Prebiotic‐Based Nanoamorphous Atorvastatin Attenuates Nonalcoholic Fatty Liver Disease by Retrieving Gut and Liver Health

The pathogenesis of nonalcoholic fatty liver disease (NAFLD) is multifactorial and composite, with the disorder of lipid metabolism‐induced lipotoxicity being one of the main risk factors. Atorvastatin (AT), the most widely prescribed lipid‐lowering drug, has pleiotropic actions benefiting NAFLD treatment. However, low absorption rate in the gut and potential disruption of AT on gut flora hindered its further applications. Notably, gut dysbiosis is involved in and is thus a promising management strategy for NAFLD. In this study, we constructed a prebiotic‐based AT nanoamorphous (PANA) to improve the efficacy of AT against NAFLD by retrieving liver and gut health. After oral administration, PANA showed superior drug accumulation in the liver tissue compared with pure AT. Moreover, PANA intervention effectively restored gut healthiness, indicated by reconstructed gut flora, and improved intestinal immunity, barrier integrity, and inflammation. Consequently, compared with AT, PANA treatment caused profound inhibition of weight gain and fat deposition, decreased plasma lipid levels, and alleviated hepatic steatosis and liver inflammation. The transcriptome analysis in the gut and liver tissues identified improved immunity and inflammation as potential mechanisms. This study suggests a promising strategy to treat NAFLD, assisted with nanotechnology in synergy with functional biomaterials.

statin exposure could reduce the risk of disease progression, hepatic decompensation, hepatocellular carcinoma (HCC) development, and death of patients with cirrhosis and pre-cirrhotic conditions. [8] Although the clinical importance of AT in NAFLD has been adequately investigated, the European Association for the Study of the Liver/Diabetes/Obesity does not suggest statin administration for NAFLD/nonalcoholic steatohepatitis [9] because of safety concerns. [10] A previous study showed that the oral bioavailability of AT was only around 14%. [11] Poor bioavailability needs a high dose but low systemic exposure, resulting in adverse effects, especially in light of frequent polypharmacy in elderly patients. [12] Most recent research indicated that high residue content in the gastrointestinal tract might cause the intervention of AT with gut microbiota, leading to gut dysbiosis, which was the potential mechanism underlying statin-associated metabolic effects. [13] Therefore, it is essential to improve the AT formulation because of its low oral bioavailability [14] and gut dysbiosis. Amorphous solid dispersion (ASD) has been considered one of the most promising strategies for improving the oral absorption of l active pharmaceutical ingredients with low aqueous solubility. [15] Among previous approaches, ASD performed using spray freeze-drying (SFD) has conventional benefits with unique characteristics, including less thermal degradation, formation of stable spherical particles with high porosity, and molecular stability inhibiting recrystallization. [16] Interestingly, a special bioactive material, prebiotics (including inulin, oligofructose, glycan, and algae), was reported as a promising reagent for SFD preparation. [17] Prebiotics are no digestible food ingredients that promote the growth of beneficial microorganisms in the gut and improve the health of individuals. Emerging evidence indicates that prebiotics can improve NAFLD by alleviating gut dysbiosis. [17d,17e] A randomized, double-blind, placebo-controlled clinical three-arm trial, including 60 patients with biopsy-proven NAFLD, showed that metronidazole followed by inulin supplementation could reduce the alanine aminotransferase (ALT) level beyond that achieved after consuming a very-lowcalorie diet. [17d] The same results were achieved by Javadi and colleagues, showing that supplementation with inulin improved the levels of liver aminotransferase enzymes and recovered the grade of fatty liver in patients with NAFLD. [18] Recently, the modulatory effect of materials on gut flora has been enthusiastically studied. [19] In a review article, Lin et al. stated that "a better understanding of the interaction of materials with the gut microbiome (GM) would promote the development of new nanomedicines." [20] The present study aimed to develop a prebiotic-based nanoamorphous for AT (PANA) to improve its effect on NAFLD treatment and elucidate the underlying improvement mechanism. We hypothesized that the anti-NAFLD effect could be amended due to the enhanced drug absorption combined with the restoration of gut homeostasis. The study planned to validate the assumption through the proposed experiments (Scheme 1).

Preparation and Characterization of PANA
In this study, PANA was prepared using SFD method to improve the absorption of AT and reduce its side effects on gut dysbiosis.
The preparation scheme is shown in Figure 1A. Furthermore, scanning electron microscopy (SEM) suggested cage-like spherical particles with porosity in PANA ( Figure 1B and S1, Supporting Information). The inulin and AT spectrums powder were acquired by optical photothermal infrared (O-PTIR). [21] The representative Raman spectrum of AT inulin is shown in Figure S2, Supporting Information, and those randomly selected samples in PANA are shown in Figure 1C (left). The carbonyl absorption at 1,523 cm À1 in AT has no signal in the inulin sample and could be used for mapping AT distribution. Following the mapping with 1,523 cm À1 (green) and 1,036 cm À1 (red), the images of active AT and prebiotic matrix showed the same distribution in the scanning area (yellow), which means the AT and inulin were mixed equally in particles Figure 1C (right). Furthermore, benzene absorption in 1,580 cm À1 for AT, strong C-C absorption in 1,457 cm À1 , and C-O stretching in 1,164 cm À1 for inulin could be observed in PANA particles. Moreover, the encapsulation and relatively uniform distribution of active drug in prebiotic matrix was clearly confirmed by confocal analysis ( Figure S3, Supporting Information). Through dynamic light scattering characterization, we found that the PANA had a particle size of 48.5 μm, with a drug loading of 8.83 AE 0.43%. In addition, the amorphous formation of active drug was further proved by Fourier transform infrared spectroscopy (FTIR), powder X-ray diffraction (PXRD), and differential scanning calorimetry (DSC) experiment ( Figure 1D-F, and Supporting Information). Also, the viability of the HepG2 cells evaluated using MTT assay reveals nontoxic effect of the experiment AT, inulin, physical mixture (PAPM), and PANA on the cells ( Figure S4, Supporting Information).

Dissolution and Absorption of at Were Improved by PANA
According to the biopharmaceutical classification system, AT is a typical class II drug, implying solubility and dissolution velocity largely determine its bio efficiency. Previous research regards the manipulation of AT particle size or crystalline state enhanced solubility. Therefore, the dissolution and absorption of AT by PANA were investigated.

Dissolution Rate
Dissolution test was carried out at different dissolution mediums (pH1.5, pH 4.5, and pH 6.8) to evaluate dissolution profile of AT formulation. As shown in Figure 1G, the release of raw drug was 15.18 AE 0.98% and 29.76 AE 2.95% in pH 1.5 and pH 4.5 media over 60 min, whereas the PANA exhibited a significant enhancement in drug release than that of AT and PAPM. In pH 6.8 media, around 90% AT was released from PANA within 10 min. It is worth of note that the release rate was quite slower in raw AT compared to PANA. The enhancement of dissolution rates could be due to nanodispersion of AT within dispersed agents and the inhibition of drug recrystallization.

In Vitro Absorption and Liver Deposition
The internalization into hepatic cells of AT, PAPM, and PANA into hepatic cells was detected using the ambient mass spectrometry imaging method described before. [22] The experimental scheme is shown in Figure S5, Supporting Information. As shown in Figure 1H,J, after 1 and 4-h incubation, the cells treated with PANA displayed higher absorption of AT than the cells treated with raw drug or PAPM, indicating an increased cellular uptake of PANA.

Liver Deposition
The liver is the target tissue for AT in the treatment of NAFLD. In the present study, the deposition of AT in liver tissue was studied after gavaging the C57 mice with raw drugs, PAPM, and PANA. Ambient mass spectrometry results ( Figure 1I,K) showed that PANA reached a higher level in the liver tissue of experimental animals than raw drug and PAPM did.

PANA Improved Gut Homeostasis in NAFLD Mice
Gut dysbiosis is closely related to the pathology of NAFLD, and we further investigated the intervention of AT, inulin, PAPM, and PANA on gut homeostasis.

PANA Modulated the Composition and Function of Gut Microbiota
Fresh fecal samples were collected from mice in six different treated groups (normal control: NC; high-fat diet (HFD)-feeding model control: MC; AT intervention: AT; inulin: INU; AT and inulin physical mixture: PAPM; prebiotic-based nanoamorphous for AT: PANA) for 16S rRNA analysis to explore the influence of Scheme 1. A) Schematic illustration of the PANA. B) A PANA was developed to improve the efficacy of AT against NAFLD by retrieving liver and gut health. After oral administration, PANA showed superior drug accumulation in the liver tissue compared with pure AT. Moreover, PANA intervention effectively restored gut healthiness, indicated by reconstructed gut flora, and improved intestinal immunity, barrier integrity, and inflammation. Consequently, compared with AT, PANA treatment caused profound inhibition of weight gain and fat deposition, decreased plasma lipid levels, and alleviated hepatic steatosis and liver inflammation. This study suggested a promising strategy to treat NAFLD by dual-targeting liver and gut homeostasis.
www.advancedsciencenews.com www.small-structures.com  As shown in Figure 2A, the richness and diversity of gut bacteria markedly decreased in HFD mice (Shannon index, P < 0.001; Chao index, P < 0.01) but were reversed by AT, INU, PAPM, and PANA treatments. The principal coordinate analysis revealed that HFD significantly shifted the GM along the PC1 level. AT treatment reshaped gut microbiota composition along the PC2 level. PAPM and PANA groups showed similar gut flora structures with NC mice. Nonmetric multidimensional scaling analysis and principal coordinates analysis (PCA) ( Figure 2B and S6, Supporting Information) revealed a similar trend.
The bacterial phenotypes of different groups were studied at different categorical levels. Further, 10 phylum, 15 classes, 47 orders, 88 families, 172 genera, and 286 species with relative abundance >0.01 were annotated, with the most abundant taxa at each level shown in Figure 2C-H. At the phylum level, HFD significantly reduced the relative abundance of Bacteroidetes but increased the richness of Firmicutes; also, the Firmicute/ Bacteroidete ratio increased. PANA treatment significantly reversed these changes ( Figure 2C). The proportion of Actinobacteria and Proteobacteria, two pathogenic phyla increased by AT, were lessened in the PANA group ( Figure S7, Supporting Information). Moreover, the correlation network between intestinal flora and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway demonstrated that Bacteroidetes had an opposite effect compared with Firmicutes, including glycan biosynthesis and metabolism, carbohydrate metabolism, lipid metabolism, digestive systems with; an opposite effect compared with Actinobacteria on xenobiotics biodegradation and metabolism pathway, biosynthesis of other secondary metabolites, and glycan biosynthesis and metabolism; and an opposite effect compared with Proteobacteria on transport and catabolism ( Figure 2D). At the class level, PANA treatment, in the same way as in NC mice ( Figure 2E), modulated the eight most richly expressed classes ( Figure S8, Supporting Information). We observed a similar trend at the order and family levels in terms of increasing the abundance of beneficial bacteria and controlling the pathogenic microorganisms ( Figure 2F,G and S9, Supporting Information). At the genus level ( Figure 2H and S10, Supporting Information); AT increased the production of probiotics, including Muribaculum, Atitipes, Enterorhabdus, Lachnospiraceae-UCG-001, Lactobacillus, Lachnospiraceae-NK4A136, and Aerococcus. However, AT increased the abundance of pathogenic bacteria, such as Acinetobacter, Clostridia, Corynebacterium, and Staphylococcus, [23] but decreased the proportion of probiotics, including Curibaculum, Prevotellaceae, Bacteroides, Alloprevotella, and Prevotellaceae. Interestingly, PANA could alleviate the disturbance of AT on specific metabolism-related bacterial phenotypes but maintained its beneficial effect in terms of boosting probiotics assisted by the inulin effect.
Bacteroides and Lactobacillus are bacteria producing shortchain fatty acids (SCFAs). SCFAs are important for maintaining the integrity of colonic epithelium and regulating glucose homeostasis, lipid metabolism, and inflammatory response. [24] Therefore, we mapped the KEGG orthology (KO) enzymes with KEGG and obtained seven enzymes involved in the SCFA metabolism pathway. The results ( Figure 2I)  3] were positively associated with beneficial bacteria, especially Lactobacillus and Lacnospiraceae, but negatively correlated with pathogenic microbiota Jeotgalicoccus and Staphylococcus. All these enzymes belonged to the carbon fixation pathway and propanoate metabolism pathway. Further analysis indicated that HFD disturbed the SCFA-generating capacity of gut microbiota ( Figure 2I-b,I-c and S11, Supporting Information), but the disorder was restored with PANA intervention. Consistently, PANA treatment increased the acetate, butyrate, and propionate levels in the feces of HFD-fed mice ( Figure S12, Supporting Information). PANA-treated mice showed longer colon lengths compared with the MC mice ( Figure 2J,L). In addition, the microvilli in NC mice were dense and even, while those in HFD-fed mice were short and sparse ( Figure 2K,M).

PANA Adjusted Gut Immunity
RNA-seq on the intestine tissues was performed to elucidate the pharmacological effect of PANA on gut health and its potential contribution to NAFLD. PANA groups showed a similarity in structure in transcriptome compared with NC mice ( Figure 3A). With a fold change of ≥2.0 and P value ≤0.05 as the selection criteria, 1,095 differentially expressed genes (DEGs, 545 upregulated and 550 downregulated) were detected from the PANA and MC groups ( Figure 3B). The gene ontology (GO) analysis revealed that the genes related to the "immune system process" and "innate immune response" were enriched ( Figure 3C) and significantly modulated by INU, PAPM, and PANA treatments in HFD-fed mice ( Figure 3D). The disruption of gut homeostasis led to an altered immune state and various liver diseases through the gut-liver axis. Multiple cell types, including macrophages, neutrophils, and other immune cell types, were involved in fatty liver disease inflammation. [25] Excessive recruitment of neutrophils and M1-type pro-inflammatory macrophages in the intestine results in mucosal damage and progression toward chronic inflammation. [26] we gated neutrophils as CD11bþ/Ly6Gþ cells from CD45þ/CD3À/CD14À populations, M1 macrophages as CD206À/Cd11cþ cells, M2 macrophages as CD206þ/Cd11c-cells from CD45þ/Ly6GÀ/F4/80þ/Cd11bþ populations. As shown in Figure 3E,F, HFD treatment induced an increased infiltration of neutrophils and M1-type macrophages; PANA suppressed the former phenomenon and improved the M2-type macrophages. Besides, the multicolor immunofluorescent of colon tissue showed consistent results ( Figure 3G).   Moreover, we also found that PANA restored T lymphocytes by decreasing the proportion of pro-inflammatory Th1 cells while increasing the population of Treg cells ( Figure 3H-J).

PANA Improved Gut Inflammation and Integrity
Numerous studies demonstrated that gut dysbiosis caused dysregulated mucosal immune response, leading to exacerbate gut inflammation and integrity loss and, subsequently, NAFLD through the gut-liver axis. Consistently, the pathways of "inflammatory response", "regulation of inflammatory response", and "cellular response to lipopolysaccharide" were enriched from GO analysis. We remapped the 122 DEGs obtained from these inflammation-related pathways to the KEGG database and observed that the genes in the "TNF-α", "NF-κB", "Toll-like receptor", "cytokine-cytokine receptor interaction", and "chemokine" signaling pathways were enriched ( Figure 4A).
We further compared the expression of key genes among the experimental groups. The clustered DEGs from the "TNF-α signaling pathway" and "Toll-like receptor signaling pathway" in the heat map ( Figure 4B) and the mRNA expression profile of key genes involved were also assessed through box diagrams ( Figure 4C). As shown in Figure 4B,C, and S13, Supporting Information, the expression of Tlr4, MyD88, Traf6, and Nfkb1 significantly increased in MC mice compared with NC mice. HFD also upregulated the expression of inflammatory cytokines and chemokines, including TNF-α, Il1b, Il6, and Mmp9, as well as that of the immunoglobulin superfamily of endothelial adhesion molecules, such as vascular cell adhesion molecule and intercellular cell adhesion molecule. PANA intervention markedly abrogated these disorders, which was probably attributed to the modulating effect of inulin. The signal transduction pathway is shown in Figure 4D; it was proved in intestinal tissues by reverse transcription-polymerase chain reaction (RT-PCR) analysis ( Figure 4E). Subsequently, the production of inflammatory cytokines in the intestine was eased with PANA intervention ( Figure 4F,J). Next, we investigated the effect of experimental agents on gut integrity and found that HFD-fed C57 mice had fewer goblet cells than mice in the NC group, while inulin and PANA treatment significantly increased the number of goblet cells ( Figure 4G). Furthermore, HFD-fed C57 mice had significantly low expression of epithelial tight junction proteins, such as ZO-1 and occludin, compared with NC mice (Figure 4H-J). The PANA supplementation prevented this reduction and maintained normal levels. These findings indicated that PANA improved intestinal health, which might benefit NAFLD treatment.

PANA Alleviated HFD-Induced Liver Injury in a Mouse Model
The pharmacological effects of PANA on the metabolic disorder and liver injury were further examined in an HFD-induced NAFLD mouse model.

PANA Alleviated Metabolic Disorders in HFD-Fed C57 Mice
As shown in Figure 5A-D, HFD supplementation for 10 weeks induced a significant increase in body weight and epididymis fat indices with no extra food consumption ( Figure S14, Supporting Information). PANA intervention significantly diminished fat deposition in HFD-fed animals compared with animals with no treatment. PAPM exposure also reduced the levels of these markers to a lesser extent. Furthermore, PANA reduced the high levels of LDL-c, TC, TG, and glucose-induced by HFD but did not influence HDL-c levels ( Figure 5E). AT and inulin were also found to reduce blood lipid levels but to a lesser extent.

PANA Alleviated Liver Injury in the Mouse Model
HFD-fed C57 mice were administered with different agents (equivalent of AT 6 mg/(kg À1 day À1 ) by oral gavage in the following 10 weeks. As shown in Figure 5F, the liver/body weight index in the HFD-fed mice was remarkably higher than that in the normal chow diet-fed mice (NC), while no differences were found between HFD-fed mice (MC) and those with no treatment. However, the liver weight increased by HFD was significantly reduced by PANA treatment ( Figure 5G). In addition, reduced hepatocyte ballooning was observed in mice treated with PANA on hematoxylin and eosin staining ( Figure 5H). Oil Red O staining showed that lipid accumulation in HFD-fed mice was significantly reduced by PANA, which was supported by the liver (TG) levels ( Figure 5I,J). The serum levels of ALT and aspartate aminotransferase were lower in PANA-treated mice than in MC mice ( Figure 5K). Furthermore, the cholesterol ester (CE) levels (including total CE, saturated CE, polyunsaturated CE, and monounsaturated CE) were significantly higher in the HFD-fed group than in the NC group and were significantly reduced by ATPM treatment (Figure 5L,M).

Mechanism Study
To interpret the mechanisms, we performed RNA-seq on the liver tissues of mice from each group. As shown in Figure 6A, 1,834 genes were differentially modulated between the drug-treated and HFD groups. Unsupervised PCA and hierarchical clustering clearly separated the samples from the MC mice and the NC mice into two clusters. PANA groups showed a similarity in structure in transcriptome with NC mice ( Figure 6B). With a fold change of ≥2.0 and P value ≤0.05 as the selection criteria, 857 DEGs (127 upregulated and 62 downregulated) were detected from the PANA and MC groups ( Figure S15, Supporting Information). The KEGG analysis revealed that the DEGs were mainly enriched in "Cholesterol metabolism", "PPAR signaling pathway", "Insulin resistance", "Nonalcoholic fatty liver disease", "AMPK signaling pathway", and "lipid metabolic pathways" (Figure 6C and S16, Supporting Information). The expression of genes related to fatty acid uptake, including Cd36, Pcsk9, Abca1, Fabp4, Fasn, Scd1, and Srebf1, increased in the MC group compared with the NC group. PANA exposure significantly reduced the disorder. Furthermore, fat degradation and insulin sensitivity genes (Irs and Cpt1α) increased after PANA treatment ( Figure 6D,E, and S17, Supporting Information). The expression of AKT, AMPK, and PPAR in liver tissue was further identified by Western blot and multifluorescent staining assay ( Figure 6F,J). www.advancedsciencenews.com www.small-structures.com www.advancedsciencenews.com www.small-structures.com www.advancedsciencenews.com www.small-structures.com

PANA Modulated Immune Cell Infiltration and Inflammation in Liver Tissue
Inflammation and hepatocyte injury are the hallmarks of NAFLD, and innate immune activation is a key factor in triggering and amplifying hepatic inflammation. [27] We further investigated the inflammation status of liver tissues from experimental mice. The flow cytometry analysis illustrated that HFD induced the infiltration of M1-type monocytes in liver tissues; PANA exposure suppressed these phenomena. Subsequently, the liver tissues in the NC mice expressed low levels of IL-1β, IL-6, and TNF-α, whereas the levels of these markers increased dramatically in MC mice ( Figure 6G,H). PANA treatment significantly reduced the expression of these pro-inflammatory factors compared with no treatment. AT and inulin were also found to reduce the inflammation status ( Figure 6I,J).

Mechanisms Were Verified using the Coculture System
The mechanisms were investigated in a triple-culture system with HepG2 cells, THP cells, and gut microbiota. The experimental scheme is shown in Figure 7A. The Trans well filters with differentiated THP cells were transferred onto well plates containing HepG2 cells. The fecal fluids and AT-containing formulations (equivalent to 10 μg mL À1 final concentration) were added to the system and incubated for 8 h at 37°C in the presence of 5% CO 2 . As shown in Figure 7B, the fecal fluids from MC mice caused marked lipid accumulation in HepG2 cells. AT intervention inhibited the increase in intracellular lipid accumulation. PANA showed a higher capacity for these effects compared with pure AT and PAPM when AT concentrations were equal. Inulin also showed an effect on lipid accumulation. These results indicated that increased absorption and modification of gut bacteria assisted by PANA might benefit pharmacological efficacy. As illustrated in Figure 7C,E, the percentage of M1 phenotype and the ratio of M1/M2 macrophages were significantly increased, while the percentage of M2-type macrophages were decreased by fecal fluids gathered from MC mice; these phenomena were inhibited by PANA intervention. AT treatment tended to increase the percentages of M2-type macrophages. A proportion of M1-type macrophages were found in the inulin group, implying the potential effect of gut microbiota. Consistently, the increased production of inflammatory factors induced by fecal fluids was greatly suppressed by PANA ( Figure 7D,F). AT, inulin and PAPM treatment tended to reduce TNF-α, IL-1β, and IL-6 levels, but the differences are not significant. The superior effect of PANA was probably attributed to the improved absorption of active drugs and the modulatory effect on gut bacteria.

In Vivo Safety
The long-term safety was investigated to determine the biocompatibility of PANA. The C57 mice received PANA (6 mg kg À1 day À1 of AT, n ¼ 7) for 10 weeks by gavage. Animals injected with saline were used as controls. As shown in Figure S18, Supporting Information, no noticeable histological differences in major organs were observed between the two groups. In addition, no significant differences in the plasma levels of ALT, AST, creatinine, and blood urea nitrogen were found between the two groups. These results indicated good safety and tissue compatibility of PANA.

Discussion
In this study, a prebiotic-based nanoamorphous for AT (PANA) was developed to improve its effect on NAFLD treatment. The results showed that the liver drug deposition of AT was increased by PANA. PANA also improved gut health, indicated by improved gut microbiota, intestinal immunity, and inflammation. In HFD-fed C57 mice, PANA exhibited a superior effect on metabolic disorders and liver injury than the pure AT did. The benefits were probably due to its synergistic effect on the gut ecosystem and liver disorders. The pathological mechanism underlying the onset and progression of NAFLD is complicated and has not been fully elucidated. Disorder of lipid metabolism-induced lipotoxicity was deemed as one main risk factors. [2] Statins, especially AT, are the most widely prescribed lipid-lowering drugs. However, low aqueous solubility and poor bioavailability limited the utilization. A previous study showed that the oral bioavailability of AT was only around 14%. [11] Thus, various methods have been applied to improve the dissolution and oral bioavailability of AT. Of these, SD, a supersaturated formulation approach, has been considered one of the most practical and effective methods widely explored in the pharmaceutical industry. Several techniques were employed to generate SDs, such as milling, spray drying, freeze-drying, use of supercritical fluids, and organic drying. SFD is a novel particle engineering method widely investigated in the food and pharmaceutical industries. [17b] It possesses the merits of the conventional lyophilization process and spray drying capability with unique characteristics as follows: SFD yields spherical particles of controllable size, lyophilization avoids heatinduced drug degradation or phase separation of the hydrophilic carrier and lipophilic drug, and porous particles accelerate the aqueous dissolution of lipophilic drugs, subsequently providing a high and reproducible bioavailability. Furthermore, it is more stable during storage, inhibiting crystallization. [17b] A special bioactive ingredient: prebiotics (including inulin, oligofructose, glycan, and algae) was reported as a promising dispersing agent for SFD preparation. [17aÀc] In the present study, AT was successfully encapsulated in cage-like spherical particles. With O-PTIR spectrum analysis, the relatively uniform distribution of active drugs in the prebiotic matrix was clearly confirmed. In addition, the amorphous formation of active drugs was further proved by FTIR, PXRD, and DSC experiments. AT is a typical class II drug, implying dissolution velocity largely determines its bio efficiency. Previous research regards the manipulation of AT particle size or crystalline state enhanced solubility. Consistently, PANA exhibited a significant enhancement in drug release than that of AT and PAPM. The improvement was probably due to the nanoamorphous status of AT within prebiotic agent and the inhibition of drug recrystallization. Consequently, augmented AT deposition in the liver tissue was detected in PANA-treated mice.
Recently, accumulating studies demonstrated that gut microbiota played a crucial role in the pathophysiology of NAFLD. [28]  The analysis of GM composition of patients with NAFLD and its comparison with that of controls revealed that GM dysbiosis and NAFLD-related GM signatures were reported in several studies, which suggested that GM is a noninvasive biomarker for NAFLD. [28,29] Indeed, the prospect of manipulating the GM offers an exciting and novel strategy for disease prevention and control of NAFLD, with the administration of prebiotics in clinical trials. The influence of statins on gut microbiota and the consequences on liver diseases have gained increasing attention but with inconsistent results. [13,30] Zhang et al. suggested that AT restored cholesterol-induced gut microbiota dysbiosis and prevented the development of NAFLD-HCC. [30a] Roessler et al. showed that the regulatory impact of AT on the gut microbial profile played a crucial role in the effects on blood lipids.
[30b] However, Su et al. showed that the cholesterollowering effect of AT did not contribute to the gut modulation effect as probiotics did. [31] Furthermore, Caparrós-Martín et al. demonstrated that statin therapy led to profound gut dysbiosis, which resulted in increased fasting blood glucose levels and weight gain. [13] Our research revealed that AT increased the proportion of some beneficial bacteria. However, AT promoted the production of many pathogenic bacteria and reduced the richness of some symbiotic microbiota. Interestingly, Inulin addition could alleviate the disturbance of AT on some metabolismrelated bacterial phenotypes but had a synergistic effect with AT in increasing probiotics and inhibiting pathogenic bacteria. PANA exposure maintained the beneficial effect of AT in terms of boosting probiotics while alleviating its disturbing effect on specific metabolism-related bacterial phenotypes. At the phylum level, PANA administration significantly decreased the relative abundance of Firmicutes, increased the relative abundance of Bacteroidetes, and then increased the ratio of Bacteroidetes/ Firmicutes compared with those in the HFD group. This was consistent with many clinical and animal studies showing that NAFLD was associated with the disruption of the balance between Firmicutes and Bacteroidetes. [32] Moreover, the proportions of actinobacteria and proteobacteria (two pathogenic phyla), increased by AT, were lessened after PANA treatment. [33] At the genus level, PANA enhanced the effect of AT on probiotics, including Muribaculum, Atitipes, Enterorhabdus, Lachnospiraceae-UCG-001, Lactobacillus, Lachnospiraceae-NK4A136, and Aerococcus. Furthermore, PANA decreased the abundance of pathogenic bacteria, such as Acinetobacter, Clostridua, Corynebacterium, and Staphylococcus enriched in the AT group and restored the proportion of probiotics, including Curibaculum, Prevotellaceae, Bacteroides, and Alloprevotella dismissed in AT mice. Bacteria such as Alloprevotella, Bacteroides, and Prevotellaceae have been reported as protective against NAFLD. [34] Bacteroides and Lactobacillus are SCFA-producing bacteria. [35] SCFAs are important for maintaining the integrity of the colonic epithelium and regulating glucose homeostasis, lipid metabolism, immunity, and inflammatory response. [24] The correlation heat-map analysis revealed that the main enzymes involved in the SCFA metabolism pathways, especially the carbon fixation pathway and propanoate metabolism pathway, positively correlated with the bacteria modulated by PANA. The gut immunity and integrity impaired by HFD were restored.
According to the alterations of gut flora, PANA significantly increased the proportion of M2-type macrophages and Treg cells and decreased the number of neutrophils, M1-type macrophages, and Th1 cells, which were associated with inflammatory activity and fibrosis grade in NAFLD. [36] Consequently, the transcriptome analysis in the gut tissues identified that PANA inhibited the TLR4/Myd88/NF-κB signaling pathway and prevented the expression of downstream inflammatory factors. The members of the immunoglobulin superfamily of endothelial adhesion molecules, vascular cell adhesion molecule (VCAM-1), and intercellular cell adhesion molecule (ICAM-1) participated in all stages of metabolic diseases and were also downregulated by PANA. [37] Interestingly, these results were in consistent with the improvement in liver inflammation by PANA. Inflammation and hepatocyte injury are the hallmarks of NAFLD, and innate immune activation is a key factor in triggering and amplifying hepatic inflammation. [27] In the present study, PANA exposure suppressed the infiltration of M1-type monocytes, decreased the M1/M2 ratio in liver tissues, and subsequently reduced the expression of pro-inflammatory factors. In vitro experiments also verified that modulation of gut bacteria contributed to the inhibition of steatosis and inflammation of liver cells. It was worth noting that PAPM group showed similar effect with PANA on gut homeostasis, which indicated the benefits might probably be attributed to the modulating effect of inulin.
The pharmacodynamic experiment showed that PANA effectively ameliorated metabolic disorders and liver injury in HFDfed C57 mice. The transcriptome analysis in liver tissue indicated that differential genes were mainly enriched in "cholesterol metabolism", "PPAR signaling pathway", "insulin resistance", "nonalcoholic fatty liver disease", "AMPK signaling pathway", and "lipid metabolic pathways". The expression of genes related to fatty acid uptake, including Cd36, Pcsk9, Abca1, Fabp4, Fasn, Scd1, and Srebf1, decreased in the PANA-treated mice compared with MC mice. Moreover, the levels of fat degradation and insulin sensitivity genes, such as Irs and Cpt1α, increased after PANA treatment. The improved effect was probably attributed to the increasing of drug absorption and retrieving of gut homeostasis.
Recently, the modulatory effect of materials on gut flora has been enthusiastically studied. [10b,11À15] In a review article, Lin et al. stated that "a better understanding of the interaction of materials with the GM would promote the development of new nanomedicines". [21] In this study, we, for the first time, elucidated the dual effect of inulin as a nanomedicine excipient on drug absorption and gut microbial homeostasis.

Conclusion
In this study, a PANA was developed to improve the therapeutic efficacy of AT on NAFLD. After oral administration, PANA showed superior drug accumulation in the liver tissue compared with pure AT. In the high-fat diet-fed C57 mice, PANA (6 mg kg À1 of AT) intervention effectively restored gut healthiness, indicated by reconstructed gut flora, and improved intestinal immunity, barrier integrity, and inflammation. Moreover, compared with AT, PANA treatment caused profound inhibition of weight gain and fat deposition, decreased plasma lipid levels, and alleviated hepatic steatosis and liver inflammation. The transcriptome analysis in the liver tissues identified improved inflammation and energy homeostasis as potential mechanisms.
In conclusion, PANA is promising for the treatment of NAFLD assisted with nanotechnology in synergy with functional biomaterials.

Supporting Information
Supporting Information is available from the Wiley Online Library or from the author.