MiRNAs are known to regulate gene expression in various organs and are recognized as important regulators of cardiac development and function
[16, 17]. The cardiac-specific transcription factors, HAND1 and HAND2, play important roles in left versus right ventricular determination. We tested a subset of differentially expressed miRNAs predicated to target CM subtype-specifying transcription factors, such as HAND1 and HAND2. In vivo regional expression, in silico predictions, and experimental validation demonstrated that miR-363 is an upstream regulator of HAND1 translation, leading to a role in left ventricular CM differentiation.
To our knowledge, regulation of HAND1 or HAND2 by miRNAs has not previously been reported. These two genes are closely related and display complementary and overlapping expression patterns in the developing heart
[18, 19]. During development, HAND1 is expressed in the inflow segment of the linear heart tube destined to become the left ventricle, whereas HAND2 is expressed much earlier throughout the linear heart tube, and then later expressed in the outflow portion of the heart tube destined to become the right ventricle and atria
. However, little is known about the posttranscriptional control of HAND gene expression, although conserved miRNA seed-pairing sequences in the 3′UTRs of both HAND1 and HAND2 suggest control by miRNAs. In this study, we sought to identify specific miRNAs that target specifically ventricular CM-determining genes. The expression profiles of known human miRNAs were analyzed to identify potential miRNA-mRNA interactions that effect hESC differentiation into CMs.
miRNAs are known to target mRNAs by imperfect base pairing with their 3′UTR
. This in turn inhibits translation and/or destabilizes the targeted mRNA
, ultimately controlling its expression. Here, we identified subsets of 137, 100, and 47 miRNAs that were highly expressed in day 8 CM precursors versus undifferentiated cells, day 14 CMs versus undifferentiated cells, and Day 8 CM precursors versus Day 14 CMs, respectively (Table
1). The left/right ventricle transcriptional determinants, HAND1 and HAND2, were identified as targets of four miRNAs, (miR-363, -367, -181a, and -181c).
In a previous report
, we showed that miR-125b regulated the development of hESC-derived early mesoderm and was highly expressed in cardiac precursors. We showed that miR-125b targets Lin28, indirectly inhibits Nanog and Oct4, and promotes onset of Brachyury, GATA4, and NKX2.5 expression to induce cardiac mesoderm formation. In both this and the current study, we predicted the target genes based on conserved pairing regions
. We are aware that the majority of these predictions rely on extensive complementarity, while accounting for other features that contribute to miRNA 3′UTR recognition
hESC differentiation in vitro leads to multiple cell lineages arising from the three embryonic germ layers, including CMs. The spontaneous differentiation of hESCs into CMs, however, leads to a heterogeneous mixture of CMs
[3, 10]. Although several protocols are used for inducing CM differentiation, we used a method that allows ongoing interaction between many CM subtypes
. In addition, we used a previously described aMHC-EGFP reporter hESC line that allowed us to sort specifically cardiac precursors and embryonic CMs
The expression of miR-363 was detected with fluorescence in situ hybridization in E10.5 mouse embryos. miR-363 was expressed in the developing limb bud, notochord, ectoderm, and brain. These results are consistent with the miR-363 expression in chick embryo reported previously
, in which miR-363 was observed in ectoderm, pharyngeal arches, notochord, and limb bud, suggestive of wide function in limb development, patterning, and central nervous system development. Tissue-specific miRNA expression implies a negative regulatory role in expression f their targets. However, many transcription factors demonstrate more-complicated, stoichiometric expression during the course of tissue development. Relevant to this study, HAND2 is initially expressed throughout the developing heart tube, but then is restricted to second heart-field structures, with HAND1 expression restricted to the developing left ventricle
. The mechanisms responsible for this developmentally regulated expression of HAND genes has until now been elusive.
Two evolutionarily conserved miR-363 seed-pairing sites in human HAND1 3′UTR suggested that the miRNA pairing sequences predicted in silico contribute to HAND1 regulation. HAND1 3′UTR reporter activity was completely abolished by miR-363 in contrast to HAND2 3′UTR activity. NKX2.5 also has been implicated in left/right asymmetric expression of HAND1 and HAND2
; however, a role for miR-363 in regulating NKX2.5 expression was not suggested by our results.
BMP signaling controls the differentiation of CMs in multiple ways
. In this work, BMP4 or BMP2 was shown to elicit induction of CM differentiation not only from precardiac mesoderm but also from tissue that is normally not cardiogenic. We tested various signaling molecules that have been implicated in cardiac induction
[30–32]. TGF-β family members such as Activin A, bFGF, and BMPs, have been identified as promoting the terminal differentiation of precardiac mesoderm; however, when used at concentrations reported to be pharmacologically effective, neither Activin A nor bFGF displayed any discernible HAND1/2-inducing effect. BMP4 is closely related to the TGF-β family member BMP2, is expressed in anterior lateral ectoderm, and is indistinguishable from BMP2 in cardiac-inducing activity
. We observed similar induction of HAND1 and HAND2 with BMP4 and BMP2. Taken together, these findings suggested a model for BMP-mediated cardiac induction and CM-subtype specification through miR-363 and differential expression of HAND1/2 (Figure
CM commitment after differentiation of hESCs was studied in the presence of a miR-363 antagomir. The isolated cardiogenic cells overexpressing anti-miR-363 expressed the left-specifying factor HAND1 at levels significantly higher than control. Cardiac progenitors that differentiate in vitro accumulate muscle-specific proteins but do not necessarily exhibit conspicuous beating or cross-striations; thus, these may be overlooked with less-sensitive assays with histology or spontaneous contractility.
We used qPCR to assess the differences in miRNA expression between beating foci and evaluated the ratio of HAND1 to HAND2 genes. miRNAs are known to act as transcriptional repressors of their target RNAs, thereby downregulating gene expression. However, it is possible that critical regulatory proteins may compensate for posttranscriptional downregulation by other mechanisms (for example, increased protein stability, decreased protein turnover). Therefore, our analysis may not identify every regulatory feature of the miRNA-CM subtype-specification pathway.
Right-ventricle CMs were devoid of HAND1 expression, which allowed us to distinguish between left and the right ventricular CMs. One explanation could be that the emergence of CMs from committed progenitors is influenced by cell density. High densities of committed progenitor cells show distinct populations of atrial and ventricular CMs
. It appears that differentiation in the presence of anti-miR-363 causes progenitor cells to differentiate into a HAND1-rich population, and left-ventricular CMs constitute the greatest percentage of cells expressing HAND1.