M.L.A., E.R., L.B., B.M., A.P., C.S., E.P., M.R., and E.L.: collection and/or assembly of data, data analysis and interpretation, and final approval of manuscript; M.G. and M.C.: provision of study material or patients and final approval of manuscript; A.B.F.: provision of study material or patients, manuscript writing, final approval of manuscript; L.L.: conception and design, manuscript writing, and final approval of manuscript; P.R.: conception and design, manuscript writing, final approval of manuscript, and financial support.
Disclosure of potential conflicts of interest is found at the end of this article.
First published online in STEM CELLSEXPRESS May 24, 2012
Recent studies implicated the existence in adult human kidney of a population of renal progenitors with the potential to regenerate glomerular as well as tubular epithelial cells and characterized by coexpression of surface markers CD133 and CD24. Here, we demonstrate that CD133+CD24+ renal progenitors can be distinguished in distinct subpopulations from normal human kidneys based on the surface expression of vascular cell adhesion molecule 1, also known as CD106. CD133+CD24+CD106+ cells were localized at the urinary pole of Bowman's capsule, while a distinct population of scattered CD133+CD24+CD106− cells was localized in the proximal tubule as well as in the distal convoluted tubule. CD133+CD24+CD106+ cells exhibited a high proliferative rate and could differentiate toward the podocyte as well as the tubular lineage. By contrast, CD133+CD24+CD106− cells showed a lower proliferative capacity and displayed a committed phenotype toward the tubular lineage. Both CD133+CD24+CD106+ and CD133+CD24+CD106− cells showed higher resistance to injurious agents in comparison to all other differentiated cells of the kidney. Once injected in SCID mice affected by acute tubular injury, both of these populations displayed the capacity to engraft within the kidney, generate novel tubular cells, and improve renal function. These properties were not shared by other tubular cells of the adult kidney. Finally, CD133+CD24+CD106− cells proliferated upon tubular injury, becoming the predominating part of the regenerating epithelium in patients with acute or chronic tubular damage. These data suggest that CD133+CD24+CD106− cells represent tubular-committed progenitors that display resistance to apoptotic stimuli and exert regenerative potential for injured tubular tissue. STEM CELLS2012;30:1714–1725
Following acute tubular injury, the kidney undergoes a regenerative response leading in most cases to recovery of renal function [1–3]. The cell source for regenerating cells is poorly understood, with the strongest evidence supporting a role for less injured tubular cells, which proliferate and migrate to replace the neighboring dead cells, and eventually reline denuded tubules restoring the structural and functional integrity of the kidney [1, 2]. Consistently, genetic tagging of renal epithelium in mouse models of acute kidney injury (AKI) indicated that repair of damaged nephrons is predominantly accomplished by intrinsic, surviving cells localized within the nephron epithelium . However, whether these regenerating cells derive from differentiated tubular cells or from a subset of renal stem/progenitor cells localized within the nephron remains unknown. In other adult organs, resident stem cell compartments critically participate in regeneration [5, 6], and several studies support the existence of a stem/progenitor cell compartment also in adult mammalian kidney [7–17]. Renal progenitors were first identified in humans, where they are characterized by coexpression of two markers, CD133 and CD24 [7–14]. CD133+CD24+ human renal progenitors are localized at the urinary pole of Bowman's capsule [7–14] and are capable of podocyte and tubular differentiation [7–10, 17]. However, the tubular portion of the nephron is very long, which suggests that the stem cell compartment localized at the urinary pole of Bowman's capsule may not be unique and that other progenitors may localize along the tubular part of the nephron. Consistently, a recent study reported the existence of scattered CD133+CD24+ cells distributed within the proximal tubules in adult human kidneys and suggested they may represent a subset of progenitors destined to tubular regeneration . However, a phenotypic and functional characterization of tubular CD133+CD24+ cells was not reported, because specific markers allowing distinction from the CD133+CD24+ cells of Bowman's capsule were unknown.
In this study, we demonstrate that CD133+CD24+ renal progenitors can be distinguished in functionally distinct subpopulations in normal human kidneys based on the surface expression of vascular cell adhesion molecule 1 (VCAM1), also known as CD106. CD133+CD24+CD106+ cells were localized at the urinary pole of Bowman's capsule, while a distinct population of CD133+CD24+CD106− scattered cells was localized in the proximal tubule as well as in the distal convoluted tubule. CD133+CD24+CD106+ cells exhibited a high proliferative rate and could differentiate toward the podocyte as well as the tubular lineage. By contrast, CD133+CD24+CD106− cells showed a lower proliferative capacity and displayed a committed phenotype toward the tubular lineage. Interestingly, both CD133+CD24+CD106+ and CD133+CD24+CD106− cells showed higher resistance to injurious agents in comparison to all other differentiated cells of the kidney. Once injected in severe combined immunodeficient (SCID) mice affected by acute tubular injury, both of these populations displayed the capacity to engraft within the kidney, generated novel tubular cells, and significantly improved renal function, a property that was not shared by other tubular cell types of the adult kidney. Finally, CD133+CD24+CD106− cells proliferated upon tubular injury and became the predominating part of the regenerating epithelium in patients with acute or chronic tubular damage, further substantiating that they represent tubular-committed progenitors.
MATERIALS AND METHODS
Overall Description of the Study Protocol
(a) Identification of markers of different kidney progenitor subpopulations by mRNA microarray; (2) validation of the selected marker with quantitative reverse transcriptase-PCR (RT-PCR) and Western blot analysis; (c) in vivo analysis of the distribution of the progenitor subpopulations within healthy adult human kidneys by confocal microscopy; (d) recovery and culture of purified progenitor subpopulations; (e) in vitro testing of the growth and differentiation capacity of the progenitor subpopulations; (f) in vivo testing of the growth and differentiation capacity of the progenitor subpopulations in murine models of acute tubular injury and focal segmental glomerulosclerosis (FSGS); (g) analysis of distribution of progenitor subpopulations in biopsies of patients affected by acute and chronic tubular damage by confocal microscopy.
Antibodies and Reagents
The following antibodies were used: mAb anti-CD24 (clone SN3), mAb anti-Vimentin (V9), pAb antinephrin (C17), pAb anti-aquaporin-2 (AQP2) (C-17), pAb anti-aquaporin-1 (AQP1) (H-55), pAb antimegalin, pAb anti-chloride channel-A (CLC-KA, clone K-16), pAb anti-thiazide-sensitive Na/Cl cotransporter (NCCT, clone N-19) (all from Santa Cruz Biotechnology, Santa Cruz, CA, www.scbt.com), mAb anti-CD133/2 (clone 293C3, Miltenyi Biotec GmbH, Bergish Gladbach, Germany, www. miltenyibiotec.com), mAb anti-CD106 (clone 1.4C1, Sigma-Aldrich, St. Louis, MO, www.sigmaaldrich.com), mAb anti-podocalyxin (PDX) (clone 222328) (R&D Systems, Minneapolis, MN, www.rndsystems.com), mAb anti-cytokeratin 7 (CK7) (clone OV-TL 12/30, Dako, Carpinteria, CA, www.dako.com), mAb anti-CD13 (22A5, Abcam, Cambridge, U.K., www.abcam.com), mAb anti-Histone-H3 (phospho S10) (Abcam), mAb antisynaptopodin (G1D4, Progen, Heidelberg, Germany, www.progen.de), mAb anti-epithelial membrane antigen (EMA) (clone E29, Dako), pAb anti-Tamm-Horsfall glycoprotein (THP) (MP Biomedicals, Verona, Italy, www.mpbio.com), phycoerythrin (PE)-conjugated anti-CD106 (IE10, R&D Systems), unlabeled anti-CD106 (1.G1B1, Southern Biotech, Birmingham, AL, www.southernbiotech.com), allophycocyanin (APC)-conjugated anti-CD133/2 (clone 293C3), anti-CD45 Microbeads (mouse IgG2a), and anti-mouse IgG1 microbeads (all purchased from Miltenyi). Secondary antibodies Alexa Fluor 488-labeled goat anti-mouse IgG2b, Alexa Fluor 488/633/546-labeled goat anti-mouse IgG1, Alexa Fluor 488/546/633-labeled goat anti-mouse IgG2a, Alexa Fluor 488/546-labeled rabbit anti-goat IgG, and Alexa Fluor 488/546/633-labeled goat anti-rabbit IgG were all from Molecular Probes (Eugene, OR, www.invitrogen.com). As control isotypes, the following antibodies were used: anti-IgG2a (clone HOPC-1, Southern Biotech), anti-IgG1 (clone MOPC-21, BD Biosciences, San Diego, CA, www.bdbiosciences.com), APC-labeled anti-IgG2b (clone IS6-11E5.11, Miltenyi) and PE-labeled anti-IgG2a (clone G155-178, BD Biosciences). FITC-labeled Lotus tetragonolobus agglutinin (LTA) and FITC-labeled Dolichos Biflorus Agglutinin (DBA) were from Vector Laboratories (Burlingame, CA, www.vectorlabs.com) To-pro-3 was from Molecular Probes.
For the isolation of CD133+CD24+ cells of glomerular and tubular origin, human kidney tissue samples were minced and recovered by a standard sieving technique through graded mesh screens (60, 80, 150, and 200 mesh). The glomerular suspension was collected from the 150 mesh and cultured in endothelial cell growth medium-microVascular (EGM-MV) (Lonza Sales Ltd., Basel, Switzerland, www.lonza.com) supplemented with 20% fetal bovine serum (FBS) (Thermoscientific Hyclone, Logan, UT, www.bioresearchonline.com) as described . The tubular suspension was recovered from the 200 mesh and digested with collagenase IV 500 U/ml in order to obtain a single-cell suspension and to recover the CD133+ fraction by immunomagnetic separation using CD133 Cell Isolation Kit (Miltenyi). The CD133+CD24+CD106− subpopulation was obtained from CD133+CD24+ tubular cells by immunomagnetic depletion for CD106 using LD columns (Miltenyi). CD133−CD24− cells were obtained as previously described . Purity of all obtained cell fractions was higher than 98% as assessed by flow cytometry. Clones were generated from CD133+CD24+CD106+ and CD133+CD24+CD106− cells by limiting dilution in 96-well plates in EGM-MV 20% FBS and renal epithelial cell growth medium (REGM) 10% FBS, respectively. Tubulogenic and podocyte differentiations were obtained as previously described [7, 17]. Assessment of cell viability following treatment with hemoglobin (0.1-0.5-1 mg/ml) was performed using an MTT (3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl tetrazolium bromide) assay (Promega, Madison, WI, www.promega.com).
TaqMan Low Density Array and Real Time Quantitative RT-PCR
Total RNA was extracted from CD133+CD24+ cells of glomerular and tubular origin (n = 5 for each cell type) using an RNeasy Microkit (Qiagen, Hilden, Germany, www.qiagen.com). Predesigned TaqMan low density array (TLDA) for human apoptosis, inflammation, and angiogenesis were obtained from Applied Biosystems (Warrington, U.K., www.appliedbiosystems.com). Relative quantities (RQ) were determined with the equation: RQ = 2−ΔCt where ΔCt = (Ct gene of interest − Ct internal control, glyceraldehyde-3-phosphate dehydrogenase (GAPDH)) and the fold change in gene expression between CD133+CD24+ cells of glomerular and tubular origin was calculated as a ratio: 2−ΔCt of glomerular cells/2−ΔCt of tubular cells. TaqMan RT-PCR was performed as described  using a 7900HT Real Time PCR System (Applied Biosystems). CD106, amino acid transporters, Na/K/Cl transporter, meprin1 subunit B (MEP1B), Na/Cl transporter, Na/H exchanger, γ-glutamyltransferase (γ-GT), nephrin, WT-1, podocin, and PDX quantification was performed using Assay on Demand (Applied Biosystems) .
A total of 13 normal adult human kidneys were analyzed in this study. Normal kidney fragments were obtained from the pole opposite to the tumor of 10 patients who had localized renal tumors and underwent nephrectomy, in agreement with the Ethical Committee on human experimentation of the Azienda Ospedaliero-Universitaria Careggi, Florence, Italy as well as from three healthy cadaveric donors in accordance with Institutional Review Board approval at Vanderbilt University. In addition, bioptic tissue specimens were obtained from four patients affected by acute tubular necrosis as well as from five patients with chronic tubulointerstitial damage related to FSGS with nephrotic syndrome and progressive renal failure.
Confocal microscopy was performed on 5-μm sections of renal frozen tissues or on cells that were cultured on chamber slides as described [7, 17] using an LSM 510 META laser confocal microscope (Carl Zeiss, Jena, Germany, www.zeiss.it).
Flow Cytometry was performed as described [7, 17] and as detailed in Supporting Information methods.
Xenograft in SCID Mouse Model of Acute Renal Failure and FSGS
Rhabdomyolysis-induced acute renal failure was studied in 6-week-old female SCID mice (Harlan, S. Pietro al Natisone, Italy, www.harlan.com), as described previously , by intramuscular injection on day 0 with hypertonic glycerol (8 ml/kg b.wt. of a 50% glycerol solution; Sigma-Aldrich) into the inferior hind limbs. Animal experiments were performed in accordance with institutional, regional, and state guidelines and in adherence with the National Institutes of Health Guide for the Care and Use of Laboratory Animals. On day 0, four groups of mice received an i.v. injection into the tail vein as follows: group 1 (n = 12), saline; group 2 (n = 12), PKH26-stained CD133+CD24+CD106+ cells (0.75 × 106 cells 4 hours after glycerol injection and 0.75 × 106 cells 20 hours after glycerol injection); group 3 (n = 12), PKH26-stained CD133+CD24+CD106− cells (0.75 × 106 cells 4 hours after glycerol injection and 0.75 × 106 cells 20 hours after glycerol injection); group 4 (n = 12), PKH26-stained CD133−CD24− cells (0.75 × 106 cells 4 hours after glycerol injection and 0.75 × 106 cells 20 hours after glycerol injection). The cells, obtained from five different human donors (two men and three women) were labeled with the PKH26 Red Fluorescence Cell Linker kit (Sigma-Aldrich) immediately before injection. Blood samples were obtained from the submandibular venous sinus at day 3, 5, 8, 11, or 15, and blood urea nitrogen (BUN) levels were measured by Reflotron System (Roche Diagnostics, Rotkreuz, Switzerland, www.harlan.com). Mice were sacrificed at day 15 and kidneys removed for morphologic assessment. Adriamycin nephropathy was induced in female SCID mice as detailed in Supporting Information methods.
Evaluation of Renal Fibrosis
Five-micrometer-thick frozen tissue sections of mouse kidney were stained with a Masson-Goldner trichromic kit (Bio-Optica, Milan, Italy, www.bio-optica.it). Hue values for bluish-green were evaluated in all cases by two independent pathologists. Nonoverlapping fields of the entire section (20 fields for each mouse) were independently analyzed using a ×20 objective. The severity of renal scarring was assessed as follows: normal tubulointerstitium scored 0; mild tubular atrophy with interstitial edema or fibrosis affecting up to 25% of the field scored 1; moderate tubular atrophy with interstitial edema or fibrosis affecting 25%–50% of the field scored 2; and severe tubular atrophy with interstitial edema or fibrosis affecting >50% of the field scored 3.
Statistical analysis was performed using SPSS software (SPSS, Inc., Evanston, IL). The expression levels of several genes were compared between glomerular and tubular cells by Mann-Whitney U test according to a nonparametric distribution. Comparison between BUN levels at different time points in treated and untreated mice was performed by analysis of variance (ANOVA) test for multiple comparisons and Mann-Whitney U test. A p-value <.05 was considered statistically significant.
CD133+CD24+ Cells Obtained from Human Glomerular or Tubular Tissue Share a Similar But Not Identical Phenotype
To perform phenotypic and functional studies separately on CD133+CD24+ cells of Bowman's capsule as well as on CD133+CD24+ tubular cells, we sought to identify a surface marker that allowed to distinguish these two populations. To this aim, CD133+CD24+ cells were prepared from outgrowth of capsulated glomeruli or from the tubular portion of renal tissue as described in Materials and Methods and screened for mRNA expression of 279 genes, using TLDA technology. More than 95% of the investigated genes showed similar expression in CD133+CD24+ cells obtained from glomerular outgrowths compared with CD133+CD24+ cells obtained from tubular tissue. However, some genes were differentially expressed, and among these genes, the expression of VCAM-1, also known as CD106, was approximately 300-fold higher in CD133+CD24+ cells obtained from glomerular outgrowths than in CD133+CD24+ cells obtained from tubular tissue. The expression ratios (glomerular/tubular) of the genes for which a significant difference was found between the two cell cultures are shown in Table 1. A strong difference in CD106 levels among CD133+CD24+ cells obtained from glomerular outgrowth and CD133+CD24+ cells obtained from tubular tissue was also confirmed at the protein level by Western blot analysis (Supporting Information Fig. S1).
Table 1. Low-density array analysis of CD133+CD24+ cells obtained from normal human glomerular or tubular tissue
Comparison of relative levels of 279 mRNAs in CD133+CD24+ cells obtained from glomerular tissue versus CD133+CD24+ cells obtained from tubular tissue was performed. Only differentially expressed genes (*p < .05) are shown.
CD133+CD24+ Tubular Cells Are Distinguished from CD133+CD24+ Cells of Bowman's Capsule by CD106 Expression and Localize in Specific Segments of the Tubule
We then analyzed the expression of CD106 on the surface of total renal cells in freshly prepared adult human kidneys, where CD133+CD24+ renal progenitors represented 0.5%–4% of total renal cells (Fig. 1A). Since CD133+ cells represent a subset of CD24+ cells , CD133 was then used in all further combined labeling to identify renal progenitors. We thus performed triple-labeling flow cytometry for CD133, CD106, as well as for the podocyte marker PDX, which specifically stains the previously described CD133+ podocyte progenitors, gated on CD133+CD106− cells, and evaluated them for PDX expression. Interestingly, CD133+CD106− cells never expressed PDX, thus suggesting that podocyte progenitors are CD106+ (Fig. 1B). This allowed us to establish the existence of three different renal progenitor subpopulations, CD133+CD106+PDX−, CD133+CD106+PDX+, and CD133+CD106−PDX− (Fig. 1C). Then, to verify whether CD106 protein expression could distinguish between CD133+CD24+ cells of Bowman's capsule and those scattered along the tubular compartment in vivo, we performed triple-label immunofluorescence for CD133, CD106, and the proximal tubule marker CD13, which demonstrated that expression of CD106 was a selective property of CD133+CD24+ cells of Bowman's capsule (Fig. 1D, G glomerulus), while CD133+CD24+ tubular cells did not express CD106 (Fig. 1D, arrows). Identical results were obtained in healthy human kidneys obtained from nephrectomies as well as in those obtained from cadaveric donors. Consistent with previous observations, CD133+CD24+ cells of Bowman's capsule as well as CD133+CD24+ tubular cells shared the coexpression of a series of other markers such as CK7 and vimentin, which were reported to selectively costain CD133+CD24+ cells in adult human kidneys  (Fig. 1E, 1F, arrows point triple-labeled cells). To better understand the localization of CD133+CD24+CD106− cells within the tubular compartment, combined immunofluorescence for CD133 and markers of different portions of the nephron were performed. The great majority of CD133+ cells were costained with AQP1, which stains proximal tubular cells as well as the thin descending limb of the loop of Henle (Fig. 1G). Interestingly, all CD133+CD24+CD106− cells that expressed AQP1 also displayed the property to bind LTA, thus suggesting that they localized within the proximal tubule and not in the thin descending limb (data not shown). In addition, CD133+CD24+CD106− cells were not observed either in the thin or in the thick ascending limb of the loop of Henle, since the Kidney-specific chloride channel-A (CLC-KA, a marker of the thin ascending limb, Fig. 1H), or the THP (a marker of the thick ascending limb, Fig. 1I), was never colocalized with CD133. However, CD133+CD24+ tubular cells were localized in the distal convoluted tubule and the connecting segment, as demonstrated by costaining of CD133 with the thiazide-sensitive NCCT (Fig. 1J). Interestingly, in the distal nephron, CD133+CD24+CD106− cells were localized as a small cluster of cells in the portion of the distal convoluted tubule that takes contact with the vascular pole of the glomerulus. In addition, double-label immunofluorescence for CD133 and AQP2 demonstrated that CD133+CD24+CD106− cells were not localized in the collecting ducts (Fig. 1K).
CD133+CD24+CD106+ and CD133+CD24+CD106− Cells Exhibit Distinct Differentiative Potential
We next analyzed CD133+CD24+ cultured cells obtained from glomerular or tubular tissue for CD106 surface expression. Cultures obtained from glomerular tissue usually consisted of 95%–100% of CD133+CD24+CD106+ cells (Fig. 2A and Supporting Information Fig. S3B). By contrast, cultures obtained from the tubular tissue and analyzed after 7 days of culture mostly consisted of CD133+CD24+CD106− cells (Supporting Information Fig. S3A), but when observed at later time, the percentage of CD133+CD24+CD106+ cells within these cultures was increased (Fig. 2B). To evaluate whether the observed increase in the percentage of CD133+CD24+CD106+ cells was related to the different proliferative potential of the two populations, rather than to an upregulation of CD106 in culture, we performed immunomagnetic depletion of CD106+ cells, which allowed us the recovery of CD133+CD24+CD106− cells with a purity >98% (Fig. 2B and Supporting Information Fig. S3C, S3D). Then, CD133+CD24+CD106− cells were stained with the fluorescent label CellTrace Violet, seeded in coculture with unlabeled CD133+CD24+CD106+ cells in a ratio of 1:1, and checked by flow cytometry, that confirmed that the two populations were clearly identifiable based on their different fluorescent labeling (Fig. 2C, top). After 5 days of culture, the distinct fluorescence labeling of the two populations was still clearly distinguishable, but the unlabeled cell population had increased to become the 70% of cultured cells (Fig. 2C, bottom). These results demonstrate that CD133+CD24+CD106+ and CD133+CD24+CD106− cells exhibited diverse growth or survival potential. Consistently, counting cell numbers of pure populations of CD133+CD24+CD106+ or CD133+CD24+CD106− cells over time in culture demonstrated that CD133+CD24+CD106+ cells had a consistently higher proliferative potential than CD133+CD24+CD106− cells (Fig. 2D). However, both of the populations showed a higher proliferative potential than CD133−CD24− cells that, as already described, consisted mostly of differentiated tubular cells  (Fig. 2D). In addition, CD133+CD24+CD106+ and CD133+CD24+CD106− cells showed a comparable resistance to cell death induced by increasing concentrations of hemoglobin, which were however considerably higher in comparison to that of CD133−CD24− cells, as assessed by MTT assay (Fig. 2E). Finally, we tried to obtain clones of the two populations to analyze their differentiative properties. Clones derived from CD133+CD24+CD106+ cells displayed an uncommitted phenotype (Fig. 3A), could be expanded in a medium that allows the growth of undifferentiated renal progenitors (EGM-MV), and differentiated toward the podocyte lineage if cultured in the podocyte-maintaining medium DMEM/F12 supplemented with Vitamin D3 and retinoic acidVRAD  as well as toward the tubular lineage if cultured in the tubular differentiating medium (REGM+HGF) (Fig. 3A′ and 3A″, respectively). Differentiation toward podocytes resulted in the upregulation of nephrin, WT1, podocin, and podocalyxin mRNA as well as protein expression of the podocyte markers nephrin and synaptopodin (Fig. 3A′). Differentiation toward tubular cells resulted in a strong upregulation of markers of different portions of the tubular nephron, such as the Na/K/Cl cotransporter, and the thiazide-sensitive Na/Cl cotransporter, amino acid transporter, MEP1B, the Na/H exchanger, or the γ-GT (Fig. 3A″). In addition, differentiated clones acquired binding properties to LTA and started to express EMA, which are specific properties of proximal and distal tubular epithelia, respectively (Fig. 3A″). On the other hand, CD133+CD24+CD106− cells could not be cloned in EGM-MV, whereas if plated in REGM+HGF medium, they generated clones that showed the phenotype of transition cells coexpressing CD133, CD24 (data not shown) as well as tubular markers at both mRNA and protein levels (Fig. 3B). Interestingly, placing clones of CD133+CD24+CD106+ cells in REGM+HGF medium resulted in a downregulation of the level of mRNA of CD106, and a relevant percentage of the cells (10%–30%) turned out to be CD106−, suggesting that CD133+CD24+CD106− cells may derive from CD133+CD24+CD106+ cells and represent a more committed step toward complete differentiation into tubular cells (Supporting Information Fig. S4).
CD133+CD24+CD106+ as well as CD133+CD24+CD106− Cells, But Not CD133−CD24− Cells, Regenerate Tubular Cells and Improve Renal Function in SCID Mice with acute kidney injury
The ability of CD133+CD24+CD106+ and CD133+CD24+CD106− cells to regenerate injured renal cells was then assessed in a model of rhabdomyolysis-induced AKI in SCID mice, generated by intramuscular injection of glycerol. To this end, CD133+CD24+CD106+, CD133+CD24+CD106− cells, or saline were injected into the tail vein of SCID mice with rhabdomyolysis-induced AKI 4 and 20 hours after glycerol injection. As an additional control, glycerol-treated SCID mice were injected with a mixture of CD133−CD24− renal cells, which represent 90%–95% of the total cells of the kidney. Measurement of BUN levels demonstrated that mice injected with CD133+CD24+CD106+ cells or CD133+CD24+CD106− cells showed a significantly reduced severity of AKI at days 5, 8, 11, and 15 that was not observed in mice treated with saline or with CD133−CD24− renal cells (Fig. 4A). Interestingly, only injection of CD133+CD24+CD106+ cells resulted in a significant reduction of the severity of AKI at day 3, as revealed by the lower BUN levels in comparison to mice treated with CD133+CD24+CD106− cells, saline, or CD133−CD24− renal cells (Fig. 4A). Treatment with CD133+CD24+CD106+ cells or CD133+CD24+CD106− cells was also associated with a better preservation of renal structure and a reduction of renal fibrosis in comparison to mice treated with CD133−CD24− cells or saline at day 15 after induction of AKI, as assessed with Masson's trichrome and scoring for the presence of renal scarring (Fig. 4B–4E). To further investigate their ability to regenerate injured tubular cells, CD133+CD24+CD106+, CD133+CD24+CD106−, or CD133−CD24− renal cells were labeled with the red fluorescent dye PKH26 before their injection into glycerol-treated SCID mice. Labeled CD133+CD24+CD106+ cells engrafted within the tissue and generated novel tubular cells, acquiring the property to bind the proximal tubule-specific marker LTA (Fig. 5A) or the distal nephron marker DBA (Fig. 5D). Identical results were obtained when CD133+CD24+CD106− cells were used (Fig. 5B, 5E). Quantitation of the number of PKH26-labeled cells over the total number of tubular cells demonstrated that in mice injected with CD133+CD24+CD106+ cells, as well as with CD133+CD24+CD106− cells, PKH26-labeled cells costained for LTA or DBA-stained tubular cells could be observed at day 15 after injury (7.81% ± 3.1% CD133+CD24+CD106+ vs. 6.7% ± 2.2% CD133+CD24+CD106− PKH26-labeled/LTA-stained cells, not significant, NS and 4.92% ± 1.8% CD133+CD24+CD106+ vs. 5.23% ± 2.13% CD133+CD24+CD106− PKH26-labeled/DBA-stained cells; NS). By contrast, red labeling was never observed in mice injected with CD133−CD24− renal cells (Fig. 5C, 5F; 7.81% ± 3.1% CD133+CD24+CD106+ vs. 0.08% ± 0.02% CD133−CD24− PKH26-labeled/LTA-stained cells, p < .05; 6.7% ± 2.2% CD133+CD24+CD106− vs. 0.08% ± 0.02% CD133−CD24− PKH26-labeled/LTA-stained cells, p < .05; 4.92% ± 1.8% CD133+CD24+CD106+ vs. 0.03% ± 0.01% CD133−CD24− PKH26-labeled/DBA-stained cells, p < .05; 5.23% ± 2.13% CD133+CD24+CD106− vs. 0.03% ± 0.01% CD133−CD24− PKH26-labeled/DBA-stained cells, p < .05) or with saline solution (data not shown). Furthermore, Y-chromosome was detected in female SCID mice injected with CD133+CD24+CD106+ cells (Fig. 5G) or CD133+CD24+CD106− cells (Fig. 5H) derived from male human subjects, as demonstrated by fluorescent in situ hybridization (FISH). Of note, PKH26-labeled CD133+CD24+CD106+ as well as CD133+CD24+CD106− cells proliferated once engrafted in the renal tissue, as demonstrated by their costaining with the mitosis marker phospho H3-histone (Fig. 5I, 5J). We then evaluated the capacity of CD133+CD24+CD106+ as well as CD133+CD24+CD106− cells to replace podocytes in an experimental model of FSGS, which is characterized by podocyte injury. To this aim, adriamycin nephropathy was induced in SCID mice, and PKH26-labeled CD133+CD24+CD106+, CD133+CD24+CD106−, or CD133−CD24− cells were injected into the tail vein of mice the day following adriamycin injection. Labeled CD133+CD24+CD106+ cells were engrafted within the tissue and generated novel tubular cells as well as podocytes, as demonstrated by costaining of PKH26-labeled cells with LTA (data not shown) or with nephrin (Supporting Information Fig. S5A). By contrast, CD133+CD24+CD106− cells only generated tubular cells, while PKH26-labeled podocytes were never observed (Supporting Information Fig. S5B), thus suggesting that only CD133+CD24+CD106+ cells could generate novel podocytes (6.42% ± 2.3% CD133+CD24+CD106+ vs. 6.83% ± 2.6% CD133+CD24+CD106− PKH26-labeled/LTA-stained cells, NS and 5.84% ± 1.4% CD133+CD24+CD106+ vs. 0.02% ± 0.01% CD133+CD24+CD106− PKH26-labeled/nephrin-stained cells; p < .05). Red labeling was never observed in mice injected with CD133−CD24− cells (Supporting Information Fig. S5C; 6.42% ± 2.3% CD133+CD24+CD106+ vs. 0.012% ± 0.005% CD133−CD24− PKH26-labeled/LTA-stained cells, p < .05; 6.83% ± 2.6% CD133+CD24+CD106− vs. 0.012% ± 0.005% CD133−CD24− PKH26-labeled/LTA-stained cells, p < .05; 5.84% ± 1.4% CD133+CD24+CD106+ vs. 0.017% ± 0.01% CD133−CD24− PKH26-labeled/nephrin-stained cells, p < .05; 0.02% ± 0.01% CD133+CD24+CD106− vs. 0.017% ± 0.01% CD133−CD24− PKH26-labeled/nephrin-stained cells, NS).
CD133+CD24+CD106+ and CD133+CD24+CD106− Cells Proliferate Following Injury in Kidney of Patients with Acute or Chronic Tubular Damage
To evaluate the status of the CD133+CD24+CD106+ and CD133+CD24+CD106− renal progenitor cells during tubular regeneration, renal biopsy cases demonstrating signs of acute or chronic tubular injury were analyzed. In all biopsies, diffuse CD133 staining was observed in tubular structures where light microscopic signs of tubular regeneration such as nuclear prominence or flattening of tubular cells were present (Fig. 6A–6D). In patients affected by acute tubular necrosis, double-labeling immunofluorescence demonstrated that CD133+CD24+CD106− cells formed 22.3% ± 6.6% of total tubular cells, while CD133+CD24+CD106+ represented 6.21% ± 2.7% (Fig. 6A, 6B; p < .05). Interestingly, in biopsies where signs of chronic tubular damage were observed, CD133+CD24+CD106− still represented 26.3% ± 4.5% of total tubular cells, while CD133+CD24+CD106+ cells were 12.3% ± 3.7% of total tubular cells (Fig. 6C; p < .05). In regions of tubular regeneration, stretches of cells bearing CD133 positivity also costained with the mitosis marker phospho H3-histone (Fig. 6D), thus demonstrating that these cells actively proliferate following tubular injury.
Most epithelia need to constantly replace damaged or dead cells throughout life [5, 6]. The process of cell replacement is typically maintained through the presence of stem or progenitor cells, which are also critical players of adult tissue repair following injury [5, 6]. Indeed, once activated, stem/progenitor cells can divide expanding the cellular pool that will then differentiate along a particular cell lineage to make the tissue [5, 6]. Consistently, the adult human kidney contains a stem/progenitor cell compartment similar to those identified in other adult organs of epithelial or nonepithelial origin [7–14]. Indeed, CD133+CD24+ renal progenitors that display the capacity to differentiate into podocytes and tubular cells [7–14], as well as podocyte-committed progenitors, were previously described within Bowman's capsule [11, 13, 17, 19]. In this manuscript, we demonstrate the existence of a population of tubular-committed progenitors localized within the proximal and distal tubule of normal human kidneys. These cells shared expression of CD133 and CD24 with the progenitors of Bowman's capsule but could be distinguished because they did not express the surface marker CD106. This allowed us to recover the two populations separately and analyze their phenotype and function. CD133+CD24+CD106+ progenitors of Bowman's capsule displayed the potential to differentiate toward the podocyte as well as the tubular phenotype, while CD133+CD24+CD106− cells already displayed a tubular-committed phenotype. More importantly, CD133+CD24+CD106+ as well as CD133+CD24+CD106− cells displayed the capacity to regenerate tubular structures and reduced the morphological and functional kidney damage in mice affected by AKI, suggesting that both cell types can potentially participate in tubular regeneration in adult human kidneys. However, our current results and previous studies suggest that the contribution of CD133+CD24+CD106+ or CD133+CD24+CD106− progenitors to tubular regeneration may be different. Indeed, several studies demonstrated that cells showing a mixed phenotype between parietal and proximal tubular epithelium increase at the tubuloglomerular junction during kidney growth, chronic renal disorders [20–23], or during ageing , suggesting that renal progenitors of Bowman's capsule may be able to change to tubular cells in these conditions. However, during AKI, repair of epithelial cells mostly depends on cells directly derived from the migration and proliferation of adjacent tubular epithelial cells [1, 2]. In addition, a recent study in rodents demonstrated that proliferation of regenerating tubular epithelium following acute ischemic injury is stochastically distributed among surviving tubular cells . However, the existence of a scattered population of CD133+CD24+ tubular-committed progenitors that is highly resistant to death induced by toxic agents is consistent with this observation. Indeed, CD133+CD24+ tubular-committed progenitors, which already represent 2%–6% of proximal tubular cells in healthy conditions, are increased when terminally differentiated tubular cells are damaged and die, and these progenitors then become a large proportion of the surviving epithelium (Fig. 6). Regenerating tubular epithelial cells in human tissues of patients with AKI consistently display long stretches of CD133+CD24+ cells , further supporting that renal progenitors localized within the tubule proliferate and differentiate to replace injured cells. Interestingly, renal progenitors in humans mostly localize in the proximal tubule and specifically in the S3 segment, which is highly susceptible to ischemia and toxic insults but has a remarkable capacity to repair its structure and function [26, 27]. This is also in agreement with a previous report, which suggested the possible presence of tubular progenitors in the S3 segment of rat kidneys . The existence of CD133+CD24+ tubular progenitors and their localization as scattered cells among more differentiated tubular cells raises the question of how these cells were generated. Interestingly, during kidney development, CD133 and CD24 coexpression identifies renal embryonic progenitors in condensed mesenchyme-derived primordial structures, such as primary vesicles, comma-shaped bodies, and S-shaped bodies [8, 10]. Previous observations suggest that while the S-shaped bodies elongate and the glomerular precursor region becomes evident, a major cluster of CD133+CD24+ progenitors remains localized at the urinary pole of Bowman's capsule [8, 10, 29], a smaller one remains localized in the distal convoluted tubule where it takes connection with the vascular pole of the glomerulus, while some CD133+CD24+ cells may remain included into the elongating nephron and constitute the scattered CD133+CD24+ tubular progenitors. This would also explain why CD133+CD24+CD106− cells selectively localize within the proximal tubule and the distal tubule that are directly generated by elongation of the S-shaped body but not in the collecting ducts that are not derived by the metanephric mesenchyme . The localization of a cluster of CD133+CD24+ tubular progenitors in the distal convoluted tubule and the connecting segment is also reminiscent of recent observations performed in fish . Indeed, Diep et al.  demonstrated that renal progenitors also exist in the adult kidney of the zebrafish, where they localize at the point of connection between the tubule and the duct.
Taken together, the results of this study provide evidence that the tubular compartment of the nephron contains scattered renal progenitors that display functional regenerating capacity for injured renal tubular cells as well as increased resistance to apoptotic stimuli. However, there are several questions that still need to be answered. First of all, genetic tagging experiments are required to provide the proof of concept that this population participates in tubular regeneration after acute and/or chronic tubular injury. Using transgenic mouse lines that express LacZ or Cre-recombinase under the control of an NFATc1 autoregulatory enhancer, a previous study already suggested the putative existence of a scattered progenitor cell population that proliferates to repopulate the proximal tubule after injury with HgCl2 . This putative progenitor population was acutely resistant to apoptosis and participated in regeneration of the damaged proximal tubule in rodents , consistently with the progenitors described here in humans. In addition, further studies are necessary to evaluate the possibility that CD133+CD24+CD106+ or CD133+CD24+CD106− cells may differentially contribute to regeneration during acute or chronic tubulointerstitial disorders in human. Finally, these findings do not exclude the possibility of further regenerative strategies to repair injured tubular cells in adult mammalian kidneys. Indeed, animals exploit different strategies for regeneration, and often more than one mechanism is involved in regeneration of a tissue , because this confers an advantage during evolution. However, the existence of tubular progenitors suggests that novel treatments aimed at promoting the regenerative capacities of the kidney could conceivably be possible and be used to prevent and treat tubular injury.
In this study, we demonstrate that CD133+CD24+ renal progenitors can be distinguished in bipotent progenitors, podocyte committed progenitors, or tubular progenitors based on surface expression of PDX or CD106. In addition, we provide evidence that tubular committed progenitors display resistance to death, higher proliferative potential than other differentiated renal cells, and can regenerate injured tubular cells following acute kidney injury. Thus, the results of this study provide further evidence for the existence of a human “renopoietic system” that accounts for the regenerative capacity of the kidney.
This manuscript is supported by the European Research Council Starting Grant under the European Community's Seventh Framework Programme (FP7/2007-2013), ERC Grant number 205027, by the European Community under the European Community's Seventh Framework Programme (FP7/2007-2013), Grant number 223007, by the Tuscany Ministry of Health (Bando Salute 2009), by the Italian Ministry of Health (RF-TOS-2008-1211098), and by the Associazione Italiana per la Ricerca sul Cancro. Maria Lucia Angelotti is recipient of a FIRC fellowship.
DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST
P.R. has intellectual property rights/inventor or patent holder, consultant/advisory role, and research funding/contracted research.