Correlation between intestinal flora and GLP-1 receptor agonist dulaglutide in type 2 diabetes mellitus treatment—A preliminary longitudinal study

Summary

GLP-1 receptor agonists (GLP-1 RA) are presently used as the first-line drugs for the clinical treatment of type 2 diabetes mellitus (T2DM). It can regulate blood glucose by stimulating insulin secretion and lowering glucagon levels. We used 16S rRNA amplicon sequencing to detect structural changes in the composition of the intestinal flora of newly diagnosed T2DM after 1 and 48 weeks of dulaglutide administration. Our research found no significant changes in the intestinal flora after the administration of dulaglutide for 1 week to subjects with newly diagnosed T2DM. Nevertheless, after 48 weeks of dulaglutide administration, the composition of the intestinal flora changed significantly, with a significant reduction in the abundance of intestinal flora. Furthermore, we found that fasting glucose levels, fasting c-peptide levels, HbA1c levels, and BMI are also closely associated with intestinal flora. This reveals that intestinal flora may be one of the mechanisms by which dulaglutide treats T2DM.

Graphical abstract

Highlights

• •Long-term use of dulaglutide can alter the intestinal flora in patients with T2DM
• •Long-term use of dulaglutide may alter metabolic pathways in people with T2DM
• •Indicators of T2DM were correlated with the abundance of certain intestinal flora

Health sciences; Microbiome; Association analysis

Introduction

The intestinal flora is considered a complex organ of the human body. Research suggests that it contains approximately 500–1,000 species and 10¹⁴ microorganisms, which is 10 times more than the total number of cells in the human body. The number of genes present in the intestinal flora is 100 times more than that of the presently known human genome. It is also known as the “second set of genetic code” of the human body.¹ Long-term studies have reported that the imbalance of the normal intestinal flora can cause acute and chronic diseases in different systems of the human body.^2,3,4 This imbalance is considered an intermediate factor in diseases associated with various systems of the human body. Therefore, it is crucial to maintain a good symbiotic relationship between the human body and the intestinal flora.⁵ Human intestinal flora can affect the abundance of good bacteria (such as those producing short-chain fatty acids [SCFAs]) and regulate the concentration of metabolites and SCFAs in the body, thus affecting glucose metabolism.⁶ The intestinal flora regulates the levels of human bile acid (BA) and endotoxin. This affects the level of intestinal hormone release by influencing intestinal permeability and inflammatory responses, ultimately affecting insulin resistance.⁷

The human gut flora is involved in the development of type 2 diabetes mellitus (T2DM). Additionally, the role of gut microbes has garnered increasing attention in the treatment of T2DM.⁸ GLP-1 receptor agonist (GLP-1 RA) is a new type of antidiabetic drug marketed in the last decade. It has gradually become the first-line drug for the clinical treatment of T2DM due to its strong hypoglycemic effect, low hypoglycemic effect, advantages for weight loss, and cardiovascular and renal benefits.⁹ GLP-1 RAs primarily act on the GLP-1 receptor (GLP-1 R) in human intestinal tissue to regulate blood glucose levels via three main mechanisms. Firstly, they exert a glucose-dependent insulin-stimulating effect, avoiding excessive strain on islet cells. Secondly, GLP-1 RA can inhibit glucagon secretion, glycogenolysis, and gluconeogenesis, except in hypoglycemic conditions. Thirdly, they can slow gastric emptying speed, thereby reducing postprandial blood sugar levels.^10,11 A few in vivo studies have reported the mechanism underlying the action of intestinal flora in the treatment of T2DM with GLP-1 RA.¹² In vivo experiments showed that gut microorganisms can alleviate diet-induced hypothalamic inflammation and improve leptin sensitivity via a GLP-1 R-dependent mechanism, ultimately affecting human metabolic function.¹³ Another study confirmed that GLP-1 RA can affect the composition and structure of the intestinal flora in mice, which has been implicated in the gut-brain axis.¹² Nevertheless, no relevant study has investigated how the gut flora changes in vivo when GLP-1 RA is used to treat T2DM and whether altered gut microbes are associated with blood glucose regulation.

Dulaglutide, a once-weekly GLP-1 RA, has been approved in many countries for treating patients with T2DM.¹⁴ Dulaglutide is gradually replacing exenatide and liraglutide in clinical practice due to its advantageous hypoglycemic properties, weight loss benefits, and long half-life. Nevertheless, the effect of liraglutide and exenatide on the intestinal microbiota has been studied,^15,16 whereas the effect of dulaglutide is unexplored.

To the best of our knowledge, this is the first study to investigate the changes in the gut flora of patients experiencing their first episode of T2DM in a clinical setting. This analysis was performed using the 16S rRNA technique before the administration of dulaglutide, after 1 week of dulaglutide administration, and after 48 weeks of dulaglutide administration. We explored the effects of dulaglutide on the biodiversity of gut flora and identified differentiating gut microorganisms. These findings serve as a theoretical basis for subsequent research to elucidate the mechanisms through which dulaglutide can alleviate T2DM through the modulation of the intestinal flora.

Results

Changes in fasting blood glucose level

The fasting blood glucose (FBG) level of the subjects was 13.19 ± 2.93 mmol/L in group F, 7.41 ± 1.11 mmol/L in group S, and 6.27 ± 0.7 mmol/L in group T. FBG level statistically decreased gradually with the administration of dulaglutide was observed (p < 0.001) (Table 1; Figure 1).

Table 1.

Clinical characteristics of the subjects

	Group F	Group S	Group T	p value
Male gender [n (%)]	26 (63.4)
Age (years)	48.12 ± 3.13
Country (China [9%])	41 (100)
Race (Han [%])	41 (100)
Height (m)	1.66 ± 0.09
Weight (kg)	75.88 ± 16.5	74.22 ± 15.12	69.56 ± 11.64	p < 0.001
BMI Index (kg/m2)	27.25 ± 4.37	26.65 ± 3.88	25.01 ± 2.74	p < 0.001
FBG (mmol/L)	13.19 ± 2.93	7.41 ± 1.11	6.27 ± 0.7	p < 0.001
HbA1c (%)	8.22 ± 1.42	8.1 ± 1.32	6.03 ± 0.61	p < 0.001
Fasting C-Peptide (mmol/L)	2.27 ± 0.9	2.37 ± 0.95	2.74 ± 0.88	p < 0.001

Figure 1.

Flowchart of subject recruitment

Changes in BMI

Subjects had a BMI of 27.25 ± 4.37 kg/m² in group F, 26.65 ± 3.88 kg/m² in group S, and 25.01 ± 2.74 kg/m² in group T. BMI significantly decreased gradually with the administration of dulaglutide (p < 0.001) (Table 1; Figure 1).

Changes in HbA1c concentration

The hemoglobin A1c (HbA1c) concentration of the subjects was 8.22% ± 1.42% in group F, 8.1% ± 1.32% in group S, and 6.03% ± 0.61% in group T. HbA1c concentration statistically decreased gradually with the administration of dulaglutide (p < 0.001) (Table 1; Figure 1).

Changes in fasting C-peptide

The fasting C-peptide level of the subjects was 2.27 ± 0.9 mmol/L in group F, 2.37 ± 0.95 mmol/L in group S, and 2.74 ± 0.88 mmol/L in group T. Fasting C-peptide level statistically increased gradually with the administration of dulaglutide (p < 0.001) (Table 1; Figure 1).

Analysis of amplicon sequence variant clustering results

Through the 16S amplicon sequencing detection, the average raw data of each sample were ultimately measured to be 115,938. After filtering, we obtained 114,924 data, which were further reduced to 111,216 after denoising. Lastly, we obtained 93,293 effective data after splicing and dechimerization, with a high effective rate of 81.44%. Subsequently, we performed the deweighting operation on the obtained valid data to obtain the deweighted sequence amplicon sequence variant (ASV), setting the standardized data depth to 95% of the minimum sample sequence volume. This resulted in a standardized sample sequencing depth of 46,606, and the final average number of ASVs per sample was 13,674 (Figure 2A).

Figure 2.

Analysis of sample size

(A) ASV Venn diagrams for groups T, F, and S; (B) the sparse curves; (C) the Shannon index curve.

Rational analysis of sample size

The sparse curves showed that the curves of all three groups, F, S, and T, gradually flattened out, with group T almost reaching a plateau (Figure 2B). Additionally, we plotted the Shannon index curve for re-verification and observed that the Shannon index curves of all three groups, F, S, and T, reached a plateau (Figure 2C). The sparse curves and the Shannon index curve serve as indicators of sample size in bioinformatics. When these curves reach a plateau, it suggests that the sample size has sufficiently covered all known microbial species. Further increasing the sample size is unlikely to significantly affect the final results.¹⁷ Our study shows that the Shannon index curve has reached a plateau, and the sparse curve has gradually reached a plateau. This implies that our sample size is adequate, and the research results are unlikely to be biased due to sample size considerations.

Alpha diversity analysis of the intestinal flora

Alpha diversity analysis of the intestinal flora before and after 1 week of dulaglutide administration

The fecal samples of the subjects, before and 1 week after using dulaglutide, underwent sequencing. The α-diversity results indicated that the Chao1 index (Figure 3A), Ace index (Figure 3B), Shannon index (Figure 3C), and Simpson’s index (Figure 3D) of the intestinal flora after 1 week of using dulaglutide did not show significant differences compared with those before using dulaglutide. This indicates no significant changes occurred in the abundance and diversity of the intestinal flora in T2DM subjects after 1 week of dulaglutide administration.

Figure 3.

α-diversity in groups F and S

(A) The Chao1 index boxplot; (B) the ACE index boxplot; (C) the Shanon index boxplot; (D) the Simpson index boxplot.

Alpha diversity analysis of intestinal flora after 48 weeks of dulaglutide administration

Subsequently, we performed an analysis of amplicon sequencing on the fecal samples from the subjects after 48 weeks of using dulaglutide and compared them with the pre-use data. The results revealed that, after 48 weeks of dulaglutide administration, the Chao1 index (Figure 4A) and the Ace index (Figure 4B) were significantly lower than those before dulaglutide administration. However, the Shannon index (Figure 4C) and Simpson’s index (Figure 4D) did not exhibit significant changes. This indicates that, after 48 weeks of treating T2DM with dulaglutide, there was a notable decrease in the abundance of gut microorganisms in subjects compared to that before treatment, while the diversity of gut microorganisms remained relatively stable.

Figure 4.

α-diversity in groups F and T

(A) The Chao1 index boxplot; (B) the ACE index boxplot; (C) the Shanon index boxplot; (D) the Simpson index boxplot.

Beta diversity analysis of gut flora

The principal coordinate analysis (PCoA) plots showed a significant gap in distance between group T samples and groups F and S samples, irrespective of the weighted UniFrac algorithm or unweighted UniFrac algorithm (Figures 5A and 5B). This indicates a high differentiation in the gut microbial community of group T compared with that of groups F and S. However, the structure of the gut flora remained quite similar between groups F and S.

Figure 5.

β-diversity analysis plots for groups F, S, and T

(A) The PCOA analysis calculated by unweighted UniFrac algorithm; (B) the PCOA analysis calculated by weighted UniFrac algorithm; (C) the boxplots for the PCOA1 axis of Plot A; (D) the boxplots for the PCOA2 axis of Plot A; (E) the B plot of the boxplot for the PCOA1 axis; (F) the boxplot for the PCOA2 axis of B plot.

Using the Kruskal-Wallis test to analyze the two axes of PCoA calculated by weighted UniFrac and unweighted UniFrac separately, we observed significant differences between group T and both groups F and S in both coordinates of the PCoA plot calculated by unweighted UniFrac (Figures 5C and 5D). In both coordinates of the PCoA plot calculated by weighted UniFrac, only group T was significantly different from group F in the PCoA1 coordinate (Figure 5E), whereas group T remained significantly different from both group F and group S in the PCoA2 coordinate (Figure 5F).

To conclude, the results of the β-diversity analysis revealed that the β-diversity of the intestinal flora after 48 weeks of using dulaglutide was significantly different from that before using dulaglutide. In contrast, the intestinal flora after 1 week of using dulaglutide showed no significant difference from that before using dulaglutide, aligning with our α-diversity results.

Anosim analysis test

Anosim analysis, a nonparametric test also known as similarity analysis, is frequently used to determine whether the differences in samples among different subgroups are significantly greater than the differences within individual groups. This helps determine whether the grouping of samples is reasonable and analytically useful. It is generally used to determine the reasonableness of the results of β-diversity analysis. In β-diversity analysis, there is no clear test conclusion to determine the significance of differences between samples of different subgroups, and it can only be judged artificially by the intuitive graph of PCoA. The R value in the Anosim analysis serves to determine the significance of grouping, falling within the range of (−1, 1). When the R value is >0, it indicates that the difference between groups is significant and the results of β-diversity between subgroups are reasonable and valid. When the R value is <0, it indicates that the difference within the group is greater than the difference between the groups, and the results of β-diversity will be biased. The confidence level of the statistical analysis is expressed by the p value, where a p value <0.05 indicates statistical significance.

Our Anosim analysis yielded results indicating that the R values for PCoA, calculated using weighted UniFrac and unweighted UniFrac algorithms, were >0, with p values <0.05 (Figures 6A and 6B). This implies that the grouping of the three samples—before dulaglutide administration, after 1 week of dulaglutide administration, and 48 weeks of dulaglutide administration—showed statistical significance. Therefore, the results of the β-diversity analysis were considered accurate and reliable.

Figure 6.

Anosim analysis test and ASV analysis of samples

(A) Anosim-analyzed boxplots computed by the unweighted UniFrac algorithm; (B) Anosim-analyzed boxplots computed by the weighted UniFrac algorithm; (C) the columnar distribution of ASV in the phylum, order, order, family, genus and species in a total of 123 fecal samples from 41 subjects in three groups of F, S and T; (D) the columnar distribution of ASV in the phylum, order, order, family, genus and species in three groups of fecal samples from F, S and T.

Distribution at the level of each species

Out of a total of 13,674 valid ASVs, 93.51% could be annotated to the database. The distribution at different taxonomic levels was as follows: 89.43% at the phylum level, 89.23% at the order level, 87.61% at the class level, 86.6% at the family level, 73.86% at the genus level, and 29.49% at the species level (Figures 6C and 6D).

At the phylum level, the dominant bacteria included Bacteroidota, Firmicutes, and Proteobacteria (Figure 7A). At the class level, the dominant bacteria included Bacteroidia, Clostridia, Gammaproteobacteria, and Gammaproteobacteria (Figure 7B). At the order level, the dominant bacteria included Bacteroidales, Lachnospirales, and Oscillospirales (Figure 7C). At the family level, the dominant bacteria included Bacteroidaceae, Lachnospiraceae, and Ruminococcaceae (Figure 7D). Dominant species at the genus level included Bacteroides, Blautia, and Faecalibacterium (Figure 7E). The dominant species were identified as Bacteroides_plebeius and Bacteroides_vulgatus (Figure 7F). Although the three groups did not exhibit significantly different dominant bacteria at all levels, there was some variability in the abundance of dominant bacteria across subgroups.

Figure 7.

Distribution at the level of each species

(A) The distribution of dominant flora in three groups in phylum level; (B) the distribution of dominant flora in three groups in class level; (C) the distribution of dominant flora in three groups in order level; (D) the distribution of dominant flora in three groups in family level; (E) the distribution of dominant flora in three groups in genus level; (F) the distribution of dominant flora in three groups in species level.

Functional prediction result

Our results indicated no significant difference in the three levels of the Kyoto Encyclopedia of Genes and Genomes (KEGG) between group F and group S. This suggests that the metabolic pathway after 1 week of GLP-1 agonist administration did not undergo significant changes compared with before administration (Figures 8A–8C). However, the Anosim analysis diagram of the KEGG pathway in the three grades of group F and group T showed significant differences, suggesting that the KEGG metabolic pathway underwent significant changes after 48 weeks of GLP-1 RA administration (Figures 8D–8F).

Figure 8.

Functional prediction result

(A) Anosim analysis of KEGG primary pathway in groups F and S; (B) Anosim analysis of KEGG secondary pathway in groups F and S; (C) Anosim analysis of KEGG tertiary pathway in groups F and S; (D) Anosim analysis of KEGG primary pathway in groups F and T; (E) Anosim analysis of KEGG secondary pathway in groups F and T; (F) Anosim analysis of KEGG tertiary pathway in groups F and T; (G) the heatmaps of KEGG primary pathway in groups F and T. (H) the heatmaps of KEGG secondary pathway in groups F and T. (I) the heatmaps of KEGG tertiary pathway in groups F and T.

Based on the functional composition and relative abundance of each sample, we performed a functional heatmap analysis. We extracted the top 50 functions with average abundance at each taxonomic level and, based on their abundance information in each sample, performed clustering of functions and samples. Heatmaps were generated to facilitate the identification of functions that clustered more or less in specific samples.

In the heatmap clustering analysis, the color indicates the functional abundance and the color gradient from green to red indicates the relative abundance from low to high. Horizontal clustering indicates the similarity of abundances of different functional classifications between samples. The closer the distance between two species, the shorter the branch length, suggesting that the abundances of these two functional classifications are more similar between samples. Because more metabolic pathways were present with differences between group F and group T, the detailed predicted functional abundance of Phylogenetic Investigation of Communities by Reconstruction of Unobserved States (PICRUSt2) is shown in Figure 9 (Figures 8G–8I).

Figure 9.

Spearman correlation heatmap of FBG, BMI, HbA1c, Fasting C-Peptide, and various bacteria

Variation in abundance of common bacteria

We assessed the variation in the abundance of individual bacteria in each sample by extracting information from the raw data and then performed statistical analysis using the paired Wilcoxon rank-sum test and Friedman rank-sum test (Tables 2 and 3). Our findings indicate that the abundance of both Firmicutes and Bacteroidota exhibited significant changes after dulaglutide administration. The abundance of Firmicutes did not show a significant change after 1 week of dulaglutide administration (p = 0.328) but exhibited a significant decrease after 48 weeks of dulaglutide administration (p < 0.001). On the other hand, the abundance of Bacteroidota did not change significantly after 1 week of dulaglutide administration (p = 0.751) but increased significantly after 48 weeks of dulaglutide administration (p = 0.005).

Table 2.

Abundance of intestinal microorganisms at the level of common phylum and genus

Bacteria	F (%)	S (%)	T (%)	p value
Bacteroidota	18.75 (8.24, 33.69)	17.94 (8.71, 34.74)	31.81 (15.68, 48.49)	0.004
Firmicutes	71.23 (55.5, 80.75)	65.2 (50.79, 82.84)	51.06 (38.92, 61.6)	0.001
Agathobacter	1.79 (0.79, 2.96)	1.14 (0.61, 2.55)	1.98 (0.89, 2.97)	0.215
Akkermansia	0.03 (0.01, 0.14)	0.06 (0.02, 0.37)	0.04 (0.02, 0.15)	0.05
Bacteroides	10.58 (5.43, 24.02)	13.17 (4.81, 26.55)	19.76 (9.03, 35.99)	0.025
Blautia	7.96 (5.24, 16.23)	7.44 (4.11, 10.97)	4.34 (3.08, 9.01)	0.012
Bifidobacterium	0.35 (0.11, 0.78)	0.14 (0.05, 0.25)	0.73 (0.39, 2.15)	0.001
Coprococcus	0.84 (0.52, 1.46)	0.58 (0.35, 1)	0.74 (0.53, 1.03)	0.532
Faecalibacterium	9.31 (3.9, 12.34)	6.73 (4.21, 10.2)	7.75 (4.81, 11.66)	0.599
Lactobacillus	0.04 (0.01, 0.22)	0.3 (0.13, 0.56)	0.2 (0.06, 0.5)	0.001
Parabacteroides	0.86 (0.34, 1.93)	1.09 (0.64, 2.13)	1.48 (0.79, 2.59)	0.249
Prevotella	0.53 (0.13, 1.32)	0.2 (0.11, 0.63)	1.51 (0.55, 3.16)	0.001
Roseburia	3.23 (1.82, 4.86)	2.41 (1.6, 4.44)	2.36 (1.54, 3.75)	0.281
Ruminococcus	0.92 (0.55, 2.09)	0.75 (0.27, 1.32)	0.37 (0.14, 0.71)	0.002

All values are retained to two decimal places.

Table 3.

Properties and mechanisms of genus-level gut microbiome in this study

Bacteria	Classification of bacteria		General mechanism of action	Reference
Agathobacter	Gram-positive anaerobic bacteria	probiotics	The main products are butyric acid and acetic acid. It is a probiotic discovered in recent years.	Rosero et al.18
Akkermansia	Gram-negative anaerobic bacteria	probiotics	1. The producer of SCFAs, such as acetic acid and propionic acid; 2. maintain the integrity of intestinal mucosal barrier; 3. suppress inflammation and regulate immune response; 4. promote the growth of other probiotics such as Faecalibacterium and Roseburia.	Zhao et al.19
Bacteroides	Gram-negative anaerobic bacteria	probiotics, opportunistic pathogen	1. It can produce acetic acid and propionic acid, maintain the stability of the intestinal environment; 2. extensive digestion of dietary fiber polysaccharides and human glycans; 3. it has anti-inflammatory effects and regulates the immune system; 4. when bacteroides are able to escape to other parts of the body beyond the gut, they can act as disease-causing bacteria, causing abscesses and other infections.	Zafar et al.20
Blautia	Gram-positive anaerobic bacteria	probiotics, opportunistic pathogen	1. Blautia ferments glucose to produce acetic acid, succinic acid, lactic acid, and ethanol; 2. it prevents colonization of pathogens by producing bacteriocins and exhibits anti-inflammatory properties and maintains glucose homeostasis by upregulating the production of regulatory T cells and SCFA; 3. some species may damage the intestinal barrier and play a negative role in the gut flora; 4. Blautia is controversial, with most studies showing that it is highly abundant in the feces of people with type 2 diabetes; however, a small number of studies have shown that decreased abundance is inversely associated with type 2 diabetes.	Liu et al.21; de Mooij et al.22
Bifidobacterium	Gram-positive anaerobic bacteria	probiotics	1. Inhibit the growth of harmful bacteria, resist the infection of pathogenic bacteria; 2. synthesize vitamins needed by the human body, promote the absorption of minerals by the human body; 3. produce acetic acid, propionic acid, lactic acid, and other organic acids to stimulate intestinal peristalsis; 4. promote defecation, prevent constipation and inhibit intestinal spoilage, purify intestinal environment, decompose carcinogens; 5. stimulate the human immune system, it plays an important role in improving disease resistance. 6. At present, studies have pointed out that Bifidobacterium may be pathogenic, but the specific mechanism is unknown.	Bottacini et al.23
Coprococcus	Gram-positive anaerobic bacteria	probiotics	1. Produce butyric and acetic acids as well as formic or propionic acid and/or lactic acid, etc; 2. it is one of the important producers of butyric acid and one of the probiotics that measure gut health; 3. helps suppress the immune response and reduce the severity of allergic reactions.	Yang et al.24
Faecalibacterium	Gram-negative anaerobic bacteria	probiotics	1. One of the main probiotics for the production of butyrate, has anti-inflammatory effects, maintains intestinal barrier permeability, and protects the digestive system from intestinal pathogens; 2. it is one of the probiotics that measure gut health and is a core bacterium in the field of future probiotic research.	Martín et al.25
Lactobacillus	Gram-positive facultative anaerobic bacteria	probiotics	1. Lactobacillus and Bifidobacterium are currently the most studied probiotics; 2. Lactobacillus can ferment glucose under anaerobic or hypoxia conditions to produce beneficial substances such as lactic acid; 3. digest and metabolize proteins and carbohydrates, synthesize B vitamins and vitamin K, catabolic bile salts, and enhance innate and acquired immunity; 4. inhibits proinflammatory mediators and has antibacterial activity against a range of pathogens.	Duar et al.26
Parabacteroides	Gram-negative anaerobic bacteria	probiotics	1. Its main products are acetic acid and succinic acid; 2. it also has a wide range of cholic acid conversion functions, can hydrolyze a variety of binding cholic acids, while converting to a variety of secondary cholic acids, improve lipid metabolism disorders.	Cui et al.27
Prevotella	Gram-negative anaerobic bacteria	probiotics, opportunistic pathogen	1. Prevotella is often considered a bacterium associated with a healthy plant-based diet, acting as a "probiotic" in the human body; 2. it mainly produces propionic acid; 3. Prevotella may interact with other flora by promoting carbohydrate fermentation, inducing visceral hypersensitivity and exacerbating IBS symptoms; 4. it contains enzymes that play an important role in mucin degradation and may lead to increased intestinal permeability, damage the intestinal barrier; 5. it has proinflammatory function, exacerbates intestinal inflammation.	Tett et al.28; Sharma et al.29
Roseburia	Gram-positive anaerobic bacteria	probiotics	1. Along with Faecalibacterium, it is an important genus of butyricogenes in the gut; 2. it can stimulate anti-inflammatory factors, regulate immune T cells, and have a strong effect on inhibiting inflammation; 3. it can improve gut biodiversity, improve glucose tolerance, aid weight loss, and rejuvenate colon cells.	Zhang et al.30
Ruminococcus	Gram-positive anaerobic bacteria	probiotics, opportunistic pathogen	1. Ruminococcus is "key bacteria in degrading resistant starch," produce acetate and formate with anti-inflammatory benefits; 2. stabilizes the intestinal barrier, reverses diarrhea, reduces colorectal cancer risk, reduces kidney stones, and increases energy; 3. some Ruminococcus genera have been shown to be proinflammatory. It can effectively induce dendritic cells to secrete inflammatory cytokine (TNF-α).	Crost et al.31

At the genus level, we observed no significant change in the abundance of Akkermansia, Parabacteroides, Faecalibacterium, Coprococcus, Roseburia, and Agathobacter. However, the abundance of Ruminococcus (p = 0.008) and Blautia (p = 0.032) significantly decreased after just 1 week of dulaglutide administration. Additionally, there was a significant increase in the abundance of Lactobacillus (p = 0.002) after 1 week of dulaglutide administration, and the abundance of Bacteroides (p = 0.011) showed a significant increase after 48 weeks of dulaglutide administration. Notably, the abundance of Bifidobacterium (p < 0.001) and Prevotella (p = 0.023) showed a significant decrease after 1 week of dulaglutide administration but increased significantly after 48 weeks of dulaglutide administration compared with the abundance before dulaglutide administration.

Spearman correlation analysis

We used Spearman correlation analysis to determine the association of FBG level, HbA1c concentration, fasting C-peptide level, and BMI with the abundance of common gut microorganisms before and after the administration of GLP-1 RAs. We found that FBG was significantly positively correlated with the abundance of Firmicutes, Ruminococcus, and Blautia, whereas FBG was significantly negatively correlated with the abundance of Bifidobacterium (Figure 9). Additionally, BMI was significantly positively correlated with the abundance of Firmicutes and Blautia and significantly negatively correlated with the abundance of Parabacteroides (Figure 9). HbA1c was significantly positively correlated with the abundance of Firmicutes, Blautia, and Ruminococcus and significantly negatively correlated with the abundance of Bacteroidota, Bacteroides, and Bifidobacterium (Figure 9). Furthermore, fasting C-peptide was significantly negatively correlated with the abundance of Ruminococcus (Figure 9).

Discussion

Nutrients, such as carbohydrates, stimulate intestinal L-cells to secrete the enteroglucagon GLP-1 into the bloodstream, which signals GLP-1R on pancreatic islet β cells to stimulate glucose-dependent insulin secretion, thereby playing an important role in glucose homeostasis.³² Additionally, GLP-1R is widely distributed in the central nervous system including the hypothalamus. GLP-1R acts on the gastrointestinal tract through the gut-brain axis when it receives the GLP-1 signal, thus inducing a sense of satiety and appetite reduction and playing a role in body weight reduction.³³ Further, intestinal flora can alter gastrointestinal motility while inhibiting GLP-1R expression in intermuscular neuronal cells throughout the gastrointestinal tract.³⁴ Grasset et al.³⁵ demonstrated that specific intestinal flora dysbiosis in T2DM mice occurs via a gut-brain axis mechanism, which results in GLP-1 resistance. Moreover, the relative abundance of Lactobacillus in the ilea of the T2DM mice decreased significantly, whereas that of Bacteroidota, Burkholderia, and Clostridium increased significantly, with a strong negative correlation between the relative abundance of Bacteroidales in Bacteroidota and the concentration of GLP-1R in the ileum.³⁵

Intestinal flora can play a role in controlling energy homeostasis via dietary fiber fermentation, thereby producing SCFAs. The human farnesoid X receptor (FXR) regulates the expression of free fatty acid receptor 2 in the SCFA receptor and Gαq signaling downstream, thereby inducing GLP-1 secretion.³⁶ Additionally, intestinal flora directly regulates FXR and G protein-coupled BA receptor 1 signaling by affecting BA levels in vivo, thereby modulating the secretion of gastrointestinal hormones such as GLP-1.³⁷ Thus, GLP-1 plays a crucial role by means of intestinal flora in affecting T2DM development.

Intestinal flora in the human body regulates GLP-1 R expression and intestinal GLP-1 secretion levels. Moreover, GLP-1 RA regulates the composition and structure of intestinal flora. Intestinal flora and GLP-1 seem to have a positive feedback relationship. Zhao et al.³⁸ found that the body weight of diabetic obese rats decreased and insulin resistance improved significantly after 12 weeks of liraglutide administration. Furthermore, 16S rRNA amplification sequencing showed a significantly decreased abundance and diversity of intestinal flora in rats, with a significantly decreased abundance of obesity-associated flora, such as Firmicutes, and a significantly increased abundance of lean-associated flora, such as Bacteroidota.³⁸ Madsen et al.³⁹ demonstrated that, after treating simple obese mice with liraglutide and dual GLP-1/GLP-2 receptor agonist (GUB09-145) for 4 weeks, α-diversity indices for both, the abundance and diversity of intestinal flora in mice, decreased significantly. The relative abundance of Firmicutes, Actinobacteria, and Proteobacteria increased, whereas that of Bacteroidota decreased significantly at the portal level.³⁹ Although animal experiments have shown that GLP-1 RA administration reduces body weight and improves insulin resistance by modulating the composition and structure of intestinal flora of mice, only a few clinical studies have been performed on humans.

A study⁴⁰ recruited 52 diabetic subjects, who were divided into efficacious and non-efficacious groups based on whether GLP-1 RA administration was efficacious or not after 12 weeks of intervention. The β-diversity analysis showed significant differences between the gut flora of the two groups. The abundance of Lachnoclostridium and Butyricicoccus was positively correlated with a decrease in blood glucose levels, whereas the abundance of Prevotella, Ruminococcaceae, Bacteroidales, Eubacterium coprostanoligenes, Dialister succinatiphilus, Alistipes obesi, Mitsuokella, and Lactobacillus mucosae was negatively correlated with insulin resistance. These results suggested that intestinal flora and glycemic response to GLP-RA treatment are closely related.⁴⁰ To the best of our knowledge, this is the first and only study on the association between GLP-RA and intestinal flora with respect to the clinical treatment of T2DM; however, it did not investigate differences in intestinal flora before and after GLP-RA treatment for T2DM.

To exclude the interference of several factors in the experiment including disease, medication, and diet,^41,42,43 we recruited newly diagnosed patients with clinical T2DM, uniformly performed physical examinations, and provided the same type of diet (Table 4). Blood and BMI indices of the patients before, after 1 week, and after 48 weeks of dulaglutide administration were determined, which confirmed the hypoglycemic and weight-loss effects of dulaglutide. Simultaneously, we collected feces samples from the patients to detect differences in the intestinal flora by 16S rRNA sequencing before, after 1 week, and after 48 weeks of dulaglutide administration.

Table 4.

Meals served at the Nutrition restaurant during the trial

Food	Weight	Energy (kcal)
Rice, steamed bread, churros	25 g	90
Soybean milk	400 mL	90
Fresh milk	160 mL	90
Edamame	70 g	90
Potato	100 g	90
Lotus root	130 g	90
Green vegetables, eggplant	400 g	90
Cabbage, celery	500 g	90
Cucumber, winter melon, tomato	720 g	90
Beef	40 g	90
Pork, chicken, eggs	50 g	90
Fish	80 g	90
Shrimp	160 g	90
Oil	10 g	90

Breakfast intake was 400 kcal; lunch intake is 800 kcal; dinner intake is 400 kcal.

The α-diversity analysis showed no significant changes in the abundance and diversity of intestinal flora before and 1 week after dulaglutide administration, in contrast to the results of the study by Kato et al..¹² They observed the significantly altered intestinal flora of wild mice after 16 h of liraglutide injection, with a significant decrease in Bacteroidota abundance and an increase in Actinobacteria abundance at the phylum level, as well as a significant decrease in Ruminococcus abundance at the genus level.¹² This discrepancy was attributed to the difference in the research subjects. These results indicated the inconsistent responsiveness of gut flora to GLP-1 RA administration in different species at different time points. Additionally, the different half-lives of liraglutide and dulaglutide probably contributed to these discrepant results.

Further, we administered dulaglutide for up to 48 weeks and observed a significant change in the α-diversity index. After 48 weeks of dulaglutide injection, microbial abundance in the intestines of T2DM subjects decreased significantly, which is consistent with previous results.^38,39 However, the present study showed that gut flora diversity was not altered, which is inconsistent with animal experiment results, most of which showed the significantly decreased abundance and diversity of gut flora after GLP-1 RA administration.^38,39

The β-diversity analysis showed little variability in the microbial communities of intestinal flora after 1 week of dulaglutide administration compared with that in the microbial communities of intestinal flora before dulaglutide administration. The results indicated no significant changes in the microbial communities of intestinal flora after 1 week of dulaglutide administration, in line with the α-diversity analysis results. After 48 weeks of dulaglutide administration, intestinal flora was significantly different from that before and after 1 week of dulaglutide administration, confirming that the long-term clinical injection of dulaglutide may alter the composition and structure of human intestinal flora.

To further validate our conjecture, we selected 12 microorganisms commonly found in the human gut and examined their abundance in each sample before, after 1 week, and after 48 weeks of dulaglutide administration by 16S high-throughput sequencing. Subsequently, we validated the obtained results by performing medical statistical analyses. The abundance of Firmicutes and Bacteroidota, which are closely associated with T2DM, did not differ significantly between before and after 1 week of dulaglutide administration, whereas their abundance changed significantly after 48 weeks of dulaglutide administration.⁴⁴ This is consistent with the α-diversity and β-diversity bioinformatics analysis results. At the genus level, the abundance of Ruminococcus and Blautia, Firmicutes that do not produce butyric acid,^31,45 decreased significantly after 1 week of dulaglutide administration. The abundance of Bacteroides and Lactobacillus increased significantly after dulaglutide administration. The abundance of Bifidobacterium and Prevotella decreased significantly after 1 week of dulaglutide administration, whereas their abundance after 48 weeks of dulaglutide administration increased significantly compared with that before dulaglutide administration. Overall, the present findings are consistent with those of a majority of previous studies.⁴⁶

Gurung et al.⁴⁶ summarized 42 studies on T2DM and intestinal microorganisms and pointed out that Bifidobacterium, Bacteroides, Faecalibacterium, Akkermansia, and Roseburia were negatively correlated with T2DM, whereas Ruminococcus, Fusobacterium, and Blautia were positively correlated with T2DM. Thereinto, Bacteroides, and Bifidobacterium were the most frequent beneficial genera, which were negatively correlated with T2DM.⁴⁶ Herein, the abundance of Bacteroides and Bifidobacterium increased significantly after 48 weeks of dulaglutide administration, whereas the abundance of Bifidobacterium and Prevotella decreased significantly after 1 week of dulaglutide administration. The study could not explain this discrepancy; thus, we speculated it as a concomitant phenomenon of GLP-1 RA. Because Bifidobacterium, Bacteroides, and Prevotella are most closely related to diet, they are susceptible to interference by various dietary factors in the intestine.⁴⁷ Additionally, the biggest clinical side effects of GLP-1 RA are gastrointestinal reactions such as nausea, vomiting, and diarrhea, which usually appear in the first and second weeks of drug administration and then gradually disappear.⁴⁸ This might have affected the gut flora results after 1 week of dulaglutide administration. Bacteroides and Prevotella, two genera belonging to Bacteroides, exhibit a competitive relationship under natural dietary conditions.⁴⁹ Bacteroides are associated with protein and fat intake, whereas Prevotella is associated with fiber and carbohydrate intake; Bacteroides and Prevotella are negatively correlated in the gut.⁵⁰ Herein, the abundance of Bacteroides and Prevotella increased significantly after the long-term use of dulaglutide, which reflected the obvious effect of GLP-1 RA on intestinal flora.

Akkermansia, a traditional probiotic, is associated with T2DM.⁵¹ Herein, the abundance of Akkermansia increased significantly after 1 week of dulaglutide administration and then decreased at 48 weeks but was higher than the abundance before dulaglutide administration. The p value of the three groups as a whole was exactly 0.05; thus, we could not strictly say that there was a significant association between Akkermansia and T2DM. Additionally, Faecalibacterium and Roseburia are the main probiotics that produce butyric acid, and several studies have reported that they are negatively correlated to T2DM.^52,53 Herein, no statistical difference was observed in the abundance of these two bacteria after dulaglutide administration. However, Lactobacillus abundance increased significantly after dulaglutide administration. Lactobacillus is one of the earliest and most widely studied probiotics; however, its role in T2DM is controversial. Among the six articles reviewed by Gurung et al., five of them reported that the Lactobacillus genus was positively correlated to T2DM. Additionally, more species-level microorganisms have been found in the Lactobacillus genus, with great differences in their abundances, which hinders the consistency of genus-level analysis.⁴⁶

Blautia, similar to Lactobacillus, is a controversial genus of microorganisms. Most studies have shown that Blautia abundance increases in T2DM;⁴⁶ however, recent studies have shown that treatment with metformin against T2DM significantly increases Blautia abundance,⁵⁴ which is in contrast with the present results. As probiotics, Blautia prevents pathogen colonization by producing bacteriocins, exhibits anti-inflammatory properties, and maintains glucose homeostasis by increasing regulatory T cell and SCFA production.⁵⁵ However, other studies have suggested that Blautia may damage the intestinal barrier.²² We believe that the relationship of Lactobacillus and Blautia with T2DM is an important direction for future research.

The Spearman’s correlation analysis showed a significant positive correlation between the levels of FBG and HbA1c and the abundance of Firmicutes, Ruminococcus, and Blautia in the intestinal tract and a significant negative correlation with the abundance of Bifidobacterium when treating patients with T2DM using dulaglutide. Additionally, a significant negative correlation was observed between HBA1c concentration and Bacteroidota and Bacteroides abundance. BMI was significantly positively correlated with Firmicutes and Blautia abundance and significantly negatively correlated with Parabacteroides abundance. Finally, fasting C-peptide levels were significantly negatively correlated with Ruminococcus abundance. Among them, the relationship of Blautia with the disease is controversial. A recent cross-sectional study in Japan confirmed Blautia as a symbiotic bacterium negatively associated with obesity and T2DM.⁵⁶ Another study showed that Blautia in the gut was negatively associated with BMI.⁵⁷ Currently, no relevant studies are available to clarify this discrepancy. However, this does not affect the result that the effect of GLP-1 RA on hypoglycemia and weight loss is closely related to the composition and structure of human intestinal flora. Additionally, our results confirmed that Ruminococcus negatively affected insulin secretion in humans. Zhai et al.⁵⁸ found that the derived metabolic scouting amine and phenylethylamine from Ruminococcus directly damaged insulin signaling in the main metabolic tissues of healthy mice and monkeys, inhibited insulin secretion, and induced insulin resistance. We believe that dulaglutide may reduce Ruminococcus abundance and its metabolites in the intestine, thereby positively affecting insulin secretion and increasing fasting C-peptide levels.

We used PICRUSt2 to predict changes in metabolic pathways after dulaglutide administration. No significant change was observed in metabolic pathways after 1 week of dulaglutide administration; however, after 48 weeks, the primary KEGG showed significantly improved pathways of environmental information processing, metabolism, organismal systems, and human diseases. We focused on the secondary KEGG pathway of the metabolic pathway analysis and found that the long-term use of dulaglutide improved pathways of cellular community-prokaryotes, energy metabolism, terpenoid and polyketide metabolism, cofactor and vitamin metabolism, carbohydrate metabolism, glycan biosynthesis and metabolism, lipid metabolism, transport, and catabolism, and biosynthesis of other secondary metabolites.

Improved cellular community-prokaryotes suggested that, after 48 weeks of dulaglutide administration, significant changes were observed in gut microbes.⁵⁹ Additionally, the weight loss effect of dulaglutide administration was confirmed because energy metabolism changed significantly.⁶⁰ Terpenoids and polyketides, secondary metabolites with diverse structures, are found in nature mostly in the form of glycosides and are widely used in medicine, healthcare products, spices, nutrients, mosquito repellents, and colorants.⁶¹ Existing studies indicate that terpenoids and polyketides may exert antitumor, antibacterial, hypoglycemic, and neuroprotective effects; however, the underlying mechanisms remain unknown.⁶²

Although humans obtain cofactors and vitamins from their daily diet, probiotic-rich gut microbiomes are additional sources of vitamins and cofactors that are vital in physiology.⁶³ Cai et al.⁶⁴ found a significant decrease in cofactor and vitamin metabolism in 3-deoxyglucosone-induced diabetic rats, whereas we found that dulaglutide reversed this effect. Additionally, owing to the significant increase in the abundance of probiotics such as Bifidobacterium and Bacteroides, after the long-term use of dulaglutide, carbohydrate metabolism, glycan biosynthesis and metabolism, and lipid metabolism were significantly improved.^46,47,65,66 Improved transport and catabolism suggested that dulaglutide enhanced intestinal epithelial permeability via intestinal bacteria-epithelial cell interactions; however, this is only our conjecture, and the specific mechanism needs further elucidation.⁶⁷

In conclusion, the intestinal flora of the human body was not significantly altered during the short-term use of dulaglutide against T2DM. However, during its long-term use, the abundance of intestinal microorganisms decreased significantly, the diversity did not change, and the overall composition and structure of the flora changed significantly compared with those before drug administration. Additionally, FBG level, BMI, HbA1c concentration, and fasting C-peptide level were closely related to changes in the abundance of certain bacteria. Nevertheless, the mechanism of the action of GLP-1 RA in T2DM treatment through gut microbes needs further investigation.

Limitations of the study

This study has some limitations. First, the sample size of the study was small. Owing to the long period, restrictive requirements, and great difficulty in clinical research on gut flora, we included only 41 subjects in this trial. Second, the mechanism of GLP-1 RA regulating microbiota was not elucidated. Because this trial is a preliminary exploratory study, we did not elucidate the specific mechanism in depth during the trial design. In the future, we will include more research subjects and thoroughly investigate the mechanism of intestinal flora in T2DM treatment using GLP-1 RA.

STAR★Methods

Key resources table

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Biological samples
Fecal samples from 41 patients with newly diagnosed type 2 diabetes mellitus	The First Affiliated Hospital of Kunming Medical University	N/A
Chemicals, peptides, and recombinant proteins
Dulaglutide	VetterPharma-FertigungGmbH & Co.KG	N/A
Critical commercial assays
TruSeq DNA PCR-Free Sample Preparation kit	Illumina, San Diego	Cat: 20015963
Phusion@High-fidelity PCR Master Mix with GC Buffer	New England Biolabs	Cat: M0532s
Deposited data
Raw 16S rRNA gene sequencing data	This paper	NCBI Sequence Read Archive (SRA) :PRJNA871098 (http://www.ncbi.nlm.nih.gov/bioproject/871098)
Oligonucleotides
Upstream primer for 16s DNA PCR: (5'CCT-ACG-GGN-GGC-WGC-AG'3)	This paper	N/A
Downstream primers for 16s DNA PCR: (5'GGA-CTA-CHV -GGGG-TWT-CTA-AT'3)	This paper	N/A
Software and algorithms
QIIME2 (2021.4)	This paper	RRID: SCR_021258 https://qiime2.org/
Lefse (1.1.2)	This paper	RRID: SCR_014609 http://huttenhower.sph.harvard.edu/galaxy
R-software (4.0.3)	This paper	RRID: SCR_001905 http://www.r-project.org/
KEGG database	This paper	RRID: SCR_001120 http://www.genome.jp/kegg/expression/
Silva database (Silva.v138)	This paper	RRID: SCR_006423 http://www.arb-silva.de
GraphPad Prism 5	This paper	RRID: SCR_002798 http://www.graphpad.com/
Other
Fully Automated Biochemistry Analyzer	C16000, Abbott, Abbott Park, IL, USA	N/A

Resource availability

Lead contact

Further information and requests for resources and reagents should be directed to and will be fulfilled by the lead contact, Prof. Xin Nian (nianxinkm@hotmail.com).

Materials availability

This study did not generate new unique reagents.

Data and code availability

16S rRNA amplicon sequencing data have been deposited at SRA and are publicly available as of the date of publication. Accession numbers are listed in the key resources table. The DOI is listed in the key resources table.

Any additional information required to reanalyze the data reported in this paper is available from the lead contact upon request.

Experimental model and study participant details

Study subjects

Following the diagnostic criteria for T2DM proposed by the World Health Organization in 1999, we recruited newly diagnosed patients with T2DM as study subjects in the inpatient Department of Endocrinology I at the First Affiliated Hospital of Kunming Medical University. The relevant informed consent form was signed after the subjects were informed in detail about the purpose and protocol of the study. The patients were included in this trial based on certain inclusion and exclusion criteria. A total of 41 subjects were included in the study, including 26 males and 15 females, with a mean age of 48.12 ±13.13 years (Figure 1) (Table 1).

The inclusion criteria were as follows: 1. Newly diagnosed patients with T2DM who have not previously consumed drugs such as GLP-1 RA, metformin, and acarbose, 2. Patients did not suffer from diabetic ketosis, 3. Patients with good physical fitness and being able to live on their own, and 4. Patients who were aware of the purpose and methodology of this trial and volunteered to participate by signing the relevant informed consent form.

The exclusion criteria were as follows: 1. Patients consuming probiotics, antibiotics, glucocorticoids, and other medications that may have affected their gut flora during the last 3 months, 2. Patients with acute and chronic diseases, such as cardiovascular and cerebrovascular diseases, inflammatory enteritis, pneumonia, cancer, Alzheimer’s disease, and autoimmune diseases, which may affect the gut microbiota, 3. Patients who were unable to consume GLP-1 RA such as those with medullary carcinoma of the thyroid gland, history of pancreatitis, pregnancy, and breastfeeding, 4. Patients with poor compliance during the trial and failure to use GLP-1 RA as required, 5. Patients with the sudden onset of acute or chronic illnesses, such as cardiovascular and cerebrovascular diseases, inflammatory enteritis, pneumonia, cancer, Alzheimer’s disease, and autoimmune diseases, may affect the gut microbiota during the trial, 6. Patients consuming drugs, such as probiotics, antibiotics, and glucocorticoids, which may affect their gut microbiota owing to sudden illnesses during the trial, 7. Patients showing significant changes in diet, lifestyle, or mental health during the trial, such as staying up too long, causing mental distress, and changing diet, and 8. Patients who voluntarily requested to be removed from the trial.

The study was conducted in strict accordance with the Declaration of Helsinki and the Four Fundamental Principles of Ethics; all subjects provided their written informed consent. The study was approved by the Ethics Committee of the First Affiliated Hospital of Kunming Medical University for ethical review and agreed to be implemented in the Department of Endocrinology of the First Affiliated Hospital of Kunming Medical University (Approval No. 2022L79). The study was successfully registered with the China Clinical Trial Registry (registration number: ChiCTR2200063198).

Diet plan

The trial was divided into two phases. The first phase was of 10 days, including a 3-day washing period and a 7-day trial period. In this phase, the subjects were residing at the Endocrinology Department of the First Affiliated Hospital of Kunming Medical University and consuming a unified diet (Table 4). The purpose of the 3-day washing period was to ensure that dietary differences in gut flora were reduced during the trial period. In the 7-day trial period, the short-term effects of GLP-1 agonists on the gut microbiota of subjects with newly diagnosed T2DM were observed. However, given that the subjects could not be observed in the hospital for a long period, they were allowed to return home to continue the trial until 48 weeks in the second phase. In this phase, the subjects were asked to continue consuming the previous diet and return visits via phone, WeChat, and other means. Subjects were excluded from the trial if they did not follow the diet prescribed in the second phase of the trial.

Grouping of samples

We grouped the collected and preserved samples such as samples of the subjects before dulaglutide administration were allotted to group F, samples after 1 week of dulaglutide administration were allotted to group S, and samples after 48 weeks of dulaglutide administration were allotted to group T.

Method details

Collection and preservation of fecal specimens

The subjects collected early morning fasting feces (inside the mid-section) in a sterile stool tube on the first day of the trial, and the specimens were stored in our laboratory at −80°C. The subjects were then injected with dulaglutide (1.5 mg, QW) under the supervision of the researchers, and after 1 week, their morning fasting feces (inside of the mid-section) were collected in the same way. After 48 weeks of dulaglutide administration, early morning fasting feces (inside the mid-section) were collected and stored in a refrigerator at −80°C.

Venous blood testing and BMI measurement

Fasting venous blood was collected early in the morning on the day of each stool sample. Fasting C-peptide, FBG, and HbA1c levels were measured using an automatic biochemical detector. The subjects were then instructed to remove their clothing and hats and their height and weight were measured to calculate a BMI.

Fecal 16S rRNA amplicon sequencing

Microbial amplicon building

The microbial amplicon library was constructed using a two-step library construction method. First, DNA was used as a template, and primers with junctions were designed to perform polymerase chain reaction (PCR). Second, the resulting PCR product was used as a template to perform PCR. The primer junctions facilitated barcode/index addition in the second step. The barcode/index was used to differentiate between the base sequences of the samples.

DNA extraction and PCR amplification of target fragments

The genomic DNA of the samples was extracted using the SDS method,⁶⁸ for which, 50 ng of genomic DNA, 10 μM Vn F (5'CCT-ACG-GGN-GGC-WGC-AG'3), 10 μM Vn R (5'GGA-CTA-CHV -GGGG-TWT-CTA-AT'3), 5 μL of KOD FX Neo Buffer, 2 μL of dNTP (2 mM each), 0.2 μL of KOD FX Neo (TOYOBO), and 10 μL of ddH₂O was added to the PCR system. A pre-denaturation of 95°C was performed for 5 min, followed by denaturation at 95°C for 30 s, annealing at 50°C for 30 s, and extension at 72°C for 40 s. Twenty-five cycles of denaturation, annealing, and extension were performed. Finally, extension was performed at 72°C for another 7 min.

Solexa PCR

The PCR products obtained were purified using a PCR product purification kit PCR amplification was performed by taking 5 μL of the purified PCR product of the target region, 2.5 μL of 2 μM MPPI-a, 2.5 μL of 2 μM MPPI-b, and 10 μL of 2×Q5 HF MM. Pre-denaturation at 98°C was performed for 30 s, followed by denaturation at 98°C for 10 s, annealing at 65°C for 30 s, and extension at 72°C for 30 s. Ten cycles of denaturation, annealing, and extension were performed. Finally, extension was performed at 72°C for 5 min.

PCR product recovery and up-sequencing

The PCR products were electrophoresed on a 1.8% agarose gel at 120 V for 40 min. The PCR products from the previous step were purified on a DNA purification column. Agarose gel (1.8%) at 120 V for 40 min was used for further electrophoresis, and the target fragments were cut and recovered using a DNA gel recovery kit (Monarch). The purified PCR product was subjected to PE250bp sequencing using the Illumina platform to obtain the 250 bp bipartite sequencing reads we needed.

Quantification and statistical analysis

Bioinformatics analysis

The DADA2 method⁶⁹ was used to perform filtering, denoizing, splicing, and dechimerization to determine the quality of the raw sequencing data in the beginning. After obtaining the valid data, deweighting was performed on the valid data to acquire the deweighted sequence ASV, which was then normalized using the diversity core-metrics-phylogenetic command in QIIME2, and the depth of the normalized data was set to be 95% of the minimum sample sequence volume. Finally, we obtained the required ASVs.

The ASVs were analyzed for alpha diversity using the ggplot2 package in R to assess variability in gut microbial abundance and diversity in each sample. Unifrac distance-based β-diversity analysis was performed using the vegan toolkit in R to assess variations in the composition and structure of the gut flora of groups F, S, and T. The reliability of β-diversity was verified by performing Adonis analysis.⁷⁰

The ASVs were compared using the Silva database in the field of microbiology expertise, and the taxonomic information of microbial species corresponding to each ASV was matched. The composition of the gut microbial community of each sample was determined at each taxonomic level (phylum, order, order, family, genus, and species), and the abundance table of the species was generated. The community structure of each sample was plotted at different taxonomic levels using the R language software.

Phylogenetic Investigation of Communities by Reconstruction of Unobserved States (PICRUSt2) predicts the functional abundance of a sample based on the abundance of labeled gene sequences in the sample. By performing the differential analysis of the KEGG database metabolic pathways, differences, and changes in the functional genes of microbial communities associated with metabolic pathways between the samples of different groups were observed, and metabolic functional changes in the community samples were studied to adapt to environmental changes.⁷¹

Statistical analysis

All data were statistically analyzed using GraphPad Prism 5. Two-group measures that conformed to normal distribution were expressed as the mean ±standard deviation (X ±S) and analyzed using paired t-tests. Multiple-group measures were also expressed as X ±S and analyzed using repeated measures analysis of variance. Measures that did not conform to normal distribution were analyzed using the paired Wilcoxon rank-sum test for two-group samples and the Friedman rank-sum test for multiple-group samples. P < 0.05 was considered statistically significant.

Published: April 22, 2024