups (P < 0.05).

Table 6 Apparent digestibility coefficients (%) of nutrient and amino acids of diets for rainbow trout
Full size table
Fig. 2
figure 2

Relationship between HFM replacement level (% of FM) in the diets and protein digestibility (%) of juvenile rainbow trout. a, bLetters denote the statistical difference of protein digestibility (P < 0.05)

Full size image

Histological analyses

Mucosa and sub‐mucosa layers of experimental fish showed general tissue integrity (Fig. 3), with no signs of tissue necrosis. The effects of dietary treatments on intestinal morphology are presented in Table 7. Villus morphologic parameters were significantly impaired by dietary inclusion of feather meal (P < 0.05) so that villus length to crypt depth decreased by 24.33, 35.27, 28.62, and 46.01% by increasing dietary feather meal from 30 to 45% compared to fish fed on the control diet. Again, villus surface area decreased by 19.44, 33.33, 44.44, and 44.44% in fish fed on different levels of dietary feather meal (HFM30, HFM35, HFM40, and HFM45, respectively) compared to the control (P < 0.05).

Fig. 3
figure 3

Villus morphology of mucosal tissues in fish fed on dietary treatments. VL, villus length; VW, villus width; CD, crypt depth. Control, fishmeal-based diet; HFM30, 30% hydrolyzed feather meal (HFM) + 70% fishmeal (FM); HFM35, 35% HFM + 65% FM; HFM40, 40% HFM + 60% FM; HFM45, 45% HFM + 55% FM

Full size image
Table 7 Effect of experimental diets on intestinal morphometry in rainbow trout
Full size table

Gene expression

Results of relative gene expression of growth-related genes in juvenile rainbow trout fed with different experimental diets are shown in Table 8. Expressions of all genes were significantly influenced by the dietary treatments. Higher expressions of IGFBP-1b, IGFr1a, GDF9, and FGF2 were observed in fish fed HFM45 diet (P < 0.05). All the experiment diets had lower GHR-I and IGF-I than the control group (P < 0.05). However, IGF-II was higher in all experiment diets compared to the control group (P < 0.05). HFM45 had significantly higher IGFBP-1b than the control (P < 0.05), while there were no significant differences observed between HFM30 and HFM35 (P > 0.05). GDF9 was highest in HFM45 (P > 0.05), followed by HFM35 which had significant difference from other groups (HFM40 > HFM30 > Control). HFM40 had a higher ERα than the control and other experimental groups (P < 0.05). HFM45 had highest FGF2, while HFM40 had the lowest among the other groups (P < 0.05).

Table 8 Relative gene expression in rainbow trout fed the experimental diets
Full size table

Discussion

Research on the use of HFM in aquafeeds is recent and limited, with notably scarce literature on its incorporation in rainbow trout diets (Bureau et al. 2000; Glencross 2011). In the conducted study, replacement of fish meal with 30, 35, and 40% HFM caused a slight decrease in growth performance compared to the control group, while the HFM45 group had the lowest growth performance. The proximate composition and EAA levels of the prepared diets met the amounts needed for juvenile rainbow trout (NRC 2011). The FCR values (1.02–1.04) of the HFM30–HFM40 groups were similar. However, feed intake of fish in the HFM45 group decreased, and their FCR values increased. Bureau et al. (2000) fed juvenile rainbow trout weighing 17 g using 24 to 40% HFM (without any supplementations such as amino acids and probiotics) as replacement of fishmeal in their diets for 8 weeks. In the study, it was reported that a maximum of 30% HFM (1.44) replacement can be used without affecting the growth and feed efficiency compared to the control (1.42) group. For carnivorous species such as Japanese flounder (Kikuchi et al. 1994), sea bream (Davies et al. 2018; Psofakis et al. 2020; Damir and Yıldız, 2022), and turbot (Cao et al. 2020), the maximum HFM replacement in diets without adversely affecting growth and feed utilization has been reported to be 25%, 25%, and 16%, respectively. The feed utilization and growth of fish in our study showed that it has better results than previous studies on the different carnivorous species fed with HFM containing diets. The use of probiotics also ensured that the whole-body proximate compositions of juvenile rainbow trout were at the optimum level for the fish, as in Niu et al. (2019), Darafsh et al. (2019), and Amit et al. (2021).

A limited growth effect of dietary lysine and methionine supplementation to diets containing HFM has been reported in other carnivorous species such as European seabass and gilthead seabream (Kikuchi et al. 1994; Pfeffer et al. 1994; Psofakis et al. 2020). In another study, Psofakis et al. (2020) stated that while up to 25% HFM replacement with the supplementations of lysine and methionine to sea bream diets improves growth performance and feed utilization, at a level of 50%, dietary HFM inclusion with the EAA supplementations decreases these parameters. In our findings, up to 40% HFM replacement with the EAA and probiotic supplementations can be used without adversely affecting growth and feed utilization of juvenile rainbow trout. As there was no prior investigation into the incorporation of HFM with probiotic supplementation in rainbow trout diets, our study’s results were documented as pioneering research in this regard. It has been reported that the use of probiotics in rainbow trout feeds coupled with alternative protein sources improves feed utilization and growth performance in fish (Panigrahi et al. 2007; Bagheri et al. 2008; Merrifield et al. 2010). Merrifield et al. (2010) reported that dietary inclusion of different probiotic species and quantities increased the growth performance and feed utilization in rainbow trout. In the present study, we report that when the amount of probiotics in the diet was doubled, the growth and feed utilization also significantly increased. We are of the view that the level of HFM can be increased further by increasing probiotic levels. In another study, Bagheri et al. (2008) reported that Bacillus subtilis with Bacillus licheniformis probiotic species added to the diets of rainbow trout fry at different levels (1.2 × 109–6.1 × 109 CFU) significantly improved the growth performance and feed utilization of the fish. Panigrahi et al. (2007) reported that the cytokine genes, immune resistance, and response of juvenile rainbow trout increased as a result of feeding Lactobacillus rhamnosus, Enterococcus faecium, and Bacillus subtilis probiotic mixtures (109 CFU g feed−1) added to the diets. In the present study, growth performance and feed utilization of juvenile rainbow trout were significantly improved as a result of including probiotic mixture (2 g/kg diet) consisting of Lactobacillus plantarum with Bacillus subtilis species in their diets (Table 4). There is an agreement between the growth performance and feed utilization findings of our study and the results reported in earlier studies (Panigrahi et al. 2007; Bagheri et al. 2008; Merrifield et al. 2010).

There was no significant difference in protein apparent digestibility coefficient (ADC) between the control group and HFM30–HFM40 groups. Although protein digestibility decreased slightly in the HFM45 group, it was found to be within acceptable limits for rainbow trout. These results indicated that the inclusion of 45% HFM with limiting amino acid and probiotic supplementation to fishmeal-based diets in juvenile rainbow trout did not present a significant negative effect on protein digestibility (87.59–91.82%). Bureau et al. (2000) reported 75% protein digestibility as a result of using 30% feather meal for juvenile rainbow trout. Glencross (2011) reported 84.80% protein and 82.50% energy digestibility as a result of replacing 30% fishmeal with HFM in rainbow trout with 291 g initial weight. Both studies stated that these percentages are acceptable for rainbow trout. Similarly, as a result of the use of alternative protein sources in juvenile rainbow trout diets, 85.30% to 93.60 protein digestibility rates have been reported to be acceptable limits in different studies (Gomes et al. 1995; Bureau et al. 1999; Palmegiano et al. 2006). In our research, protein digestibility in the HFM45 group was calculated as 87.59%. Thus, in this research presented, it is seen that it is appropriate to include 45% HFM in juvenile rainbow trout diets. The dietary energy and lipid balance were not affected (89.9–92.3%) by the HFM level in the present study. These are above the range (82–88%) reported in earlier studies for this species with fishmeal-based diets (Peres and Oliva-Teles, 1999, 2005; Santos et al. 2010). Our findings indicated that inclusion of HFM in diets up to 45% did not adversely affect protein and energy digestibility. Although a previous study (Campos et al. 2017) on the use of HFM reported that amino acid digestibility was reduced in European seabass, our study showed that the supplementation of probiotics to the diet improved amino acid digestibility. With the additions of EAAs and probiotics, we aimed to improve growth performance, feed utilization, and digestibility for juvenile rainbow trout to increase replacement level of HFM in diets. Amino acid analysis of HFM indicated that lysine and methionine levels were lower than the requirements of juvenile rainbow trout (Table 2). Therefore, lysine and methionine supplements were added to diets. Our findings showed that HFM can successfully replace fishmeal in diets at 45% without compromising the amino acid digestibility of fish. We think that the main factors contributing to the high digestibility of amino acids in juvenile rainbow trout in our study are the inclusion of probiotics and essential amino EAAs in the diets. It is seen that more than 45% HFM can be used instead of fishmeal in diets by increasing probiotic levels. It has been reported that probiotics used in diets improve the intestinal flora of fish and therefore increase the digestibility of feed (Panigrahi et al. 2005; 2007; Merrifield et al. 2010; González-Félix et al. 2018).

The intestinal histological analyses of the study indicated that feeding juvenile rainbow trout with diets containing up to 40% HFM can replace fishmeal without any negative effects. However, some negativities were observed in the data regarding CD, VW, and CD to VW in the intestinal morphometry of fish in the HFM45 group. We think that it is possible to improve morphometric parameters in the intestines by adding higher amounts and/or different species of probiotics to diets (Panigrahi et al. 2005; 2007; Merrifield et al. 2010; González-Félix et al. 2018; Adelina et al. 2021). The findings of the current study showed some similarities with a recent study (Adelina et al. 2021) on juvenile silver pompano (Trachinotus blochii) with an omnivorous feeding habit. It has been reported that replacing fishmeal with 50% HFM fermented with Bacillus subtilis probiotic in the diets of juvenile silver pompanos did not damage the intestinal tissue; when the HFM level is increased to 75%, it damaged the intestinal tissue. Merrifield et al. (2010) fed rainbow trout weighing 100 g with Bacillus licheniformis + B. subtilis, E. faecium, Pediococcus acidilactici (each containing log 7 CFUg−1) probiotics for 5 weeks and indicated improvement in proximal intestinal and distal intestinal mucosa. González-Félix et al. (2018) reported that feeding totoaba (Totoaba macdonaldi) with probiotics belonging to the Bacillus genus resulted with improvements in intestinal histology compared to diets not containing probiotics. Additionally, the authors stated that maintaining a healthy gut microbiota can positively affect the intestinal epithelial structure of fish, restricting the presence of pathogens that produce virulence factors such as extracellular enzymes and pore-forming toxins within the gastrointestinal tract. This can reduce mucosal damage and maintain the integrity of the intestinal epithelium, which in turn improve the absorptive surface. Similarly, Peredo et al. (2015) reported that feeding Oreochromis niloticus with GroBiotic®-A added to the diet revealed improvements in digestive system histology and especially reported that the use of probiotics significantly increased microvilli height. Besides, enzymes added into feed to improve the digestion of nutrients in fish have been reported to cause improvements in intestinal histology (Abo‐Norag et al. 2018; Goda et al. 2020). In our literature review, we could not find any study that investigated the effect of sole or probiotic supplemented HFM on intestinal morphology. However, there are some studies (Ramos et al. 2015; 2017; Gaffar et al. 2023) investigating the effect of probiotic-containing (Bacillus subtilis, B. polymyxa, B. licheniformis, B. pumilus, B. coagulans, B. amyloliquefaciens, B. megaterium, Lactobacillus buchneri, L. plantarum, Aspergillus niger, A. oryzae, Rhodococcus spp., Rhodobacter spp., Nitrosomonus, Nitrobacter, Pedicoccus sp., P. acidilactici, Enterococcus sp.,) diets on intestinal morphology for different species. Ramos et al. (2015) reported that by adding two commercial probiotics at two concentrations in the diet, the villus length in intestinal morphology of rainbow trout fed with these diets was higher compared to fish fed with diets not containing probiotics. Similarly, when three different probiotic mixtures were added to Ganges mystus (Mystus cavasius) diets, it was reported that there was an increase in the villus length, width, area, and crypt depth parameters compared to the probiotic-free diet group (Gaffar et al. 2023). In another study on Nile tilapia (Ramos et al. 2017), one commercial multi-species probiotic was used at two different concentrations in a diet containing high amounts of plant protein sources. It was stated that the height of the villi in the intestinal morphology of fish fed these diets increased significantly compared to the probiotic-free group. When research on the effects of probiotics in fish intestinal morphology are examined, increases in villus length, height, width, surface area, and crypt depth parameters are characterized as improvement. Similarly, in the present study, villus length, width, surface area, and crypt depth parameters in the intestines of fish fed with diets supplemented with probiotics showed integrity. However, this integrity was observed to be less in the intestines of fish fed the HFM45 diet than in other groups. These results indicate that HFM can be used in diets at 40% replacement rate without any problems, but when used at a rate of 45%, it causes low-level damage to intestinal morphology.

Growth-related genes in fish are significantly influenced by the dietary treatments. TOR, GHR-1, and IGFs play a key role in feed intake and regulating fish growth (Pierce et al. 2012; Psofakis et al. 2020; Irm et al. 2020; Zhang et al. 2022). In the present study, the findings showed that as HFM level increased, TOR, IGF-II, IGFBP-1b, and IGFr1a increased, but GHR-1 and IGF-I decreased. Psofakis et al. (2020) observed downregulation in these growth-related genes (IGF-I, IGF-II, and IGFBP) when levels of HFM increased from 25 to 50% in the juvenile sea bream diets. Irm et al. (2020) also stated a downregulation in TOR and IGF-I genes when increasing amounts of HFM were used in the diet of juvenile black sea bream (Acanthopagrus schlegelii). However, in the present study, the gene expression results of the control and experiment groups showed upregulation. We think that the reason for the downregulation observed in other researches (Psofakis et al. 2020; Irm et al. 2020) but did not occur in our research may be due to the lysine, methionine levels, and probiotics added to the diets. When our findings are generally examined, growth-related genes of all HFM groups, especially TOR, IGF-II, IGFBP-1b, and IGFr1a genes, indicated better results than the control group. ERα participates in the regulation of the immune response in mammals and teleosts (Iwanowicz et al. 2014). The results in our study revealed that treatment groups had no immune problems (Erα, 0.75–1.71). Moreover, the immune response (Erα, 1.71) of HFM40 enhanced with the treatment. Result of the immune-related gene indicated that 40% replacement level of HFM is optimum for fish.

Although similar studies on juvenile rainbow trout (Bureau et al. 2000; Glencross 2011) have shown that the HFM replacement level is limited to 30%, the present study indicated that supplementation of HFM with lysine, methionine, and probiotics increased this level up to 40%. Moreover, these supplementations have been revealed to improve growth performance, feed intake, and digestibility in juvenile rainbow trout. In addition, the replacement of HFM up to 40% increased disease resistance and immune system and had no harm on histology in the fish.

In conclusion, HFM was found to be an important protein source for juvenile rainbow trout diets. However, future research should investigate the additional ratios and/or species of probiotics needed to increase the HFM replacement rate in the fish diets above 40%.