Published online Dec 28, 2017. doi: 10.3748/wjg.v23.i48.8489
Peer-review started: October 31, 2017
First decision: November 14, 2017
Revised: November 21, 2017
Accepted: November 28, 2017
Article in press: November 28, 2017
Published online: December 28, 2017
Currently, the pathogenesis of primary biliary cholangitis (PBC) is still unclear. The appearance and fluctuation of autoantibodies not only represent the disease status but also further our understanding of the pathogenesis of the disease. The specific diagnostic biomarker of PBC, anti-mitochondrial antibody (AMA), was confirmed to participate in injuring intrahepatic biliary epithelial cells (iBECs) and perpetuating the destructive process. Exploring novel autoantibodies for PBC will help unravel the underlying mechanism of pathogenesis.
AMA cannot be detected in 10% of PBC cases, which prompted us to determine whether other pathogenesis-associated and AMA-like autoantibodies exist in AMA-negative PBC patients. This may aid in decrypting the pathogenesis of PBC.
We sought to carry out an exploratory study using a novel proteomics approach that combines the enrichment of antigens from human BECs (HiBECs), label-free mass spectrometry, and bioinformatics to identify candidate autoantibodies for AMA-negative PBC. In this pilot study, we aimed to ascertain the feasibility of this experimental approach and obtain preliminary results that will guide future larger-scale and in-depth studies. In addition, the novel experimental roadmap established here can be applied in related studies in future.
Eighteen PBC patients were recruited from the Peking Union Medical College Hospital. They included nine AMA-positive and nine AMA-negative PBC cases, with pairs matched for age and sex. Sera were collected from patients and frozen at -80°C until use. Human iBECs (HiBECs) were subcultured and lysed to provide sufficient target antigens for candidate autoantibodies enriched from the sera of PBC patients. Cell lysates and enriched autoantibodies were mixed and incubated under various conditions. Candidate autoantigens were eluted and subsequently identified by label-free mass spectrometry. Student’s t-test was performed for comparison of differences between AMA-positive and -negative proteins, with P < 0.05 set as the significance cut-off. Enrichment analyses were conducted for the Gene Ontology (GO) biological processes (GOBP), cellular components (GOCC), and molecular functions (GOMF) categories. Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways were analyzed through the DAVID platform, using the complete UniProt human protein database as a background and an FDR < 0.05 as a cut-off. Heatmap analyses and visualization of significantly different proteins were conducted by complete linkage hierarchical clustering. Significantly differentially expressed proteins were used as the input for STRING analysis and a network was built based on high-confidence (0.8) evidence of experimental protein-protein interactions provided by the STRING database. The network was visualized using Cytoscape v.3.0. This is the first time that enrichment of antigens from HiBECs, label-free mass spectrometry, and bioinformatics were combined to explore candidate pathogenesis-associated autoantibodies for AMA-negative PBC.
In total, 1081 autoantigen candidates were identified from all PBC patients. Among these, 371 were significantly differentially expressed between AMA-positive and -negative groups. Fisher’s exact test for the enrichment of GO protein annotations in the set of significantly different proteins revealed a range of protein categories (FDR < 0.05). As expected, the mitochondria-related biological process term was highly enriched. In addition, proteins that participate in the positive regulation of B-cell activation and phagocytic recognition and engulfment were highly enriched. Clustering analysis was applied to examine the rationality and accuracy of the selected proteins. Cluster 1 represents significantly upregulated proteins in AMA-negative PBC patients. They mainly participate in positive regulation of B-cell activation, recognition of phagocytosis, and complement activation. Cluster 2 represents significantly downregulated proteins in AMA-negative PBC patients that are upregulated in AMA-positive PBC patients. These downregulated proteins mainly participate in electron transport and cellular metabolic processes, which are mostly restricted to mitochondria.
This pilot study exploring PBC-related autoantigens demonstrated that the dysfunction of three pathways in HiBECs might be causative in the pathogenesis of AMA-negative PBC, including phagocytosis recognition, positive regulation of B-cell activation, and complement activation. More interestingly, the controversy of the existence of AMA-negative PBC was illustrated based on bioinformatics analysis. AMA cannot be detected in sera of PBC patients because the method applied in clinical practice is not sensitive enough. However, AMA-negative PBC might exist as a unique clinical subset of PBC. This is the first study to combine enrichment of antigens from HiBECs, label-free mass spectrometry, and bioinformatics to explore candidate autoantibodies for AMA-negative PBC. Of note, the results of the current pilot study are consistent with those in research publications, confirming the feasibility and reliability of the design and technology we applied.
The existence of serum autoantibody-negative autoimmune disease can be verified using proteomic technology, including enrichment of antigens from the HiBECs, label-free mass spectrometry, and bioinformatics. For future studies, the design needs to be optimized, including a larger sample size containing healthy and cholestatic cohorts, and the candidate autoantigens should be verified by Western blot analysis or enzyme-linked immunosorbent assay.