MicroRNAs make a difference in cardiovascular robustness

Abstract

Invertebrate microRNAs can suppress developmental variability that is caused by environmental and genetic variation. In this issue of Developmental Cell, Kasper et.al. (2017) show that zebrafish miRNAs suppress variability in cardiovascular development during embryogenesis, providing insight into the conserved link between miRNAs and robustness.

Organisms are naturally subject to variable conditions of existence, and yet their morphological development is generally robust to such challenges. Indeed, developmental robustness is an emergent property of living systems that has long been remarked upon (Waddington, 1942). How developmental robustness is achieved is a question of keen interest.

One class of gene regulators that have been proposed to provide robustness is the microRNA (miRNA) family of non-coding RNAs (Ebert and Sharp, 2012; Hornstein and Shomron, 2006). Most miRNAs elicit weak repression of target gene expression. Weak and tunable repression has led to the proposition that miRNAs generally elicit two distinct effects on their molecular targets (Ebert and Sharp, 2012). In the first type of effect, a miRNA reduces the level of target protein below a threshold that might act like a switch. In the second type of effect, a miRNA buffers fluctuations in the target protein, limiting undesired signal propagation. Each of these effects can potentially be harnessed to provide robustness to gene regulation.

These regulatory properties are consistent with many genetic studies in invertebrate model organisms. There, loss of a miRNA typically has little or no effect on average developmental outcomes but dramatically increases the phenotypic variability between individuals (Vidigal and Ventura, 2015). Such effects are predicted if developmental robustness is compromised. Surprisingly, this has not been rigorously tested in vertebrate models until now.

In this issue of Developmental Cell, Stefania Nicoli and colleagues describe how several miRNAs reduce phenotypic variation in zebrafish lab populations, and may thus canalize developmental traits related to the cardiovasculature (Kasper et al., 2017). The authors identified miR-139, miR-223, and the miR-24 family, which were highly enriched in cardiovascular endothelia of zebrafish embryos. Using TALEN- and CRISPR-induced knockout mutations, they found that only 1-2% of computationally predicted target mRNAs became upregulated as a result of loss of miRNA activity.

Next, the authors wanted to examine the developmental phenotypes in the miRNA mutants. Loss of miR-139 modestly increased the average number of filopodia membrane extensions on developing blood vessel cells. However, the mutants showed strikingly greater variation in number of filopodia from cell to cell. This variation occurred in two directions: some cells had an excess of filopodia whereas others had no filopodia at all. The number of filopodia in vascular cells affects their ability to migrate in response to signals; thus alterations in this trait could affect cell behavior during vascular formation.

As there are four genes in the miR-24 family, the authors generated mutants with differing numbers of family members, and were able to observe effects on the length of the hypobranchial artery, which supplies blood to craniofacial tissues. Although a significant change in average artery length could not be detected until three family members were lost, there was much greater variation in artery length between individuals even when just one or two miR-24 genes were mutated. It appears that reproducibility of artery development is sensitive to miR-24.

Why did these miRNA mutants increase phenotypic variability? The authors hypothesized that without miR-139 and miR-24, development is more sensitive to environmental variation. To test this hypothesis, they applied mild stress to embryos in the form of angiogenesis drugs, elevated temperature, or mild hypoxia. They found that while these conditions had no effect on cardiovascular development of wild-type embryos, the cardiovascular system of miR-139 and miR-24 mutants developed abnormally.

Interestingly, miR-223, which is expressed in hematopoietic stem/progenitor cells, did not appear to impact robustness. Variation in stem cell number was unaffected whereas the average number of stem cells was elevated in miR-223 mutants. Moreover, mild environmental stress did not exacerbate the developmental defect. Thus, not all miRNAs play a role in suppressing developmental variation.

Although the role of miRNAs in suppressing developmental variation is established for invertebrates, and now for vertebrates (Figure 1), it remains poorly understood how miRNA regulation of gene expression is linked to developmental robustness. How does the weak repression of target genes produce a buffering effect at the macroscopic level? Presumably, miRNAs do so by stabilizing variable gene expression, but this needs to be broadly tested. Experiments in tissue culture suggest that mouse miRNAs can reduce variability in protein expression from transfected reporter genes (Schmiedel et al., 2015). It remains to be seen if this is a general principle for animal miRNAs and their natural targets in vivo. Do miRNAs mediate phenotypic robustness by buffering the levels of one or two important genes, or many genes in networks? Experiments in Drosophila suggest that at least for one miRNA, a single target mediates suppression of phenotypic variation (Cassidy et al., 2013). It is important to determine if this is a general principle. Finally, what are the sources of variation that are buffered by miRNAs? As now shown in zebrafish, as well as invertebrate models, environmental differences are one plausible source (Cassidy et al., 2013; Kasper et al., 2017; Ren and Ambros, 2015; Yatsenko and Shcherbata, 2014). Another source, at least in Drosophila, is the natural variation in genome sequence between individuals (Cassidy et al., 2013). Finally, variation in gene expression caused by stochastic processes may be another target of miRNA action. Quantitative measurements within model systems will need to be combined with mathematical modeling to investigate how miRNAs work to generate developmental robustness.

Figure 1. Bottom-up effect of miRNAs on zebrafish developmental robustness.

Suppression of variation in protein expression from miRNA target genes presumably leads to a Waddington landscape in which cell fate decisions and morphogenesis are rendered robust. Waddington envisioned a developing embryonic cell like a ball rolling down a landscape, staying in valleys and seeking the lowest point (Waddington, 1942). The height of barriers (hills) correlates with the stability of developmental decisions that cells make. In the case of zebrafish, miR-24 and miR-139 increase the height of barriers, resulting in phenotypic homogeneity for cardiovascular traits.