Lack of WWC2 Protein Leads to Aberrant Angiogenesis in Postnatal Mice

Abstract

The WWC protein family is an upstream regulator of the Hippo signalling pathway that is involved in many cellular processes. We examined the effect of an endothelium-specific WWC1 and/or WWC2 knock-out on ocular angiogenesis. Knock-outs were induced in C57BL/6 mice at the age of one day (P1) and evaluated at P6 (postnatal mice) or induced at the age of five weeks and evaluated at three months of age (adult mice). We analysed morphology of retinal vasculature in retinal flat mounts. In addition, in vivo imaging and functional testing by electroretinography were performed in adult mice. Adult WWC1/2 double knock-out mice differed neither functionally nor morphologically from the control group. In contrast, the retinas of the postnatal WWC knock-out mice showed a hyperproliferative phenotype with significantly enlarged areas of sprouting angiogenesis and a higher number of tip cells. The branching and end points in the peripheral plexus were significantly increased compared to the control group. The deletion of the WWC2 gene was decisive for these effects; while knocking out WWC1 showed no significant differences. The results hint strongly that WWC2 is an essential regulator of ocular angiogenesis in mice. As an activator of the Hippo signalling pathway, it prevents excessive proliferation during physiological angiogenesis. In adult animals, WWC proteins do not seem to be important for the maintenance of the mature vascular plexus.

Keywords: Hippo signalling pathway; WWC protein; angiogenesis; endothelial cell specific knock-out; hypersprouting.

Figure 1

Retinal whole mounts from P6 mice with blood vessels labelled with isolectin B4. The control is shown in (A), and the three knock-out groups in (B–D). The latter did not express WWC2 and differed in their WWC1 expression as indicated. They exhibited a dense vascular network in the peripheral plexus, the so-called, “hypersprouting areas”, and a smaller portion of the retinal vascularisation.

Figure 2

Example for evaluation of a retinal whole-mount. Digital image of a complete retinal whole-mount is shown in (A). In this example, size of total retinal area was determined to be 19,752,079 pixels. Size of vascularised area was 13,578,789 pixels (B), and size of hypersprouting area 4,607,985 pixels (C). According to this, sprouting had a 33.9% share of vascularised area. Staining against the endothelial cell-specific molecule 1 (ESM1) displays the tip cells which helped to determine the hypersprouting areas (D).

Figure 3

Results of evaluation of vascularised and hypersprouting areas as indicated in P6 retinal whole-mounts. In this and the following diagrams, data are presented of n = 42 control eyes, n = 8 WWC1^+/+WWC2^−/− eyes, n = 56 WWC1^+/−WWC2^−/− eyes, and n = 23 WWC1^−/−WWC2^−/− eyes. Statistical significance of differences between the groups is denoted by asterisks: *** p < 0.001, **** p < 0.0001.

Figure 4

Example for the evaluation of a retinal leaf using the AngioTool software. The retinal leaf (A) was divided into a peripheral part (B) and a central part (C). Vessels (red) and junctions (blue dots) were detected by the software in the digital images (D,E).

Figure 5

Diagrams showing values of total vessel lengths, normalised to the analysed area, and average vessel lengths between junction points as indicated. Circles, triangles and squares indicate outliers according to Tukey’s presentation of box plots. Statistical significance of differences between the groups is denoted by asterisks: ** p < 0.01, **** p < 0.0001.

Figure 6

Diagrams showing values of junction density, lacunarity, and vascular endpoints in P6 retinas as indicated. Junction density and number of endpoints were normalised to the total vessel length in the analysed areas. Triangles and squares indicate outliers according to Tukey’s presentation of box plots. Statistical significance of differences between the groups is denoted by asterisks: * p < 0.05, ** p < 0.01, *** p < 0.001, **** p < 0.0001.

Figure 7

Examples of in vivo imaging of Cre-negative mice (controls) and Cre-positive mice (knock-outs) as indicated. On top, images are shown obtained by infrared scanning laser ophthalmoscopy (left) and fluorescein angiography (right). Below, optical coherence tomography images are shown, with both horizontal and vertical scans.

Figure 8

Values of different ERG parameters obtained in Cre-positive (knock-outs) and Cre-negative (controls) WWC1^−/−WWC2^−/− mice. On top, latencies and amplitudes in scotopic ERG are shown depending on intensity of light stimulus. Below, amplitudes and latencies of scotopic and photopic oscillatory index, photopic b-wave, photopic 30-Hz Flicker ERG, and scotopic b/a ratio are shown as indicated.

Figure 9

Retinal whole mounts from 12-week-old mice with blood vessels labelled with Isolectin B4.

Figure 10

Diagrams showing values of total vessel length normalised to analysed area, lacunarity, average vessel length, and junction density normalised to total vessel length in 12-week-old retinas of Cre-positive mice (knock-outs) and Cre-negative mice (controls) as indicated. Triangles and circles indicate outliers according to Tukey’s presentation of box plots. Statistical significance of differences between the groups is denoted by asterisks: ** p < 0.01.