1. General information
Nineteen IMNM patients and 23 healthy individuals were included in our research. The demographic characteristics, clinical features and laboratory data of the two groups are shown in Table 1. The average ages of the IMNM and HC groups were 56.50 ± 7.72 and 50.35 ± 8.55 years old, respectively. No significant difference was found between the groups (p = 0.383). The ratio of women to people in the IMNM group was 1.119/19 (47.37), and that in the healthy group was 11/23 (47.82). The gender distribution of the two groups was also verified to have no difference (p > 0.999).
Table 1
Clinical manifestations and laboratory data of IMNMs and HCs.
Characteristics | HC(n = 23) | IMNM(n = 19) | p |
Demographic data | | | |
Age (years), mean ± SD | 50.35 ± 8.55 | 56.50 ± 7.72 | 0.383 |
Female sex, no./total no (%) | 11/23 (47.82) | 9/19 (47.37) | > 0.999 |
BMI (kg/m2), mean ± SD | / | 21.35 ± 3.35 | / |
Disease course (months), median (range) | / | 18.25 (1, 108) | / |
Clinical manifestation, no./total no (%) | / | | / |
myasthenia | / | 15/19 (78.95) | / |
myalgia | / | 10/19 (52.63) | / |
amyotrophy | / | 4/19 (21.05) | / |
Laboratory data | / | | / |
ANA (+), no./total no (%) | / | 15/19(78.95) | / |
Anti-SSA (+), (no./total no (%)) | / | 5/19(26.32) | / |
Anti-Ro-52 (+), (no./total no (%)) | / | 6/19(31.58) | / |
HMGCR (+), (no./total no (%)) | / | 3/19(15.79) | / |
SRP (+), (no./total no (%)) | / | 9/19(47.37) | / |
ALT (IU/L), median(range) | / | 112.47 (19, 334) | / |
AST (IU/L) median(range) | / | 106.16 (12, 302) | / |
CK (IU/L), median(range) | / | 2621.58 (38, 9453) | / |
LDH (IU/L), median(range) | / | 637.58 (127, 1484) | / |
HBDH (IU/L), median(range) | / | 500.95 (91, 1199) | / |
MYO (ng/mL), median(range) | / | 1045.817 (21, 3000) | / |
CKMB (ng/mL), median(range) | / | 134.11 (1.75, 300) | / |
cTnT (ng/L), median(range) | / | 688.77 (9.9, 2131) | / |
ESR (%), median(range) | / | 31 (2, 113) | / |
CRP (mg/L), median(range) | / | 8.74 (1, 47.4) | / |
Occurrence of ILD, no./total no (%) | / | 8/19 (42.11) | / |
Abbreviation: IMNM: immune-mediated necrotizing myopathy; HC: healthy control; SD: standard deviation; BMI: body mass index; ANA: antinuclear antibody; SSA: Anti-SSA antibody; Anti-Ro-52: Anti-Ro-52 antibody; SRP: signal recognition particle; HMGCR: 3-hydroxy-3-methylglutaryl-coenzyme A reductase; ALT, AST, CK: creatine kinase; MYO: myoglobin; CKMB: creatine kinase muscle and brain isoenzyme; ; cTnT: cardiac troponin T; ESA: erythrocyte sedimentation rate; CRP: C-reactive protein; ILD: interstitial lung disease.
2. Diversity Analysis Of Gut Microbiota In The Imnm Group And Healthy Group
A total of 42 samples were used for 16S rDNA gene sequencing in our study. The rarefaction curves of all samples on the Shannon index reached a plateau, suggesting that the sequencing depth commonly covered all species and that the sequencing data had reached saturation, which indicates that the sequencing depth could reflect the community structure and diversity of the species to some extent (Fig. 1A).
The gut microbiota diversity and richness of the HC and IMNM groups were evaluated by the Shannon, Simpson and Chao1 indexes based on the OTU species and abundance. The results showed that the community diversity indexes, including the Shannon index (p = 0.0002) (Fig. 1B) and Simpson index (p = 0.002) (Fig. 1C), were significantly different between the HC and IMNM groups. Additionally, the Chao1 index (p = 0.001) indicated that the OTU richness was obviously different between the two groups (Fig. 1D). The results showed that the gut microbiota diversity and richness of IMNM patients were lower than those of healthy people.
To prove whether the overall structure of the gut microbiota of the HC group differed from that of the IMNM group, beta diversity was illustrated by PCoA based on Bray‒Curtis distances. We found that beta-diversity plots could differentiate the gut microbiota of the two groups (ANOSIM test; R = 0.288, P = 0.001) (Fig. 1E).
The number of common OTU units in the IMNM group and the HC group was 576, the number of unique OTU units in the HC group was 188, and the number of unique OTU units in the IMNM group was 394; the difference was statistically significant (P < 0.001) (Supplementary Fig. 1).
3. Composition And Abundance Distribution Of Gut Microbiota In Imnm Patients
First, we analyzed microbial richness at the phylum level. The pie charts revealed that the dominant phyla across all subjects were Firmicutes, Bacteroidota, Actinobacteriota and Proteobacteria, whose relative abundance together accounted for more than 99% of bacterial sequences (Fig. 2A). Among these four dominant bacteria, the relative abundance of Bacteroidota distinctly decreased from the HC group to the IMNM group (p = 0.003), while alterations in the others were not statistically significant (Fig. 2B).
Then, we focused on specific differences in microbiota composition at the genus level between the two groups. The considerable difference in the distribution of gut microbiota at the genus level in the two groups was shown by the genus bar plot, and the top dominant genus in the HC group was Bacteroides, while that in the IMNM group was Lactobacillus (Supplementary Fig. 2A). Furthermore, we applied Welch’s T test to compare the differences in the genera between the two groups (Fig. 2C). In IMNM patients, the relative abundances of Bacteroides, Roseburia and Coprococcus were decreased compared to those in the HC group (P < 0.05) (Supplementary Fig. 2B), while the relative abundances of Lactobacillus and Streptococcus were relatively increased (P < 0.05) (Supplementary Fig. 2C). The above findings were further verified by LefSe, as shown in the cladogram, and the contributory discriminate taxa with LDA scores more than 4 were plotted for each group (Fig. 3A, B).
4. Relationships Between Gut Microbiota At The Genus Level And Clinical Indicators In Imnm Patients
To assess the association between gut microbiota changes and clinical indicators, Spearman correlation analysis was performed to analyze the correlation between the top five differential genera, including Lactobacillus, Bacteroides, Streptococcus, Roseburia and Coprococcus, and 22 clinical indicators with (1) sex, BMI and disease course of IMNM; (2) myositis antibodies, including anti-SRP antibodies, anti-HMGCR antibodies, ANA, SSA, and Ro52; (3) clinical symptoms and signs, including myasthenia, myalgia and amyotrophy; (4) disease activity parameters, including AST, ALT, CK, LDH, HBDH, MYO, CKMB, cTnT, ESR, and CRP; and (5) involvement of ILD (Fig. 4).
The study revealed that in IMNM patients, the relative abundance of Lactobacillus was positively correlated with the levels of anti-SRP antibodies and anti-Ro52 antibodies and ESR, while Streptococcus was positively correlated with anti-HMGCR antibodies and C-reactive protein (CRP). Both genera were significantly increased in IMNM patients. The results also showed that Roseburia, which was decreased in IMNM patients, was negatively correlated with the serum levels of MYO, cTnT, ESR, and CRP and the occurrence of ILD, which suggested that Roseburia was associated with the disease activity of IMNM and lung involvement. Bacteroides was negatively correlated with ESR and CRP, while Coprococcus was negatively correlated with ESR, and both genera were significantly decreased in IMNM patients.
Usually, Lactobacillus was reported that is one of the probiotics present in the human intestine to improve intestinal microenvironment, regulate immunity and promote nutrient absorption(18). It was reported that Lactobacillus decreased significantly in SLE (23) patients and IIM patients(22). Surprisingly, our results released the abundance of Lactobacillus increased and was positively correlated with the level of anti-SRP antibodies that play an important role in IMNM patients. Therefore, we performed the subgroup analysis according to the expression of anti-SRP antibodies. All participants were divided into three groups: healthy individuals (HC), SRP-negative IMNM patients (IMNM_SRP_negative), and SRP-positive IMNM patients (IMNM_SRP_positive). The result released that there was significant statistical difference among the three groups and the abundance of Lactobacillus was the most in SRP-positive IMNM patients, the second in SRP- negative patients, and the least in healthy individuals (Fig. 4B).
5. Examination Of Interactions Among Differential Genera
We established a correlation network to explore the coabundance and coexclusion interactions between 37 differential genera, as shown in Fig. 5A. The correlated genera were from four phyla, including Firmicutes, Bacteroidota, Actinobacteriota and Proteobacteria, and the majority of genera belonged to Firmicutes. The results showed that Lactobacillus was negatively associated with norank_f_Lachnospiraceae and norank_f_Ruminococcaceae. Additionally, a significant negative correlation was found between Bacteroides and Streptococcus. The other correlations within the community were positive.
6. Prediction model of the gut microbiota biomarker profile in IMNM patients
We next constructed a prediction model using the top five differential genera, including Lactobacillus, Bacteroides, Streptococcus, Roseburia and Coprococcus, to distinguish IMNM patients from healthy individuals. The efficacy of these differential genera in discriminating between IMNM patients and healthy people was set up using an ROC curve. The area under the ROC curve (area under curve, AUC) was 87%, and the 95% confidence interval (CI) was 73–100% (Figure 5B).