A combined bioinformatics, experimental and clinical approach to identify novel cardiac-specific heart failure biomarkers: is Dickkopf-3 (DKK3) a possible candidate?

Aims: Cardiac specificity provides an advantage in correlating heart failure (HF) biomarker plasma levels with indices of cardiac function and remodelling, as shown for natriuretic peptides. Using bioinformatics, we explored the cardiac specificity of secreted proteins and investigated in more detail the relationship of Dickkopf-3 (DKK3) gene expression and DKK3 plasma concentrations with cardiac function and remodelling in (pre)clinical studies.

Methods and results: The cardiac specificity of secreted proteins was determined using RNAseq data for a large panel of organs and tissues. This showed that natriuretic peptides (NPPA and NPPB) are highly cardiac-specific (>99%), whereas other HF biomarkers, including galectin-3 (Gal-3, LGALS3) and growth differentiation factor-15 (GDF-15), lack cardiac specificity (<4%). DKK3 was cardiac-enriched (44%), warranting further investigation. In three different HF mouse models, cardiac Dkk3 expression was altered, but DKK3 plasma concentrations were not. In humans, DKK3 plasma concentrations were higher in HF patients (n = 2090) in comparison with age- and sex-matched controls without HF (n = 240) (46.4 ng/mL vs. 36.3 ng/mL; P < 0.001). Multivariate regression analysis revealed that DKK3 was strongly associated with HF risk factors and comorbidities, including age, kidney function and atrial fibrillation. After correction for existing prediction models, DKK3 did not independently predict HF outcome [all-cause mortality/HF hospitalization, hazard ratio 1.13 (0.79-1.61) per DKK3 doubling; P = 0.503].

Conclusions: Of actively secreted HF biomarkers, only natriuretic peptides showed high cardiac specificity. Despite a cardiac specificity of 44%, secreted DKK3 had limited additional diagnostic and prognostic value.