Article Text
PDF
Abstract
Objective Intestinal flora and metabolites are associated with multiple systemic diseases. Current approaches for acquiring information regarding microbiota/metabolites have limitations. We aimed to develop a precise magnetically controlled sampling capsule endoscope (MSCE) for the convenient, non-invasive and accurate acquisition of digestive bioinformation for disease diagnosis and evaluation.
Design The MSCE and surgery were both used for sampling both jejunal and ileal GI content in the control and antibiotic-induced diarrhoea groups. The GI content was then used for microbiome profiling and metabolomics profiling.
Results Compared with surgery, our data showed that the MSCE precisely acquired data regarding the intestinal flora and metabolites, which was effectively differentiated in different intestinal regions and disease models. Using MSCE, we detected a dramatic decrease in the abundance of Bacteroidetes, Patescibacteria and Actinobacteria and hippuric acid levels, as well as an increase in the abundance of Escherichia–Shigella and the 2-pyrrolidinone levels were detected in the antibiotic-induced diarrhoea model by MSCE. MSCE-mediated sampling revealed specific gut microbiota/metabolites including Enterococcus, Lachnospiraceae, acetyl-L-carnitine and succinic acid, which are related to metabolic diseases, cancers and nervous system disorders. Additionally, the MSCE exhibited good sealing characteristics with no contamination after sampling.
Conclusions We present a newly developed MSCE that can non-invasively and accurately acquire intestinal bioinformation via direct visualization under magnetic control, which may further aid in disease prevention, diagnosis, prognosis and treatment.
- gastrointesinal endoscopy
- intestinal bacteria
- inflammatory bowel disease
Data availability statement
Data are available upon reasonable request.
Statistics from Altmetric.com
Significance of this study
What is already known about this subject?
Intestinal flora and metabolites are associated with the prevention, diagnosis, prognosis and treatment of multiple important diseases.
Capsule endoscopy has been used to visualise the entire GI tract; however, it lacks the capability of acquiring intestinal bioinformation.
Few studies have combined the use of an endoscope and gut microbiota or metabolite sampling capsule.
What are the new findings?
We originally developed a magnetically controlled sampling capsule endoscope (MSCE) for the non-invasive and accurate acquisition and demonstration of findings regarding the intestinal content through direct visualisation.
The newly developed MSCE was capable of acquiring and presenting GI content through direct visualisation with a high success rate.
The MSCE was capable of accurately reflecting gut microbiota and metabolites regardless of the intestinal regions or disease models.
The MSCE was able to accurately reflect certain disease-related gut microbiota and metabolites.
The MSCE exhibited efficient sealing characteristics with no contamination after sampling.
How might it impact on clinical practice in the foreseeable future?
The MSCE could be used to non-invasively sample and present the GI content through direct visualisation with a high success rate in disease models, and with high accuracy, as evidenced by the results of intestinal flora and metabolomics analyses.
The MSCE proposed in our study may open a new chapter for the prevention, diagnosis and individualised treatment of diseases.
Introduction
The intestinal flora and metabolites of the digestive tract can furnish biological information, which is associated with an individual’s health status. For instance, specific microbial strains alterations of Ruminococcus gnavus and adherent invasive Escherichia coli are associated with IBD.1 Interestingly, faecal samples obtained from patients with different stages of colorectal neoplasia provide an evidence of distinct stage-specific phenotypes of faecal microorganisms.2 Fusobacterium nucleatum orchestrates a molecular network to control colorectal cancer chemoresistance.3 Moreover, gut microbiota that start in the gut have far-reaching effects beyond the gut. After nearly a decade of research, experts from 25 different institutions jointly demonstrated the existence of enterotypes in the human gut microbiome.4 Gut microbiota dysbiosis has been shown to influence the course of not only digestive, but also cardiovascular and metabolic diseases, as well as cancers.5–8 Therefore, intestinal flora analysis is considered useful in the screening, diagnosis, severity evaluation, treatment choice selection and prognosis assessment of numerous human disorders.
Meanwhile, metabolites are potential biomarkers for various diseases, for example, oleic acid, a fatty acid, reduces intestinal inflammation and decreases the likelihood of colorectal cancer development.9 Additionally, linoleic acid accumulation in patients with non-alcoholic fatty liver disease (NAFLD) results in the progression of NAFLD-related hepatocellular carcinoma.10 Furthermore, a clinical study demonstrated that the profiles of five metabolites were strongly correlated with clinical markers of kidney disease.11 Remarkably, metabolites could also be used to predict treatment response, that four metabolites were increased in resistant hypertension.12
Currently, there are three main approaches, namely breath hydrogen testing,13 faecal microbial examination14 and invasive endoscopy for digestive sample acquisition.15 However, the principal disadvantages of these approaches include a low specificity16; a non-detailed representation of microbiota by breath testing, an inability to reveal the bioinformation of specific intestinal regions in their real state by faecal sample examination14; and the invasiveness, inconvenience and laborious GI preparation requirement of endoscopy.15 In particular, the small bowel, termed as the ‘black box’ with a total length of 4–6 m and tortuous anatomy, cannot be easily examined by gastroscopy and colonoscopy.17 The evolution of capsule endoscopy enables the visualisation of the entire GI tract with minimal discomfort18; however, the ability to provide sampling systems attached to the capsule remains a challenge.19
In this study, we developed a new type of magnetically controlled sampling capsule endoscope (MSCE) from NaviCam endoscopy series. Apart from direct imaging, it can acquire intestinal bioinformation in a non-invasive, contactless and an accurate manner through direct visualisation under magnetic control. MSCE has a high success rate in different diseases, and may play an important role in disease prevention, diagnosis and treatment.
Methods
Characteristics of MSCE
The overall design of the MSCE is illustrated in figure 1A. The MSCE (32×11.6 mm) is composed of five main parts: an imaging module consisting of a camera and four light-emitting diodes for visualisation, batteries for powering electrical components, a microcontroller unit for data acquisition, an instruction processing and sampling module for intestinal content collection, and an absorbent resin ball for closing sampling port. The sampling module consists of three sampling ports (0.5 mm) on the side wall of the MSCE, an anticlogging ring (not shown in figure 1A) inside the MSCE, a microvalve serving as the switch of the sampler, a permanent magnet for locomotion and posture adjustment, a bottom sampling chamber and a pressure monitoring module to monitor the pressure and temperature of the sampling chamber. The detailed information of the MSCE is shown in online supplemental figure 1.
Supplemental material
The magnetically controlled sampling capsule endoscope (MSCE) is capable for sampling intestinal content. (A) MSCE model. (B) The process of MSCE sampling by two operators, one operator (left) used endoscope to track the capsule endoscope movement, while another operator (right) controlled the movement and sampling process of the capsule by magnetic control according to the capsule endoscope visual field, and not endoscopic field. (C) The endoscope visual field tracking the capsule endoscope movement (up), and the corresponding MSCE visual field taken by its forward camera (down). (D) The layout of treatment and sampling. Surgery group served as the gold standard group for the MSCE. LEDs, light-emitting diodes.
After inserting the MSCE via laparotomy and enterotomy, the camera at the top of MSCE acquires digestive tract images at a frame rate of up to 6 frames/s. The images are wirelessly transmitted to the receiving device outside. MSCE can reach the target for sampling via rolling (sample port orientation adjustment), horizontal rotation, movement along the length of the intestine and height adjustment by magnetic control (online supplemental videos 1–4). Once the capsule reaches the target region, the operator can use the accelerometer to calculate the orientation of the sample port (capsule altitude, pitch angle and roll angle). The sampling port is fixed in the 6 o’clock direction of the lens; therefore, the operator can adjust the posture of the capsule to ensure that the liquid plane is aligned with the sampling port direction.
Supplementary video
Supplementary video
Supplementary video
Supplementary video
Active MSCE suction can occur through the pressure difference inside and outside the sampling chamber. Prior to intestinal content collection, the sampling chamber (0.4 mL) was in a vacuum. The sampling tube inside the capsule was clamped by a valve made of a low-melting-point metal. After receiving the external sampling command, the internal heating element heated the metal to partially melt it. Thereafter, the valve is opened by the tube’s resilience, and the liquid flows into the sampling chamber under the action of external pressure.
Additionally, phenolphthalein and sodium hydroxide solutions were used to test the sealing characteristics of the MSCE (online supplemental figure 2A). The detailed procedures were illustrated by figure legend. We agitated the solution to mimic the behaviour of the capsule motion and left it still for 24 hours. Due to the long duration of entire process, online supplemental video 2 shows only the flow of this experiment. Further, we put the MSCE in a plastic bottle filled with black ink. After shaking the bottle for 7 hours, the liquid in the MSCE remained transparent (online supplemental figure 2B and online supplemental video 5).
Supplementary video
Supplementary video
Animal experiments
Eight juvenile male Guangxi Bama mini-pigs (5–6 months old, 25–30 kg; Department of Animal Science and Technology, Guangxi University) were housed in individually ventilated cages at the SPF facility of Huazhong University of Science and Technology under controlled environmental conditions (temperature 22°C±2°C; relative humidity 60%–70%). They were randomly allocated into two groups (n=4 for each group) by a coworker blinded to the experimental protocol. The experimental group received antibiotic treatment (6 mg/kg/day of sulfadimidine, 6 mg/kg/day of amoxicillin and 10 mg/kg/day of doxycycline) by oral gavage for 10 days to establish the antibiotic-induced diarrhoea model, whereas the control group received an equal volume of saline. The layout of treatment and sampling layout is shown in figure 1D. A total of 32 samples were tested for the microbiome profiling metabolomics profiling.
Intestinal content sampling procedure
To verify whether the capsule could accurately acquire the intestinal content for the determination of intestinal flora and metabolomics in the small intestines, we sampled jejunal and ileal contents, respectively. First, in order to avoid contamination, the enterotomy position was located upstream, approximately 20 cm from the target site after laparotomy operation performed, 80–100 cm from the distal end of the pylorus in the jejunum and 20–30 cm from the proximal cecum in the ileum. After enterotomy, the aseptic MSCE was placed in the small intestine using long aseptic surgical forceps. To simulate the real intestinal environment to the fullest possible extent, we clamped the intestinal and the abdominal incisions using forceps. The capsule was then flipped to the target site under external magnetic control. After the MSCE reached to the target position with the most abundant GI content, the surrounding environment around it could be observed to confirm the location of fluid accumulation. The external magnetic control device adjusted the position of the sampling port on the MSCE side wall to improve the acquisition success rate, and the MSCE internal sensor (acceleration sensor) provided real-time feedback on capsule posture information. After the external sampling command was provided, the valve was gradually opened, and the capsule commenced sampling through the side wall sampling port. Following sampling process completion, a sterilised endoscope was inserted to reach the MSCE position and suctioned the intestinal fluid through the aseptic sprinkler tube. This group was called the surgery group and considered as the control of the MSCE group (online supplemental figure 3).
To evaluate the average success rate of MSCE sampling and the learning curve among different operators, we recruited 10 operators to participate in the experiment. Initially, the 10 operators were trained using less than 20 samples and had a sampling success rate of 70%. After sampling training for less than and more than 50 times, the average sampling success rate of the participants increased to 84% (p<0·05) and 96% (p<0·05), respectively (online supplemental figure 4A). According to the sampling results, we summarised the probability of failure in online supplemental table 1, and determined the corresponding solutions in online supplemental figure 4B, thereby significantly improving the sampling success rate.
Supplemental material
At the time of collection, the well-trained (after 50 times) operators were selected to ensure the high acquisition success rate. The collection volume could reach up to at least 0.35 mL.
Microbiome profiling
The detailed information is shown in the online supplemental materials.
Supplemental material
Metabolomics profiling
The detailed information is shown in the online supplemental materials.
Statistical analysis
The detailed information is shown in the online supplemental materials.
Results
MSCE can acquire intestinal contents through direct visualisation with a high success rate
The overall design of the MSCE is shown in figure 1A. The feasibility of sampling intestinal content using the MSCE was tested with an in vitro model. As shown in figure 1B, the capsule endoscope operator independently controlled the direction of capsule movement and rotation according to the visual field of capsule endoscope, while the endoscope operator acted as a third-party observer tracking the movement of capsule endoscope. First, we demonstrated the flexible movement of the MSCE controlled by an external magnetic field in the intestinal, as shown in the video. Second, when the capsule reached specified site, it could rotate to place its sampling port on intestinal content. Subsequently, the capsule was used to sample through the sampling port after the external sampling command was provided. The video shows the entire sampling process recorded by MSCE outside and inside views (online supplemental video 1), and the corresponding images are displayed in figure 1C.
Supplementary video
MSCE and surgery can sample comparable gut microbiota and metabolites in different intestinal regions
Our experimental layout is shown in figure 1D. We first evaluated the jejunal and ileal content microbiota of the control group. The sequencing results indicated that α-diversity indices of the microbiota were comparable between MSCE and surgery (figure 2A–B) groups. Additionally, principal coordinate analysis (PCoA) based on unweighted UniFrac distances indicated that the content samples were mainly clustered by different intestinal regions; however, these findings were similar between MSCE and surgery (figure 2C) groups. Furthermore, the results showed that ileal content had relatively decreased abundance of Proteobacteria and increased abundance of Firmicutes compared with the jejunal content, and MSCE and surgery reflected similar results (figure 2D).
The magnetically controlled sampling capsule endoscope (MSCE) and surgery can sample comparable gut microbiota and metabolite profiles of different intestinal regions in pigs. (A) Chao1 index of the indicated groups (con-J-surg, con-J-MSCE, con-I-surg, con-I-MSCE). (B) Shannon index of the indicated groups. (C) Principal coordinate analysis of faecal microbiota of indicated groups based on Bray-Curtis distance. (D) Average phylum, family and genus distribution of gut microbiomes of the indicated groups. n.s., no significance. con-J-surg, collecting jejunum content after surgery in control group; con-I-surg, collecting ileum content after surgery in control group; con-J-MSCE, collecting jejunum content by MSCE in control group; con-I-MSCE, collecting ileum content by MSCE in control group. (E) Principal component analysis (PCA) scores plots are color-coded based on the different groups. (F) Volcano plot comparing metabolite abundances between indicated groups (jejunum group and ileum group). (G) Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analysis of metabolites between indicated groups. The ordinate represents the distinct KEGG pathways, and the abscissa represents rich factor (rich factor=amount of differentially expressed metabolites in the pathway/amount of all metabolites in background metabolite set). The colours of the dots represent the p values of enrichment. Red and blue colours indicate high and low enrichments, respectively. Pathway terms were sorted by p value in ascending order, and were marked in bold and underlined when p<0.05. The sizes of the dots represent the gene number of enrichments.
We then investigated the metabolite profiles. The dominant source of variation in the data set was attributed to different regions but not different sampling methods, as demonstrated in the principal component analysis (PCA) score plot of nuclear magnetic resonance content spectral data (figure 2E).
To determine the differential metabolites between surgery and MSCE groups, differences in metabolite screening were further illustrated in the form of volcano plots (figure 2F). In the jejunum, 5 metabolites from samples collected during surgery were downregulated and 9 were upregulated than those collected by MSCE, whereas 415 metabolites showed no significant difference. In the ileum, results were much more similar (6 downregulated, 6 upregulated and 342 no difference). Regarding differential metabolites between the jejunum and ileum, that the surgery and MSCE groups reflected similar number of down/insignificant/up metabolites (249 vs 250, 167 vs 170, 5 vs 7; figure 2F). The differential metabolites of the jejunum and ileum sampled by the two different methods clustered to similar Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways (figure 2G).
Venn diagrams indicated that overlapping regions included common metabolites. Results showed that the surgery and MSCE groups had large overlapping regions (online supplemental figure 5A). Then, data showed that sampling method had little influence on the numbers of different metabolites (online supplemental figure 5B). The aforementioned results above indicated that these two different sampling methods did not influence the metabolite profiles.
MSCE and surgery can sample comparable gut microbiota in different disease models
We evaluated the alterations of microbiota diversity and composition in both groups, and found that antibiotic usage significantly decreased microbiota diversity. Compared with samples collected by surgery, those collected by MSCE had similar α-diversity indices of the same state (figure 3A–B). The β-diversity Bray Curtis distance PCoA analysis indicated that region and model differences, instead of method differences, were responsible for the variability in the microbial communities (figure 3C). Similar results are shown in online supplemental table 2.
The magnetically controlled sampling capsule endoscope (MSCE) and surgery can sample comparable gut microbiota of different disease models in pigs. (A) Chao1 index of the indicated groups (con group and antibiotics-induced diarrhoea group). (B) Shannon index of the indicated groups. (C) Principal coordinate analysis of faecal microbiota of indicated groups based on Bray-Curtis distance. (D) Average phylum of gut microbiomes of the indicated groups. *P<0.05. **p<0.01. n.s., no significance. anti-J-surg, collecting jejunum content after surgery in antibiotics-induced diarrhoea group; anti-I-surg, collecting ileum content after surgery in antibiotics-induced diarrhoea group; anti-J-MSCE, collecting jejunum content by MSCE in antibiotics-induced diarrhoea group; anti-I-MSCE, collecting ileum content by MSCE in antibiotics-induced diarrhoea group.
Furthermore, we measured gut microbiota composition in disease models. Antibiotic treatment included a decrease in the abundance of Bacteroidetes phyla, along with abundance of Patescibacteria and Actinobacteria phyla and a significantly increase in Escherichia–Shigella genera abundances especially in ileum. However, samples collected by MSCE showed comparable microbiota composition compared with those obtained by surgery (figure 3D, online supplemental figure 6A). Further, regardless of the region or model, no significant differences existed between surgery and MSCE (online supplemental table 3).
Supplemental material
Moreover, the abundance of the top 30 genera is displayed as a heatmap. Pairwise comparison of bacterial taxa abundance revealed slight differences in different sampling methods (online supplemental figure 6B). These results indicated that the different sampling methods did not affect the diversity or composition of gut microbiota.
MSCE and surgery can sample comparable metabolites in different disease models
To determine the technical reproducibility of the two methods, we evaluated the metabolite profiles in both groups. We found a clear clustering of samples according to the different models, whereas the surgery and MSCE groups clustered together indicating that variability was largely explained by the different models, instead of methods (figure 4A).
The magnetically controlled sampling capsule endoscope (MSCE) and surgery can sample comparable metabolite profiles of different disease models in pigs. (A) The principal component analysis (PCA) scores plots are color coded based on the different groups. (B) Volcano plot comparing metabolites between indicated groups (control group and antibiotics-induced diarrhoea group). (C) Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analysis of metabolites between indicated groups. The ordinate represents the distinct KEGG pathways, and the abscissa represents rich factor (rich factor=amount of differentially expressed metabolites in the pathway/amount of all metabolites in background metabolite set). The dot colours represent the p values of enrichment. Red and blue colours indicate high and low enrichments, respectively. Pathway terms were sorted by p value in ascending order, and were marked in bold and underlined when p<0.05. The sizes of the dots represent the gene number of enrichments.
Similarly, variable importance in projection (VIP) scores were used to determine the differential metabolites between the surgery and MSCE groups. We found that the surgery group had similar results with the MSCE group (figure 4B). Moreover, the two sampling methods had little effect on the differential metabolites evaluation of the different models in the different regions (figure 4B). Furthermore, KEGG pathway enrichment analysis was used to estimate the differential metabolites between different models. Results showed that the differential metabolites of the different models sampled by the two different methods clustered to similar KEGG pathways (figure 4C).
Additionally, results from Venn diagram analyses of differentially expressed metabolites demonstrated a large overlapping region in both the surgery group and MSCE groups (online supplemental figure 7A). Moreover, regarding the numbers of different metabolites, both groups showed little difference (online supplemental figure 7B).
The abundance of all metabolites is displayed as a heatmap. The pairwise comparison of metabolites abundance revealed little differences between the two sampling methods (online supplemental figure 7C). The results indicated that the two sampling methods did not influence the metabolite profiles.
MSCE and surgery can sample comparable disease-related gut microbiota and metabolites
We analysed specific disease-related bacterial genera in intestinal content in order to evaluate MSCE sampling accuracy. We found that Escherichia–Shigella genera abundance increased after antibiotic treatment, especially in the ileum. Further, fewer short-chain fatty acids (SCFA)-producing genera including Enterococcus, Lachnospiraceae and Ruminococcaceae, and anti-inflammatory genera such as Lactobacillus and Ruminococcus_gauvreauii were decreased in the antibiotics-induced diarrhoea group. Nevertheless, the above-mentioned genera abundances were comparable between the MSCE and surgery sampling groups (figure 5A).
The magnetically controlled sampling capsule endoscope (MSCE) and surgery can sample comparable disease related gut microbiota and metabolites in pigs. (A) The amount of disease-related gut microbiota in indicated groups. (B) The amount of disease-related metabolites in indicated groups. *P<0.05. **P<0.01. n.s., no significance.
Furthermore, specific disease-related metabolites in intestinal content were analysed to estimate the accuracy of MSCE sampling. Antibiotic treatment increased proinflammatory metabolite abundance; 2-pyrrolidinone increased, as well as the abundance of other colitis-related metabolites including 5-hydroxyindole-3-acetic acid and succinic acid. In contrast, the abundances of anti-inflammatory metabolites including acetyl-L-carnitine, glycerophosphatidylcholine, Sn-glycero-3-phosphocholine, hippuric acid and 6-methylnicotinamide decreased after antibiotic treatment (figure 5B). Nonetheless, the above-mentioned metabolite abundances were not influenced by the sampling methods (figure 5B). Our results showed that MSCE could accurately reveal the disease-related intestinal flora and metabolite abundance, thereby playing an important role in disease diagnosis.
Discussion
The biological information of the intestinal track is reported to reflect various important diseases.5 7 20 In this study, we validated the novel method of sampling intestinal content with our newly developed MSCE. Several highlights should be emphasised. First, we originally developed the MSCE, which could successfully sample specific intestinal content using direct visualisation. Second, the MSCE can sample biological information in different intestinal regions or different diseases. Third, intestinal content obtained by the MSCE could truly reflect the specific disease-related gut microbiota and metabolite abundance.
A recent study showed that an ingestible, biocompatible, 3D-printed microengineered pill can sample using an integrated osmotic sampler and microfluidic channels.21 Another study has demonstrated a passive 3D-printed gut sampling capsule that uses the swelling property of highly absorbent hydrogel for the sampling.22 However, they cannot observe the mucous membrane through direct visualisation. They ‘passively sample’ using osmotic pressure or material swelling materials instead of ‘actively suctioning’; moreover, they cannot collect intestinal flora in specific areas. To our largest knowledge, no research has combined the use of an endoscope and a capsule in gut microbiome sampling. Therefore, we developed the MSCE to bridge this gap. We consider that our MSCE is completely different from the above-mentioned devices in principle and design. MSCE has several advantages: (1) it can observe the GI mucous membrane through direct visualisation to determine the best sampling site, thereby facilitating disease diagnosis; (2) it can rotate clockwise and counterclockwise, and flip along the long axis through magnetic control to the target site; (3) it can perform actively suctioning via an external sampling command, and this feature can help to collect intestinal flora in specific areas and avoid the contamination of the entire digestive tract. Recent studies reported that the main challenge in capsule endoscopy is the ability to provide sampling systems attached to itself.19 23 24 However, this newly developed MSCE circumvents this challenge.
Nevertheless, regarding the success rate of digestive content sampling for experienced MSCE operators was 96%. Although there were certain requirements for operators, those who had performed the procedure more than 50 times had a significantly higher success rate. Several factors leading to operation failure were listed in online supplemental table 1. Efforts were then made to improve the procedure success rate with regard to two aspects. Regarding MSCE, anticlogging ring and multiple sampling ports distributing into an arc along the surface have been devised to prevent sampling ports from clogging. Further, the sampling delay time may be reduced from approximately 40 s to less than 15 s after upgrading the capsule system. Regarding the collection procedures, the operator needs to check the liquid surface direction from MSCE videos and adjust the capsule altitude to ensure sampling ports immersion in the liquid using external magnetic control. The operator can use the MSCE internal sensor (acceleration sensor) to get real-time information of the MSCE posture and calculate the orientation of the sampling port. The sampling port is fixed in the 6 o’clock direction of the lens; therefore, the operator can adjust the posture of the capsule to ensure that the liquid plane is aligned with the direction of the sampling port. Details are shown in online supplemental table 1. In order to mitigate the cost of future training, the operators can undergo a series of training sessions to rotate the capsule very skillfully under magnetic control and to adjust the altitude of the capsule before sampling. After operators can freely control the capsule to an appropriate sampling altitude, they can trigger the sampling switch to verify if the acquisition was successful. This change can mitigate the training cost to the largest extent.
Moreover, it is necessary to ensure that the collected sample uncontaminated. Although the online supplemental video 2 lasted only for 5 min, no liquid leakage was found in our 24 hours test, indicating that the sampling port in the capsule was properly sealed after sampling. Effective sealing could provide a basis for the authenticity of the results in the process of widespread clinical application.
With regard to sample storage, when the MSCE is put into clinical application, fixing the bacteria and metabolites will be a good way to prevent them from evolving inside the capsule as it takes time for the MSCE to be excreted. As reported in previous studies, microbiota and metabolomic measurements appear reproducible and stable in samples collected with 95% ethanol.25–29 In the clinical setting, MSCE will be excreted by intestinal peristalsis, and the fixing method using 95% ethanol can reflect the status of intestinal flora and metabolites to the largest extent.
In our experiment, we demonstrated that the MSCE could collect sufficient intestinal content for microbiota detection. Our results showed that after antibiotic treatment, diversity indices significantly decreased, which was similar to findings of Lankelma et al.30 After antibiotic administration, the abundances of the phyla Bacteroidetes, Patescibacteria and Actinobacteria decreased, consistent with the findings of the study by Looft.31 In addition, the abundance of the Escherichia–Shigella genera in the jejunum and ileum was significantly increased after early antibiotic administration, which corroborated with the findings of Zhang et al.32 The abundances of all the observed phyla showed no significant differences between different sampling methods, thereby demonstrating the accuracy of MSCE sampling. In this study, the MSCE was also demonstrated to sample enough digestive content for metabolite profile detection. Additionally, PCA score plot, VIP and KEGG pathway enrichment analyses showed that sampling methods did not influence the metabolite profiles. We believe that the digestive contents collected by the MSCE can accurately reveal gut microbiota and metabolites, thus providing reliable evidence for disease evaluation.
Previous studies have shown that some specific intestinal flora are strongly related to different diseases.5–8 In our study, the abundances of SCFA-producing genera including Enterococcus, Lachnospiraceae and Ruminococcaceae were decreased after antibiotic treatment, which is similar to the findings of a recent study that antibiotic treatment depleted the abundance of Lachnospiraceae in the jejunum.33 Importantly, these bacteria have been proven to prevent the IBD and type 2 diabetes mellitus.34 Enterococcus strains have probiotic attributes and are frequently reduced in patients with diabetes, obesity and cancers.35 Additionally, our results indicated that the abundances of Lactobacillus and Ruminococcus-gauvreauii with antibacterial, anti-inflammatory and antimicrobial activities36 decreased after antibiotic treatment, which was consistent with the findings of Pi et al, who showed that antibiotics markedly decreased the populations of Bifidobacterium, Lactobacillus and Ruminococcus in the ileum.37 Our results showed that the MSCE could accurately reveal the disease-related intestinal flora, indicating its possible role for disease diagnosis.
Furthermore, our results showed that antibiotic treatment increased and decreased, respectively, the levels of colitis-related and anti-inflammatory metabolites, leading to a greater risk of colitis, which was in agreement with previous findings.38 As for disease-related metabolites, 2-pyrrolidinone, 5-hydroxyindole-3-acetic acid and succinic acid have been reported to be positively related with colitis. Acetic acid-induced colitis was previously used as a perfect model for colitis investigation.39 Additionally, other endogenous metabolites can govern proinflammatory responses, as recently reported for succinate.40 Moreover, some other metabolites including acetyl-L-carnitine, glycerophosphatidylcholine, Sn-glycero-3-phosphocholine and hippuric acid have been reported to play anti-inflammatory roles.41 The MSCE was demonstrated to precisely disclose the amount of disease-related metabolite levels, thus denoting its potential diagnostic value.
In particular, in addition to its application in the above-mentioned important systemic diseases, the MSCE can be used in the current COVID-19 pandemic, as a non-contact endoscope that can provide an evidence of digestive tract viral infection and investigate the digestive tract environment without the risk of aerosol transmission and endoscope decontamination.
Conclusions
In summary, we present a newly developed MSCE that can sample intestinal content via direct visualisation in a non-invasive manner using precise magnetic control with a high success rate in disease diagnosis; moreover, it can verify the result accuracy of intestinal flora and metabolomics. Gut microbiota can well reflect the state of the digestive tract, and MSCE that observes the gut has a far-reaching view beyond gut diseases. We believe that the MSCE proposed in our study may open a new chapter and change the scheme of capsule endoscopy in the prevention, diagnosis and individualised treatment of multiple diseases.
Supplemental material
Supplemental material
Supplemental material
Supplemental material
Supplemental material
Data availability statement
Data are available upon reasonable request.
Ethics statements
Patient consent for publication
Ethics approval
All animal studies were approved by the Animal Experimentation Ethics Committee of Huazhong University of Science and Technology.
Acknowledgments
The authors wish to thank all study participants, researchers, clinicians, technicians and administrative staff who contributed to this study.
References
Supplementary materials
Supplementary Data
This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
Supplementary Data
This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
Footnotes
ZD, WW and KZ contributed equally.
Correction notice This article has been corrected since it published Online First. The provenance and peer review statement has been included.
Contributors RL, XH, ZD, WJW and KZ designed the study. WW, KZ, FM, TY, TX and HY did the experiment and collected the samples. FM, TY, YB and HP contributed to operate the machine. HS, CH, WJ and JL interpreted the data and analyses. WW and KZ wrote the first draft of the manuscript, and all authors reviewed, contributed to, and approved the manuscript.
Funding This study was supported by the National Natural Science Foundation of China (Nos. 81770539, 81330014, 81572428, 81272656, 81974068 and 81900580), and the National Key Research and Development Program of China (No. 2017YFC0110003).
Competing interests None declared.
Provenance and peer review Not commissioned; externally peer reviewed.
Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.