Duchenne muscular dystrophy (DMD) is caused by loss of dystrophin in muscle, and while all patients share the primary gene and biochemical defect, there is considerable patient-patient variability in clinical symptoms. We sought to develop multivariate models of serum protein biomarkers that explained observed variation, using functional outcome measures as proxies for severity. Serum samples from 39 steroid-naïve DMD boys 4 to < 7 years enrolled into a clinical trial of vamorolone were studied (NCT02760264). Four assessments of gross motor function were carried out for each participant over a 6-week interval and their mean was used as response for biomarker models. Weighted correlation network analysis was used for unsupervised clustering of 1305 proteins quantified using SOMAscan® aptamer profiling to define highly representative and connected proteins. Multivariate models of biomarkers were obtained for time to stand performance (strength phenotype; 17 proteins) and 6-minute walk performance (endurance phenotype; 17 proteins) including some shared proteins. Identified proteins were tested with associations of mRNA expression with histological severity of muscle from dystrophinopathy patients (n = 28) and normal controls (n = 6). Strong associations predictive of both clinical and histological severity were found for ERBB4 (reductions in both blood and muscle with increasing severity), SOD1 (reductions in muscle and increases in blood with increasing severity), and CNTF (decreased levels in blood and muscle with increasing severity). We show that performance of DMD boys was effectively modeled with serum proteins, with proximal strength associated with growth and remodeling pathways, and muscle endurance centered on TGFβ and fibrosis pathways in muscle.

Introduction: One of the hallmarks of injured skeletal muscle is the appearance of elevated skeletal muscle proteins in circulation. Human skeletal muscle generally consists of a mosaic of slow (type I) and fast (type IIa, IIx/d) fibers, defined by their myosin isoform expression. Recently, measurement of circulating fiber-type specific isoforms of troponin I has been used as a biomarker to suggest that muscle injury in healthy volunteers (HV) results in the appearance of muscle proteins from fast but not slow fibers. We sought to understand if this is also the case in severe myopathy patients with Becker and Duchenne muscular dystrophy (BMD, DMD).Methods: An enzyme-linked immunosorbent assay (ELISA) that selectively measures fast and slow skeletal troponin I (TNNI2 and TNNI1) was used to measure a crosssection of patient plasma samples from HV (N = 50), BMD (N = 49), and DMD (N = 132) patients. Creatine kinase (CK) activity was also measured from the same samples for comparison.Results: TNNI2 was elevated in BMD and DMD and correlated with the injury biomarker, CK. In contrast, TNNI1 levels were indistinguishable from levels in HV. There was an inverse relationship between CK and TNNI2 levels and age, but no relationship for TNNI1.Discussion: We define a surprising discrepancy between TNNI1 and TNNI2 in patient plasma that may have implications for the interpretation of elevated muscle protein levels in dystrophinopathies.

Duchenne muscular dystrophy is initiated by dystrophin deficiency, but downstream pathophysiological pathways such as membrane instability, NFĸB activation, mitochondrial dysfunction, and induction of TGFβ fibrosis pathways are thought to drive the disability. Dystrophin replacement strategies are hopeful for addressing upstream dystrophin deficiency; however, all methods to date use semi-functional dystrophin proteins that are likely to trigger downstream pathways. Thus, combination therapies that can target multiple downstream pathways are important in treating DMD, even for dystrophin-replacement strategies. We sought to define blood pharmacodynamic biomarkers of drug response in the mdx mouse model of Duchenne muscular dystrophy using a series of repurposed drugs. Four-week-old mdx mice were treated for four weeks with four different drugs singly and in combination: vehicle, prednisolone, vamorolone, rituximab, β-aminoisobutyric acid (BAIBA) (11 treatment groups; n = 6/group). Blood was collected via cardiac puncture at study termination, and proteomic profiling was carried out using SOMAscan aptamer panels (1,310 proteins assayed). Prednisolone was tested alone and in combination with other drugs. It was found to have a good concordance of prednisolone-responsive biomarkers (56 increased by prednisolone, 39 decreased) focused on NFκB and TGFβ cascades. Vamorolone shared 45 (80%) of increased biomarkers and 13 (33%) of decreased biomarkers with prednisolone. Comparison of published human corticosteroid-responsive biomarkers to our mdx data showed 14% (3/22) concordance between mouse and human. Rituximab showed fewer drug-associated biomarkers, with the most significant being human IgG. On the other hand, BAIBA treatment (high and low dose) showed a drug-associated increase in 40 serum proteins and decreased 5 serum proteins. Our results suggest that a biomarker approach could be employed for assessing drug combinations in both mouse and human studies.