Large scale serum proteomics identifies proteins associated with performance decline and clinical milestones in Duchenne muscular dystrophy

Abstract

Serum biomarkers are promising minimally invasive outcome measures in clinical studies in Duchenne muscular dystrophy (DMD). However, biomarkers strongly associated with clinical progression and predicting performance decline are lacking. In this study we aimed to identify serum biomarkers associated with clinical performance and able to predict clinical milestones in DMD. Towards this aim we present a retrospective multi-center cohort study including serum samples and clinical data collected in research participants with DMD as part of a natural history study at the University of Florida (UF) and real-world observations at Leiden University Medical Center (LUMC) between 2009-2022. The 7K SomaScan^® assay was used to analyse protein levels in in individual serum samples. Serum biomarkers predicted age at loss of ambulation (LoA), age at loss of overhead reach (OHR) and age at loss of hand to mouth function (HTM). Secondary outcomes were the association of biomarkers with age, corticosteroid (CS) usage, and clinical performance based on the North Star Ambulatory Assessment (NSAA), 10 meter run velocity (10mrv), 6 minute walk (6MWT) and Performance of the Upper Limb (PUL2.0). A total of 716 serum samples were collected in 79 participants at UF and 74 at LUMC (mean[SD] age; 10.9[3.2] vs 8.4[3.4]). 244 serum proteins showed an association with CS usage in both cohorts independent of CS type and regimen, including MMP3 and IGLL1. 318 probes (corresponding to 294 proteins) showed significant associations with NSAA, 10mrv, 6MWT and/or PUL2.0 across both cohorts. The expression of 38 probes corresponding to 36 proteins such as RGMA, EHMT2, ART3, ANTXR2 and DLK1 was associated with risk of both lower and upper limb clinical milestones in both the LUMC and UF cohort. In conclusion, multiple biomarkers were associated with CS use, motor function and upper lower and upper limb disease milestones in DMD. These biomarkers were validated across two independent cohorts, increasing their likelihood of translation for use within the broader DMD population.

Figure 1.

Associations of Serum Proteins with Age in Individuals with DMD. Volcano plots demonstrate the proteins from DMD serum that are associated with age in A) the LUMC cohort, and B) the UF cohort. C) Venn diagram showing the number of proteins associated with age in common between the LUMC and UF cohorts, with D) a plot of coefficients, indicating the direction of protein association with age, in the LUMC versus UF cohorts. Plots E-H highlight the longitudinal trajectories of protein levels over time in select proteins that decrease with increasing age such as CKM, MYOM3, TTN and TNNI2, while plots I-L show trajectories of select proteins that increase with increasing age such as CNTN3, CNDP1, LEP and TIMP4. Each trajectory plot includes the gene symbol and aptamer identification number. DMD, Duchenne muscular dystrophy; LUMC, Leiden University Medical Center; UF, University of Florida.

Figure 2.

Associations of Serum Proteins with Corticosteroid use in Individuals with DMD. Volcano plots demonstrate the proteins from DMD serum that are associated with corticosteroid use in A) the LUMC cohort, and B) the UF cohort. C) Venn diagram showing the number of proteins associated with corticosteroid use in common between the LUMC and UF cohorts, with D) a plot of coefficients, indicating the direction of protein association with corticosteroid use, in the LUMC versus UF cohorts. Plots E-H show differences in protein levels between serum from CS treated versus untreated individuals with DMD in two previously identified proteins and two proteins identified in this study. I) Scatterplot showing the relationship between age and steroid treatment coefficients for proteins significantly associated with CS treatment. Proteins showing discordant coefficients for age and CS treatments were considered as efficacy biomarkers, while with discordant coefficients were considered as safety biomarkers (grey shaded). Orange dots represent estimates for the LUMC cohort, while blue dots represent estimates for the UF cohort. CS, corticosteroid; DMD, Duchenne muscular dystrophy; LUMC, Leiden University Medical Center; UF, University of Florida.

Figure 3.

Associations of Serum Proteins with Performance on Tests of Motor Function in Individuals with DMD. A-D) Longitudinal trajectory plots of NSAA results, 10MRV, 6MWD, and PUL 2.0 score from both cohorts. E) Plots of coefficients, indicating the direction of protein association with motor function test performance in the LUMC versus UF cohorts. Highlighted data points represent significant associations. The number of significant associations is mentioned for each panel. F) UpSet plot demonstrating the intersection of significantly associated probes with each of the four tests of motor function. G-H) Trajectory plots showing the relationship between ART3 and either NSAA (G) or 6MWT (H) over the age in the LUMC cohort. I) Scatterplot showing the relationship between the coefficient estimates for proteins significantly associated with steroid treatment and performance test. Proteins showing concordant coefficients for performance tests and CS treatments were considered as efficacy biomarkers (e.g. a protein associated with both high functional scores and further elevated by steroid treatment), while with concordant coefficients were considered as safety biomarkers (grey shaded). Orange dots represent estimates for the LUMC cohort, while blue dots represent estimates for the UF cohort. Shapes indicate the association for each performance test. 6MWD, 6-minute walk distance; 10MRV, 10-meter run velocity; CS, corticosteroid; DMD, Duchenne muscular dystrophy; LUMC, Leiden University Medical Center; NSAA, North Star Ambulatory Assessment; PUL, Performance of the Upper Limb; UF, University of Florida.

Figure 4.

Associations of Serum Proteins with Clinical Milestones. A) Kaplan Meier curves showing the probability of LoA, loss of OHR, and loss of HTM by age in each cohort. B) UpSet plot of the intersection of significant probes with each clinical milestone. C) Log hazard ratios of the protein probes significantly associated with at least 1 clinical milestone. D) Kaplan Meier curves for loss of each clinical milestone for RGMA by years since first sample collection. E) Correlations between protein probe log hazard ratios and use of corticosteroids. HTM, hand-to-mouth; LoA, loss of ambulation; OHR, overhead reach.