Study of growth, metabolism, and morphology of Akkermansia muciniphila with an in vitro advanced bionic intestinal reactor

As a kind of potential probiotic, Akkermansia muciniphila abundance in human body is directly causally related to obesity, diabetes, inflammation and abnormal metabolism. In this study, A. muciniphila dynamic cultures using five different media were implemented in an in vitro bionic intestinal reactor for the first time instead of the traditional static culture using brain heart infusion broth (BHI) or BHI + porcine mucin (BPM). The biomass under dynamic culture using BPM reached 1.92 g/L, which improved 44.36% compared with the value under static culture using BPM. The biomass under dynamic culture using human mucin (HM) further increased to the highest level of 2.89 g/L. Under dynamic culture using porcine mucin (PM) and HM, the main metabolites were short-chain fatty acids (acetic acid and butyric acid), while using other media, a considerable amount of branched-chain fatty acids (isobutyric and isovaleric acids) were produced. Under dynamic culture Using HM, the cell diameters reached 999 nm, and the outer membrane protein concentration reached the highest level of 26.26 μg/mg. This study provided a preliminary theoretical basis for the development of A. muciniphila as the next generation probiotic.

In terms of in vitro culture of A. muciniphila, the traditional way was to use brain heart infusion broth (BHI) or BHI + porcine mucin (BPM) in an anaerobic bottle and placed it in an anaerobic environment for static fermentation culture [29][30][31]. Porcine mucin (PM) was the key component of A. muciniphila media in most reports. By adding a certain amount of mucin to BHI to cultivate A. muciniphila, the nutritional requirements for its growth and metabolism could be met. For simulation of A. muciniphila physiological state in human body, using human mucin (HM) instead of PM for in vitro culture is one key issue of great significance. However, as far as we know, the in vitro culture of A. muciniphila using HM has not been reported yet.
For culture method, the currently used A. muciniphila static culture is easy to cause accumulation of metabolites, which in turn inhibits the growth of A. muciniphila. In addition, A. muciniphila cells are difficult to harvest because many of them adhere to bottle bottom and wall. The in vitro dynamic fermentation reactor is a recently developed tool for studying the intestinal flora, mainly through bionic technology to simulate the real environment in digestive system [32]. In this study, an advanced bionic intestinal reactor (ABIR) was built to simulate the peristalsis of the real human colon, the micro-ecological environment of intestinal and the absorption of flora metabolites. As the traditional static culture state is far from the real A. muciniphila growth environment in human body, using ABIR as the A. muciniphila culture equipment is another key issue of great significance for simulating in vivo A. muciniphila physiological environment.
In this study, in order to better understand the growth, metabolism and morphology of A. muciniphila in real human body, in vitro dynamic A. muciniphila fermentation was implemented using five different media including HM and BHI + human mucin (BHM) in ABIR. The discovery was an improvement in culture method and media for the development of A. muciniphila as the next generation probiotic.

Results
Effects on cell growth and metabolites of A. muciniphila under static and dynamic cultures As shown in Fig.1a, the logarithmic growth periods of static cultures and ABIR dynamic culture were both 6-10 h, then the bacteria entered a stable stage after 12 h. The biomass under the ABIR dynamic culture increased steadily after 37 h and reached 1.92 g/L at 48 h, which was 44.36% higher than that of the static culture (1.33 g/L). For both static and dynamic cultures, the type and concentration of metabolites were shown in Fig. 1b. Under dynamic culture, the metabolites were mainly acetic acid, propionic acid, butyric acid, isobutyric acid, and isovaleric acid yields, which reached 47.05, 10.21, 9.87, 2.78, and 4.06 mM, respectively. The differences in yields of butyric acid between two cultures were significant (one-way ANOVA analysis with Tukey's test, p < 0.05, the same below). Simultaneously, considerable differences were Fig. 1 Effects on cell growth and metabolites of A. muciniphila under static culture and dynamic culture. Note: The data are shown as the mean ± SD (n = 3) and analyzed using one-way ANOVA with Tukey's test, *p < 0.05, **p < 0.01 observed in terms of yields of propionic acid, isobutyric acid and isovaleric acid (p < 0.01).
Effects of different media on A. muciniphila growth under dynamic culture As shown in Fig. 2a, using PM and HM, A. muciniphila grew slowly at 0-18 h, and the biomass of A. muciniphila grown on PM (0.68 g/L) was higher than that on HM (0.35 g/L). During this period, the secretion rates of glycosidase and protease accelerated, resulting in increased degradation of mucin. After this period, A. muciniphila adapted to the growth environment of medium. The growth rates of A. muciniphila on PM and HM were both in the logarithmic period from 18 h to 36 h. From 36 to 48 h, the biomass of A. muciniphila on HM exceeded that on PM and reached 2.89 g/L at 48 h. The results proved that A. muciniphila culture medium based on human mucin was an ideal medium for cell growth. Moreover, growth performances of A. muciniphila on BHI, BPM, and BHM were studied. As shown in Fig. 2b, the A. muciniphila grown on BPM and BHM were in logarithmic growth phase at 4-8 h, and the A. muciniphila biomass on BPM (1.26 g/L) was higher than that on BHM (0.85 g/L). The A. muciniphila grown on BHI was in the logarithmic growth phase at 8-12 h, and A. muciniphila biomass of 0.61 g/L was obtained, which was lower than that on BPM and BHM. All three batches subsequently entered a stable growth phase. At 48 h, the biomass on BHI, BPM, and BHM were 1.17, 1.92, and 1.96 g/L, respectively.

Effects of different media on A. muciniphila metabolites under dynamic culture
As shown in Fig. 3. The metabolites of A. muciniphila on PM and HM were SCFAs (acetic acid and butyric acid), whereas the metabolites of A. muciniphila on BHI, BPM, and BHM were SCFAs (acetic, propionic, butyric, valeric) and BCFAs (isobutyric and isovaleric acids). For the batches grown on PM and HM, the concentrations of acetic acid (4-6 mM) were not significantly different (Fig. 3b), but the concentrations of butyric acid were significantly different (  Effects of different media on the thickness of A. muciniphila outer membrane and diameter of cell under dynamic culture Figure 5 showed the TEM and SEM images of A. muciniphila using five different culture media under dynamic culture. The cells of A. muciniphila were round or elliptical. The cells of the batch grown on BHI grew alone or Fig. 3 Effects of different media on metabolites of A. muciniphila under dynamic culture. In this figure, the brain heart infusion broth (BHI), porcine mucin (PM), human mucin (HM), BHI+ porcine mucin (BPM), and BHI+ human mucin (BHM). Note: The data are shown as the mean ± SD (n = 3) and analyzed using one-way ANOVA with Tukey's test, *p < 0.05, **p < 0.01, ***p < 0.001 in pairs, whereas the cells of the batches grown on PM, HM, BPM, and BHM containing mucin grew in pairs or chains and even formed aggregates. As shown in Table 1, the relative thicknesses of the outer membranes of A. muciniphila in the BHI, PM, HM, BPM, and BHM were 71.66, 101.20, 104.50, 71.46, and 72.11 nm, respectively. Significant differences between the BHI and PM, PM and BPM, PM and BHM, BHI and HM, HM and BPM, and HM and BHM were observed (p < 0.01). No significant difference was observed in the other remaining groups (p > 0.05). The above results were consistent with those of the outer membrane protein concentration of A. muciniphila, indicating that increased outer membrane thickness of the cells resulted in high outer membrane protein concentration. The cell diameters of A. muciniphila grown on BHI, PM, HM, BPM, and BHM were 871, 985, 999, 980, and 971 nm, respectively. The cell diameter on BHI was significantly different with those on the other four media (p < 0.05), and no significant difference was observed between the other groups (p > 0.05). On the mucin medium, A. muciniphila had a diameter and length of 640 and 690 nm, respectively. On BHI medium, A. muciniphila had a diameter and length of 830 nm and 1000 nm, respectively.

Discussion
The static culture and ABIR dynamic culture were compared in terms of biomass and metabolites. Compared with the static culture, the ABIR simulated physic-chemical and physiological processes in the real human colon, which was an idea tool for investigating the effects of the media components on A. muciniphila fermentation. In static culture, the metabolites of A. muciniphila accumulated partially at the late stage of the culture process and high level of SCFAs (> 80 mM) inhibited cell growth [8,33]. In ABIR dynamic culture, simulation of real human colon peristalsis was achieved, and the cells, media and metabolites could be well mixed. Thus, the inhibition effects of metabolites could largely be released. Under dynamic culture, isobutyric acid and isovaleric acid yields were relatively higher than those under static culture. In the late fermentation stage of dynamic culture, the carbon source was exhausted and A. muciniphila turned to protein fermentation, thereby generated corroded toxic metabolites, such as isobutyric acid and isovaleric acid [34,35].
Using five different media (BHI, PM, HM, BPM and BHM), A. muciniphila biomass under dynamic cultures were compared. The mucin in the PM and HM were the only carbon and nitrogen sources for the growth of A. muciniphila. When A. muciniphila grew on these media, glycosidase was secreted to degrade the glycoprotein exposed at the terminal of mucin to produce Nacetylgalactosamine, N-acetylglucosamine, fucose, and galactose components, which could be used as nutrients for cells [36][37][38]. The BHI contained rich carbon and nitrogen sources, which supposed to fully support the growth of A. muciniphila. However, the biomass on BHI Fig. 4 Effects of different media on concentration of A. muciniphila outer membrane protein under dynamic culture. In this figure, the brain heart infusion broth (BHI), porcine mucin (PM), human mucin (HM), BHI+ porcine mucin (BPM), and BHI+ human mucin (BHM). Note: The data are shown as the mean ± SD (n = 3) and analyzed using one-way ANOVA with Tukey's test, *p < 0.05, **p < 0.01, ***p < 0.001 was the lowest among the three batches. When mucin was added in BHI, the biomass of A. muciniphila increased about 65%, which demonstrated that mucin was a key component for A. muciniphila growth. To find other cheaper sources instead of mucin, Derrien and Colleagues discovered that A. muciniphila could also grow on a limited amount of carbon source, including N-acetylglucosamine, N-acetylgalactosamine, and glucose [6,31]. Based on this, a mixed medium instead of mucin to cultivate A. muciniphila was formed by adding glucose, soy protein, threonine, N-acetylglucosamine and N-acetylgalactosamine together. The biomass of A. muciniphila grown on the mixed medium reached the same level of the batches grown on BPM or BHM [14,39]. In addition, studies found that adding fructooligosaccharides [40,41], metformin [42,43], polyphenols [44][45][46][47], probiotics [48][49][50] and fish oil unsaturated fatty acids [51] could increase the abundance of in vivo A. muciniphila.
Using five different media, the type and the concentration of A. muciniphila metabolites (SCFAs and BCFAs) were further studied, and an interesting conclusion was obtained. At HM and PM, the metabolites of A. muciniphila are acetic and butyric acids. The latest research showed the causal relationship between the butyric acid produced by intestinal microorganisms and the risk of diabetes [28]. Butyric acid can promote the secretion of insulin by β cells, regulate blood sugar, and improve insulin response [34,35]. Further research was needed on the mechanism of A. muciniphila main metabolites  figure, a, b, c, d, e represents SEM images of the brain heart infusion broth (BHI), porcine mucin (PM), human mucin (HM), BHI+ porcine mucin (BPM), and BHI+ human mucin (BHM), respectively. f, g, h, i, j represents TEM images of the brain heart infusion broth (BHI), porcine mucin (PM), human mucin (HM), BHI+ porcine mucin (BPM), and BHI+ human mucin (BHM), respectively (acetic acid and butyric acid) on PM and HM. Studies showed that BCFAs like isobutyrate acid and isovalerate acid, were derived from the fermentation of branched-chain amino acids. In contrast to straight chain SCFAs, these compounds were considered detrimental to colonic and metabolic health [52,53]. Compared with SCFAs, these compounds were considered harmful to the colon and metabolic health. Isobutyric acid and isovaleric acid were produced on BHI, BPM and BHM. The contents of isobutyrate acid and isovalerate acid on BPM and BHM were relatively high, indicating that the glucose or mucin used by A. muciniphila was used simultaneously with glucose.
Outer membrane protein of A. muciniphila had a specific protein named Amuc_1100, which can maintain a stable state at the temperature of pasteurization and improve the intestinal barrier [14]. Through daily quantitative A. muciniphila feeding to mice, A. muciniphila can offset the increase in body weight and fat caused by high-fat diet (HFD) and improve glucose tolerance and insulin resistance, indicating that Amuc_1100 from pasteurization A. muciniphila outer membrane protein had beneficial effects on HFD-induced metabolic syndrome. Using five different media, we quantitatively detected the content of the outer membrane protein of A. muciniphila. Results showed the concentration of the outer membrane protein significantly increased in the batches grown on media containing only PM or HM. For these batches, the Amuc_1100 protein of A. muciniphila increased accordingly (Supplementary Fig. 1). The result provided an alternative way for the improvement of Amuc_1100 protein content of A. muciniphila outer membrane.
The cell morphology and cell size were observed by electron microscope. Found that the cell size differed depending on the medium under dynamic culture and static culture. Derrien et al. found that using porcine mucin medium under static culture [6], A. muciniphila could grow in single cells or in pairs and rarely grow in chains; it usually forms aggregates, and a translucent layer of material could be observed. Using BHI, this phenomenon was rarely observed, and cells appeared alone or in pairs but rarely in groups [31]. The results were different from our discoveries because of the culture methods (the dynamic culture was carried out in this study). Note: a: Relative thickness of the outer membrane of A. muciniphila, b: Diameter of A. muciniphila. The brain heart infusion broth (BHI), porcine mucin (PM), human mucin (HM), BHI+ porcine mucin (BPM), and BHI+ human mucin (BHM). The data are shown as the mean ± SD (n = 4) and analyzed using one-way ANOVA with Tukey's test, *p < 0.05, **p < 0.01

Conclusions
Dietary changes can reveal the products and functions of the gut microbiota and their influence on human health. This study explained the growth, metabolism, and morphology of Akkermansia muciniphila in different nutrient media with an in vitro advanced bionic intestinal reactor. Our results indicated that the nutritional composition has a presumptive relationship with A. muciniphila biomass, outer membrane protein concentration and thickness, and cell diameter. At conditions containing mucin as sole carbon and nitrogen sources, the metabolites of A. muciniphila were acetic and butyric acids. This study provided a certain reference for the mechanism of A. muciniphila action in the host.

Strains and media
A. muciniphila AT56 were isolated from Chinese human feces. 16S rDNA sequencing showed that A. muciniphilas AT56 had 99% similarity with the type stain A. muciniphila ATCC BAA-835. Phylogenetic information for the type strain was reported in NCBI Sequence Read Archive (SRA) database with the Bioproject number PRJNA331216. The BHI contained 10.0 g/L tryptone, 2.5 g/L dibasic sodium phosphate, 17.5 g/L beef heart infusion, 5.0 g/L sodium chloride, and 2.0 g/L glucose and was maintained at pH 6.5.
The PM contained 4 g/L porcine mucin, 2.5 g/L disodium hydrogen phosphate, and 5.0 g/L sodium chloride and was maintained at pH 6.5(porcine mucin, Kuer, Beijing).
The HM contained 4 g/L human mucin, 2.5 g/L disodium hydrogen phosphate, and 5.0 g/L sodium chloride and was maintained at pH 6.5 (human mucin, extracted from pancreatic myxoma).
The BPM contained 10.0 g/L tryptone, 2.5 g/L disodium hydrogen phosphate, 17.5 g/L bovine heart dip powder, 5.0 g/L sodium chloride, 2.0 g/L glucose, and 4 g/L pig-derived mucin and was maintained at pH 6.5.
The BHM contained 10.0 g/L tryptone, 2.5 g/L disodium hydrogen phosphate, 17.5 g/L bovine heart dip powder, 5.0 g/L sodium chloride, 2.0 g/L glucose, and 4 g/L human mucin and was maintained at pH 6.5.
All media had the same concentration, magnetically stirred for 2 h, mixed well, and sterilized at 121°C for 30 min before use (Mucins was filtered by a 0.22 μm water membrane to isolate bacteria).
Our study was performed in accordance with the principles of the Declaration of Helsinki with regard to ethical research involving human subjects, and the protocols were approved by Medical Ethics Committee of Affiliated Hospital of Jiangnan University (approval No. S2019-012-05).

Device introduction
An in vitro ABIR was developed to simulate the environment of A. muciniphila in the real large intestine [54]. As shown in Fig. 6, the bioreactor was composed of three interconnected reaction bottles with a flexible and transparent silica gel intestine that simulated the internal intestinal wall and the shape of the bionic intestine. The silica gel membrane between the reaction flask and the large intestine (Fig. 6d) was filled with deionized water at 37°C. The simulated large intestine contracted and caused peristaltic waves by on line controlling the pressure of the water flow through a circulating water pump. Thus, the materials in the simulated large intestine were mixed and moved in the system. This mixing was better than the mixing done in the fermenter, where the phase separation of fluid and solid occurred. The reaction bottle was equipped with vacuum ( Fig. 6g) and mixed gas (Fig. 6h) control devices to continuously maintain the oxygen-free environment in the simulation chamber. The system was also added with a water absorption device that simulated the function of the large intestine (Fig. 6j). The mixture in the cavity was absorbed to simulate the formation of feces. By continuously measuring the pH (Fig. 6l) and secreting alkaline solution (Fig.  6e) to neutralize the acid, the system pH was maintained to prevent the accumulation of microbial metabolites. The system was equipped with a dialysis device, which consisted of hollow fiber nanofiltration membranes that filtered the metabolic products of microorganisms. Thus, the microbial metabolites in the reactor cavity were maintained at physiological concentrations.

Dynamic and static culture methods
The specific operation under dynamic culture was as follows. At first, the air tightness of the system was checked, and the whole reactor was immersed in water. The mixed gas aerator was opened to fill the reactor cavity with gas. If no bubble emerged, the air tightness of the system was good. The connection seals of each reaction bottle were detected to ensure that no bubble emerged. An aliquot of frozen stock culture of A. muciniphila (100 μL) was inoculated in 5 mL medium and incubated at 37°C for 24 h under strict anaerobic conditions (hydrogen, 5%; carbon dioxide, 10%; nitrogen, 85%). The configured medium was added into the ABIR through the sample port, placed in a sterilization pot maintained at 121°C for 30 min, and cooled to 37°C. The extraction valve was opened, and the air in the reactor was extracted. The extraction valve was closed, and the inflation valve was opened. Anaerobic gas (hydrogen, 5%; carbon dioxide, 10%; nitrogen 85%) was poured into the reactor, and the process was repeated thrice for the reactor to have an anaerobic environment. The seed fluid flow of the incubated strain was added to the reactor cavity through a peristaltic pump. The reactor peristaltic system and the dynamic cultivation of the strain were started. In addition, the OD 600 module (Fig. 6m) can be added to this reactor to monitor the OD 600 value of the fermentation broth in real time. For static culture, the same cultured seed liquid (5 mL) was transferred to an anaerobic bottle with BPM medium volume of 300 mL. The culture was carried out by static incubation for 48 h under the same conditions.

Biomass determination
The fermentation broth (10 mL) was collected and centrifuged, washed twice with deionized water, moved to the centrifuge tube after weighed, and centrifuged to obtain the cells. The cells were dried at 105°C and weighed [55].

Determination of the short-chain fatty acids of metabolites
The fermentation broth (1 mL) was collected, added with 10 μL 2-methylbutyric acid (1 M) as an internal standard, and slowly added with 250 μL concentrated hydrochloric acid. The mixture was mixed well, added with 1 mL diethyl ether, vortexed for 1 min, and allowed to stand until the organic and the water phases separated. The supernatant (organic phase) was carefully collected, added with anhydrous sodium sulfate, vortexed, and passed through a 0.22 μm organic filter. The supernatant (5 μL) was injected into the Agilent 7890 gas chromatograph equipped with an electron capture detector (Agilent, USA), which was used to determine the short-chain fatty acids of A. muciniphila.

Bacterial outer membrane protein extraction method and concentration determination
A. muciniphila was subjected to outer membrane protein extraction using the bacterial outer membrane protein extraction kit (BIOLABO HR0095, Beijing), and the protein content was estimated using the enhanced BCA protein assay kit (Beyotime Biotechnology, Beijing).

FSEM and FTEM
The cultured strains were washed with PBS buffer; fixed with 2.5% glutaraldehyde solution at 4°C for 2 h; washed again with PBS buffer; dehydrated using 30, 50, 70, 80, 90, and 100% alcohol; dried; sprayed with gold; and lyophilized using the SU8200 (Japan) equipment. FSEM analysis was performed, and the morphology of A. muciniphila was observed at 3.0 KV × 10.0 k. The cell suspension was dropped on the copper grid and then dried at room temperature. TEM observations were taken by a transmission electronic microscope (JEM-I010, Hitachi, Tokyo, Japan) in 120 kV.
A. muciniphila outer membrane relative thickness and diameter measurement The TEM and SEM images were imported into the Adobe Photoshop CC 20.0.4, and the ruler tool was used to measure the relative thickness and diameter of the adventitia. Each cell was measured four times at different locations and the averaged values were presented.

Statistical analysis
Data were expressed as mean ± SEM. Differences between the two groups were assessed using the unpaired twotailed Student t test. The data sets that involved more than two groups were assessed using ANOVA. In the figures, data with * were significantly different at p < 0.05 in accordance with posthoc ANOVA. Data were analyzed using the GraphPad Prism version 8.00 for Windows (GraphPad Software).