mVps34 Deletion in Podocytes Causes Glomerulosclerosis by Disrupting Intracellular Vesicle Trafficking

Recent studies have suggested that autophagy is a key mechanism in maintaining the integrity of podocytes. The mammalian homologue of yeast vacuolar protein sorting defective 34 (mVps34) has been implicated in the regulation of autophagy, but its role in podocytes is unknown. We generated a line of podocyte-specific mVps34-knockout (mVps34^pdKO) mice, which were born at Mendelian ratios. These mice appeared grossly normal at 2 weeks of age but exhibited growth retardation and were significantly smaller than control mice by 6 weeks of age, with no difference in ratios of kidney to body weight. mVps34^pdKO mice developed significant proteinuria by 3 weeks of age, developed severe kidney lesions by 5–6 weeks of age, and died before 9 weeks of age. There was striking podocyte vacuolization and proteinaceous casts, with marked glomerulosclerosis and interstitial fibrosis by 6 weeks of age. Electron microscopy revealed numerous enlarged vacuoles and increased autophagosomes in the podocytes, with complete foot process effacement and irregular and thickened glomerular basement membranes. Immunoblotting of isolated glomerular lysates revealed markedly elevated markers specific for lysosomes (LAMP1 and LAMP2) and autophagosomes (LC3-II/I). Immunofluorescence staining confirmed that the enlarged vacuoles originated from lysosomes. In conclusion, these results demonstrate an indispensable role for mVps34 in the trafficking of intracellular vesicles to protect the normal cellular metabolism, structure, and function of podocytes.The podocyte plays an essential role in establishment of the size- and charge-selective permeability of the glomerular filtration barrier and in maintenance of the glomerular structural integrity. Although alterations in structural proteins of the podocyte are now recognized to contribute to kidney disease, much of the podocyte’s functions remain incompletely understood.^1,2Autophagy is a tightly regulated intracellular process in which portions of cytoplasm, including proteins and organelles, are sequestered within double-membrane vesicles termed autophagosomes and are delivered to lysosomes for degradation and recycling of cellular components.^3,4 Mammalian cells are postulated to use autophagy as a mechanism for turnover of long-lived proteins and removal of protein aggregates and damaged organelles, and as a survival strategy under metabolic stress, including conditions of nutrient deprivation.^3,5 Recent studies suggest that autophagy is a key mechanism maintaining the homeostasis and integrity of podocytes.^6–9The vacuolar protein sorting defective 34 (Vps34) was originally cloned from yeast and found to be essential for the sorting of hydrolases to the yeast vacuole.¹⁰ It was subsequently identified as the only phosphatidylinositol 3-kinase (PI3K) in yeast.¹¹ The mammalian homologue of yeast Vps34 (mVps34) is also known as class III PI3K. Unlike the class I and class II PI3Ks, the class III PI3K, mVps34, can use only phosphatidylinositol as a substrate to generate a single product, phosphatidylinositol-3-phosphate, by specifically phosphorylating the D-3 position on the inositol ring of phosphatidylinositol.^11,12 Of interest, mVps34 has been implicated in the regulation of autophagy,^13–15 but the role of mVps34 in podocytes has not previously been explored.To determine the potential roles of mVps34 in podocytes, we inactivated the mouse mVps34 gene, Pik3c3, by creating a Pik3c3/mVps34-floxed mouse (mVps34^flox/flox). Figure 1A depicts the structure of the mVps34 gene–targeting vector and the Cre-LoxP strategy for generating a conditional mVps34 knockout mouse. The LoxP sites were inserted to delete Pik3c3 exons 20–21, which code for the entire catalytic core and the key AsparDic acid-Fenylalanine-Glycine (DFG) motif,¹⁶ and the targeting vector was designed to render all the distal exons out of reading frame so the catalytic domain and the ATP binding domain of this kinase were deleted. As shown in Figure 1B, to delete mVps34 in podocytes, we crossed our mVps34^flox/flox mice with a podocin-Cre mouse (Pod-Cre), which expresses Cre-recombinase exclusively in podocytes starting from the capillary loop stage during glomerular development,¹⁷ and generated Pik3c3/mVps34^flox/flox;Pod-Cre⁽⁺⁾ mice (subsequently called mVps34^pdKO), which were compared with their sex-matched littermates with a genotype of Pik3c3/mVps34^flox/flox;podocin-Cre⁽⁻⁾ as control mice (hereafter called mVps34^Ctrl). PCR of the genomic DNA from isolated glomeruli verified expected homologous recombination of mVps34 gene by podocin-Cre in mVps34^pdKO, but not in mVps34^Ctrl (Figure 1C). Immunoblotting confirmed significant deletion of mVps34 protein in the glomeruli of mVps34^pdKO mice (Figure 1D).Open in a separate windowFigure 1.Generation of an mVps34 gene (Pik3c3)-floxed mouse and deletion of mVps34 selectively in podocytes. (A) Strategy for making a conditional mVps34 knockout mouse. (B) Generation of renal glomerular podocyte-specific mVps34 knockout (mVps34^pdKO) mice. (C) Podocin-Cre–mediated homologous recombination of mVps34 gene verified by PCR using isolated glomerular genomic DNA as template. (D) Effective deletion of mVps34 protein by podocin-Cre recombinase confirmed by immunoblotting analysis of isolated glomerular homogenates. FLPe, enhanced FLP; WT, wild type.Homozygous mVps34^pdKO pups were born at expected Mendelian ratios (data not shown) and were indistinguishable from their mVps34^Ctrl littermates at birth. No apparent phenotype was seen by 2 weeks of age. However, mVps34^pdKO mice exhibited growth retardation and were significantly smaller by 6 weeks of age (Figure 2A), with a lower body weight (Figure 2B); however, there was no difference in the ratio of kidney to body weight (Figure 2C). The BUN levels of mVps34^pdKO mice were statistically higher by 3 weeks of age, and all mVps34^pdKO mice developed renal failure (Figure 2D) and died before 9 weeks of age. Furthermore, mVps34^pdKO mice developed proteinuria by 3 weeks of age (Figure 2E). SDS-PAGE assays revealed that albumin is the major protein species in the urine, although other plasma proteins also contributed to proteinuria (Figure 2F).Open in a separate windowFigure 2.Deletion of mVps34 in podocytes causing significant phenotypes after 2 weeks of age. (A–C) mVps34^pdKO mice exhibited growth retardation and were significantly smaller by 6 weeks of age (A), compared with same-sex mVps34^Ctrl littermates; body weight was significantly decreased (B) but ratio of kidney to body weight was similar (C). NS, not significant. (D) By 3 weeks of age, mVps34^pdKO mice developed statistically higher BUN levels (knockout, 67.29±11.83 mg/dl; control, 31.80±3.91 mg/dl), which continued to increase further; values are means ± SEM (n=4 for each time point group; *P<0.05 and **P<0.0001 versus mVps34^Ctrl groups at 3–7 weeks of age; no significant difference was seen among the different age groups between 3-week-old and 7-week-old mVps34^Ctrl mice). (E and F) Deletion of mVps34 specifically in podocytes caused massive proteinuria by 3 weeks of age. (E) Urinary protein-to-creatinine ratios were significantly increased by 3 weeks of age in mVps34^pdKO mice (values are means ± SEM [n=4 for each group; **P<0.0001 versus 1- or 2-weeks-of-age groups]). (F) SDS-PAGE assay (with Coomassie blue staining) confirmed that albumin is the major protein species in the urine, although other plasma proteins also contributed to the proteinuria. One microliter of urine from an individual 6-week-old mouse in each genotype group was loaded into each lane of the SDS-PAGE.mVps34^pdKO mice showed normal glomeruli at 9 days of age, similar to mVps34^Ctrl littermates (Figure 3A), suggesting that mVps34 is not essential for early podocyte development. However, mVps34^pdKO mice developed severe kidney lesions by 5–6 weeks of age. Their glomerular podocytes showed striking vacuolization (Figure 3, B–D), with focal segmental (Figure 3, E and F) and global (Figure 3G) glomerulosclerosis. mVps34^pdKO mice also had renal tubular dilation, numerous proteinaceous casts, and mild to moderate interstitial inflammation and fibrosis (Figure 3, B and C), which were confirmed by Masson trichrome staining (Figure 4A). Immunohistochemistry revealed massive fibronectin deposition along the Bowman capsule but moderate increases in the glomeruli and some periglomerular areas (Figure 4B).Open in a separate windowFigure 3.Podocyte-specific loss of mVps34 causing severe glomerulosclerosis. mVps34^pdKO mice had normal renal histologic findings at 4 days of age (data not shown) and 9 days of age (A) but developed severe kidney lesions by 6 weeks of age, including striking glomerular podocyte vacuolization (B–D), focal segmental (E and F) and global (G) glomerulosclerosis, renal tubular dilation, proteinaceous casts, and mild to moderate interstitial inflammation and fibrosis (B and C), as indicated by hematoxylin and eosin (H-E) (A, B, and D) and periodic acid-Schiff (PAS) staining (C, E, F, and G). Original magnification, ×100 in A, B, and Upper panel C; ×200 in lower panel C; ×400 in D–G.Open in a separate windowFigure 4.Confirmation of renal fibrosis using Masson trichrome staining and immunohistochemical staining for fibronectin. (A) Masson trichrome staining highlighted tubulointerstitial fibrosis and glomerulosclerosis in mVps34^pdKO mice at 6 weeks of age. (B) Immunohistochemical staining with an antibody specific for fibronectin revealed massive fibronectin deposition along the Bowman capsule, with a moderate increase in the glomeruli and some periglomerular areas of mVps34^pdKO mice, compared with mVps34^Ctrl mice.Electron microscopy revealed complete podocyte foot process effacement, with irregular and thickened glomerular basement membrane (Figure 5, A and B). The podocytes of mVps34^pdKO mice contained numerous enlarged vacuoles (Figure 5, D and E) and increased number of double-membraned cytoplasmic vesicles characteristic of autophagosomes, which, however, had irregular inner membranes (Figure 5, C and F). Accumulation of these aberrant autophagosomes in mVps34^pdKO mice is consistent with the fact that the enzymatic product of mVps34, phosphatidylinositol-3-phosphate, is enriched on the inner membranes of autophagosomes.^18–20Open in a separate windowFigure 5.Micrographs of electron microscopy from mVps34^Ctrl and mVps34^pdKO glomeruli at 6 weeks of age. (A and B) A representative micrograph revealing complete podocyte foot process effacement (black arrowheads), with irregular and thickened glomerular basement membrane in mVps34^pdKO mice compared with mVps34^Ctrl littermates (black arrows). (D and E) Lower-magnification image reveals numerous enlarged vacuoles (indicated by black asterisks) in the podocytes of mVps34^pdKO mice, but not in the podocytes of mVps34^Ctrl littermates. (C and F) Higher-magnification image shows the accumulation of aberrant double-membraned autophagosomes with irregular inner membranes in the podocytes of mVps34^pdKO mice (small arrows). Magnifications are indicated by the scale bars in the corresponding micrographs.In an attempt to determine the origin of these enlarged vacuoles (Figure 5E), we determined that the lysosome-associated membrane proteins 1 and 2 (LAMP1 and LAMP2) were both markedly elevated (Figure 6A). Immunofluorescence staining revealed that the enlarged vacuoles were positive for both LAMP1 (Figure 6B) and LAMP2 (Figure 6C). In contrast, podocin levels were decreased markedly (Figure 6, B and C). Thus, the enlarged vacuoles may have originated from lysosomes. During autophagy, the cytosolic form of microtubule-associated protein 1 light chain 3 (LC3)-I forms LC3-phosphatidylethanolamine conjugate (LC3-II), which is recruited to autophagosomal membranes and plays a critical role in autophagy and thus has become a reliable marker for autophagosomes.^21,22 We found LC3-II and even the cytosolic LC3-I were both markedly increased in the mVps34^pdKO mice (Figure 6A).Open in a separate windowFigure 6.Increased protein levels of LAMP1 and LAMP2 localized to the membranes of the enlarged vacuoles in the podocytes of mVps34^pdKO mice. (A) Immunoblotting of isolated glomerular homogenates revealed markedly increased protein levels of the lysosome markers, LAMP1 and LAMP2, as well as the autophagosome marker, LC3-II and the cytosolic LC3-I. (B and C) Immunofluorescence staining localized the increased protein levels of LAMP1 (B) and LAMP2 (C) to the enlarged vacuoles in the podocytes of mVps34^pdKO mice; in contrast, podocin levels were markedly diminished compared with mVps34^Ctrl mice (B and C). DAPI, 4′,6-diamidino-2-phenylindole. Original magnification, ×400 in B and C.Recent studies indicate that inactivation of the mammalian target of rapamycin (mTOR) by either double knocking out Raptor (an essential protein for mTOR complex 1 activity) and Rictor (key for mTORC2 activity)²³ or directly deleting the mTOR gene in podocytes²⁴ caused a phenotype similar to what we have seen in our mVps34^pdKO mice, including podocyte vacuolization, massive proteinuria, progressive glomerulosclerosis, and renal failure within 6 weeks after birth.^23,24 Podocyte-specific mTOR deletion also caused an accumulation of autophagosomes and autolysosomes by disrupting autophagic flux in podocytes.²⁴ Because mVps34 has been shown to lie upstream of the mTOR pathway,^25,26 it is possible that mVps34 deletion in podocytes may result in the similar phenotype by inactivating the mTOR pathway. Surprisingly, when we examined the phosphorylation levels of the ribosomal protein S6 (rpS6), a downstream target of mTOR commonly used as an in vivo readout for mTOR activities, we found that the mTOR pathway was activated, rather than inactivated, along with increased levels of both LAMP1 and LAMP2 (Figure 7B). Proteinuria was not detected until 17 days of age (data not shown), but significant increases in both LAMP1 and LAMP2 had already occurred by 15 days of age, without increases in rpS6 phosphorylation (Figure 7A). Thus, mTOR activation, rather than inactivation, seen in 6-week-old mVps34^pdKO mice (Figure 7B) might be secondary to proteinuria or dysregulated autophagy. However, our data in this study cannot rule out the possibility that mTOR inactivation by mVps34 knockout may be an early signaling event that leads to a phenotype similar to that seen in the podocyte-specific mTOR knockout mice.^23,24 Future studies are required to clarify the precise interplay between mVps34 and mTOR.Open in a separate windowFigure 7.Deletion of mVps34 in podocytes increases the protein levels of LAMP1, LAMP2, and LC3 but does not activate the mTOR pathway before the onset of proteinuria. (A) Immunohistochemistry showed no apparent alterations in the phosphorylation levels of rpS6 in 15-day-old mVps34^pdKO mice compared with their mVps34^Ctrl littermates; however, these 15-day-old mVps34^pdKO mice had increased levels of the lysosome-associated membrane proteins, LAMP1 and LAMP2. (B) The phosphorylation levels of rpS6 were increased selectively in the vacuolated podocytes of mVps34^pdKO mice at 6 weeks of age when massive proteinuria had already developed (as shown in Figure 2, E and F) and glomerulosclerosis occurred (as shown in Figure 3, B–G). (C and D) Immunofluorescence staining revealed markedly elevated levels of both LC3-II/I and LAMP-1 in mVps34^pdKO mice by 15 days of age before the onset of proteinuria and remained elevated at 6 weeks of age when detrimental podocyte vacuolization was observed and severe glomerulosclerosis occurred. DAPI, 4′,6-diamidino-2-phenylindole. Original magnification, ×400 in A–D.Additional experiments with immunofluorescence staining revealed markedly increased levels of both LC3 and LAMP1 in mVps34^pdKO mice at 15 days of age (Figure 7C) as well as 6 weeks of age (Figure 7D), consistent with the immunoblotting data shown in Figure 6A. As indicated in Figure 7C, some of the increased LC3-positive granules or punctate dots were co-localized with the increased LAMP1 (suggesting increases in autolysosomes), whereas others were not co-localized with LAMP1, indicating increases in autophagosomes. Double immunofluorescence staining confirmed that the increases in both LC3 (Figure 8A) and LAMP1 (Figure 8B) occurred only in nephrin-positive glomerular cells (the podocytes). Indeed, electron microscopy further confirmed that formation of numerous vacuoles (Figure 9, E and F) and an increase of autophagosomes (Figure 9, G and H) occurred only in the podocytes, with occasional focal foot process effacement (Figure 9, E and F), compared with mVps34^Ctrl littermates (Figure 9, A–D). Thus, both marked elevations of LC3 and LAMP1/2 and the earliest ultrastructural alterations (podocyte vacuolization, focal foot process effacement, and increased autophagosome formation) had occurred before the onset of proteinuria.Open in a separate windowFigure 8.Podocyte-specific mVps34 knockout increases the protein levels of both LC3 and LAMP1, but not that of nephrin, by 15 days of age before the onset of proteinuria. (A) Triple immunofluorescence staining with the indicated antibodies verified that compared with their mVps34^Ctrllittermates, mVps34^pdKO mice showed markedly increased LC3 protein levels specifically in the nephrin-positive glomerular cells, which are podocytes. (B) Triple immunofluorescence staining also confirmed that increased LAMP1 protein levels were confined to the podocytes. In contrast, the nephrin protein level of Vps34^pdKO mice was not increased, compared with that of mVps34^Ctrl mice. DAPI, 49,6-diamidino-2-phenylindole. Original magnification, ×400.Open in a separate windowFigure 9.Lack of functional mVps34 in the glomerular podocyte disrupts the dynamics of intracellular vesicle forming and processing. Podocyte-specific mVps34 knockout resulted in striking ultrastructural podocyte vacuolization (E and F) and increased autophagosome formation (G–H) by 15 days of age before the onset of proteinuria, compared with mVps34^Ctrl littermates (A–D). Focal foot process effacement was seen occasionally as indicated by a black arrowhead (F). Shown are representative micrographs of electron microscopy from three mice in each group with similar results. A scale bar is shown in each corresponding micrograph to indicate the magnification, with black asterisks representatively indicating enlarged vacuoles (F and H) and small white arrows representatively pointing to increased number of autophagosomes (F–H) in the podocytes of mVps34^pdKO mice.Of note, previous studies suggested that mVps34 plays an indispensable role at the initiation step of autophagosome formation,^13–15,27 which would not be consistent with the increases in autophagosome formation (Figure 9) and a striking accumulation of autophagic vesicles in the enlarged autolysosomes of the mVps34-deleted podocytes (Figure 5, C and F). Thus, our results suggest the existence of an mVps34-independent mechanism that can initiate the formation of autophagosomes in the podocyte. Follow-up studies are required to define the precise roles of mVps34 and its interactions with many other autophagy-related genes (Atg). It is also important to determine whether endocytic pathway has also been disrupted and contributed to the enlarged vacuoles in the mVps34^pdKO podocytes.In summary, this study represents the first report of podocyte-specific mVps34 knockout. The major findings include markedly increased LC3-II/I and LAMP1/2 but decreased podocin levels, causing a lethal accumulation of aberrant autophagosomes and enlarged autolysosomes, resulting in striking podocyte vacuolization, foot process effacement, glomerulosclerosis, and interstitial fibrosis, consequently leading to massive proteinuria and renal failure. Thus, there is no compensation by related genes in the mVps34^pdKO mice, which all died before 9 weeks of age. In contrast, podocyte-specific Atg5 knockout mouse caused age-dependent late-onset glomerulosclerosis.^6–9mVps34^pdKO mice reported herein exhibited a much more pronounced and severe phenotype than did the Atg5 knockout mice, even when 6-week-old mVps34^pdKO mice were compared with the 24-month-old Atg5^pdKO mice.⁹ In future studies, it will be important to determine whether mVps34 inactivation contributes to the pathogenesis of proteinuria and glomerulosclerosis in some cases of glomerulosclerosis, such as FSGS and diabetic nephropathy.