Fish meal (FM) is the gold dietary protein in aquafeeds (Tacon and Metian, 2015; Ytrestøyl et al., 2015), but the global increase in aquaculture production needs to be supported by alternative feed ingredients, which very often, have a negative impact on growth, intestinal health, and immuno-competence of farmed marine fish (Conceição et al., 2012; Krogdahl et al., 2015; Estensoro et al., 2016; Aragão et al., 2020). Traditionally, plant proteins have been considered as the most obvious FM alternative, but high levels of replacement can induce different signs of enteritis, including the shortening of mucosal folds, thickening of the lamina propia and submucosa, and infiltration of the distal intestine by inflammatory cells (Urán et al., 2009; Romarheim et al., 2013, Booman et al., 2018). These drawback effects are differentially exacerbated in each farmed fish species, but the supplementation of plant-based diets with fish protein hydrolysates or short-chain fatty acids (SCFAs) helps to mitigate most gut health detrimental effects in salmonid and non-salmonid fish (Estensoro et al., 2016; Piazzon et al., 2017; Egerton et al., 2020). However, novel ingredients and formulations need to be investigated to produce improved, efficient, and sustainable fish aquafeeds.

Seaweed polysaccharides have the capacity to regulate the non-specific host immunity, and supplementation of practical diets, with Gracilaria sp. improving the resistance of European sea bass (Dicentrarchus labrax) to Photobacterium damselae (Peixoto et al., 2019). In the same line, antioxidant extracts of Gracilaria protected against oxidative stress, improving the stress resilience of gilthead sea bream during acute hypoxia (Magnoni et al., 2017). Moreover, the supplementation of FM-free diets with marine diatomeas (Phaeodactylum tricornutum) stimulated the immune system of gilthead sea bream, which might be relevant as a prophylactic measure before predictable stressful events (Reis et al., 2021). Additionally, algae or algae-based products are currently used to replace or reduce the inclusion level of fish oil (FO) in aquafeeds, assuring the dietary supply of n-3 polyunsaturated fatty acids to meet the nutritional requirements of fish, as well as the healthy value of seafood products for human consumption (Maldonado-Othón et al., 2020; Pereira et al., 2020).

Processed animal proteins (PAPs), such as poultry meal, blood meal, and feather meal, are also used as a replacement for FM in practical fish diets (Wu et al., 2018; Ferrer Llagostera et al., 2019; Solé-Jiménez et al., 2021). Likewise, insect meal products, such as black soldier fly meal and defatted Tenebrio meal, did not compromise gut health and fish performance (Welker et al., 2007; Sánchez-Muros et al., 2016; Magalhães et al., 2017; Basto et al., 2021) and in some cases, even improved performance. Microbial biomasses have also been proposed as an effective replacement for marine feedstuffs in fish and shrimp diets (Delamare-Deboutteville et al., 2019; Simon et al., 2019, 2020). Such new and emerging ingredients have, in fact, been on the agenda of formulators for some time, but they are still available in low amounts, and their costs are often too high in comparison with most conventional feed ingredients. Moreover, many of these ingredients have been tested on a one-by-one basis rather than with different formulation combinations, although by this way important advances were achieved. Thus, a recent study on gilthead sea bream pointed out that sustainable PAP- and non-PAP-based feed formulations are able to support high growth rates when theoretical nutrient requirements are met by the diet (Fernandes et al., 2021). However, the use of PAP-based feed formulations was related to a slight impairment in feed conversion ratio (FCR) in association with a progressive down-regulated expression of hepatic insulin-like growth factor-I, a well-known marker of growth performance in fish, gilthead sea bream in particular (Pérez-Sánchez et al., 2018). Additionally, fish fed with the highest amount of PAP feed ingredients showed a pro-inflammatory status, supported by the up-regulated expression in the head kidney of several cytokines (il1β, tnfα), chemokines (ck8), and T-cell markers (cd3ζ, cd4-1, cd8α).

In this study, we focused on gut health indicators using samples from the study of Fernandes et al. (2021) to address the main effects of PAP- and non-PAP feed formulations on the composition of mucosal adherent bacteria from the anterior intestine (AI) in association with changes in the host transcriptomic profile of the intestine liver, and head kidney. Over the course of the last 5 years, the number of studies on fish, gilthead sea bream in particular, addressing the nutritional regulation of gut microbiota has increased exponentially (e.g., Piazzon et al., 2017, 2020; Rimoldi et al., 2019, 2020; Firmino et al., 2021; Moroni et al., 2021; Pelusio et al., 2021; Solé-Jiménez et al., 2021). Although our study did not include probiotics, previous studies on fish supported the gut-to-brain communication, and interestingly, the dietary administration of Lactobacillus rhamnosus interfered in shoaling behavior and brain expression levels of genes involved in serotonin signaling and metabolism (Borrelli et al., 2016). Recently, significant differences in the intestinal expression pattern of key genes involved in innate and acquired immunity were reported when gilthead sea bream juveniles fed with probiotic (Lactococcus lactis) and control diets (Moroni et al., 2021) were compared. However, to our knowledge, this is the first report to link fish alternative feed formulations with changes in gut microbiota and host transcriptomics at the local and/or systemic level in typical farmed marine fish.

Fish production continues to be strongly dependent on FM (Ytrestøyl et al., 2015; Hua et al., 2019), but there is now evidence that the use of a combination of plant proteins, poultry meal, and insect proteins is able to support optimal growth in a wide range of farmed marine fish, such as gilthead sea bream (Basto et al., 2021; Reis et al., 2021). Moreover, there is now evidence that the reshaping of gut microbiota by the replacement of FM with poultry meal and microbial biomass would serve to exert an adaptive and counter-regulatory inflammatory action at the local intestine level (Solé-Jiménez et al., 2021). Certainly, a number of factors (e.g., age, sex, genetics, and environment) regulate the composition of the microbial population, which could eventually influence growth performance and health (Nayak, 2010; Cordero et al., 2015). However, diet-associated factors are perhaps one of the most important (Silva et al., 2011; Ghanbari et al., 2015). For instance, in zebrafish, a gluten-formulated diet displayed heightened abundances of Legionellales, Rhizobiaceae, and Rhodobacter, as compared with the control diet (Koo et al., 2017). Regarding aquaculture species, in yellowtail kingfish (Seriola lalandi), diet- and diet-associated bacteria shaped gut microbiota through development (Wilkes Walburn et al., 2019). Legume-based diets increased the abundance of lactic acid bacteria in Atlantic salmon, Salmo salar (Gajardo et al., 2017). Likewise, long-term feeding trials with plant-based diets displayed a shift in the resident intestinal microbiota of gilthead sea bream, driven by a dramatic increase in the genus Photobacterium that was partially reversed by dietary butyrate supplementation (Piazzon et al., 2017). However, interactions among microbial community, fish metabolism, fish performance, and health are still largely unknown, and this study serves to highlight the different reshaping of gut microbiota with the PAP and non-PAP diets. Additionally, several correlations between microbial populations and host gene expression were disclosed, with a more pronounced local intestinal action, and with important interactions at the systemic level.

At a closer look, the non-PAP diet decreased ACE and diversity Shannon indices, resulting in reduced richness and alpha diversity, in line with the study of Solé-Jiménez et al. (2021) using poultry meal and microbial biomass as FM replacements. The general thinking is that the occurrence of less diverse bacterial populations is a bad sign, because it is prone to increasing the chance of opportunistic pathogenic bacteria to proliferate and cause host damage (Sekirov et al., 2010; Apper et al., 2016). However, a deeper and experiment-specific understanding of the changes in gut microbiota is needed before reaching such conclusion for a given fish species and experimental condition. For instance, the gut microbiota of fish selected for growth became more stable but at the same time was functionally more plastic against changes in diet compositions (Piazzon et al., 2020). In this study, the microbial composition at the phylum level also remained relatively stable (Figure 3), showing the typical predominant phyla of gilthead sea bream within the expected range of variations for Proteobacteria (38–50% vs. 40–60%), Firmicutes (20–38% vs. 15–40%), and Bacteroidetes (3% vs. 2–5%), and only some deviations for Actinobacteria (11–18% vs. 15–25%) (Parma et al., 2016; Piazzon et al., 2019; Rimoldi et al., 2020). Since this occurred under brackish water conditions (18 ppt), this would be indicative that the gut microbiota of an euryhaline fish, such as gilthead sea bream, has the capacity to remain almost unaltered at the phylum level in the range of 18–37 ppt, which is in contrast with the dramatic changes observed with the sea water transfer in Atlantic salmon (Dehler et al., 2017). In this scenario, the PLS-DA disclosed 135 OTUs that drove (VIP ≥ 1) the separation among the groups (Figure 4). When this discriminant fraction of OTUs was considered more in-depth, the LEfSe analysis displayed a considerable number of taxa commonly related with opportunistic pathogenic bacteria in the gut of both the non-PAP- (Streptococcaceae, 3%; Flavobacteriaceae, 1%; Sphingomonas, 4.5%; Staphylococcus,2.5%) and PAP-fed fish (Clostridium_sensu_stricto, 5.5%) (Starliper, 2011; Sabry et al., 2016; Seghouani et al., 2017). However, assignments at high taxonomic levels, such as genus or families, do not necessarily reflect a pathogenic population, as these OTUs have been commonly detected in the microbial population of gilthead sea bream, and wide and diverse groups with several genus or species are implicated in diverse functions (Estruch et al., 2015; Parma et al., 2016; Nikouli et al., 2018; Piazzon et al., 2019).

Despite the increasing research on gut microbial composition, the understanding of how this population influences and is influenced by host gene expression remains relatively unexplored (Nichols and Davenport, 2021). Model organisms, such as zebrafish, C. elegans, Drosophila melanogaster, and mice are widely used to investigate these associations, with researchers having the control over the environment and complete control over diet, and importantly, the ability to study all tissues with advanced genomic tools (Borrelli et al., 2016; Wang et al., 2017). For aquaculture species, there is still a lack of information regarding the correlation between microbial populations and host gene expression, with some exceptions in black tiger shrimp (Penaeus monodon) (Uengwetwanit et al., 2020). Here, we take advantage of the genomic tools available for gilthead sea bream (Calduch-Giner et al., 2013; Pérez-Sánchez et al., 2019) to intersect the gene expression patterns of the intestine (from this study), liver, and head kidney (from Fernandes et al., 2021) with gut microbial community abundances (Figures 6, 7 and Supplementary Figure 5), putting together the paired samples from CTRL and fish fed with the PAP and non-PAP diets. It is worth noting that the high percentage of the OTUs (43 out of 135; ~32%) driving the separation among the dietary groups in the PLS-DA was significantly correlated with DE genes at the local and systemic levels. This fact supports the usefulness of the PLS-DA filter strategy (VIP ≥ 1) for the detection of transcriptome-associated taxa, as reported in other research animal models correlating metagenomics with host transcriptomics and metabolomics (Yan et al., 2020; Liu and Zhang, 2021; Zhao et al., 2021). Nonetheless, there is an important gap in information on the long-range action of the microbiota, mainly due to the assumption that microorganisms establish their niches in the intestine and they and their response do not spread, and that when an asystemic response exists, it is not always of sufficient clinical relevance (Chiu et al., 2017). However, both assumptions are questionable, because several studies stated the ability of resident microbiota to exert important systemic effects (Brenner et al., 2015; Ho et al., 2015; Grigg and Sonnenberg, 2017). Thus, in this fish study, the number of discriminant OTUs correlating with DE genes decreased from 28 in the intestine to 16 in the liver, and 9 in the head kidney. Among them, 10 out of 48 were involved in local and systemic correlations at the same time. According to this, the host and microbial interactions were more pronounced at the local (gut) level, but the systemic action cannot be underestimated, as further discussed below. Besides, our correlation analysis was based on a targeted gene approach with different tissue-specific PCR arrays, but it remains elusive if the use of a massive gene expression approach (RNA-seq) could modify the established trend for the network of gut microbiota and host transcriptomics.

The homoeostasis of the constantly renewing intestinal epithelium relies on integrated control of proliferation, differentiation, and apoptosis (Macara et al., 2014). Thus, the down-regulation of pcna in gilthead sea bream fed with FM-based diets supplemented with olive oil bioactive compounds has been related to a lower epithelial turnover in association with a better gut health condition (Gisbert et al., 2017). The pcna gene was also down-regulated in fish fed with plant-based diets, and the restoration of control values by butyrate supplementation was considered part of the mechanisms driven by this gut health factor in gilthead sea bream (Estensoro et al., 2016; Piazzon et al., 2017). Accordingly, in this study, the PAP-based feed formulation, and, only to a lower extent (non-significantly), the non-PAP diet (with a higher content of plant proteins), triggered the up-regulation of pcna in the AI of the fish in these two groups. This feature will be prone to promoting cell proliferation rather than cell differentiation with the replacement of FM with non-conventional dietary protein sources in the PAP and non-PAP diets. Of note, we found that the pcna gene expression was inversely correlated to Nocardioides (CTRL.003%; non-PAP and PAP <0.001%) and Sphingomonas oligophenolica (CTRL 2%; ~0.003% in the non-PAP and PAP fish), which suggests a role of these taxa in fish epithelial cell proliferation and regulation.

The integrity of the intestinal epithelium also relies on the maintenance of the mucus layer that is mainly composed of mucins, O-glycosylated glycoproteins that are present on the apex of all wet-surfaced epithelia, protecting epithelial cells from bacteria-, virus- or pH-derived damages and participating in cell signaling processes (Lang et al., 2007). Up to six mucins have been molecularly and transcriptionally characterized in gilthead sea bream, with muc13 being extensively and constitutively expressed along the entire intestinal tract (Pérez-Sánchez et al., 2013). However, in this study, the expression of muc13 was down-regulated in the non-PAP-fed fish, and to a lower extent (non-significantly), in the PAP fish. Besides, the correlation analysis highlighted a positive association of muc13 with OTUs assigned to the phylum Chloroflexi (CTRL,0.58%; absent in PAP and non-PAP) and Anoxybacillus flavithermus (CTRL 0.01%; non-PAP and PAP <0.001%). Since Chloroflexi and the majority of species of the Anoxybacillus genus are described as anaerobic bacteria (Pikuta et al., 2000; Speirs et al., 2019), the up-regulation of muc13 gene expression will tend to produce a thicker layer with less available oxygen that will favor intestinal colonization by these specific taxa. Intriguingly, we also evidenced a negative association of muc13mRNA transcripts with OTUs assigned to Pelosinus and Psychrosinus genera (CTRL, < 0.0001; NoPAP, 0.87%; PAP, 1.13%). To our knowledge, there is no described correlation between these taxa and host gene expression. However, species belonging to these taxa are capable of altering metal speciation, being well-known iron reducers (Ray et al., 2018). Besides, the higher bioavailability of iron in the intestinal lumen of mice has been related to the action of these bacteria on the reduction of intestinal oxidative DNA damage (Li et al., 2015; Eteshola et al., 2020). In our case, the up-regulated expression of the hepatic superoxide dismutase [Mn] (mn-sod/sod2) positively correlated with the higher abundance of Pelosinus and Psychrosinus genera in the AI of the non-PAP- and PAP-fed fish, pointing out a possible systemic cross-talk between fish antioxidant defense and gut microbial community.

As for pcna, the intestinal expression of alpi increased from the CTRL fish to the PAP fish, with intermediate expression values in the fish fed with the non-PAP diet. ALPI is found in high concentrations in the brush border of intestinal epithelial cells of both fish and mammals (Estensoro et al., 2016), regulating the rate of lipid absorption (Mahmood et al., 2003). Additionally, ALPI has a role in the disassembly of lipopolysaccharides (LPS) of Gram-negative bacteria (Poelstra et al., 1997), reducing inflammatory response (Millán, 2006; Rader, 2017). Thus, as in other animal models, a decline in ALPI activity is generally associated with malnutrition in fish (Bakke-McKellep et al., 2000; Ducasse-Cabanot et al., 2007), and perhaps to a less effective response for preventing bacterial invasion across the gut mucosal barrier. In this regard, we found that the up-regulated expression of alpi was closely related with a high representation of Psychrobacter piscatorii (CTRL and non-PAP <0.005%; PAP1.6%) in the intestinal microbiota of gilthead sea bream. Overall, this can be viewed as a positive feature, because some species of the Psychrobacter genus are tolerant to alkaline phosphatase (Denner et al., 2001), and their use as a probiotic reduced the proportion of harmful and pro-inflammatory species in the gastrointestinal tract of cultured orange-spotted grouper (Epinephelus coloides) (Yang et al., 2011). However, the species of this genus also activate the pro-inflammatory NF-kB pathway (Chow et al., 1999; Yu et al., 2016), and a recent study on gilthead sea bream (Solé-Jiménez et al., 2021) associated the lower abundance of Psychrobacter with the reshaping of gut microbiota to preserve an anti-inflammatory gut profile in fish fed with an FM replacement with a high content of poultry meal.

The expression of intestinal igm was markedly up-regulated in the AI of fish fed with the PAP diet. Earlier studies on this fish species highlighted the up-regulation of the intestinal igm in response to changes in the nutritional background, and bacteria or parasite challenges (Estensoro et al., 2012; Piazzon et al., 2016; Simó-Mirabet et al., 2017). Therefore, it is not surprising that this gene correlated with up to six different discriminant OTUs (i.e., Firmicutes, Sphingomonas, Escherichia/Shigella, Pseudomonas matsuisoli, Staphylococcus petrasii, and Streptococcus cristatus). Among these, only the OTUs assigned to the Escherichia/Shigella genus were found in a considerable proportion (1.6%) in the gut of fish fed with the PAP diet, establishing a direct correlation. This genus is defined as an activator of the inflammasome system (Liu et al., 2020), suggesting the driving role of this bacteria in the host immune response. In the head kidney, the expression trend for igm in the PAP fish was also up-regulated. However, the opposite was found for the igt gene expression, which showed a more down-regulated pattern in the head kidney than in the AI, which confirms and extends the idea of differential regulation of these two types of immunoglobulins in fish, gilthead sea bream in particular (Reyes-Cerpa et al., 2014; Piazzon et al., 2016). Besides, both the systemic and local igt expressions were negatively correlated with the abundance of intestinal Psychrobacter piscatorii, reinforcing the aforementioned relationship of this bacteria with gut health in gilthead sea bream.

Another pro-inflammatory marker, il8, was up-regulated in the AI of fish fed with the PAP diet, although in this case, we failed to establish any significant correlation with the intestinal microbial population. However, at the systemic level, the expression of the head kidney il8 was negatively correlated with Bacilli. This link is not surprising, as bacteria from this class are known to have anti-inflammatory and anti-oxidant properties (Giri et al., 2019), being used as probiotics to improve growth and intestinal health in mammals (Rhayat et al., 2019; Zhou et al., 2020) and livestock fish (Mingmongkolchai and Panbangred, 2018), such as gilthead sea bream (Simó-Mirabet et al., 2017; Moroni et al., 2021). In this regard, it must be noted that the alleviation of the intestinal pro-inflammatory pattern of non-PAP fish was concomitant with a higher abundance of the OTUs assigned to Bacilli class (non-PAP 1.63%; CTRL and PAP <0.001%). In the head kidney, this was extended to the expression pattern of tnfα, which achieved the highest expression level in fish fed with the PAP diet and with intermediate values in the non-PAP fish. TNFα has a central role in inflammation (Bradley, 2008), and microbiota have been addressed to block (Yang et al., 2020) or activate its gene expression (Mendes et al., 2019). Hence, in the study, tnfα expression was positively correlated with the abundance of Paracoccus yeei, not present in the gut of CTRL fish but increasing by up to 0.3 and 1.03% in the non-PAP and PAP fish, respectively. This might serve to exert a counter-regulatory inflammatory response that is supported by the inferred metagenomics pathway analysis, where P. yeei is the only contributor in the inferred over-representation of the TNF signaling pathway. Thus, although bacteria-host gene expression interactions can be sometimes difficult to interpret, the combination of different analyses provided support to the results.

Another sign of an anti-inflammatory profile in the non-PAP fish was the intestinal up-regulation of il10, a key anti-inflammatory cytokine used in both fish and mammals as an important marker for health status of the host (Levast et al., 2015; Piazzon et al., 2015). Some increase in the expression level of il10 was also seen in the head kidney, although not significant, confirming the stronger modulation of this cytokine at the local level, as reported in earlier gilthead sea bream studies (Pérez-Cordón et al., 2014). Moreover, regarding microbiota and gene expression associations, the intestinal expression of il10 positively correlated with the abundant Serratia nematodiphila in non-PAP fish (1.02%), while being poorly represented in the microbial population (<0.001%) of CTRL and PAP fish. In any case, the role of the Serratia genus in inflammatory processes remains controversial, because some species act as pro-inflammatory agents through protease-activated receptors (Kida et al., 2007), whereas others are anti-inflammatory factors by means of enzymes with potent anti-oxidant power (Saeki et al., 1974; El-Abd and Ibrahim, 2020).

In a healthy gut, the recognition of microorganisms by PRRs has a primary role in the activation or repression of innate immunity (Fukata and Arditi, 2013), exerting diverse functions depending on the offending factor (Boltaña et al., 2011). In our experiment model, several PRRs, such as galectins and toll-like receptor genes, were monitored. Among them, it is worth noting that the pronounced down-regulation of intestinal lgals8 in both PAP and non-PAP fish would contribute to the maintenance of the repressed immune system. This is in line with several pro-inflammatory molecules abundantly produced in mice by LGAL8-stimulated endothelial cells (Cattaneo et al., 2014). Otherwise, the intestinal expression of lgals8 was significantly down-regulated in healthy gilthead sea bream juveniles fed with practical diets that are supplemented with sodium salt medium-chain fatty acids or Bacillus-based probiotic (Simó-Mirabet et al., 2017). Conversely, LGALs1 has a recognized role in the control of chronic inflammations, weakening cytokine synthesis and de-activating antigen presenting cells, causing an overall immune repression in mice (Seropian et al., 2018). Therefore, the up-regulated expression of intestinal lgals1 in the non-PAP fish, but not in the PAP fish, reinforced the anti-inflammatory expression pattern of fish fed with the non-PAP diet. This was not the case for the tlr genes that showed an enhanced expression in the PAP and non-PAP fish (tlr2) or only in the NoPAP fish (tlr9). The functional significance of this finding remains unclear, but TLR signaling pathways play a key role in the regulation of the immune system, preventing autoimmune and inflammatory diseases in humans and rodents (Tang et al., 2012; Kawasaki and Kawai, 2014). Likewise, tlr genes are often up-regulated when fish deal with a bacterial infection (Reyes-Becerril et al., 2015), but there is now evidence that spirulina feed significantly enhanced the immune response of gibel carp (Carassius auratus gibelio) through the Tlr2 pathway (Cao et al., 2018). Since the non-PAP-based feed formulation also contained Spirulina as a microalgae meal, it is likely that the enhanced expression of tlrs in fish fed with the non-PAP diet was due, at least in part, to the use of this alternative dietary protein source. In any case, the number of OTUs (eight) associated with changes in the expression pattern of PPRs was relatively high. Besides, the correlated taxa belonged to different and diverse assignations (Firmicutes, Clostridium, Nocardioides, Azotobacter, Pseudomonas, and Phenylobacterium), showing the wide range of bacterial types that could be interacting with PRRs.

In a back-and-forth response, gut microbiota can exert a substantial influence on host lipid metabolism, and mechanistic links involving the microbial generation of SCFAs, microbial processing of bile acids, and bacterial-derived pro-inflammatory factors have been reported in mouse models (Schoeler and Caesar, 2019; Lamichhane et al., 2021). This highlights a gut microbiota-liver axis where SCFAs generated by the gut bacterial fermentation of dietary fiber fuel the SCD1 (Stearoyl CoA desaturase)-mediated lipogenesis in the liver of mice (Singh et al., 2015). Intriguing associations between host lipid metabolism and the composition of gut microbiota have also been reported in Atlantic salmon (Dvergedal et al., 2020) and rainbow trout (Yildirimer and Brown, 2018). However, in fish, these relationships remain in an infancy state, reinforcing the value of this study where, for the first time in a typical marine fish, a possible link between enzymes of lipid metabolism and gut microbiota was evidenced. Here, this can be exemplified by the lipogenic scd1a and scd1b enzymes, which becomes especially interesting in the case of scd1a given that a recent study revealed reliable epigenetic mechanisms (changes in DNA methylation rates of the scd1a promoter) by which the nutrition of parents can shape the scd1a gene expression in the gilthead sea bream offspring (Perera et al., 2020). The extent to which this can be favored by changes in the gut microbiota remains unknown, but interestingly, the up-regulation of scd1 genes in the fish fed with the non-PAP diet was associated with OTUs assigned to Serratia nematodiphila (NoPAP, 1.02%) and Brevinema andersonii (NoPAP 0.1%; CTRL and PAP <0.01%), giving these taxa a role as intestinal SCFA producers in humans (Parada Venegas et al., 2019) and enhancers of fatty acid metabolism in insects (Zhou et al., 2021). The case of B. andersonii merits further attention, as this species belongs to Spirochaetesphylum, which colonizes the gut gilthead sea bream microbiome to a large extent with advancing age (11% in 4-year-old individuals) (Piazzon et al., 2019). The physiological consequences of this finding merit further research, because studies on mice highlighted alterations of the hepatic lipid profile and fatty acid synthesis in response to gut colonization by aged microbiota (Albouery et al., 2020). Future studies should focus on enzymatic levels and activities to corroborate these results further.

Like hepatic SCD1, fatty acid elongase 6 (ELOVL6) is a rate-limiting enzyme of fatty acid synthesis and lipogenic activity, being triggered by the gut microbial production of SCFAs in mice (Kindt et al., 2018). This might also occur in our gilthead sea bream model, and the hepatic expression of elovl6 was positively correlated with the non-pathogenic species Staphylococcus carnosus (non-PAP 0.03%; PAP <0.00001%), applied as a starter culture in industrialized processes (Janssens et al., 2012). Another type of association of lipid metabolism and gut microbiota was represented by the peroxisome proliferator-activated receptor α (pparα). This transcription factor of lipid metabolism is a well-known lipolytic factor (Mottillo et al., 2012; Pawlak et al., 2015), and its hepatic down-regulated expression in the fish fed with the non-PAP diet negatively correlated with OTUs assigned to Moraxellaceae, Haemophillus pittmaniae, and Celeribacter marinus. Therefore, the trade-off of host lipogenic and lipolytic pathways becomes related to the changes in the gut microbial population of a farmed marine fish.

In summary, even in the absence of major changes in growth performance, a number of processes related to epithelial cell turnover, immune response, and lipid metabolism were affected by dietary intervention. Both pro- and anti-inflammatory responses were triggered by the non-PAP- and PAP-based feed formulations, although the net effect would be prone to a slight pro-inflammatory status that appears mostly attenuated in the fish fed with the non-PAP diet. Thus, the formulation of the non-PAP diet arises as an attractive formulation to be further studied for its use in gilthead sea bream aquaculture. Indeed, the fish fed with the non-PAP diet shared an enhanced lipogenic activity, which might suggest an enhanced microbial production of SCFAs. As summarized in Figure 8, remarkable correlations between changes in gut microbial populations and host gene expression were unveiled at the local and systemic levels. By correlating microbiome and host gene expression, we offer new insights into the physiological processes promoting both metabolic and gut homeostasis and, ultimately, the health of farmed fish. All this reinforces the action of the gut microbiome as a “second genome,” being involved in—and/or being influenced by—the regulation of the transcriptomic response in fish fed with new feed formulations based on increased circularity and resource utilization. Future studies should focus on the effect of improved non-PAP-based formulations on fish growth and health, with a special focus on disease resistance.

The datasets presented in this study can be found in the Sequence Read Archive (SRA) under Bioproject accession number PRJNA745265 (BioSample accession numbers: SAMN20157689-709).

The animal study was reviewed and approved by Fernando Bernardo, Direçao Geral de Alimentaçao e Veterinaria, Lisboa, Portugal.

FN-C, GV, MP, AF, and JC-G: formal analysis. FN-C, GV, MP, and JP-S: writing—original manuscript. AS-B, LC, and JP-S: conceptualization. All authors contributed to the experimental investigation, writing—review and editing, read, and approved the final version of the manuscript.

This study was supported by the EU H2020 GAIN Project (Aquaculture intensification in Europe, contract 773330). This study reflects only the views of the authors, and the European Union cannot be held responsible for any use that may be made of the information contained herein. Additional funding was obtained by the EU H2020 Research Innovation Program under the TNA Program (project AE150004) at the IATS-CSIC Research Infrastructure within the AQUAEXCEL²⁰²⁰ Project (652831), and the Spanish MICINN Project (Bream-AquaINTECH, RTI2018–094128-B-I00). MP was funded by a Ramón y Cajal Postdoctoral Research Fellowship (RYC2018-024049-I/AEI/10.13039/501100011033, co-funded by the European Social Fund (ESF) and ACOND/2020 Generalitat Valenciana).

GV, AF, and LC were employed by the company SPAROS (Olhăo, Portugal). The remaining authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.