Bmp7 maintains undifferentiated kidney progenitor population and determines nephron numbers at birth

Abstract

The number of nephrons, the functional units of the kidney, varies among individuals. A low nephron number at birth is associated with a risk of hypertension and the progression of renal insufficiency. The molecular mechanisms determining nephron number during embryogenesis have not yet been clarified. Germline knockout of bone morphogenetic protein 7 (Bmp7) results in massive apoptosis of the kidney progenitor cells and defects in early stages of nephrogenesis. This phenotype has precluded analysis of Bmp7 function in the later stage of nephrogenesis. In this study, utilization of conditional null allele of Bmp7 in combination with systemic inducible Cre deleter mice enabled us to analyze Bmp7 function at desired time points during kidney development, and to discover the novel function of Bmp7 to inhibit the precocious differentiation of the progenitor cells to nephron. Systemic knockout of Bmp7 in vivo after the initiation of kidney development results in the precocious differentiation of the kidney progenitor cells to nephron, in addition to the prominent apoptosis of progenitor cells. We also confirmed that in vitro knockout of Bmp7 in kidney explant culture results in the accelerated differentiation of progenitor population. Finally we utilized colony-forming assays and demonstrated that Bmp7 inhibits epithelialization and differentiation of the kidney progenitor cells. These results indicate that the function of Bmp7 to inhibit the precocious differentiation of the progenitor cells together with its anti-apoptotic effect on progenitor cells coordinately maintains renal progenitor pool in undifferentiated status, and determines the nephron number at birth.

Figure 1. Acceleration of nephron maturation and apoptosis of cap mesenchyme at E14.5 in Bmp7 knockout kidneys.

(A) Pregnant mothers bearing Bmp7^LacZ/fl;Gt(ROSA)26Sor^CreERT2 and Bmp7^+/fl;Gt(ROSA)26Sor^CreERT2 embryos were administered tamoxifen at E12.5, and sacrificed at E14.5. (B) The expression of Bmp7 mRNA was significantly reduced at E14.5 in Bmp7^LacZ/fl;Gt(ROSA)26Sor^CreERT2 (Bmp7 knockout kidneys) (n = 4, white column) compared to control kindeys (n = 5, black column). Data are represented as mean ± SD. (C and D) Cap mesenchyme was maintained in both Bmp7 knockout kidneys and control kidneys. (E) Cap mesenchyme positive for WT1 was maintained in both genotypes. (F) The distribution and the number of phospho-Histone H3-positive cells were comparable in both genotypes. pHistone H3 denotes phospho-Histone H3. (G) TUNEL-positive cells were increased in Bmp7 knockout kidneys. (H) The number of TUNEL-positive mesenchymal cells per section was increased in knockout kidneys (white column) compared to control kidneys (black column). Data are represented as mean ± SD. The average numbers of TUNEL-positive mesenchymal cells in three sections were used as the values for each embryo. The mean of the values from three embryos is presented in the graph. (I) Bmp7 knockout kidneys exhibited inappropriately “mature” glomeruli at E14.5. (J) Glomeruli in the mature stages were increased in Bmp7 knockout kidneys (white column) compared to the control kidneys (black column). Criteria for the assessment of glomerular maturity are detailed in method section. Because C stage is the most common in control kidneys at E14.5, we summed up the glomeruli of stages I, II, and III for comparison of maturation. Glomeruli were counted in 5 slices of each kidney. The mean of the values ± SD is presented in the graphs (n = 7 for control embryos, and 5 for knockout embryos). Scale bars: 100 µm (C, F, G) or 50 µm (D, E, I). n.s.: not significant.

Figure 2. Acceleration of nephron maturation and reduction of cap mesenchyme at E18.5 in Bmp7 knockout kidneys.

(A) Pregnant mothers bearing Bmp7^LacZ/fl;Gt(ROSA)26Sor^CreERT2 and Bmp7^+/fl;Gt(ROSA)26Sor^CreERT2 embryos were administered tamoxifen at E12.5, and sacrificed at E18.5. (B) The expression of Bmp7 mRNA was significantly reduced in knockout kidneys (n = 4, white column) compared to control kidneys (n = 3, black column). Data are represented as mean ± SD. (C and D) Bmp7 knockout embryos at E18.5 exhibited small kidneys. (E and F) The number of cap mesenchymal cells, as shown by the immunostaining of WT1, was decreased in Bmp7 knockout kidneys. (G) The number of phospho-Histone H3-positive cells was reduced in Bmp7 knockout kidneys. pHistone H3 denotes phospho-Histone H3. (H) TUNEL-positive mesenchymal cells were increased in Bmp7 knockout kidneys. (I) The number of TUNEL-positive mesenchymal cells per section was increased in knockout kidneys (white column) compared to control kidneys (black column). Data are represented as mean ± SD. The average numbers of TUNEL-positive mesenchymal cells in three sections were used as the values for each embryo. The mean of the values from three embryos is presented in the graph. (J) Inappropriately “mature” glomeruli were observed in Bmp7 knockout kidneys. (K) Glomeruli in the mature stages were increased in Bmp7 knockout kidneys (white column) compared to the control kidneys (black column). Criteria for the assessment of glomerular maturity are detailed in method section. Because stage I is the most common in control kidneys at E18.5, we summed up the glomeruli of stages II and III for comparison of maturation. Glomeruli were counted in 3 slices of each kidney. The mean of the values ± SD is presented in the graphs (n = 4 for control embryos, and 6 for knockout embryos). (L) Kidneys were stained with LTA (green) and WT1 (red) to label proximal tubules and glomeruli, respectively. The volume of LTA-positive proximal tubule sections in knockout kidneys was comparable to the control kidneys, whereas the number of glomeruli was significantly reduced. (M) The number of LTA⁺ proximal tubule cross sections normalized by the number of WT1⁺ glomeruli was increased in knockout kidneys (white column) compared to control kidneys (black column). Data are represented as mean ± SD. At least three sections were stained for each kidney. The sum of the number of LTA-positive proximal tubule cross sections was divided by the sum of the number of WT1-positive glomeruli. The mean of the values from four embryos is presented in the graph. Scale bars: 100 µm (D, G, L) or 50 µm (E, F, H, J).

Figure 3. Bmp7 inhibits the differentiation of cap mesenchyme in kidney explant culture.

Kidney explants were taken from Bmp7^LacZ/fl;Gt(ROSA)26Sor^CreERT2 embryos at E12.5 and cultured for 72 h in the presence or absence of 4-OHT. The expression of Bmp7 mRNA was significantly reduced in Bmp7^LacZ/fl;Gt(ROSA)26Sor^CreERT2 explants treated with 4-OHT compared to the explants from the same embryos treated with a vehicle (n = 3). Data are represented as mean ± SD. Whole explants were costained with pan-cytokeratin (green) to label ureteric buds and Jagged1 (red) to label developing nephrons. The Jagged1-positive red area was measured utilizing Photoshop software. In Bmp7^LacZ/fl;Gt(ROSA)26Sor^CreERT2 explants treated with 4-OHT, Jagged1-positive regions were significantly expanded, indicating accelerated differentiation. Scale bars: 100 µm.

Figure 4. Bmp7 inhibits the epithelialization of cap mesenchyme in colony-forming assay.

(A) Administration of Bmp7 dose-dependently inhibited the formation of sheetlike colonies of cap mesenchyme (n = 3). Data are represented as mean ± SD. Scale bars: 100 µm. The expression of E-cadherin, but not vimentin, in these colonies was confirmed by immunostaining. Because the colonies were formed on the feeder cells, some DAPI-positive nuclei were observed outside of the colony. (B) Simultaneous administration of Noggin (300 ng/ml) reversed the inhibitory effect of Bmp7 (10 ng/ml) on colony formation (n = 3). Data are represented as mean ± SD. *: P<0.001. Scale bars: 100 µm.