Familial clustering of dysbiotic oral and fecal microbiomes in juvenile dermatomyositis

doi:10.21203/rs.3.rs-1620914/v1

Download PDF

Research Article

Familial clustering of dysbiotic oral and fecal microbiomes in juvenile dermatomyositis

https://doi.org/10.21203/rs.3.rs-1620914/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted

You are reading this latest preprint version

Background: Juvenile dermatomyositis (JDM) is a rare immune-mediated disease of childhood that is thought to result from genetic predisposition and environmental drivers, with documented links to microbial exposures. In this multi-center, prospective, observational cohort study, we evaluated whether JDM is associated with discrete oral and gut microbiome signatures.

Methods: We generated 16S rRNA sequencing data from fecal, saliva, supragingival, and subgingival plaque samples from JDM probands (n=28, age range 3-18 years, mean age 10 years, 46% female). To control for genetic and environmental determinants of microbiome community structure to the greatest extent possible, we also profiled microbiomes of their unaffected family members (n=27 siblings, n=26 mothers, and n=17 fathers). We performed paired within-family comparisons as well as unpaired analyses of different cohorts.

Results: Sample type (oral vs fecal) and nuclear family unit were the two most predominant variables explaining variance in microbiome diversity, more so than having a diagnosis of JDM. The oral and gut microbiomes of JDM probands were more similar to their own unaffected siblings than they were to the microbiomes of other JDM probands. Nonetheless, in a sibling-paired analysis, several potentially immunomodulatory bacterial taxa were differentially abundant in the microbiomes of JDM probands compared to their unaffected siblings, including Faecalibacterium in the gut and Streptococcus in the oral cavity. Further, compared to published microbiome datasets generated from healthy children and adults, fecal microbiomes of both JDM probands and their unaffected family members were markedly different.

Conclusions: Family unit (which reflects the confluence of genetic and household effects) has a significant effect on microbiome community structure. Features of the dysbioses seen in oral and fecal microbiomes of JDM probands were shared by their unaffected family members, suggesting familial microbiome signatures. The loss or gain of specific fecal and oral bacteria may potentially play a role in disease pathogenesis and/or be secondary to immune dysfunction in susceptible individuals.

juvenile dermatomyositis

gut microbiome

oral microbiome

pediatrics

Juvenile dermatomyositis (JDM) is a rare disease with increased morbidity and mortality that is characterized by stereotypical rashes and proximal muscle weakness, with an estimated incidence of 3.2 per million children in the United States [1]. While JDM primarily affects skin and muscle, it can also carry significant morbidity and mortality due to calcinosis and cardiac, respiratory, and gastrointestinal organ system involvement [2,3]. The etiology of JDM is thought to reflect the confluence of genetic predisposition, environmental exposures, and dysregulated immunity [4]. Immune-mediated diseases such as rheumatoid arthritis (RA), psoriasis, lupus, inflammatory bowel disease (IBD), and type 1 diabetes have been linked to alterations in fecal and oral microbiota [5–8]. Gingival inflammation is observed in 10-45% of JDM patients [9], suggesting that plaque-induced gingival inflammation from oral bacteria may contribute to a systemic inflammatory response in JDM. However, there have been no studies investigating the role of the oral or gut microbiome in JDM patients. Identification of bacterial communities contributing to chronic inflammation in JDM could lead to an improved understanding of pathogenesis and offer new approaches for monitoring and treatment.

Disease cohort studies of the microbiome are statistically likely to identify differences based on the combination of multiple hypothesis testing and tremendous interpersonal microbiome differences [10]. In order to adjust for interpersonal differences and control for genetic predisposition to some extent, we recruited children with JDM and their unaffected family members (siblings and parents) in this unique multicenter, prospective, observational cohort study. We reasoned that by controlling for shared family attributes, we would be able to identify microbiome signatures of JDM more clearly. This study design permitted (i) paired statistical analyses in which each JDM proband had familial controls and (ii) disease versus healthy cohort-based analyses.

Human subjects. This prospective observational study was conducted in 2018-2019. Patients and their families were recruited from the Seattle Children’s Hospital (SCH) Rheumatology Clinic (protocol ID 1044) and the National Institutes of Health (NIH) Clinical Center (as part of the NIEHS Twins and Siblings Discordant for Systemic Rheumatic Diseases study; protocol ID 03-E-0099) using institutional review board-approved research protocols for informed consent, medical record review, clinical sample collection, and use of samples. JDM probands were children 2-18 years of age who met probable or definite criteria for JDM (as defined by Bohan and Peter criteria for myositis [11,12]). Siblings in both cohorts had no underlying autoimmune disease and had an age difference of ≤5 years, except for 2 siblings from the SCH cohort (JDM0019R1 and JDM0020R1) who were within 10 years of age. Exclusion criteria for JDM probands were other co-existing immune-mediated diseases; for siblings, any autoimmune or immune-mediated disease diagnoses; and for all participants, antibiotics received within the past 3 months. A total of 28 children with JDM (probands), 27 unaffected siblings (representing 26 families; in one family, a JDM proband had 2 unaffected siblings), 26 mothers, and 17 fathers were enrolled, resulting in a collection of 33 fecal, 95 saliva, 94 supragingival swab, 19 anterior subgingival dental plaque, and 19 posterior subgingival dental plaque samples. Fecal samples were collected using DNA Genotek OMNIgene Gut (OMR-200). Oral samples were collected in two ways: (i) anterior and posterior subgingival dental plaque samples were collected with sterilized Gracey currettes from 3 anterior teeth and 3 posterior teeth, and (ii) supragingival swabs were collected with OMNIgene Oral (OM-505) kits. All samples were placed in the same buffer for stabilization of nucleic acids and stored at -80°C until use. Medication use and laboratory parameters of myositis disease activity were extracted from patient charts (SCH) or prospectively collected at study visits (NIH). Myositis disease activity was assessed using two validated metrics: Manual Muscle Testing 8 (MMT-8), with bilateral testing of the muscle groups (score 0- 150) and the Cutaneous Dermatomyositis Disease Area and Severity Index (CDASI) [13,14].

Genomic DNA extraction. Samples (250 µL per stool sample, 1 mL per saliva sample, or 1 mL per plaque sample) was homogenized at 25-30 Hz for 10 minutes using a Qiagen TissueLyser II. Genomic DNA was then extracted and purified using either the Qiagen PowerFecal Pro DNA Kit (fecal samples; as per manufacturer’s instructions) or the Qiagen QIAamp DNA Microbiome Kit (saliva and plaque samples; host DNA depletion step skipped to minimize concomitant loss of bacterial DNA). DNA was quantified using a Qubit fluorometer.

16S rRNA sequencing and analysis. Hypervariable region 4 (V4) of 16S ribosomal RNA genes present in samples was PCR amplified using 515F (5′-GTGCCAGCMGCCGCGGTAA-3′) and 806R (5′-GGACTACNVGGGTWTCTAAT-3′) primers using Taq high-fidelity polymerase (Invitrogen). Sample-specific barcoded adapters were used to prepare and pool sequencing libraries as previously described [15–17]. To minimize batch effects and bias, we (i) employed pseudo-randomization during library preparation, (ii) minimized the number of sequencing runs (2 total runs: all stool samples in 1 run and all oral samples in the other run), and (iii) used an automated liquid handler for multiple steps during library preps. Researchers generating the sequencing data were blinded to clinical metadata associated with samples. Multiplexed sequencing was performed on the Illumina MiniSeq platform (150 nucleotide paired-end reads; 20% PhiX). 12 samples (8 saliva from 4 patients, 3 siblings, and 1 father; 4 supragingival swabs from 2 patients and 2 siblings) were excluded due to low numbers of remaining high-quality reads, likely attributable to insufficient amounts of bacterial DNA captured in the original sample. Quality filtering, merging of mate pairs, and determination of amplicon sequence variants (ASVs) [18] was performed using DADA2 [19] via QIIME2 [20]. Sequencing depth of retained samples was 131,263 ± 7,141 (mean ± SEM) high-quality reads per sample. ASVs with a relative abundance <0.001% and those present in only 1 sample throughout the entire dataset were removed prior to further analysis. Taxonomic classifications were performed in QIIME2 using a classifier trained on 515F-806R sequences from the Silva database [20]. Selected sequences classified as “unannotated bacteria” were subjected to an NCBI BLAST search against the nr database using a sequence identity threshold of >98%. Diversity analyses were performed using QIIME2 and the R package phyloseq (version 1.30.0). Alpha diversity calculations were performed after rarefaction to 6,100 reads per sample. Beta diversity calculations were made using UniFrac, a phylogenetic metric that calculates the proportion of a phylogenetic tree shared by two samples [21], both with and without rarefaction of sequencing data; the latter is recommended by Holmes, Willis, and colleagues [22,23] and is reported here.

Myositis autoantibodies. Sera were tested for myositis-specific (MSA) and myositis-associated autoantibodies (MAA) using validated immunoprecipitation (IP) and IP-immunoblotting methods at the Oklahoma Medical Research Foundation laboratory [24].

Statistics. Statistical analysis was performed in R (version 4.0.3) using the specific tests named in the text. To identify significantly differentially abundant bacterial taxa while adjusting for family as a covariate of interest, we used corncob, an R package that leverages a beta-binomial regression model and uses the Wald test to calculate significance [25]. In our analysis, we also adjusted for sequencing depth, a potential source of measurement noise [26]. The effects of covariates on UniFrac distances were calculated using permutational multivariate analysis of variance (PERMANOVA). In follow-up analyses, specific comparisons were performed using a Student’s t-test. A p-value of less than 0.05 was considered statistically significant. Plots were generated in R using ggplot2 (version 3.3.5) and phyloseq packages (version 1.30.0) and then assembled in Adobe Illustrator (2022 version).

Altered representation of immunomodulatory taxa in JDM probands

We enrolled 28 children with JDM (mean ± SEM 10.0 ± 0.7 years of age, 46% female) and unaffected family members as controls (27 healthy siblings, 26 mothers, and 17 fathers; Table 1) across two study sites (SCH, NIH). The JDM probands had mostly well-controlled disease activity and relevant disease data are depicted in Table 1. MMT-8 scores were available in 25 probands (89%) and ranged from 91-150 (maximum 150). Based on these scores, 3 of 28 (11%) probands had abnormal muscle strength, most of which were in a range indicative of mild weakness (score greater than 135). Seven out of 28 probands had mild cutaneous disease activity (median score 2). Antinuclear antibody (ANA) was positive in 70% of JDM probands. Sixteen JDM probands were receiving one or more immunosuppressive medications, including methotrexate (n=12), daily oral corticosteroids (n=8), hydroxychloroquine (n=7), intravenous immunoglobulin (IVIG) (n=5), mycophenolate mofetil (n=3), abatacept (n=2), and/or tacrolimus (n=1). Of the remaining probands, 10 were not receiving medical therapy for JDM, and 2 did not provide medication history.

To identify key differences in microbiome diversity and community structure in JDM, we performed 16S rRNA amplicon sequencing of the V4 hypervariable region of oral and gut samples collected from these family-based cohorts (Tables S1, S2A). After filtering, we observed a total of 3,304 ASVs throughout our dataset (Table S2B), of which 1,448 ASVs were detected in gut microbiomes, 1,928 ASVs were detected in oral microbiomes, and 72 ASVs were seen in subsets of both fecal and oral gut microbiomes. Firmicutes and Bacteroides predominated in the fecal microbiomes (p<10^-10), whereas Proteobacteria, Fusobacteria, and Actinobacteria were also highly prevalent in the oral microbiomes (p<10^-21; Figure 1A, B). The 72 ASVs that were shared between fecal and oral microbiomes comprised of diverse phyla, constituted <1% of total relative abundance in samples (on average), and clustered predominantly by sample type (i.e., oral versus fecal) rather than by individual or family (Figure S1). Principal coordinates analysis (PCoA) based on weighted UniFrac distances demonstrated stark differences in oral and fecal microbiota community structures (Figure 1C). Matching expectations of greater diversity and biomass in fecal than in oral samples, fecal microbiomes were characterized by greater alpha diversity than oral microbiomes (p<10^-13, two-tailed Student’s t-test; Figure S2A); similarly, genomic DNA concentrations were significantly higher in fecal than oral samples (p<10^-11, two-tailed Student’s t-test; Figure S2B).

JDM was a significant factor explaining variance in weighted UniFrac distances in fecal and supragingival swab samples (p<0.03, PERMANOVA). Therefore, we next sought to identify specific bacterial taxa differentiating children with JDM from their unaffected siblings in all fecal and oral samples, reasoning that differences in specific individual taxa may be biologically important despite a lack of globally significant differences (i.e., UniFrac distances) in saliva or subgingival dental plaque samples. Our unique study design permitted us to adjust for microbiome biases attributable to family and age, e.g. via paired comparisons of JDM probands to their siblings.

Approximately 4% of bacterial ASVs were differentially abundant in fecal microbiomes between JDM probands and their siblings (Figure 2A). Most of these ASVs were rare Firmicutes, namely Clostridia and Bacilli. Two ASVs significantly enriched in the stool of JDM probands compared to their healthy siblings were classified as unannotated bacteria. An NCBI BLAST search of these sequences suggested that these reads represent an unknown Ruminococcaceae or Oscillospiraceae species (0.08 ± 0.03% JDM vs 0.05 ± 0.03% siblings) and an uncultured Faecalibacterium (which is in the Oscillospiraceae family; 1.13 ± 0.15% JDM vs 0.95 ± 0.12% siblings). Additionally, an unknown Roseburia species (0.15 ± 0.05% JDM vs 0.02 ± 0.02% siblings) and Ruthenibacterium lactatiformans (0.03 ± 0.007% JDM vs 0.007 ± 0.005% siblings) were significantly enriched in the stool of JDM probands. In contrast, unknown species of the genera Agathobacter (0.02 ± 0.02% JDM vs 0.07 ± 0.04% siblings), Anaerostipes (0.55 ± 0.13% JDM vs 0.83 ± 0.25% siblings), Fusicatenibacter (0.9 ± 0.2% JDM vs 1.7 ± 0.3% siblings), Streptococcus (3 ASVs: 0.1 ± 0.05% JDM vs 0.4 ± 0.2% siblings, 0.007 ± 0.008% JDM vs 0.02 ± 0.01% siblings, and 0.007 ± 0.004% JDM vs 0.02 ± 0.02% siblings), and Subdoligranulum (0.0004 ± 0.0004% JDM vs 0.01 ± 0.01% siblings) were significantly depleted in JDM probands.

A greater proportion and diversity of ASVs were differentially abundant in the oral microbiomes between JDM probands and their siblings. Approximately 8% of ASVs identified in the anterior subgingival dental plaque were significantly different between JDM probands and their healthy siblings (Figure 2B). The most enriched ASV in JDM probands was an unknown Parvimonas species (0.03 ± 0.03% JDM vs 0.006 ± 0.006% siblings). Meanwhile, Corynebacterium matruchotii (2 ASVs: 1.9 ± 0.3% JDM vs 2.4 ± 0.1% siblings and 1.4 ± 0.2% JDM vs 1.8 ± 0.08% siblings), Porphyromonas pasteri (2 ASVs: 0.95 ± 0.1% JDM vs 0.1 ± 0.1% siblings and 0.6 ± 0.1% JDM vs 0.7 ± 0.1% siblings), Granulicatella (0.3 ± 0.03% JDM vs 0.4 ± 0.04% siblings), Prevotella loescheii (0.3 ± 0.1% JDM vs 0.4 ± 0.2% siblings), and an unclassified species with high sequence identity to Prevotella (0.1 ± 0.07% JDM vs 0.3 ± 0.1% siblings) were significantly depleted in JDM probands.

In posterior subgingival dental plaque samples, approximately 7% of identified ASVs were significantly differentially abundant between JDM probands and their siblings (Figure 2C). ASVs representing Actinomyces species (2 ASVs: 1.0 ± 0.2% JDM vs 0.8 ± 0.1% siblings and 0.6 ± 0.1% JDM vs 0.5 ± 0.07% siblings), Gemella (0.7 ± 0.1% JDM vs 0.6 ± 0.1% siblings), Streptococcus pneumoniae (3.6 ± 0.5% JDM vs 3.3 ± 0.4% siblings), and an unclassified ASV with high sequence identity to Lautropia (0.08 ± 0.08% JDM vs 0.03 ± 0.03% siblings) were enriched in JDM probands. Additionally, Actinomyces dentalis (0.01 ± 0.01% JDM vs 0.05 ± 0.01% siblings), Leptotrichia (0.03 ± 0.03% JDM vs 0.09 ± 0.09% siblings), and an unclassified ASV with high identity to Prevotella (0.05 ± 0.03% JDM vs 0.2 ± 0.09% siblings) were significantly depleted in JDM probands.

We also observed significantly differentially abundant ASVs in the saliva and supragingival swab samples from JDM probands, all of which were found in very low abundances. In the saliva, JDM probands were enriched with Veillonella (0.07 ± 0.06% JDM vs 0.05 ± 0.02% siblings) and Prevotella shahii (0.02 ± 0.008% JDM vs 0.009 ± 0.009% siblings). In contrast, Prevotella melaninogenica (0.005 ± 0.003% JDM vs 0.01 ± 0.01% siblings) was depleted in JDM proband saliva samples. In the supragingival swab samples, the most significantly enriched ASVs in JDM probands were Neisseria oralis (0.2 ± 0.07% JDM vs 0.09 ± 0.06% siblings), 2 unclassified species with high sequence identity to Corynebacterium (0.03 ± 0.01% JDM vs 0.02 ± 0.01% siblings), and Prevotella (0.04 ± 0.02% JDM vs 0.02 ± 0.01% siblings).

Oral and fecal microbiomes clustered by family unit and particularly by sibling relationship

To characterize similarities of microbiomes within and between families, we calculated pairwise distances between samples using UniFrac. Family unit was the predominant factor explaining variance in both weighted and unweighted UniFrac distances between samples for each sample type (p<0.001 in fecal, saliva, supragingival swab, and posterior subgingival dental plaque samples; p<0.01 in anterior subgingival dental plaque samples, permutational multivariate analysis of variance [PERMANOVA]).

Paired unweighted UniFrac distances between JDM probands and their own siblings were significantly smaller than the paired distances between JDM probands and their own parents (p<0.05, two-tailed Student’s t-test; Figure 3A). In fact, the unweighted UniFrac distances between siblings was significantly smaller than the distances between JDM probands in our cohort (p<0.002, two-tailed Student’s t-test). This trend was also seen in a comparison of weighted UniFrac distances (p<0.08, two-tailed Student’s t-test; Figure 3B). In other words, siblinghood was the most reliable of three indicators of fecal microbiome similarity in our study, which also included family unit and JDM diagnosis.

Saliva, supragingival swab, and subgingival dental plaque microbiomes were highly similar between members of families (Figures 3C-H). The UniFrac distances between siblings’ oral microbiomes were not significantly different from the distances between JDM probands and their own parents (or between the healthy siblings and their own parents; two-tailed Student’s t-tests; Figures 3C-G). Intra-familial distances were significantly smaller than all inter-familial distances (p<0.005, two-tailed Student’s t-test; Figures 3C-H). As with gut microbiomes, the UniFrac distances between oral microbiomes of JDM probands and their healthy siblings were significantly less than the distances between all JDM probands (p<0.02, two-tailed Student’s t-test), further supporting the notion of a stronger effect of sibling relationship than JDM diagnosis. UniFrac distances between oral microbiomes of parents within a household were significantly smaller than the distances between adults in different households (p<0.05 for all comparisons, except weighted UniFrac distances of saliva samples, two-tailed Student’s t-test; Figures 3C-F) — a trend seen in fecal microbiomes as well (Figure 3A).

Family units of JDM probands exhibit dysbiotic microbiomes

We next asked whether the familial clustering seen in our cohort of JDM probands and their nuclear families reflected familial dysbiosis. To compare gut microbiomes with healthy individuals lacking any known genetic or familial predisposition to JDM, we performed a cross-dataset analysis of adults [27,28] and children (selecting data from children ≥3 years of age) [29] whose gut microbiomes had been previously profiled by other groups using the same sequencing strategy and otherwise similar methods (including use of the same sample collection kit in one study [27]; Table S2C). In PCoA analyses of both weighted and unweighted UniFrac distances, the stool samples from individuals in our study (children with JDM, their unaffected siblings, and/or their parents) all clustered separately from samples from these three published datasets (separation along PC1 accounted for 70% [unweighted] and 93% [weighted] of variance in the aggregate dataset; p<0.001, PERMANOVA; Figure 4A). Potential confounders in this analysis may be biases imparted by batch effects or methodological differences. Indeed, study site (SCH, NIH, and sites from the published studies whose data we re-analyzed) was a significant covariate explaining variance in both unweighted and weighted UniFrac distances (p<0.001, PERMANOVA; Figure 4B). However, the effect size of having a familial relationship to a child with JDM was 1-2 orders of magnitude greater than the effect size of study site (unweighted UniFrac distances: 35% vs 5%; weighted UniFrac distances: 76.7% vs 1.2%; PERMANOVA). Therefore, this suggests that microbiomes of not only JDM probands, but entire JDM family units, are markedly different from healthy microbiomes.

Reasoning that the microbiomes of JDM probands and their families may be enriched for novel or atypical organisms, we then further investigated contributions of unannotated bacteria to our diversity analyses. We rank-ordered ASVs that were classified via QIIME2 simply as “Bacteria.” The 20 most abundant ASVs, which together accounted for >50% of unannotated bacteria (Figure S3), were not observed in the downloaded samples, but altogether accounted for 14.6% ± 0.7% relative abundance in stool samples from families in our cohort. Remarkably, UniFrac distances between “core” microbiomes (defined here as the sum of all genera represented in at least one sample from each study; Figure 4C) significantly increased with the addition of these 20 ASVs (Figure 4D). To investigate whether these sequences had been previously reported, we performed an NCBI BLAST search of the sequencing reads associated with these ASVs (Table S2D). Many of the sequences associated with these 20 ASVs, which largely accounted for the significant separation between fecal microbiomes in our dataset and the published comparator microbiomes, had been previously reported in studies profiling inflammatory and immune-mediated conditions (PMIDs reported in Table S2D). Twelve of these ASVs mapped to the Bacteroides genus, 5 mapped to Lachnospiraceae, 2 mapped to Faecalibacterium, and 1 mapped to Ruminococcus.

Alpha diversity was significantly increased in our JDM cohort than these published cohorts (287.5 ± 7.2 vs 147.6 ± 4.9; p<10^-22 [observed species]; 306.3 ± 7.7 vs 152.8 ± 5.0; p < 10^-23 [Chao1]; two-tailed Student’s t-test; Figure S4). Alpha diversity was significantly lower in children compared to adults (271.4 ± 8.2 vs 308.5 ± 10.9 [observed species]; 288.0 ± 8.3 vs 325.1 ± 11.9 [Chao1]; p<0.02, two-tailed Student’s t-test). Consistent with our findings of sibling-associated microbiome similarities, alpha diversity was not significantly different in a paired analysis comparing JDM probands to their unaffected siblings (p>0.05, paired two-tailed Student’s t-test; both observed species and Chao1 metrics).

Our study examined oral and fecal microbiomes in JDM, a rare pediatric multi-systemic autoimmune disease of unclear etiology. We studied JDM probands in comparison to their unaffected family members, employing healthy siblings as the best possible controls to match for both age and environmental influences. The fact that JDM patients and their siblings were found to have such similar microbiomes proved advantageous: controlling (to some extent) for microbiome effects related to shared genetic and environmental contexts (including genetic determinants of the microbiome community structure, vertically transmitted microbiomes, shared dietary and cultural influences, and co-habitation) permitted us to identify microbiome features with a greater likelihood of being clinically pertinent to JDM. Despite JDM probands in our study having very mild disease, we identified several bacteria that were significantly enriched or depleted in JDM, and each was observed in low abundance, consistent with findings of prior studies investigating the microbiome in immune-mediated diseases [30].

We observed significant enrichment of Faecalibacterium and Ruminococcaceae (which are both members of the Lachnospiraceae family) in the fecal microbiomes of children with JDM. Faecalibacterium are typically associated with anti-inflammatory properties, and while adult studies elucidate their impact on inflammation [31,32], their immune-modulating effects in children are unknown. Intriguingly, Faecalibacterium are enriched in the gut microbiomes of children with Crohn's disease [33], suggesting age-dependent physiologic effects of this microbe. Altered representation of Faecalibacterium has also been reported in Sjögren’s syndrome [34], RA [35], and juvenile idiopathic arthritis (JIA) [36,37]. Fecal abundances of Ruminococcus species are greater in individuals with systemic lupus erythematosus and also correlated with disease activity [38]. These findings add to a body of evidence linking Lachnospiraceae bacterial species to immune-mediated diseases.

Additionally, we observed a significant enrichment of Roseburia and Muribaculaceae in the fecal microbiomes of children with JDM. Roseburia species have previously been associated with other autoimmune diseases such as type 1 diabetes [39] and have the potential to activate autoantibody responses in vivo [40]. Levels of Muribaculaceae have previously been shown to positively correlate with disease activity as well as pro-inflammatory cytokines IL-17, TNF-α, and IFN-γ in RA patients [41].

Subdoligranulum, on the other hand, was a depleted taxon, consistent with a previous study that reported lower Subdoligranulum in Crohn's disease, suggesting it may be a putative probiotic in multiple contexts [42]. Indeed, Subdoligranulum variabile may be a potentially key member of the healthy human gut microbiome [43].

Interestingly, we observed significant decreases in 3 different Streptococcus sequences in the fecal samples of JDM probands, which contradicts previous reports of positive associations with autoimmune diseases, such as RA [35], type 1 diabetes [39], and Sjögren’s syndrome [34]. This discrepancy could be due to differences in the disease nature of JDM, the age of affected individuals, use of immunosuppressive medications, or in the Streptococcus species themselves.

Different bacteria comprised JDM signatures of oral samples; in fact, few ASVs were found in both oral and stool samples, consistent with a recent study suggesting that oral bacteria do not colonize the distal gut [44]. In the anterior subgingival dental plaque, two Fusobacterium ASVs were significantly enriched in JDM probands, consistent with previous studies reporting increased abundance in RA [45] and periodontal disease [46], which has been linked to JDM [47]. Additionally, posterior subgingival dental plaque from JDM probands were enriched for S. pneumoniae, which displays a plasma binding protein on its cell wall that may be immunogenic [48]. We also observed a significant enrichment of Neisseriaceae and Laurotrpia in JDM probands which have both been associated with RA [45]. Other notable enriched genera in JDM probands included Campylobacter and Gemella which have been associated with periodontal disease and JIA, respectively [49,50]. Conversely, we observed a loss of Corynebacterium matruchotii abundance in JDM probands, consistent with reports of lower levels in RA, JIA, and periodontal disease [51,52]. JDM probands also had a significant reduction in Porphyromonas pasteri (the most abundant and prevalent Porphyromonas species in healthy adults [53,54]) compared to their healthy siblings. Porphyromonas has been associated with RA in the subgingival plaque [45,51]. A recent report of Sjögren’s syndrome found higher abundance of Porphyromonas pasteri in the oral microbiome of healthy controls [55]. Though our findings in the saliva and supragingival swab samples indicated that the microbial differences between JDM probands and their healthy siblings are among low abundant bacteria, we still observed interesting differences. For instance, JDM probands were enriched with Veillonella and Prevotella species in their saliva. A recent report of RA patients suggested that these microbes could predispose patients to the development of this autoimmune disease in the early stages [56], consistent with our observations. Additionally, the increased abundance of a Neisseria species in the supragingival swab samples of JDM probands corroborates similar findings in the anterior subgingival dental plaque. The observation that species from the same genera (e.g. Streptococcus and Prevotella) were both enriched and depleted in JDM probands suggests that microbial-disease associations may be species-specific. Interestingly, each of these species has been correlative with autoimmune and periodontal diseases in other studies [45,57]. Thus, analyses using higher-level taxonomic classifications may lose the granularity needed to determine disease associations with specific bacterial species.

We found that the nuclear family unit had a very strong impact on oral and gut microbiomes, as expected. In our cohort, children had lower alpha diversity than adults, consistent with prior reports demonstrating increasing alpha diversity throughout childhood until a plateau in early adulthood [29,58]. Unweighted UniFrac distances of fecal samples revealed JDM probands were more similar to their siblings than to their parents and other probands. These observations suggest that the sibling relationship has a larger effect on gut microbiome similarity than does carrying a diagnosis of JDM, and the similarities among siblings are most pronounced with respect to less abundant taxa. As the youngest child recruited was 3.5 years, this observation is not confounded by the rapidly evolving developmental microbiome program that occurs in the first 3 years of life [29,58]. Together, these findings offer further support to the generally accepted notion that environmental exposures play major roles in microbiome assembly and establishment in childhood. Overall, we observed smaller distances between family members living in the same household compared to unrelated individuals – evidencing convergence of microbiomes of unrelated individuals attributable to living in the same household, particularly with respect to rarer microbes [59]. Oral microbiomes of children within a household (i.e., JDM probands and their siblings) were no more similar than the microbiomes of adults in the same household. This finding may reflect effects of kissing, common meals and similar dietary habits, direct sharing of utensils, etc., among family members, although these specific aspects of family life were not queried in our study.

While some studies of immune-mediated diseases such as RA [60] show lower alpha diversity in disease cohorts, studies of JIA [36], Sjögren’s syndrome [61], and type 1 diabetes [62] have found increased alpha diversity among diseased patients, which is in line with our findings in the comparison to previous microbiome surveys.

Finally, we identified a panel of ASVs enriched in JDM families that accounted for significant variance between our cohort and previously studied individuals. These ASVs were classified as unannotated bacteria by the SILVA classifier in QIIME2. A NCBI BLAST search suggests that these ASVs included representatives of Bacteroides, Lachnospiraceae, Faecalibacterium, Ruminococcus and have been previously described in the context of immune-mediated diseases [38]. Future studies involving cultivation and characterization of these, and other novel bacterial isolates derived from JDM patient samples, may provide insights into pathogenesis. Our findings suggest that in depth microbial investigations of fecal and subgingival plaque samples (and perhaps less so salivary or supragingival swab samples) may be particularly revealing.

Our study has several limitations. First, we lacked control families without any immune-mediated diseases. We attempted to overcome this limitation by analyzing published datasets and found evidence of familial dysbiosis. However, based on the available data and our current analysis, we are unable to draw conclusions regarding immunological effects of these familial microbiome differences. One compelling hypothesis is that interactions between one’s genetics and microbiome drive immune dysfunction a la the “ecological model of dysbiosis” [63]. Second, our sample collection methods were designed solely for DNA-based profiling. Therefore, we could not culture bacterial isolates for direct testing of specific bacteria or bacterial metabolites in model systems of dysregulated immunity to study their functional impact. Third, our study was cross-sectional in design, a feature often inherent to studying a rare disease. Enrolled subjects largely had well-controlled or mildly active disease and were receiving immunosuppressive therapies. We were unable to distinguish potential disease-driving bacteria from those whose representation was altered because of microenvironmental changes. An ideal future study would be a longitudinal multicenter study where we could sample the microbiome of new-onset JDM patients just prior to initiation of therapy, at timepoints during therapy, and once in remission to assess for changes in the microbiome. Fourth, due to limitations in sample size and medication use history, we were unable to assess or control for the effects of specific drugs and drug classes. It known that immunosuppressive drugs such as methotrexate [64], oral corticosteroids (i.e., prednisone) [65], and mycophenolate mofetil [66] can alter the structure of bacterial communities. Despite these therapies, JDM probands and their siblings displayed high similarities in overall composition of oral and gut microbiomes.

Rare chronic pediatric diseases such as JDM are often challenging to diagnose and treat, and their diagnosis impacts entire family units. In this novel study of the oral and fecal microbiomes in JDM, in which we profiled JDM probands and their unaffected family members, we showed that family is a major influence on microbiome variability. This challenges the clinical paradigm of focusing on affected individuals and raises the prospect of whether practitioners ought to focus on family units in characterizing the microbiomes of patients. More broadly, our findings raise the question of whether microbiome-based therapies can ultimately succeed in treating individuals in isolation, or whether familial and social network units need to be addressed as a whole. In an analysis adjusting for microbiome differences attributable to family, we nevertheless found differences in several potentially immunomodulatory bacteria in patients with JDM. Further research will be needed to understand the role of these bacteria in disease pathogenesis and whether targeting them with microbiome-based interventions may be of clinical therapeutic value.

JDM, juvenile dermatomyositis; NIDCR, National Institute of Dental and Craniofacial Research; NIH, National Institutes of Health; NIEHS, National Institute of Environmental Health Sciences; SCH, Seattle Children’s Hospital; MMT, Manual Muscle Testing; CDASI, Cutaneous Dermatomyositis Disease Area and Severity Index; MSA, myositis-specific autoantibodies; MAA, myositis-associated autoantibodies; ANA, anti-nuclear antibody; IP, immunoprecipitation; PCoA, principal coordinates analysis; PERMANOVA, permutational multivariate analysis of variance; IBD, inflammatory bowel disease; ASV, Amplicon Sequence Variant; RA, rheumatoid arthritis; JIA, juvenile idiopathic arthritis.

Ethics approval and consent to participate. Informed written consent was obtained from and documented for all study participants. The research methods used were approved by the Institutional Review Boards at Seattle Children’s Hospital (protocol ID 1044) and at the National Institutes of Health (protocol ID 03-E-0099).

Consent for publication. Not applicable.

Availability of data and material. The datasets generated and analyzed during the current study are available in the ENA repository PRJEB50544.

Competing interests. The authors declare that they have no competing interests.

Funding. This work was supported by grants from the NIH (NIDDK K08 DK111941), the CARRA-Arthritis Foundation, the Cure JM Foundation grant as part of the Cure JM - Seattle Children’s Myositis Center of Excellence, the Office of the Clinical Director and the intramural research program at the National Institute of Dental and Craniofacial Research, and the intramural research program at the National Institute of Environmental Health Sciences at the National Institutes of Health (ZIAES101074). Nastaran Bayat was supported by a clinical research support services contract for NIEHS (Contract HHSN2732016000021).

Authors' contributions (using CRediT taxonomy).

Conceptualization, S.T.K., A.C., P.L., J.S.M., L.R., S.S., A.M.S., N.D.;

Data curation, S.T.K., A.C., M.E., P.L., L.R., S.S., A.M.S., N.D.;

Formal analysis, S.T.K., A.C., E.P., L.R., S.S., A.M.S., N.D.;

Funding acquisition, L.R., S.S., A.M.S., N.D.;

Investigation, S.T.K., A.C., E.P., P.L., M.E., L.B., P.G., H.J.S., N.L., A.S., R.V., N.B., L.R., S.S., A.M.S., N.D.;

Methodology, S.T.K., A.C., L.R., P.L., S.S., A.M.S., N.D.;

Project administration, S.T.K., A.C., L.R., S.S., A.M.S., N.D.;

Resources, L.R., S.S., A.M.S., and N.D.;

Software, S.T.K., A.C., E.P., and N.D.;

Supervision, P.L., L.R., S.S., A.M.S., and N.D.;

Validation, S.T.K., A.C., L.R., S.S., A.M.S., N.D.;

Visualization, S.T.K., A.C., L.R., S.S., A.M.S., N.D.;

Writing - original draft, S.T.K., A.C., L.R., S.S., A.M.S., N.D.;

Writing - review & editing, S.T.K., A.C., E.P., P.L., M.E., L.B., P.G., H.J.S., J.S.M., N.L., A.S., R.V., N.B., L.R., S.S., A.M.S., N.D.

Acknowledgements. We thank study participants, clinical study coordinators, and administrators who made this multi-institutional study possible. We are also indebted to our colleagues in Seattle (including Sam Minot, Amy Willis, and other members of the University of Washington Microbiome Center) and at the NIH (Drs. Robert Colbert and Jacqueline Mays) for helpful suggestions in the preparation of this paper.

Mendez EP, Lipton R, Ramsey-Goldman R, Roettcher P, Bowyer S, Dyer A, et al. US incidence of juvenile dermatomyositis, 1995-1998: Results from the National Institute of Arthritis and Musculoskeletal and Skin Diseases Registry. Arthritis Rheum. 2003;49:300–5.
Feldman BM, Rider LG, Reed AM, Pachman LM. Juvenile dermatomyositis and other idiopathic inflammatory myopathies of childhood. Lancet Lond Engl. 2008;371:2201–12.
Rider LG, Katz JD, Jones OY. Developments in the classification and treatment of the juvenile idiopathic inflammatory myopathies. Rheum Dis Clin North Am. 2013;39:877–904.
McCann LJ, Livermore P, Wilkinson MGL, Wedderburn LR. Juvenile dermatomyositis. Where are we now? Clin Exp Rheumatol. 2022;40:394–403.
Dey N, Soergel DA, Repo S, Brenner SE. Association of gut microbiota with post-operative clinical course in Crohn’s disease. BMC Gastroenterol. 2013;13:131.
Vieira SM, Pagovich OE, Kriegel MA. Diet, microbiota and autoimmune diseases. Lupus. 2014;23:518–26.
Franzosa EA, Sirota-Madi A, Avila-Pacheco J, Fornelos N, Haiser HJ, Reinker S, et al. Gut microbiome structure and metabolic activity in inflammatory bowel disease. Nat Microbiol. 2018;1.
Wu W-JH, Zegarra-Ruiz DF, Diehl GE. Intestinal Microbes in Autoimmune and Inflammatory Disease. Front Immunol. 2020;11:597966.
Savioli C, Silva CAA, Fabri GMC, Kozu K, Campos LMA, Bonfá E, et al. Gingival capillary changes and oral motor weakness in juvenile dermatomyositis. Rheumatol Oxf Engl. 2010;49:1962–70.
Willis AD. Rigorous Statistical Methods for Rigorous Microbiome Science. mSystems. 2019;4:e00117-19.
Bohan A, Peter JB. Polymyositis and Dermatomyositis: (First of Two Parts). N Engl J Med. 1975;292:344–7.
Bohan A, Peter JB. Polymyositis and Dermatomyositis: (Second of Two Parts). N Engl J Med. 1975;292:403–7.
Rider LG, Werth VP, Huber AM, Alexanderson H, Rao AP, Ruperto N, et al. Measures of adult and juvenile dermatomyositis, polymyositis, and inclusion body myositis: Physician and Patient/Parent Global Activity, Manual Muscle Testing (MMT), Health Assessment Questionnaire (HAQ)/Childhood Health Assessment Questionnaire (C-HAQ), Childhood Myositis Assessment Scale (CMAS), Myositis Disease Activity Assessment Tool (MDAAT), Disease Activity Score (DAS), Short Form 36 (SF-36), Child Health Questionnaire (CHQ), physician global damage, Myositis Damage Index (MDI), Quantitative Muscle Testing (QMT), Myositis Functional Index-2 (FI-2), Myositis Activities Profile (MAP), Inclusion Body Myositis Functional Rating Scale (IBMFRS), Cutaneous Dermatomyositis Disease Area and Severity Index (CDASI), Cutaneous Assessment Tool (CAT), Dermatomyositis Skin Severity Index (DSSI), Skindex, and Dermatology Life Quality Index (DLQI). Arthritis Care Res. 2011;63 Suppl 11:S118-157.
Ahmed S, Chen KL, Werth VP. The validity and utility of the Cutaneous Disease Area and Severity Index (CDASI) as a clinical outcome instrument in dermatomyositis: A comprehensive review. Semin Arthritis Rheum. 2020;50:458–62.
Dey N, Wagner VE, Blanton LV, Cheng J, Fontana L, Haque R, et al. Regulators of Gut Motility Revealed by a Gnotobiotic Model of Diet-Microbiome Interactions Related to Travel. Cell. 2015;163:95–107.
Koester ST, Li N, Lachance DM, Morella NM, Dey N. Variability in digestive and respiratory tract Ace2 expression is associated with the microbiome. Foster J, editor. PLOS ONE. 2021;16:e0248730.
Li N, Koester ST, Lachance DM, Dutta M, Cui JY, Dey N. Microbiome-encoded bile acid metabolism modulates colonic transit times. iScience. 2021;24:102508.
Callahan BJ, McMurdie PJ, Holmes SP. Exact sequence variants should replace operational taxonomic units in marker-gene data analysis. ISME J. 2017;11:2639–43.
Callahan BJ, McMurdie PJ, Rosen MJ, Han AW, Johnson AJA, Holmes SP. DADA2: High-resolution sample inference from Illumina amplicon data. Nat Methods. 2016;13:581–3.
Bolyen E, Rideout JR, Dillon MR, Bokulich NA, Abnet CC, Al-Ghalith GA, et al. Reproducible, interactive, scalable and extensible microbiome data science using QIIME 2. Nat Biotechnol. 2019;37:852–7.
Lozupone C, Knight R. UniFrac: a new phylogenetic method for comparing microbial communities. Appl Environ Microbiol. 2005;71:8228–35.
McMurdie PJ, Holmes S. Waste Not, Want Not: Why Rarefying Microbiome Data Is Inadmissible. McHardy AC, editor. PLoS Comput Biol. 2014;10:e1003531.
Willis AD. Rarefaction, Alpha Diversity, and Statistics. Front Microbiol. 2019;10:2407.
Shah M, Mamyrova G, Targoff IN, Huber AM, Malley JD, Rice MM, et al. The clinical phenotypes of the juvenile idiopathic inflammatory myopathies. Medicine (Baltimore). 2013;92:25–41.
Martin BD, Witten D, Willis AD. Modeling microbial abundances and dysbiosis with beta-binomial regression. Ann Appl Stat [Internet]. 2020 [cited 2022 Jan 20];14. Available from: https://projecteuclid.org/journals/annals-of-applied-statistics/volume-14/issue-1/Modeling-microbial-abundances-and-dysbiosis-with-beta-binomial-regression/10.1214/19-AOAS1283.full
McLaren MR, Willis AD, Callahan BJ. Consistent and correctable bias in metagenomic sequencing experiments. eLife. 2019;8:e46923.
ENDIA Study Group*, Penington JS, Penno MAS, Ngui KM, Ajami NJ, Roth-Schulze AJ, et al. Influence of fecal collection conditions and 16S rRNA gene sequencing at two centers on human gut microbiota analysis. Sci Rep. 2018;8:4386.
Mori G, Orena BS, Cultrera I, Barbieri G, Albertini AM, Ranzani GN, et al. Gut Microbiota Analysis in Postoperative Lynch Syndrome Patients. Front Microbiol. 2019;10:1746.
Roswall J, Olsson LM, Kovatcheva-Datchary P, Nilsson S, Tremaroli V, Simon M-C, et al. Developmental trajectory of the healthy human gut microbiota during the first 5 years of life. Cell Host Microbe. 2021;29:765-776.e3.
Chen J, Wright K, Davis JM, Jeraldo P, Marietta EV, Murray J, et al. An expansion of rare lineage intestinal microbes characterizes rheumatoid arthritis. Genome Med. 2016;8:43.
Sokol H, Pigneur B, Watterlot L, Lakhdari O, Bermúdez-Humarán LG, Gratadoux J-J, et al. Faecalibacterium prausnitzii is an anti-inflammatory commensal bacterium identified by gut microbiota analysis of Crohn disease patients. Proc Natl Acad Sci U S A. 2008;105:16731–6.
Sokol H, Seksik P, Furet JP, Firmesse O, Nion-Larmurier I, Beaugerie L, et al. Low counts of Faecalibacterium prausnitzii in colitis microbiota. Inflamm Bowel Dis. 2009;15:1183–9.
Hansen R, Russell RK, Reiff C, Louis P, McIntosh F, Berry SH, et al. Microbiota of de-novo pediatric IBD: increased Faecalibacterium prausnitzii and reduced bacterial diversity in Crohn’s but not in ulcerative colitis. Am J Gastroenterol. 2012;107:1913–22.
de Paiva CS, Jones DB, Stern ME, Bian F, Moore QL, Corbiere S, et al. Altered Mucosal Microbiome Diversity and Disease Severity in Sjögren Syndrome. Sci Rep. 2016;6:23561.
Picchianti-Diamanti A, Panebianco C, Salemi S, Sorgi ML, Di Rosa R, Tropea A, et al. Analysis of Gut Microbiota in Rheumatoid Arthritis Patients: Disease-Related Dysbiosis and Modifications Induced by Etanercept. Int J Mol Sci. 2018;19:E2938.
Dong Y-Q, Wang W, Li J, Ma M-S, Zhong L-Q, Wei Q-J, et al. Characterization of microbiota in systemic-onset juvenile idiopathic arthritis with different disease severities. World J Clin Cases. 2019;7:2734–45.
Qian X, Liu Y-X, Ye X, Zheng W, Lv S, Mo M, et al. Gut microbiota in children with juvenile idiopathic arthritis: characteristics, biomarker identification, and usefulness in clinical prediction. BMC Genomics. 2020;21:286.
Azzouz D, Omarbekova A, Heguy A, Schwudke D, Gisch N, Rovin BH, et al. Lupus nephritis is linked to disease-activity associated expansions and immunity to a gut commensal. Ann Rheum Dis. 2019;78:947–56.
Vatanen T, Franzosa EA, Schwager R, Tripathi S, Arthur TD, Vehik K, et al. The human gut microbiome in early-onset type 1 diabetes from the TEDDY study. Nature. 2018;562:589–94.
Ruff WE, Dehner C, Kim WJ, Pagovich O, Aguiar CL, Yu AT, et al. Pathogenic Autoreactive T and B Cells Cross-React with Mimotopes Expressed by a Common Human Gut Commensal to Trigger Autoimmunity. Cell Host Microbe. 2019;26:100-113.e8.
Wang Q, Zhang S-X, Chang M-J, Qiao J, Wang C-H, Li X-F, et al. Characteristics of the Gut Microbiome and Its Relationship With Peripheral CD4+ T Cell Subpopulations and Cytokines in Rheumatoid Arthritis. Front Microbiol. 2022;13:799602.
Kaakoush NO, Day AS, Huinao KD, Leach ST, Lemberg DA, Dowd SE, et al. Microbial dysbiosis in pediatric patients with Crohn’s disease. J Clin Microbiol. 2012;50:3258–66.
Kim ES, Tarassishin L, Eisele C, Barre A, Nair N, Rendon A, et al. Longitudinal Changes in Fecal Calprotectin Levels Among Pregnant Women With and Without Inflammatory Bowel Disease and Their Babies. Gastroenterology. 2021;160:1118-1130.e3.
Rashidi A, Ebadi M, Weisdorf DJ, Costalonga M, Staley C. No evidence for colonization of oral bacteria in the distal gut in healthy adults. Proc Natl Acad Sci. 2021;118:e2114152118.
Chen B, Zhao Y, Li S, Yang L, Wang H, Wang T, et al. Variations in oral microbiome profiles in rheumatoid arthritis and osteoarthritis with potential biomarkers for arthritis screening. Sci Rep. 2018;8:17126.
Han YW, Shi W, Huang GT, Kinder Haake S, Park NH, Kuramitsu H, et al. Interactions between periodontal bacteria and human oral epithelial cells: Fusobacterium nucleatum adheres to and invades epithelial cells. Infect Immun. 2000;68:3140–6.
Kozu KT, Silva CA, Aikawa NE, Pereira RMR, Sallum AM, Savioli C, et al. Juvenile dermatomyositis: is periodontal disease associated with dyslipidemia? Adv Rheumatol. 2018;58:28.
Bergmann S, Rohde M, Chhatwal GS, Hammerschmidt S. alpha-Enolase of Streptococcus pneumoniae is a plasmin(ogen)-binding protein displayed on the bacterial cell surface. Mol Microbiol. 2001;40:1273–87.
Macuch PJ, Tanner AC. Campylobacter species in health, gingivitis, and periodontitis. J Dent Res. 2000;79:785–92.
Frid P, Baraniya D, Halbig J, Rypdal V, Songstad NT, Rosèn A, et al. Salivary Oral Microbiome of Children With Juvenile Idiopathic Arthritis: A Norwegian Cross-Sectional Study. Front Cell Infect Microbiol. 2020;10:602239.
Scher JU, Ubeda C, Equinda M, Khanin R, Buischi Y, Viale A, et al. Periodontal disease and the oral microbiota in new-onset rheumatoid arthritis. Arthritis Rheum. 2012;64:3083–94.
Grevich S, Lee P, Leroux B, Ringold S, Darveau R, Henstorf G, et al. Oral health and plaque microbial profile in juvenile idiopathic arthritis. Pediatr Rheumatol Online J. 2019;17:81.
Yasunaga H, Takeshita T, Shibata Y, Furuta M, Shimazaki Y, Akifusa S, et al. Exploration of bacterial species associated with the salivary microbiome of individuals with a low susceptibility to dental caries. Clin Oral Investig. 2017;21:2399–406.
Lenartova M, Tesinska B, Janatova T, Hrebicek O, Mysak J, Janata J, et al. The Oral Microbiome in Periodontal Health. Front Cell Infect Microbiol. 2021;11:629723.
Rusthen S, Kristoffersen AK, Young A, Galtung HK, Petrovski BÉ, Palm Ø, et al. Dysbiotic salivary microbiota in dry mouth and primary Sjögren’s syndrome patients. PloS One. 2019;14:e0218319.
Kroese JM, Brandt BW, Buijs MJ, Crielaard W, Lobbezoo F, Loos BG, et al. Differences in the Oral Microbiome in Patients With Early Rheumatoid Arthritis and Individuals at Risk of Rheumatoid Arthritis Compared to Healthy Individuals. Arthritis Rheumatol Hoboken NJ. 2021;73:1986–93.
Maeda N, Okamoto M, Kondo K, Ishikawa H, Osada R, Tsurumoto A, et al. Incidence of Prevotella intermedia and Prevotella nigrescens in periodontal health and disease. Microbiol Immunol. 1998;42:583–9.
Yatsunenko T, Rey FE, Manary MJ, Trehan I, Dominguez-Bello MG, Contreras M, et al. Human gut microbiome viewed across age and geography. Nature. 2012;486:222–7.
Song SJ, Lauber C, Costello EK, Lozupone CA, Humphrey G, Berg-Lyons D, et al. Cohabiting family members share microbiota with one another and with their dogs. eLife. 2013;2:e00458.
Sun Y, Chen Q, Lin P, Xu R, He D, Ji W, et al. Characteristics of Gut Microbiota in Patients With Rheumatoid Arthritis in Shanghai, China. Front Cell Infect Microbiol. 2019;9:369.
Mendez R, Watane A, Farhangi M, Cavuoto KM, Leith T, Budree S, et al. Gut microbial dysbiosis in individuals with Sjögren’s syndrome. Microb Cell Factories. 2020;19:90.
Mrozinska S, Kapusta P, Gosiewski T, Sroka-Oleksiak A, Ludwig-Słomczyńska AH, Matejko B, et al. The Gut Microbiota Profile According to Glycemic Control in Type 1 Diabetes Patients Treated with Personal Insulin Pumps. Microorganisms. 2021;9:155.
Gill T, Asquith M, Brooks SR, Rosenbaum JT, Colbert RA. Effects of HLA-B27 on Gut Microbiota in Experimental Spondyloarthritis Implicate an Ecological Model of Dysbiosis. Arthritis Rheumatol. 2018;70:555–65.
Nayak RR, Alexander M, Deshpande I, Stapleton-Gray K, Rimal B, Patterson AD, et al. Methotrexate impacts conserved pathways in diverse human gut bacteria leading to decreased host immune activation. Cell Host Microbe. 2021;29:362-377.e11.
Huang EY, Inoue T, Leone VA, Dalal S, Touw K, Wang Y, et al. Using corticosteroids to reshape the gut microbiome: implications for inflammatory bowel diseases. Inflamm Bowel Dis. 2015;21:963–72.
Jardou M, Provost Q, Brossier C, Pinault É, Sauvage F-L, Lawson R. Alteration of the gut microbiome in mycophenolate-induced enteropathy: impacts on the profile of short-chain fatty acids in a mouse model. BMC Pharmacol Toxicol. 2021;22:66.

Table 1. Summary of clinical and demographic characteristics of study participants.

	JDM probands (n=28)	Siblings (n=27)	Parents (n=43)
Age, years	10.0 ± 0.7	11.3 ± 0.9	41.4 ± 1.3
Gender, female	13 (46%)	11 (41%)	26 (60%)
Race*
Caucasian	20 (72%)
Mixed race	4 (14%)
Not reported	4 (14%)
Selected labs and disease activity scores	Median (IQR)
MMT-8 (maximum score 150)	150 (144 - 150)
CDASI activity score (maximum score 96)	2 (0 - 5)
CDASI damage score (maximum score 32 )	0.5 (0 - 2)
CRP (normal range <5 mg/dl)	0.8 (0.2 - 1)
ESR (normal range <20 mm/hr)	7 (7 - 8)
Aldolase (normal range <7 U/L)	5.8 (4 - 7.8)
CK (normal range 35 - 230 U/L)	118 (90 - 159)
LDH (normal range 370 - 840 U/L)	606 (557 - 678)
AST (normal range 5 - 41 U/L)	32.5 (23 - 36)
ALT (normal range 5 - 30 U/L)	18.5 (13.75 - 28.5)
Autoantibodies**
Non-specific autoantibodies
ANA (n=24)	17 (70%)
Anti-dsDNA antibody (n=17)	1 (6%)
Myositis-specific autoantibodies (MSA)
Anti-MJ (NXP2) antibody (n=8)	3 (38%)
Anti-p155 (TIF1) antibody (n=8)	2 (25%)
Anti-MDA5 antibody (n=8)	2 (25%)
Anti-Mi2 antibody (n=23)	1 (4%)
Myositis-associated autoantibodies (MAA)
Anti-Ro antibody (n=23)	3 (13%)
Anti-U1RNP (n=24)	2 (8%)
Anti-U2RNP (n=24)	1 (4%)
Negative all autoantibodies (n = 8)	1 (13%)
Negative MSA/MAA (n = 8)	4 (50%)
Medications***
Methotrexate	12 (46%)
Oral corticosteroids (i.e., prednisone)	8 (31%)
Hydroxychloroquine	7 (27%)
Intravenous immunoglobulin G (IVIG)	5 (19%)
Mycophenolate mofetil	3 (12%)
Abatacept	2 (8%)
Tacrolimus	1 (4%)
None	5 (19%)
* Race and ethnicity were self-reported. Mixed race denotes combinations of Caucasian, Native American, Alaskan Native, Black/African American, Native Hawaiian, and/or Pacific Islander.
** Out of 28 JDM probands, 8 had all 19 myositis autoantibodies assessed in the serum, 2 had no autoantibodies assessed, and the remainder had between 3 and 15 autoantibodies assessed. The numbers of JDM probands who had a given antibody checked are shown in brackets following each antibody name. The following antibodies are not included here because no JDM probands were positive for them: anti-Jo1, anti-PL-7, anti-PL-12, anti-EJ, anti-OJ, anti-KS, anti-SRP, anti-PM-Scl, and anti-Ku.
*** No medication data were available for 2 JDM probands; therefore, percentages reflect a denominator of 26. Total percentages do not add to 100% because several JDM probands were receiving multiple medications.

No competing interests reported.

KoesterChowetal2022.FigureS1.v220419.pdf
Figure S1. Characteristics of ASVs detected in both oral and fecal microbiomes.
KoesterChowetal2022.FigureS2.v220419.pdf
Figure S2. (A) Alpha diversity, as calculated using observed species and Chao1 metrics, and (B) genomic DNA concentrations of fecal vs oral samples.
KoesterChowetal2022.FigureS3.v220407.pdf
Figure S3. ASVs classified as unannotated bacteria, rank ordered by mean relative abundance across all stool samples in our cohort.
KoesterChowetal2022.FigureS4.v220407.pdf
Figure S4. Alpha diversity of stool samples in our study and 3 other studies employing the same 16S rRNA sequencing strategy.
KoesterChowetal2022.TableS1.v220420.xlsx
Table S1. Catalog of samples collected at SCH and NIH as part of the present study. Study participants from both sites were ultimately assigned a person ID of the same format ("JDM" followed by 4-digit family ID followed by relationship code; "A" = proband, "M" = mother, "F" = father, "R" = sibling). “X” indicates that a sample was collected. Dark gray filled boxes indicate samples that were not collected. Light gray boxes indicate samples that were collected but excluded from the final analysis due to low genomic DNA yield and/or poor-quality sequencing data.
KoesterChowetal2022.TableS2.v220430.xlsx
Table S2. DNA sequencing datasets generated in this study and downloaded from others' studies that were used in the analyses presented herein. (A) Bacterial V4-16S rRNA sequencing datasets generated from samples collected at SCH and NIH. (B) All ASVs seen in our dataset and their corresponding taxonomic classifications. (C) Downloaded bacterial V4-16S rRNA sequencing data. (D) ASVs corresponding to the most abundant unannotated bacteria in stool samples.

Download PDF

Version 1

posted

You are reading this latest preprint version

Familial clustering of dysbiotic oral and fecal microbiomes in juvenile dermatomyositis

Status:

Version 1

Abstract

Figures

Background

Methods

Results

Discussion

Conclusion

Abbreviations

Declarations

References

Tables

Additional Declarations

Supplementary Files

Status:

Version 1