Nonmuscle Myosin II Regulates the Morphogenesis of Metanephric Mesenchyme–Derived Immature Nephrons

Abstract

The kidney develops from reciprocal interactions between the metanephric mesenchyme and ureteric bud. The mesenchyme transforms into epithelia and forms complicated nephron structures, whereas the ureteric bud extends its pre-existing epithelial ducts. Although the roles are well established for extracellular stimuli, such as Wnt and Notch, it is unclear how the intracellular cytoskeleton regulates these morphogenetic processes. Myh9 and Myh10 encode nonmuscle myosin II heavy chains, and Myh9 mutations in humans are implicated in congenital kidney diseases and focal segmental glomerulosclerosis in adults. Here, we analyzed the roles of Myh9 and Myh10 in the developing kidney. Ureteric bud-specific depletion of Myh9 resulted in no apparent phenotypes, whereas mesenchyme-specific Myh9 deletion caused proximal tubule dilations and renal failure. Mesenchyme-specific Myh9/Myh10 mutant mice died shortly after birth and showed a severe defect in nephron formation. The nascent mutant nephrons failed to form a continuous lumen, which likely resulted from impaired apical constriction of the elongating tubules. In addition, nephron progenitors lacking Myh9/Myh10 or the possible interactor Kif26b were less condensed at midgestation and reduced at birth. Taken together, nonmuscle myosin II regulates the morphogenesis of immature nephrons derived from the metanephric mesenchyme and the maintenance of nephron progenitors. Our data also suggest that Myh9 deletion in mice results in failure to maintain renal tubules but not in glomerulosclerosis.

The kidney develops from reciprocal induction between two precursor tissues: the mesenchyme and ureteric bud. At embryonic day 11.5 (E11.5) in the mouse, the ureteric bud invades the metanephric mesenchyme, and Wnt9b secreted from the bud induces Six2-positive nephron progenitors in the mesenchyme to transform into epithelia.^1,2 The differentiating cells lose Six2 expression and begin to express Wnt4, which further enhances differentiation.³ After the immature epithelia emerge, Notch2 signaling specifies the nascent nephrons into the distal and proximal regions.^4,5 Thus, progenitors sequentially transit to pretubular aggregates, renal vesicles, and then, C- and S-shaped bodies, which eventually develop into nephron epithelia. The proximal region of the S-shaped bodies becomes glomerular podocytes and proximal renal tubules, whereas the distal region becomes the distal renal tubules. Although the roles are well established for Wnt and Notch, it is still largely unknown how the cytoskeleton regulates cellular polarity and the luminal space, which leads to drastic morphologic changes in nascent nephrons. Simultaneously, the ureteric bud continues to branch and form collecting ducts and ureters. Because this process involves the extension of pre-existing epithelial tube structures, a different mechanism may operate from that in the mesenchyme that requires mesenchymal-to-epithelial transition.

Nonmuscle myosin II is a motor protein family that modulates cell adhesion and polarity.^6,7 These proteins work in association with actin filaments, and actomyosin underpins cadherins at the adherens junctions through β- and α-catenins. The establishment and proper maintenance of adherens junctions lead to the formation of tight junctions and separation of apical and basal domains of the epithelia, thus establishing apical-basal polarity. Furthermore, actomyosin-mediated constriction of the apical domain is a prerequisite for tubule morphogenesis. In vertebrates, there are three isoforms with the nonmuscle myosin heavy chains (NMHCs) (IIA, IIB, and IIC), which are encoded by three different genes (Myh9, Myh10, and Myh14, respectively).⁶ Myh9 deficiency results in the loss of cell adhesion in the visceral endoderm of peri-implantation embryos, whereas Myh10-null mice show defects in the heart and brain during midgestation.^8,9 Mice lacking Myh14 alone show no apparent phenotypes,¹⁰ suggesting predominant roles of Myh9 and Myh10 over Myh14 in embryonic development. In humans, a number of infantile disorders are caused by mutations in the Myh9 gene.^11–13 Such disorders include Fechtner and Epstein syndromes, which are platelet abnormalities accompanied by kidney symptoms. Myh9 is also associated with adult kidney diseases on the basis of genome-wide association studies.^14,15 African Americans are reportedly more predisposed to nephropathy, including focal segmental glomerulosclerosis (FSGS), because of certain race-related single-nucleotide polymorphisms in the Myh9 gene. However, other reports argue that ApoL1, a neighbor gene of Myh9, is more directly linked to these diseases.^16,17 Despite these findings, it is still unclear how Myh9 mutations affect the kidneys. Even the simple question of whether Myh9 is important in mesenchyme- or ureteric bud–derived lineages has not been addressed.

We have previously shown that a newly recognized kinesin, Kif26b, is essential for kidney development and identified NMHCs IIA and IIB as proteins that interact with Kif26b.¹⁸ Therefore, in this study, we examined the in vivo roles of nonmuscle myosin II in kidney development and its relationship to human diseases. We deleted Myh9 and Myh10 specifically in the metanephric mesenchyme and found that these genes are essential for this lineage.

Results

Expression Patterns of Myosin Isoforms during Kidney Development

At E14.5, NMHC IIA (encoded by Myh9) was expressed in the nascent epithelialized nephrons and ureteric bud and less abundantly in the cortical metanephric mesenchyme (Figure 1A). NMHC IIA was still widely distributed in both the cortex and medulla in the neonatal kidney (postnatal day 0 [P0]) but expressed weakly at 8 weeks after birth. NMHC IIB (encoded by Myh10) was expressed abundantly in cortical and medullary interstitial cells at E14.5 and P0 (Figure 1B), whereas it was less abundant in the metanephric mesenchyme and ureteric bud and undetectable in the adult. NMHC IIC (encoded by Myh14) was expressed in the glomerular podocytes at E14.5 and the cortex at P0 but weakly detectable in the adult (Figure 1C). Therefore, Myh9 and Myh10 encode the major myosins during embryonic kidney development.

Figure 1.

Expression patterns of myosin isoforms during kidney development. (A) Expression of NMHC IIA (Myh9) at E14.5, P0 (newborn), and 8 weeks after birth. Myh9 is expressed in the embryonic and neonatal kidneys but only weakly in the adult. (B) Expression of NMHC IIB (Myh10). Myh10 is expressed in the interstitial cells of embryonic and neonatal kidneys but absent in the adult. (C) Expression of NMHC IIC (Myh14). Arrows point to glomerular podocytes. Signals in the ureteric bud stalk (asterisks) may result from background staining of the antibody, which is not consistent with the data obtained by in situ hybridization. Scale bars, 100 μm in E14.5 and P0; 1 mm in 8 w.

Metanephric Mesenchyme-Specific Myh9 Deletion Causes Proximal Tubule Dilation, Leading to Renal Failure in Adults

We first crossed the floxed allele of Myh9 with the Hoxb7Cre mouse expressing Cre recombinase specifically in the ureteric bud.¹⁹ In the Hoxb7Cre;Myh9^flox/flox mutant, Myh9 disappeared only in the ureteric bud–derived epithelia, indicating successful deletion of Myh9 (Figure 2A). However, the mutant mice showed no obvious defects in their kidney morphology. Immunostaining of cytokeratin, a ureteric bud marker, revealed similar patterns of expression in both the control and mutant. Levels of blood urea nitrogen (BUN), creatinine, and albumin in the sera from 8-month-old mice were not significantly different in the two groups (Table 1). Myh10 expression in the ureteric bud–derived epithelia was slightly upregulated in the newborn Myh9 mutants, whereas no alteration was found in Myh14 expression (Figure 2B).

Figure 2.

Myh9 is dispensable for ureteric bud development. (A) Kidneys of mice lacking Myh9 in the ureteric bud lineage at birth. (Left panels) Control mice. (Row 1) Myh9 immunostaining. (Row 2) magnified images. Note that the Myh9 expression is absent in mutant ureteric bud–derived epithelia (asterisks). (Row 3) Hematoxylin and eosin staining. (Row 4) Immunostaining of cytokeratin (CK), a ureteric bud marker. Dark purple indicates a positive signal. Sections are counterstained with nuclear fast red. Scale bars, 100 μm. (B) Expression of Myh10 and Myh14 in the ureteric bud-specific Myh9 mutants at birth. Scale bars, 100 μm.

Table 1.

Renal functions in adult mutant mice

Parameter	Control	Mutant	P Value
	Myh9 flox/flox (n=3)	Six2Cre;Myh9 flox/flox (n=4)
Age (mo)	10	10
BUN (mg/dl)	24.8 (5.3)	102.2 (2.1)	<0.01
Creatinine (mg/dl)	0.17 (0.01)	0.48 (0.09)	<0.01
Albumin (g/dl)	3.2 (0.8)	3.4 (0.1)	0.68
	Myh10 flox/+ (n=3)	Six2Cre;Myh10 flox/flox (n=4)
Age (mo)	8	8
BUN (mg/dl)	28.4 (2.1)	36.2 (9.1)	0.20
Creatinine (mg/dl)	0.11 (0.01)	0.14 (0.02)	0.03
Albumin (g/dl)	3.2 (0.2)	3.3 (0.3)	0.60
	Myh9 flox/flox (n=3)	Hoxb7Cre;Myh9 flox/flox (n=3)
Age (mo)	8	8
BUN (mg/dl)	33.9 (1.1)	35.5 (4.2)	0.59
Creatinine (mg/dl)	0.13 (0.01)	0.10 (0.01)	0.05
Albumin (g/dl)	3.4 (0.3)	3.7 (0.1)	0.19

Mean (standard deviation).

We next deleted Myh9 in the metanephric mesenchyme using the Six2Cre mouse.¹ Six2Cre;Myh9^flox/flox mutant mice were viable, but the kidneys of 8-week-old adults showed signs of cysts and tubular dilations (Figure 3A, upper panels). There was also enlargement of Bowman’s space of glomeruli (Figure 3A, lower panels). The mutant mice survived for at least 10 months (our observation period), but there was a significant elevation of BUN and creatinine (Table 1). We also observed histologically severe dilations of the medulla and renal tubules (Figure 3B, upper panels). Although Bowman’s space and the capillary loops were also dilated, there was no sign of glomerular sclerosis (Figure 3B, lower panels). The serum albumin level was not reduced in the mutants (Table 1), and we did not detect any proteinuria. Therefore, Myh9 deletion in the metanephric mesenchyme leads to kidney failure, which likely results from tubular dilations and not glomerulosclerosis.

Figure 3.

Metanephric mesenchyme–specific Myh9 deletion causes proximal tubule dilation, leading to renal failure in adults. (A) Kidneys of mice lacking Myh9 in the metanephric mesenchyme at 8 weeks after birth. (Left panels) Control mice. (Upper panels) Hematoxylin and eosin staining. (Lower panels) Magnified images of glomeruli. Podocytes are immunostained with an anti-Wt1 antibody. Asterisks show the Bowman’s spaces. Scale bars, 1 mm in upper panels; 100 μm in lower panels. (B) Hematoxylin and eosin staining of control and mutant kidneys at 10 months after birth. (Lower panels) Magnified images of glomeruli. Asterisks show the Bowman’s spaces. Scale bars, 1 mm in upper panels; 100 μm in lower panels. (C) Kidneys of control and mutant mice at birth. Hematoxylin and eosin staining and immunostaining with the indicated lineage markers. Scale bars, 100 μm. (D) Intact apicobasal polarity of proximal tubule epithelia in (left panel) the control and (right panel) the mutant shown by Aqp1 staining (green). Scale bars, 50 μm. (E) Expression of Myh10 and Myh14 in the metanephric mesenchyme-specific Myh9 mutants at birth. Scale bars, 100 μm. CK, cytokeratin; LTL, lotus tetragonolobus lectin; THP, Tamm–Horsfall protein.

Newborn Myh9 mutant kidneys already exhibited tubular dilations (n=7), although the glomeruli (Wt1 positive) were relatively intact (Figure 3C). Lotus tetragonolobus lectin staining and immunostaining of aquaporin 1 (Aqp1) showed that the dilated regions were segments of the proximal tubules and descending limb of Henle. In contrast, the thick ascending limb of Henle (positive for Tamm–Horsfall protein) and distal tubules (positive for solute carrier family 12, member 3: Slc12a3) showed normal diameters. The ureteric bud epithelia and collecting ducts marked by cytokeratin were not dilated. Thus, the proximal tubules and descending limb of Henle are initially dilated in the absence of Myh9. A magnified view showed that Aqp1 was confined to the apical domain in Myh9 mutants, suggesting that the apicobasal polarity is maintained in the absence of Myh9 (Figure 3D). Myh10 but not Myh14 was slightly upregulated in the mesenchyme-derived tissues in the newborn Myh9 mutants (Figure 3E). Considering the analysis of Myh9/Myh10 mutants described below, we speculate that dilations of proximal tubules and the descending limb of Henle may result from reduced apical tension, although no measurement methods are available.

Mesenchyme-Specific Myh9/Myh10 Mutants Die Shortly after Birth and Lack Most Nephron Components

Metanephric mesenchyme–specific deletion of Myh10 resulted in similar but much milder tubular dilations (Supplemental Figure 1A) with normal kidney functions (Table 1), which was consistent with its weak expression in the mesenchyme (Figure 1B, Supplemental Figure 1B). Thus, we generated mice lacking both Myh9 and Myh10 in the metanephric mesenchyme. The double-mutant mice (Six2Cre;Myh9^flox/flox Myh10^flox/flox) were born but died shortly after birth, and their kidneys were smaller than those of the control (n=8) (Figure 4A). Hematoxylin and eosin staining of kidney sections showed reduced formation of the cortical nephron layer (n=3) (Figure 4B). Immunostaining of Six2 and Cited1 revealed a reduction of the cap mesenchyme that contains nephron progenitors (Figure 4B). Staining of various markers for nephron segments confirmed a significant reduction of the metanephric mesenchyme–derived nephron components, whereas the ureteric bud-derived structure was relatively intact (Figure 4C). Therefore, Myh9 and Myh10 have redundant functions in the metanephric mesenchyme, and loss of both genes leads to severe nephron loss.

Figure 4.

Mesenchyme-specific Myh9/Myh10 mutants die shortly after birth and lack most nephron components. (A) Kidneys of mice lacking Myh9 and Myh10 in the metanephric mesenchyme at birth. (Left panel) Control mice. b, Urinary bladder; k, kidney; t, testis. (B) Reduction of nephron progenitors in newborn mutant kidneys. (Top panels) Hematoxylin and eosin staining. (Middle and bottom panels) Immunostaining of Six2 and Cited1 (progenitor markers). Scale bars, 100 μm. (C) Reduction of mesenchyme-derived nephron components in mutant kidneys at birth. Immunostaining of the indicated lineage markers. Scale bars, 100 μm. CK, cytokeratin; LTL, lotus tetragonolobus lectin; THP, Tamm–Horsfall protein.

Myh9/Myh10 Mutant Nascent Nephrons Are Irregularly Shaped and Exhibit Apoptosis

At E14.5, the characteristic nascent nephrons, ranging from pretubular aggregates, renal vesicles, and C-shaped bodies to S-shaped bodies, were observed in control mice (Figure 5A). However, the formation of these structures was significantly reduced in the double mutants, and we detected irregularly formed nephron structures (n=6). Double staining of cleaved caspase-3 and Sall1, which marks mesenchyme derivatives,²⁰ indicated increased apoptosis in the irregularly shaped nascent nephrons in the mutants (Figure 5B). Therefore, the abnormal morphogenesis of nascent nephrons accompanied by apoptosis may lead to the nephron loss in the absence of nonmuscle myosin II. Indeed, Myh9 was expressed in these nascent nephrons as well as the ureteric bud and less abundantly in the metanephric mesenchyme (Figure 5C). Myh9 expression in these structures, except the ureteric bud, was absent in Six2Cre;Myh9^flox/flox;Myh10^flox/flox mice, indicating successful deletion of Myh9 in the mesenchyme-derived lineages. However, we observed some residual expression in mutant nascent nephrons, which is consistent with the mosaic recombinase activity of Six2Cre mice. Although Myh10 expression in the nascent nephrons was so weak that no significant differences in staining were observed in the control and mutant (Supplemental Figure 1B), the exacerbation of the phenotypes underscores the importance of Myh10 in nascent nephrons.

Figure 5.

Myh9/Myh10 mutant nascent nephrons are irregularly shaped and exhibit apoptosis. (A) Hematoxylin and eosin staining of the kidneys of mice lacking Myh9 and Myh10 in the metanephric mesenchyme at E14.5. (Left panels) Control mice. (B) Apoptosis in mutant nascent nephrons at E14.5. Immunostaining of cleaved caspase-3 (red) and Sall1 (green). (Lower panels) Higher magnification of the indicated regions. (C) Immunostaining of Myh9 in E14.5 kidneys. Myh9 expression is absent in nephron progenitor regions and nascent nephrons but present in the ureteric bud. (D) Immunostaining of Wnt4, Lef1, and Jag1 at E14.5. Arrowheads show renal vesicles, and arrows show S-shaped bodies. Scale bars, 50 μm. u, Ureteric bud. *Progenitor region.

The metanephric mesenchyme undergoes mesenchyme-to-epithelial transition induced by Wnt9b-mediated signaling. Wnt4 is expressed in pretubular aggregates to promote additional epithelization, and Lef1 is used as an indicator of canonical Wnt activity. There was no reduction in the expression of Wnt4 or Lef1 in double-mutant kidneys (Figure 5D), suggesting that the initiation of mesenchyme-to-epithelial transition is not affected. Subsequently, Notch signaling specifies the proximodistal patterning of the nephrons. Jagged1 (Jag1) was expressed in the proximal to middle domains of immature nephrons in control mice (Figure 5D). Jag1 was also expressed in the double-mutant nephrons, suggesting that proximodistal specification of the nephron also occurs in the absence of nonmuscle myosin II.

Nonmuscle Myosin II Is Required for Proper Lumen Elongation of Nascent Nephrons

Morphogenetic changes require proper formation of intercellular adherens junctions accompanied by the establishment of apical and basolateral domains of the epithelia. Atypical protein kinase C (aPKC) was confined to the apical domains that face the lumen of the control renal vesicle, whereas neural cell adhesion molecule (NCAM) was expressed in the basolateral domains (Figure 6A) as reported previously.²¹ In addition, the apical domains were narrower than the basal domains, and the cells formed rosette-like structures. The apicobasal distributions of aPKC and NCAM were maintained to some extent in the double mutants, but not all of the cells were aligned to form the rosette or face the lumen surrounded by the aPKC-positive apical domain (Figure 6A, upper panels). At the S-shaped body stage, a single continuous lumen was delineated by the aPKC domain in the control, whereas the mutant lumina were discontinuous and branched irregularly (Figure 6A, lower panels). F-actin (shown by phalloidin staining) and R-cadherin should be concentrated to apical adheres junctions and, indeed, were expressed at the apical domains in the control (Figure 6B and C, left panel). Although F-actin and R-cadherin were also expressed at the apical domains in double mutants, the lumina lined with F-actin were again branched abnormally. Myh9 was expressed evenly in the cytoplasm of the control, but its deletion in the mutant was mosaic, reflecting the recombinase activity of the Six2Cre allele (Figures 5C and 6C). It is noteworthy that R-cadherin expression was detected even in the region with clustered Myh9-negative cells (Figure 6C). Therefore, myosin II is not required for the initial establishment of apicobasal polarity but the recruitment and integration of new apical domains into the elongating lumen of nascent nephrons. We also examined the process of lumen formation using the Madin–Darby canine kidney (MDCK) cell line embedded in Matrigel.²² Similar to embryonic kidneys, aPKC was detected in the apical domain (Figure 6D). Although Myh9 and Myh10 were ubiquitously expressed in the cytoplasm, they were enriched close to the apical domain (Figure 6D). In addition, F-actin had accumulated significantly in the apical domain (Figure 6E), indicating the existence of abundant actomyosin complexes in this domain facing the lumen. In the presence of blebbistatin, a specific myosin II inhibitor, the lumen lined with actin-positive apical domains was irregularly shaped and fragmented as observed in the Myh9/10 mutant kidneys (Figure 6E). The milder phenotype in vivo may at least partly result from incomplete deletion of Myh9/10. These data suggest that mechanical tension evoked by actomyosin in the apical domain (i.e., apical constriction) contributes to correct tubular morphogenesis of the nascent nephrons.

Figure 6.

Nonmuscle myosin II is required for proper lumen elongation of nascent nephrons. (A, upper panels) Renal vesicles and (A, lower panels) S-shaped bodies stained for aPKC (red) in mice with mesenchyme-specific deletion of Myh9 and Myh10 at E14.5. NCAM is stained green to visualize nascent nephrons. (B) Phalloidin (red) and NCAM (green) staining of the nascent nephrons. (C) Immunostaining of R-cadherin (red) and Myh9 (green) of the nascent nephrons. Scale bars, 20 μm. u, Ureteric bud. *Myh9-positive region; **Myh9-negative region. (D) Immunostaining of the MDCK cysts. (Top panel) aPKC (red) and integrin-β1 (green). (Middle panel) Myh9 (red). (Bottom panel) Myh10 (red). Scale bars, 20 μm. (E) Phalloidin (red) and β-catenin (green) staining of MDCK cysts in (upper panel) the absence or (lower panel) presence of (−)-blebbistatin. Scale bars, 20 μm.

Nonmuscle Myosin II Is Required for the Maintenance of Nephron Progenitors

We previously identified NMHCs IIA and IIB as interacting partners of Kif26b.¹⁸ Because Kif26b is exclusively expressed in the cap mesenchyme, myosin should also have a role in this cell population, including nephron progenitors. Indeed, Myh9 was expressed in the progenitor region, although at a lower level than that in epithelialized nascent nephrons (Figure 5C), and progenitors were significantly reduced in newborn Myh9/Myh10 mutants (Figure 4B). At E14.5, Six2-positive progenitors accumulated around the ureteric buds in the control, whereas Myh9/Myh10-deficient progenitors were scattered slightly (Figure 7A). Indeed, N-cadherin, which was clearly localized in the lateral domains of the nephron progenitors, was obscure in the mutant (Figure 7A), suggesting impaired cell-to-cell adhesion of the progenitors in the absence of myosin activity. Cited1 expression in nephron progenitors was lost in the mutants, despite the existence of the cells. Because there was a reduction of integrin-α8 in the mesenchyme (Figure 7A), myosin may also be involved in the integrin-mediated adhesion of mesenchymal nephron progenitors to the ureteric bud, which expresses nephronectin, the ligand for integrin-α8.²³ However, survival or proliferation defects were not detected in the mutant nephron progenitors (Figure 5B, Supplemental Figure 2A). Therefore, a reduction of lateral and basal adhesion of nephron progenitors might impair their self-renewal followed by marked apoptosis of the nascent nephrons, leading to the significant nephron loss at birth.

Figure 7.

Nonmuscle myosin II is required for the maintenance of nephron progenitors. (A) Nephron progenitors of mice with mesenchyme-specific deletion of Myh9 and Myh10 at E14.5. Immunostaining of Six2, Cited1, integrin-α8 (Itga8), and N-cadherin. Scale bars, 50 μm. Images in row 4 show N-cadherin (red) and Sall1 (green) staining. Scale bars, 20 μm. (B) Nephron progenitors of mice with mesenchyme-specific deletion of Kif26b at E14.5. (C) Nephron progenitors of mice with mesenchyme-specific deletion of Kif26b at birth. Scale bars, 50 μm. u, Ureteric bud.

We next generated a floxed allele of Kif26b (Supplemental Figure 3) and crossed these mice with Six2Cre. In this mouse strain, we could bypass the severe phenotype observed in the conventional Kif26b deletion, namely failure of ureteric bud attraction that leads to kidney agenesis even at E14.5. In Six2Cre;Kif26b^flox/flox kidneys at E14.5, we found scattered Six2-positive progenitors, an absence of Cited1, and reductions of N-cadherin and integrin-α8 (Figure 7B), all of which were similar to the phenotypes in the conventional Kif26b mutants¹⁸ and Six2Cre;Myh9^flox/flox;Myh10^flox/flox mice. As observed in the Myh9/Myh10 mutants, survival or proliferation defects were not detected in the Kif26-deficient nephron progenitors (Supplemental Figure 2B). Thus, Kif26b and myosin might function together to maintain nephron progenitors. However, Six2Cre;Kif26b^flox/flox mice were born with apparently normal-sized kidneys and survived to adulthood, although the numbers of nephron progenitors were reduced at birth (Figure 7C). The irregularly shaped nascent nephrons or dilated tubules, which were observed on myosin deletion, were never detected in the absence of Kif26b (Supplemental Figure 2B). Therefore, the morphogenetic roles of myosin played in nascent nephrons are independent of Kif26b.

Discussion

In this study, we investigated the role of nonmuscle myosin II in the kidney by tissue-specific deletion of myosin heavy chain genes. Myh9 and Myh10 have redundant functions in the metanephric mesenchyme. We have shown that nonmuscle myosin II is essential for nascent nephron morphogenesis, especially lumen formation. Force evoked by actomyosin in the apical domain (apical constriction) is likely to contribute to the correct tubular morphology and integration of new apical domains into the elongating lumen of nascent nephrons. Recently, it was shown that proamniotic cavity formation in peri-implantation egg cylinder–stage embryos is mediated by rosette formation and actomyosin-dependent apical constriction of epiblasts.²⁴ This process is called hollowing. Because we found rosette formation and myosin dependence in our study, the lumen in renal tubules is likely also formed by hollowing, which is strikingly similar to that in peri-implantation embryos. However, renal tubules are subjected to more complicated morphogenesis and elongation to form S-shaped bodies and finally, nephrons. Pronephric tubule elongation in Xenopus is driven by multicellular rosette–based convergent extension that is dependent on myosin.²⁵ It is possible that convergent extension mediated by myosin II activity may also be involved in elongation of mammalian renal tubules that are derived from the mesenchyme. Genetically labeled mouse kidneys combined with time-lapse analyses would be needed to precisely address this point.

In humans, Myh9 mutations are known to be the cause of infantile disorders that mainly affect platelets, which are accompanied by symptoms in the kidneys.^11–13 Myh9 is also associated with FSGS in adults,^14,15 although some controversy exists. Our findings on the basis of deletion of Myh9 in the Six2-positive metanephric mesenchyme indicate that Myh9 mutations are, indeed, directly involved in the progression of kidney disease. However, the main symptoms are dilations of proximal tubules and the descending limb of Henle. Although glomerular podocytes are derived from Six2-positive mesenchyme, the mutant glomeruli are relatively intact until the end stage of the disease, although tubule dilation may secondarily affect glomerular function and eventually, structure. This result is partly consistent with a previous study on podocyte-specific deletion of Myh9 on a C57BL/6 background.^26,27 These mutant mice are viable with normal kidneys under standard conditions but predisposed to drug-induced glomerulopathy. However, spontaneous glomerulosclerosis was reported on podocyte-specific Myh9 deletion on a mixed genetic background.²⁸ Our Six2Cre-mediated mutants, which were backcrossed to C57BL/6 several generations, may exhibit podocyte dysfunctions under stressed conditions, but the severe defects in the renal tubules would hinder our analysis. Nonetheless, our data clearly indicate the importance of Myh9 in tubule integrity, at least in mice, whereas the glomerular phenotypes are less apparent. In contrast, even heterozygous mice carrying point mutations in Myh9 that mimic human patients exhibit glomerulosclerosis instead of tubular dilations, which phenocopies Myh9-related autosomal dominant human hereditary diseases.²⁸ Therefore, some Myh9 mutations found in humans are unlikely to result in hypomorphic alleles but may exert a gain-of-function effect that preferentially manifests glomerulosclerosis, although species difference should be considered between mice and humans. Recently, we generated three-dimensional glomeruli and renal tubules in vitro from human induced pluripotent stem cells.²⁹ Nephrons formed by induced pluripotent stem cells derived from patients with Myh9 mutations will help address these points.

Our data also show that myosin and Kif26b play a role in the maintenance of nephron progenitors. Reductions of integrin-α8 and N-cadherin in both mutants suggest that adhesion to the ureteric bud niche and cell-to-cell adhesion between the progenitors (i.e., mesenchymal condensation) are required for the progenitor maintenance. However, progenitor depletion at E14.5 was not as severe as that observed in Sall1-deficient mice, which we reported recently,³⁰ and Kif26 mutants showed no kidney size reduction and survived to adulthood. It is likely that the progenitor pool is larger than that required and that its mild reduction alone, without any defects in nascent nephrons, does not result in a severe loss of nephrons. Therefore, the marked apoptosis accompanied by morphologic abnormalities in the Myh9/10-null nascent nephrons is likely to play a major role in the nephron loss, although the mild progenitor depletion may also contribute to the final phenotype to some extent.

Myh9 is not required for ureteric bud development, although Myh9 is expressed abundantly in this lineage. Ureteric bud development involves the extension of pre-existing tubes with lumens, which is different from the mesenchymal-to-epithelial transition that occurs in the metanephric mesenchyme. It is possible that ureteric bud elongation and branching require more fine-tuned reorganization of adherens junctions. Alternatively, other myosins could compensate for Myh9 deletion. Because Myh10 expression is slightly upregulated on ureteric bud–specific Myh9 deletion, it is necessary to generate mice lacking both Myh9 and Myh10 in this lineage.

In summary, we have revealed the essential roles of nonmuscle myosin II in nascent nephron development and the maintenance of renal tubules. Additional studies on cytoskeletal regulation of nephron development would provide important information on the complex three-dimensional structures of the kidney as well as human diseases.

Concise Methods

Generation of the Mutant Mice

Myh9^flox/flox and Myh10^flox/flox mice (available from MMRRC 32096 and 16981, respectively) are described elsewhere.^31,32 Bacterial artificial chromosome-mediated transgenic Six2GFPCre mice were provided by Andrew P. McMahon (University of Southern California), and the Hoxb7Cre mouse was obtained from the Jackson Laboratory.^1,19 These mouse strains were backcrossed to C57BL/6 at least three generations. The Kif26b-flox targeting vector was constructed by incorporating the 5′ and 3′ homology arms into a vector containing an loxP-loxP-frt-Neo-pA-frt-pA cassette and pMC1DTA. The loxP sequences were positioned so that exon 7 encoding the kinesin motor domain was excised by Cre-mediated recombination. Southern blot analysis confirmed correct targeting of four embryonic stem cell clones (flox/+). Two Kif26b^lox/+ embryonic stem cell clones (17 and 72) were injected into eight cell-stage embryos to generate chimeric mice. The chimeras were crossed with C57BL/6 mice to obtain Kif26b floxed heterozygous offspring (Kif26b^flox/+ mice; CDB0800K; http://www.cdb.riken.jp/arg/mutant%20mice%20list.html). When crossed with the mouse expressing Cre recombinase ubiquitously, kidney agenesis or hypoplasia was observed in both strains, which is consistent with the phenotypes of conventional Kif26b-null mice.¹⁸ The primers used for genotyping were as follows: Cre1 (5′-AGGTTCGTTCACTCATGGA-3′) and Cre2 (5′-TCGACCAGTTTAGTTACCC-3′) for the Cre allele (250 bp); Myh9-F (5′-GGGACACAGTTGAATCCCTT-3′) and Myf9-R (5′-GGGCAGGTTCTTATAAGG-3′) for the Myh9 allele (wild type, 600 bp; mutant, 770 bp); Myh10-F (5′-GACCGCTACTATTCAGGACTTATC-3′) and Myh10-R (5′-CAGAGAAACGATGGGAAAGAAAGC-3′) for the Myh10 allele (wild type, 250 bp; mutant, 350 bp); and Kif26bEx7-F(5′-CTTGCTGAGGAGGCTGACA-3′), Kif26bEx7-R(5′-CCAGACATCCATCCCATGAT-3′), and Neo-R2 (5′-CTGTCCATCTGCACGAGACT-3′) for the Kif26b allele (wild type, 750 bp; mutant, 500 bp). PCR amplifications were performed under identical conditions using GoTaq DNA polymerase (Promega) by denaturation at 95°C for 5 minutes followed by 35 cycles of 95°C for 30 seconds, 58°C for 60 seconds, and 72°C for 30 seconds and a final extension at 72°C for 7 minutes. The PCR products were analyzed by electrophoresis in a 1.2% agarose gel and visualized by ethidium bromide staining. BUN, creatinine, and albumin in serum samples were analyzed by the standard methods using ultraviolet, hydroxyl triiodobenzoic acid, and bromocresol green, respectively. All animal experiments were performed in accordance with institutional guidelines and ethical review committees.

Immunostaining

Mouse embryos or kidneys were fixed with 10% formalin or 4% paraformaldehyde and then embedded in paraffin or OCT compound, respectively. Immunostaining was carried out automatically using a BlueMap Kit and automated Discovery System (Roche) or manually for immunofluorescence staining. The following primary antibodies were used: anti-Myh9 (Sigma-Aldrich), anti-Myh10 (Developmental Studies Hybridoma Bank), anti-Myh14 (Covance), anti-Six2 (Proteintech), anti-Sall1³³ (PPMX Perseus Proteomics), anti-Cited1 (Thermo Scientific), anti-Wt1 (Santa Cruz Biotechnology), lotus tetragonolobus lectin (Vector), anti-Aqp1 (Abcam), anti-Tamm–Horsfall protein (Santa Cruz Biotechnology), anti-Slc12a3 (Millipore), anti-cytokeratin (Sigma), anti-cleaved caspase-3 (Cell Signaling), anti-Wnt4 (R&D Systems), anti-Lef1 (Cell Signaling), anti-Jag1 (Santa Cruz Biotechnology), anti-NCAM (Millipore, GeneTex, and Developmental Studies Hybridoma Bank), anti–p-aPKC (Abcam), R-cadherin (Developmental Studies Hybridoma Bank), anti-pHH3 (Millipore), anti–N-cadherin (Santa Cruz), and anti–integrin-α8 (Sigma-Aldrich). Immunofluorescence was visualized with an LSM780 confocal microscope (Zeiss).

Lumen Formation of MDCK Cells

MDCK II cells, which were provided by Akira Kikuchi (Osaka University), were embedded in growth factor–reduced Matrigel (BD Biosciences) as described previously.²² The cells were cultured for 5 days with or without 10 µM (−)-blebbistatin (Wako) and then stained with phalloidin (Invitrogen) or the following antibodies: anti-Myh9 (Covance), anti-Myh10 (Covance), anti-aPKC (Santa Cruz), anti–β-catenin (Cell Signaling), and anti–integrin-β1 (Millipore). Representative data from four independent experiments are shown.

Disclosures

None.