Response to Thum et al.

Abstract

Thum et al. conclude that microRNA-21 (miR-21) is essential for cardiac hypertrophy and fibrosis in response to pressure overload (1). They also claim that our failure to observe a blockade to these processes in mice treated with an 8-mer locked nucleic acid–modified oligonucleo­tide against miR-21 (called Anti-21) (2) is due to the ineffectiveness of such inhibitors. We wish to point out several caveats to their study regarding the role of miR-21 in cardiac hypertrophy and their conclusions regarding the efficacy of the Anti-21 oligonucleotide. First, we find that Anti-21 inhibits miR-21 with a half-maximal inhibitory concentration of 0.9 nM, indicating the efficacy of Anti-21. Second, Thum et al. do not state the method they used to measure miR-21 inhibition, though we assume it to be quantitative PCR (qPCR). In our hands, qPCR alone is unreliable for measuring miRNA inhibition, especially for 8-mer inhibitors, since they may be displaced during qPCR and thereby give an underrepresentation of miRNA inhibition. To demonstrate functional inhibition of a miRNA, it is important to show data from multiple assays, such as small RNA Northern blots, luciferase reporter assays, and target derepression, as shown in our study (2). Such data are lacking in the Thum et al. rebuttal, which makes comparison of the different chemistries impossible. Thum et al. also state that we measured miR-21 inhibition on day 2 after dosing with Anti-21, when in fact we measured inhibition 3 weeks after dosing. At this time point, we observed inhibition of miR-21 in pressure-overloaded hearts at a level significantly below that of control mice. Thus, their approach for inhibition and/or measurement of miR-21 by their 8-mer inhibitors differs markedly from ours, since we observed robust miR-21 inhibition 3 weeks after injection, as demonstrated by multiple readouts (2). Using mismatched oligonucleotide controls is also important for interpreting miRNA inhibition studies in vivo, as described in our paper, rather than using PBS as a control, as reported by Thum et al. (1). Finally, Thum et al. postulate that constitutive genetic deletion of miR-21 in mice may not reveal the functions of miR-21 in cardiac disease because of compensatory events that mask such functions. If such compensation occurs, it must be specific for the cardiac functions of miR-21, since miR-21 null mice are resistant to lung tumorigenesis (3), consistent with the documented pro-oncogenic functions of miR-21. To further address the possibility of genetic compensation, we have deleted a floxed miR-21 allele immediately prior to thoracic aortic constriction in mice using a ubiquitously expressed tamoxifen-regulated Cre transgene. These animals show cardiac hypertrophy and fibrosis comparable to that of their Cre-negative littermates. Genetic compensation therefore cannot account for the normal pathological cardiac remodeling response in miR-21 null mice. Moreover, functions of other miRNAs in heart disease can be revealed by genetic deletion in mice, as shown for miR-208 (4). Thus, while 22-mer oligonucleotide inhibitors against miR-21 are efficacious in inhibiting cardiac hypertrophy (1), other loss-of-function approaches appear ineffective. We remain enthusiastic about miRNAs as therapeutic targets and welcome such dialog.