|Home | About | Journals | Submit | Contact Us | Français|
Mutations in ACTN4, encoding the actin-binding protein α-actinin-4, cause a form of familial focal segmental glomerulosclerosis. We had developed two strains of transgenic mice with distinct alterations in the expression of α-actinin-4. One strain carried a human disease-associated mutation in murine Actn4 while a knockout strain did not express α-actinin-4 protein. Most adult homozygous Actn4 mutant and knockout mice developed collapsing glomerulopathy. Homozygous Actn4 mutant mice also exhibited actin and α-actinin-4-containing electron dense cytoplasmic structures that were present but less prominent in heterozygous Actn4 mutant mice and not consistently seen in wild type or knockout mice. Heterozygous Actn4 mutant mice did not develop glomerulosclerosis, but did exhibit focal glomerular hypertrophy and mild glomerular ultrastructural changes. The ultrastructural abnormalities seen in heterozygous Actn4 mutant mice suggest low-level glomerular damage which may increase susceptibility to injury caused by genetic or environmental stressors. Our studies show that different genetic defects in the same protein produce a spectrum of glomerular morphologic lesions depending upon the specific combination of normal and/or defective alleles.
Mutations in ACTN4, the gene encoding the actin binding protein α–actinin-4, can cause familial focal segmental glomerulosclerosis (FSGS) in humans (1). This familial form of FSGS exhibits autosomal dominant transmission and is highly, but not fully, penetrant. Individuals heterozygous for these mutations exhibit slowly progressive disease that often results in kidney failure. Morphologic changes in the kidney are typical of FSGS, and include segmental podocyte foot process effacement (1).
In order to study the role of α–actinin-4 protein in normal podocyte function and genetic disease in vivo, we developed two strains of transgenic mice with distinct alterations in the expression of α–actinin-4. One strain carries a specific mutation in the murine Actn4 gene encoding α–actinin-4, corresponding to the K255E disease-associated mutation in humans (2). The other strain exhibits complete absence of measurable α–actinin-4 expression (3). Animals homozygous for either of these altered Actn4 genotypes develop a rapidly progressive kidney disease phenotype characterized by increased perinatal lethality, severe glomerular disease, kidney failure, and death by approximately 12 weeks. The resulting disease phenotype in these homozygous animals is more aggressive than the heterozygous human disease, and the absence of normal α–actinin-4 protein in these animals suggests a distinct disease mechanism. Therefore, one objective of this study was to better characterize the morphologic pattern of kidney disease observed in these transgenic animals, to enhance our understanding of glomerular disease developing in the absence of wild type α–actinin-4.
The heterozygous Actn4 mutant animal represents a more genetically faithful model of the autosomal dominant ACTN4-mediated human disease, as this animal expresses both wild type and mutant α–actinin-4. Initial studies of young heterozygous Actn4 mutant mice did not reveal a kidney disease phenotype. This is not inconsistent with the human disease, as the first manifestations of kidney damage in humans are not apparent until early adulthood (1). However, a transgenic animal developed by Michaud et al. that expresses both endogenous wild type and a K256E-mutant α–actinin-4 transgene, showed focal glomerulosclerosis developing at approximately 10 weeks (4). Since humans with (heterozygous) ACTN4 mutations develop kidney disease later in life, it is plausible that heterozygous Actn4 mutant mice may also not develop overt kidney disease until later in life. Therefore, another objective of this study was to examine the kidneys of older mice, heterozygous for the Actn4 mutant allele, to look for features of slowly developing glomerular disease that were not apparent in younger mice.
Kidneys from 23 wild type (WT), 17 heterozygous Actn4 K256E mutant (HET), 12 homozygous Actn4 K256E mutant (KI), and 18 homozygous Actn4 null (KO) mice, age range 1 to 60 weeks, were evaluated by light microscopy. Only WT and HET mice were available at later time points; KI and KO mice did not survive past 21 and 15 weeks, respectively.
No significant morphologic changes were seen among juvenile (age 0 to 7 weeks) or young adult (age 7 to 21 weeks) WT mice (Figures 1 and 2a,b). Examination of kidneys from old adult (age > 21 weeks) WT mice revealed minimal focal glomerular hypertrophy, but no other glomerular lesions (Figure 1). HET mice showed no evidence of overt glomerular injury, but did show focal glomerular hypertrophy at all ages, occasionally accompanied by focal hyaline cast formation in the tubules, or focal interstitial inflammation (Figures 1 and 2c,d).
Glomerular injury (segmental or global sclerosis or collapse) was observed in 4 of 12 KI mice as early as 5.4 weeks, including 1 of 6 juvenile mice, and 3 of 6 young adult mice. Similarly, 9 of 18 KO mice showed glomerular injury as early as 3.7 weeks, including 4 of 12 juvenile and 5 of 6 young adult mice. Among the homozygous mice not showing overt glomerular injury, 5 of 8 KI mice and 2 of 9 KO mice (including all young adult KI and KO mice) showed focal glomerular hypertrophy involving a small proportion of glomeruli, beginning as early as 4.4 and 3.4 weeks, respectively (Figure 2e,g). Three juvenile KI mice and seven juvenile KO mice showed no glomerular lesions. Averaged over all mice in each category, 2.2% and 52.3% of glomeruli in juvenile and young adult KI mice, respectively, and 15.7% and 70.4% of glomeruli in juvenile and young adult KO mice, respectively, showed some form of glomerular lesion (Figure 1).
Of the homozygous mice exhibiting overt glomerular injury, 4 of 4 KI mice (youngest 5.4 weeks) and 8 of 9 KO mice (youngest 3.7 weeks) showed a histologic pattern of collapsing glomerulopathy. Features of collapsing glomerulopathy included: focal glomerular lesions composed of segmental or global glomerular capillary collapse and epithelial cell proliferation, focally prominent protein reabsorption granules in glomerular and tubular epithelial cells, and tubular “microcysts” (Figures 2f,h and and3).3). One KO mouse (6.7 weeks) exhibited diffuse segmental and global glomerulosclerosis, but without collapsing lesions. Within age groups, collapsing lesions involved a greater proportion of glomeruli in KO mice than in KI mice. Averaged over all mice in each category, glomerular collapsing lesions involved 0.1% and 7.0% of glomeruli in juvenile and young adult KI mice, respectively, while 4.7% and 26.9% of glomeruli in juvenile and young adult KO mice, respectively, were involved by collapsing lesions (Figure 1). Homozygous mice also showed a varying degree of segmental and global glomerulosclerosis in addition to the collapsing lesions. Averaged over all mice in each category, sclerosing lesions (segmental or global) in KO mice involved 10.6% of glomeruli in juveniles and 43.1% of glomeruli in young adults (Figure 1). In KI mice, no sclerosing lesions were seen in juveniles, while 43.4% of glomeruli in young adults exhibited sclerosing lesions (Figure 1). On average, sclerosing lesions involved a greater proportion of lesional glomeruli in young adult KI mice, as compared to KO mice (Figure 1).
Kidneys with evidence of glomerular injury also exhibited a proportional degree of tubulointerstitial damage (Figure 2f,h). The tubulointerstitial damage was characterized by mononuclear interstitial inflammation and features of acute tubular injury (e.g., epithelial flattening, loss of brush border, cytoplasmic vacuolization, and rare mitotic figures). Prominent tubular microcyst formation was noted in all homozygous KO kidneys where collapsing lesions were identified, as well as in the previously noted 6.7 week homozygous KO specimen with diffuse global and segmental glomerulosclerosis that did not show collapsing lesions. Microcysts were not as prominent in the homozygous KI mice, and were only observed in two of the four specimens with collapsing lesions. Hyaline casts were seen in almost all homozygous specimens with glomerular damage. Chronic tubulointerstitial changes (advanced tubular atrophy and interstitial fibrosis) were a minor feature of the tubulointerstitial damage, even in cases with advanced glomerular disease. No significant vascular lesions were observed in any animals.
A total of 18 WT, 21 HET, 12 KI, and 17 KO animals, ranging in age from 1 to 93 weeks, were evaluated at the ultrastructural level. Of the WT animals, 8 of 18 showed no ultrastructural abnormalities (Figures 4 and 5a,b). However, 8 of 18 showed mild glomerular basement membrane (GBM) abnormalities as early as 6.3 weeks, but most commonly in adult animals (Figures 4a and and5c).5c). GBM abnormalities consisted of segmental areas of subepithelial thickening and redundancy, often protruding from the subepithelial surface. A smaller subset of WT animals (5 of 18) revealed mild podocyte foot processes (FP) abnormalities (Figure 4b). Notably, one of the WT mice showing mild FP abnormalities was the youngest mouse in this group (age 1.4 weeks). Only one older adult WT animal (60 weeks) showed mild podocyte cell body (PCB) abnormalities (Figure 4c).
HET animals showed a pattern of ultrastructural findings similar to WT animals, but abnormalities were somewhat more severe, and affected a greater proportion of animals (Figures 4 and 5d-f). The GBM in 15 of 21 HET animals, including 13 of 14 adults, showed mild to moderate abnormalities, again characterized by segmental subepithelial thickening and redundancy (Figures 4a and 5e,f). FP in HET animals were mildly but almost uniformly abnormal; 16 of 21 animals including all adults showed segmental effacement and morphologic irregularity (Figures 4b and 5e,f). The difference in mean FP abnormality scores between WT and HET mice was statistically significant for both the young adult (p=0.05) and older adult (p=0.01) cohorts (Figure 4b). PCB also exhibited mild ultrastructural abnormalities in 12 of 21 HET animals (Figures 4c and 5e,f). The difference in mean PCB abnormality scores between the WT and HET mice in the older adult cohort was also statistically significant (p=0.03) (Figure 4c).
Nearly all KI mice (11 of 12), with the exception of the youngest KI mouse evaluated at the ultrastructural level (2.1 weeks), showed GBM abnormalities. These changes were mild in the juvenile KI mice, but were moderate to severe in the adult KI mice, and again consisted of segmental subepithelial thickening and redundancy, often with numerous protrusions (Figures 4a and 5g,h). The difference in mean GBM abnormality scores between WT and KI mice was statistically significant for both juvenile (p=0.03) and young adult (p=0.01) cohorts (Figure 4a). Podocytes in almost all KI mice (11 of 12), again excepting the youngest (2.1 week) mouse, exhibited progressive degenerative changes that are characteristic of primary podocyte injury, including FP and PCB abnormalities (Figure 5g,h). Juvenile KI mice showed a variable but overall mild degree of podocyte degenerative changes, first detected at 4 weeks, while all young adult KI mice showed more severe changes (Figure 4b,c). The difference in mean FP and PCB abnormality scores between WT and KI mice were both statistically significant (p<0.01) for the young adult cohort (Figures 4b,c). In a small proportion of glomeruli from adult KI mice examined at the ultrastructural level, ultrastructural changes characteristic of collapsing glomerulopathy were observed, including capillary collapse and wrinkling of the GBM, diffuse FP effacement, loss of primary processes and cell simplification, distension of Bowman’s space by cuboidal cells, and accumulation of electron-dense lysosomes. In all of the juvenile and adult KI mice in which podocyte injury was present, a limited degree of actin condensation was noted adjacent to the GBM. Basal actin condensation was always focal, mild, and much less extensive than the degree of foot process effacement, with many effaced segments showing no actin condensation.
Most KO mice (14 of 17) showed moderate to severe GBM abnormalities (Figures 5i,j). The difference in mean GBM abnormality score between WT and KO mice was statistically significant for both juvenile (p=0.01) and young adult (p<0.01) cohorts (Figure 4a). Podocytes in nearly all KO mice (16 of 17) exhibited degenerative changes similar to those seen in KI mice (Figure 5i,j). The difference in both FP and PCB abnormality scores between WT and KO mice was statistically significant (p<0.01) for both juvenile and young adult cohorts (Figures 4b,c).
Glomerular ultrastructural abnormalities in KO mice, although qualitatively similar to KI mice, were characterized by more severe changes at an earlier age, as compared to KI mice, and more frequently included ultrastructural changes characteristic of collapsing glomerulopathy (Figures 4 and 5g-j). One exception to this pattern was the tendency for basal actin condensation in injured podocytes to appear focal and mild, as in KI mice.
A unique ultrastructural finding seen primarily in KI mice was the presence of cytoplasmic electron densities (Figures 5g,h and and6).6). These structures were seen in all KI mice regardless of age or degree of podocyte injury. Almost all HET mice (except two 4 week-old mice) demonstrated these structures, with small but diffusely distributed electron densities apparent in most adult HET mice. These structures were also observed sporadically and more focally in older WT mice, and focally in KO mice exhibiting concurrent severe podocyte injury. These irregularly-shaped, uniformly electron-dense structures ranged from 100 nm to greater than 1 μm in greatest dimension, and were usually associated with the cell membrane in podocyte cell bodies or major processes. These structures were not observed in any other cell type examined, including mesangial cells, endothelial cells, parietal glomerular epithelium, vascular smooth muscle, tubular epithelium, erythrocytes, platelets, and leukocytes.
Immunogold transmission electron microscopy was performed to better characterize the composition of these electron dense structures. Transmission electron micrographs of immunogold-stained sections from juvenile and adult KI mouse kidneys revealed high concentrations of actin and α–actinin-4 together in these electron dense structures in podocytes (Figure 6). Lower concentrations of β-actin and α-actinin-4 were also observed in podocyte foot processes and the cytoplasm of endothelial cells.
Kidneys from eighteen mice (6 WT, 5 HET, 4 KI, and 3 KO), age 4.6 to 70.1 weeks, were examined and morphometrically evaluated at the light microscopic level, using immunohistochemistry for markers of podocyte injury, differentiation and proliferation. KI and KO mice exhibited an immunohistochemical phenotype characteristic of collapsing glomerulopathy. In most KI mice and all KO mice, morphometric analysis revealed loss of expression of markers of podocyte differentiation, including WT1 and synaptopodin, and increased expression of desmin (a marker of podocyte injury) in extracapillary glomerular cells (Table 1, Figure 7). Additionally, all KI and KO mice revealed increased expression of markers of proliferation including cyclin D1 and Ki-67 (Table 1, Figure 7). Although one KI mouse (2373) did not exhibit loss of podocyte markers WT1 and synaptopodin, or evidence of podocyte injury in the form of desmin staining, this mouse did show increased staining for cyclin D1 and Ki-67 in podocytes, suggesting early activation of proliferative mechanisms. In contrast to the KI and KO mice, the staining pattern in HET mice, including specimens as old as 60.1 and 70.1 weeks, was comparable to that in WT mice (Table 1, Figure 7). Endocapillary cells (endothelial cells and/or mesangial cells) were weakly positive for cyclin D1 in all mice, as has been reported previously by other investigators (5).
These findings demonstrate that the homozygous Actn4 K256E mutant (KI) and Actn4 null (KO) mice developed in our laboratory represent genetically new and distinct models of collapsing glomerulopathy. Several murine models of collapsing glomerulopathy have been described, including transgenic animals exhibiting podocyte-specific over expression of vascular endothelial growth factor (VEGF) (6), podocyte expression of HIV genes (7-9), and mitochondrial defects (5). Some of these murine models also demonstrate a role for glomerular epithelial proliferation in the pathogenesis of the characteristic glomerular lesions, as in humans (5,9-11). In the Actn4 KO and KI mice, collapsing glomerulopathy is associated with the absence of wild type α-actinin-4, a cytoskeletal protein. How might absence of wild type α-actinin-4 precipitate collapsing glomerulopathy, even if a mutant protein is expressed? One possibility is that the collapsing lesion may be a stereotypical response to certain forms of severe and acute podocyte injury or loss. Work in our laboratory suggests that the absence of α-actinin-4 may be associated with increased podocyte shedding (12). Thus, lack of wild type α-actinin-4 protein may lead indirectly to the collapsing phenotype, through a mechanism that promotes podocyte detachment and loss.
Our results show considerable variability in the timecourse of glomerular lesions in the KI and KO mice. For instance, although almost all KO mice with glomerular lesions exhibited collapsing lesions, a single juvenile KO mouse (age 6.7 weeks) exhibited sclerosing lesions involving most glomeruli, without morphologically distinctive collapsing lesions. At the other extreme, a small subset of KI and KO animals in the 7 to 21 week cohort exhibited few or no glomerular lesions. These findings suggest variability in the rate of progression from “active” collapsing lesions to glomerulosclerosis. This phenotypic heterogeneity may be secondary to genotypic variation in the mixed background mice used for this study. Further studies with pure strains may help to clarify the nature of this phenotypic variability.
Although the overall pattern of kidney damage observed in the KI and KO mice was similar, there are two distinguishing features. One consistent feature of our results was the tendency for KO mice to show features of glomerular injury that were more prominent and earlier appearing than that observed in KI mice (Figures 1 and and4).4). This could be accounted for by the presence of mutant α-actinin-4 in KI mice, which may retain some of the function of wild type α-actinin-4. A second finding that distinguished the KI mice from the other genotypes was the presence of prominent intracytoplasmic electron densities in all KI mice. Similar but smaller structures were also observed in almost all HET mice, and although they generally remained small, they became more diffusely distributed with age. These structures were also observed sporadically and more focally in older WT mice, and sporadically in KO mice with severe podocyte injury. Immunogold staining demonstrates that the aggregates in KI mice contain both β-actin and (mutant) α-actinin-4 (Figure 6). Work in our laboratory suggests that mutant α-actinin-4 exhibits increased binding affinity for actin (1,13). Therefore, these aggregates may represent accumulations of bound actin and α-actinin-4 within the cytoplasm resulting from increased actin binding, mediated by mutant α-actinin-4.
In contrast to the KI and KO animals, the heterozygous mutant (HET) animals did not develop overt glomerular lesions. However, there were several minor findings noted in HET animals. Although focal glomerulosclerosis was absent, a small proportion of glomeruli in HET animals were hypertrophic at all ages (Figures 1 and 2c,d). Mild ultrastructural abnormalities were also noted in older HET mice, including mild foot process abnormalities and mild but diffuse electron dense aggregate accumulation (Figures 4 and and5f).5f). In comparison, wild type (WT) animals did not exhibit the same degree of glomerular hypertrophy or ultrastructural abnormalities as HET mice (Figures 1, 2a,b, ,4,4, and 5a-c). Although mild histologic and ultrastructural changes are noted in the glomeruli of HET mice, immunohistochemistry in HET mice revealed a staining pattern consistent with podocyte differentiation and comparable to that observed in WT mice, even up to age 70.1 weeks (Table 1). Thus, unlike KI and KO mice, the changes developing with time in HET mice do not appear to involve podocyte “de-differentiation”. The mild glomerular changes seen in HET mice may represent the cumulative effect of an intrinsic cellular abnormality associated with the presence of mutant α-actinin-4. The consistent finding of cytoplasmic electron densities in older HET animals supports this. Alternatively, the observed podocyte ultrastructural changes may be an indirect result of other preceding functional changes that lead to adaptive glomerular hypertrophy early in life, and podocyte structural changes that are observable at the ultrastructural level later on. Further studies, potentially using podocyte-specific conditional transgenic animals, will be required to clarify the nature of any functional deficit associated with the Actn4 mutation.
Another transgenic mouse model exhibiting podocyte-specific over expression of K256E mutant α-actinin-4 protein has been shown to develop features of podocyte injury and focal glomerulosclerosis in a subset of proteinuric animals at age 10 weeks (4). Why do these transgenic animals, with heterozygous expression of mutant and wild type Actn4, often exhibit glomerular disease early in life, in contrast to our HET animals? One key finding from the Michaud study is that the relative expression level of mutant α-actinin-4 in their mice correlated with the degree of proteinuria and podocyte injury. In comparison, proteinuria was only observed in a subset of our HET animals greater than 42 weeks of age (data not shown). This suggests that the HET animals in our study may exhibit relatively low levels of mutant Actn4 expression, in comparison to the proteinuric transgenic animals of Michaud et al. The development of a mutant-specific antibody that might be useful for direct quantification of protein expression levels in our HET animals is essentially impossible, as there is only a single amino acid difference between wild-type and mutant α-actinin-4. Therefore it is not possible to compare expression of mutant and wild-type protein directly. Another important difference between these animal models involves control of gene expression. In the heterozygous mutant (HET) mice we describe, both the mutant and wild-type alleles are under the control of the endogenous promoter. In the transgenic model that Michaud described, mutant Actn4 expression is under the control of a podocyte-specific promoter. In these mice, both of the normal wild-type Actn4 alleles remain intact. Thus there are important differences in the number of copies of both the mutant and wild-type alleles in these models and in the way that the expression of the mutant alleles is regulated. It seems very likely that these differences will result in differences in quantities of expressed protein and in the temporal pattern and degree of variability of protein expression. Indeed, the Michaud paper reported a good deal of variability in the phenotype of the transgenic mice derived from different founders.
GBM abnormalities were observed as a feature of aging in all genotypes, but were more prominent in KI and KO mice, and tended to correlate with the extent of podocyte injury (Figures 4 and and5).5). Distinct structures composed of redundant GBM were often seen protruding from the subepithelial surface, and the podocyte processes overlying these structures almost always exhibited loss of foot process organization. Similar structures have been described in another transgenic mouse model that exhibits podocyte alterations (14). The presence of foot process disorganization in areas where GBM abnormalities occur in KI and KO mice suggests that the GBM abnormalities may be a manifestation of podocyte injury, perhaps indicating areas where loss of foot process organization has resulted in uneven deposition of basement membrane material. Thus, although these structures may arise with time in all mice, it is conceivable that they might occur with increased frequency in animals with overt and widespread podocyte injury, such as our KI and KO mice.
Since the HET mice represent the most faithful genetic model of ACTN4-mediated human disease, it is of interest to compare the morphologic findings in the HET mice to those seen in the kidneys of humans with ACTN4 mutations. Four biopsies from humans who carry ACTN4 mutations and exhibit clinical evidence of kidney disease were available for our examination. In contrast to the HET mice, all four human kidneys show focal to extensive global and segmental glomerulosclerosis, mild to moderate chronic tubulointerstitial damage, and segmental podocyte injury (foot process effacement, microvillous change, etc.) However, there are some common findings in the murine and human mutant kidneys. One similarity is the presence of mild alterations of the glomerular basement membrane, seen in all human biopsies, consisting of segmental redundancy of the lamina densa (i.e. “double contours”). The second similarity is the presence of mild to extensive accumulation of intracytoplasmic electron densities in all four of the human biopsies, which were morphologically identical to those seen in the transgenic mice.
As with the mice, the extent of involvement was variable, with the most prominent aggregates observed in a patient with morphologically mild disease. To our knowledge, such structures have not been reported in human collapsing glomerulopathy, or in other forms of podocyte injury. A third similarity was the presence of focal glomerular hypertrophy in 3 of 4 human cases. This finding was most pronounced in two patients with moderate chronic damage. Of the two remaining patients, both of whom exhibited mild (and presumably early) kidney damage, one patient also showed focal glomerular hypertrophy, while the other did not. In the two patients with more advanced chronic damage, the glomerular hypertrophy may represent a secondary compensatory response to loss of functional renal mass. It would be necessary to examine kidney tissue from these patients earlier in the disease course to determine if focal glomerular hypertrophy is present in the intact kidney, as is the case in the HET mice. Nevertheless, these similar findings are consistent with the notion that the human disease pathology observed represents further progression of the same mechanism operating in the murine HET model.
In humans that carry ACTN4 mutations, functional glomerular disease is rarely observed before early adulthood, and there is considerable variability in the age of onset and penetrance of glomerular disease (1). Thus additional modulating factors (genetic or otherwise) appear to play a role in the development of ACTN4-mediated human disease. Overt glomerular disease does not develop spontaneously in HET mice up to at least two years of age, suggesting the absence of necessary modifying factors in the laboratory animal. The findings of Michaud et al. suggest that variability in expression levels of mutant ACTN4 could contribute to phenotypic heterogeneity in Actn4 transgenic mouse models, and correspondingly in humans (4). Currently, specific genetic modifiers of human ACTN4-mediated disease have not been conclusively identified. However, if such factors are identified in the future, it may be possible to study their effect using the HET animal. Similarly, environmental, behavioral or physiologic factors, such as hypertension or diet, may be studied in the HET mouse. Further studies such as these may show that additional “stressors” are required to precipitate overt kidney disease in heterozygous Actn4 mutant mice.
In summary, our findings show that genetic defects affecting a single cytoskeletal protein, α-actinin-4, can cause different morphologic patterns of injury, depending upon the nature of the genetic defect. Further study of the mechanism by which Actn4 deficiency leads to podocyte injury and glomerular collapse, and comparison with other models of glomerular collapse will help to better define the basis of this lesion. Our findings also suggest that additional modifying factors play a role in the development of glomerular disease in the heterozygous Actn4 mutant animal. This may account for phenotypic variability seen in human ACTN4-mediated disease. Subsequent studies will be required to identify these modifying factors, and Actn4 transgenic animal models will play an important role in that process.
Previously, we described the development of genetically distinct Actn4 K256E mutant and Actn4 null transgenic strains (2,3). Both of these strains, bred on a mixed 129/SvJ and C57BL/6 genetic background, were used for these studies. A total of 123 mice were examined. Mice were genotyped using PCR amplification of DNA isolated from mouse tails (3).
All experimental procedures were performed in accordance with the guidelines of the Institutional Animal Care and Use Committee at Harvard University. Transgenic mice exhibiting any of four genotypes including wild type (WT), heterozygous K256E mutant Actn4 (HET), homozygous K256E mutant Actn4 (KI), and homozygous Actn4 null (KO) were sacrificed at various ages from 1 to 93 weeks, so as to obtain an evenly distributed age representation of each genotype across the life cycle. Sampling was more concentrated in the juvenile (0 to 7 weeks) and young adult (7 to 21 weeks) age ranges, where all four genotypes were represented (KI and KO mice did not survive beyond 21 weeks), and where the most dramatic and rapidly evolving changes were noted. Kidneys were harvested, divided and placed in formalin, Bouin’s and Karnovsky’s fixatives for light and electron microscopy. Selected specimens retained for immunogold studies were fixed in paraformaldehyde/lysine/periodate fixative according to standard protocol.
Fixed kidney tissue was paraffin-processed, and 4 μm sections were stained with H&E and PAS stains. Light microscopic evaluation included quantification of the total number of glomeruli present in one microscopic section, as well as quantification of lesional glomeruli (global sclerosis, segmental sclerosis, glomerular collapse, and glomerular hypertrophy). Glomerular size (i.e., normal vs. hypertrophic glomeruli) was evaluated subjectively. The tubulointerstitial and vascular compartments were also evaluated.
Fixed specimens of kidney cortex were trimmed into small fragments, post-fixed with osmium tetroxide, dehydrated in serial ethanols, and embedded in epoxy resin. Semi-thin sections were cut at 1 μm and stained with toluidine blue for light microscopic examination. Ultra-thin sections were cut at 80 nm, mounted on 200 mesh copper grids, treated with uranyl acetate and lead citrate, and examined in a JEOL 1010 transmission electron microscope (Tokyo, Japan).
Electron micrographs showing glomerular ultrastructure (500x to 50,000x) were collected and evaluated in a blinded fashion by two kidney pathologists (J.M.H. and S.A.) At least three mature glomeruli from the deep cortex were evaluated in each specimen. For each specimen, scores were assigned from zero (no abnormality) to 4 (severe abnormality) for two categories of ultrastructural changes indicating podocyte injury, including: abnormalities of the podocyte cell body, such as microvillous degeneration, vacuolization, and increased number of cytoplasmic organelles; and abnormalities of the podocyte foot process, including effacement and morphologic irregularity. Glomerular basement membrane abnormalities were also scored for each specimen in a similar fashion, as a third category of ultrastructural changes. Within each age group, differences between mean scores for WT and other genotypes were compared using Student’s t-test assuming unequal variances. Other ultrastructural changes in the podocyte, including actin condensation at the basal foot processes, and the presence of cytoplasmic electron dense aggregates, were assessed non-quantitatively.
Formalin-fixed mouse kidney tissue was paraffin-processed and sectioned at 4 μm. After processing for antigen retrieval, sections were treated with antibodies against WT-1 (clone 6F-H2), synaptopodin (mouse monoclonal, gift of Dr. Peter Mundel, Mount Sinai School of Medicine, New York, NY), desmin (clone D33, Dako, Carpenteria, CA), cyclin D1 (clone SP4, Abcam, Cambridge, MA) or Ki-67 (polyclonal, Vector, Burlingame, CA), followed by ABC-HRP (Vectastain Elite ABC Kit, Vector Laboratories). All sections were counterstained with hematoxylin. Immunohistochemical phenotype was evaluated by determining the proportion of glomeruli in the section exhibiting the altered podocyte phenotype characteristic of collapsing glomerulopathy. Glomeruli were considered to exhibit an altered podocyte phenotype if they exhibited desmin in at least one glomerular segment, cyclin D1 or Ki-67 in one or more extracapillary (i.e., podocyte) cell nuclei in the glomerular tuft, absence of WT-1 in extracapillary nuclei of at least one glomerular segment, or absence of synaptopodin throughout the glomerulus. Endocapillary (i.e., mesangial or endothelial) cell staining for desmin, cyclin D1 or Ki-67 was not considered to represent an altered podocyte phenotype. No fewer than 40 glomeruli were counted in any section.
Mouse kidney tissue retained for immunogold electron microscopy was trimmed, dehydrated in serial ethanols, and embedded in LR White resin (EMS, Hatfield, PA). Ultra-thin sections were cut at 90 nm and collected on formvar-coated grids. Sections were blocked for 10 min with 5% normal goat serum (Sigma, St. Lois, MO), and incubated for 1 h on anti-β-actin (Sigma) or anti-α-actinin-4 (gift of A. Beggs) diluted in DAKO diluent (Carpinteria, CA.) Antibodies were labeled with goat anti-mouse/anti-rabbit IgG gold, 10 nm and 15 nm, respectively (Ted Pella, Redding, CA) in DAKO diluent for 1 h. Grids were stained with 2% aqueous uranyl acetate for 10 min, rinsed and air dried. Sections were examined and digital photomicrographs recorded using a JEOL 1011 transmission electron microscope.
This work was supported by NIH research grant DK66017 (to M.P.), NIH NRSA Fellowships T32HL07627 and F32DK72625 (to J.H.), an Educational Grant from the Ministry of Health of the State of Kuwait (to S.A.), a Research Fellowship from the NKF (to A.W.), and a Predoctoral Fellowship from the AHA (to S.D.) We thank M. McLaughlin of the EM Core of the Program in Membrane Biology, Massachusetts General Hospital, Boston, MA, and A. Shahsafaei of the Department of Pathology, Brigham and Women’s Hospital, Boston, MA, for technical assistance with immunohistochemical studies. The P.M.B. EM Core is supported by an NIH Center for the Study of Inflammatory Bowel Disease Grant (DK43351) and an NIH Boston Area Diabetes Endocrinology Research Center Award (DK57521). We also thank Dr. A. Beggs (Children’s Hospital, Boston, MA) for the gift of anti-α-actinin-4 antibodies, and Drs. H. Rennke (Brigham and Women’s Hospital, Boston) and L. Barisoni (New York University) for advice.
Potential conflicts of interest: None.