Figure 1.
The generation of Vldlr-/- Lrp5-/- (DKO) mice via crossover mice containing both the Lrp5 knockout allele (Lrp5-) and the Vldlr knockout allele (Vldlr-).
The Lrp5 and Vldlr genes are both located on mouse chromosome 19. Double heterozygous Vldlr+/- Lrp5+/- mice were generated by mating the Vldlr-/- and Lrp5-/- mice, and they were then intercrossed. A Vldlr+/- Lrp5-/- male mouse with a crossover was identified from 124 offspring mice. This male mouse was mated with Lrp5-/- female mice to generate both female and male Vldlr+/- Lrp5-/- mice. The DKO mice were generated by an intercross between Vldlr+/- Lrp5-/- mice.
Figure 2.
Fundus photos and fluorescein angiograms of 6-week-old wild-type (WT), Vldlr-/-, Lrp5-/- and Vldlr-/- Lrp5-/- or DKO mice.
The upper panels are fundus photos, and the lower panels are angiograms. Hypopigmented yellowish patches in the fundus of Vldlr-/- (arrows) were not observed in Lrp5-/- or DKO fundus. The Vldlr-/- angiogram displayed hyperfluorescent spots (arrows) that corresponded to hypopigmented yellowish patches on the fundus. Both Lrp5-/- and DKO angiograms showed homogeneous leakage of retinal vasculature without hyperfluorescent spots.
Figure 3.
The deletion of LRP5 prevents the overgrowth of retinal vessels in the Vldlr-/- mice.
(A) Immunostaining of 3-week-old WT, Vldlr-/-, Lrp5-/- and Vldlr-/- Lrp5-/- DKO retinal sections. Retinal vessels were visualized with an anti-CD31 antibody (green), and cell nuclei were labeled by DAPI (blue). In the WT control retina, vascular ECs were observed in the outer-plexiform layer (OPL), inner-plexiform layer (IPL), and ganglion cell layer (GCL). In the Vldlr-/- retina, CD31 signals were not only present in the OPL, IPL and GCL, but also in the photoreceptor outer nuclear layer (ONL) and in the subretinal space (white arrow). In the Lrp5-/- retina, very few CD31-stained cells were detected in the OPL, and CD31-positive ECs formed clusters in the IPL (white arrowhead). A CD31 staining pattern similar to Lrp5-/- was observed in the DKO retina, where ECs also formed clusters (white arrowheads). INL: inner nuclear layer. Scale bar: 50µm. (B) Retinal histology of 3-week-old WT, Vldlr-/-, Lrp5-/-, and DKO mice. Compared to the WT retinal section, the Vldlr-/- section displayed a disorganized ONL with photoreceptor cells moving along the abnormal vessels (white arrowheads). Both Lrp5-/- and DKO sections showed toluidine blue-stained cell clusters (white arrows) in the IPL. Scale bar: 50µm.
Figure 4.
3D reconstruction of retinal vasculature in 3-week-old WT, Vldlr-/-, Lrp5-/- and DKO mice with the Sca1-GFP transgene.
The 3D images were constructed from Z-stack images of GFP-positive ECs, starting from the retinal surface (GCL) through the OPL. The upper panels show a top view of 3D images while the lower panels show a 90 degree–rotated side view of the upper 3D images. The top-view panel (the upper left) reveals wild-type retinal vasculature consisting of large vessels, small vessels and capillaries while the side-view panel (the lower left) shows a typical interconnected 3-layer of retinal vasculature, corresponding to enriched vessels in the GCL (upper), IPL (middle) and OPL (lower), in the wild-type retina. Compared to a clear 3-layer retinal vascular network in the WT control, the Vldlr-/- retina shows abnormal vessel growth originated from the OPL vessels (white arrowheads). The Lrp5-/- and DKO retinas display very similar phenotypes, such as incomplete retinal vasculature with only the surface vessel network and the formation of EC clusters in the IPL (white arrows).
Figure 5.
Transcription levels of angiogenesis-related molecules examined by semi-quantitative PCR and quantitative real-time PCR.
(A) Representative semi-quantitative RT-PCR results from 3-week-old WT, Lrp5-/-, Vldlr-/- and DKO mice. Results from retinal RNAs of two littermate mice for each genotype were shown. As expected, Lrp5-/- retinas (L1 and L2) lacked detectable LRP5 transcript and Vldlr-/- retinas (V1 and V2) showed no expression of VLDLR. Both LRP5 and VLDLR transcripts were undetectable in DKO retinas (D1 and D2). Moreover, transcription levels of both Lrp6 and FZD4 remained unchanged among WT, Lrp5-/-, Vldlr-/- and DKO retinas. The G3PDH levels from the same RNA samples were used as quantification controls. (B) Relative expression levels of FZD4, Norrin, β-catenin, Sox17, Ang1 and Slc38a5 in retinas of 3-week-old WT (W), Lrp5-/- (L), Vldlr-/- (V) and DKO (D) mice were examined by RT-qPCR. The bar graphs show relative changes of these molecules compared to WT. Average WT levels normalized to G3PDH were set as 1. Retinas from three littermate mice of same genotype (n=3) were used for the study. *P<0.05.
Figure 6.
Expression of PLVAP and claudin-5 in WT, Lrp5-/-, Vldlr-/- and DKO retinas.
(A) Immunostaining of 3-week-old WT, Lrp5-/-, Vldlr-/- and DKO retinal frozen sections. An antibody against PLVAP (green) was co-stained with a vascular EC-specific Tie2 antibody (red). In the WT section, PLVAP expression was undetectable in retinal ECs. PLVAP staining signals was obviously increased in retinal ECs of both Lrp5-/- and DKO mice (arrowheads). However, the PLVAP staining signals was restrictively detectable only in abnormal overgrown vessels located in the ONL in the Vldlr-/- retina (arrowhead). Scale bar: 50µm. (B) Relative expression of PLVAP and claudin-5 transcripts in WT (W), Lrp5-/- (L), Vldlr-/- (V) and DKO (D) retinas examined by qRT-PCR. PLVAP expression was significantly increased in both Lrp5-/- and DKO retinas, as well as in Vldlr-/- retinas. Significantly reduced claudin-5 expression was observed in both Lrp5-/- and DKO retinas, while the claudin-5 expression in Vldlr-/- retinas seemed to be slightly increased although this change is not statistically significant. *p<0.05, n=3 for each genotype.
Figure 7.
A schematic summary of retinal vasculature phenotypes in Vldlr-/-, Lrp5-/-, and Vldlr-/- Lrp5-/- DKO mice.
Compared to the typical three-layer retinal vasculature network (green) in GCL, IPL and OPL of the wild-type control retina, the Vldlr-/- retina displays abnormal vessel growth into the photoreceptor nuclear layer, which originated from retinal vessels in the OPL layer; the Lrp5-/- retina has the surface vessel layer in the GCL but has no vascular networks extending to the IPL and OPL, and ECs form cell clusters in IPL. Similar to the Lrp5-/- retina, the DKO retina does not form a complete retinal vascular network except the surface layer in the GCL, and the DKO retina also forms large ECs clusters in IPL.