Mouse models of hereditary hemorrhagic telangiectasia: recent advances and future challenges

Abstract

Hereditary hemorrhagic telangiectasia (HHT) is a genetic disorder characterized by a multi-systemic vascular dysplasia and hemorrhage. The precise factors leading to these vascular malformations are not yet understood and robust animal models of HHT are essential to gain a detailed understanding of the molecular and cellular events that lead to clinical symptoms, as well as to test new therapeutic modalities. Most cases of HHT are caused by mutations in either endoglin (ENG) or activin receptor-like kinase 1 (ACVRL1, also known as ALK1). Both genes are associated with TGFβ/BMP signaling, and loss of function mutations in the co-receptor ENG are causal in HHT1, while HHT2 is associated with mutations in the signaling receptor ACVRL1. Significant advances in mouse genetics have provided powerful ways to study the function of Eng and Acvrl1 in vivo, and to generate mouse models of HHT disease. Mice that are null for either Acvrl1 or Eng genes show embryonic lethality due to major defects in angiogenesis and heart development. However mice that are heterozygous for mutations in either of these genes develop to adulthood with no effect on survival. Although these heterozygous mice exhibit selected vascular phenotypes relevant to the clinical pathology of HHT, the phenotypes are variable and generally quite mild. An alternative approach using conditional knockout mice allows us to study the effects of specific inactivation of either Eng or Acvrl1 at different times in development and in different cell types. These conditional knockout mice provide robust and reproducible models of arteriovenous malformations, and they are currently being used to unravel the causal factors in HHT pathologies. In this review, we will summarize the strengths and limitations of current mouse models of HHT, discuss how knowledge obtained from these studies has already informed clinical care and explore the potential of these models for developing improved treatments for HHT patients in the future.

Keywords: Bmp/Smad signaling; TGFβ signaling; angiogenesis; arteriovenous malformation; vascular development; vascular disease.

FIGURE 1

Role of endoglin and ACVRL1 in TGFβ family signaling. TGFβ superfamily ligands bind to specific serine/threonine kinase receptors and induce formation of a heteromeric receptor complex. In ECs, TGFβ1 ligand activates ACVRL1/ALK5/TGFBR2 receptor complexes leading to phosphorylation of SMAD1/5/8 which then interacts with SMAD4 protein and moves to the nucleus to regulate transcription of target genes. BMP9 ligand promotes Smad1/5/8-mediated signaling through ACVRL1-BMPR2 receptor complexes. Endoglin promotes signaling through ACVRL1, potentially by facilitating ligand binding to the signaling receptor complex. Under particular environmental conditions (e.g., during inflammation) the extracellular domain of endoglin is cleaved to form soluble endoglin (Sol-Eng), which may sequester circulating BMP9/10 ligands and reduce receptor binding. Sol-Eng has also been shown to inhibit binding of TGFβ to TGFBR2 (Venkatesha et al., 2006) potentially by a sequestering mechanism (as shown).

FIGURE 2

Floxed mice used to generate models of HHT. The floxed endoglin mouse (A) and the floxed Acvrl1 mouse (B) were generated for conditional knockout studies using Cre-LoxP recombination (Allinson et al., 2007; Park et al., 2008). The floxed endoglin mouse (A) was designed with the position of loxP sites (arrowheads) to allow conditional deletions of exons (boxed regions) 5 and 6 of the Endoglin gene, which also leads to a frameshift mutation in exon 7 to generate a truncated non-functional protein. Cre mediated recombination of the floxed Acvrl1 allele (B) leads to removal of exons 4–6 which includes the essential transmembrane domain. To achieve cell-type specific knockdown of endoglin or Acvrl1, these floxed mice were crossed with specific Cre-lines (see Tables 1 and 2).

FIGURE 3

Three event hypothesis for AVM formation in HHT. Individual endothelial cells in HHT patients may undergo local loss of endoglin or ACVRL1 protein either due to shedding or somatic mutation. If vessels containing these transient or permanent null cells are exposed to pro-angiogenic triggers then the cells may undergo further proliferation and contribute to vessel enlargement. We suggest that once blood flow increases in this enlarged vessel there is a positive feedback loop leading to further endothelial cell proliferation and vessel enlargement.

FIGURE 4

Arteriovenous malformations in mouse neonatal retinas following endothelial specific loss of Eng or Acvrl1 expression. Endothelial cells of the neonatal retinal are stained with isolectin to reveal normal two-dimensional architecture of the vascular plexus at postnatal day 7 (P7). Note the regular organization of arteries (a) and veins (v) with intervening capillary networks (A), and at higher magnification (B). Arteries are readily recognized by the capillary free zones around them (A,B). Following endothelial specific depletion of endoglin (by tamoxifen treatment of Eng^fl/fl; Cdh5-Cre^ERT2 neonates), Eng-iKO^e retinal vessels at P7 show delayed progression of the vascular plexus toward the edge of the retina as well as abnormal connections between arteries and veins to generate multiple AVMs [red arrows in (C), and shown in higher power images in (D)]. Endothelial specific loss of Acvrl1 leads to AVMs (red arrows in E,F), hyperbranching (F, asterisks) and enlargement of veins. Panels (A,C) are reproduced from a previous study where further details of the experimental approach can be found (Mahmoud et al., 2010; Tual-Chalot et al., 2014).

FIGURE 5

Stages of AVM development in mouse skin AVMs following loss of Acvrl1 combined with wounding. The novel connections between nascent arteries and veins can be seen by following hyperspectral imaging of the wound area in a dorsal window chamber in real time. The color bar indicates the relative oxygen saturation levels of the hemoglobin which allows tracking of arterial (red) and venous (blue) blood flow. Four novel AV shunts (AV1–4) appear over the course of 7 days following wounding. Arterial blood flowing into venous branches through AV shunts can be visualized. For instance, “blue” venous branches (marked by a blue asterisk or star) in day 2 (A) turned to “red” by connecting to adjacent arterial branches (marked by red asterisk for AV1 and star for AV2) in day 3 (B), indicating establishment of AV shunts. A portion of this data and further methodological details can also be found in a previous publication (Han et al., 2014).