How Does Pikeperch Sander lucioperca Respond to Dietary Insect Meal Hermetia illucens? Investigation on Gut Microbiota, Histomorphology, and Antioxidant Biomarkers

Effects of feeding dietary defatted black soldier fly (Hermetia illucens) larvae meal (HI) on intestine microbiota, and on histomorphology, oxidative enzyme activities in liver and intestine of pikeperch (Sander lucioperca) were investigated. Four isoproteic (45% crude protein) and isolipidic (18% ether extract) diets were formulated to include 0% (CO), 9% (HI9), 18% (HI18) and 36% (HI36) of HI as replacement for fishmeal at 0, 25, 50, and 100%, respectively, and were fed to triplicate groups of juvenile pikeperch (initial body weight, 68.7 ± 7.1 g) for 84 days. No adverse effects were detected on the intestine of pikeperch fed diet groups, in terms of histomorphology (P > 0.05), while fish fed free or low levels of HI (≤ 9% in diet) showed significant liver degeneration (P < 0.05). Dietary HI significantly affected the oxidative enzyme activities of catalase and glutathione peroxidase in the liver, and glutathione S-transferase in the intestine (P < 0.05), while activity of superoxide dismutase in both liver and intestine was HI-dose independent (P > 0.05). Feeding HI-containing diets positively modulated the richness and diversity of intestinal microbiota, especially for HI18 group (P < 0.05). Inclusion HI up to 18% (50% fishmeal replacement) in pikeperch diets increased abundance of Clostridium, Oceanobacillus, Bacteroides, and Faecalibacterium genera, whereas the predominant bacterium, Cetobacterium was found in control and HI36 groups. This study reveals the potential of HI as an immune and health booster for juvenile pikeperch.


INTRODUCTION
Aquaculture is the largest global consumer of fishmeal production, accounting for 68-73% (Shepherd and Jackson, 2013;Tacon and Metian, 2015). Fishmeal is mainly derived from marine capture fisheries (70% in 2018) (FAO, 2020a), which has reached a plateau since the 2000s (Shepherd and Jackson, 2013) and has been projected that the ecological limits of stock will be reached by 2037 (Froehlich et al., 2018a). Therefore, the current fastest growth of aquaculture in food-producing sectors (FAO, 2020a) and the continuous increasing trend, requires the development of novel aquafeed ingredients. Terrestrial crops have been used in aquafeeds more than other alternatives until recent (Tacon et al., 2011;Tacon and Metian, 2015) and, by 2050, the use of these feedstuffs in aquaculture will rise to twice the current level in a business-as-usual scenario, reaching 91 million tonnes (Froehlich et al., 2018b). However, crop-based feeds for aquatic animals introduce concerns regarding their nutritional properties and environmental consequences. An unbalanced essential amino acid profile, low palatability, and the presence of anti-nutritional substances could impair their inclusion in aquafeeds (Gatlin et al., 2007). Moreover, the expansion and intensification of the production of terrestrial crops will lead to tremendous environmental burdens pertaining to climate change, biodiversity loss, and increasing demand for arable land and water. Among such burdens, land use is considered the one that entails the greatest pressures on the planet (Foley et al., 2005(Foley et al., , 2011Boissy et al., 2011). Beyond terrestrial plant ingredients, fishery by-products and insect meals have shown the greatest potential to be protein-supplied to aquafeeds in the coming years (Hua et al., 2019;Gasco et al., 2020a). Although approximately 34% of the world's fishmeal production will be derived from fish by-products by 2030 (FAO, 2020a), this potential protein source will still not be able to meet the projected aquafeed demand by 2050 (Froehlich et al., 2018a). The efficiency of insect meal as a future aquafeed ingredient has already been identified, especially concerning the feasibility of costs, scalability, and processing technology (Hua et al., 2019). Globally, insect production is on the rise, and will reach approximately 1.2 million tonnes by 2025 and become price-competitive with fishmeal by 2023 (Hua et al., 2019;Gasco et al., 2020a). In addition, the development of production facilities and processing techniques would help to improve the environmental performance of insect meal as a sustainable aquafeed ingredient (van Huis and Oonincx, 2017). The use of seven insect species (two flies, two mealworms, and three cricket species) in fish diets has been authorised by the European Commission (Regulation No. 2017/893). Among these species, black soldier fly (Hemertia illucens), which belongs to the Diptera order, has received the most research interest (Hua, 2021). Hemertia illucens larvae meal possesses important nutritional profiles, especially amino acid profile which is close to that of fishmeal (Nogales-Mérida et al., 2019). As far as environmental impact is concerned, H. illucens production, if obtained using non-valorised substrates, entails significantly less arable land and water use than soybean meal (Smetana et al., 2019;Gasco et al., 2020b). Moreover, H. illucens meal-containing diets have shown lower environmental impacts associated with abiotic depletion, acidification potential, eutrophication potential, climate change, human toxicity potential, and marine aquatic ecotoxicity potential for Arctic char (Salvelinus alpinus) (Smárason et al., 2017) and lower water use for European perch (Perca fluviatilis) (Stejskal et al., 2020) than insectfree diets.
Pikeperch (Sander lucioperca) is one of the main percid species that has drawn a great deal of attention in aquaculture (Schulz et al., 2006). Aquaculture production of pikeperch reached 1557 tonnes in 2018, which was doubled that of 2009 (750 tonnes) (FAO, 2020b), and has mainly been established in intensive recirculation systems (Dalsgaard et al., 2013). However, pikeperch and other percid fish have so far received very little attention from feed manufacturers (Bochert, 2020). Although some commercial aquafeeds for percids have become available, salmonids-targeted feeds are more widely used in practice (Stejskal et al., 2016). Since European pikeperch aquaculture is moving toward an established freshwater aquaculture sector (Policar et al., 2019), it will be necessary to develop suitable and sustainable feeds for aforementioned sector. Dietary protein requirements of at least 43% have been reported for appropriate growth performance and feed utilization of pikeperch fingerling (Nyina-Wamwiza et al., 2005). In the nature, aquatic insects, i.e., larvae of lake flies (Chironomidae) (Diptera order), play an important role as food sources for the early ontogenetic stages of pikeperch (Vinni et al., 2009;Ginter et al., 2011;Kashinskaya et al., 2018;Huuskonen et al., 2019). Therefore, the use of H. illuces larvae meal has been hypothesised to be suitable for pikeperch aquaculture. The aim of present study is to investigate the effects of dietary defatted black soldier fly (H. illucens) (HI) on the diets of juvenile pikeperch (S. lucioperca) on intestinal microbiota, histomorphology, and oxidative enzyme activities. The outputs could provide information in the choice of an alternative aquafeed ingredient for the emerging percid farming industry in Europe.

Ethics Statement
The experimental procedures were performed under European Communities Directive (No. 2010/63/EU) on the protection of animals used for scientific purposes and have been approved by the Czech Ministry of Health (MSMT-6744/2018-2).
The prepared diets were fed to triplicate groups of juvenile pikeperch (initial body weight 68.7 ± 7.1 g, with 50 individuals per tank) for 84 days. A combined feeding protocol of four meals per day, provided at 07.00, 09.00, 11.00, 13.00, by automatic feeders (EHEIM Twins, Deizisau, Germany), and one hand feeding, at 15.00 was adopted during the trial. Any unconsumed feeds were collected by siphoning and dried in an oven to calculate the exact feed intake.

Fish Biometry
At the start and the end of the feeding trial, fish were individually weighed to calculate weight gain (WG) and feed conversion ratio (FCR): WG (g) = final body weight-initial body weight FCR = total feed supplied (g, Dry Matter)/WG

Antioxidative Enzyme and Histo-Morphological Analysis
After 84 days of the experiment, a total of 45 fish (3 individuals/tank) were randomly sampled, after 24 h of feed deprivation, and were euthanised by means of overdose anaesthesia (MS222, 125 mg/l). Dissected livers and intestines from 15 fish/group were stored at −80 • C for further antioxidative enzyme analysis. A similar number of samples, taken from another 15 fish/group, were fixed by immersion in a 10% buffered formalin solution for histo-morphological analysis.

Intestinal Microbiota
At the end of the experiment, three fish were randomly taken from each tank and euthanised by means of overdose anaesthesia (MS222, 125 mg/l). In order to ensure that all sampled fish had digesta throughout the intestinal tract, fish were deprived of feeds 12 h prior to sampling time. Fish exterior was wiped with 70% ethanol before abdomen was opened, whole intestine from each fish was removed from the abdominal cavity and digesta from proximal to distal intestine was squeezed gently into a 1.5 ml aseptic Eppendorf and immediately stored at −80 • C for further analysis.

Diet Chemical Composition
Analysis of HI defatted meal and experimental diets for dry matter, crude protein, crude lipid, ash, and fatty acids (FAs) were performed as described elsewhere (Tran et al., 2021). Gross energy was determined by mean of a calorimetric bomb (IKA C7000, Stufen, Germany).

Oxidative Stress in Livers and Intestines
Oxidative stress biomarkers were evaluated in liver and intestine of each fish sample by means of spectrophotometer analysis (Varian Cary spectrophotometer, Santa Clara, CA, United States) as previously described by Elia et al. (2018). Briefly, superoxide dismutase (SOD) activity was measured in 50 mM Na 2 CO 3 , pH 10, 0.1 mM EDTA, 500 mM cytochrome C, and 1 mM hypoxanthine and xanthine oxidase. Reduction of cytochrome C by the xanthine/hypoxanthine system was measured versus a standard curve of SOD units at 550 nm. Catalase (CAT) activity was measured as the decrease in absorbance at 240 nm due to the consumption of H 2 O 2 . The assay was performed in an NaH 2 PO 4 + Na 2 HPO 4 buffer (100 mM, pH 7) and 12 mM H 2 O 2 . Glutathione peroxidase (SeGPx's) activities were measured by following the oxidation of NADPH at 340 nm and using 0.6 mM H 2 O 2 or 0.8 mM cumene hydroperoxides (tot GPx) as substrates. Glutathione S-transferase (GST) was measured at 340 nm using as a substrate 1-chloro-2,4-dinitrobenzene (CDNB).

Histo-Morphological Analysis of Intestine and Liver
Samples of the anterior intestine were excised and flushed with a 0.9% saline solution to remove all the content. The collected samples were fixed in a 10% buffered formalin solution, routinely embedded in paraffin wax blocks, sectioned at a 5 µm thickness, mounted onto glass slides and stained with Haematoxylin & Eosin (HE). One slide per intestinal segment was examined by means of light microscopy and captured with a Nikon DS-Fi1 digital camera, coupled to a Zeiss Axiophot microscope, using a 2.5× objective lens. NIS-Elements F software was used to capture images.
Morphometric analysis was performed using Image R -Pro Plus software on ten well-oriented and intact villi. The evaluated morphometric indices were villi height (from the villus tip to submucosa) and villi width (across the base of the villus, but not including the brush border).
The observed histopathological findings were evaluated in all the organs, using a semi-quantitative scoring system as follows: absent (score = 0), mild (score = 1), moderate (score = 2), and severe (score = 3). Histopathological findings in intestine were assessed separately for each segment for mucosa (inflammatory infiltrates) and submucosa [inflammatory infiltrates and Gut-Associated Lymphoid Tissue (GALT) activation]. The total score of each gut segment was obtained by adding to the mucosa and submucosa scores. All the slides were blind assessed by two independent observers, and any discordant cases were reexamined, using a multi-head microscope, until unanimous consensus was reached.

DNA Extraction and 16S rRNA Amplicon Target Sequencing
Nucleic acid was extracted from the intestine content (500 mg as starting materials). Total DNA from the samples was extracted using a RNeasy Power Microbiome KIT (Qiagen, Milan, Italy), according to the manufacturer's instructions. One microlitre of RNase (Illumina Inc, San Diego, CA, United States) was added to *Only FAs > 10 mg/g total FAs (except for EPA) are presented; Different letters denote significant differences among the experimental groups (P < 0.05).
Frontiers in Marine Science | www.frontiersin.org digest the RNA in the DNA samples for an incubation period of 1 h at 37 • C. DNA was quantified using Qubit ds and standardised at 5 ng/µ l. DNA extracted directly from digesta samples was used to assess the microbiota, through amplification of the V3-V4 region of the 16S rRNA gene (Klindworth et al., 2012). The PCR products were purified according to the Illumina metagenomic standard procedure (Illumina Inc, San Diego, CA, United States). Sequencing was performed with an MiSeq Illumina instrument, with V3 chemistry, and 250 bp paired-end reads were generated according to the manufacturer's instructions.

Statistical Analysis
All data for antioxidative enzyme activities were tested for homogeneity of variance using Cochran, Hartley, Bartlett test. The effects of diet on oxidative stress in different organs were analysed separately, by means of one-way ANOVA, followed by Tukey test. Statistical analyses were performed using STATISTICA 12.0, with P-value < 0.05 as the significant difference.
Raw reads of microbiota were first joined, after sequencing, using FLASH software (Magoč and Salzberg, 2011), with default parameters, and were filtered, using QIIME 1.9.0 software and the pipeline as recently described . Briefly, shorter reads (<300 bp) were discarded, using Prinseq. USEARCH software (version 8.1) was used for chimera filtering, and the Operational Taxonomic Units (OTUs) were picked, at a threshold of 97% similarity, using UCLUST algorithms. Taxonomy was assigned against 16S rRNA from Greengenes. The OTU table was rarefied at 10,144 sequences/sample. The OTU table displays the highest taxonomy resolution that was reached. When the taxonomy assignment was not able to reach the genus level, the family or phyla were displayed. R software was used to calculate the alpha diversity, while Weighted and Unweighted UniFrac distance matrix and OTUs table were used to find differences between samples, using permutational multivariate analysis of variance (Anosim) and analysis of similarity (Adonis) statistical test, considering the same function in R environment. Pairwise Wilcoxon test were used to determine any significant differences in alpha diversity or OTU abundance as a function of dietary insect meal. Principal component analysis (PCA) were plotted, using the dudi.pca function, through the made4 package of R environment. Non-normally distributed variables were presented as median values (interquartile range, IR), and box plots represented the interquartile range between the first and the third quartile, with the error bars showing the lowest and the highest value. Pairwise Kruskal-Wallis tests were used to find any significant differences in microbial taxa abundance according to the dietary treatment. P-values were adjusted for multiple testing, and a false discovery rate (FDR) < 0.05 was considered as significant. The data generated from sequencing were deposited in the NCBI Sequence Read Archive (SRA) and are available under the BioProject Accession Number PRJNA704237.
GraphPad Prism R software (version 8.0) was used to perform statistical analysis, for histo-morphometrical investigations. The Shapiro-Wilk test was used to test the normality of the data distribution before statistical analyses. Data were described by mean and standard deviation (SD), or median and IR depending on data distribution. Bivariate analysis was performed, by means of one way-ANOVA or Kruskall Wallis tests, to compare the intestine morphology and organs histopathology among different diet groups. P-values < 0.05 were considered statistically significant.

Oxidative Stress in Liver and Intestine
The results of oxidative biomarkers, SOD, CAT, SeGPx, and GST, in liver and intestine of pikeperch fed experimental diets are depicted in Figure 1. Dietary HI did not alter the SOD activities in either liver or intestine, CAT activities in liver, SeGPx activities in intestine, or GST activities in liver of pikeperch (P > 0.05). No significant difference was observed across experimental groups (P > 0.05) for liver, as regards CAT activities, whereas this biomarker was significantly lower in HI18 and HI36 than in HI9 (P < 0.05), but remained similar to CO (P > 0.05) in intestine. Even if did not differ from the CO group, among fish fed HIcontaining diets, HI9 produced highest SeGPx activity in liver (P < 0.05), while the lowest activity was found in HI36 group (P < 0.05). A significant increase in the GST concentration was observed in intestine of pikeperch fed HI-containing diets, compared to CO (P < 0.05). Of the different insect-fed groups, HI9 showed a higher GST than HI18 (P < 0.05), while HI36 was remained intermediate position.

Histo-Morphology
Data regarding histopathological evaluation are reported in Table 3. Only few differences were observed for morphometry and histopathology of intestine among diet groups. Although there was no significant difference, a trend could be observed (P = 0.065) with HI36 group recording wider villi than the other groups. Thus, dietary HI inclusion did not induce any significant morphological changes in the pikeperch intestine, thereby suggesting no negative influence of such dietary HI on the physiological development of intestine.
Mild to severe multifocal to diffuse liver vacuolar degeneration was recorded in all treatments, and it was found to be greater in CO and HI9 group than in the HI18 and HI36 ones. Dietary HI did not show any evidence of inflammation of the liver of pikeperch (Table 3 and Figure 2).

Microbiota
The total number of high-quality paired-end sequences obtained from 16S rRNA sequencing reached 1.916.822 raw reads. After the filtering, 1.295.693 reads passed the filters applied by QIIME, with a median value of 37.559 ± 15.565 reads/sample, and a mean sequence length of 443 bp. The rarefaction analysis and Good's coverage, expressed as a median percentage (97%), also indicated satisfactory coverage of all samples.
The result of the OTUs analysis showed that there was no significant difference in Shannon index (P > 0.05) among diet groups, while alpha-diversity of intestinal bacteria, associated with Chao1 and observed OTUs, in fish fed HI18 significantly increased relative to CO diet (P < 0.05) (Figure 3).
Adonis and Anosim statistical tests, based on weighted and on unweighted UniFrac distance matrix using the OTUs table, showed significant differences between diet groups as a administration of HI (P < 0.002). These differences were also observed when the PCA plot was produced at a genus level (Figure 4). It was also possible to observe a certain degree of separation, following diet groups. Microbiota of CO diet was near to the insect meal inclusion of 9%, while the microbiota of fish fed with 18 and 36% of HI was well separated (Figure 4).
The dominant OTUs, at the phyla level, were Firmicutes (mean values, 45-75%), regardless to dietary HI. Perch fed CO diet was enriched with Proteobacteria (26%), while Bacteroidetes (7-13%) was the prevalent phyla in fish fed HIcontaining diets. As a result, Clostridiaceae, Enterococcaceae, and Bacillaceae were found to be the predominant families across fish fed diet groups. Clostridium, Acetobacter, Cetobacterium, Plesiomonas, Acetobacter, Peptostreptococcaceae, Bacteroides, and Oceanobacillus were, at the genus level, the most abundant genera found in intestine of perch considered in our study (Figure 5).
Dietary HI positively affected relative abundance of almost OTUs, compared with CO (FDR < 0.05), excepted for Bacillus, Burkholderia, and Sporosarcina, which were dominant in the CO group (Figure 6).

Oxidative Enzymes
Reactive oxygen species (ROS) is the production of aerobic metabolism processes, including superoxide, hydrogen peroxide, and lipid peroxides (Buetler et al., 2004). Excessive ROS compounds cause cellular and tissue damages (Rosa et al., 2008). The balance of ROS production ensures the normal physical function of any organism and is regulated by antioxidant systems (Rosa et al., 2008) involving two mechanisms, (i) enzymes that remove ROS, including SOD, CAT, and SeGPx; and (ii) antioxidative compounds, i.e., ascorbate, glutathione, scavenge free radicals (Passi et al., 2002). Antioxidative enzyme activities were documented to be tissue-specific in pikeperch, and liver was the most sensitive organ to the diet manipulation under recirculating aquaculture system . In the case of detoxification in the intestine, however, certain enzymes such as SOD were known to play a vital role (Tang et al., 2013). This study indicates that in liver of pikeperch dietary HI did not alter the SOD, CAT, or GST oxidative enzymes, while significantly reduced SeGPx activity, a result that is in agreement with those of previous study (Elia et al., 2018), who performed a trial on rainbow trout fed dietary HI. The significant reduction in the catalytic SeGPx efficiency in liver of pikeperch fed dietary HI could be explained by the presence of chitin (Elia et al., 2018). Indeed, increasing inclusion levels of HI increased chitin levels in diets (Table 1). In addition, declining in SeGPx activities, as a result of increasing dietary HI, could be attributed to different dietary PUFA levels ( Table 2), which are highly susceptible to oxidation. In fact, Tocher et al. (2002) reported that a high dietary PUFA content increased lipid peroxidation in fish tissues, and consequently the SeGPx enzyme activity involved in reducing peroxides, including FA hydroperoxides and hydrogen peroxide, will be also high (Passi et al., 2002).
The present study indicates that the CAT activity in intestine of pikeperch was significantly higher for HI9 than Frontiers in Marine Science | www.frontiersin.org for HI18 and HI36 groups. A similar phenomenon was reported for CAT activity in the intestine of rainbow trout fed insect meal (T. molitor), where a substitution level of 25% fishmeal displayed higher activity than the 50% level (Henry et al., 2018a). The CAT and SeGPx activities in the present study were similar for CO and HI9, and lower than for HI18, HI36 groups. This result indicates that substantial substitution of fishmeal with HI reduced antioxidant enzyme activities in pikeperch. This is in line with a previous finding pertaining to rainbow trout (O. mykiss) (Elia et al., 2018). The decline of these biomarkers in HI18 and HI36 groups could be related to an imbalance between ROS production and antioxidant capacity. A suitable concentration of antioxidants, such as chitin and other bioactive compounds (Ngo and Kim, 2014), may support antioxidant enzyme activities in HI9 compared to the other HI-contained diets (Henry et al., 2018a).
Glutathione S-transferase plays an essential role in scavenging free radicals and xenobiotics detoxification (Aksnes and Njaa, 1981;Li et al., 2010). Increased glutathione S-transferase activity in intestine, but not liver, was observed across diet groups in the present study (Figure 1), thus implicating that some of the compounds in HI may have stimulated the biotransformation pathway in intestine of pikeperch, which was also found in liver of tilapia (Oreochromis niloticus) fed cricket-based feeds (Ogunji et al., 2007). In fact, insect meals may contain harmful substances, i.e., heavy metals and pesticides (van der Spiegel et al., 2013). The absence of an alteration of the hepatic GST activities after administration of HI could be the result of factors other than xenobiotics (Collier and Varanasi, 1991) or tissue-specific response (Martínez-Álvarez et al., 2005).
We also observed numerically higher oxidative biomarkers in liver of pikeperch than in intestine (Figure 1), which was in agreement with recent findings , reporting that liver was one of the most susceptible tissue in response to artificial nutrition and controlled conditions.

Histo-Morphology
Dietary HI in our study did not induce any morphological or inflammatory changes in the intestine of pikeperch, a result that is in agreement with previous studies conducted on different fish species fed dietary insect meals (Elia et al., 2018;Zarantoniello et al., 2019;Zarantoniello et al., 2020a). The absence of intestinal and hepatic inflammation could be linked to anti-inflammatory properties regulated by dietary saturated fatty acids content, especially lauric acid (C12:0) and chitin FIGURE 3 | Boxplots showing the alpha diversity rarefaction index across experimental diets. The boxes represent the interquartile range (IR) between the first and third quartiles, and the line inside the boxes represents the median. Whiskers denote the lowest and the highest values within 1.56 IR from the first and third quartiles, respectively. Circles represent outliers beyond the whiskers. Different superscripts within boxplot indicate significant differences. component (Henry et al., 2018b;Vargas-Abundez et al., 2019;Zarantoniello et al., 2019;Gasco et al., 2020b,c) which were found to be particularly high in HI and HI-containing diets in the present study. Although there were no significant differences (at P-value < 0.05), the villi were more expanded in the HI36 group than in the other groups (Table 3), and this was attributed to the presence of chitin. Chitin could stimulate the growth of villi thickness in tilapia (O. niloticus), probably due to its viscosity and water holding capacity (Kihara and Sakata, 1997). Chitin also induced the production of short-chain fatty acids, such as acetate, propionate and n-butyrate, and n-butyrate in particular was observed in intestine of tilapia (Kihara and Sakata, 1997), thereby increasing intestinal histo-morphology of fish, e.g., villi length and weight (Dawood, 2021). The large quantity of Paenibacillus genus in intestinal digesta of fish fed HI36 (Figure 6) could act as a probiotic for aquatic animal species (Midhun et al., 2017;Chen et al., 2019;Amoah et al., 2020), consequently enhancing intestinal health indices, including histomorphology (Dawood, 2021).
In contrast to recent findings, which reported that an increasing inclusion of insect meals induced a higher degree of hepatic vacuolization degeneration in fish (Li et al., 2017;Zarantoniello et al., 2019), the present study indicates that feeding pikeperch with ≤ 9% HI caused more severe hepatic degeneration than 18 or 36% did (Table 3), which could be related to a fatty liver status. Schulz et al. (2005) reported that a low level of palmitic acid (C16:0) yielded a higher hepatic lipid content. In the present study, the significantly lower palmitic acid in the control group than in the HI-containing groups could partly explain the hepatocellular vacuolization phenomenon. The mechanism to which palmitic acid affecting hepatic tissues remained to be elucidated. However, this FA promotes hepatocyte proliferation (Wang et al., 2011) and possess anti-inflammatory and antiviral effects (Librán-Pérez et al., 2019). On the other hand, the high content of dietary lauric acid (C12:0), high oxidation and low tissue deposition, was found to decrease liver lipid storage in freshwater Atlantic salmon (Belghit et al., 2019). This could explain the reduction in the adipose liver in pikeperch fed HI18 and HI36, compared to the control and HI9 diets ( Table 3). Two FAs, linoleic and oleic acids, were confirmed to induce the occurrence of hepatic steatosis in sea bream (Sparus aurata) (Caballero et al., 2004). Moreover, owing to large molecular weight, oleic acid could produce a large lipid droplet while inrush hepatocyte (Bradbury, 2006).These FAs were found to  Only bacteria with an overall abundance of ≥ 1% and ≥ 0.5% at phylum and family/genus level, respectively, were presented. The bacteria were pool as "Others," when lower than aforementioned abundance. be significantly higher in CO than in HI18, HI36 (Table 2), which could indicate severe steatosis in livers of the former group (Table 3). High intakes of eicosapentaenoic acid (EPA) and docosahexaenoic acid (DHA) are known to an inhibitor of lipid accumulation in livers of sea bream (S. aurata) (Caballero et al., 2004). Therefore, the change in the percentage of the different FAs in the experimental diets, due to the inclusion of HI, could further explain the severity of hepatic vacuolization degeneration observed in perch fed CO and H9 diets.

Microbiota
The present study reveals that dietary HI enhanced microbial biodiversity indices in intestine of pikeperch, compared with insect-free diet, a result that is in line with recent findings on rainbow trout (Bruni et al., 2018;Huyben et al., 2019;Rimoldi et al., 2019;Terova et al., 2019), thereby contributing to gut health and health status of the host.
In agreement with previous studies on intestinal microbiota of percid fish and freshwater species, the present study reveals that Firmicutes, Proteobacteria, Bacteroidetes were the most dominant phyla in the intestine of pikeperch, regardless of the HI inclusion level (Li et al., 2014;Kashinskaya et al., 2018;Terova et al., 2019).
Our results show an abundance of Clostridium genus in fish fed HI9 and HI18, which was even greater than those fed CO and HI36. Members of the Clostridium genus are common effective microorganism used as probiotics in aquaculture (Nayak, 2010a,b). Clostridium butyricum has been shown to possess a pathogenic inhibition capacity in farmed fishes (Pan et al., 2008a,b;Gao et al., 2013), improve feed efficiency in shrimp (Duan et al., 2017;Li et al., 2019), and to be suitable for use as probiotics in farmed fish (Hai, 2015;Zorriehzahra et al., 2016). The greater prevalence of Clostridium and other probiotic-used bacteria in HI9, such as Lactobaccillus and Bacillus genera, than in HI36, could explain the difference in feed conversion ratio between these diets in present study. The Bacteroides and Clostridium genera are known to be the main taxa involved in production of fatty acids and vitamins (Balcázar et al., 2006). The abundant presence of these taxa could partially compensate for nutritional insufficiencies in HIcontaining diets, and consequently resulted in a comparable growth rate among control, HI9 and HI18 diets, yet the offset may be not efficient for HI36 group.
It is worth noting that Cetobacterium, the most predominant bacterium in intestine of natural pikeperch (Kashinskaya et al., 2018) and other freshwater fish (Larsen et al., 2014), was detected in our captive pikeperch fed dietary HI. Similar findings were also observed in rainbow trout (Etyemez and Balcázar, 2015), common carp (van Kessel et al., 2011), and giant arapaima (Ramírez et al., 2018) fed commercial aquafeeds. It seems relevant that Cetobacterium is among the core bacteria in pikeperch.
Insect meal, in general, is a chitin-rich ingredient. The degradation and digestion of this substance require binary enzymes, including chitinase and β−N−acetylglucosaminidase, and involve various microbacteria derived from digestive tract of fish with a chitinase-produced capacity (Ray et al., 2012;Ringø et al., 2012). Among these chitin-degraded bacteria, the Plesiomonas and Bacillus genus were detected across treatment groups at a particularly low abundance (Figure 4). This finding implicates that pikeperch may not be able to degrade chitin. A limited presence of chitinase-producing bacteria was also observed in rainbow trout (Bruni et al., 2018;Huyben et al., 2019;Rimoldi et al., 2019;Terova et al., 2019) and this may help to explain the low or absent chitin digestibility in this species (St-Hilaire et al., 2007;Henry et al., 2015;Renna et al., 2017;Caimi et al., 2020).
In conclusion, HI, fed as a partial or total replacement of fishmeal did not induce any inflammation of liver or intestine, or any intestine degeneration, but did show signs of severe hepatic steatosis of pikeperch fed CO and HI9 groups. Dietary HI promotes antioxidative enzyme activities of CAT, GPx and GST, but not of SOD, in liver and, to a lesser extent, in intestine of pikeperch. The inclusion of HI up to 18% or 50% fishmeal replacement in pikeperch diets increased abundance of Clostridium, Oceanobacillus, Bacteroides, and Faecalibacterium, whereas the predominant bacterium, Cetobacterium was found in the control and HI36 groups. Because of the absence of inflammation in tissues, the evolution of antioxidative enzyme, and modification of the favourable microbiota observed in the present study, it is possible to assume that defatted HI could have an immunological effect on juvenile pikeperch. Further study on immune response and disease resistance of pikeperch fed insect meal could help to explore this efficiency.

DATA AVAILABILITY STATEMENT
The original contributions presented in the study are included in the article, further inquiries can be directed to the corresponding author.

ETHICS STATEMENT
The animal study was reviewed and approved by the Czech Ministry of Health (MSMT-6744/2018-2).

AUTHOR CONTRIBUTIONS
LG and VS: planning the experiment and editing the manuscript. HT: data analysis, writing, and editing of the manuscript. MP, MZ, and TG: wrote the manuscript. AE, EC, IF, CC, and FG: analysis and the first draft. All authors contributed to the article and approved the submitted version.

FUNDING
The study was financially supported by the Ministry of Agriculture of the Czech Republic, project NAZV (QK1810296). This JU-IAPW research manuscript is part of a project that has received funding from the European Union's Horizon 2020 Research and Innovation Programme under grant agreement No. 652831 (AQUAEXCEL2020), the TNA project ID number: AE070026.