A human-mouse chimera of the alpha3alpha4alpha5(IV) collagen protomer rescues the renal phenotype in Col4a3-/- Alport mice

Abstract

Collagen IV is a major structural component of basement membranes. In the glomerular basement membrane (GBM) of the kidney, the alpha3, alpha4, and alpha5(IV) collagen chains form a distinct network that is essential for the long-term stability of the glomerular filtration barrier, and is absent in most patients affected with Alport syndrome, a progressive inherited nephropathy associated with mutation in COL4A3, COL4A4, or COL4A5 genes. To investigate, in vivo, the regulation of the expression, assembly, and function of the alpha3alpha4alpha5(IV) protomer, we have generated a yeast artificial chromosome transgenic line of mice carrying the human COL4A3-COL4A4 locus. Transgenic mice expressed the human alpha3 and alpha4(IV) chains in a tissue-specific manner. In the kidney, when expressed onto a Col4a3(-/-) background, the human alpha3(IV) chain restored the expression of and co-assembled with the mouse alpha4 and alpha5(IV) chains specifically at sites where the human alpha3(IV) was expressed, demonstrating that the expression of all three chains is required for network assembly. The co-assembly of the human and mouse chains into a hybrid network in the GBM restores a functional GBM and rescues the Alport phenotype, providing further evidence that defective assembly of the alpha3-alpha4-alpha5(IV) protomer, caused by mutations in any of the three chains, is the pathogenic mechanism responsible for the disease. This line of mice, humanized for the alpha3(IV) collagen chain, will also provide a valuable model for studying the pathogenesis of Goodpasture syndrome, an autoimmune disease caused by antibodies against this chain.

Figure 1.

SDS-PAGE analysis of proteinuria. Lanes 1 to 4: Bovine serum albumin: 0.5, 1, 5, and 10 μg; lane 5: 8-week-old Col4a3^−/− mouse; lane 6: 8-week-old Col4a3^+/+ (wild-type control) littermate; lane 7: 8-week-old TG/Col4a3^−/− littermate. Heavy albuminuria was observed in Col4a3^−/− mice, but not in TG/Col4a3^−/−.

Figure 2.

Comparative histology of Col4a3^−/− (a) and TG/Col4a3^−/− littermate (b) mice at 8 weeks of age. Severe lesions, including glomerular sclerosis, glomerular crescents, tubular dilatations, and interstitial fibrosis were evident in the Col4a3^−/− mouse kidneys, whereas the renal morphology was preserved in the TG/Col4a3^−/− mice (trichrome light green stain). c to e: Electron microscopy of 8-week-old mice: Col4a3^+/+ mouse (c); Col4a3^−/− mouse (d); and TG/Col4a3^−/− littermate (e). In the Col4a3^+/+ mouse, the GBM was globally regular, except for some very focal thickening (arrow). Foot processes were normal. In the Col4a3^−/− mouse, an irregular thickening of the GBM with focal loss of foot processes was observed. In the TG/Col4a3^−/− mouse, the normal structure of the GBM and the foot processes were preserved, albeit very focal thickening of the GBM was observed (arrow). Original magnifications: ×200 (a); ×400 (b); ×4350 (c); ×4800 (d); ×4500 (e).

Figure 3.

Immunohistological analysis of human Col4a3^+/+ mouse (middle) and TG/Col4a3^+/+ mouse (right) kidneys using antibodies recognizing human (Hu) α3 and α4(IV) chains. Antibodies strongly stained the human GBM and distal tubule BM, but did not label wild-type Col4a3^+/+ control mouse BMs. A discrete GBM and focal TBM labeling was observed in TG/Col4a3^+/+ mice.

Figure 4.

Immunohistological analysis of Col4a3^+/+ mouse (left), Col4a3^−/− mouse (middle), and TG/Col4a3^−/− mouse (right) kidneys using antibodies recognizing human (Hu) and mouse (Ms) α1.α2(IV), human and mouse α3(IV), human and mouse α4(IV), human α3(IV), human α4(IV), mouse α4(IV), mouse α5(IV), and human and mouse α3/α4/α5(IV) containing native hexamers. The human α3-α4(IV) were expressed in the TG/Col4a3^−/− mouse GBM, but not in TBM. The mouse α4(IV) and α5(IV) chains were expressed in GBM and TBM in Col4a3^+/+ mice, but were absent in all renal BM [except focally in the Bowman capsule for α5(IV)] in Col4a3^−/− mice, and were expressed, along with human α3 and α4(IV), in the GBM in TG/Col4a3^−/− mice.

Figure 5.

Western blot analysis of collagen IV chain expression in the kidneys of Col4a3^+/+ and Col4a3^−/− mice, in the presence or absence of the human transgene (TG). Collagenase-solubilized NC1 hexamers from murine renal BMs were resolved by SDS-PAGE into component NC1 monomers (M) and dimers (D), transferred to PVDF membranes, and immunoblotted with chain-specific mAbs recognizing both human and murine NC1 domains (α1 to α5), as well as with mAbs specific for human α3 and α4(IV) NC1 domains (Hu α3, Hu α4). Human GBM hexamers were used as positive control.

Figure 6.

Affinity fractionation of murine NC1 hexamers from Col4a3^+/+ mice (top) and TG/Col4a3^−/− mice (bottom), using mAb 26-20 [which bound hexamers containing α3/α4/α5(IV) NC1 domains] and mAb 15-47 [which bound hexamers containing α1(IV) NC1 domain]. The composition of the NC1 hexamer fractions bound or not bound to the immobilized mAbs was analyzed by Western blot with chain-specific mAbs recognizing both human and murine NC1 domains (α1 to α5), as well as with mAbs specific for human α3 and α4(IV) NC1 domains (Hu α3, Hu α4).

Figure 7.

Immunohistological analysis of kidneys from Col4a3^+/+ mice (left), Col4a3^−/− mice (middle), and TG/Col4a3^−/− mice (right) using antibodies recognizing mouse α2 (α2LN), β1 (β1LN), and γ1 (γ1 LN) laminin chains, perlecan, and fibronectin (FN). These proteins are mainly expressed in the mesangium of control Col4a3^+/+ mice, but strongly stained in the GBM of Col4a3^−/− mice. In TG/Col4a3^−/− mice, the expression is comparable to control Col4a3^+/+ mice.

Figure 8.

Western blot analysis of collagen IV chain expression in the extrarenal BMs of Col4a3^+/+ and TG/Col4a3^−/− mice. Collagenase-solubilized NC1 hexamers from murine lung and bladder BMs were resolved by SDS-PAGE into component NC1 monomers (M) and dimers (D), transferred to PVDF membranes, and were immunoblotted with chain-specific mAbs recognizing both human and murine NC1 domains (α1 to α5), as well as mAbs specific for human α3 and α4(IV) NC1 domains (Hu α3, Hu α4).

Figure 9.

Top: Indirect immunofluorescence staining with Goodpasture autoantibodies of native (A–E) and acid urea-treated (F) kidney sections. The sections were from Col4a3^−/− mouse (A), human (B), TG/Col4a3^−/− mouse (C), and Col4a3^+/+ mouse (D–F). Under native conditions, antibodies bound in a linear manner to human and TG/Col4a3^−/− mouse GBM but not to Col4a3^+/+ mouse GBM. In Col4a3^+/+ mice, only podocytes were clearly labeled under native conditions, whereas GBM and TBM were stained after acid urea treatment. Bottom: The reactivities with Goodpasture antibodies of native (solid bars) and dissociated (open bars) NC1 hexamers from human GBM and mouse kidneys were compared by enzyme-linked immunosorbent assay. Under native conditions, the Goodpasture epitopes were much less accessible in the murine than in the human NC1 hexamers, consistent with the results of immunofluorescence staining.