Influence of CTG repeats from the human DM1 locus on murine gut microbiota

Abstract

Myotonic Dystrophy type 1 (DM1) is caused by a CTG repeat expansion in the 3′ untranslated region of the DMPK gene. This expansion leads to the production of toxic RNA transcripts, which accumulate in the nucleus and interfere with normal RNA processing. DM1 affects a broad range of tissues and systems such as the skeletal muscle, the central nervous system, cardiac, visual, reproductive, and gastrointestinal (GI) system. GI dysfunction is a significant but poorly understood aspect of DM1. Particularly, it is unknown if there are alterations in the intestinal microbiome in DM1. Here, we used a transgenic humanized mouse model (DMSXL) to explore how the gut microbiome may be linked to GI issues in DM1. For this purpose, 68 stool samples from Homozygous, Heterozygous, and Wild-Type (WT) mice were collected. These samples were sequenced by MiSeq and analyzed with DADA2 to generate taxonomic profiles. Our analysis indicated that the overexpression of CTG repeats significantly influences the bacterial structure of the gut microbiome in Homozygous mice samples, especially in terms of the relative abundance of the Patescibacteria and Defferibacterota Phyla. These results provide valuable information about the gut microbiota structure thus improving the understanding of the role of these changes in the pathogenicity as well as GI problems of DM1 patients.

Graphical abstract

1. Introduction

Trinucleotide repeat expansions are the underlying cause of a group of genetic disorders, primarily affecting the nervous system, known as trinucleotide repeat disorders. These include Huntington's disease, fragile X syndrome, and myotonic dystrophy type 1 (DM1) [1]. DM1 is caused by an expansion of a triplet (CTG)_n repeats in the 3’ untranslated region of DMPK gene encoding a serine–threonine protein kinase [2]. As a result, the expanded DMPK transcript forms hairpin structures that bind and alters the function of RNA-binding proteins such as Muscleblind-like 1 (MBNL1) and CUGBP Elav-Like Family Member 1 (CELF1) [3]. These events lead to RNA foci formation, alternative splicing defects, and cellular dysfunction. For instance, splicing defects in skeletal muscle lead to misregulated chloride channel (CLCN1) expression, causing myotonia, a classic symptom of DM1 that contributes to the pathogenesis of the disease [3].

The clinical features of DM1 patients include progressive muscle weakness, atrophy, and myotonia as well as prominent multisystem involvement including heart conduction defects, and gastrointestinal (GI) alterations [4]. GI symptoms include constipation, diarrhea, cramping, abdominal pain, or bloating, occurring at various levels from the pharynx to the anal sphincter [5], [6], [7]. Little is known about the origin of GI manifestations in DM1 [8]and recommended care is based on clinical practice only [9]. Previous reports implied that GI symptoms may be due to alterations in hormonal secretion [10], [11] and myoelectric activity, which could impair the coordinated contractions required for proper GI motility [11], [12]. Moreover, smooth muscle dysfunction, a hallmark of DM1, is believed to contribute to dysmotility and other GI complications [13]. Recent reports suggested that Small Intestinal Bacterial Overgrowth (SIBO) may cause some of the GI symptoms, such as bloating, diarrhea, and malabsorption, in patients with DM1 [14], [15].

A previous study indicated a connection between the gut microbiota and the skeletal muscles that regulate muscle mass and function [16]. Dysregulation of the microbiota could be implicated in the pathogenesis of muscular dystrophies. For instance, it was shown in a mouse model of Duchenne Muscular Dystrophy that dystrophic mdx mice have a reduction in the number and abundance of different operational taxonomic units (OTUs), as well as large taxonomic modification of many phyla: Actinobacteria, Proteobacteria, Tenericutes, and Deferribacteres [17]. These changes were associated with an upregulation of circulating inflammatory markers. Dysbiosis was shown to correlate with the dystrophic symptoms and influence muscle immunity and fibrosis [18]. Despite these connections between microbiota and muscular dystrophies, the impact of DM1 on the microbiota and its potential role in contributing to symptoms remains unexplored.

Different mouse models of DM1 have been instrumental in providing detailed insights into the disease's pathogenic mechanisms [19], [20], [21], [22]. These models have enabled researchers to study specific features of disease progression, including molecular, cellular, and systemic abnormalities. Additionally, they serve as robust platforms for evaluating pharmacological interventions [23], [24], [25], [26], [27], [28], and RNA-based therapeutic strategies [29], which target specific targets of the pathobiological cascade. To our knowledge, this is the first study aiming to find whether the expression of CTG repeats in the DM1 mouse model is associated with bacterial gut dysbiosis, as we recently described in human DM1 patients [30]. Here, we sequenced the 16S rRNA gene of stool samples from DMSXL mice carrying a mutated human DMPK transgene containing > 1000 CTG repeats [31]. DNA sequencing data revealed significant differences in bacterial taxa abundance between DMSXL mice and Wild Type (WT) groups. These findings indicated a probable association between gut microbiome structure and DM1 status, offering new perspectives into the microbiome’s role in GI-related symptoms of DM1 patients.

2. Methods

2.1. Mice

In this study, we used the DMSXL mice carrying a 45 kb expanded human DMPK DNA fragment containing between 1000 and 1600 CTG repeats. These mice exhibit systemic abnormalities, such as growth retardation (30–50 % reduction in body weight in the first few months of life) due to the expression of mutant DMPK transcripts, as described previously [31]. A total of 68 mice samples were used in the study. Among them, 49 samples were from DMSXL mice (on C57BL/6 background) including 26 Homozygous and 23 Heterozygous, and 19 samples were from wild type C57BL/6 littermates. Mice were housed on a 12:12-hour light: dark cycle, with a controlled temperature of 21°C and a humidity level of 40 % in pathogen-free cages within the Animal Holding Facility at the Centre Hospitalier Universitaire CHU Sainte-Justine in Montréal, Canada. Fresh fecal samples were collected from adult mice with average ages of 9–10 weeks old (wild-type: 9.0, Heterozygous: 10.8, and Homozygous; 9.6 weeks old). Each mouse was isolated in an empty sterile cage until they defecated. The feces were then collected using sterile forceps, carefully placed in sterile Eppendorf tubes, and promptly stored at a temperature of −20°C [32], [33]. Long-term storage of the samples was done at −80°C until DNA isolation. All animal experiments received approval from the CHU Sainte-Justine Research Ethics Committee and were conducted in accordance with the regulations established by the Comité Institutionnel des Bonnes Pratiques Animales en Recherche (CIBPAR), with approval number 2022–3608. These experiments were carried out in strict compliance with the guidelines set forth by the Canadian Council on Animal Care.

2.2. 16S rRNA gene library preparation

Fecal DNA was isolated by using a QIAamp DNA stool mini kit (QIAGEN). Genomic DNA was amplified with primers 515 F (5′-GTGCCAGCMGCCGCGGTAA-3′) and 806 R (5′-GGACTACHVGGGTWTCTAAT-3′) targeting the V4 regions of microbial small subunit ribosomal RNA genes generated using PCR amplification protocol as described previously [34], [35], [36]. PCR amplifications were performed in 10 μl reactions in 96-well plates, using high-fidelity 2 × master mix (NEBNext). PCR conditions were 98°C for 30′′, followed by 28 cycles of 98°C for 10′′, 65°C for 15′′ and 72°C for 10′′. Lastly, the pooled and indexed libraries were sequenced in paired-end modus on an Illumina MiSeq [37] in RNomics Platform of the University of Sherbrooke (https://rnomics.med.usherbrooke.ca/). The raw sequencing data were presented at SRA database of NCBI GenBank database (BioProject ID: PRJNA1182323, BioSample: SAMN44591550)”.

2.3. Microbiome analysis

To investigate the relationship between the mice DM1 status and the gut microbiome changes, we used several methods including analyses of alpha diversity and beta diversity, as well as ANCOM-BC differential abundance analyses. Alpha diversity analyses provide an assessment for overall microbiome diversity while beta diversity analyses measure microbiome profile similarity or dissimilarity of the compared communities in each group. Conversely, ANCOM-BC analyses test for the differential abundance of bacterial taxa between experimental groups of mice.

Microbiome analyses were performed in four comparison groups: 1- WT vs. Homozygous mice, 2- WT vs. Heterozygous mice, 3- WT vs. DM1 mice (including both Homozygous and Heterozygous mice), and 4- Homozygous vs. Heterozygous. In each comparison set, the quality of sequences was checked, then trimmed and merged, and chimeras were removed using the DADA2 v1.22.0 pipelines [38] for 16S library and phyloseq v.1.38.00 [39] as packages for R (R Development Core Team; http://www.R-project.org) in RStudio v.4.1.3 [40]. SILVA v.138 bacterial databases [41] was employed to assign taxonomy to the bacterial community matrices from the resulting unique Amplicon Sequence Variant (ASV). All steps in the study were performed according to the workflow tutorials at https://benjjneb.github.io/dada2/tutorial.html.

Alpha diversity was assessed using the Chao1 [42], Shannon [43], and Simpson [44] indices with statistical significance determined by non-parametric Wilcoxon rank sum exact tests. Using multiples alpha diversity indices provides complementary information on the changes more specific to species richness and evenness in the community. For beta diversity, Bray–Curtis dissimilarity, as well as weighted and unweighted Unifrac indices were calculated, visualized with non-metric multidimensional scaling Bray Curtis (NMDS), and used to identify the key drivers by Permutational Multivariate Analysis of Variance (PERMANOVA) [45]. The UniFrac distance was determined based on a phylogenetic tree constructed from the raw sequencing data. The phylogenetic tree was made using the phyloseq, DECIPHER, phangorn, and ggplot2 packages in R and was based on raw sequence data. This tree was then used to calculate the weighted and unweighted UniFrac distance matrixes to calculate beta-diversity among the groups. The weighted UniFrac distance considers the relative abundance of taxa and impacts the branch length with abundance variation. By contrast, an unweighted UniFrac distance uses only the presence and absence of species and counts the fraction of branch length distinctive to each community. PERMDISP, a resemblance-based permutation test, was also used for comparisons testing the null hypothesis of homogeneity of multivariate dispersions. This test can detect if the dispersion of the group data from the centroids contributes to the significant difference detected by the PERMANOVA. Bacterial taxonomic levels that showed differential abundance between DM1 status were determined using ANCOM-BC (Analysis of Communities of the Microbiome) [46]. A false discovery rate (FDR) approach was used to correct for multiple testing. It is important to note that ANCOM-BC identifies taxa that are differentially abundant but does not provide directionality at the phylum level, particularly when relative abundances are very small. Therefore, significant changes in the relative abundance of phyla, such as Patescibacteria, were inferred based on statistical significance, rather than fold changes or effect sizes. This limitation does not apply to the genus-level analysis, where directionality was more directly assessed through the statistical output of the analysis.

2.4. qPCR

Intestinal samples (duodenum or jejunum) were collected at 13–14 weeks of age from WT and DMSXL mice. The dissected intestines were immediately frozen in liquid nitrogen and stored at −80°C. Total RNA was extracted from 20 to 30 mg of intestinal samples with 1 ml of TRIzol (ThermoFisher, 15596018) and RNeasy Mini Kit (Qiagen, 74104). The purity and concentration of RNA were determined with a Thermo Scientific™ NanoDrop™ 8000 Spectrophotometer. Then, 2 μg of RNA was used for cDNA synthesis by Reverse Transcriptase kit (Abm, G592) according to the manufacturer’s instruction. Gene amplification for quantitative PCR was performed with 50 ng of synthesized cDNA, 1 μM of Forward and Reverse primers, and the BlasTaq™ 2X qPCR MasterMix (Abm, G892) on a Roche LightCycler® 480 Instrument II. Primers are indicated in Table 1. Each sample was run in triplicate and the average cycle threshold (CT) value was used for data analysis. Each sample was normalized to the expression level of housekeeping gene, Hprt, and relative gene expression was determined according to the 2^−ΔΔCt equation against the value of their wild-type group.

Table 1.

The sequences of primers used for qPCR are listed below.

Gut inflammation marker	Used Primers	Sequence of primers
Hprt	Forward	CAGTCCCAGCGTCGTGATTA
	Reverse	GGCCTCCCATCTCCTTCATG
Ccl20	Forward	ACTACGACTGTTGCCTCTCG
	Reverse	TTCTTGACTCTTAGGCTGAG
Cxcl1	Forward	GCCACACTCAAGAATGGTCG
	Reverse	TACTTGGGGACACCTTTTAG
Tgfb1	Forward	CGCAACAACGCCATCTATGA
	Reverse	TTCCGTCTCCTTGGTTCAGC
Il1b	Forward	TGCCACCTTTTGACAGTGAT
	Reverse	TGATGTGCTGCTGCGAGATT

3. Results

3.1. Alpha and beta diversity analyses

We performed alpha diversity analyses using three metrics: Chao1, Shannon, and Simpson for all the comparison sets. No significant differences were observed (P ≥ 0.05) for all the comparisons indicating that the overall diversity and richness of the gut microbiome in mice. This suggests that, despite potential links between DM1 and the gut microbiota, the CTG repeat expansion alone does not appear to alter the overall microbial composition. Alpha diversity analyses did not show statistically significant differences in overall diversity and richness of the gut microbiome across the comparison groups. While alpha diversity of Heterozygous mice appeared to shift towards higher Shannon diversity, this trend was not statistically significant. Thus, we cannot conclude that these mice harbor a more diverse or evenly distributed gut microbiome. As for beta diversity analyses, which assesses the divergence in microbial community composition, small significant differences were observed between groups, except for the comparison group WT vs Heterozygous (P ≥ 0.05). (Suppl. Fig. 1). Notably, significant differences were observed between the WT and Homozygous groups, as demonstrated by Bray-Curtis (P = 0.03, R²=0.043) and Unifrac (P = 0.003, R²= 0.046) (Fig. 2). These findings indicate that Homozygous mice exhibit a distinct gut microbiota composition compared to WT mice.

Fig. 1.

Alpha-diversity indices (Shannon, Simpson, and Chao1) for comparing gut microbiome of Homozygous (n = 26) and Heterozygous DMSXL mice (n = 23). In blue, Heterozygous mice and in red, Homozygous mice (P ≥ 0.05).

Fig. 2.

Beta-diversity results between Wild Type (WT, n = 19) and Homozygous DMSXL mice (n = 26). (a) Bray-Curtis (P ≤ 0.05, R²=0.043) (b) PCoA based on weighted UniFrac (P ≤ 0.05, R²= 0.066) and (c) unweighted UniFrac (P ≤ 0.05, R²= 0.046). In red, Homozygous mice and in blue, WT.

Moreover, small significant differences were also detected in the WT vs DMSXL (Homozygous + Heterozygous) group, but only with the Unifrac (P = 0.049, R²= 0.032) and not with the Bray-Curtis outcome (P = 0.07, R²= 0.02) (Fig. 3). Homozygous vs Heterozygous analysis revealed a significant difference for the Bray-Curtis (P = 0.04, R²= 0.040) and the unweighted Unifrac (P = 0.007, R²= 0.040), but not with the weighted Unifrac P = 0.226, R²= 0.040) (Fig. 4). Fig. 5 shows the comparison of beta diversity for all WT, Homozygous, and Heterozygous mice. This figure illustrates the significant difference in beta diversity among the groups (P = 0.01, R²= 0.054). Specifically, the bacterial community structures of WT, Homozygous, and Heterozygous mice appear distinct from one another, suggesting that genotype plays an important role in shaping the gut microbiota. The overall separation observed in Fig. 5 indicates that each genotype exhibits a unique bacterial profile, and these differences are statistically significant when comparing the groups. Previously, we conducted pairwise comparisons between specific genotypes, including WT vs Homozygous, Homo vs Heterozygous, WT vs Heterozygous, and WT vs both Homozygous and Heterozygous. In Fig. 5, we made a comparison between WT, Homozygous, and Heterozygous genotypes collectively. A summary of the results for both alpha and beta diversity analysis in all four-comparison series can be found in Table 2. Considering that GI symptoms could be different between males and females in DM1, further analyses based on biological sex were performed in Homozygous mice. No significant differences were found for both alpha and beta diversity analyses between male and female groups in Homozygous mice (Supplementary Fig 2). The PERMDISP analysis specified no significant dispersion between all comparison groups (P ≥ 0.05) suggesting the PERMANOVA results are not affected by differences in dispersion across phenotypes. This indicates that the reported changes in bacterial community composition are not simply resulting from variations in the dispersion of data points but signifying actual differences in the gut bacterial structure between DMSXL and WT mice.

Fig. 3.

Beta-diversity results between Wild Type (WT, n = 19) and all DMSXL mice (n = 49) including both Homozygous and Heterozygous mice. (a) Bray-Curtis (P ≥ 0.05) (b) PCoA based on weighted UniFrac (P ≤ 0.05, R²= 0.032) and (c) unweighted UniFrac (P ≤ 0.05, R²= 0.022). In red, DM1 mice and in blue, WT.

Fig. 4.

Beta-diversity results between Homozygous (n = 26) and Heterozygous DMSXL mice (n = 23). (a) Bray-Curtis (P ≤ 0.05, R²= 0.040). (b) PCoA based on weighted UniFrac (P ≥ 0.05, R²= 0.040).and (c) unweighted UniFrac (P ≤ 0.05, R²= 0.040). In red, Heterozygous DMSXL mice and in blue, Homozygous mice.

Fig. 5.

Beta-diversity results for WT (n = 19), Homozygous (n = 26) and Heterozygous DMSXL mice (n = 23). Bray-Curtis (P ≤ 0.01, R²= 0.054). In blue, Heterozygous DMSXL mice, in red, Homozygous mice and in green, WT.

Table 2.

The full details of results for alpha and beta diversity analysis in all comparison groups of mice.

Comparison groups	WT/Homo	WT/Hetero	WT/DM11	Homo/Hetero
Numbers per group	19/26	19/23	19/49	26/23
Alpha diversity	NS	NS	NS	NS
Beta diversity	S P = 0.03, R2= 0.0431	NS	NS	S P = 0.02 R2= 0.04032
Bray-Curtis
Weighted Unifrac	S P = 0.01 R2= 0.066	NS	S P = 0.049 R2= 0.03233	NS
Unweighted Unifrac	S P = 0.003 R2= 0.047	NS	S P = 0.041 R2= 0.02291	S P = 0.007 R2= 0.04009
PERMDISP	NS P = 0.17	-	NS P = 0.07	NS P = 0.58

*P value more than 0.05 is non-Significant (NS) and lower than 0.05 is Significant (S).

1- Both Homozygous and Heterozygous mice.

3.2. Differential abundance analysis

Taxa with differential abundance in the comparison groups of mice between DMSXL and WT mice were characterized by ANCOM-BC, a methodology for differential abundance and correlation analyses in microbiome data based on the compositional log-ratios. Our results showed that, at the phylum level, Patescibacteria exhibited significant changes across three of the four comparison groups: DMSXL versus WT, Homozygous versus WT, and Homozygous versus Heterozygous mice. No significant differences were observed between WT and Heterozygous mice. This indicates that Patescibacteria has a notable shift in abundance in each of these comparisons. Further analyses comparing the microbiomes of WT and DMSXL mice including both Homozygous vs Heterozygous mice (Fig. 6A) and WT vs Homozygous mice (Fig. 6B) revealed that the two phyla of Patescibacteria (P = <0.01, RA: 1.5 %, LFC= - 0.05) and Defferibacterota (P = 0.01, RA: 1.25 %, LFC= - 0.2) were significantly altered in Homozygous mice. Moreover, the comparison between the WT and Homozygous groups also showed a trend in the differential abundance of Cyanobacteria (P = 0.05, RA: 0.48 %, LFC=0.75), Verrocumicrobiota (P = 0.06, RA: 0.05 %, LFC=1.3) and Firmicutes (P = 0.06, RA: 41 %, LFC=0.5). Finally, comparing the gut microbiomes of Homozygous mice vs Heterozygous mice, our results showed a significant difference in the detection of the phyla Patescibacteria (P = <0.01, RA:1.1 %, LFC=0.2) and Verrocumicrobiota (P = 0.02, RA:1.5 %, LFC=0.3) in Homozygous mice (Fig. 6C). These findings emphasize the distinct microbial profiles associated with DM1 genotype variations.

Fig. 6.

ANCOM-BC analysis in a) WT (n = 19) and DM1 mice (n = 49), b) WT (n = 19) vs Homozygous mice (n = 26) and c) Heterozygous (n = 23) vs Homozygous mice (n = 26). Significant taxa at the phylum level with their p value as well as the most significantly differed genera with higher and lower abundance of each group has been shown. The ANCOM analyses was performed at a P = 0.05 with Benjamini–Hochberg FDR correction.

At the genus level, the comparison between WT and DMSXL groups revealed statistically significant differences in the relative abundance of specific gut bacterial taxa in WT mice. Notably, Eubacterium nodatum (P = 0.005) was significantly more abundant in WT mice, along with Streptococcus (P = 0.002) and Lachnospiraceae (P = 0.001) (Fig. 6A). Alternatively, in the WT vs DMSXL mice, the highest level of statistical significance was observed for Candidatus Saccharimonas (P = 0.001), followed by Eubacterium brachy (P = 0.002). The comparison of WT vs Homozygous groups showed that the relative abundance of Eubacterium nodatum (P = <0.01) is increased in WT mice compared to Homozygous mice (Fig. 6B). On the other hand, the relative abundance of Lachnospiraceae NK4B4 (P = 0.0002), Christenseenellaceae R-7 (P = <0.01) and Romboutsia (P = <0.01) was higher in Homozygous mice compared to WT mice (Fig. 6B). Lastly, in the Homozygous vs Heterozygous mice groups (Fig. 6C), our findings displayed that Eubacterium nodatum (P = <0.01) abundance declined in the Homozygous mice, similar to what was observed in the two previous comparison groups. Lactobacillus HT002 (P = <0.01) abundance was also reduced in the Homozygous mice compared to Heterozygous mice. Finally, the relative abundance of Romboutsia (P = <0.01) as well as Christenseenellaceae R-7 (P = <0.01) was elevated in Homozygous mice, similar to what was observed in the comparison between the WT vs Homozygous mice.

To determine if changes in microbiome observed in the DM1 mice may be associated with the inflammatory response, the expression levels of various markers associated with gut inflammation (CCL20, CXCL1, TGFβ, and IL-1β) were assessed in intestine samples using qPCR [47], [48], [49], [50]. Our findings indicate that CCL20 levels are upregulated in DM1 mice (P = 0.03) (Fig. 7). These findings suggest possible alterations in the microbiota and gut inflammation in Homozygous DMSXL mice, a model for DM1 patients. This relationship highlights an opportunity to utilize the DMSXL mouse model to gain deeper insights into the complex interaction between gut dysbiosis and inflammatory processes in DM1.

Fig. 7.

qPCR analysis of inflammatory markers in intestine samples of WT (n = 8–9) and DM1 Homozygous mice (n = 11–13). Mann Whitney test with false discovery rate correction. *p < 0.05, Mean±SEM.

4. Discussion

The principal finding reported here is that the structure of the murine gut microbiome is significantly associated with the DM1 status. While no significant differences in both alpha and beta diversity analysis for the gut microbiome of Heterozygous mice compared to WT, Homozygous mice showed significant alterations of their gut microbiome composition. Beta diversity analysis revealed significant alterations in the microbial community of Homoxygous mice when compared with WT, with specific genera like Eubacterium nodatum, Romboustia and Lachnospirace sp showing distinct differences. Moreover, significant changes were observed in the gut microbiota when comparing Homozygous mice with Heterozygous mice. These changes included differences in beta diversity as well as notable alterations in the abundance of specific phyla, namely Patescibacteria and Verrucomicrobiota (Figs. 4, 6C- Table 2). This suggests that both alleles of mutated DMPK gene with CTG trinucleotides expansion are necessary to induce measurable changes in the gut microbiome composition of DM1 mice. Importantly, sex did not influence the gut microbiome structure in Homozygous DMSXL mice (Supplementary Fig 2). Therefore, Homozygous DMSXL mice emerges as a more suitable model for studying GI symptoms associated with DM1. This is in agreement with other studies showing that Heterozygous DMSXL mice do not display an obvious phenotype [31], [51].

The association of gut microbiome with DM1 status was observed exclusively in beta diversity analysis, highlighting differences across multiple bacterial lineages. These findings underscore the importance of examining beta diversity as a more sensitive indicator of gut microbiota shifts in DM1 models. ANCOM-BC analyses identified certain taxa that were associated with the genotype. Generally, these findings are in accordance with our earlier report that GI problems of human DM1 patients are associated with changes in their gut microbiota [30]. In this previous study, the relative abundance of Bacteroidota, Euryarchaeota, Fusobacteriota, and Cyanobacteria phyla in human DM1 samples differed from that in healthy controls, without significant alterations across DM1 phenotypes [30]. As for mice samples the results were different in each group of comparison, but Patescibacteria phylum was the most common phylum of bacteria with significant changes in Homozygous mice. At the genus level, Eubacterium nodatum, Streptococcus, and Lachnospiraceae found with a decrease in their relative abundance of Homozygous mice. Alternatively, Lachnospiraceae NK4B4, Christenseenellaceae R-7 and Romboutsia were reported with higher levels than WT. There are several factors implicated in microbial groups differences between human DM1 samples and the DM1 mouse model. Variations at species-specific level in microbiome composition between humans and mice can be a cause for differential responses in DM1 status. Besides, specific environmental and housing conditions related to the animal models may contribute to distinct microbial profiles, which then impact their responses to DM1-related alterations.

The genotype was not associated significantly with alpha diversity analyses as shown by the Shannon, Simpson, and Chao1 metrics in all comparison groups, indicating that the overall diversity of gut microbiome in DMSXL mice is not significantly different from WT mice. Alpha diversity is a measure of the total amount of different species and the evenness of these numbers through species in a sample. The lack of significant association between the genotype and alpha diversity of mice gut microbiome is consistent with prior studies having reported no correlation with the trinucleotide repeat expansion (measured in blood) and GI symptoms [52].

The main finding of this study is that the genotype was significantly associated with gut microbiota modifications in DMSXL mice as demonstrated through beta diversity analyses. Beta diversity measures the similarity or dissimilarity between two communities based on their microbial composition, highlighting structural differences that arise due to genetic factors. In fact, the gut microbiome was significantly affected in Homozygous DMSXL mice in terms of both relative abundance of bacteria and the abundance of various phylogenic lineage of taxa. These findings support the results of our previous study on DM1 human samples [30].

Another important finding is the identification of significant differences in the relative abundance of gut microbiota in DMSXL mice using ANCOM-BC differential abundance analysis. While ANCOM-BC does not provide direct fold-change values, our results indicate that the Patescibacteria phylum was significantly different in DM1 compared to controls. Interestingly, Ravegnini et al. also reported gastric cancer patients were significantly enriched in this phylum [53]. This similarity suggests a potential shared mechanisms between DM1 and gastric cancer. For instance, immune dysregulation or gut barrier dysfunction might contribute to microbiota alterations in both conditions. Further studies are needed to explore this possible relation. Furthermore, the increase in the relative abundance of Lachnospiraceae in Homozygous DMSXL mice is similar to gut microbiota changes in the gastric cancer patients [54], [55]. In microbiome analysis of DMSXL mice, we observed a reduction in the relative abundance of genera Lactobacillus HT002 and Streptococcus. Lactic acid producing bacteria from the genera Lactobacillus and Bifidobacteria as well as Streptococcus thermophilus have demonstrated beneficial health effects in humans [56]. The decline in these beneficial bacteria in human gut microbiota can cause diarrhea, bloating, and cramps for Lactobacillus [57], and colitis for Streptococcus [58]. Additional genera, Christenseenellaceae R-7, Romboutsia, and the Verrocumicrobiota phylum were also identified as significantly different in the DMSXL mouse model. Other studies also indicated that high-fat diet increased the relative abundance of the Romboutsia [59], [60], [61]. It was significantly associated with triglyceride and free fatty acids in the high fat diet group of rats [62]. Therdtatha et al., reported the gut microbiome of Indonesian obese people associated with obesity and type 2 diabetes, in which Romboutsia is abnormally enhanced in relation to fat intake [63]. Romboutsia is recognized as an obesity-related genus that positively correlates with BMI [64], as well as lipid profiles and lipogenesis in the liver [65]. Altogether, increase in the relative abundance of Romboustia can lead to obesity and type 2 diabetes. Noteworthy, insulin resistance and diabetes mellitus are more frequently associated with DM1 [66]. In addition, lower abundance of Eubacterium nodatum in DMSXL mice is in agreement with Baima et al., which has showed changes in the gut microbiome of Inflammatory Bowel Diseases (IBD) patients compared to healthy controls [67]. Other alterations of gut microbiome in DMSXL mice were characterized by increased abundance of the bacterial Deferribacterota phylum which has been shown to be positively associated with proinflammatory chemokines [68]. Noteworthy, our findings indicate that the pro-inflammatory chemokine CCL20 is upregulated in DMSXL Homozygous mice. This chemokine can attract cells expressing its receptor, CCR6, including immature dendritic cells, regulatory T cells, T helper 17 cells, and B cells [69]. Many studies have implicated the participation of CCL20 in the pathogenesis of IBD [48], [70], [71], and showed that its expression is upregulated in Crohn’s disease and ulcerative colitis [72]. Furthermore, it has been demonstrated that CCL20 possesses antimicrobial effects against various bacteria, including Streptococcus [73]. Notably, Streptococcus levels are reduced in DMSXL mice. These findings suggest a potential link between the microbiota and gut inflammation in the DMSXL model, which may provide insights into similar mechanisms in DM1 that will need to be validated functionally in future studies. Similarly, the abundance of Deferribacteres, a core member of murine gut microbiota, has been shown to be reduced in mdx mice, a model of Duchenne Muscular Dystrophy, and it was associated with an increase in the systemic inflammatory markers [17]. It has also been detected in association with intestinal inflammation in colitis models [74]. Ulcerative colitis also causes inflammation and ulcer sores in gastrointestinal tract [75]. These detrimental changes in the gut microbiota suggest that DMSXL mice might be a good model to study therapeutic approaches for GI symptoms in DM1.

Treatment targeting the gut-muscle axis has shown promising therapeutic potential in other muscular dystrophy models. Fecal microbiota transplant of eubiotic microbiota from healthy WT mice into dystrophic mdx mice was shown to reduce inflammation and partially restore muscle function [18]. Similarly, supplementation with sodium butyrate to increase the growth of bacteria producing short-chain fatty acids was shown to dampen inflammatory markers and increase physical function in dystrophic mdx mice [76]. The current findings open the way to studies investigating the therapeutic potential of drugs targeting the microbiota for the treatment of the different symptoms in DM1.

Although, the present findings are novel and useful to understanding the potential roles of the gut microbiome in DM1, a limitation of this study is that our findings were not extended to a functional analysis. Another limitation of the study is the limits of 16S rRNA gene sequencing providing data on live, dormant, and inactive microbes. this approach cannot detect and verify which microbial species are dead or active. Extension of these findings to functional analysis of the significantly associated taxa would be necessary. Further metabolomic studies would provide information regarding differences in the gut microbiome as the consequences of CTG repeats. Another limitation was that we did not monitor stool characteristics, defecation frequency, or toxin levels produced by bacteria in WT and DMSXL mice. These parameters indeed could provide valuable insights into how well the DMSXL model reflects DM1-related GI issues, as well as other additional metrics for assessing therapeutic efficacy in future studies. Furthermore, in our current study, we utilized 16S rRNA sequencing for microbiome analysis and qPCR specifically for assessing gut inflammatory markers while our focus was on the associations in the context of DM1. Future research incorporating a time-course design like Mossad et al., highlighting the value of examining microbiota diversity over time [77] could provide valuable insights into the dynamics of microbiota changes and their relationship with GI symptoms. In addition, implementing longitudinal sampling and monitoring a broader range of inflammatory markers would allow for a more comprehensive characterization of the inflammatory responses.

5. Conclusion

This study provides evidence of gut microbiota dysbiosis in DMSXL mice, which resembles the changes observed in human affected by DM1. Our findings lay the foundation for future multi-omics research and functional investigations. Furthermore, these findings shed light on the roles of gut microbiota in DM1 to improve our understanding of DM1 pathogenesis and the development of more accurate diagnostic methods. Lastly, this may help in developing novel therapeutic strategies targeting significant taxa in the gut microbiota of DM1 patients.

Funding

This work was supported by Canadian Institutes of Health Research [MOP389354].

Author statement

The authors have stated that nothing exists to disclose.

CRediT authorship contribution statement

Kim Tae-Yeon: Methodology, Investigation. Massé Éric: Writing – review & editing, Supervision, Resources, Project administration, Investigation, Funding acquisition. Mahdavi Mazdeh Manijeh: Writing – original draft, Software, Methodology, Investigation. A. Dumont Nicolas: Writing – review & editing, Resources, Project administration, Investigation. Laforest-Lapointe Isabelle: Writing – review & editing, Software, Project administration, Data curation. Fissette-Paul Hus Isabelle: Resources, Investigation. Gagné-Ouellet Valérie: Resources, Investigation. Balthazar Philippe: Software, Methodology. Prévost Karine: Methodology, Investigation. Gagnon Cynthia: Writing – review & editing, Supervision, Resources, Project administration, Investigation. Harvey Séréna: Investigation, Conceptualization, Data curation, Writing – review & editing. Duchesne Élise: Resources, Investigation.

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Appendix A. Supplementary material

Supplementary Fig 1

Supplementary material