Differentiated kidney epithelial cells repair injured proximal tubule

Abstract

Whether kidney proximal tubule harbors a scattered population of epithelial stem cells is a major unsolved question. Lineage-tracing studies, histologic characterization, and ex vivo functional analysis results conflict. To address this controversy, we analyzed the lineage and clonal behavior of fully differentiated proximal tubule epithelial cells after injury. A CreER(T2) cassette was knocked into the sodium-dependent inorganic phosphate transporter SLC34a1 locus, which is expressed only in differentiated proximal tubule. Tamoxifen-dependent recombination was absolutely specific to proximal tubule. Clonal analysis after injury and repair showed that the bulk of labeled cells proliferate after injury with increased clone size after severe compared with mild injury. Injury to labeled proximal tubule epithelia induced expression of CD24, CD133, vimentin, and kidney-injury molecule-1, markers of putative epithelial stem cells in the human kidney. Similar results were observed in cultured proximal tubules, in which labeled clones proliferated and expressed dedifferentiation and injury markers. When mice with completely labeled kidneys were subject to injury and repair there was no dilution of fate marker despite substantial proliferation, indicating that unlabeled progenitors do not contribute to kidney repair. During nephrogenesis and early kidney growth, single proximal tubule clones expanded, suggesting that differentiated cells also contribute to tubule elongation. These findings provide no evidence for an intratubular stem-cell population, but rather indicate that terminally differentiated epithelia reexpress apparent stem-cell markers during injury-induced dedifferentiation and repair.

Fig. 1.

Validation of recombination specificity and efficiency in SLC34a1^GCE mouse. (A) Low-power pictures of kidney from SLC34a1^GCE/+; R26R^tdTomato/+ mice given either 3 mg of tamoxifen five times every other day or vehicle reveals diffuse recombination throughout the cortex, but no other kidney region. (Scale bars, 400 μm.) (B–D). Recombined, tdTomato-positive cells uniformly costained with the proximal tubule marker LTL. There was no recombination in DBA-positive collecting duct or nephrin-positive podocytes. (Scale bars, 15 μm in B and D and 30 μm in C.) (E and F). A single dose of 3 mg of tamoxifen induced partial recombination in cortical S1/S2 and S3 segments. (Scale bars, 15 μm.) (G) Tamoxifen dose–response in S1/S2 segment compared with cortical or outer medullary S3 segment. Recombination was nearly complete in S1/S2 segments. There is reduced labeling efficiency in cortical S3 segment, and almost no labeling in outer medullary S3 segment. Average ± SE.

Fig. 2.

Clones of differentiated proximal tubule cells expand after injury and repair. (A) SLC34a1^GCE/+; R26R^tdTomato/+ were administered submaximal tamoxifen and subjected to either mild (26-min) or severe (35-min) IRI injury. Fourteen days later, there was expansion of tdTomato-positive clones in the inner cortex of mildly injured kidney and in both the inner and outer cortex of severely injured kidney. (B) Clone size and frequency was quantitated in the outer cortex vs. the inner cortex. In both mild and severe injury, there was a reduction in the fraction of single-cell clones and a substantial increase in the number of multicellular clones. Severe injury caused a greater number clones with three, four, or five or more cells in both mild and severe injury. (C) Average clone size before and after injury increased in the inner cortex of mildly injured kidneys and in both the inner and outer cortex of severely injured kidney. Average ± SE. (Scale bars, 15 μm.)

Fig. 3.

Fully differentiated cells accomplish proximal tubular repair after acute injury. (A) Experimental scheme. SLC34a1^GCE; R26tdTomato mice were placed on a low-phosphorus diet (0.06%) 5 d before tamoxifen administration. After they returned to a normal diet, severe IRI was performed at day 0, with BrdU administration daily for 7 d, after which the mice were killed. (B–D) Proximal tubules were labeled efficiently, and only a few cells incorporated BrdU in uninjured CLK. In IRI kidneys, there was substantial BrdU incorporation in tdTomato+ proximal tubular cells without dilution of label upon quantitation. (E) Demonstration of the gating strategy for the FACS of tdTomato-positive cells. Kidney cell suspensions of mice not injected with tamoxifen served as a negative control (Right). (F–J) Quantitative PCR of tdTomato-positive cells showed that CD133 (F), CD24 (G), vimentin (H), and KIM-1 (I) mRNA levels were substantially increased in tdTomato-positive cells in injured kidney. In addition, SLC34a1 mRNA level decreased in injured kidney (J), indicating injured, fully differentiated tubular epithelial cells lose a differentiated phenotype after injury. Bone marrow cells serve as a positive control for stem cell markers. Average ± SE, *P < 0.01, **P < 0.05. (Scale bars, 15 μm.)

Fig. 4.

Fully differentiated proximal tubular epithelia dedifferentiate and proliferate during injury and repair. (A–C) Labeled clones did not stain for Ki67 before injury; however, Ki67+ clones could be detected on day 2 after IRI, but by day 14 labeled cells no longer expressed Ki67. Similar patterns of transient expression of vimentin (D and E), Pax2 (G–I), and KIM-1 (J–L) were observed in tdTomato+ cells. (Scale bars, 15 μm.)

Fig. 5.

Differentiated proximal tubular epithelia contribute to tubule elongation during nephrogenesis and kidney growth. (A) A section of P0 kidney from pups that were exposed to 0.1 mg of tamoxifen at 14.5 dpc. There were rare, scattered single labeled cells in tubules. (Scale bar, 100 μm.) (B and C) The distance to the nearest labeled cell was measured for 93 clones at P0. The average distance between clones was 713.2 + 54.5 μm. (D) A fraction of tdTomato+ clones proliferated during development. (Scale bar, 30 μm.) (E) Clone size-frequency analysis at P0, P14, and P42. Clone size increased from P0 to P14, but there was no further clonal proliferation at P42.