Animal models of Alport syndrome

This is the authors' abstract. We don't add key points for this paper.

The last decade of the twentieth century was a very productive period for the study of Alport syndrome. Alport syndrome was shown to result from mutations in certain members of the type IV collagen family of proteins, the α3(IV), α4(IV), and α5(IV) chains. Several hundred different mutations in the COL4A5 gene, which encodes the α5(IV) chain, were described in patients with X‐linked Alport syndrome [1]. A few dozen mutations were found in the COL4A3 and COL4A4 genes, which respectively encode the α3(IV) and α4(IV) chains, in patients with autosomal recessive Alport syndrome [2,3]. Autosomal dominant Alport syndrome, due to heterozygous mutations in COL4A3 or COL4A4, was established as an entity, and distinguished from Fechtner and Epstein syndromes, which arise from mutations in a non‐collagen locus, MYH9 [4–7]. Investigators established that mutations in the α3(IV), α4(IV), or α5(IV) chain typically prevent expression of any of these proteins in glomerular basement membranes (GBM), because these chains must interact with each other in networks in order to be deposited in the GBM [8]. These chains are replaced in the GBM by α1(IV) and α2(IV) chains, which normally disappear, for the most part, from the GBM during the capillary loop stage of glomerular maturation [9–11]. Other extracellular matrix proteins, such as types V and VI collagen and the laminin α2 chain, are aberrantly expressed in the Alport GBM [9,12]. Despite these advances, key questions about Alport syndrome remain. Except perhaps for anterior lenticonus, for which it can be plausibly argued that reduced tensile strength of the lens capsule basement membrane prevents maintenance of normal lens shape, we have very little understanding of the pathogenesis of the extrarenal manifestations of Alport syndrome, in particular sensorineural deafness. We do not know how the Alport GBM loses its permselectivity, or why Alport kidneys fail. Of greatest importance, we do not have treatments of proven efficacy, and the nature of the disease—its relative rarity, its slow progression, and the strong influence of genotype on the severity of the Alport phenotype—will complicate efforts to conduct randomized clinical trials. Animal models of genetic disorders provide opportunities for investigating both pathogenesis and treatment of disease. Fortunately, several excellent animal models of Alport syndrome have been developed (Table 1). Thus far, these models are limited to spontaneous Alport syndrome in dogs and transgenic models of Alport syndrome created in mice. This review provides a brief discussion of these models and their applicability to investigations of the pathogenesis and treatment of Alport syndrome, and of the biology of type IV collagen. Animal models of Alport syndrome Animal models of Alport syndrome About 80% of patients with Alport syndrome have the X‐linked form of the disease. Two canine models of XLAS have been identified. The first of these was discovered in a strain of Samoyed dogs, and was found to result from a G to T substitution in exon 35 (of 51) of COL4A5, changing a glycine codon to a stop codon [13,14]. The second, found in a kindred of mixed‐breed dogs from Navasota, Texas [15], arises from a 10‐bp deletion in exon 9 of COL4A5 that creates a frame‐shift and a stop codon in exon 10 (Cox M., Lees G., Kashtan C., Murphy K., submitted for publication). Affected animals in both models display ultrastructural changes in the GBM, and alterations in the GBM expression of collagen IV α chains, that are indistinguishable from those found in human Alport patients. Overt proteinuria appears in affected males during the first 6 months of life, and uraemia by 12–18 months. Affected males do not have clinically detectable deafness or ocular abnormalities. At this time, there is no reported murine model of XLAS. A murine XLAS model would be valuable because XLAS is the predominant human form of the disease, and would facilitate investigation of the role of X‐inactivation in determining outcome in female heterozygotes. ARAS accounts for about 15% of patients with the disease. The existence of an inherited nephropathy in English cocker spaniels was known for many years, and in the 1980s this was shown to be a progressive glomerular disorder with autosomal recessive inheritance [16,17]. GBM of affected dogs showed changes similar to Alport syndrome by electron microscopy, and was found to lack the α3(IV) and α4(IV) chains [18,19]. Expression of the α5(IV) chain in the GBM is reduced, but not absent. The mutation responsible for ARAS in English cocker spaniels has not yet been identified, although efforts to sequence the COL4A3 and COL4A4 genes in normal and affected dogs are under way. Several transgenic models of ARAS have been developed. Two of these models involve knockouts of the murine col4a3 gene [20,21]. In the third model, the proximal portions of both the col4a3 and the col4a4 gene were deleted (this is possibly because these genes are located adjacent to each other, in a 5′ to 5′ orientation) [22]. The GBM of homozygous ARAS mice lacks the α3(IV), α4(IV), and α5(IV) chains, and by electron microscopy shows the basket‐weave transformation characteristic of human Alport syndrome [20–22]. Overt proteinuria and uraemia develop within a few months of birth in murine ARAS (Table 1). Hood and colleagues [23,24] have described an autosomal dominant nephropathy in bull terriers with GBM changes similar to those occurring in Alport syndrome. The mutation responsible for this disorder has yet to be identified. The GBM of affected dogs does not exhibit abnormal collagen IV α‐chain expression, but this does not exclude a heterozygous mutation in COL4A3 or COL4A4 as the cause of the disease. Both canine and murine models of Alport syndrome are associated with advantages and disadvantages. Canine breeding programmes are expensive to maintain, and gestation periods are relatively long. An advantage of canine models is the larger size of the animal, which will facilitate development of gene transfer protocols for the treatment of Alport syndrome that are technically applicable to the human disease [25]. Murine models have the advantage of a short gestation period. The murine forms of Alport syndrome progress very rapidly (Table 1). On the one hand, this minimizes the time needed to determine the effect of a therapeutic intervention. On the other hand, while prolongation of survival of a few weeks might reach statistical significance in a murine model, such an effect might be difficult to confirm in human subjects. Some apparent therapeutic effects may be strain‐specific, due to the influence of a predominant modifying gene allele (see Note added in proof). Animal models of Alport syndrome will be of great value to investigators interested in the functions of the α3(IV), α4(IV), α5(IV), and α6(IV) chains. The α1(IV) and α2(IV) chains have been essential components of multi‐cellular organisms for hundreds of millions of years [26]. The α3(IV)–α6(IV) chains are relatively recent additions to the mammalian protein repertoire, arising from duplication and modification of the α1(IV) and α2(IV) genes [27]. What purposes do the α3(IV)–α6(IV) chains serve? They are clearly not required for fetal survival or for the formation of kidneys, ears and eyes that are, from birth and for a substantial time thereafter, essentially normal in structure and function. The α3(IV)–α4(IV)–α5(IV) chains may protect GBM from proteolysis, since GBM composed of these chains is more resistant to proteolytic degradation than GBM containing only α1(IV)–α2(IV) chains [10]. The supramolecular assembly of α3(IV)–α4(IV)–α5(IV) protomers is stabilized by disulphide bonds not present in α1(IV)–α2(IV) networks, which may confer greater mechanical stability [8]. Adhesion of glomerular epithelial cells to GBM appears to be mediated, at least in part, by interaction between the α3β1 integrin, the major integrin expressed by these cells, and the α3(IV) chain [28]. Podocytes may be able to adhere to an α3(IV)‐deficient GBM, but maintenance of the fully differentiated podocyte phenotype may require interaction with the α3(IV) chain. In summary, animal models of Alport syndrome offer excellent opportunities for investigation of pathogenesis and treatment of the disease, and for answering basic questions about the roles of type IV collagen chains in the biology of the kidney, cochlea, eye, and other tissues. Andrews et al. [29] have reported that quantitative trait loci influence disease progression in murine Alport syndrome. The Navasota COL4A5 mutation was presented in abstract by Melissa Cox at the Keystone Symposium ‘Genotype to Phenotype Focus on Disease’, Santa Fe, New Mexico, February 19–24, 2002. Correspondence and offprint requests to: Clifford E. Kashtan MD, University of Minnesota Medical School, Department of Pediatrics, 420 Delaware Street SE, MMC 491, Minneapolis, Minnesota, 55455 USA. Email: kasht001@umn.edu Dr Kashtan's work is supported by the National Institutes of Health (DK57676).