Sox9-Positive Progenitor Cells Play a Key Role in Renal Tubule Epithelial Regeneration in Mice

Abstract

The kidney has a tremendous capacity to regenerate following injury, but factors that govern this response are still largely unknown. We isolated cells from mouse kidneys with high proliferative and multi-lineage differentiation capacity. These cells expressed a high level of Sox9. In regenerating kidneys, Sox9 expression was induced early, and 89% of proliferating cells were Sox9 positive. In vitro, Sox9-positive cells showed unlimited proliferation and multi-lineage differentiation capacity. Using an inducible Sox9 Cre line and lineage-tagging methods, we show that Sox9-positive cells can generate new daughter cells, contributing to the regeneration of proximal tubule, loop of Henle, and distal tubule segments but not to collecting duct and glomerular cells. Furthermore, inducible deletion of Sox9 resulted in reduced epithelial proliferation, more severe injury, and fibrosis development. In summary, we demonstrate that, in the kidney, Sox9-positive cells show progenitor-like properties in vitro and contribute to epithelial regeneration following injury in vivo.

Figure 1. Isolation of cells with progenitor properties from mouse kidneys

A, Schematic diagram of isolating cells with high proliferative capacity and clonogenicity from mouse kidneys. B, Morphology of tubular epithelial cells during ex vivo expansion. C, Cumulative doubling numbers for six different cell lines throughout in vitro expansion. D, Relative mRNA levels of progenitor cell markers in whole kidneys and cultured tubule epithelial cells (Sox9; sex determining region Y-box 9, Lgr4; Leucine-Rich Repeat Containing G Protein-Coupled Receptor 4, Lgr5; Leucine-Rich Repeat Containing G Protein-Coupled Receptor 5, Pax2; paired box 2, Pax8; paired box 8, Wt1; Wilms tumors 1). E, Relative transcript levels of mesenchymal stem cell markers (Scf; stem cell factor). All data are presented as means as ± S.E.M and analyzed by t-test, * represent p<0.05. F, Mouse model to follow progeny of Sox9 positive cells. G, Representative images of kidney sections of non-inducible Sox9^IRESCre-Gt(ROSA)26Sor^{tm4(ACTB-tdTomato,-EGFP)} mice on day 0 and day 42 after birth. H, Quantification of GFP positive cells in non-inducible Sox9^IRESCre-Gt(ROSA)26Sor^{tm4(ACTB-tdTomato,-EGFP)}. I, Double immunostaining of GFP and LTL; marker of proximal tubule, Calbindin; marker of distal tubule, and Aquaporin 2; marker of collecting duct in non-inducible Sox9^IRESCre-Gt(ROSA)26Sor^{tm4(ACTB-tdTomato,-EGFP)} at day 6 weeks of age. Scale bar, 20 μm.

Figure 2. Sox9 expression reemerges following acute kidney injury

A-E, Relative transcript levels of Sox9, Cd133, Foxd1, Lgr4 and Pax8 in kidneys of mice injected with folic acid (FAN) or sham 2 hr to 7 days following the injection. All data are presented as means as ± S.E.M and analyzed by t-test, * represent p<0.05 compare to control. F, Representative images of Sox9 immunostaining of mouse kidney samples. Animals were injected with sham (CTL), or with 250mg/kg folic acid and sacrificed 7 days later. G, Double immunostaining of Sox9 (red) and (green) LTL (a marker of proximal tubule), PNA (marker of loop of henle/distal tubule), and DBA (a marker of collecting duct) 7 days after folic acid injection. Scale bar, 20 μm.

Figure 3. Sox9 expressing cells proliferate and expand following injury

A, Quantitative RT-PCR analysis of transcripts of cycline family genes (Ccna2; cycline a2, Ccnb1; cycline b1, Ccnd2; cycline d2, Ccne1; cycline e1) in control and FAN mice. All data are presented as means as ± S.E.M and analyzed by t-test, * represent p<0.05 compare to control. B, Percent of Ki67 positive cells in Sox9 expressing cells (left) and Sox9 positive cells in proliferating Ki67 expressing cells (right). C, Representative double immunofluorescence staining images of Sox9 (green) and Ki67 (red) in control and FAN injected mice. D, Representative images of GFP (red), Ki67 (green) and DAPI (blue) in inducible Sox9 Cre^ERT2-Gt(ROSA)26Sor^{tm4(ACTB-tdTomato,-EGFP)} mice injected with sham or folic acid. Scale bar, 20 μm. E, Percent Ki67 positive cells in GFP labeled cells in control and FA injected kidneys.

Figure 4. Sox9 positive cells expand after injury

A, Generation of tamoxifen-inducible Sox9 Cre^ERT2-Gt(ROSA)26Sor^{tm4(ACTB-tdTomato,-EGFP)} mice for genetic lineage tracing. B, Quantification of GFP expressing cells (of Sox9 progeny) 1, 4, 12 or 20 weeks after sham or folic acid injection. C, Representative images of kidney sections of sham or folate injected mice. All data are presented as means and ± S.E.M and analyzed by t-test, * represent p<0.05 compare to CTL. Scale bar, 20 μm.

Figure 5. Sox9 positive cells contribute to regeneration of multiple tubule segments

Double immunostaining images of kidney section of GFP (Sox9 linage marker) and tubule segment markers in tamoxifen-inducible Sox9 Cre^ERT2-Gt(ROSA)26Sor^{tm4(ACTB-tdTomato,-EGFP)} mice injected with sham or folate 2 weeks after injury. (A) LTL; proximal tubule marker, (B), PNA; loop of Henle/distal tubule, (C) calbindin; distal tubule marker (D) Aquaporin 2 (AQP2) was used for double labeling. Scale bar, 20 μm.

Figure 6. Sox9 positive cells show high proliferative and multilineage differentiation potential in vitro

A-B, Quantitative RT-PCR analysis of transcripts of Notch pathway related gene (A), and Wnt pathway related genes (B), in Sox9 negative and positive fractions derived from control and FAN mice. C, Representative bright field microscopy images and crystal violet staining and CFU (colony forming unit) of Sox9 positive and negative cells after 3^rd passage. D-E, Adipogenic differentiation of Sox9 positive cells; representative oil red O (D) staining and transcript levels of adipocyte markers (E). F-G, Osteogenic differentiation of Sox9 positive cells; representative images of von Kossa staining (F) and transcript levels of osteoblast markers (G). H-I, Chondrogenic differentiation of Sox9 positive cells; representative images of Alcian blue staining (H) and transcript levels of chondrocyte markers (I). J, E-cadherin, aquaporin 1, n-cadherin, calbindin, and aquaporin 2 immunofluorescence images of Sox9 positive cells cultured in tubular epithelial cell differentiation medium for 3 weeks. Scale bar, 20 μm. K, Relative transcript levels of cadherins in control and differentiated cells. All data are presented as means and ± S.E.M and analyzed by t-test, * represent p<0.05 compare to CTL.

Figure 7. Loss of Sox9 exacerbates tubulointerstitial fibrosis following injury

A, Representative double immunofluorescence staining images of Sox9 (red) and Pax8 (green) in control mouse kidney sections. B, Representative images of PAS and Sox9 immunostaining of kidney sections mice injected with sham or folic acid, 7 days after injury. We used WT and single transgenic animals as control (CTL and CTL FA), and Pax8rtTA-TetOcre-Sox9 f/f transgenic mice treated with folic acid (Sox9f/f FA). C-D, Relative transcript levels of fibrosis markers (C) and proliferation associated markers (D) in control (CTL), CTL with folic acid treated (CTL, FA) and Pax8 rtTA-TetO cre-Sox9 f/f with folic acid injected (Sox9 f/f FA) mice. E, Serum creatinine levels of CTL, CTL FA, and Pax8 rtTA-TetO cre-Sox9 f/f with folic acid injected (Sox9 f/f FA) mice. F, Representative PAS stained images of kidney sections of sham injected and folate injected animals 12 weeks after injury. Scale bar, 20 μm. G-H, Relative transcript levels of fibrosis markers (G) and proliferation associated markers (H) in control (CTL), CTL with folic acid treated (CTL FA) and Pax8 rtTA-TetO cre-Sox9 f/f with folic acid injected (Sox9 f/f FA) mice. I, Serum creatinine levels in CTL, CTL FA, and Pax8 rtTA-TetO cre-Sox9 f/f FA mice. All data are presented as means and ± S.E.M and analyzed by t-test, * represent p<0.05 compare to CTL, + represent p<0.05 compare to CTL FA.