Skip to main content
NCBI home page
Rheumatology (Oxford, England) logoLink to Rheumatology (Oxford, England)
. 2024 Feb 5;63(SI2):SI240–SI248. doi: 10.1093/rheumatology/keae082

Genetic changes from type I interferons and JAK inhibitors: clues to drivers of juvenile dermatomyositis

Lauren T Covert 1,✉, Joseph A Prinz 2, Devjanee Swain-Lenz 3,4, Jeffrey Dvergsten 5, George A Truskey 6
PMCID: PMC11381683  PMID: 38317053

Abstract

Objective

To better understand the pathogenesis of juvenile dermatomyositis (JDM), we examined the effect of the cytokines type I interferons (IFN I) and JAK inhibitor drugs (JAKi) on gene expression in bioengineered pediatric skeletal muscle.

Methods

Myoblasts from three healthy pediatric donors were used to create three-dimensional skeletal muscle units termed myobundles. Myobundles were treated with IFN I, either IFNα or IFNβ. A subset of IFNβ-exposed myobundles was treated with JAKi tofacitinib or baricitinib. RNA sequencing analysis was performed on all myobundles.

Results

Seventy-six myobundles were analysed. Principal component analysis showed donor-specific clusters of gene expression across IFNα and IFNβ-exposed myobundles in a dose-dependent manner. Both cytokines upregulated interferon response and proinflammatory genes; however, IFNβ led to more significant upregulation. Key downregulated pathways involved oxidative phosphorylation, fatty acid metabolism and myogenesis genes. Addition of tofacitinib or baricitinib moderated the gene expression induced by IFNβ, with partial reversal of upregulated inflammatory and downregulated myogenesis pathways. Baricitinib altered genetic profiles more than tofacitinib.

Conclusion

IFNβ leads to more pro-inflammatory gene upregulation than IFNα, correlating to greater decrease in contractile protein gene expression and reduced contractile force. JAK inhibitors, baricitinib more so than tofacitinib, partially reverse IFN I-induced genetic changes. Increased IFN I exposure in healthy bioengineered skeletal muscle leads to IFN-inducible gene expression, inflammatory pathway enrichment, and myogenesis gene downregulation, consistent with what is observed in JDM.

Keywords: juvenile dermatomyositis, dermatomyositis, interferon, tissue engineering, gene expression, pediatric rheumatology, muscle biology

Rheumatology key messages.

  • IFNβ leads to a greater pro-inflammatory gene response than IFNα in pediatric skeletal muscle.

  • Type I interferons downregulate oxidative phosphorylation, myogenesis, and contractile protein gene expression in skeletal muscle.

  • In a model of juvenile dermatomyositis, JAK inhibitors partially reverse IFNβ-induced genetic changes.

Introduction

Juvenile dermatomyositis (JDM) is characterized by a type I interferon (IFN I) gene signature in blood, skin and muscle [1–3]. Because type I interferons, including IFNα and IFNβ, activate the Janus kinase (JAK)-signal transducer and transcription (STAT) pathway, JAK inhibitors (JAKi) such as baricitinib and tofacitinib have therapeutic potential for JDM [4–7]. However, the mechanistic role of IFN I in skeletal muscle dysfunction is not understood. Furthermore, the rarity of JDM and difficulty obtaining routine muscle biopsies from patients with JDM make uncovering IFN I’s role in muscle weakness challenging.

Using an in vitro pediatric-derived skeletal muscle model termed myobundles, we have shown that IFNβ, not IFNα, leads to reduced skeletal muscle contractile force [8]. Both types of IFN I led to upregulation of MHC class I and donor-specific expression of myositis-specific autoantigens in myobundles [8]. Our aim in the current study is to analyse how IFN I and JAKi influence healthy pediatric skeletal muscle gene expression. Using bulk mRNA sequencing, we analysed genetic alterations after IFN I exposure with and without JAKi in myobundles. We hypothesized that myobundles treated with IFN I would have transcriptomic profiles consistent with those seen in JDM muscle biopsies: ones with upregulation of IFN-stimulated and proinflammatory genes [2, 9, 10]. We also expected that JAKi treatment would normalize IFN I-induced change in gene expression.

Materials and methods

Human primary myoblast isolation and culture

Human skeletal muscle samples were obtained from three healthy pediatric donors as surgical waste during operative treatment for orthopedic non-muscle diagnoses. Sample collection occurred under Duke University IRB approved protocols. Primary myoblasts were isolated, expanded and cryopreserved according to previously described methods [11, 12]. Myoblasts were thawed, plated on matrigel-coated flasks and grown for up to 1 week in human growth media (hGM) until they reached 80% confluency.

Human myobundle fabrication

The protocol for myobundle fabrication has been previously described [8, 12]. Briefly, each myobundle consisted of myoblasts (7.5 × 105 cells) mixed with fibrinogen, matrigel, thrombin and hGM. This solution polymerized into myobundles within polydimethylsiloxane (PDMS) molds between beams of a nylon frame. Myobundles were cultured for 4 days in hGM containing 1.5 mg/ml 6-aminocaproic acid (ACA) prior to removal from the molds and differentiation over 7 days in differentiation media (DM).

Treatment with type I interferon and JAK inhibitors

Mature myobundles were treated with one or a combination of the following reagents added to DM exchanged every other day: IFNα (Sigma-Aldrich, St Louis, MO, USA; Interferon-alpha 2A human, 10 µg/ml in 0.1% BSA in PBS), IFNβ (PeproTech, Cranbury, NJ, USA; 10 µg/ml in 0.1% BSA in PBS), tofacitinib (Millipore Sigma, Burlington, MA, USA; suspended in H2O), and/or baricitinib (Sigma-Aldrich; suspended in DMSO 0.01%). Myobundles were treated with IFNα or IFNβ diluted to 0 (control), 5, 10 or 20 ng/ml for 7 days then flash frozen in liquid nitrogen on day 18. There is a paucity of literature on the concentration of IFN I found in muscle of JDM patients; this is likely due to technical and biological limitations, such as the cytokines’ nonspecific immunoreactivity, transient expression and presence at low concentrations [3, 13]. Concentrations of IFN I used in this study were determined based on initial dose-dependent responses observed in contractile force production in myobundles [8] and informed by concentrations of IFNγ used in a prior myobundle study, in which up to 20 ng/ml was used [14].

Because decreased force production and slowed twitch kinetics were seen only in IFNβ-exposed myobundles [8], a subset of myobundles exposed to IFNβ, not IFNα, received additional treatment with JAK inhibitors. Clinically relevant, non-toxic concentrations of tofacitinib and baricitinib were determined based on previous work in myobundles and available pharmacodynamic data [15–17]. Final dose of tofacitinib and duration of JAKi treatment was determined based on preliminary studies previously described [8]. To test the effect of JAK inhibitors, myobundles were treated with IFNβ 0 (control) or 20 ng/ml every other day on days 11 through 21 after myobundle fabrication. A subset of IFNβ-exposed myobundles were treated with concurrent tofacitinib 1 µM or baricitinib 500 nM on days 15 through 21. Controls were treated with DMSO 0.01% on days 15 through 21. On day 21, these myobundles were flash frozen. Experimental schema is illustrated in Supplementary Fig. S1, available at Rheumatology online.

RNA sequencing and data analysis

RNA extraction from flash frozen myobundles was performed using an RNeasy Mini Kit (Qiagen 74104, Hilden, Germany). Samples were stored at –80°C until assayed. RNA Integrity Number (RIN) was between 8.4 and 10 (Supplementary Table S1, available at Rheumatology online.). RNA libraries were built by the Duke Sequencing and Genomics Technologies Core Facility using stranded mRNA kit (Roche: 07962207001, Indianapolis, IN, USA), and libraries were sequenced as 50 pair-ended base pairs with a NovaSeq 6000 S-Prime. The average number of reads mapped was 47.61 M reads per sample.

RNA-seq data was processed using the fastp toolkit [18] to trim low-quality bases and sequencing adapters from the 3’ end of reads, then mapped to GRCh38 (downloaded from Ensembl, version 106) [19] using the STAR RNA-seq alignment tool [20] and reads aligning to a single genomic location were summarized across genes. For genes having an overlap of at least 10 reads, gene counts were normalized and differential expression was carried out using the DESeq2 [21] Bioconductor [22] package implemented for the R programming environment. Consistent with the recommendation of the DESeq authors, independent filtering [23] was utilized prior to calculating adjusted P-values [24] and moderated log2 fold changes were derived using the ashr package [25]. Gene set enrichment analysis [26] was performed to identify gene ontology terms and pathways associated with altered gene expression for each of the comparisons performed. The data discussed in this publication have been deposited in NCBI’s Gene Expression Omnibus [27] and are accessible through GEO Series accession number GSE243339.

The IFN score of treated myobundles was calculated as the mean log2FC relative to untreated myobundles for the following IFN I activated genes: ISG15, IFI6, MX1, RSAD2, MX2, OAS1, IRF9, IFITM1, OAS3 and IFI35 [28]. Log2FC of these genes were compared across treatment groups using one-way analysis of variance with statistical significance of P ≤0.05.

Ethics

This study complied with the Declaration of Helsinki, had approval from Duke University IRB, and written informed consent was obtained from the participants (or their legally authorized representative).

Results

Effect of IFNα and IFNβ concentration on global genetic expression

The gene expression profiles of 76 myobundles were analysed. To observe general patterns in gene regulation, we performed Principal Component Analysis (PCA) of all samples across donors (Fig. 1). PC1 explains 34.27% of variance in the dataset and separates samples by treatment condition; this indicates that IFNα and IFNβ-exposed myobundles were distinct from controls. Of note, increasing IFN I concentration explains significant variance in PC1. Notably, IFNβ-exposed myobundles across all donors exhibit a greater rightward shift in PC1 than IFNα-exposed myobundles regardless of specific donor response.

Figure 1.

Figure 1.

Open in a new tab

Principal component analysis of myobundles. Each point represents an individual myobundle. Clusters of similar gene expression are observed based on donor and treatment. Dose response regardless of donor is evident in a rightward shift in PC1 with IFNβ with greater effect than IFNα. JAK inhibitor treatment with tofacitinib (tofa) or baricitnib (bari) in IFNβ-exposed myobundles are more similar to myobundles treated with IFNα alone. C, untreated control; α, IFNα; β, IFNβ

In contrast to concentration-dependent separation along PC1, PC2 explains 15.87% of variance and separates variation by donor. As expected, myobundles derived from the same donor in similar treatment groups cluster together. There are clusters of similar gene expression across IFNα-exposed myobundles, which is distinct from that of IFNβ-exposed myobundles. Additionally, IFNβ-exposed myobundles treated with JAKi have gene expression profiles more similar to myobundles treated with IFNα alone.

There are clear distinctions in the top 100 differentially expressed genes between myobundles treated with either IFN I relative to controls (Fig. 2A). For most genes, IFNα produced no change or a modest decrease in gene expression relative to control. In contrast to IFNα, IFNβ treatment led to more downregulation and upregulation of different sets of genes. Addition of tofacitinib or baricitinib moderated the gene expression induced by IFNβ (Fig. 2B).

Figure 2.

Figure 2.

Open in a new tab

Heatmap of top 100 differentially expressed genes from IFNα and IFNβ exposure. (A) and by JAK inhibitor treatment with either tofacitnib (tofa) or baricitinib (bari) after IFNβ 20 ng/mL exposure (B). α, IFNα; β, IFNβ

Gene set enrichment analysis

To understand large gene regulatory changes from IFNα and IFNβ, we conducted Gene Set Enrichment Analysis (GSEA) of Hallmark Pathways. Compared with control myobundles, IFNα-exposed myobundles had 8/50 pathways (16%) significantly enriched (Normalized Enrichment Score, NES = 1.49–3.00; Padj = 5.59 × 10−39–0.03); IFNβ-exposed myobundles had 15/50 pathways (30%) enriched (NES = 1.37–2.89; Padj = 2.12 × 10−50–0.05). Of these pathways, seven were significantly enriched for myobundles treated with both cytokines, consisting of: IFNα, IFNγ, and inflammatory response, complement, allograft rejection-related, IL2/STAT5 and IL6/JAK/STAT3 pathways (Fig. 3).

Figure 3.

Figure 3.

Open in a new tab

Gene set enrichment analysis. GSEA of Hallmark pathways in myobundles exposed to any concentration of IFNα (top) or IFNβ (bottom). Interferon response and inflammatory pathways were highly upregulated across all myobundles. Pathways involved in oxidative phosphorylation, myogenesis and cell proliferation (MYC target genes) were downregulated across all myobundles

Downregulated pathways common to both IFNα and IFNβ included oxidative phosphorylation, myogenesis and MYC target pathways (Fig. 3). The most downregulated pathways included oxidative phosphorylation for IFNα-exposed myobundles (NES = -2.68; Padj = 4.17 × 10−17) and myogenesis genes for IFNβ-exposed myobundles (NES = -2.58; Padj = 2.69 × 10−18). IFNα uniquely downregulated the transcription factor E2F target and fatty acid metabolism pathways with Padj < 0.05. Finally, when GSEA was performed on the Log2FC IFNβ/IFNα, 15 pathways were statistically significantly enriched (NES > 0; Padj ≤ 0.05) with seven of those pathways exclusively enriched in IFNβ-exposed samples and seven enriched in both IFNα and IFNβ treated samples (Supplementary Fig. S2, available at Rheumatology online.).

Differential expression of individual genes

To examine specific differential expression on an individual gene level, we identified the most differentially expressed genes and their associated pathway, if applicable, in myobundles exposed to 20 ng/ml IFNα or IFNβ vs untreated controls. The most upregulated individual gene across myobundles exposed to either IFNα or IFNβ was HLA complex P5 (HCP5), a long non-coding RNA (lncRNA), thought to regulate adaptive and immune responses [29] (Supplementary Table S2, available at Rheumatology online).

There were more significantly upregulated and downregulated genes in response to any IFNβ concentration than IFNα (log2FC ≥ and ≤ 2, Padj ≤ 0.05, respectively). There was also significant overlap between IFNα and IFNβ-induced upregulated genes (Fig. 4A). A similar trend was observed for downregulated genes (Fig. 4B). Interferon-related genes that are dysregulated in JDM patient whole blood transcriptomic data [30] were similarly altered in myobundles exposed to either IFN I, yet IFNβ led to greater change than IFNα.

Figure 4.

Figure 4.

Open in a new tab

Venn diagrams of number of up and downregulated genes after IFN I exposure. (A) Number of significantly upregulated genes (log2FC ≥ 2, padj ≤ 0.05) in healthy donor-derived myobundles after any IFNα or IFNβ exposure and when directly comparing log2FC IFNβ/IFNα. (B) Number of significantly downregulated genes (log2FC ≤ –2, padj ≤ 0.05) in healthy donor-derived myobundles after any IFNα or IFNβ exposure and when directly comparing log2FC IFNβ/IFNα

To investigate why IFNβ-exposed myobundles have more decreased contractile force than IFNα-exposed myobundles [8], we examined the expression of specific genes involved in muscle growth, structure, contraction and calcium release. Compared with IFNα, IFNβ led to greater downregulation of RYR1, a gene involved in sarcoplasmic reticular calcium release in muscle contraction (Supplementary Table S3, available at Rheumatology online.). IFNβ also led to further decreased expression of other protein-encoding genes involved in muscle contraction and regeneration, including MYH2, MYH3, MYH8 and TNNI2 (log2FC < –1, Padj ≤ 0.05) (Supplementary Table S3, available at Rheumatology online). MYH2, MYH3 and MYH8 encode for myosin heavy chains, which generate mechanical force in skeletal muscle contraction [31]. MYH2 encodes a myosin heavy chain that is highly expressed in fast type IIx muscle fibers [31], whereas MYH3 and MYH8 encode embryonic and neonatal myosin heavy chains, respectively; they are seen in normally developing muscle fibers and regenerating fibers of pathologic skeletal muscle [32]. Finally, TNNI2 encodes for a fast-twitch skeletal muscle protein involved in calcium-dependent regulation of striated muscle contraction [33].

Genetic alterations after JAK inhibitor treatment

To understand the effect of JAK inhibition, we conducted a differential gene expression analysis between IFNβ-exposed myobundles with and without tofacitinib or baricitinib treatment. Differential gene expression analysis demonstrated that baricitinib led to greater number of genes significantly upregulated (59) compared with tofacitinib [34] (log2FC ≥ 2, Padj ≤ 0.05) with 28 of those genes upregulated by both drugs. Similarly, baricitinib led to a greater number of downregulated genes (135) than tofacitinib (107) with 84 of those genes overlapping (log2FC ≤ –2, Padj ≤ 0.05).

JAK inhibitor treatment of IFNβ-exposed myobundles led to significantly less pathway enrichment than those treated with IFNβ alone. Baricitinib resulted in 2/50 (4%) pathway enrichment (NES = 1.58–2.45; Padj = 4.12 × 10−16–0.002), which consisted of myogenesis and oxidative phosphorylation pathways. Tofacitinib led to 1/50 (2%) pathway enrichment (NES = 2.47; Padj = 7.98 × 1 0−13), which included only the myogenesis pathway. Both baricitinib and tofacitinib significantly downregulated 13 similar pathways, which included IFNα, IFNγ, TNFα, inflammatory, JAK-STAT, hypoxia-related, apoptosis and complement signalling pathways. Log2FC tofacitinib/baricitinib comparison revealed that only one pathway, the angiogenesis pathway, was significantly more downregulated by tofacitinib (NES = –2.42; Padj = 0.02).

Interferon score and its relationship with key genes

To further characterize the effect of IFNα and IFNβ at various concentrations on gene expression, we examined the relationship between affected genes and IFN score, calculated as the mean log2FC of 10 IFN I-inducible genes (ISG15, IFI6, MX1, RSAD2, MX2, OAS1, IRF9, IFITM1, OAS3 and IFI35) [28]. When the log2FC for these 10 IFN I-inducible genes are plotted across treatment conditions, there is a statistically significant, non-dose-related upregulation induced by IFNβ compared with untreated myobundles (Fig. 5). IFNα exposure led to less upregulation than IFNβ exposure, yet still a significant increase in log2FC relative to untreated myobundles. Thus, IFN score was higher in all IFNβ-exposed myobundles than IFNα-exposed myobundles. Because there was no dose-dependent effect of IFNβ on IFN I-induced gene expression at the tested concentrations, we treated only myobundles exposed to 20 ng/ml IFNβ with JAK inhibitor drugs to examine JAKi effects. Treatment of IFNβ-exposed myobundles with either tofacitinib or baricitinib caused a statistically significant decrease in most IFN I pathway genes relative to untreated conditions (Fig. 5).

Figure 5.

Figure 5.

Open in a new tab

Association between interferon treatment and IFN I-inducible genes. IFN pathway genes log2 fold change associated with IFN I treatment. α or β refer to IFNα or IFNβ treatment, respectively, at 5, 10 or 20 ng/mL, Tofa represents tofacitinib treatment and Bari represents baricitinib treatment of myobundles exposed to IFNβ 20 ng/mL. Data were analysed using a one-way analysis of variance. Treatments with similar capital letters are not significantly different

Next, we analysed the role that IFN I-induced gene alteration may play in myogenesis, functional contraction and oxidative phosphorylation. IFN scores were inversely related to changes in expression of leading-edge genes within the myogenesis hallmark pathway (Fig. 6A), skeletal muscle contractile protein genes and the gene for the extracellular matrix protein titin (TTN) (Fig. 6B). IFN scores were also inversely related to the expression of mitochondrial genes COX17, ATP5ME and COX7A2, which are leading-edge genes of the oxidative phosphorylation pathway (Fig. 6C).

Figure 6.

Figure 6.

Open in a new tab

Correlation between IFN score and expression of key genes. Association of IFN score, calculated as the mean log2 fold change of 10 IFN I-induced genes as referenced [28], with log2 fold change gene expression of leading-edge myogenesis-related genes (A), contractile protein genes and the titin gene (TTN) (B) and select mitochondrial gene expression (C) across treatment groups and donors

Discussion

In evaluation of IFN I and JAKi influence on gene expression in skeletal muscle, we found that when exposed to IFN I, healthy pediatric-derived bioengineered muscle has expected positive IFN scores and upregulation of pathways involving IFNα response, IFNγ response, inflammatory, JAK-STAT and complement signalling genes. Compared with untreated myobundles, IFNβ leads to more profound genetic alteration than IFNα at the individual gene and pathway level. Regardless of donor, JAKi partially normalizes IFNβ-induced downregulation of myogenesis and oxidative phosphorylation pathways, shifting gene expression towards that of myobundles treated with IFNα alone.

Interestingly, both types of IFN I led to significant upregulation of a lncRNA, HCP5. Microarray analysis of DM muscle biopsies has shown that while dysregulated IFN I signalling is common among all DM subgroups, there are differentially expressed lncRNAs across clinical subtypes [34]. The current study suggests that lncRNAs play a critical role in the immune system by regulating the expression of IFN I-inducible proteins. lncRNAs possibly contribute to DM pathogenesis in addition to established IFN I upregulation. Our result of upregulated lncRNA HCP5 insinuates that lncRNA dysregulation may be a downstream effect of IFN I.

A key downregulated pathway across myobundles exposed to either IFNα or IFNβ was oxidative phosphorylation. This observation is consistent with prior studies establishing mitochondrial respiratory chain dysfunction in JDM. DM and JDM muscle biopsies exhibit depletion of mitochondrial DNA and deficiencies of complex I and IV [35], and patients with DM have demonstrated reduced oxidative phosphorylation and rate of ATP production by quantitative MRI and phosphorus magnetic resonance spectroscopy [36]. Furthermore, studies of DM and JDM patient blood have revealed dysregulated transcripts of mitochondrial and oxidative pathways as well as elevated mitochondrial DNA and anti-mitochondrial antibodies, which correlate with interferon-stimulated gene expression [37–39]. Our findings in IFN I-exposed bioengineered muscle corroborate with those seen in JDM and DM patients.

In our study, the JAK inhibitor baricitinib enriched the oxidative phosphorylation pathway in IFNβ-exposed myobundles more significantly than tofacitinib relative to untreated controls. Baricitinib also led to a greater number of up and downregulated genes than tofacitinib did. While both drugs led to a shift towards the gene profile of myobundles exposed to IFNα alone, results indicate that baricitinib may be more effective at decreasing IFN I-induced changes in myositis. This difference may be due to drug selectivity and potency of JAK inhibition between tofacitinib and baricitinib. Tofacitinib is a selective inhibitor of JAK1 and JAK3 with limited inhibition of JAK2 and tyrosine kinase (TYK) 2, whereas baricitinib selectivity inhibits JAK1 and JAK2 with moderate inhibition of TYK2 and limited inhibition of JAK3 [40]. Further, in vitro high-throughput screening of IFN I inhibition in skeletal myoblasts shows that baricitinib was more potent than tofacitinib [41].

While most pathways were more affected by IFNβ than IFNα, fatty acid metabolism was downregulated to a greater extent after IFNα relative to IFNβ treatment. Fatty acids and lipid derivatives have important regulatory roles in skeletal muscle mass and function, and chronic lipidomic abnormalities can lead to muscle impairment [42]. Adult DM patients have altered lipidomic serum profiles [43], and a metabolomic study of DM patient plasma has shown altered fatty acid biosynthesis and beta-oxidation of very long chain fatty acids [44]. Further, metabolomic analyses of treatment naive JDM patient sera revealed greater concentrations of acyl carnitines, mainly long-chain acyl carnitines and ceramides, implicating dysregulation of mitochondrial fatty acid β-oxidation [45]. As our study isolates the effect of IFN I on muscle, our findings indicate that the observations seen in prior work may have been downstream effects of upregulated IFNα.

Both types of IFN I led to downregulation of myogenesis pathways. IFNβ led to more downregulation than IFNα. Decreased gene expression of myogenesis-related transcripts correlates to prior studies showing the effect of IFN I on human myoblasts and mouse models; IFN I decreases myoblast differentiation, myogenin expression and myotube area [46, 47]. Additionally, relative to healthy muscle stem cells, muscle stem cells from DM patients have been observed to have reduced proliferating capacities, less myotube formation and greater senescence [48].

Expectedly with downregulated myogenesis pathways in IFN I-exposed myobundles, there was also a statistically significant decrease in the expression of protein coding myosin heavy and light chain genes as well as contraction-related troponin I isoform and RYR1 genes. Compared with myobundles exposed to IFNα, myobundles treated with IFNβ had greater downregulation of fast twitch myofiber genes MYH2 and TNNI2 as well as genes related to regenerating muscle, such as MYH3 and MYH8. This decrease may indicate a shift from glycolytic metabolic pathways towards oxidative aerobic metabolism of slow twitch fibers, which would alter muscle fatigue. Findings are consistent with the greater reduction in contractile function seen in myobundles exposed to IFNβ compared with those exposed to IFNα [8]. Myogenesis pathway and contractile gene downregulation seen here further support the pathogenic role of IFN I in pediatric myositis, with IFNβ perhaps contributing more than IFNα in skeletal muscle contractile dysfunction. JAKi treatment of IFNβ-exposed myobundles led to partial reversal of downregulated myogenesis pathways, implicating a direct therapeutic effect of these novel drugs in JDM patients.

To our knowledge, this is the first transcriptomic study of functional skeletal muscle in response to type I interferons. While several studies have addressed human muscle cell growth, repair, and respiratory capacity in response to IFN I in 2D culture [46–48], this study demonstrates similar findings in a 3D functional model that details genetic changes in skeletal muscle in isolation of other factors including immune cell signalling and vascular compromise, both of which are implicated in JDM pathogenesis as well. Thus, the myobundle model helps to distinguish the response of the muscle itself from IFN I exposure. Additional strengths of this study include: (i) use of human pediatric donor tissue; (ii) analysis of technical replicates with demonstrated reproducibility in gene expression; (iii) use of bulk RNA-seq to gain knowledge of novel transcripts including long non-coding RNAs; and (iv) analysis of key pathways with aberrant expression in states of upregulated IFN I.

Limitations of this study include the use of non-diseased skeletal muscle tissue from a small number of healthy pediatric donors. While the purpose of this study was to investigate IFN I effect on gene expression in pediatric muscle using a bioengineered model, we acknowledge that comparison with myobundles derived from diseased tissue from children with JDM would be beneficial, and future study of JDM patient-derived myobundles is planned. However, JDM is a complex, heterogeneous disease that also can be clinically amyopathic, which limits the generalizability of a muscle model system to all patients with JDM regardless. Additionally, this model includes only muscle cells without an immune or endothelial cell component, which will be critical to systematically evaluate in the future. We also recognize that cytokines do not work in isolation in autoimmune disease pathogenesis; this study only addresses the isolated effects of individual type I interferons on skeletal muscle gene expression. Future study is planned to examine genetic expression of myobundles after exposure to a combination of IFNα and IFNβ as well as other proinflammatory cytokines implicated in JDM.

In summary, our results demonstrate that IFNβ leads to a greater pro-inflammatory gene response than IFNα in pediatric skeletal muscle. Downregulated oxidative phosphorylation, myogenesis, fatty acid metabolism and contractile protein gene expression are implicated as downstream effects of IFN I and may be important in JDM pathology. These genetic changes are partially reversed by JAK inhibitors, with baricitinib leading to greater genetic normalization than tofacitinib. Our findings show that analysis of differential gene expression in bioengineered pediatric skeletal muscle can inform our knowledge of JDM pathogenesis and therapeutic effects.

Supplementary Material

keae082_Supplementary_Data
keae082_supplementary_data.zip (227KB, zip)

Acknowledgements

The authors wish to acknowledge Hailee Patel, MS for her contribution to the experimental results of this work and the Cure JM Foundation for ongoing financial support. Support was also provided through the local efforts of the Duke Children’s Office of Development and its Children’s Miracle Network Hospitals fundraising corporate partnerships and programs.

Contributor Information

Lauren T Covert, Department of Pediatrics, Duke University Health System, Durham, NC, USA.

Joseph A Prinz, Sequencing and Genomics Technologies Core Facility, School of Medicine, Duke University, Durham, NC, USA.

Devjanee Swain-Lenz, Sequencing and Genomics Technologies Core Facility, School of Medicine, Duke University, Durham, NC, USA; Department of Molecular Genetics and Microbiology, School of Medicine, Duke University, Durham, NC, USA.

Jeffrey Dvergsten, Department of Pediatrics, Duke University Health System, Durham, NC, USA.

George A Truskey, Department of Biomedical Engineering, Duke University, Durham, NC, USA.

Supplementary material

Supplementary material is available at Rheumatology online.

Data availability

The data discussed in this publication have been deposited in NCBI’s Gene Expression Omnibus [27] and are accessible through GEO Series accession number GSE243339.

Funding

This work was supported by the local efforts of the Duke Children’s Office of Development and its Children’s Miracle Network Hospitals fundraising corporate partnerships and programs [L.T.C.], the Cure JM Foundation [63022 to J.D.], the National Center for Advancing Translational Sciences [UH3TR002142 to G.A.T.], and the NIH Eunice Kennedy Shriver National Institute of Child Health & Development T32 [HD094671 to L.T.C].

Disclosure statement: The authors have declared no conflicts of interest.

References

  • 1. Turnier JL, Pachman LM, Lowe L  et al.  Comparison of lesional juvenile myositis and lupus skin reveals overlapping yet unique disease pathophysiology. Arthritis Rheumatol  2021;73:1062–72. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 2. Moneta GM, Pires Marafon D, Marasco E  et al.  Muscle expression of type I and type II interferons is increased in juvenile dermatomyositis and related to clinical and histologic features. Arthritis Rheumatol  2019;71:1011–21. [DOI] [PubMed] [Google Scholar]
  • 3. Baechler EC, Bilgic H, Reed AM.  Type I interferon pathway in adult and juvenile dermatomyositis. Arthritis Res Ther  2011;13:249. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4. Huang B, Wang X, Niu Y  et al.  Long-term follow-up of Janus-kinase inhibitor and novel active disease biomarker in juvenile dermatomyositis. Rheumatology (Oxford)  2023;62:1227–37. [DOI] [PubMed] [Google Scholar]
  • 5. Yu Z, Wang L, Quan M, Zhang T, Song H.  Successful management with Janus kinase inhibitor tofacitinib in refractory juvenile dermatomyositis: a pilot study and literature review. Rheumatology (Oxford)  2021;60:1700–7. [DOI] [PubMed] [Google Scholar]
  • 6. Sabbagh S, Almeida de Jesus A, Hwang S  et al.  Treatment of anti-MDA5 autoantibody-positive juvenile dermatomyositis using tofacitinib. Brain  2019;142:e59. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 7. Papadopoulou C, Hong Y, Omoyinmi E, Brogan PA, Eleftheriou D.  Janus kinase 1/2 inhibition with baricitinib in the treatment of juvenile dermatomyositis. Brain  2019;142:e8. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 8. Covert LT, Patel H, Osman A  et al.  Effect of type I interferon on engineered pediatric skeletal muscle: a promising model for juvenile dermatomyositis. Rheumatology (Oxford)  2023;63:209–17. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 9. Tezak Z, Hoffman EP, Lutz JL  et al.  Gene expression profiling in DQA10501+ children with untreated dermatomyositis: a novel model of pathogenesis. J Immunol  2002;168:4154–63. [DOI] [PubMed] [Google Scholar]
  • 10. Musumeci G, Castrogiovanni P, Barbagallo I  et al.  Expression of the OAS gene family is highly modulated in subjects affected by juvenile dermatomyositis, resembling an immune response to a dsRNA virus infection. Int J Mol Sci  2018;19:2786. doi: 10.3390/ijms19092786. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11. Davis BNJ, Santoso JW, Walker MJ  et al.  Modeling the effect of TNF-alpha upon drug-induced toxicity in human, tissue-engineered myobundles. Ann Biomed Eng  2019;47:1596–610. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12. Madden L, Juhas M, Kraus WE, Truskey GA, Bursac N.  Bioengineered human myobundles mimic clinical responses of skeletal muscle to drugs. eLife  2015;4:e04885. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 13. Greenberg SA.  Type 1 interferons and myositis. Arthritis Res Ther  2010;12(Suppl 1):S4. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14. Zhan RZ, Rao L, Chen Z, Strash N, Bursac N.  Loss of sarcomeric proteins via upregulation of JAK/STAT signaling underlies interferon-γ-induced contractile deficit in engineered human myocardium. Acta Biomater  2021;126:144–53. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15. Chen Z, Li B, Zhan RZ, Rao L, Bursac N.  Exercise mimetics and JAK inhibition attenuate IFN-γ-induced wasting in engineered human skeletal muscle. Sci Adv  2021;7:eabd9502. doi: 10.1126/sciadv.abd9502. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 16. Dowty ME, Jesson MI, Ghosh S  et al.  Preclinical to clinical translation of tofacitinib, a Janus kinase inhibitor, in rheumatoid arthritis. J Pharmacol Exp Ther  2014;348:165–73. [DOI] [PubMed] [Google Scholar]
  • 17. Shi JG, Chen X, Lee F  et al.  The pharmacokinetics, pharmacodynamics, and safety of baricitinib, an oral JAK 1/2 inhibitor, in healthy volunteers. J Clin Pharmacol  2014;54:1354–61. [DOI] [PubMed] [Google Scholar]
  • 18. Chen S, Zhou Y, Chen Y, Gu J.  fastp: an ultra-fast all-in-one FASTQ preprocessor. Bioinformatics  2018;34:i884–90. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 19. Kersey PJ, Staines DM, Lawson D  et al.  Ensembl Genomes: an integrative resource for genome-scale data from non-vertebrate species. Nucleic Acids Res  2012;40:D91–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20. Dobin A, Davis CA, Schlesinger F  et al.  STAR: ultrafast universal RNA-seq aligner. Bioinformatics  2013;29:15–21. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21. Love MI, Huber W, Anders S.  Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2. Genome Biol  2014;15:550. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 22. Huber W, Carey VJ, Gentleman R  et al.  Orchestrating high-throughput genomic analysis with Bioconductor. Nat Methods  2015;12:115–21. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23. Ignatiadis N, Klaus B, Zaugg JB, Huber W.  Data-driven hypothesis weighting increases detection power in genome-scale multiple testing. Nat Methods  2016;13:577–80. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 24. Benjamini Y, Hochberg Y.  Controlling the false discovery rate: a practical and powerful approach to multiple testing. J R Stat Soc Ser B (Methodol)  1995;57:289–300. [Google Scholar]
  • 25. Stephens M.  False discovery rates: a new deal. Biostatistics  2017;18:275–94. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 26. Subramanian A, Tamayo P, Mootha VK  et al.  Gene set enrichment analysis: a knowledge-based approach for interpreting genome-wide expression profiles. Proc Natl Acad Sci USA  2005;102:15545–50. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 27. Edgar R, Domrachev M, Lash AE.  Gene Expression Omnibus: NCBI gene expression and hybridization array data repository. Nucleic Acids Res  2002;30:207–10. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 28. Pinal-Fernandez I, Casal-Dominguez M, Derfoul A  et al.  Identification of distinctive interferon gene signatures in different types of myositis. Neurology  2019;93:e1193–204. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29. Kulski JK.  Long noncoding RNA HCP5, a hybrid HLA class I endogenous retroviral gene: structure, expression, and disease associations. Cells  2019;8:480. doi: 10.3390/cells8050480. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 30. Kim H, Gunter-Rahman F, McGrath JA  et al.  Expression of interferon-regulated genes in juvenile dermatomyositis versus Mendelian autoinflammatory interferonopathies. Arthritis Res Ther  2020;22:69. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 31. Stuart CA, Stone WL, Howell ME  et al.  Myosin content of individual human muscle fibers isolated by laser capture microdissection. Am J Physiol Cell Physiol  2016;310:C381–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 32. Schiaffino S, Rossi AC, Smerdu V, Leinwand LA, Reggiani C.  Developmental myosins: expression patterns and functional significance. Skeletal Muscle  2015;5:22. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 33. Sheng JJ, Jin JP.  TNNI1, TNNI2 and TNNI3: evolution, regulation, and protein structure-function relationships. Gene  2016;576:385–94. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 34. Peng QL, Zhang YM, Yang HB  et al.  Transcriptomic profiling of long non-coding RNAs in dermatomyositis by microarray analysis. Sci Rep  2016;6:32818. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 35. Hedberg-Oldfors C, Lindgren U, Visuttijai K  et al.  Respiratory chain dysfunction in perifascicular muscle fibres in patients with dermatomyositis is associated with mitochondrial DNA depletion. Neuropathol Appl Neurobiol  2022;48:e12841. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 36. Cea G, Bendahan D, Manners D  et al.  Reduced oxidative phosphorylation and proton efflux suggest reduced capillary blood supply in skeletal muscle of patients with dermatomyositis and polymyositis: a quantitative 31P-magnetic resonance spectroscopy and MRI study. Brain  2002;125:1635–45. [DOI] [PubMed] [Google Scholar]
  • 37. Baechler EC, Bauer JW, Slattery CA  et al.  An interferon signature in the peripheral blood of dermatomyositis patients is associated with disease activity. Mol Med  2007;13:59–68. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 38. Duvvuri B, Pachman LM, Hermanson P  et al.  Role of mitochondria in the myopathy of juvenile dermatomyositis and implications for skeletal muscle calcinosis. J Autoimmun  2023;138:103061. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 39. Wilkinson MGL, Moulding D, McDonnell TCR  et al.  Role of CD14+ monocyte-derived oxidised mitochondrial DNA in the inflammatory interferon type 1 signature in juvenile dermatomyositis. Ann Rheum Dis  2023;82:658–69. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 40. Miyazaki Y, Nakano K, Nakayamada S  et al.  Efficacy and safety of tofacitinib versus baricitinib in patients with rheumatoid arthritis in real clinical practice: analyses with propensity score-based inverse probability of treatment weighting. Ann Rheum Dis  2021;80:1130–6. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 41. Kinder TB, Dranchak PK, Inglese J.  High-Throughput Screening to Identify Inhibitors of the Type I Interferon-Major Histocompatibility Complex Class I Pathway in Skeletal Muscle. ACS Chem Biol  2020;15:1974–86. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 42. Funai K, Lodhi IJ, Spears LD  et al.  Skeletal muscle phospholipid metabolism regulates insulin sensitivity and contractile function. Diabetes  2016;65:358–70. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 43. Raouf J, Idborg H, Englund P  et al.  Targeted lipidomics analysis identified altered serum lipid profiles in patients with polymyositis and dermatomyositis. Arthritis Res Ther  2018;20:83. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 44. Zhao Q, Hu Q, Meng S  et al.  Metabolic profiling of patients with different idiopathic inflammatory myopathy subtypes reveals potential biomarkers in plasma. Clin Exp Med  2023;23:3417–29. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 45. Dvergsten JA, Reed AM, Landerman L  et al.  Metabolomics analysis identifies a lipidomic profile in treatment-naïve juvenile dermatomyositis patients vs healthy control subjects. Rheumatology (Oxford)  2022;61:1699–708. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 46. Franzi S, Salajegheh M, Nazareno R, Greenberg SA.  Type 1 interferons inhibit myotube formation independently of upregulation of interferon-stimulated gene 15. PLoS One  2013;8:e65362. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 47. Ladislau L, Suarez-Calvet X, Toquet S  et al.  JAK inhibitor improves type I interferon induced damage: proof of concept in dermatomyositis. Brain  2018;141:1609–21. [DOI] [PubMed] [Google Scholar]
  • 48. Gallay L, Fermon C, Lessard L  et al.  Involvement of Type I Interferon Signaling in Muscle Stem Cell Proliferation During Dermatomyositis. Neurology  2022;98:e2108–19. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

keae082_Supplementary_Data
keae082_supplementary_data.zip (227KB, zip)

Data Availability Statement

The data discussed in this publication have been deposited in NCBI’s Gene Expression Omnibus [27] and are accessible through GEO Series accession number GSE243339.

Articles from Rheumatology (Oxford, England) are provided here courtesy of Oxford University Press

RESOURCES