A micro-ribonucleic acid signature associated with recovery from assist device support in 2 groups of patients with severe heart failure

Abstract

Objectives: This study was conducted to test the hypothesis that cardiac micro-ribonucleic acid (miR) profiling in severe heart failure patients at the time of ventricular assist device (VAD) placement would differentiate those who remained VAD-dependent from those with subsequent left ventricular (LV) recovery.

Background: The relationship of myocardial miR expression to ventricular recovery is unknown.

Methods: We studied 28 patients with nonischemic cardiomyopathy requiring VAD support consisting of test and validation cohorts from 2 institutions: 14 with subsequent LV recovery and VAD removal and 14 clinically matched VAD-dependent patients. Apical core myocardium was studied for expression of 376 miRs by polymerase chain reaction (PCR) array and real-time-PCR methods. Samples from 7 nonfailing hearts were used in confirmatory studies.

Results: By PCR array, 10 miRs were differentially expressed between LV recovery and VAD-dependent patients in the test cohort. The real-time PCR confirmed lower expression in LV recovery patients for 4 miRs (15b, -1.5-fold; 23a, -2.2-fold; 26a, -1.4-fold; and 195, -1.8-fold; all p < 0.04 vs. VAD dependent). The validation cohort similarly showed lower miRs expression in LV recovery patients (23a, -1.8-fold; and 195, -1.5-fold; both p < 0.03). Furthermore, miR 23a and 195 expression in nonfailing hearts was similar to LV recovery patients (both p < 0.04 vs. VAD dependent). The LV recovery patients also had significantly smaller cardiomyocytes by quantitative histology in both cohorts.

Conclusions: Lower cardiac expression of miRs 23a and 195 and smaller cardiomyocyte size at the time of VAD placement were associated with subsequent LV functional recovery. Differential expression of miRs at VAD placement may provide markers to assess recovery potential.

Figure 1. miRs differentially expressed between Recovered and Dependent patients as detected by RT-PCR Array in Test cohort

Threshold detection values after normalization to small RNAs U6 and RNU44 expression. Box and whisker plot, where the range of values are represented by the whiskers, the upper and lower quartile by the box, and the median value by the bar. N=6 per group; *, p<0.05.

Figure 2. Recovered patients in the Test Cohort have decreased expression of selected miRs

Individual miRs found differentially expressed in the screening array were tested using individual PCR. miRs-15b, -23a, -26a and -195 were confirmed to be differentially expressed. *p<0.04 (n=7 per group)

Figure 3. Recovered patients in the Validation Cohort have decreased expression of miRs-23a and -195

miRs confirmed to be differentially expressed in the Test cohort were tested in the Validation cohort (n=7 per group). miRs-23a and -195, but not -15b and 26a were differentially expressed. *p<0.03

Figure 4. Expression of miRs 23a and 195 was similar in Nonfailing hearts compared with Recovered, but was significantly increased in Dependent vs. Nonfailing

Fig 4A and B represent data from the Test Cohort, while Figs 4C and D are from the Validation cohort. *p<0.04, Dependent vs. Recovered; †p<0.03, Dependent vs. Nonfailing control Hearts.

Figure 5

Figure 5a. Recovered patients have smaller cardiomyocytes at the time of VAD implantation. Representative images (20X) of wheat germ agglutinin (green; nuclei blue) stained VAD core samples from Test Cohort Recovered and Dependent samples, with calculated cross sectional area in μM². (n=3 per group; *p=0.01). 5b. Representative (H+E) images and measurements of cardiomyocyte diameter (μM) in the Validation Cohort Recovered and Dependent samples (n=7 per group, *p = 0.01).

Figure 6

Summary of experimental protocol and major findings.