Selective and specific changes in gene expression characterize the end-stage failing heart. However, the pattern and relation of these changes to evolving systolic and diastolic dysfunction during development of heart failure remains undefined. In the present study, we assessed steady-state levels of mRNAs encoding a group of cardiac proteins during the early development of left ventricular dysfunction in dogs with pacing-induced cardiomyopathy. Corresponding hemodynamic assessments were made in the conscious state in the same animals and at the same time points at baseline, after 1 week of ventricular pacing, and at the onset of clinical heart failure. Systolic dysfunction dominated after 1 week of pacing, whereas diastolic dysfunction was far more pronounced with the onset of heart failure. Atrial natriuretic factor mRNA was undetectable in 7 of 12 hearts at baseline but was expressed in all hearts at 1 week (P < .01 by chi 2 test), and it increased markedly with progression to failure (P = .05). Creatine kinase-B mRNA also rose markedly with heart failure (P < .01). Levels of mRNA encoding beta-myosin heavy chain, mitochondrial creatine kinase, phospholamban, and sarcoplasmic reticulum Ca(2+)-ATPase did not significantly change from baseline, despite development of heart failure. Additional analysis to determine if these mRNA changes were related to the severity of diastolic or systolic dysfunction revealed that phospholamban mRNA decreased in hearts with larger net increases in end-diastolic pressure (+19.2 +/- 1.9 mm Hg) compared with those hearts in which it did not change (+4.0 +/- 4.9, P < .02).(ABSTRACT TRUNCATED AT 250 WORDS)