A conditional null allele of Mylk3 was generated by introducing loxP sites spanning exon 5, which was done through homologous recombination in ES cells as described previously (Warren et al., 2012). Floxed-Mylk3 homozygous mice (Mylk3^flox∕flox) (Massengill et al., 2016) were bred with transgenic mice carrying the Cre-ER™ gene under CMV promoter, Tg(CAGGS-Cre-ER™) (Hayashi and McMahon, 2002). Subsequent matings between offspring generated Mylk3^flox∕flox/Tg-CAGGS-Cre-ER™, Mylk3^wild∕wild/Tg-CAGGS-Cre-ER™, and Mylk3^flox∕flox on a mixed genetic background, mainly C57BL/6J. For perinatal deletion of the floxed-Mylk3 genes, a single injection of tamoxifen at 50 mg/kg body weight, was administered intra-peritoneally to pregnant mice on the 19th day of gestation, which is the final day of gestation in the mouse. All animal experiments were performed using protocols reviewed and approved by the University of Florida Institutional Animal Care and Use Committee.

Mice were anesthetized with 1.5–2% isoflurane for M-mode ultrasound imaging of the left ventricles using Vevo700 as described previously (Briggs et al., 2008; Takeda et al., 2009; Warren et al., 2012; Massengill et al., 2016).

Cardiomyocytes were isolated by retrograde perfusion of collagenase as described previously (Takeda et al., 2009; Warren et al., 2012; Massengill et al., 2016). Briefly, the hearts were perfused by Langendorff system with the collagenase digestion buffer (type 2 collagenase 2 mg/ml, in MEM with 10 mM taurine, 3.8 mM creatine, 30 mM 2,3-Butanedione 2-Monoxime, and 20 units of insulin, pH 7.3) at 37⋅C for 30 min. Then the hearts were minced into 8–10 pieces, and further digested with the second collagenase digestion (type 2 collagenase 2 mg/ml, in MEM with 10 mM taurine, 3.8 mM creatine, 30 mM 2,3-Butanedione 2-Monoxime, 20 units of insulin, 0.5% bovine serum albumin, and 0.3 mM CaCl₂, pH 7.3) at 37⋅C for 10 min. The cells were spun down at 46 × g for 2 min, resuspended in the second digestion buffer without collagenase and loaded onto a 36% Percoll gradient (GE Healthcare). After centrifugation at 290 × g for 10 min, cardiomyocytes were pelleted and non-myocytes were floated on top. The cardiomyocytes were washed twice with the second digestion buffer without collagenase, suspended and plated on the laminin-coated coverglass for 15 min. The cells were immediately fixed with 4% PFA and analyzed for measurement of cell size as described previously (Takeda et al., 2009; Warren et al., 2012; Massengill et al., 2016). All the reagents except for Percoll were obtained from Sigma.

Western blot analyses and immunostaining were performed with GAPDH (MAB374, EMD Millipore), phospho-MLC2v (gift from Dr. N. Epstein, NIH) (Davis et al., 2001), MLC2 (F109/3E1. ALX-BC-1150-S-L005, Enzo Life Science), and cMLCK (Chan et al., 2008) antibodies. The rabbit affinity-purified cMLCK antibody production has been described previously (Chan et al., 2008). Briefly, the antigen was purified from GST-cMLCK(aa 28-463) following cleavage of GST-tag using thrombin and SDS-PAGE separation followed by electro-elution. Rabbits were injected with the antigen and adjuvants 4 times, and the serum from these animals were affinity-purified using GST-cMLCK(aa 28-463) covalently-coupled beads. 1:4000-diluted antibody was used for Western blotting.

Histological sections (5 μm) were analyzed by hemotoxylin and eosin staining, as well as for fibrosis after Picrosirius red staining, which was performed by heating the tissue sections at 60⋅C for 45 min before deparaffinization and then immersing them in a combination of 0.1% direct red 80 and 0.1% fast green FCF in 1.2% picric acid for 1 h. TUNEL staining was performed using the In Situ Cell Death Detection Kit (Roche 11684795910).

Real-time RT-PCR was performed using inventoried Taqman Gene Expression Assays (Thermo Fisher Scientific): Atrial natriuretic factor (ANF) Mm01255748, brain natriuretic peptide (BNP) Mm00435304, and β-myosin heavy chain (βMHC) Mm00600555. All were normalized to β-actin expression (No. 4352933E). Duplicated experiments were averaged.

Data presented are expressed as mean values ± S.E.M. Results were compared using T-test or ANOVA with Fisher's post-hoc test (SPSS version 24). p < 0.05 was considered significant.

To examine the effects of cMLCK in perinatal cardiac function, we generated tamoxifen-inducible Mylk3 knockout mice that carry homozygous floxed-Mylk3 alleles and heterozygous Cre-ER™ transgene under the control of the CMV enhancer and the chicken β-globulin promoter (CAGGS-Cre-ER™) (Figure 1A). Deletion of floxed-exon 5 resulted in elimination of the first coding exon of the catalytic domain, and a frame-shift of the subsequent downstream exons. In addition, deletion of exon 5 resulted in reduction of cMLCK mRNA (Warren et al., 2012), attributable to a non-sense-mediated mRNA decay (Conti and Izaurralde, 2005), with targeted cMLCK mRNA containing a premature termination codon. Hereafter, conditional deletion of Mylk3 gene in Mylk3^flox∕flox/CAGGS-Cre-ER™ mice with tamoxifen injection at embryonic day 19, will be described as perinatal Mylk3-KO. Of note, this model produces a general knockout of the Mylk3 gene, although the expression of cMLCK is known to be restricted to the myocardium (Chan et al., 2008).

A single tamoxifen injection into pregnant mice on gestational day 19, at 50 mg/kg, given intraperitoneally, reduced the expression of cMLCK protein and MLC2 phosphorylation by postnatal day 3 (P3), such that it could no longer be detected (Figures 1B,C). During the initial experiment, we noticed that perinatal Mylk3-KO mice appeared unhealthy with growth retardation on P19, 2 days before weaning, and there was concern that these mice would die after weaning. Indeed, the ratio of heart weight/body weight (HW/BW) was increased in the perinatal Mylk3-KO mice compared to the two control groups, represented by the Mylk3^flox∕flox and Mylk3^wild∕wild/Tg-CAGGS-Cre-ER™ mice, mainly due to a decrease in body weight (Figure 1D). The remaining groups of mice, therefore, were analyzed using the same experimental timeline. In Figure 1E, we show representative hearts dissected from the control and perinatal Mylk3-KO mice. Expression of fetal genes, often observed in failing hearts, was increased in perinatal Mylk3-KO mice, including atrial natriuretic factor (ANF), BNP, and β-myosin heavy chain (βMHC) (Figure 1F). These results indicate that the perinatal Mylk3-KO induced heart failure, as shown by the increase of HW/BW ratio, increased fetal gene expression, and the loss of body weight as compared to the controls.

Echocardiography performed on the 19th postnatal day demonstrated a significant reduction in cardiac contractility in perinatal Mylk3-KO mice, as measured by %fractional shortening (%FS), and increased systolic dimensions of the left ventricular cavity (Figures 2A,B). Cardiomyocytes isolated from these mice were significantly longer than their controls, albeit without any change in the diameter of the short-axis (Figures 2C,D). Histological analyses, however, failed to show any interstitial fibrosis, nor an increase in TUNEL-positive cells (Figures 3A,B).

Together with our previous studies (Warren et al., 2012; Massengill et al., 2016), we have now analyzed three different Mylk3 knockout mouse models under similar experimental conditions, with the knockout induced in the germline, in the perinatal period, or at adulthood. The populations show both phenotypic similarities and differences, as we have summarized in Table 1. All three models demonstrate a reduction in cardiac contractility, accompanied with elongation of the cardiomyocytes. On the other hand, the findings regarding the fetal gene expression, interstitial fibrosis, and thickening of the cardiomyocytes were different in the three models. When the knockout was introduced in the germline, it failed to induce expression of fetal genes and interstitial fibrosis, but the cardiomyocytes demonstrated thickening. Knockout in the perinatal period, in contrast, did produce an increase in the expression of fetal genes, but again in the absence of fibrosis, and without producing any change in the short-axis dimensions of cardiomyocytes. Knockout in adult mice produced expression of fetal genes, fibrosis, and cardiomyocyte atrophy. These results showed that, when the Mylk3 gene is eliminated in early stages of development, it is most effectively compensated. Compensation, in contrast, is ineffective when the Mylk3 gene is knocked-out in adult mice. Notably, we included two controls in the inducible knockout studies, namely flox/flox and wild/wild/Cre mice with tamoxifen injection in the present and previous studies (Massengill et al., 2016), so as to eliminate any effects of the potential toxicity of tamoxifen and/or Cre transgene for cardiac contractility. Neither of the control populations showed any significant differences in HW/BW ratio or cardiac contractility.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.