|Home | About | Journals | Submit | Contact Us | Français|
The mammalian metanephric kidney is derived from the intermediate mesoderm. In this report, we use molecular fate mapping to demonstrate that the majority of cell types within the metanephric kidney arise from an Osr1+ population of metanephric progenitor cells. These include the ureteric epithelium of the collecting duct network, the cap mesenchyme and its nephron epithelia derivatives, the interstitial mesenchyme, vasculature and smooth muscle. Temporal fate mapping shows a progressive restriction of Osr1+ cell fates such that at the onset of active nephrogenesis, Osr1 activity is restricted to the Six2+ cap mesenchyme nephron progenitors. However, low-level labeling of Osr1+ cells suggests that the specification of interstitial mesenchyme and cap mesenchyme progenitors occurs within the Osr1+ population prior to the onset of metanephric development. Furthermore, although Osr1+ progenitors give rise to much of the kidney, Osr1 function is only essential for the development of the nephron progenitor compartment. These studies provide new insights into the cellular origins of metanephric kidney structures and lend support to a model where Osr1 function is limited to establishing the nephron progenitor pool.
The nephron is the basic functional unit of the metanephric kidney. Depending on the mammalian species, tens to hundreds of thousands of nephrons are formed over an extensive period of fetal and neonatal nephrogenesis (Cullen-McEwen et al., 2001; Nyengaard and Bendtsen, 1992). The main body of the nephron is a tubular epithelial network comprised of specialized segments along it proximal to distal axis (Yu et al., 2004). The interface of the vasculature and the nephron epithelium occurs at the proximal end of the nephron in the glomerulus where podocytes of the nephron epithelia lie opposed to a specialized vascular and supporting cell (mesangial cell) network (Spinelli, 1974). Blood filtrate travels along the nephron tubule where all essential small molecules in the filtrate are resorbed into a closely associated vasculature (Reeves et al., 2001). The distal end of the nephron fuses with the epithelial collecting system enabling the passage and final adjustment of the water/salt balance of the urine prior to its removal via the ureter into the bladder (Meneton et al., 2004). Smooth muscle surrounding the medullary collecting system and ureter facilitate the fluid movement in these final stages (Yu et al., 2002). In addition to the cell populations above, the nephrons are embedded within an interstitial mesenchyme whose function is poorly understood (Hatini et al., 1996; Levinson et al., 2005).
Recent studies have started to probe the cellular origins and cellular relationships among the component parts of the kidney to determine the developmental mechanisms that establish a functional upper urinary system (Boyle et al., 2008; Kobayashi et al., 2008; Yu et al., 2002). The metanephric kidney of the mouse arises from within the posterior intermediate mesoderm (IM) around embryonic day 10.5 (E10.5) (Saxén, 1987; Vize et al., 2002). At this time, Gdnf expression within pre-specified metanephric IM mesenchyme at the hindlimb level induces an outgrowth of the adjacent nephric duct, the ureteric bud (Grieshammer et al., 2004; Kume et al., 2000; Sanchez et al., 1996b). On entering the metanephric mesenchyme, cell interactions between the ureteric bud and adjacent mesenchyme drive assembly of the functional kidney. In this, the mesenchyme continues to stimulate branching growth of the epithelium, generating the arborized network of the urine transporting collecting system (Grieshammer et al., 2004; Pepicelli et al., 1997; Sanchez et al., 1996b; Yu et al., 2002).
With each branching event, a portion of the metanephric mesenchyme adjacent to the ureteric bud tips, the cap mesenchyme, is induced by the ureteric epithelium to undergo a mesenchymal to epithelial transition (MET). This induction establishes the renal vesicle, the epithelial precursor of the main body of the nephron (Carroll et al., 2005; Dressler, 2002; Park et al., 2007; Stark et al., 1994). The cap mesenchyme is a self-renewing population of renal vesicle progenitors and expresses the transcriptional regulator Six2, which is required for its self-renewal (Boyle et al., 2008; Kobayashi et al., 2008; Self et al., 2006). From E11.5 through post-natal day 3, this process reiterates in conjunction with branching growth of the ureteric epithelium to generate the full complement of nephrons (Hartman et al., 2007).
A second population of Foxd1+ mesenchyme surrounds Six2+ nephron progenitors (Hatini et al., 1996; Levinson et al., 2005). While the fate of this cell population has not been addressed, they are thought to give rise to the interstitial components of the metanephric kidney. Origins of the vascular, mesangial and smooth muscle compartments are not fully resolved, though lineage-tracing studies suggest that the vasculature may arise intrinsically within the early metanephros or extrinsically by migration into the developing kidney (Bernstein et al., 1981; Loughna et al., 1997; Loughna et al., 1998; Sariola et al., 1984).
The transcriptional regulator Osr1 is expressed broadly in the early IM mesenchyme beginning at E7.5 and its activity is essential for development of multiple IM derived structures (James et al., 2006; Wang et al., 2005). Notably, in the presumptive metanephric mesenchyme of E10.5 Osr1 mutant embryos, the expression of several key regulators of kidney development, including Pax2, Six2, Sall1, Eya1 and Gdnf, is absent and the metanephric mesenchyme undergoes apoptosis. Thus, Osr1 has an essential function at the outset of metanephric kidney development (James et al., 2006; Wang et al., 2005).
In this study, we have used a temporal fate mapping approach to address the relationship between Osr1+ cells and multiple cell lineages in the mouse kidney. We demonstrate that many distinct cellular compartments emerge from the Osr1+ population prior to the appearance of the metanephric anlagen. Furthermore, there is a separation of molecularly distinct Six2+ and Foxd1+ progenitor pools prior at the onset of metanephric development. Finally, Osr1 function is itself required for formation of the Six2+ renal vesicle progenitors, but not for the formation of interstitial precursors or ureteric epithelium components.
To examine the relationship between Osr1+ descendants and distinct cell lineages in the kidney, we targeted a cDNA encoding an eGFPCreERt2 (GCE) cassette to the endogenous Osr1 locus, replacing the start codon of Osr1 with the start codon of the GCE (Fig. S1A) (Mugford et al., 2008). The use of GCE allows for both a spatial and temporal analysis of Osr1+ cell fate. Though this strategy creates a null mutation in one Osr1 allele, Osr1eGFPCreERt2/+ (Osr1GCE/+) heterozygotes were both viable and fertile (Wang et al., 2005). Whole mount analysis of wild-type and Osr1GCE/+ E8.5 (Fig. S1B–D) and E9.5 (Fig. S1F–H) embryos demonstrated that GFP and Osr1 mRNA co-localize to IM and dorsal lateral plate mesoderm as expected (Fig. S1E, I). Furthermore, GFP in the Six2+ cap mesenchyme at E15.5 (Fig. S1K) mirrors endogenous Osr1 expression (Fig. S1J). Thus, the GCE cassette faithfully recapitulates the endogenous expression pattern of Osr1.
To identify the fate of Osr1+ cells at specific developmental time points, we intercrossed Osr1GCE/+ animals with Gt(ROSA)26tm1Sor (R26R) (Soriano, 1999) animals and dams were injected with Tamoxifen (TM) at 24 hour intervals from E6.5 through E11.5 with a TM dose that enables Cre activity for less than 24 hours (Kobayashi et al., 2008) (see Materials and Methods). This pulse labeling strategy covers time points prior to the onset of Osr1 expression through the first inductive events within the metanephric kidney. The contribution of Osr1+ descendants to various renal cell types was determined at E15.5.
Low resolution analysis of X-gal staining demonstrates that Cre activity was dependent upon TM as expected; no β-gal activity was present in Osr1GCE/+;R26R embryos of dams injected with carrier (oil) alone between E7.5 and E9.5 (Fig. S2A). When TM was injected at E6.5 (early gastrulation), a small number cells within the ureteric epithelium were X-gal+ in E15.5 Osr1GCE/+;R26R kidneys (data not shown). When Osr1GCE/+;R26R embryos were exposed to TM from E7.5 through E11.5, X-gal+ cells a variable contribution was observed to multiple cell types of the E15.5 metanephric kidney, gonad and adrenal gland (Fig. S2B–F).
To identify the exact cell types to which Osr1+ descendents contributed, we compiled a set of cell specific markers (Table 1) and performed dual immunofluorescence with antibodies directed against β-galactosidase (β-gal) focusing initially on the progenitor compartments of the nephrogenic zone; the ureteric epithelium, cap mesenchyme and cortical interstitial mesenchyme.
Six2 and Foxd1 are present within renal vesicle and putative interstitial progenitors, respectively (Hatini et al., 1996; Self et al., 2006). In order to establish Six2 and Foxd1 as definitive markers of the cap mesenchyme and cortical interstitial mesenchyme, respectively, we generated antibodies against Six2 and Foxd1 and examined their distribution relative to each other and Pax2, a marker of both the nephric duct (Fig. 1B, F, J, N, R, V, dashed lines) and the cap mesenchyme (Fig. 1B, F, J, N, R, V) (Torres et al., 1995). Foxd1 was present in cells of the interstitial mesenchyme progenitors from E10.5 through E15.5 (Fig. 1A, E, I) and Pax2+ cells of the nephric duct and the cap mesenchyme were Foxd1− (Fig. 1C, D, G, H, K, L). At all time points examined, the majority of Foxd1+ cells surrounded Pax2+ cells of the cap mesenchyme and after branching had occurred, Foxd1+ cells were found in the cleft of the branching ureteric epithelium (Fig. 1E, I, G, K and data not shown). In contrast, Six2+ cells were present within the cap mesenchyme and Six2 was down-regulated as these cells underwent MET (Fig. 1M, Q, U) (Self et al., 2006). Unlike Foxd1, Six2 and Pax2 were present in all cap mesenchymal cells, but Six2 was absent from the Pax2+ collecting duct epithelium (Fig. 1O, P, S, T, W, X). Thus, from the onset of metanephric development at E10.5, Six2 and Foxd1 demarcate mutually exclusive progenitor compartments; Six2+ cap mesenchyme renal vesicle progenitors and Foxd1+ putative interstitial mesenchyme progenitors.
At E15.5, cytokeratins are present solely within the cells of the presumptive collecting duct epithelium (Fig. 2A, E, I, M, Q) (Fleming and Symes, 1987). If TM was administered at either E7.5 (Fig. S2B and Fig. 2A–D) or E8.5 (Fig. S2C and Fig. 2E–H) β-gal+, cytokeratin+ cells were detected in E15.5 Osr1GCE/+;R26R kidneys. Conversely, if TM was administered at E9.5 or later (Fig. S2D–F and Fig. 2I–L, M-P, Q–T), E15.5 cytokeratin+ cells were β-gal−. Thus, the collecting duct lineage arises from an Osr1+ cell-type prior to E9.5.
Foxd1+, β-gal+ cells were detected in E15.5 Osr1GCE/+;R26R kidneys following GCE induction between E7.5 and E10.5 but not at E11.5 (Fig. S2B–F and Fig. 3A–D, E–H, I–L, Q–T, arrowheads). Thus, Osr1+ cells contribute to the Foxd1+ compartment up until the onset of nephrogenesis. Finally, Six2+, β-gal+ cells were detected following TM administration at all time points consistent with the continued expression of Osr1 within Six2+ renal vesicle progenitors after nephrogenesis commences (Fig. S1K, S2B–F and Fig. 4A–D, E–H, I–L, J–K, Q–T, arrowheads).
Though our earlier data indicates that Foxd1 and Six2 are mutually exclusive populations of cells (Fig. 1), descendants of Osr1+ cells contributed to both Foxd1+ and Six2+ compartments upon E10.5 TM administration. This suggests that Osr1 is not restricted to Six2+ cap mesenchyme until after E11.5. Indeed, at E10.5, GFP was clearly present in both Foxd1+ (Fig. S3A–D, arrowheads) and Six2+ cells (Fig. S3E–H, arrowheads); however, the intensity of the GFP signal was considerably higher in Six2+ cells. At E11.5, Foxd1+ cells were GFP− (Fig. S3I–L, arrowheads); only Six2+ cells remained GFP+ (Fig. S3M–P, arrowheads).
Taken together, these results demonstrate that Osr1+ cells have the potential to give rise to all cortical progenitor compartments of the metanephric kidney. Of note, the Osr1+ population exhibits its broadest potential between E7.5 and E9.5. The restriction of the Osr1+ population to the cap mesenchyme mirrors the spatial restriction of Osr1 to Six2+ cells by E11.5. However, this restriction was delayed relative to the apparent separation of the nephron and interstitial lineages since Six2+ cells do not contribute to the Foxd1+ compartment (Kobayashi et al., 2008) and Six2+ and Foxd1+ cells are mutually exclusive (Fig. 1). Thus, Osr1 regulation is likely distinct from the mechanisms that control lineage specification within the mesenchymal progenitor compartments.
To examine more broadly the contribution of β-gal+ cells in E15.5 Osr1GCE/+;R26R kidneys, we next focused on the vascular-renal tubule interface within the glomerulus, which consists of Flk1+ vascular endothelial cells (Robert et al., 1998), PDGFRb+ mesangial support cells (Hellstrom et al., 1999) and WT1+ podocytes (Mundlos et al., 1993).
As expected, WT1+ podocytes, which are derivatives of the Six2+ cap mesenchyme (Boyle et al., 2008; Kobayashi et al., 2008), were β-gal+ at all TM pulse time points (Fig. 5E–G, I–K, M–O, Q–S, concave arrowheads, Fig. S4 and data not shown). TM administration at E7.5 (Fig. 5A–D, arrowheads) and E8.5 (Fig. 5E–G, arrowheads) generated β-gal+, Flk1+ endothelial cells in the glomerulus. In addition, the extraglomerular renal vasculature and the vasculature in other organs, including the neural tube, lung and limbs, was also β-gal+ (data not shown). However, no β-gal+, Flk1+ cells were detected following TM administration at or after E9.5 (Fig. 5I–L, M–O, Q–T, arrowheads). In contrast, Osr1+ descendants contributed to PDGFRb+ mesangial cells at all but the latest time point (E11.5) (Fig. S4A–D, E–G, I–L, M–O, arrowheads). Furthermore, PDGFRb+, β-gal+ pericytes were detected adjacent to non-glomerular associated kidney vasculature at the same pulse time points (Fig. S4E–G, I–K, Q–T concave arrowheads and data not shown).
The kidney capsule, a layer of epithelial cells surrounding the cortical surface of the metanephric kidney (Levinson et al., 2005), is recognized by the cell surface antigen Liv2 (Fig. S5A, E, I, M, Q) (Y.Q. Soh and A.P. McMahon, unpublished observation). Like Foxd1+ interstitial progenitors, PDGFRb+ pericytes and mesangium, Liv2+, β-gal+ cells were only labeled on TM induction between E7.5 and E10.5 (Fig. S5A–D, E–H, I–L, M–P, Q–T arrowheads). That these diverse cell types show a common labeling property may reflect a similar lineage relationship from Osr1+ cells.
In contrast, smooth muscle progenitors displayed a distinct temporal lineage relationship from the Osr1 population. At E15.5, smooth muscle identified by Smooth Muscle α-Actin (SMA) surrounds the ureter (Yu et al., 2002). This population was labeled upon TM administration from E8.5 through E10.5 (Fig. S6E–H, I–L, M–O, arrowheads), but not at E7.5 (Fig. S6A–D, arrowheads) or E11.5 (Fig. S6Q–T). This indicates that the Osr1+ population is not be fixed at E7.5. Rather, new cells with smooth muscle generating potential are recruited by E8.5.
Finally, macrophages reside among cells of the interstitial stroma and can be identified by the G-protein coupled receptor F4/80 (Challen et al., 2006). No F4/80+, β-gal+ macrophages were detected in E15.5 Osr1GCE/+; R26R kidneys when TM was administered anytime between E7.5 and E11.5 (Fig. S7A–D, E–G, I–L, M–O, Q–T, arrowheads). Taken together, these results demonstrate that descendants of Osr1+ cells give rise to a broad array of metanephric cell types including the ureteric epithelium, nephron and interstitial mesenchyme progenitors, vasculature, pericytes, mesangium, the kidney capsule and smooth muscle. Thus the Osr1+ population is the early progenitor for most kidney populations. However, descendants of Osr1+ cells are not restricted to the kidney; notably, descendants contribute to the vasculature of multiple organs that are not IM-derived.
Although the Osr1+ population is multi-potent, the potential of an individual Osr1+ cell remains an open question. Development of the metanephric kidney mesenchyme is first evident at E10.5. At this time, Six2+ and Foxd1+ cells form mutually exclusive compartments, though both cell types express Osr1 (Fig. 1). Thus, a single Osr1+ cell prior to E10.5 may generate both progenitor cell types. Alternatively, distinct Osr1+ cells may independently generate the Six2+ and Foxd1+ cell compartments.
Previous reports have demonstrated that low doses of TM can activate a limited number of CreERt2+ cells, allowing for clonal analysis (Kobayashi et al., 2008; Zhou et al., 2007). In an attempt to address the potential of a single Osr1+ cell, we injected dams with low doses of TM at E9.5 (see Materials and Methods) and performed immunofluorescence for β-gal, Pax2 and Foxd1 at E10.5. If a single Osr1+ cell is multipotent, then adjacent Pax2+, β-gal+ and Foxd1+, β-gal+ cells may be detected at the interface of the E10.5 interstitial and cap mesenchyme. Alternatively, if the Osr1+ population is segregated into Six2+, Pax2+ and Foxd1+ cellular compartments, then adjacent β-gal+ cells should both be either Pax2+ or Foxd1+. In the absence of information regarding cell motility or precise mitotic indices, adjacent cells are defined as two nuclei within a distance no greater than 10µm between the centers of their nuclei.
Most β-gal+ cells did not occur as adjacent pairs; single β-gal+ cells were generally spaced at distances greater than 15µm from each other. Only 16 pairs of adjacent β-gal+ in 5 embryos were identified and most were not located at the border between the stromal and cap mesenchyme. No adjacent β-gal+, Foxd1+ cells were detected. Adjacent β-gal+ cells within the Foxd1+ compartment consisted of one β-gal+, Foxd1+ cell and one β-gal+, Foxd1− cell (5/11 pairs in 5 embryos) (Fig. 6A–D, brackets E, arrows) or both cells were β-gal+, Foxd1− (4/11 pairs in 5 embryos) (Fig. 6F–I, brackets I, arrows). Within the Pax2+ cap mesenchyme, β-gal+, Pax2+ adjacent pairs were identified; however, in all cases both nuclei were β-gal+, Pax2+ (5/5 pairs in 5 embryos) (Fig. 6K–N, brackets O, arrowheads). No adjacent β-gal+ pairs were detected where one cell was Foxd1+ and the other Pax2+. Taken together, this data suggests that if a single Osr1+ cell has the potential to give rise to both cap mesenchyme and interstitial mesenchyme progenitors, this cell is a rare member of the Osr1+ population between E9.5 and E10.5. Thus, interstitial mesenchyme and cap mesenchyme are likely separate lineages prior to the onset of metanephric development.
Given these results, we next addressed the requirement for Osr1 in development of the interstitial mesenchyme population. Kidneys do not develop in Osr1 mutants; the absence of a number of genes essential for normal cap mesenchyme development including Pax2, Sall1, Six2, Eya1 and Gdnf indicates that cap mesenchyme progenitors are absent at E10.5 (James et al., 2006; Wang et al., 2005). However, the Osr1 locus is still active in Osr1 mutants (James et al., 2006); consequently, Osr1 mutant IM cells can be identified by the presence of GFP in E10.5 Osr1 mutants. In order to generate embryos lacking all Osr1 activity, we removed the PGK Neo cassette from the GCE allele (GCEN) and intercrossed Osr1GCE/+ and Osr1GCEN/+ animals (see Materials and Methods).
In agreement with previously reports (James et al., 2006; Wang et al., 2005), Pax2 was present in the nephric duct epithelium in embryos lacking Osr1 activity (dashed lines in Fig. 7B and L, G and Q), but was absent from the GFP+ posterior IM mesenchyme in Osr1GCE/GCEN mutants (compare Fig. 7B–D, L–N with G–I, Q–S). Six2 was also absent from the GFP+ E10.5 posterior IM of Osr1GCE/GCEN mutants (data not shown). Conversely, Foxd1+, GFP+ cells were present in the posterior IM of E10.5 Osr1GCE/+ (Fig. 7A, C, D, arrowheads) and E10.5 Osr1GCE/GCEN embryos (Fig. 7F, H, I arrowheads).
The Hox11 paralogs play a crucial role in the axial specification and patterning of the posterior IM mesenchyme (Mugford et al., 2008; Patterson et al., 2001; Wellik et al., 2002). To determine whether the metanephric mesenchyme is incorrectly specified in Osr1 mutants, Hoxd11 activity was examined. Hoxd11+, GFP+ cells were detected throughout the posterior IM of E10.5 Osr1GCE/+ embryos, including the Pax2+ cap mesenchyme and the Pax2− interstitial mesenchyme (Fig. 7K–O, concave arrowheads and arrowheads, respectively). In E10.5 Osr1GCE/GCEN embryos, Hoxd11 was present within mesenchymal GFP+ cells (Fig. 7P–T, arrowheads), but no Hoxd11+, Pax2+, GFP+ cells were detected (Fig. 7P–T). Thus, an appropriate Foxd1+, Hoxd11+ interstitial compartment appears to form in the absence of Osr1 function, but the cap mesenchyme progenitor compartment was not established.
We have demonstrated that descendants of Osr1+ cells initially contribute to a variety of IM derived cell types including cells within the adrenal gland, gonad and metanephric kidney (Fig. 8). Ultimately, Osr1 expression and consequently Osr1+ descendant cell fates are restricted to metanephric nephron progenitors at the onset of active nephrogenesis; Osr1 function is essential for the establishment of this progenitor pool. Analysis of Osr1+ descendant cell fate and restriction gives new insights regarding the specification and cellular behaviors of metanephric progenitor domains.
The nephric duct is thought to be the first committed cell population to emerge from the IM mesenchyme (Grote et al., 2006; Pedersen et al., 2005). In agreement with this interpretation, Osr1+ descendants only contribute to the ureteric epithelium, an outgrowth of the nephric duct, prior to E9.5 coincident with the initial appearance of this epithelial structure. Thus, it is likely that growth rather than continued cell recruitment from an Osr1+ mesenchymal population accounts for posterior extension of the nephric duct.
Interestingly, between E7.5 and E11.5, Osr1+ descendants contribute to interstitial mesenchyme precursors, pericytes, mesangial cells and capsule cells; a shared temporal labeling consistent with a possible common lineage amongst these cell types. We also demonstrate that descendants of Osr1+ cells give rise to renal smooth muscle lining the ureter. The contribution of Osr1+ descendants to smooth muscle did not temporally correlate with their contribution to any other metanephric progenitor population, suggesting that metanephric smooth muscle does not arise from the cap mesenchyme, ureteric epithelium or interstitial mesenchyme progenitors. This is in agreement with past reports demonstrating that descendants of these progenitor populations are excluded from smooth muscle (Boyle et al., 2008; Kobayashi et al., 2008; Yu et al., 2002). Thus, the incorporation of a smooth muscle population likely reflects de novo expression of Osr1 in a new cell population around E8.5. Of note, Osr1 is present in both IM and dorsal lateral plate mesoderm, but the precise timing of expression within these compartments is hard to assess due to the absence of tightly restricted lineage markers.
Previous fate mapping experiments have used Cre driver lines to directly label the Six2+ cap mesenchyme and Hoxb7+ ureteric epithelial progenitor compartments within the metanephric kidney and have demonstrated that the descendants of these populations are exclusive of each other (Boyle et al., 2008; Kobayashi et al., 2008; Yu et al., 2002). Though these progenitor populations are spatially and molecularly distinct at the onset of metanephric development, they arise from a common Osr1+ precursor population. Within the mesenchyme, labeling studies suggest that the Osr1+ IM mesenchyme is predominantly a heterogeneous population of cells already committed to either cap mesenchyme or interstitial progenitor cell fates by E10.5. It should be noted that our analysis cannot rule out the possibility of a rare multipotent Osr1+ cell, nor can we exclude the possibility of rapidly dividing and/or migrating IM mesenchyme progenitors. How distinct cell fates emerge from an Osr1+ progenitor pool remains to be determined, but signaling from adjacent tissues (e.g. somatic mesoderm) may play a significant role.
Vascular precursors are thought to arise from the splanchnic mesoderm of the lateral plate (Poole et al., 2001; Saha et al., 2004), while the Aorta-Gonad-Mesonephros (AGM) region of the developing embryo is the origin of definitive heamatopoeitic stem cells (HSCs) (Cumano et al., 2000; Kumaravelu et al., 2002; Medvinsky and Dzierzak, 1996; Minehata et al., 2002; Robin and Dzierzak, 2005; Sanchez et al., 1996a). A progenitor cell, the hemangioblast, has been proposed to give rise to both vasculature and hematopoeitic lineages, though the cells existence is somewhat controversial (Xiong, 2008).
Osr1 is expressed in the presumptive AGM and lateral plate (James et al., 2006; Wang et al., 2005); this study) and we have previously shown that descendants of Osr1+ cells contribute to the mesonephros (Mugford et al., 2008). Here, we demonstrate that while Osr1 descendants give rise to the vasculature of many organ systems, they do not contribute to renal macrophage, derivatives of HSCs (Gordon and Taylor, 2005). This suggests that Osr1+ cells and their descendants, though located within the developing AGM, are excluded from the heamatopoeitic lineage. Thus, if there is a cell type generating the blood and vascular lineages, Osr1 is only activated after the separation of the vascular lineage. Various studies have suggested an IM source and a non-IM source (Bernstein et al., 1981; Loughna et al., 1997; Loughna et al., 1998; Sariola et al., 1984) of vascular progenitors in the kidney. Osr1 is expressed in both the IM and LPM potentially supporting either view. However, definitive evidence would require an independent analysis of the vascular forming potential of the IM and LPM components of the Osr1 expression domain.
Functionally, Osr1 is upstream of multiple transcriptional regulators required for the development of the cap mesenchyme (James et al., 2006; this study). However, its function is not required for the formation of the nephric duct or for the specification of Hoxd11+, Foxd1+ interstitial mesenchyme progenitors (James et al., 2006; this study). Together, this suggests that despite its early broad expression in cells that generate multiple cellular compartments in the adult kidney, Osr1 is not a master regulator of the IM. Rather Osr1 is required specifically for the development of the renal vesicle forming cap mesenchyme lineage.
Recent reports suggest that Osr1 encodes a transcriptional repressor that may recruit Groucho-like co-repressors to promote a mesenchymal precursor state (Goldstein et al., 2005; James et al., 2006; Tena et al., 2007). Once nephron and interstitial progenitors are separated, Osr1 may function to activate a cap mesenchyme specific program, transcriptionally suppressing an interstitial progenitor program. While this interpretation remains speculative in the absence of direct information of Osr1 targets, its critical role in establishing the renal vesicle progenitor pool is clear.
Osr1 repressor function is essential for the specification of the cap mesenchyme. In this, Osr1 is genetically upstream of Pax2, Eya1, Six2 and Gdnf (James et al., 2006; this study). Six2 function is required within the cap mesenchyme, not for its specification, but to maintain this renal vesicle forming progenitor population in a self-renewing mesenchymal state (Kobayashi et al., 2008; Self et al., 2006). Gdnf is required for the initial invasion and continual branching of the ureteric epithelium within the metanephric mesenchyme (Costantini, 2006). Six2 and Gdnf are both synergistically activated by a Pax2-Eya1-Hox11 complex (Gong et al., 2007). Thus, Osr1 may provide a molecular link between progenitor maintenance and ureteric epithelial branching through the regulation of a Pax2-Eya1-Hox11 activation complex and in turn, the regulation of Six2 and Gdnf expression. Furthermore, the maintenance of Osr1 in the cap mesenchyme throughout the period of nephrogenesis may ensure both the continued renewal of nephron progenitors and branching morphogenesis throughout kidney organogenesis.
Animal care and research protocols were performed in accordance with Harvard University’s institutional guidelines, following approval by Harvard University’s institutional committee on animal use. For staging of embryos, the morning of vaginal plug was designated as embryonic day 0.5 (E0.5). Gt(ROSA)26tm1Sor (Soriano, 1999) (Jackson Laboratories) were genotyped as previously described. The Osr1eGFPCreERt2/+ line (Mugford et al., 2008) (Osr1GCE/+) was generated by knocking an eGFPCreERt2 (GCE) construct into the endogenous Osr1 locus. 3’ and 5’ homology arms were PCR amplified from BAC clone RP23-398M9. All primer sequences used to generate homology arms are available upon request. Osr1GCE/+ heterozygotes were genotyped as previously described (Mugford et al., 2008). Tamoxifen was administered at a dose of either 1mg/40g body weight (general cell fate studies) or 0.05mg/40g body weight (low-level cell labeling studies) of the dam.
The PGK Neo selection cassette was removed by intercrossing Osr1GCE/+ heterozygotes with the Gt(ROSA)26Sortm1(FLP1)Dym line (Farley et al., 2000). PGK Neo removal in the resulting allele Osr1GCEN/+ was confirmed by PCR using the Osr1ER3’GenoRv reverse primer (Mugford et al., 2008) with primer Osr1NoNeoFw: 5’-CGTGGAGGAGACGGACCAAA-3’. Osr1GCEN/+;R26R animals demonstrated ectopic Cre activity in the absence of Tamoxifen (data not shown) and were therefore not used for fate mapping studies. Osr1 mutant embryos were generated by intercrossing Osr1GCE/+ heterozygote animals with Osr1GCEN/+ heterozygote animals. Embryos were genotyped for the presence of both the GCE and GCEN alleles and Osr1GCE/GCEN embryos were obtained at expected Mendelian ratios.
Rabbits were immunized with a KHL-conjugated peptide RLPSSAQRRRRSYAGEDDLE corresponding to amino acids 48–67 of mouse Foxd1 (Covance Research Products). The antiserum was tested by transiently transfecting (Lipofectamine 2000, Invitrogen) COS7 cells with a mouse Foxd1 expression plasmid followed by immunostaining. Immunofluorescence was also performed on E15.5 metanephric cryosections and compared to section in situ hybridization for Foxd1 transcript.
All tissue preparation, whole mount analysis and immunofluorescence was carried out as previously described (Mugford et al., 2008). Confocal images were acquired on a Zeiss LSM510 META confocal microscope (Zeiss). The following antibodies and dilutions were used: anti-Hoxd11 (1:500) (Mugford et al., 2008), anti-Foxd1 (1:2000), anti-Six2 (1:1000) (Kobayashi et al., 2008), anti-GFP (1:500, AvesLabs), anti-cytokeratin (1:500, Sigma), anti-β-galactosidase (1:500, AbCam), anti-β-gal (1:50, DSHB), anti-Liv2 (1:200, MBL International), anti-F4/80 (1:1000, eBiosciences), anti-Smooth Muscle α-actin (1:2000, Sigma), anti-Flk1 (1:1000, BDPharmigen), anti-PDGFRb (1:100, eBiosciences). Appropriate Cy2, Cy5 (1:500, Jackson Immuno), Alexa488, Alexa568 or Alexa647 (1:500, Invitrogen) conjugated secondary antibodies were used to detect primary antibodies. Nuclei were stained with Hoechst 33342 (Invitrogen). Experiments requiring the simultaneous use of two rabbit antibodies were conducted using Zenon dual labeling kits (Invitrogen).
(A) Targeting strategy for eGFPCreERt2 into the endogenous Osr1 locus. (B–I) Comparison of Osr1 mRNA expression pattern with eGFP in Osr1GCE/+ embryos. Whole mount analysis comparing Osr1 mRNA in WT E8.5 (B) and E9.5 (F) embryos with eGFP in Osr1GCE/+ E8.5 (C, D) and E9.5 (G, H) embryos. Transverse sections of E8.5 (E) and E9.5 (I) Osr1GCE/+ embryos immunostained for GFP. Section analysis in E15.5 metanephric kidneys of WT embryos for Osr1 mRNA (J) and in Osr1GCE/+ embryos immunostained for GFP (green), Six2 (red) and Cytokeratin (purple). Scale bars = 100µm.
(A–F) X-gal staining in frontal sections of Osr1GCE/+;R26R E15.5 metanephric kidneys injected with Oil alone between E7.5 and E9.5 (A), Tamoxifen at E7.5 (B), E8.5 (C), E9.5 (D), E10.5 (E) and E11.5 (F). Scale bar = 500µm.
(A–P) Immunofluorescent confocal microscopy of transverse sections of the E10.5 metanephric blastema (A–H) and saggital sections of E11.5 metanephric kidneys (I–P). Samples immunostained for Foxd1 (red in A, E), Six2 (red in I, M), GFP (green in B, F, J, N) and Cytokeratin (blue in A, B, E, F, I, J). Nuclei stained for Hoechst 33258 (D, H, L, P). Merged images (C, G, K, O). Arrowheads (A–C) indicate Foxd1+, GFP+ cells. Arrowheads (I–K) indicate Foxd1+, GFP− cells. Arrowheads (E–G, M–O) indicate Six2+, GFP+ cells. Scale bars = 50µm.
(A–T) Immunofluorescent confocal microscopy of glomeruli within E15.5 Osr1GCE/+;R26R metanephric kidneys after Cre activation at E7.5 (A–D), E8.5 (E–H), E9.5 (I–L), E10.5 (M–P) and E11.5 (Q–T). Samples immunostained for Flk1 (green in A, E, I, M, Q), WT1 (blue in A, B, E, F, I, J, M, N, Q, R) and β-gal (red in B, F, J, N, R). Nuclei stained for Hoechst 33258 (D, H, L, P, T). Merged images (C, G, K, O, S). Arrowheads (A–C, E–G, I–K, M–O) indicate PDGFRb+, β-gal+ mesangial cells. Concave arrowheads (E–G, I–K) indicate PDGFRb+, β-gal+ pericytes. Scale bar = 50µm.
(A–T) Immunofluorescent confocal microscopy of E15.5 Osr1GCE/+;R26R metanephric kidneys after Cre activation at E7.5 (A–D), E8.5 (E–H), E9.5 (I–L), E10.5 (M–P) and E11.5 (Q–T). Samples immunostained for Liv2 (A, E, I, M, Q) and β-gal (B, F, J, N, R). Nuclei stained for Hoechst 33258 (D, H, L, P, T). Merged images (C, G, K, O, S). Arrowheads (A–C, E–G, I–K, M–O) indicate Liv2+, β-gal+ cells. Arrowheads (Q–S) indicate Liv2−, β-gal+ cells. Scale bar = 50µm.
(A–T) Immunofluorescent confocal microscopy of E15.5 Osr1GCE/+;R26R ureters after Cre activation at E7.5 (A–D), E8.5 (E–H), E9.5 (I–L), E10.5 (M–P) and E11.5 (Q–T). Samples immunostained for SMA (A, E, I, M, Q) and β-gal (B, F, J, N, R). Nuclei stained for Hoechst 33258 (D, H, L, P, T). Merged images (C, G, K, O, S). Arrowheads (E–G, I–K, M–O) indicate SMA+, β-gal+ cells. Arrowheads (A–C) indicate SMA−, β-gal+ cells. Scale bar = 50µm.
(A–T) Immunofluorescent confocal microscopy of E15.5 Osr1GCE/+;R26R metanephric kidneys after Cre activation at E7.5 (A–D), E8.5 (E–H), E9.5 (I–L), E10.5 (M–P) and E11.5 (Q–T). Samples immunostained for F4/80 (A, E, I, M, Q) and β-gal (B, F, J, N, R). Nuclei stained for Hoechst 33258 (D, H, L, P, T). Merged images (C, G, K, O, S). Arrowheads (A–C, E–G, I–K, M–O, Q–S) F4/80−, β-gal+ cells. Scale bar = 50µm
We thank Dr. Akio Kobayashi for help in generating the eGFPCreERt2 cassette and Ying Qi (Shirleen) Soh for communicating her Liv2 data. Work in A.P.M.’s laboratory was supported by a grant from the NIH (DK054364). P.S. was funded by grants from the Finnish Academy (#107827), Helsingin Sanomain 100-vuotis and Alfred Kordelin Foundations.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.