Isolation and Characterization of Multipotent Progenitor Cells from the Bowman’s Capsule of Adult Human Kidneys

Abstract

Regenerative medicine represents a critical clinical goal for patients with ESRD, but the identification of renal adult multipotent progenitor cells has remained elusive. It is demonstrated that in human adult kidneys, a subset of parietal epithelial cells (PEC) in the Bowman’s capsule exhibit coexpression of the stem cell markers CD24 and CD133 and of the stem cell–specific transcription factors Oct-4 and BmI-1, in the absence of lineage-specific markers. This CD24⁺CD133⁺ PEC population, which could be purified from cultured capsulated glomeruli, revealed self-renewal potential and a high cloning efficiency. Under appropriate culture conditions, individual clones of CD24⁺CD133⁺ PEC could be induced to generate mature, functional, tubular cells with phenotypic features of proximal and/or distal tubules, osteogenic cells, adipocytes, and cells that exhibited phenotypic and functional features of neuronal cells. The injection of CD24⁺CD133⁺ PEC but not of CD24⁻CD133⁻ renal cells into SCID mice that had acute renal failure resulted in the regeneration of tubular structures of different portions of the nephron. More important, treatment of acute renal failure with CD24⁺CD133⁺ PEC significantly ameliorated the morphologic and functional kidney damage. This study demonstrates the existence and provides the characterization of a population of resident multipotent progenitor cells in adult human glomeruli, potentially opening new avenues for the development of regenerative medicine in patients who have renal diseases.

Chronic renal failure is a leading cause of mortality and morbidity in Western countries (¹). The number of patients with ESRD is growing consistently, and the cumulative ESRD costs are even greater than the direct treatment costs of cancer (¹). Therefore, the potential use of stem cells (SC) for regenerative medicine to treat kidney diseases represents a critical clinical goal (²).

The postnatal kidney has a high capacity to regenerate and repair, as illustrated by its functional recovery after glomerular or tubular injury (^2,3); however, the origin of newly generated renal cells has not yet been defined. Some cells seem to derive from the division of fully differentiated cells, and recent reports suggested that these cells might represent tubular progenitors, expressing lineage-specific markers (^4–8). Another study reported that potential tubular progenitors are present in renal interstitium (⁹). Moreover, although glomerular injury is critical for initiation of irreversible renal failure, the existence of progenitors within glomerular structures has not yet been described (^2,4–9). Previous reports suggested that bone marrow may be a source of progenitors for tubule turnover and/or repair (^2,10,11). However, recent studies have shown that unidentified intrarenal, not bone marrow–derived, cells mostly are responsible for regeneration in ischemic acute renal failure (ARF) (^12,13).

The best strategy to identify and amplify multipotent progenitors and/or SC has been their selection on the basis of functional properties of self-renewal, clonogenicity, multidifferentiation, and/or expression of specific markers (^14,15). To identify multipotent progenitors and/or SC in adult human kidney, we assessed the presence of both CD24, a surface molecule that has been used to identify different types of human SC (^16,17) and also is expressed by uninduced metanephric mesenchyme during renal embryogenesis (¹⁸), and CD133, a marker of adult tissue SC (^19,20). The results showed that both markers were coexpressed by a subset of parietal epithelial cells (PEC) in the Bowman’s capsule. Once isolated, CD24⁺CD133⁺ PEC were found to lack lineage-specific markers; to express transcription factors that are characteristic of multipotent SC; and to exhibit self-renewal, high clonogenic efficiency, and multidifferentiation potential. When injected intravenously in SCID mice that had ARF, CD24⁺CD133⁺ PEC regenerated tubular structures in different portions of the nephron and also reduced the morphologic and functional kidney damage with important implications for the development of regenerative medicine in patients who have renal diseases.

Materials and Methods

Antibodies

The following antibodies (Ab) were used: Anti-CD24 and anti-vimentin (Santa Cruz Biotechnology, Santa Cruz, CA); anti-CD133/1 (clone AC133) and anti-CD133/2 (clone 293C3; Miltenyi Biotec GmbH, Bergisch Gladbach, Germany); anti-CD105, anti-CD31, anti-CD34, anti-CD35, anti-CD45, and mouse IgG1 and IgG2b (BD Biosciences, San Diego, CA); anti-CD29 mAb, rabbit anti–choline acetyl-transferase (ChAT), anti–neurofilament M (NFM), and anti–microtubule-associated protein-2 (MAP-2; Chemicon International, Temecula, CA); anti-CD106 and anti–epithelial membrane antigen-1 (EMA-1; Dako, Glostrup, Denmark); anti-cytokeratin, anti–human HLA-I, anti–α-smooth muscle actin (α-SMA), and rabbit anti–neurofilament H (NF200; Sigma-Aldrich, St. Louis, MO); anti-CD54, PE-conjugated anti-CD106 and IgG2a mAb, goat anti-mouse IgG1, and rabbit anti-goat IgG Ab (Southern Biotech, Birmingham, AL). PE-conjugated anti-CD105 mAb was from Ancell Corp. (Bayport, MN). The anti-LAP (TGF-β1) mAb was from R&D Systems (Minneapolis, MN). The anti–Tamm-Horsfall glycoprotein (THG) goat polyclonal Ab was from MP Biomedicals (Verona, Italy). Alexa Fluor 488–, 546–, 633–, and 647–labeled goat anti-mouse IgG1; Alexa Fluor 488–labeled goat anti-mouse IgG2a or goat anti-rabbit IgG; and Alexa Fluor 488–and 546–labeled goat anti-mouse IgG2b or rabbit anti-goat IgG Abs were from Molecular Probes (Eugene, OR).

Tissues

Normal kidney fragments were obtained from 20 patients who had localized renal tumors and underwent nephrectomy, in accordance with the recommendations of the Regional Ethical Committee on human experimentation.

Isolation and Culture of CD24⁺CD133⁺ PEC

To obtain CD24⁺CD133⁺ PEC, we minced the cortex and isolated glomeruli by a standard sieving technique through graded mesh screens (60, 80, and 150 mesh). The glomerular suspension was collected, washed with endothelial growth medium–microvascular (EGM-MV; Cambrex Bio Science, East Rutherford, NJ) without serum, and plated on fibronectin-coated dishes (10 μg/ml; Sigma-Aldrich) at a density of 200 glomeruli/100-mm plate. To save the Bowman’s capsule, we performed no enzymatic digestion. After 4 to 5 d of culture, isolated glomeruli adhered to the plate, resulting in cellular outgrowth that usually was detectable after 5 d of culture. Glomeruli then were detached, and adherent cells were cultured as bulk. Several culture media were compared. EGM-MV 20% FBS (Hyclone, Logan, UT) yielded the highest degree of purity and the best amplification efficiency and therefore was used in subsequent experiments. Bulk cultures were checked for simultaneous expression of CD133 and CD24 by flow cytometry and then used for cloning. Generation of clones from CD24⁺CD133⁺ PEC that were obtained from glomerular outgrowth was achieved by limiting dilution in fibronectin-coated 96-well plates in EGM-MV 20% FBS.

CD24⁺CD133⁺ PEC also were maintained in culture as bulk, and routine cell passaging was performed. Medium was changed twice a week. The cell counts and cellular dilution factor were recorded at each passage. This process was repeated for a 4-mo period. The number of population doublings (PD) was calculated by solving the following equation: n of PD = log₂(N_i/N_o), where N_i is the number of cells yielded and N_o is the number of cells plated.

Cell Cultures

Human mesangial cells and human glomerular visceral epithelial cells were obtained as described (^21,22). Human renal proximal tubular cells, human microvascular endothelial cells, and human aortic smooth muscle cells were obtained from Cambrex Bio Science, and the HEK-293 cell line was obtained from ECACC (Sigma-Aldrich).

Immunomagnetic Cell Sorting

Single-cell suspensions were obtained from kidney cortical tissue specimens by mechanical disaggregation using the Medimachine System (BD Biosciences). Anti-CD45 MicroBeads, anti–glycophorin A MicroBeads, anti-FITC Multisort Kit, anti-PE Multisort Kit, and CD133 Cell Isolation Kit were obtained from Miltenyi Biotec GmbH.

Isolation of CD24⁺CD133⁺ and CD24⁻CD133⁻ cells was performed by high-gradient magnetic cell sorting (²³). The positive cell fractions consisted of >97% of CD24⁺CD133⁺ cells. Generation of clones from CD24⁺CD133⁺ and CD24⁻CD133⁻ cells was achieved by limiting dilution in fibronectin-coated 96-well plates in EGM-MV 20% FBS.

Confocal Microscopy

Confocal microscopy was performed on 5-μm sections of renal frozen tissues or on cells that were cultured on chamber slides as described (²⁴) by using an LSM 510 META laser confocal microscope (Carl Zeiss, Jena, Germany).

Staining with Alexa Fluor 488 Phalloidin (Molecular Probes), FITC-labeled Dolichos Biflorus Agglutinin (DBA), and FITC-labeled Lotus Tetragonolobus lectin (LTA; Vector Laboratories, Burlingame, CA) were performed following the manufacturers’ instructions.

For quantification of fibrosis in glycerol-injected SCID mice (see later), four random sections of kidney tissue that stained for α-SMA or TGF-β1 were recorded using a ×20 objective and scanned from each tissue of 16 mice (eight mice that were treated with CD24⁺CD133⁺ PEC and eight mice from the saline-treated group). All random scans of the kidney tissue for each treatment group were recorded at the same photo multiplier tube, pinhole aperture, and laser voltage setting and analyzed using LSM 510 confocal microscopy software 3.0. This analysis resulted in a data set expressed as fibrotic tissue (α-SMA and TGF-β1 positive) area in μm² per image field.

Real-Time Quantitative Reverse Transcriptase–PCR

Taq-Man reverse transcriptase–PCR (RT-PCR) was performed as described (²⁵). BmI-1, Tau protein, MAP-2, necdin, neural enolase, nestin, β-tubulin III, Na/H exchanger, aminopeptidase A, Na/glucose co-transporter (Na/Gluc1), γ-glutamyltransferase (γ-GT), aquaporin-1 (AQP1), aquaporin-3 (AQP3), Na/Cl transporter, Runx2, and adiponectin quantification was performed using Assay on Demand kits (Applied Biosystems, Warrington, UK). Oct-4 mRNA expression and quantification were performed as described (²⁴).

Flow Cytometry

Flow Cytometry was performed as described (²³).

In Vitro Differentiation

Tubulogenic differentiation was obtained by culturing clones of CD24⁺CD133⁺ PEC in commercially available REBM medium that contained SingleQuotes (hydrocortisone, hEGF, FBS, epinephrine, insulin, triiodothyronine, transferrin, and gentamicin/amphotericin-B; Cambrex Bio Science) and was supplemented with 50 ng/ml hepatocyte growth factor (HGF) (Peprotech, Rocky Hill, NJ) for 30 d.

Osteogenic, adipogenic, or neurogenic differentiation of CD24⁺CD133⁺ PEC–derived clones was induced as described elsewhere (^24,26,27). For osteogenic induction, CD24⁺CD133⁺ PEC were cultured in α-MEM and 10% horse serum that contained 100 nM dexamethasone, 50 μM ascorbic acid, and 2 mM β-glycero-phosphate (all from Sigma-Aldrich). The medium was changed twice a week for 3 wk. For adipogenic differentiation, CD24⁺CD133⁺ PEC were incubated in DMEM high glucose (hg; Invitrogen, Carlsbad, CA) that contained 10% FBS, 1 μM dexamethasone, 0.5 μM 1-methyl-3-isobutylxanthine, 10 μg/ml insulin, and 100 μM indomethacin (all from Sigma-Aldrich). After 72 h, the medium was changed to DMEM hg, 10% FBS, and 10 μg/ml insulin for 24 h. These treatments were repeated three times. The cells then were maintained in DMEM hg, 10% FBS, and 10 μg/ml insulin for one additional week. For neurogenic differentiation, CD24⁺CD133⁺ PEC were plated in DMEM hg and 10% FBS. After 24 h, medium was replaced with DMEM hg, 10% FBS that contained B27 (Invitrogen), 10 ng/ml EGF (Peprotech), and 20 ng/ml basic fibroblast growth factor (Peprotech). Five days later, cells were washed and incubated with DMEM that contained 5 μg/ml insulin, 200 μM indomethacin, and 0.5 mM 1-methyl-3-isobutylxanthine in the absence of FBS for 5 h. Alizarin red, Oil-Red O, and alkaline phosphatase (AP) staining was performed as described (^24,26,27).

Electrophysiologic Analysis

The whole-cell patch-clamp technique was performed in voltage-clamp conditions, as described in detail previously (²⁸). Cells in the recording chamber were superfused at a rate of 1.8 ml/min at 22 to 24°C with the following bath solution: 122.5 mM NaCl, 2 mM CaCl₂, 10 mM HEPES, and 20 mM TEA-OH as K⁺ channel blocker. For blocking of Na⁺ and L-type Ca²⁺ channels, 1 μM Tetrodotoxin (TTX), 10 μM nifedipine, and 100 μM Cd²⁺ (added as CdSO₄) were used. Pipette solution contained 150 mM CsBr, 5 mM MgCl₂, 10 mM EGTA, and 10 mM HEPES. For bath and pipette solution, pH was titrated to 7.4 with NaOH and to 7.2 with TEA-OH, respectively. Pipettes resistance was 2 to 3 MΩ.

Determination of [Ca²⁺]_i

[Ca²⁺]_i was determined by a laser confocal microscope (LSM 510 META, Zeiss), as described (²⁹).

Xenograft in SCID Mice Model of ARF

Models of rhabdomyolysis-induced ARF were performed in female SCID mice (Harlan, S. Pietro al Natisone, Italy), as described previously (^30,31), by intramuscular injection with hypertonic glycerol (8 ml/kg body 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 to the National Institutes of Health Guide for the Care and Use of Laboratory Animals. Two groups of mice on days 3 and 4 after glycerol received an intravenous injection into the tail vein as follows: Group 1, saline (n = 32); and group 2, CD24⁺CD133⁺ PEC (n = 32; 0.75 × 10⁶ on day 3 and 0.75 × 10⁶ on day 4) obtained from five different human donors (two men and 3 women). Mice were killed at different time intervals (days 3, 7, 11, and 14), and samples for blood urea nitrogen (BUN) determination were collected. BUN levels were measured in heparinized blood by the Aeroset c8000 test (Abbott, Wiesbaden, Germany). BUN levels that exceeded 40 mg/dl were considered abnormal. Normal range in our experiments was between 30 and 37 mg/dl, as calculated in 16 additional untreated mice that were killed when the other mice received injection of glycerol (day 0). Two additional groups of mice were treated as follows: Group 3 (n = 8) received an intravenous injection of CD24⁺CD133⁺ cells (0.75 × 10⁶ on day 3 and 0.75 × 10⁶ on day 4) that were labeled with the PKH26 Red Fluorescence Cell Linker Kit (Sigma-Aldrich); and group 4 (n = 8) received an intravenous injection of CD24⁻CD133⁻ cells (0.75 × 10⁶ on day 3 and 0.75 × 10⁶ on day 4) that were labeled with the red fluorescent dye PKH26 and obtained from three female human donors. Mice were killed 10 d later (day 14). BUN levels were measured as described above. Kidneys were collected from all groups of animals.

Immunohistochemistry and Fluorescence In Situ Hybridization Analysis

Immunohistochemical analysis was performed as described previously (²⁵). Sections for fluorescence in situ hybridization analysis were prepared using the Paraffin Pretreatment Reagent Kit (Vysis, Olympus, Milan, Italy), and Y chromosome was revealed with whole chromosome paint Y spectrum orange probe (Vysis, Olympus), following the manufacturer’s instructions.

Statistical Analyses

The results are expressed as mean ± SD. Comparison between groups was performed by the Mann-Whitney test. P < 0.05 was considered to be statistically significant.

Results

Subset of PEC of the Bowman’s Capsule Coexpresses CD24 and CD133

Analysis by confocal microscopy of tissue specimens from normal adult human kidneys revealed that CD24, a marker of the renal embryonic progenitor population, was expressed not only by a subset of distal tubules that localized mainly in the medulla but also by a subset of PEC of the Bowman’s capsule (Figure 1, A and B). Of note, an anti-CD133 mAb (clone 293C3), which recognizes an epitope that is selectively expressed by adult SC, co-stained the same subset of PEC (Figure 1, A and B), in addition to some interstitial cells, and rare tubular structures. Such a distribution of CD133 was confirmed by using another anti-CD133 mAb (clone AC133; Figure 1C). Only PEC that localized in close proximity to the tubule/glomerular junction at the urinary pole showed the co-staining for CD24 and CD133 (Figure 1A). A triple staining for CD24, CD133, and CD29 (which allowed clear definition of glomerular morphology) confirmed that CD24⁺CD133⁺ PEC localized opposite to the vascular pole (Figure 1D). CD24⁺CD133⁺ PEC also coexpressed CD106 (Figure 1E), CD105, CD54, and CD44 (data not shown). Triple-label immunofluorescence demonstrated co-localization of CD24 and CD133 in the cytoplasm and on the membrane of PEC facing the glomerulus (G), whereas only CD24 was expressed on the basal membrane of the cells (Figure 1B), where it co-localized with CD106 (Figure 1E).

Isolation and Characterization of CD24⁺CD133⁺ PEC

To obtain CD24⁺CD133⁺ PEC free from any other renal cell type, we plated isolated capsulated glomeruli in fibronectin-coated dishes. After 5 d of culture, cellular outgrowth from scattered adherent glomeruli was observed (Figure 2A). Confocal microscopy demonstrated that these cells originated from the outgrowing of CD24⁺ PEC of the Bowman’s capsule (Figure 2B). Cells that grew out from plated glomeruli that stained with an isotype control yielded negative results (Figure 2C). FACS analysis showed that the recovered population homogeneously exhibited the presence of CD24, CD133, CD106, CD105, and CD44 (Figure 2D), whereas surface molecules that were expressed by endothelial cells (CD31, CD34) were not detectable (Figure 2D). FACS analysis for the podocyte marker CD35 yielded negative results (Figure 2E). Expression of the podocyte markers WT-1 and synaptopodin was very weak or absent (Figure 2F), whereas the same markers strongly stained cultured podocytes (Figure 2G), as assessed by confocal microscopy. The absence of contaminating cells that originated from distal tubules or collecting ducts was confirmed at the protein level by the lack of EMA-1 (Figure 2H) and THG expression (Figure 2I, left). Furthermore, the negative staining with LTA (Figure 2I, middle) and for AP excluded the possible contamination by proximal tubules (Figure 2I, right).

To address further the nature of CD24⁺CD133⁺ cells, we also assessed and compared mRNA levels of several lineage-specific renal cell markers in freshly isolated CD24⁺CD133⁺ and CD24⁻CD133⁻ cells by real-time RT-PCR. To this aim, total cell suspensions of digested cortical renal tissue were sorted with immunomagnetic beads into CD24⁺CD133⁺ and CD24⁻CD133⁻ cells. CD24⁺CD133⁺ cells represented 0.5 to 4% of cortical renal cells, whereas CD24⁻CD133⁻ cells were on average 95 to 99% of them. As expected, in agreement with their nature of fully differentiated cells, CD24⁻CD133⁻ cells were highly enriched in differentiation markers that were specific for different portions of the nephron (Figure 2J), such as aminopeptidase A and Na/Gluc1 (proximal tubules), γ-GT (proximal convoluted tubules), the Na/H exchanger (proximal tubule and apical membrane of thin and thick limbs of the loop of Henle), AQP1 (proximal tubule and descending thin-limb epithelial cells), AQP3 (principal cells in collecting ducts), or the thiazide-sensitive Na/Cl transporter (distal convoluted tubule). By contrast, even with such a highly sensitive technique, irrelevant levels of all of the renal differentiation markers analyzed were found in CD24⁺CD133⁺ cells, further suggesting the absence of contaminating tubular cells of any portion of the nephron and/or their nature of undifferentiated cells (Figure 2J).

CD24⁺CD133⁺ PEC Exhibit Self-Renewal and High Clonogenic Potential

The growth properties and clonogenic potential of CD24⁺CD133⁺ PEC then were assessed and compared with those of CD24⁻CD133⁻ cells. The proliferative potential of CD24⁺CD133⁺ PEC was consistently higher than that of CD24⁻CD133⁻ cells (Figure 3A). CD24⁺CD133⁺ PEC were grown continuously for 60 to 90 PD, depending on the donor (Figure 3B), during a period of 4 mo. When assessed at 50 PD, cells exhibited diploid DNA content (Figure 3C). Accordingly, CD24⁺CD133⁺ PEC expressed high levels of the SC-specific (^32,33) transcription factors BmI-1 (Figure 3D) and Oct-4 (Figure 3E). Comparable levels of Oct-4 and BmI-I were observed in the human embryonic kidney cell line HEK (Figure 3, D and E). By contrast, several types of human primary cultures, as well as CD24⁻CD133⁻ cells, displayed irrelevant levels of Oct-4 and BmI-1, in agreement with their nature of differentiated cells (Figure 3, D and E).

The clonogenicity of CD24⁺CD133⁺ PEC that were obtained from glomerular outgrowth was assessed by limiting dilution in a 96-well plate. Two hours after plating, the wells that contained only one cell were identified by microscopy examination and marked. Plating of 2700 CD24⁺CD133⁺ PEC, obtained from eight different donors, was required to seed 894 single cells. A total of 370 clones were collected, with a cloning efficiency equal to 41.3 ± 14% (Figure 4A). The cloning efficiency of CD24⁺CD133⁺ cells that were obtained by immunomagnetic sorting was similar (37 ± 11%), whereas CD24⁻CD133⁻ cells generated rare clones (2.1 ± 1.9%; Figure 4A). Figure 4B shows an example of single CD24⁺CD133⁺ PEC deposition followed by its subsequent proliferation. Among the 370 clones that were obtained from CD24⁺CD133⁺ PEC, 50 clones that showed the highest amplification potential were selected and analyzed in detail. These clones stably maintained the double CD24⁺CD133⁺ phenotype on 100% of their cells (Figure 4C), remained in undifferentiated state, were long surviving, and could be amplified up to 1 × 10⁸ cells. The same clones coexpressed cytokeratin and vimentin (Figure 4D) but did not express renal lineage markers, such as the presence of AP, the binding of LTA, and the expression of THG (Figure 5, A through C, day 0), EMA-1, or CD35 (data not shown). Importantly, seeding of single-cell suspensions that were generated from each clone generated secondary clones that were identical to primary clones, and seeding of single cells that were generated from secondary clones led to the generation of tertiary clones, further demonstrating that CD24⁺CD133⁺ PEC exhibited self-renewal in culture.

Single CD24⁺CD133⁺ PEC Exhibit Multilineage Differentiation Potential

CD24⁺CD133⁺ PEC that were obtained from the same single clone were cultured under conditions that were favorable for tubulogenic, osteogenic, adipogenic, or neurogenic differentiation; the same experimental procedure was repeated with 50 different clones. Differentiation toward tubular cells resulted in the acquisition of markers of proximal tubular epithelia, such as AP (Figure 5A), and the binding of LTA (Figure 5B). Furthermore, THG, a marker of distal tubules, also appeared on the cells after their differentiation (Figure 5C). When double labeling for LTA and THG was performed, coexistence in the same clone of cells that coexpressed markers of multiple tubule segments and of others that expressed only proximal or distal tubule markers were observed (Figure 5D). Real-time RT-PCR demonstrated strong upregulation of other markers of different portions of the nephron, such as aminopeptidase A and Na/Gluc1, γ-GT, the Na/H exchanger, AQP1, AQP3, or the thiazide-sensitive Na/Cl transporter (Figure 5E). Furthermore, differentiated tubular cells acquired the capacity to respond to angiotensin II with intracellular calcium influx (Figure 5F), a property that was not shared by undifferentiated CD24⁺CD133⁺ PEC.

After osteoinduction, cloned CD24⁺CD133⁺ PEC formed AP-positive colonies (Figure 6A, left) that, during differentiation, transformed into mineralized nodules, as assessed by Alizarin Red staining (Figure 6A, left). Accordingly, osteogenesis associated with upregulation of Runx2 mRNA levels (Figure 6A, right). Adipogenic differentiation of CD24⁺CD133⁺ PEC was demonstrated by Oil Red O staining of lipid vacuoles, which were completely absent from undifferentiated cells (Figure 6B, left), and upregulation of adiponectin mRNA levels (Figure 6B, right).

CD24⁺CD133⁺ PEC that were cultured under neurogenic conditions acquired the expression of NF-200, neurofilament M, choline acetyl-transferase, MAP-2 (Figure 7, A and B), and a neuron-like morphology (Figure 7C). Furthermore, differentiated cells exhibited strong upregulation of mRNA levels of τ-protein, MAP-2, necdin, neural enolase, nestin, and β-tubulin III (Figure 7D). Patch-clamp recordings that were performed on CD24⁺CD133⁺ PEC–derived neural cells put in evidence a slow high-voltage activated inward current (Figure 7E). The voltage threshold of its occurrence (approximately −40 mV) and its typical time course suggested the activation of the high-voltage operated L-type Ca²⁺ channel. Accordingly, the specific channel blockers nifedipine and Cd²⁺ completely abolished such a current. Moreover, the I-V relation showed maximal current amplitude at 20 mV and could be fitted best by a Boltzmann function with parameters in agreement with the presence of L-type Ca²⁺ channel of neuronal type (Figure 7F). In the presence of L-type Ca²⁺ channel blockers, a fast, transient inward current could be recorded in the first 10 ms of the test pulse (Figure 7G). This current resolved in approximately 7 ms at −10 mV, peaking at 0.7 ms. It activated from −40 mV and showed a maximum at −5 mV. Addition of 1 μM TTX completely but reversibly abolished this current (Figure 7G, red line). Figure 7I shows current traces that were recorded during inactivation. The I-V activation and inactivation curve (Figure 7, H and J) were fitted by Boltzmann function with parameters in agreement with TTX-sensitive Na⁺ channels of neuronal type.

Intravenously Injected CD24⁺CD133⁺ PEC Regenerate Tubular Cells in SCID Mice with ARF

To test the ability of CD24⁺CD133⁺ PEC to participate to renal repair, we used an in vivo model of rhabdomyolysis-induced ARF in SCID mice, generated by intramuscular injection of glycerol. Compared with normal renal tissue (Figure 8A), kidneys from glycerol-treated mice showed vacuolization, widespread necrosis of tubular epithelial cells, and tubular hyaline cast formation (Figure 8B). Proximal and distal tubules displayed loss of brush border and flattening of epithelial cells (Figure 8B). At the peak of tubular injury, CD24⁺CD133⁺ PEC as well as CD24⁻CD133⁻ cells were labeled with the red fluorescent dye PKH26, and each of the two cell populations was injected into the tail vein of glycerol-treated SCID mice. Ten days later, both kidneys were harvested from each mouse, and sections were analyzed for the presence of labeled cells. Labeled cells were never detected in control mice that received injections of CD24⁻CD133⁻ cells (Figure 8C) or of saline solution (data not shown), whereas in mice that received injections of CD24⁺CD133⁺ PEC, they spread to the cortex and the medulla (Figure 8D, red). Most injected PEC localized inside the tubules, where they expressed specific markers of different portions of the nephron (Figure 8, D through F, arrows), even when some cells also were observed in the interstitium (Figure 8D) and a very few in the glomeruli (data not shown). Quantification of the number of PKH26-positive cells that expressed markers of differentiated tubular cells was performed on sections that were stained with LTA or DBA. The number of PKH26-labeled/LTA-stained tubular cells was equal to 6.48 ± 3.4% of all LTA-stained proximal tubular cells, whereas proportions of PKH26-labeled/DBA-stained cells corresponded to 5.8 ± 2.6% of all DBA-stained distal tubules/collecting ducts. In SCID mice that received injections of CD24⁻CD133⁻ cells (Figure 8G, left) or saline (data not shown), HLA-I human antigen expression was never found, whereas human HLA-I antigen was detected consistently in mice that received injections of CD24⁺CD133⁺ cells (Figure 8G, middle and right). Double-label immunohistochemistry for human HLA-I antigen and cytokeratin confirmed the engraftment of CD24⁺CD133⁺ PEC into tubular structures (Figure 8G, middle and right). In addition, whereas kidney cells of mice that received injections of saline (Figure 8H, left) never exhibited the presence of the Y chromosome (Figure 8H, left), this could be detected clearly in kidney cells from mice that received injections of CD24⁺CD133⁺ PEC that were derived from male human donors (Figure 8H, middle and right, red).

Effects of CD24⁺CD133⁺ PEC on Renal Function and Structure

To determine whether CD24⁺CD133⁺ PEC also could influence the renal function of mice with ARF, we measured BUN levels in mice that had glycerol-induced ARF and received injections of CD24⁺CD133⁺ PEC or saline. In our setting, injection of glycerol induced significant increases in serum BUN, which peaked at day 3, declined at day 7, and stabilized at days 11 and 14 to values that were significantly higher than those in healthy mice (Figure 9, A and B), as described previously (^34,35). Intravenous injection of CD24⁺CD133⁺ PEC on days 3 and 4 seemed to be protective of renal function, as reflected by significantly lower BUN values on days 11 and 14 in comparison with mice that received injections of saline or CD24⁻CD133⁻ cells (Figure 9, A and B). It is interesting that on day 14, mice that were treated with CD24⁺CD133⁺ PEC displayed completely restored renal function, with BUN levels that were not significantly different from those of healthy mice, whereas mice that were treated with saline or CD24⁻CD133⁻ cells showed significantly higher levels of BUN in comparison with both CD24⁺CD133⁺ PEC–treated mice or healthy mice (Figure 9B).

Next, we investigated whether the improvement of renal function by CD24⁺CD133⁺ PEC treatment was also associated with a better preservation of renal structure. Assessment of renal histology on day 14 after injury provided clear evidence for tubular repair, although it was not uncommon to observe a small proportion of tubules with abnormal morphology, together with areas of tubulointerstitial and periglomerular fibrosis. Fibrotic areas therefore were quantified through direct measurement of green fluorescence area for α-SMA and TGF-β1 expression by image analysis. In Figure 9, C and D, representative confocal micrographs of a single optical section of kidney parenchyma from the two groups of mice are depicted. Kidney tissue from the CD24⁺CD133⁺-treated group was normal (Figure 9D). By contrast, kidneys from mice that were treated with glycerol and received saline (Figure 9C) were characterized by focal areas of interstitial and periglomerular fibrosis, as illustrated by α-SMA staining (Figure 9C). In mice that were treated with CD24⁺CD133⁺ cells, there was a significant decrease in α-SMA–stained tissue area (181.8 ± 54.4 versus 3124 ± 808.7 μm²; P < 0.001) and in TGF-β1–stained tissue area (1233.9 ± 205.3 versus 4608.9 ± 1311.6 μm²; P < 0.01) compared with mice that were treated with saline, as quantified by confocal microscopy.

Discussion

There is increasing evidence that cells that show at least multipotentiality (²⁴) and possibly pluripotentiality (²⁰) exist in different adult organs. We demonstrate here that in adult normal kidney, a subset of PEC in the Bowman’s capsule is the only cell type that shows coexpression of the SC markers CD24 and CD133 and of the SC-specific transcription factors Oct-4 and BmI-1 but a lack of lineage-specific markers. Differently from all other types of renal cells, CD24⁺CD133⁺ PEC also expressed CD106, a surface molecule that together with CD105, CD54, and CD44 is usually coexpressed by adult SC types that grow adherent, such as mesenchymal SC or multipotent adult progenitor cells (MAPCS) (^20,27). Purified CD24⁺CD133⁺ PEC could be obtained directly from outgrowth of isolated glomerular structures and exhibited high clonogenic efficiency and self-renewal potential. Moreover, under appropriate culture conditions, clones that were derived from single CD24⁺CD133⁺ PEC could be induced to differentiate into tubular epithelial cells that showed markers of cells from different portions of the nephron. Differentiation toward tubular cells resulted in the acquisition of high mRNA levels of markers that are characteristic of fully differentiated tubular epithelia, such as aminopeptidase A and Na/Gluc1, γ-GT, the Na/H exchanger, AQP1, AQP3, or the thiazide-sensitive Na/Cl transporter, with a prominent increase in AQP1 levels, consistently with the high proportions of cells in differentiated clones that acquire protein markers of proximal tubular cells. Importantly, cells that were derived from the same clones also could be differentiated in extrarenal cell types, such as adipocytes, osteoblasts, or cells that show phenotypic markers and functional properties of neurons, as it has already been shown for other types of adult human SC (^36,37). Taken together, these results strongly suggest that CD24⁺CD133⁺ PEC represent a population of multipotent progenitor cells. Accordingly, CD24⁺CD133⁺ but not CD24⁻CD133⁻ human renal cells engrafted into the kidney of SCID mice that had glycerol-induced ARF and also improved the morphologic and functional kidney damage.

To our knowledge, this is the first report to show that a population of resident renal cells of human origin ameliorate the structural recovery of the kidney after the induction of ARF and, more important, that they exert therapeutic effects on renal function. Acute tubular necrosis is the most common form of ARF and is considered a potentially reversible process. However, high percentages of patients (approximately 40%) fail to recover their renal function completely, and at discharge, they show mild to moderate renal failure (^38,39). It is interesting that recent follow-up studies also demonstrated that approximately 10% of these patients require renal replacement therapy for ESRD after 5 yr because of progressive renal fibrosis and chronic dysfunction (⁴⁰). Therefore, the observation that treatment of mice that were affected by acute tubular necrosis with CD24⁺CD133⁺ cells led to a complete recovery of renal function and to a significant reduction of renal fibrosis whereas control mice did not completely recover renal function and developed large areas of interstitial and periglomerular fibrosis is of potential clinical relevance. Functional protection by CD24⁺CD133⁺ cells probably is due to the capacity of these cells to engraft the damaged kidney and to integrate/differentiate within tubules, as shown by the demonstration that CD24⁺CD133⁺ but not CD24⁻CD133⁻ renal cells repopulated the tubule, exploiting their potential to generate tubular epithelial cells of different portions of the nephron.

One possible explanation for this phenomenon is that it is the result of a cell fusion. Recently, indeed, cell fusion between transplanted cells and recipient tissue has been claimed as an alternative novel mechanism to differentiation, which can occur in vivo and produce functional cells (^41–43). However, in other experimental systems, the cell fusion process has been excluded as a way to explain bone marrow SC plasticity (^44,45). Whether in our setting CD24⁺CD133⁺ cell–driven regeneration of tubular cells also might result from the fusion with resident cells cannot be ruled out completely, at least in vivo. However, our in vitro results strongly suggest that CD24⁺CD133⁺ PEC are plastic and acquire phenotypic and functional properties of renal and extrarenal cell types through differentiation and not through cell fusion, because multidifferentiation was achieved in clonal progenies that were derived from single CD24⁺CD133⁺ cells. Taken together, these results suggest that CD24⁺CD133⁺ cells represent a previously unidentified population of resident renal multipotent progenitors and thus can be named “adult parietal epithelial multipotent progenitors” (APEMP).

The results of this study not only demonstrate the existence of a population of renal multipotent progenitors within glomerular structures but also suggest that the urinary pole of the Bowman’s capsule may represent a renal SC niche (⁴⁶). This hypothesis is supported by the observation that when renal tubular cells are differentiated from embryoid bodies in the presence of nephrogenic factors and injected in developing kidney rudiments, they selectively localize to the glomerular/proximal tubule junction (⁴⁷). Another study suggested the possible existence of SC niche in the renal papilla (⁵). Given the complex embryologic origin of the kidney, it is possible that two distinct SC niches exist in the medulla and in the cortex and that proliferating amplifying renal cell progenitors likely localize to the proximal and distal tubules. In agreement with this hypothesis, possible tubular progenitors that express lineage markers and show limited differentiation potential (^4,5,9) were identified recently in normal kidneys at the level of proximal and distal tubular structures or at the interstitial level (^4–9).

The demonstration that APEMP represent a population of multipotent progenitors may provide an important contribution to the understanding of the pathogenesis of nephron loss. As known, the majority of diseases that progress to chronic renal failure start at the glomerular level, in the endocapillary compartment, where the inflammatory process involves the capillaries and/or the mesangium. As long as a glomerular disease remains restricted to the endocapillary compartment, restitution or repair is possible, even in the case of massive lesions. By contrast, spreading of the inflammation to the extracapillary compartment (Bowman’s space and Bowman’s capsule) results in dramatic kidney injury (^48,49). The glomerulus most likely dies, and the nephron is lost. On the basis of the results of this study, we suggest that these irreversible processes might reflect the loss of the renal SC niche that occurs only when the extracapillary compartment is affected, thus impairing the possibility of repair that may be provided by APEMP. By contrast, when the injury is limited to the endocapillary compartment and the SC niche is not affected, the glomerular damage can be repaired.

The identification of a subset of multipotent progenitors in the Bowman’s capsule also may provide an intriguing explanation for the genesis of crescents, which are known to reflect uncontrolled proliferation of PEC and their transdifferentiation into mesenchymal and myeloid cells during rapidly progressive glomerulonephritis (^48,49). We suggest that crescent formation might reflect a dysregulated activation of APEMP in response to chronic inflammatory stimulation. The nature of a multipotent progenitor of a subset of PEC also may provide a reasonable explanation for another renal disorder, such as embryonal hyperplasia of Bowman’s capsular epithelium (EHBCE) (⁵⁰). EHBCE is a lesion that occurs in kidneys of patients who are maintained on chronic dialysis, which consists of poorly differentiated cells that proliferate around sclerosed or obsolescent glomeruli (⁵⁰). EHBCE is considered a reversion of Bowman’s capsular PEC to the state of embryonic cell (⁵⁰). APEMP might represent such a previously unidentified population of embryonic progenitor-like cells. Taken together, the results of this study provide the first description of a multipotent progenitor cell in adult human glomeruli, thus opening new avenues for the development of autologous cell therapies in renal disorders.

Figure 1: Coexpression of the stem cell (SC) markers CD24 and CD133 identifies a subset of parietal epithelial cells (PEC) in the Bowman’s capsule of adult human kidney. (A) Double-label immunofluorescence showing expression of CD24 (red) and CD133 (green) by PEC in the Bowman’s capsule of an adult human kidney. Merged image (yellow) demonstrates coexpression of CD24 and CD133 by a subset of PEC localized at the urinary pole (UP; bar = 50 μm). To-pro-3 counterstains nuclei (blue). (B) High-power magnification of a double-label immunofluorescence showing expression of CD24 (red) and CD133 (green) by PEC. Merged image demonstrates co-localization of CD24 and CD133 in the cytoplasm and on the membrane of PEC facing the glomerulus (G), whereas only CD24 is expressed on the basal membrane of the cells (bar = 10 μm). To-pro-3 counterstains nuclei (blue). (C) CD133 detection with two different anti-CD133 mAb. Both 293C3 (red) and AC133 (green) mAb stain a subset of PEC in the Bowman’s capsule. Merged image demonstrates co-staining of the same cells (yellow; bar = 50 μm). To-pro-3 counterstains nuclei (blue). (D) Detection of CD24 (red), CD133 (green), and CD29 (blue) at kidney glomerular level. CD29 staining allows identification of the afferent arteriola (AA). Merged image shows that CD24 and CD133 selectively co-stain a subset of PEC localized opposite to the vascular pole (yellow; bar = 50 μm). (E) High-power magnification of a triple-label immunofluorescence showing expression by PEC of CD24 (red), CD133 (green), and CD106 (blue). Merged image demonstrates co-localization of CD24 and CD133 (yellow) in the cytoplasm and on the membrane of PEC facing the glomerulus, whereas CD24 and CD106 (purple) are coexpressed on the basal membrane. Apical membrane of PEC is indicated by the arrow. Areas of coexpression among CD24, CD133, and CD106 appear white (bar = 10 μm).

Figure 2: Isolation and characterization of CD24⁺CD133⁺ PEC. (A) Light microscopy image of cells outgrowing from seeded capsulated glomeruli. (B) Laser confocal microscopy demonstrates CD24 expression by all cells outgrowing from glomeruli (green). To-pro-3 counterstains nuclei (blue; bar = 100 μm). (C) Laser confocal microscopy demonstrates absence of green signal in all cells outgrowing from glomeruli when stained with an isotype-matched control antibody. To-pro-3 counterstains nuclei (blue; bar = 100 μm). (D) CD24⁺CD133⁺ PEC that were derived from glomerular outgrowth represent a homogeneous population that is composed of virtually 100% of cells that express CD24, CD133, CD106, CD105, and CD44, but all are negative for the endothelial markers CD31 and CD34. Flow cytometry analysis of a representative bulk culture is shown. (E) CD24⁺CD133⁺ PEC that were derived from glomerular outgrowth do not express the podocyte marker CD35, as assessed by flow cytometry. (F) Confocal microscopy demonstrates that CD24⁺CD133⁺ do not express the podocyte markers synaptopodin and WT-1 (bar = 100 μm). (G) Primary cultures of podocytes express high levels of synaptopodin at the cytoplasmic level (red) and WT-1 at the nuclear level (light blue; bar = 100 μm). (H) CD24⁺CD133⁺ PEC that were derived from glomerular outgrowth do not express the distal tubules/collecting ducts marker epithelial membrane antigen-1 (EMA-1), as assessed by flow cytometry. (I) CD24⁺CD133⁺ PEC that were derived from glomerular outgrowth lack the distal tubule marker Tamm-Horsfall glycoprotein (THG; left), as well as fluorescence staining for the proximal tubule markers Lotus Tetragonolobus lectin (LTA; middle; bar = 100 μm), as assessed by confocal microscopy. Negative histochemical staining for alkaline phosphatase (AP; right). A representative bulk culture is shown. (J) Comparison by quantitative reverse transcriptase–PCR (RT-PCR) of mRNA levels for markers of differentiated tubular cells in CD24⁺CD133⁺ versus CD24⁻CD133⁻ renal cells. Columns represent mean values ± SD as obtained from three different donors. Magnification, ×40 in A.

Figure 3: Growth properties of CD24⁺CD133⁺ cells and comparison with CD24⁻CD133⁻ cells. (A) Representative growth curves of the CD24⁺CD133⁺ cells (•) or CD24⁻CD133⁻ (□) cells that were obtained through immunomagnetic sorting from total renal cell suspensions. The results represent mean values ± SD of cell counts that were obtained in four different experiments from four different donors in the first 10 d of culture. (B) CD24⁺CD133⁺ cells in culture were expanded for 60 to 90 population doublings (PD) during a 4-mo period. Results are mean ± SD obtained from four different donors. (C) Flow cytometric analysis of DNA content performed on bulk cultures of CD24⁺CD133⁺ cells at 50 PD, demonstrating 100% diploid cells. One representative of four separate experiments is shown. (D) Assessment of mRNA levels for BmI-1 by real-time quantitative RT-PCR in cultures of human microvascular endothelial cells (HMVEC), human renal proximal tubular cells (HRPTEC), human mesangial cells (HMC), human glomerular visceral epithelial cells (HGVEC), human aortic smooth muscle cells (HASMC), CD24⁺CD133⁺ cells, CD24⁻CD133⁻ cells, and HEK cells. Results are expressed as mean ± SD of triplicate assessment in primary cultures from five different donors. (E) Assessment of mRNA levels for Oct-4 by real-time quantitative RT-PCR in cultures of HMVEC, HRPTEC, HMC, HGVEC, HASMC, CD24⁺CD133⁺ cells, CD24⁻CD133⁻ cells, and HEK cells. Results are expressed as mean ± SD of triplicate assessment in primary cultures from five different donors.

Figure 4: Characterization of single clones of CD24⁺CD133⁺ PEC. (A) Cloning efficiency of CD24⁺CD133⁺ cells in comparison with CD24⁻CD133⁻ cells. Results are expressed as mean ± SD of the percentage of clones over the number of single cells plated, as obtained from eight different donors (*P < 0.05). (B) Representative image of single-cell deposition (red circle, top left) obtained by limiting dilution of CD24⁺CD133⁺ PEC, followed by its proliferation and formation of a clone (subsequent panels). (C) Coexpression of CD24 and CD133 by 100% of cells that were derived from one representative clone of CD24⁺CD133⁺ PEC, as assessed by FACS analysis. (D) Coexpression of vimentin (red) and cytokeratin (green) by a representative clone of CD24⁺CD133⁺ PEC, as assessed by confocal microscopy (merged image, yellow). To-pro-3 counterstains nuclei (bar = 100 μm). Magnification, ×50 in B.

Figure 5: Differentiation of CD24⁺CD133⁺ PEC–derived clones into tubular epithelial cells. (A) Representative micrographs of histochemical staining for AP activity in CD24⁺CD133⁺ PEC before (day 0) and after (day 30) culture in tubular differentiation medium. (B) Staining for LTA before (day 0) and after (day 30) culture in tubular differentiation medium, as assessed by confocal microscopy (green). To-pro-3 counterstains nuclei (bar = 100 μm). (C) Expression of THG before (day 0) and after (day 30) culture in tubular differentiation medium, as assessed by confocal microscopy (green). To-pro-3 counterstains nuclei (bar = 100 μm). (D) Double labeling for LTA (green) and THG (red) showing coexistence in the same clone of cells that coexpress markers of multiple tubule segments (merged, yellow) and of cells that express only proximal or distal tubule markers (bar = 100 μm). One representative of five experiments is shown. (E) Assessment by quantitative RT-PCR of mRNA levels fold increase for tubular markers after 30 d of culture in tubular differentiation medium compared with values that were obtained in the same cells before differentiation. Columns represent mean values ± SD as obtained after differentiation of 50 different clones. (F, left) Representative micrographs of confocal fluo-4 fluorescence images recorded at 488-nm excitation before and after angiotensin II (1 μM) treatment (bar = 20 μm). (Right) Time course of change in fluorescence intensity recorded from five single cells from 10 different clones examined is shown. Magnifications: ×65 in A, left; ×80 in A, middle; ×320 in A, right.

Figure 6: Differentiation of CD24⁺CD133⁺ PEC–derived clones in osteoblasts and adipocytes. (A, left) Representative micrographs of histochemical staining for Alizarin red and AP before (day 0) and after (21 d) CD24⁺CD133⁺ PEC culture in osteogenic differentiation medium. (Right) Assessment of mRNA levels of Runx2 before (day 0) and after (21 d) culture in the same medium. Columns represent mean values ± SD obtained from 50 different clones. (B, left) Representative micrographs of histochemical staining for Oil Red-O before (day 0) and after (21 d) CD24⁺CD133⁺ PEC culture in adipogenic differentiation medium. (Inset) High-power magnification of some differentiated cells. (Right) Assessment of mRNA levels of adiponectin at 0 d and after 21 d of culture in the same medium. Columns represent mean values ± SD obtained from 50 different clones. Magnifications: ×100 in A; ×200 in B; ×320 in B, inset.

Figure 7: Acquisition by CD24⁺CD133⁺ PEC–derived clones of phenotypic and functional properties of neural cells. (A) Absence of the neural markers neurofilament 200 (NF200), neurofilament M (NFM), choline acetyl-transferase (ChAT), and microtubule-associated protein-2 (MAP-2) before culturing PEC in neurogenic differentiation medium, as assessed by confocal microscopy. To-pro-3 counterstains nuclei (bar = 100 μm). One representative clone is shown. (B) Strong expression of the neural markers NF200, NFM, ChAT, and MAP-2 after differentiation in the same medium (green). To-pro-3 counterstains nuclei (bar = 100 μm). One representative clone is shown. (C) High-power magnification of a representative image showing acquisition of a typical neuronal morphology and staining for ChAT (green) by CD24⁺CD133⁺ PEC cultured under neurogenic conditions (bar = 100 μm). (D) Assessment by real-time quantitative RT-PCR of mRNA levels fold increase of several neural markers after differentiation under neurogenic conditions compared with values that were obtained in the same cells before differentiation. Columns represent mean values ± SD obtained from 50 different clones. (E through H) Inward Ca²⁺ and Na⁺ currents in CD24⁺CD133⁺ PEC–derived neurons. Representative current traces recorded at a holding potential of −90 mV; 1-s step pulses from −80 to 50 mV were applied in 10-mV increments. Data were acquired with different sampling time (50 μs in the first 100 ms and 1 ms for the remaining duration of the test pulse) to highlight fast or slow phenomena. (E) Time course of L-type Ca²⁺ current (I_Ca); for clarity, only current traces that were recorded at −60, −40, −20, 0, 20, 30, and 40 mV are presented. (F) I_Ca-V curve determined at the current peak (n = 26). (G) Time course of Na⁺ current (I_Na); only current traces that were recorded at −60, −40, −30, −20, −10, 0, 20, and 30 mV are presented; red line is I_Na elicited at 0 mV in the presence of Tetrodotoxin (1 μM). (H) I_Na-V curve determined at the current peak (n = 26). (F and H) Continuous line superimposed through the data are the fitted Boltzmann function for activation: I_a(V) = G_max(V − V_rev)/{1 + exp[(V_a − V)/k_a]}, where G_max is the maximal conductance, V_rev is the apparent reversal potential, V_a is the potential that elicits the half-maximal increase in conductance and k_a is the slope factor. The best-fit parameters for I_Ca were G_max = 6 ± 1 nS, V_a = 0 ± 2 mV, k_a = 8.4 ± 2 mV, and V_rev = 79 ± 8 mV; those for I_Na were G_max = 28 ± 7 pS, V_a = −18 ± 2 mV, k_a = 6.0 ± 1 mV, and V_rev = 47 ± 4 mV. (I) I_Na inactivation evoked from holding potential of −90 mV; test pulse to 0 mV prepulsed from −90 to 30 mV in 10-mV increments, only traces without prepulse (−90 mV) and prepulsed at −70, −60, −30, −40, and −30 mV are depicted. (J) Normalized inactivation curve for I_Na; continuous line superimposed through the data are the fitted Boltzmann function for inactivation: I_h(V) = 1/{1 + exp[−(V_h − V)/k_h]}, where V_h is the potential eliciting the half-maximal current and k_h is the slope factor for inactivation. The best-fit parameters were V_h = −58 ± 5 mV and k_h = 6.0 ± 6 mV. For comparison, the curve reported on the right is that for activation.

Figure 8: Engraftment of CD24⁺CD133⁺ PEC in kidneys of SCID mice with acute renal failure (ARF) and generation of different types of renal tubular cells. (A) Light micrograph showing normal mouse renal tissue stained with hematoxylin and eosin (H&E; left) or with phalloidin (green, right; bar = 50 μm). (B) Tubulonecrotic injury observed after an intramuscular injection of glycerol, as assessed with H&E staining (left) or with phalloidin (right); the latter reveals the loss of brush border and the flattening of epithelial cells (green; bar = 50 μm). (C) Representative micrograph of kidney sections of control SCID mice that received injections of CD24⁻CD133⁻ cells and stained with LTA showing the absence of red-stained cells, as assessed by confocal microscopy (bar = 20 μm). (D) Representative micrograph of kidney sections of mice that had ARF and received injections of PKH26-labeled CD24⁺CD133⁺ PEC (red) and stained with LTA (green), as assessed by confocal microscopy. Small arrows point to multiple red cells. The larger arrow points to a proximal tubule (bar = 20 μm). (E) High-power magnification of the kidney section shown in D, which demonstrates regeneration of a proximal tubule structure (bar = 20 μm). (F) High-power magnification of another kidney section obtained from a SCID mouse that had ARF and received an injection of PKH26-labeled CD24⁺CD133⁺ PEC (red) and stained with Dolichos Biflorus Agglutinin (DBA) on the basal surface of two tubular structures (green), which demonstrates regeneration of a collecting duct structure (arrow). Other tubular structures that are stained with PKH26 but not with the collecting ducts marker DBA are visible (bar = 20 μm). (G) Double-label immunohistochemistry for cytokeratin (blue) and HLA-I human antigen (red) in kidneys of SCID mice with glycerol-induced ARF. (Left) Absence of red signal in tubules of a kidney section from a mouse that received an injection of CD24⁻CD133⁻ cells. (Middle and right) Human HLA class I–positive cells (red, arrows) in cytokeratin-expressing (blue) tubules of SCID mice with glycerol-induced ARF after injection of CD24⁺CD133⁺ PEC. (H) Detection of chromosome Y by the fluorescence in situ hybridization technique in control mice that received injections of saline solution (left) and in kidneys from female mice that received injections of CD24⁺CD133⁺ PEC from human men (red, middle and right; bar = 20 μm).

Figure 9: CD24⁺CD133⁺ cells protect glycerol-treated mice from renal structure and function deterioration. (A) Blood urea nitrogen (BUN) levels as measured in untreated (○) or in glycerol-treated mice that received saline (red-filled circle) or CD24⁺CD133⁺ cells (•). Arrows point to the days of injection of saline or CD24⁺CD133⁺ cells. Data are expressed as mean values ± SD. *P < 0.01 and **P < 0.001 versus glycerol+saline at the same time. (B) Comparison of BUN levels among healthy mice (□), mice that were treated with saline (light gray), mice that were treated with CD24+CD133+ cells (dark gray), and mice that were treated with CD24⁻CD133⁻ cells (▪) at day 14. Data are expressed as mean values ± SD. (C) Representative micrographs of kidneys that were treated with saline and stained for α-smooth muscle actin (α-SMA; green). Nuclei are stained with To-pro-3 (bar = 100 μm). (D) Representative micrographs of kidneys that were treated with CD24⁺CD133⁺ cells and stained for α-SMA (green). Nuclei are stained with To-pro-3 (bar = 100 μm).

This study was supported by the Tuscany Ministry of Health, by Ministero dell’Istruzione, dell’Università e della Ricerca (MIUR), and by the Research Institute “BIOAGROMED” of the University of Foggia. B.M. is the recipient of a Fondazione Italiana per la Ricerca sul Cancro (FIRC) fellowship.

We thank Melissa Poggesi for assistance.