Extracellular matrix complexity in biomarker studies: a novel assay detecting total serum tenascin-C reveals different distribution to isoform-specific assays.

ABSTRACT

Serum biomarkers are the gold standard in non-invasive disease diagnosis and have tremendous potential as prognostic and theranostic tools for patient stratification. Circulating levels of extracellular matrix molecules are gaining traction as an easily accessible means to assess tissue pathology. However, matrix molecules are large, multimodular proteins that are subject to a vast array of post-transcriptional and post-translational modifications. These modifications often occur in a tissue- and/or disease-specific manner, generating hundreds of different variants, each with distinct biological roles. Whilst this complexity can offer unique insight into disease processes, it also has the potential to confound biomarker studies. Tenascin-C is a pro-inflammatory matrix protein expressed at low levels in most healthy tissues but elevated in, and associated with the pathogenesis of, a wide range of autoimmune diseases, fibrosis, and cancer. Analysis of circulating tenascin-C has been widely explored as a disease biomarker. Hundreds of different tenascin-C isoforms can be generated by alternative splicing, and this protein is also modified by glycosylation and citrullination. Current enzyme-linked immunosorbent assays (ELISA) are used to measure serum tenascin-C using antibodies, recognising sites within domains that are alternatively spliced. These studies, therefore, report only levels of specific isoforms that contain these domains, and studies on the detection of total tenascin-C are lacking. As such, circulating tenascin-C levels may be underestimated and/or biologically relevant isoforms overlooked. We developed a highly specific and sensitive ELISA measuring total tenascin-C down to 0.78ng/ml, using antibodies that recognise sites in constitutively expressed domains. In cohorts of people with different inflammatory and musculoskeletal diseases, levels of splice-specific tenascin-C variants were lower than and distributed differently from total tenascin-C. Neither total nor splice-specific tenascin-C levels correlated with the presence of autoantibodies to citrullinated tenascin-C in rheumatoid arthritis (RA) patients. Elevated tenascin-C was not restricted to any one disease and levels were heterogeneous amongst patients with the same disease. These data confirm that its upregulation is not disease-specific, instead suggest that different molecular endotypes or disease stages exist in which pathology is associated with, or independent of, tenascin-C. This immunoassay provides a novel tool for the detection of total tenascin-C that is critical for further biomarker studies. Differences between the distribution of tenascin-C variants and total tenascin-C have implications for the interpretation of studies using isoform-targeted assays. These data highlight the importance of assay design for the detection of multimodular matrix molecules and reveal that there is still much to learn about the intriguingly complex biological roles of distinct matrix proteoforms.