AChR antibodies show a complex interaction with human skeletal muscle cells in a transcriptomic study

Abstract

Acetylcholine receptor (AChR) antibodies are the most important pathogenic marker in patients with myasthenia gravis (MG). The antibodies bind to AChRs on the postsynaptic membrane, and this leads to receptor degradation, destruction, or functional blocking with impaired signal at the neuromuscular junction. In this study, we have explored the effects of AChR antibodies binding to mature human myotubes with agrin-induced AChR clusters and pathways relevant for AChR degradation using bulk RNA sequencing. Protein-coding RNAs and lncRNAs were examined by RNA sequencing analysis. AChR antibodies induced marked changes of the transcriptomic profiles, with over 400 genes differentially expressed. Cholesterol metabolic processes and extracellular matrix organization gene sets were influenced and represent AChR-trafficking related pathways. Muscle contraction and cellular homeostasis gene sets were also affected, and independently of AChR trafficking. Furthermore, we found changes in a protein-coding RNA and lncRNA network, where expression of lncRNA MEG3 correlated closely with protein-coding genes for cellular homeostasis. We conclude that AChR antibodies induce an active response in human skeletal muscle cells which affects key intra- and extracellular pathways.

Figure 1

Skeletal muscle cell differentiation and AChR antibody binding to AChRs. (A) Three groups (Agrin−/Ab−, Agrin+/Ab− and Agrin+/Ab+) of differentiated skeletal muscle cells were examined. (B) Agrin and AChR antibodies affects AChR dynamics. (C) Differentiated and undifferentiated skeletal muscle cells in culture. (D) AChR antibodies bound to both clustered (agrin-treated) and unclustered (no agrin) AChRs on myotubes. AChRs were stained with the AChR antibody (mAb198). (E) Mature myotubes expressed myosin proteins.

Figure 2

Transcriptome profiles of differentiated skeletal muscle cells and the effect of AChR antibodies at widespread level. (A) The principle component analysis showed both a batch effect (i) and an AChR antibody treatment effect (ii). Within each batch, the antibody treated samples shift to the left for the first component (x-axis) compared with non-antibody treated groups (ii). Contr: Agrin−/Ab−; Agrin: Agrin+/Ab−; Ab: Agrin+/Ab+. (B) Transcriptome similarities between samples (i) and groups (ii) were calculated based on Pearson correlation coefficient. Incubation with AChR antibodies gave marked changes in the transcriptomic profiles. (C) The heatmap showed the DE genes between Agrin+/Ab− and Agrin+/Ab+ groups.

Figure 3

Gene set enrichment analysis of DE genes. (A) 29 significantly enriched gene sets for cellular components were identified. The top 15 sets are listed (i) and all 29 illustrated by a semantic similarity-based scatter plot (ii). Enriched gene sets included extracellular matrix, actin cytoskeleton and the myosin gene family. (B) 225 significantly enriched gene sets for biological processes were identified. The top 15 sets are listed (i) and the top 30 illustrated by a semantic similarity-based scatter plot (ii). Enriched gene sets included rhythmic processes, and cholesterol and lipid metabolic processes. (C) DE genes reflecting extracellular matrix (i), actin cytoskeleton (ii), cholesterol metabolic processes (iii), and circadian rhythms (iv). ***Padj < 0.001, **Padj < 0.01, *Padj < 0.05.

Figure 4

Short time series analysis in three consecutive groups of differentiated skeletal muscle cells. Contr: Agrin−/Ab−; Agrin: Agrin+/Ab−; Ab: Agrin+/Ab+. Ten clusters were generated using the K-means clustering algorithm based on the consecutive expression pattern of group Agrin−/Ab−, Agrin+/Ab−, and Agrin+/Ab+. Cluster 1–4 (in light blue) had opposite trends in group Agrin+/Ab− (versus Agrin−/Ab−) and group Agrin+/Ab+ (versus Agrin+/Ab−), and were classified as AChR trafficking related clusters; Cluster 5–7 (in violet) had a deviated trend only in Agrin−/Ab− group when compared to the other two groups, and were classified as muscle physiology pathway associated clusters. Cluster 8–10 (in green) were the remaining clusters without a distinct relation to AChR antibodies.

Figure 5

AChR antibodies induced changes of AChR trafficking and muscle physiology related pathways. (A, B) 51 genes were identified by intersecting between the AChR trafficking related gene clusters and the DE protein-coding genes (A). They were enriched for cholesterol metabolic processes, and glycosaminoglycan binding and transcription factor activity (B). (C,D) 94 genes were identified by intersecting between the muscle physiology associated clusters and DE protein-coding genes (C). They were enriched for cellular homeostasis (rhythmic process, cellular response to wound and lipid, etc.) (D-i) and muscle contraction related pathways (D-i,ii).

Figure 6

LncRNA and protein-coding RNA co-expression network. (A) A co-expression network was built based on the correlation efficient between all DE protein-coding RNAs and lncRNAs. LncRNA MEG3, SNHG3, and RP11-184M15.1 were correlated with several protein-coding RNA groups. (B,C) Gene set enrichment analysis of lncRNA correlated with protein-coding RNAs. 29 enriched gene sets for MEG3 correlated protein-coding RNAs, and one gene set (circadian rhythm) for SNHG3 correlated protein-coding RNAs were identified (B). The 29 enriched gene sets for MEG3 correlated protein-coding mRNAs were illustrated by a similarity-based scatter plot (C). The enriched gene sets for lncRNA MEG3 included circadian rhythm, response to biotic stimuli, and positive response to external stimuli. (D) 19 out of 46 MEG3 correlated protein-coding RNAs belonged to cellular homeostasis-related gene sets.

Figure 7

RT-qPCR validation for RNA sequencing results. Six genes (ACTA2, ACTC, ACTG1, ARNTL, PER3, NR1D2) from relevant pathways (actin cytoskeleton, circadian rhythm), and one lncRNA (MEG3) were validated.

Figure 8

AChR antibody-induced effects on myotubes expressing AChR clusters. 1. Cholesterol is involved in AChR internalization, modulation and recycling. 2. The dynamic process of actin polymerization and turnover is involved in AChR internalization. 3. Multiple myosin family genes and muscle contraction pathway genes were differentially expressed. 4. AChR antibodies induced an active response of extracellular matrix molecules. 5. Dysregulation of cellular homeostasis genes correlated closely with changes in lncRNA MEG3 gene expression.