Astaxanthin-Shifted Gut Microbiota Is Associated with Inflammation and Metabolic Homeostasis in Mice

ABSTRACT

Background

Astaxanthin is a red lipophilic carotenoid that is often undetectable in human plasma due to the limited supply in typical Western diets. Despite its presence at lower than detectable concentrations, previous clinical feeding studies have reported that astaxanthin exhibits potent antioxidant properties.

Objective

We examined astaxanthin accumulation and its effects on gut microbiota, inflammation, and whole-body metabolic homeostasis in wild-type C57BL/6 J (WT) and β-carotene oxygenase 2 (BCO2) knockout (KO) mice.

Methods

Six-wk-old male and female BCO2 KO and WT mice were provided with either nonpurified AIN93M (e.g., control diet) or the control diet supplemented with 0.04% astaxanthin (wt/wt) ad libitum for 8 wk. Whole-body energy expenditure was measured by indirect calorimetry. Feces were collected from individual mice for short-chain fatty acid assessment. Hepatic astaxanthin concentrations and liver metabolic markers, cecal gut microbiota profiling, inflammation markers in colonic lamina propria, and plasma samples were assessed. Data were analyzed by 3-way ANOVA followed by Tukey's post hoc analysis.

Results

BCO2 KO but not WT mice fed astaxanthin had ∼10-fold more of this compound in liver than controls (P < 0.05). In terms of the microbiota composition, deletion of BCO2 was associated with a significantly increased abundance of Mucispirillum schaedleri in mice regardless of gender. In addition to more liver astaxanthin in male KO compared with WT mice fed astaxanthin, the abundance of gut Akkermansia muciniphila was 385% greater, plasma glucagon-like peptide 1 was 27% greater, plasma glucagon and IL-1β were 53% and 30% lower, respectively, and colon NOD-, LRR- and pyrin domain-containing protein 3 (NLRP3) inflammasome activation was 23% lower (all P < 0.05) in male KO mice than the WT mice.

Conclusions

Astaxanthin affects the gut microbiota composition in both genders, but the association with reductions in local and systemic inflammation, oxidative stress, and improvement of metabolic homeostasis only occurs in male mice.

Introduction

Obesity and associated metabolic disorders, such as diabetes, are major health issues across the world (1). The etiology of obesity-related metabolic disorders is multifactorial and can be linked to dietary, genetic, and environmental factors. Carotenoids, a large family of fat-soluble pigments found in fruits, vegetables, and egg yolks, and seafood such as salmon and shrimp, have been recognized as essential nutrients with important health benefits (2). Epidemiological studies have suggested that low blood concentrations of carotenoids increase the risk of developing metabolic disorders, such as diabetes (3–5).

The gut microbiota is an ecosystem with the highest density of bacteria in the colon. It is an integral part of the human body and exhibits an impact on numerous metabolic and biological functions. Environmental factors, such as diet, can induce changes in microbial community composition, metabolism, and function, thereby altering the host immune responses and influencing the development of metabolic disorders (6, 7). There are several possible mechanisms through which an imbalance in gut microbiome homeostasis contributes to inflammation and metabolic disorders. For example, increased release of lipopolysaccharides (LPS) from gram-negative gut bacteria stimulates inflammation; decreased production of short-chain fatty acids (SCFAs) results in energy starvation of gut beneficial microbes and host intestinal epithelial cells. Reduced secretion of gut hormones, such as glucagon-like peptide 1 (GLP-1), can lead to damage to the pancreas (8–10).

There are 2 major groups of dietary carotenoids, carotenes and xanthophylls. Carotenes, such as α-carotene and β-carotene, contain no oxygen atoms and are major dietary sources of vitamin A. Xanthophylls, such as zeaxanthin, lutein, and astaxanthin, contain ≥1 oxygen atom. Because humans and animals are unable to synthesize carotenoids de novo, the diet is the only source for these essential nutrients. With regular meals, only 10–20% of the ingested carotenoids are absorbed in the small intestine in humans (11). As a result of their relatively poor bioavailability, the large majority of these dietary pigments reach the large intestine. In vitro studies suggested that only 10–50% of these carotenoids are recovered in the colonic fraction (3, 12). Thus, carotenoids are likely “fermented” in the gut, likely through the influence of the gut microbiome, though studies that can fully identify these mechanisms are lacking.

Astaxanthin, a dietary xanthophyll predominantly found in marine foods (or seafood) exhibits superb antioxidant properties (13). Indeed, the beneficial health effects of astaxanthin include energy metabolism, cardiovascular health, immunity, and vision (14, 15). However, despite these beneficial health effects, our knowledge about the function of astaxanthin remains limited. In the literature, a majority of published animal work relays on the findings from a single-gender animal model. Gender-specific differences in microbiome structure and function are still overlooked. In this study, we sought to examine the extent to which astaxanthin impacts both the gut microbiota and host metabolism in male and female wild-type C57BL/6 J (WT) and β-carotene oxygenase 2 (BCO2) knockout (KO) mice. BCO2 cleaves carotenoids at the 9, 10, and/or 9′,10′ double bonds to yield apocarotenoid products (16–18). The BCO2 enzyme is highly expressed and active in the intestine in rodents (19, 20) and resides at the inner membrane of the mitochondria (21). However, the enzymatic role of BCO2 in carotenoid metabolism in humans is controversial (22–24). Mice deficient in BCO2 accumulate parent carotenoids such as lutein and zeaxanthin apart from their apocarotenoids metabolites, whereas WT mice lack accumulation of the parent carotenoids due to BCO2-dependent conversion into apocarotenoids (17). The hypothesis of this study is that astaxanthin supplementation alters the gut microbiota and that the mouse BCO2 genotype and gender will impact astaxanthin metabolism and modulate this interaction.

Methods

Animals, husbandry, and dietary treatments

Six-wk-old male and female KO and the genetic background WT mice were randomly assigned to 1 of 2 dietary treatments (n = 10 mice/group): the control diet (CONT) consisting of AIN93M, with 10% kcal from fat, and the astaxanthin diet (ASTX) of CONT supplemented with 0.04% astaxanthin (wt/wt) for 8 wk. Thus, there was a total of 8 treatment groups, and mice of each gender [female (F) and male (M)] were grouped into KO-CONT (BCO2 KO mice fed CONT), KO-ASTX (BCO2 KO fed ASTX), WT-CONT (WT fed CONT), and WT-ASTX (WT fed ASTX), respectively. The original KO mice (17) were transferred from the Case Western Reserve University to the Lin Laboratory at Oklahoma State University. KO animals were backcrossed to WT (purchased from The Jackson Laboratory, stock #000664) for 10 generations to achieve the BCO2 KO strain in a C57BL/6 J background. The littermates of KO and WT mice were then used for studies. Genotyping was ensured by PCR validation performed by Transnetyx, Inc. For the current study, animals in each experimental group were randomly selected from 3–5 litters of WT and/or KO mice to maximize the intragroup representativeness. AstaREAL^®L10, an extract from the microalgae Haematococcus pluvialis, containing 10% ASTX and approximately 90% triglyceride (TG), was provided by Fuji Chemical Industry Corporation Limited. The diet composition is provided in Supplemental Table 1. Unless otherwise indicated, mice were group-housed and maintained at the Oklahoma State University Laboratory Animal Research facility under humidity- and temperature-controlled conditions and a 12-h light–12-h dark cycle with free access to food and water. Animals were individually identified by ear tagging in the study. Body weight and food intake were monitored weekly. All animal experiments and procedures were performed in accordance with the protocol approved by the Oklahoma State University Institutional Animal Care and Use Committee (OSU ACUP #HS-14-4). Due to the greater responses to the dietary treatments, only male mice were further subject to the assessment of metabolism homeostasis.

Measurement of energy expenditure by indirect calorimetry

Assessment of whole-body substrate oxidation and energy expenditure were performed using an indirect calorimetry system (TSE PhenoMaster, TSE Systems) as described previously (25). O₂ consumption (VO₂) and CO₂ production (VCO₂) were measured at every 15-min interval for a total of 48 h. The respiratory exchange ratio (RER), the ratio of VCO₂ to VO₂ [or RER = V(CO₂)/V(O₂)] was measured.

Fecal and tissue collection

At wk 8 of the supplementation, mice were singly housed overnight to collect feces. At the termination of dietary supplementations, mice were deprived of food for 3 h prior to being killed. Whole-body PixiMus scans (GE Lunar) were used to assess body composition, e.g., whole-body fat percentage (Fat%) and bone mass density (BMD). Whole blood was collected from the carotid artery into EDTA-coated tubes. Plasma samples were obtained via centrifugation (3000 × g, 25 min, 4°C) and stored at −80°C until analyzed. Liver tissues were weighed and frozen in liquid nitrogen for biochemical analysis, and a portion was maintained on ice for the mitochondrial fragmentation assay. The cecal contents were pelleted by flushing with sterile ice-cold PBS followed by centrifugation. In addition, the colon tissues were flushed with sterile ice-cold PBS and incised, and lamina propria tissues were collected into microcentrifuge tubes and stored at −80°C for protein extraction and subsequent immunoblotting analysis (26).

Astaxanthin HPLC quantification

Hepatic tissue (∼20 mg) was pulverized in liquid nitrogen. Astaxanthin extraction and HLPC identification were performed, as previously described (25). Due to a lack of standards for astaxanthin metabolites and isomers, the molar amounts of these compounds (including meso-, cis, and all-trans astaxanthins), were calculated based on the standard of all-trans astaxanthin.

Gut microbiome profiling

Genomic DNA was extracted from frozen cecal contents using the PowerFecal DNA kit following the standard protocol (Qiagen, catalog #12830). Cecal microbiome 16S rRNA sequencing was conducted by Novogene Bioinformatics Technology Co., Ltd. Sequencing data bioinformatics was performed as we previously reported (27). The 16S V4 rRNA region was amplified by PCR with primers (515F: GTGCCAGCMGCCGCGGTAA; 806R: GGACTACHVGGGTWTCTAAT). The raw operational taxonomic unit (OTU) abundance was normalized for subsequent analysis of α and β diversities using a standard sequence number corresponding to the sample with the least sequences.

The α diversity (observed species, Chao1, Shannon, Simpson, ACE, good coverage), for the complexity of species diversity for an individual sample, was calculated with QIIME (V1.7.0) and displayed with R software (V2.15.3). The β diversity, for evaluating differences among samples in species complexity, was calculated by QIIME software. Cluster analysis was performed by principal coordinate analysis (PCoA) and displayed by use of the WGCNA package, stat package, and ggplot2 package in R software. P values were calculated by the method of the permutation test, whereas q values were calculated by the Benjamini and Hochberg false discovery rate method (28).

Short-chain fatty acid analysis

Fecal short-chain fatty acids (SCFAs) were determined according to previous publications (29, 30). A 5-point calibration was performed using standard solutions of butyric acid, acetic acid, propionic acid, and valeric acid (Sigma-Aldrich).

Measurement of plasma parameters

Plasma concentrations of C-reactive protein (CRP), alanine aminotransferase (ALT), aspartate aminotransferase (AST), lactate dehydrogenase (LDH), total cholesterol, HDL cholesterol, LDL cholesterol, and triglyceride (TG) were measured by BioLis 24i automated analyzer (Carolina Liquid Chemistries Co.). Plasma LPS binding protein (LBP) was tested by ELISA. Plasma inflammatory markers including tumor necrosis factor-alpha (TNF-α), IL-6, IL-17A, IL-1β, and diabetes markers including GLP-1 and glucagon were measured by Bio-Plex MAGPIX Multiplex Reader (Bio-Rad) and data were processed using Bio-Plex Manager 6.1 software.

SDS-PAGE and immunoblotting

Frozen liver samples were used for the preparation of total proteins as previously described (31). Total proteins from colon lamina propria samples were extracted using the standard TRIzol-protein extraction protocol according to the manufacturer's instruction (Thermo Fisher Scientific). About 20 μg total proteins were separated by SDS-PAGE and followed by immunoblotting (31). Primary and secondary antibodies used in the current study are listed in Supplemental Table 2.

Statistical analyses

All statistical computation was conducted by SAS (Version 9.4; SAS Institute) and was set as the significance level. Equal variance was checked using Levene's test. Three-way ANOVA and Tukey's post hoc procedures were conducted to evaluate the main effects of the BCO2 genotype, diet, gender, and their interactions in the datasets of gut microbiota, SCFAs, and plasma metabolic parameters. To further explore the interactive effect between diet and genotype, we focused on male mice only because the male mice had greater responses to the astaxanthin supplementation. Two-way ANOVA was employed in the data sets of astaxanthin accumulation and animal energy metabolism and inflammation. Spearman correlation analysis was performed to calculate the correlation between the relative abundance of different microbiota and clinical parameters. Data were presented as means ± SDs (n = 10) unless otherwise stated in the text and figure legends.

Results

Astaxanthin accumulation and its associated metabolic homeostasis in KO mice

Neither abnormal feeding behaviors nor other health problems, such as changes in food intake or water consumption, were observed in mice during the astaxanthin supplementation. Due to the deletion of BCO2, astaxanthin supplementation resulted in astaxanthin accumulation subcutaneously and also in the adipose and liver tissues as visualized by the naked eye (Supplemental Figure 1A). The orange-red color of the hepatic mitochondrial fractions indicated astaxanthin accumulation in mitochondria of BCO2-deficient mice (KO) (Supplemental Figure 1B). HPLC analyses of the carotenoid composition and concentrations of hepatic lipid extracts are shown in Figure 1. As expected, hepatic astaxanthin concentrations were significantly higher by a factor of 10-fold in KO mice compared with WT mice. Parent astaxanthins existed in different geometric forms (for example, cis, all-trans, and meso- astaxanthins).

FIGURE 1.

Representative HPLC traces at 420 nm (A) and hepatic astaxanthin concentrations (B) in male WT and BCO2 KO mice fed a control or 0.04% astaxanthin diet for 8 wk. Both WT and KO mouse liver tissues contain all-trans astaxanthins (peak 2). Additionally, they contain peaks that resemble from spectral characteristics geometric (cis and meso-) isomers of astaxanthins. The spectral characteristics of peaks 1, 2, and 3 are shown underneath the blot (A). Astaxanthin concentration data were analyzed by 2-way ANOVA (genotype and diet). Values are means ± SD, n = 5. Within sex, labeled means without a common letter differ, P < 0.05. Abs, absorbance; ASTX, astaxanthin (diet); CONT, nonpurified AIN93M diet; KO, BCO2 knockout mice; nd, not detectable (<0.001 nmol/g liver tissues); WT, wild type C57BL/6 J mice.

Although daily food intake was not different among groups (data not shown), male KO-CONT had a significantly higher total body weight gain (BWG) than male KO-ASTX, WT-CONT, or WT-ASTX mice (Figure 2A and Supplemental Table 3). Female WT mice fed CONT exhibited a significantly lower fasting blood glucose (FBG) than male WT mice fed CONT. Male KO-ASTX trended toward a lower FBG than male KO-CONT (P = 0.09) (Figure 2B). In contrast, female KO-AST exhibited a higher FBG than female KO-CONT (Figure 2B). Similarly, in male mice, the whole-body fat percentage (Fat%) was significantly higher in KO-CONT mice than WT-CONT mice. Astaxanthin supplementation increased Fat% levels in male WT but not KO mice. No distinct changes in Fat% were identified in female groups (Figure 2C and Supplemental Table 3).

FIGURE 2.

Total body weight gain (A), fasting blood glucose (B), total body fat percentage (C), and concentrations of plasma cholesterol (D), HDL cholesterol (E), and LDL cholesterol (F) in male and female WT and BCO2 KO mice fed control or 0.04% astaxanthin diet for 8 wk. Values are means ± SDs, n = 10. Labeled means without a common letter differ, P < 0.05. Statistical outputs for the parameters analyzed by 3-way ANOVA (gender, genotype, and diet) are shown in Supplemental Table 3. BWG, total body weight gain (g/8 wk); CHOL, Cholesterol; Fat%, percentage body fat; FBG, fasting blood glucose; KO-ASTX, BCO2 knockout (KO) mice fed astaxanthin diet (ASTX); KO-CONT, KO mice fed control nonpurified AIN93M diet (CONT); WT-ASTX, wild-type C57BL/6 J (WT) fed ASTX; WT-CONT, WT fed CONT.

Plasma concentrations of total cholesterol and HDL cholesterol were higher in male and female KO mice supplemented with astaxanthin than in the mice in the other groups. KO-CONT mice had a higher LDL cholesterol concentration compared with WT-CONT. Astaxanthin supplementation elevated LDL cholesterol concentrations in male WT and female KO mice (Figure 2D and F and Supplemental Table 3). No significant differences in TG and CRP concentrations were observed in all 8 groups (Supplemental Figure 2B and C). Significant elevations of AST and ALT concentrations in plasma were observed in female KO-CONT mice (Supplemental Figure 2D and E). The elevated LDH concentrations were only observed in male KO mice, and these concentrations were diminished by astaxanthin supplementation (Supplemental Figure 2F and Supplemental Table 4).

Taken together, the accumulation of parent astaxanthins was associated with the overall alteration of glucose and lipid homeostasis in KO mice.

Astaxanthin modulates gut microbiota compositional shifts

The cecal microbiota 16S rRNA sequencing results showed that the α diversity and Chao 1 indexes of the observed species were not significantly different among treatment groups (Figure 3A and B). The results of the principal coordinates analysis (PCoA) showed significant rearrangements of the gut bacterial structure among the groups (Figure 3C). Male mice presented significantly different cluster separations from female mice. The multiresponse permutation procedure (MRPP) statistical analysis also revealed significant microbiome structural differences (P < 0.05) among all experimental groups (Supplemental Table 5). As shown in male mice, 1038 OTUs were shared among all 4 groups (Figure 3D). Only 149 OTUs were found in WT-CONT, 28 OTUs in KO-CONT, 36 OTUs in WT-ASTX, and 633 OTUs in KO-ASTX (Figure 3D). Distinct from male mice, 998 OTUs were shared among all 4 female groups, with 359 OTUs found in WT-CONT, 238 OTUs in KO-CONT, 18 OTUs in WT-ASTX, and 121 OTUs in KO-ASTX (Figure 3E).

FIGURE 3.

The gut microbiota composition at α-diversity levels, including observed species (A), Chao 1 (B), PCoA (C), and Venn diagram showing the unique and shared OTUs in the different groups (D, male; E, female) in male and female WT and BCO2 KO mice fed control or 0.04% astaxanthin diet for 8 wk. Values in A and B were analyzed by 3-way ANOVA (gender, genotype, and diet) and presented as means ± SDs; n = 5. Labeled means without a common letter differ, P < 0.05. F, female mice; KO-ASTX, BCO2 knockout (KO) mice fed astaxanthin diet (ASTX); KO-ASTX-F, female KO-ASTX; KO-ASTX-M, male KO-ASTX; KO-CONT, KO mice fed control nonpurified AIN93M diet (CONT); KO-CONT-F, female KO-CONT; KO-CONT-M, male KO-CONT; M, male mice; NS, not significant (P ≥ 0.05); PCoA, principal coordinate analysis; WT-ASTX, wild type C57BL/6 J (WT) fed ASTX; WT-ASTX-F, female WT-ASTX; WT-ASTX-M, male WT-ASTX; WT-CONT, WT fed CONT; WT-CONT-F, female WT-CONT; WT-CONT-M, male WT-CONT.

Heatmap analysis of the top 10 abundances of microbiota at the phylum level demonstrated a distinct microbiome shift among these 8 treatment groups (Figure 4A). At the phylum level, Bacteroidetes and Proteobacteria were significantly higher, whereas the Firmicutes/Bacteroidetes (F/B) ratio and Actinobacteria abundance were significantly lower in the female than the male mice, e.g., gender specificity. Dietary astaxanthin significantly increased the abundance of Cyanobacteria and decreased the abundance of Firmicutes and Spirochaetes. Further, the deletion of BCO2 was associated with significant changes in the abundances of Deferribacteres (increased by 610%) and Actinobacteria (decreased by 290%) (Figure 4A). In male mice, KO-CONT mice had a significantly lower F/B ratio than KO-ASTX mice. However, in female mice, astaxanthin supplementation significantly lowered the F/B ratio in WT (Figure 4B). Moreover, the significantly altered gut microbiota composition at a species level is shown in Figure 4C. The relative abundance of Mucispirillum schaedleri (phylum: Deferribacteres) was significantly higher in KO mice than WT in both genders fed either the CONT or ATSX diet. Astaxanthin supplementation significantly elevated the abundance of Akkermansiaa muciniphila (phylum: Verrucomicrobia) in male KO mice compared with male WT mice (Figure 4C and 9). In contrast, astaxanthin supplementation significantly decreased abundance of Parabacteroides distasonis in KO mice, but not WT mice (Figure 4C).

FIGURE 4.

Abundance distribution of the gut microbiota of the top 10 phyla (A), the F/B ratio (B), significantly changed species among the top 10 most abundant microbiota species (C), and correlations between clinical metabolic parameters and species of the gut microbiota (D) in male and female WT and BCO2 KO mice fed control or 0.04% astaxanthin diet for 8 wk. For A, *P < 0.05, male mice compared with female mice; +P < 0.05, WT mice compared with BCO2 KO mice; #P < 0.05, CONT diet compared with ASTX diet. Statistical outputs for panels B and C analyzed by 3-way ANOVA (gender, genotype, and diet) are shown in Supplemental Table 6. P value was calculated by the method of the permutation test. Values for B and C are means ± SDs; n = 5. Labeled means without a common letter differ, P < 0.05. Spearman's correlation coefficients between the relative abundances of the top 50 enriched species and the concentration of clinical metabolic parameters were calculated and shown in D. Only significant correlations are presented. Blue color, negative correlation; Red color, positive correlation. The phylum of each species related is differentially colored. n = 5; ⁺q < 0.05, *q < 0.01. F/B, Firmicutes/Bacteroidetes; KO-ASTX, BCO2 knockout (KO) mice fed astaxanthin diet (ASTX); KO-CONT, KO mice fed control nonpurified AIN93M diet (CONT); WT-ASTX, wild type C57BL/6 J (WT) fed ASTX; WT-CONT, WT fed CONT.

The 3-way ANOVA analysis of the relative abundance of the top 35 abundance microbiota at different taxonomic ranks revealed specific associations of gut microbiota with gender (e.g., male compared with female), genotype (e.g., WT compared with KO), and/or diet (e.g., CONT compared with ASTX). For example, there was a 340% increase in the abundance of Lachnospiraceae bacterium 609 (phylum: Firmicutes) and a 190% decrease in Ruminococcus 1 (phylum: Firmicutes) in female mice compared with male mice (Table 1). Deletion of BCO2 was associated with a significant increase in the abundance of M. schaedleri (phylum: Deferribacteres) and a significant decrease in Bifidobacteriales (phylum: Actinobacteria) (Supplemental Table 7). Astaxanthin supplementation enriched the abundance of Gastranaerophilales (phylum: Cyanobacteria) by 220% but decreased the abundance of Spirochaetales (phylum: Spirochaetes) by 150%, compared with the control diet (Supplemental Table 8).

TABLE 1.

Gut microbiota shifts among different taxonomic ranks associated only with gender in male and female WT and BCO2 KO mice fed a control or 0.04% astaxanthin diet for 8 wk

Microbial rank and name	Ratio (female/male)	P value
Phylum
Saccharibacteria	0.61	0.0104
Class
Unidentified Firmicutes	0.63	0.0016
Unidentified Saccharibacteria	0.56	0.0027
Family
Rikenellaceae	1.46	<0.0001
Clostridiales vadinBB60 group	0.48	<0.0001
Genus
Blautia	3.05	<0.0001
Eubacterium coprostanoligenes	1.46	0.0027
Oscillibacter	1.25	0.0129
Lachnoclostridium	1.21	0.0075
Ruminococcaceae UCG014	0.76	0.0124
Ruminococcus 1	0.54	0.0124
Species
Lachnospiraceae bacterium 609	3.37	<0.0001
Oscillibacter sp. 1–3	1.47	<0.0001
Sutterella sp. YIT 12,072	0.84	0.0334
Clostridiales bacterium enrichment culture clone06-1,235,251-76	0.78	0.0242

BCO2, β-carotene oxygenase 2; KO, knockout; WT, wild type C57BL/6 J.

In summary, dietary astaxanthin had a significant impact on the microbiota composition at phylum/class/order/family/genus/species levels, depending on gender and genotype.

Astaxanthin supplementation alters fecal SCFA production in WT and BCO2 KO mice

Female mice had significantly higher fecal concentrations of acetic acid (by 130%), butyric acid (by 170%), propionic acid (by 170%), and valeric acid (by 180%) than male mice (Supplemental Figure 3A–D, Supplemental Table 9). There were no significant changes in SCFAs caused by ASTX supplementation. No significant changes were observed in plasma concentrations of LBP among treatment groups (CONT and ASTX) regardless of genotype (WT and KO) (Supplemental Figure 3E and Supplemental Table 9).

Gut microbial taxa are correlated with metabolic parameters

Spearman's correlation analysis results revealed that there were many associations between microbiota and particular metabolic parameters (such as BWG, Fat%, FBG, and plasma cholesterol concentrations). For instance, at a species level, Clostridium cocleatum (phylum: Firmicutes) had a strong negative correlation with most metabolic parameters, including Fat%, FBG, cholesterol, LDL cholesterol, and HDL chlosterol. In contrast, Marvinbryantia schaedleri showed a strong positive correlation with significant metabolic markers, such as Fat%, FBG, cholesterol, LDL cholesterol, HDL chlosterol, LDH, AST, and ALT (Figure 4D). At a genus level, Marvinbryantia (phylum: Firmicutes) and Erysipelatoclostridium (phylum: Firmicutes) exhibited strong negative correlations with most above markers, whereas Mucispirillum (phylum: Deferribacteria) showed a strong positive correlation with the same metabolic parameters above (Supplemental Figure 4).

Astaxanthin supplementation promotes hepatic mitochondrial biogenesis and attenuates local and systemic inflammation in male BCO2 KO mice

As shown above, male mice exhibited higher responses to the astaxanthin supplementation than female mice. Therefore, we further assessed markers of mitochondrial biogenesis, oxidative stress, and inflammation in male mice (Figure 5). As shown in Figure 5A, supplementation with astaxanthin restored the protein abundance levels of hepatic peroxisome proliferator-activated receptor gamma coactivator 1-alpha (PGC-1α) and heat shock protein 60 (HSP60) in KO mice. Astaxanthin-induced sirtuin 3 (Sirt3) protein abundance was higher in KO-ASTX mice than that in WT-ASTX. On the other hand, loss of BCO2 resulted in elevated protein abundance of NOD-, LRR- and pyrin domain-containing protein 3 (NLRP3) in colonic lamina propria, and this increased abundance was attenuated by astaxanthin supplementation (Figure 5B). NRLP3 concentrations were not altered in WT mice.

FIGURE 5.

Protein markers of hepatic mitochondrial biogenesis and stress (A), colonic lamina propria NLRP3 protein expression (B), and plasma glucagon (C), glucagon-like protein-1 (D), TNF-α (E), IL-6 (F), IL-17A (G), and IL-1β (H), and animal whole-body respiratory exchange ratio (I) in male WT and BCO2 KO mice fed control or 0.04% astaxanthin diet for 8 wk. Values are means ± SDs, n = 10. Within sex, labeled means without a common letter differ, P < 0.05. Data were analyzed by 2-way ANOVA (genotype and diet). GLP-1, glucagon-like protein-1; HSP60, heat shock protein 60; KO-ASTX, BCO2 knockout (KO) mice fed astaxanthin diet (ASTX); KO-CONT, KO mice fed control nonpurified AIN93M diet (CONT); NLRP3, NLR family pyrin domain containing 3; PGC-1α, peroxisome proliferator activated receptor gamma coactivator 1 α; RER, respiratory exchange ratio; Sirt3, sirtuin 3; WT-ASTX, wild type C57BL/6 J (WT) fed ASTX; WT-CONT, WT fed CONT.

BCO2 deficiency significantly increased the concentration of fasting blood glucagon in the KO-CONT compared with the WT-CONT mice, and this was reversed by astaxanthin supplementation in KO mice (KO-ASTX) (Figure 5C). No changes were observed in plasma concentrations of GLP-1 between WT-CONT and KO-CONT groups. However, astaxanthin supplementation significantly elevated plasma GLP-1 in KO mice compared with WT mice (Figure 5D). Plasma concentrations of IL-1β and TNF-α were higher (P < 0.05) and those of IL-6 (P = 0.09) and IL-17A (P = 0.08) tended to be higher in KO-CONT mice than WT-CONT mice. The anti-inflammatory effect of astaxanthin was also evident in the reduction of plasma IL-1β in KO-ASTX mice compared with that in WT-ASTX mice (Figure 5E–H).

In order to understand if these biochemical changes in male mice were linked to alterations in whole-body energy metabolism, male mouse energy expenditure was monitored, and the results are shown in Figure 5I. Overall, RER over a period of 24 h (in 12-h light and 12-h night cycles) was not different between WT and KO mice with or without astaxanthin supplementation. However, during the night cycle (19:00 to 07:00), KO-CONT mice showed a significantly lower RER than WT-CONT but not WT-ASTX mice.

Discussion

In this study we investigated the effects of astaxanthin on gut microbiota and host nutrient metabolism in male and female WT and BCO2 KO mice. The study demonstrated an interplay between gender, diet, and genotype in gut microbiota homeostasis, host metabolism, and immune responses. Our major findings included the following: 1) male and female mice exhibited different gut microbiota structures that may represent distinct responses to the astaxanthin supplementation; 2) BCO2 deletion led to a gut microbiota shift, which was associated with inflammation, mitochondrial oxidative stress, and metabolic dysfunction; 3) dietary astaxanthin-induced alteration of the gut microbiota was correlated with metabolic homeostasis, and that depended on BCO2 genotype; 4) Mucispirillum schaedleri (phylum: Deferribacteres) was positively and Marvinbryantia (phylum: Firmicutes) was negatively associated with metabolic disorder markers. The major findings are now discussed and put into context with what has been published in the literature.

Our study revealed that gender was associated with alterations in the gut microbiota structure, which might present distinct responses to dietary astaxanthin supplementation. Female mice exhibited higher concentrations of fecal SCFAs and a significantly higher abundance of Lanchnospiraceae bacterium 609, belonging to the family Lachnospiraceae of phylum Firmicutes. Previous publications show that several members of Lanchnospiraceae function to ferment undigested dietary fibers to produce SCFAs (32). SCFAs stimulate commensal bacterial growth and host epithelial regeneration, leading to inhibition of systemic inflammation (33) and restoration of hepatic macronutrient metabolism and blood glucose concentrations (34). Further, female mice exhibited significantly lower BWG, FBG, and Fat%. Therefore, we conclude that the significance of gender difference is a critical factor to consider in studies regarding gut microbiome and host nutrient metabolism.

We have previously shown that deletion of BCO2 causes metabolic impairment that is characterized by mitochondrial dysfunction and alteration of the metabolome (31, 35). As we show here, these changes in glucose and lipid metabolism correlated with a shift in the gut microbiota. In BCO2-deficient mice, M. schaedleri abundance was significantly increased regardless of gender and highly associated with several metabolic markers, such as FBG, Fat%, and cholesterol. Previous publications have reported that increased abundance of M. schaedleri in the gut leads to local and systemic inflammatory responses (36, 37). Our previous publications and current data indicated that the deletion of BCO2 is associated with hepatic oxidative stress, mitochondrial dysfunction, and altered energy metabolism. Thus, changes in the abundance of M. schaedleri could be a potential biomarker for BCO2 protein expression status in mice.

Astaxanthin supplementation caused gut microbiota compositional shifts, particularly in the phyla of Cyanobacteria and Spirochaetes, though broad shifts of the gut microbiota were observed in mice fed the astaxanthin-supplemented diet. Of note, the source of astaxanthin used in the current study was extracted from the green microalga Haematococcus pluvialis and was a mix of astaxanthin isomers, including but not limited to all-trans astaxanthin, cis astaxanthin, meso-astaxanthin, and possible other esterified forms of astaxanthins (38, 39). The green microalga originality of astaxanthin might induce the expansion of Cyanobacteria growth (38). Given the distinct effects of astaxanthin supplementation on the gut microbiota in WT compared with KO mice, data suggested that parent astaxanthins, but not their BCO2-cleaved products (e.g., metabolites), could specifically shift the gut microbiota, for example, astaxanthin supplementation-decreased abundance of Parabacteroides distasonis, in KO mice but not WT mice.

To date, neither astaxanthin metabolism nor astaxanthin function have been well studied and characterized in mammals. Similar to the carotenoids lutein and zeaxanthin, astaxanthin may be cleaved by BCO2 in the mitochondria (21, 40–42). Parent astaxanthin and/or its metabolites play roles in many aspects of cell growth and survival. For example, astaxanthin attenuates bacterial infection–induced gastric inflammation and suppresses chemical-induced carcinogenesis (43, 44). We observed much higher astaxanthin concentrations in the liver tissues and hepatic mitochondrial fractions in KO mice. These results suggest that BCO2-mediated catabolism of astaxanthin occurs predominantly in the mitochondria.

In male mice, astaxanthin supplementation resulted in restoration of PGC-1α, HSP60, Sirt3, and NLRP3, suggestive of improved hepatic mitochondrial biogenesis, oxidative stress, and inactivation of the colonic NLRP3 inflammasome in KO mice. Recent publications reported the interaction of the gut microbiome with the NLRP3 inflammasome as a new therapeutic target through regulation of inflammatory IL-1β and IL-18 concentrations (45–47). The association of astaxanthin accumulation with NLRP3 inflammasome inactivation, as indicated by the decreased IL-1β level, but not with other inflammatory cytokines, suggests that this inflammasome is a possible target of the interaction between astaxanthin and the microbiota. Further research is warranted to explore the underlying mechanism. These results suggested that astaxanthin largely attenuated the metabolic changes associated with the genetic disruption of the BCO2 gene.

The astaxanthin supplementation–associated changes in the gut microbiota composition were linked to attenuation of colonic (or local) and systemic inflammation in KO mice. A. muciniphila is known as a mucin-degrading bacteria that plays a critical role in protecting the gut mucus layer and also the mucosal layer. Indeed, serum carotenoid concentrations are tightly associated with the growth of A. muciniphila in the gut (48). Enrichment of A. muciniphila protects the mouse white adipose tissues from high saturated lipid-induced inflammation (49). The population of A. muciniphila was decreased following the progression of obesity induced by a high-fat diet (50). Our current results on the abundance of A. muciniphila suggested that the accumulation of astaxanthin, and not the apocarotenoid metabolites, promoted the expansion of A. muciniphila in both male and female KO mice. Another example is the decreased abundance of Proteobacteria in KO mice. Increased Proteobacteria population is associated with inflammation and is considered an indicator of microbiome imbalance and a casual factor of cecal inflammation (51, 52). Therefore, we conclude that astaxanthin was beneficial for gut microbiota homeostasis by enhancing the beneficial bacterial abundance (such as A. muciniphila) and suppressing pathogenic bacterial growth (such as Proteobacteria). That led to amelioration of local and systemic inflammation in KO mice.

Many studies have reported that gut microbial fermentation is linked to increased secretion of GLP-1 and other enteroendocrine peptides by the enterocytes (53, 54), which in turn suppresses glucagon secretion from pancreatic α-cells and consequently benefits the maintenance of blood glucose homeostasis (55). In male KO mice, astaxanthin supplementation elevated GLP-1 production and suppressed glucagon secretion into the circulation (Figure 5C and D). These results could also suggest that astaxanthin can act as a prebiotic and helps in blood glucose control. Recent studies suggest that astaxanthin is associated with the alteration of the gut microbiota and lipid metabolism in rodent models of obesity and fatty liver disease (56, 57).

We demonstrated that astaxanthin is associated with the gut microbiota composition shift, inflammation, and nutrient metabolism in mice, depending on BCO2 expression. There are several major limitations in the current study. First, the interactions among housing, environment, diet, gut microbiota, and host responses are extremely complex in microbiome research, particularly in animal studies. Mice are sensitive to other minor changes in the environment, for example, bedding and light exposure (58–61). Therefore, we used both strains from our long-term–maintained colonies of littermates to minimize the confounding variables caused by the environment. Second, several questions about the functions of distinctly changed microbiota under different treatments need to be further elucidated by functional microbiome analysis. Third, the roles of astaxanthin catabolism and its potential metabolites warrant further studies. Fourth, fecal transplantation could be used to confirm whether the change in gut microbiota is a causal factor in astaxanthin regulation of nutrient metabolism. We are also aware of the relatively high dose of astaxanthin that was used in the current feeding study and the extent to which these amounts of astaxanthin may be attainable in humans strictly through dietary intake. There is no specific recommendation for human astaxanthin intake in the literature. According to much higher metabolic rates in mice, a high dose for a short time period may help improve the gut microbiome in rodents, though it should not be assumed that similar dietary interventions would not yield similar results in humans. And finally, longitudinal studies in mice and humans would promote a better understanding of astaxanthin as a potential prebiotic in human health.

In summary, astaxanthin supplementation resulted in changes in the gut cecal microbiota composition, the host oxidative stress condition, inflammatory cytokines, mitochondrial biogenesis, and energy expenditure. These effects of astaxanthin depended on BCO2 protein expression. Our studies suggest that astaxanthin has a profound effect on the gut microbiota and attenuates local and systemic oxidative stress and inflammation. These properties of astaxanthin control inflammation, energy metabolism, and blood glucose concentrations, with an overall beneficial effect on metabolic homeostasis in mice.

Notes

This work was partially supported by the USDA/NIFA (2020-67017-30842 to DL). YL was a visiting graduate student supported by the Priority Academic Program Development of Jiangsu Higher Education Institutions (PAPD), China (to XS).

Author disclosures: The authors report no conflicts of interest.

Availability of Data and Materials: Raw gut microbiome sequencing files have been deposited in NCBI's Sequence Read Archive (https://www.ncbi.nlm.nih.gov/sra). The accession number for BioProject is PRJNA529189. The accession numbers for 40 BioSamples are SAMN11258200 to SAMN11258239. All other original data files are available upon request.

Supplement Sponsorship and Disclosure: Supplemental Tables 1–9 and Supplemental Figures 1–4 are available from the “Supplementary data” link in the online posting of the article and from the same link in the online table of contents at https://academic.oup.com/jn/.

Abbreviations used: ALT, alanine aminotransferase; AST, aspartate aminotransferase; ASTX, astaxanthin diet; BCO2, beta-carotene oxygenase 2; BMD, bone mass density; BWG, body weight gain; CHOL, cholesterol; CONT, control nonpurified AIN93M diet; CRP, C-reactive protein; F/B ratio, Firmicutes/Bacteroidetes ratio; FBG, fasting blood glucose; GLP-1, glucagon-like peptide 1; HSP60, heat shock protein 60; KO, knockout; KO-ASTX, BCO2 knockout mice fed astaxanthin diet; KO-CONT, BCO2 knockout mice fed control nonpurified diet; LBP, lipopolysaccharide binding protein; LDH, lactate dehydrogenase; LPS, lipopolysaccharide; NLRP3, NOD-, LRR- and pyrin domain–containing protein 3; OTU, operational taxonomic unit; PGC-1 α, peroxisome proliferator-activated receptor gamma coactivator 1-alpha; RER, respiratory exchange ratio; SCFA, short chain fatty acid; Sirt3, sirtuin 3; TG, triglyceride; TNF-α, tumor necrosis factor alpha; VCO₂, CO₂ production; VO₂, O₂ consumption; WT, wild type (C57BL/6 J) mice; WT-ASTX, wild type (C57BL/6 J) mice fed astaxanthin diet; WT-CONT, wild type (C57BL/6 J) mice fed control nonpurified diet; 16S rRNA, 16S ribosomal RNA.