Mouse Embryonic Stem Cell–Derived Embryoid Bodies Generate Progenitors That Integrate Long Term into Renal Proximal Tubules In Vivo
Cécile Vigneau*,,
Katalin Polgar*,
Gary Striker*,
Justine Elliott*,
Deborah Hyink*,
Odile Weber,
Hans-Joerg Fehling,
Gordon Keller,
Christopher Burrow* and
Patricia Wilson*
Departments of * Medicine and Gene and Cell Medicine, Mount Sinai School of Medicine, New York, New York; Department of Medicine, University Pierre and Marie Curie, Paris, France; and Department of Immunology, University of Ulm, Ulm, Germany
Address correspondence to: Dr. Patricia Wilson, Mount Sinai School of Medicine, Division of Nephrology, Department of Medicine, 1425 Madison Avenue, New York, NY 10029. Phone: 212-659-9383; Fax: 212-849-2434; E-mail: pat.wilson{at}mssm.edu
Received for publication October 2, 2006.
Accepted for publication March 18, 2007.
The metanephric kidney is a mesodermal organ that develops asa result of reciprocal interactions between the ureteric budand the blastema. The generation of embryonic stem (ES) cell–derivedprogenitors offers potential for regenerative therapies butis often limited by development of tumor formation. Becausebrachyury (T) denotes mesoderm specification, a mouse ES cellline with green fluorescence protein (GFP) knocked into thefunctional T locus as well as lacZ in the ROSA26 locus (LacZ/T/GFP)was used in cell selection and lineage tracing. In the absenceof leukemia inhibitory factor, mouse ES cells give rise to embryoidbodies that can differentiate into mesoderm. Culture conditionswere optimized (4 d, 10 ng/ml Activin-A) to generate maximalnumbers of renal progenitor populations identified by expressionof the specific combination of renal markers cadherin-11, WT-1,Pax-2, and Wnt-4. LacZ/T/GFP+ cells were further enriched byFACS selection. Five days after injection of LacZ/T/GFP+ cellsinto embryonic kidney explants in organ culture, -galactosidaseimmunohistochemistry showed incorporation into blastemal cellsof the nephrogenic zone. After a single injection into developinglive newborn mouse kidneys, co-localization studies showed thatthe LacZ/T/GFP+ cells were stably integrated into proximal tubuleswith normal morphology and normal polarization of alkaline phosphataseand aquaporin-1 for 7 mo, without teratoma formation. It isconcluded that defined differentiation of ES cells into embryoidbodies with Activin-A and selection for T expression providesa means to isolate and purify renal proximal tubular progenitorcells with the potential for safe use in regenerative therapies.
In metanephric kidney development, both the epithelial and themesenchymal inductive components are derived from the intermediatemesoderm. Onset of expression of Brachyury (T) marks the specificationof mesoderm at the time of gastrulation and is transiently expressedfrom embryonic day (E) 7 to 8.5 in the mouse intermediate andaxial mesoderm (1), whereas onset of Pax-2 expression is firstseen after gastrulation in the mediolateral mesoderm (2). Developmentof the mature mammalian kidney results from reciprocal signalingbetween the in-growing, branching ureteric bud tips and theundifferentiated metanephric mesenchyme that leads to the aggregationand condensation of renal epithelial progenitor cells followedby S-body intermediate formation and finally differentiationof the distinct glomerular and tubular epithelia of the nephron(3). The uninduced metanephric mesenchyme expresses cadherin-11,WT-1, and Pax-2 (4–6), whereas Wnt-4 is expressed in mesenchymalcondensates shortly after induction (7). Although there areno known specific single markers of tubular versus glomerularepithelial cell progenitors, expression of WT-1, cadherin-11,Pax-2, and Wnt-4 is a characteristic combination that can beused to define renal progenitor cells (8–10).
Embryonic stem (ES) cells that are isolated from blastocystsremain undifferentiated in culture in the presence of leukemiainhibitory factor (LIF) (11,12). When LIF is withdrawn, ES cellsgive rise to embryoid bodies (EB) that can be induced to differentiateinto several different cells types of mesodermal, endodermal,and neuroectodermal lineages, including those that specify hematopoietic,cardiomyocyte, pancreatic, hepatocyte, or neuronal fates (13–19).Although undifferentiated ES cells are multipotent and thusmight offer important potential for application in regenerativetherapies, a significant drawback of their use in vivo is thecommon development of proliferative and inflammatory abnormalities,including teratoma development (11).
Several strategies have been used to identify renal stem cells,including the use of fetal renal progenitors (20) and more recentlyES cells (21–24). Although undifferentiated ES cells thatwere injected into ex vivo mouse embryonic kidneys and culturedfor short periods localized to tubular epithelial structures(22,24), injection of EB-derived cells that overexpressed Wnt-4,under the renal capsule in vivo, formed teratomas (23). Previousstudies showed that differentiation of ES/EB cells in specificconcentrations of growth factors such as Activin-A or bone morphogeneticprotein-7 (BMP-7) can induce endodermal or mesodermal lineagecommitment in vitro (16,22).
For application for regenerative therapies, it is essentialto develop methods to derive specific renal epithelial progenitorsthat integrate into normal nephron segmental epithelia overlong periods of time, without the development of tumors or otherabnormalities. In these studies, we used a novel ES cell linewith a selectable marker to show that the combination of preciselycontrolled differentiation and selection results in the isolationand enrichment of specific renal epithelial progenitors thatare capable of long-term integration into stable, normally differentiatedproximal tubules after a single injection into live mice invivo.
Generation and Culture of LacZ/T/Green Fluorescence Protein ES Cell Lines
The generation of T/green fluorescence protein (T/GFP) "knock-in"ES cells has been described previously (25). A lacZ-neo fusiongene was inserted into the ROSA26 locus of T/GFP ES cells byhomologous recombination ("gene targeting"). The lacZ-neo expressioncassette was derived from plasmid pSA-geo (26) and consistedof an adenovirus splice acceptor sequence, a -galactosidasegene fused to the neomycin resistance gene (-geo), and a polyadenylationsite from the bovine growth hormone gene. This cassette wascloned into the unique XbaI site of targeting vector pROSA26-1(27), in which we had replaced the diphtheria toxin A negativeselection cassette by the thymidine kinase gene, giving riseto the final targeting vector pOW-AA. (The complete sequenceof pOW-AA can be obtained from H.J.F. upon request.) PlasmidspSA-geo, pROSA26-1, and pROSA-5' (carrying an external probeto confirm correct homologous recombination events) were providedby Dr. Soriano (Fred Hutchinson Cancer Research Center, Seattle,WA). Of 80 neomycin-resistant T/GFP ES colonies, 24 had correctlyinserted the LacZ-neo cassette in one ROSA26 allele, givingrise to LacZ/T/GFP ES cell lines.
These cells were maintained on gelatin-coated dishes in feeder-free,serum-free (FFSF) medium that contained 50% neurobasal supplementedwith N2/50% DMEM/F12 supplemented with B27, 10% BSA, 2 mM glutamine,and 50 U/ml penicillin/streptomycin (all Life Technologies,Rockville, MD), 4.5 x 104 M monothioglycerol (Sigma, St. Louis,MO), 1% LIF, and 10 ng/ml human recombinant BMP-4 (R&D Systems,Minneapolis, MN). For generation of EB, ES cells were dissociatedwith 0.25% trypsin-EDTA (Cellgro; MediaTek, Indianapolis, IN)and 105 were plated per 60-mm Petri-dish (VWR, West Chester,PA) in differentiation medium that contained 50% neurobasal,50% DMEM/F12, 2 mM glutamine, 50 U/ml penicillin/streptomycin,4.5 x 104 M monothioglycerol, 5 µl/ml transferrin (Roche,Nutley, NJ), and 10 µl/ml ascorbic acid (Sigma). EB cellswere cultured for 1 to 8 d in the presence or absence of 2 to100 ng/ml Activin-A (R&D Systems).
Flow Cytometry and Cell Sorting
EB were harvested after dissociation with 0.25% trypsin for3 min and passed through a 20-G needle before resuspension in(DMEM + 5% FBS (Gemcell, Pennant Hills, Australia). A totalof 1 to 2 x 106 cells/ml were analyzed for GFP expression usinga FACScan (Becton Dickinson, Franklin Lakes, NJ) or were sortedusing a MoFlo cell sorter (Cytomation Systems, Glostrup, Denmark;in Mount Sinai School of Medicine shared facility) to generatecells for PCR analysis or injection into mouse kidneys.
RNA Extraction and PCR
RNA was extracted (RNAeasy minikit; Qiagen, Germantown, MD),treated with DNAse, and reverse-transcribed (Sensiscript; Qiagen),and PCR was performed (Hot Star Taqmaster mix; Qiagen) usingvarious primers (Table 1) under the following conditions: 95°Cdenaturation for 15 min, 30 to 40 amplification cycles (94°Cdenaturation for 1 min, 60°C annealing for 1 min, 72°Celongation for 1 min), and 72°C for 10 min. PCR productswere analyzed on 2% agarose gels. cDNA from E13 mouse metanephrickidneys or adult kidneys were used as positive controls. Genomiccontamination of RNA samples was ruled out by omission transcriptaseor by the use of water alone.
Table 1. Reverse transcription–PCR primer sequences and conditions
Embryonic Mouse Kidney Organ Culture and Microinjection
E11.5 kidneys were dissected from outbred CD-1 mice (CharlesRivers Laboratories, Wilmington, MA) using a dissecting stereomicroscope(Olympus, Center Valley, PA), and paired kidneys were placedon 24-mm Transwell Clear membrane inserts, 0.4-mm pore size,in six-well cluster plates (Corning, Corning, NY). A total of1 ml of supplemented serum-free organ culture medium (50% DMEM,50% Ham F12, 15 mM HEPES, 2 mM l-glutamine [Cellgro, Mediatek],4.5 g/L glucose, 45 mM sodium bicarbonate, and 1x insulin/transferrin/selenium[MediaTek]) was added to the lower chamber, allowing a thinfilm to cover the apical surface of the explant. Kidney explantswere injected 1 to 2 h after dissection using an automatic microinjectorattached to a Zeiss inverted microscope (Oberkochen, Germany;injection pressure of 100 hPa, compensation pressure of 10 hPa,injection time 0.5 s). Fifteen metanephric kidneys were injectedat two sites with 150 LacZ/T/GFP+ cells, and 10 metanephrickidneys were injected with 150 LacZ/T/GFP– cells. In eachexperiment, the contralateral control kidney rudiment was eitheruninjected (n = 19) or sham injected with PBS (n = 6). All kidneyrudiments were incubated in a tissue culture incubator (5% CO2)for 4 d with daily medium changes.
Live Newborn Mouse Kidney Injection In Vivo
Eighty-five Newborn (2 to 5 d old) mice were anesthetized for10 s by isofluorane before percutaneous injection of 30,000cells into the left kidney through a 26-G needle; right kidneys(controls) were uninjected. Thirty-three mice were administeredan injection of LacZ/T/GFP+ cells, 32 mice were administeredan injection of with LacZ/T/GFP– cells, and 20 mice wereadministered an injection of undifferentiated LacZ/ES cells.All mice survived the procedure and were observed daily beforebeing killed by CO2 after 4 h; 1, 5, 15, 30, and 90 d; and 6and 7 mo according to the regulations of the Animal Care Committeeof Mount Sinai School of Medicine.
Morphologic Analysis
Embryonic metanephric kidney explants and newborn kidneys werefixed in 4% paraformaldehyde (PFA; EM Sciences, Pasadena, CA)in PBS (pH 7.4) that contained calcium and magnesium for 4 hand rinsed in PBS three times for 5 min before embedding inparaffin. Five-micron sections were cut, deparaffinized, rehydrated,and stained with Harris's Hematoxylin and Eosin (Sigma); dehydratedin graded ethanols and xylene; mounted in Vectamount (VectorLaboratories, Burlingame, CA); and viewed under a Nikon FXA-Mikrophotmicroscope using bright-field illumination (Tokyo, Japan). Allsections were blinded and viewed independently by three investigatorsand scored for pathologic characteristics, including glomerularand tubular integrity, presence or absence of interstitial fibrosis,inflammatory reaction, and tumor or other lesions.
Immunolocalization
For -galactosidase staining, 5-µm sections were deparaffinized,rehydrated through a graded series of ethanols, and incubatedfor 30 min in 7% H2O2aq to block endogenous peroxidase activityand in 10% normal goat serum in PBS overnight. Sections werethen coated with rabbit anti–Escherichia coli-galactosidaseantibody (MP Biomedicals, Solon, OH) diluted 1:500 in PBS thatcontained 2% BSA for 2 h, washed twice for 5 min in 0.01% PBS-Tween20 and once for 5 min in PBS, incubated for 45 min with biotinylatedgoat anti-rabbit IgG (Vectastain; Vector Laboratories), washedtwice for 5 min in PBS-Tween and once in PBS, incubated for45 min with avidin-biotin peroxidase, washed three times for5 min in 0.05 M Tris-buffered saline, and color-developed for30 min using red aminoethylcarbazole as substrate (Vector Laboratories).
For double immunostaining, after the rehydration and blockingsteps, 5-µm sections were incubated with rabbit anti–aquaporin-1or aquaporin-2 antibodies (Alpha Diagnostic, Caguas, PuertoRico; 1:100) for 45 min, washed, incubated for 45 min with biotinylatedgoat anti-rabbit IgG, washed, incubated for 45 min with avidin-biotinperoxidase, washed, and color-developed for 10 min using bluechromogen as substrate (Vector Laboratories). After this firststaining, slides were processed for -galactosidase localizationas described previously. For alkaline phosphatase double staining,a 10-min incubation was carried out immediately after the rehydrationstep using the enzymatic substrate 5-bromo-4-chloro-3-indolylphosphate/nitro-blue tetrazolium (BCIP/NBT; Vector Laboratories)before blocking and anti–-galactosidase immunohistochemistry.
Cytospin-collected EB-derived progenitor cells were fixed for5 min in 4% PFA at room temperature or in ice-cold methanoland washed three times for 5 min in PBS, and some PFA-fixedcells were permeabilized with 0.01% Triton. After blocking in3% BSA for 20 min, the slides were incubated with one or twoof the following primary antibodies: Goat anti–Wnt-4 (1:100;Abcam, Cambridge, MA), mouse anti–WT-1 (1/100; Santa CruzBiotechnology, Santa Cruz, CA), rabbit anti–Pax-2 (1:100;Covance, Princeton, NJ), and goat anti–cadherin-11 (1:100;Santa Cruz Biotechnology) for 45 min. After three 5-min washesin PBS, slides were incubated with one or more of the followingsecondary fluorophores: Donkey anti-mouse Alexa-488 (1:100;Molecular Probes, Eugene, OR), donkey anti-rabbit Alexa-594(1:100; Molecular Probes), and donkey anti-goat Alexa-568 (1:100;Molecular Probes). In some cases, DAPI (1:5000; Molecular Probes)or TO-PRO-3 (1:5000; Molecular Probes) was added to the secondaryantibody incubation step. After three 5-min washes in PBS, slideswere mounted with Vectamount and fluorescence was viewed undera Zeiss Axioplan-2 Microscope (Mount Sinai School of Medicineshared facility). Immunohistochemistry sections were examinedusing a Nikon FXA-Microphot microscope equipped with Nomarskioptics.
Optimization of Culture Conditions for Generation of EB-Derived Renal Progenitor Cells and Enrichment by FACS Selection
The mouse LacZ/T/GFP ES cell line that is maintained in culturein an undifferentiated state by culture on gelatin-coated platesin FFSF medium that contained LIF and BMP-4 were converted tonumerous, small spherical EB by the removal of LIF and BMP-4.Because mesoderm is specified by T expression (1), FACS analysiswas carried out to identify the time point of maximal GFP/Texpression in EB after various lengths of time in culture. Figure 1Ashows a sharp peak of GFP expression after 4 d of culture. Inlight of studies showing that specific concentrations of Activin-Acan direct lineage-specific differentiation of EB in culture(16), FACS analysis for T/GFP was carried out on LacZ/T/GFPEB that were cultured for 4 d in the presence of 2 to 100 ng/mlActivin-A. Figure 2B shows maximal GFP expression in the presenceof 10 ng/ml Activin-A under FFSF conditions. Reverse transcription–PCR(RT-PCR) analysis was then used to determine the gene expressionprofiles of these optimized enriched populations of EB-derivedcells. Figure 1C confirms maximal expression of T after 4 din culture in the presence of 10 ng/ml Activin-A, as well asthe expression of renal progenitor markers including cadherin-11,Wnt-4, WT-1, and Pax-2. However, these cells also expressedPax-6, a marker of neuro-ectoderm. For further enrichment forrenal progenitors, FACS selection of T/GFP was carried out andLacZ/T/GFP+ and LacZ/T/GFP– populations were compared(Figure 1D). RT-PCR expression analysis confirmed that withoutGFP/FACS selection, the EB-derived cell population expressedcadherin-11, Wnt-4, WT-1, Pax-2, Oct-4, CD24, CD133, and Pax-6and showed that cell populations that were treated with 10 ng/mlActivin-A also expressed T. Importantly, selection of LacZ/T/GFP+cells by FACS allowed for the enrichment of populations thatexpressed renal progenitor markers but not Pax-6 (Figure 1D,lane 3). CD133 was also localized to differentiated proximaltubules in normal E17 fetal kidneys (Supplementary Figure 1).By contrast, FACS-selected Lac/Z/T/GFP– cell populationsexpressed Pax-6; high levels of Pax-2; and low levels of Oct-4,CD24 and CD133, cadherin-11, Wnt-4, and WT-1 (Figure 1D, lane4). Double immunofluorescence studies showed that both T/GFP+and T/GFP– cells contained combinations of the renal progenitorcell markers WT-1, Pax-2, and Wnt-4 (Figure 1E). By comparisonwith DAPI or TO-PRO-3 universal nuclear staining, it was concludedthat 85% of T/GFP+ cells were Pax-2/WT-1/Wnt-4 positive. Theseresults suggested that FACS selection of T+ cells from EB cellsthat were cultured for 4 d in the presence of 10 ng/ml Actvin-Aeffected a significant enrichment of a renal progenitor cellpopulation. By contrast, FACS-selected T cells that were culturedunder the same conditions, although not homogeneous, were enrichedfor renal progenitors that contained high levels of Pax-2 expression.These isolation and enrichment procedures were highly reproducible,showing consistent results in multiple experiments.
Figure 1. (A) FACS analysis of brachyury/green fluorescence protein (T/GFP) expression: Effect of days in culture. LacZ/T/GFP embryoid bodies (EB) were differentiated by removal of leukemia inhibitory factor (LIF) in feeder-free, serum-free conditions in the absence (dotted line) or presence of 10 ng/ml Activin-A (solid line) and cultured for 0 to 6 d. Maximal expression of T/GFP was seen after 4 d of culture. (B) FACS analysis of T/GFP expression: Dosage response of Activin-A. EB were cultured for 4 d in the presence of 2 to 100 ng/ml Activin-A. Maximal expression of T/GFP was seen after incubation in 10 ng/ml Activin-A. (C) Reverse transcription–PCR (RT-PCR) analysis of EB: Effects of days in culture. EB-derived cells that were cultured in the presence of 10 ng/ml show maximal expression of T on day 4. Cadherin-11, Wnt-4, WT-1, Pax-2, and Pax-6 were also expressed. (D) RT-PCR analysis of EB-derived cells before and after FACS selection for T/GFP. EB were cultured for 4 d in the absence (lane 1) or presence (lanes 2 to 4) of 10 ng/ml Activin-A. Some LacZ/T/GFP– EB that were treated with 10 ng/ml Activin-A for 4 d were harvested and selected by FACS into T/GFP+ (lane 3) and T/GFP– (lane 4) populations. All cells were analyzed by RT-PCR. Note the absence of Pax-6 from LacZ/T/GFP+ sorted cells in lane 3. (E) Immunofluorescence analysis of FACS-selected LacZ/T/GFP+ and LacZ/T/GFP– cells. Both populations express WT-1 (left, green), Pax-2 (middle, cyan), and Wnt-4 (right, red). Note co-localization of cytoplasmic/membrane Wnt-4 with nuclear Pax-2 and WT-1.
Figure 2. Localization of FACS-selected EB-derived LacZ/T/GFP+ and LacZ/T/GFP– cells in embryonic day 11.5 (E11.5) mouse kidneys after 4 d in organ culture. A total of 150 cells were injected at two injection sites per kidney on day 0. Avidin-biotin--galactosidase immunohistochemistry of sectioned rudiments. Aminoethylcarbazole (AEC) substrate, red reaction product. (A) LacZ/T/GFP+ cells are seen in the cytoplasm of small undifferentiated blastemal cells in the nephrogenic zone (arrows). The ureteric bud (arrowhead) is unstained. (B) LacZ/T/GFP+ cells are seen (red reaction product) in the cytoplasm of blastemal cells and pretubular aggregates (arrow) adjacent to the tips of ureteric bud (stained light blue). (C) -Galactosidase staining (red) in the cytoplasm of proliferating cell nuclear antigen–positive cells with dark blue nuclei (arrow) adjacent to the tip of the ureteric bud (light blue cytoplasm). (D) No staining was seen in uninjected contralateral kidneys. (E) CD133-positive cells (red) in cells at the tip of the ureteric bud and in a pretubular aggregate (arrows). (F) LacZ/T/GFP– cells are seen in the ureteric bud epithelium of fetal kidneys (arrowheads). Magnifications: x40 in A, D, and F (Nomarski illumination); x60 in B (bright-field illumination) and C; x80 in E (Nomarski illumination).
Incorporation of EB-Derived Renal Progenitors into Mouse E11.5 Kidneys in Organ Culture
At E11.5, the ureteric bud has invaded the undifferentiatedmetanephric mesenchyme and undergone one to two bifurcations.These explants when grown in organ culture in specific supplemented,serum-free conditions undergo normal renal glomerular and tubuleepithelial differentiation and morphogenesis after 5 to 10 d(28). These ex vivo preparations were previously used to studythe effects of gene manipulation by microinjection (29). Fordetermination of whether optimized EB-derived, FACS-selectedLac-Z/T/GFP+ or Lac-Z/T/GFP– cells could be incorporatedinto early metanephric progenitor cell pools, E11.5 mouse embryonickidneys were injected on day 0 with 150 cells at two sites,and their fates were determined after 4 d of organ culture byimmunodetection of the distribution of LacZ-encoded E. coli-galactosidase (Figure 2). After 4 d of organ culture, E11.5mouse metanephroi show some glomerular differentiation but noclear proximal tubule maturation. Four days after microinjectionof FACS-selected LacZ/T/GFP+ cells into these E11.5 metanephricexplants, incorporation was seen exclusively in the cytoplasmof small cells in the cortical nephrogenic zone, consistentwith the previously reported morphology and distribution patternsof proliferating cell nuclear antigen (PCNA)-positive, cytokeratin-negativeblastemal cells at the highly proliferative cortical rim ofthe kidney (Supplementary Figure 2) as well as at the tips ofthe ureteric bud, where condensation of epithelial precursorsto form pretubular aggregates takes place (5–7,20) (Figure 2,A through C). No staining was seen in the ureteric bud (Figure 2A,arrowhead), and no LacZ/T/GFP+ cells were ever detected in uretericbud–derived structures. Co-distribution with PCNA-positive,CD133-positive cells (Figure 2, C and E) further suggested incorporationof LacZ/T/GFP+ cells into progenitor cells of a proximal tubulelineage. By contrast, 3 to 4 d after microinjection of FACS-selectedLacZ/T/GFP– cells, the majority of staining was seen inthe epithelial cells of the ureteric bud (Figure 2F), consistentwith the RT-PCR and immunostaining evidence that isolated LacZ/T/GFP–cells express the ureteric bud markers Pax-2 and L1-CAM (Figure 1,D and E; data not shown). Occasional localization of LacZ/T/GFP–cells in the blastema and low levels of expression of additionalearly progenitor cell markers (Figures 1, D and E, and 2F) mightreflect delamination of ureteric bud progenitor cells, but additionalstudies would be needed for confirmation of this conclusion.Contralateral uninjected (Figure 2D) or sham-injected E11.5kidneys showed no reaction product after 4 d of culture (datanot shown).
LacZ/T/GFP+ Progenitors Integrate into Renal Proximal Tubules of Newborn Mice and Are Detectable for 7 Months
Because nephron induction is not complete until 2 wk after birthin mice, it was reasoned that renal progenitors that are injecteddirectly into the kidneys of live newborn mice in vivo mightintegrate into the newly forming metanephric kidney. This newtechnique allows long-term lineage tracing of LacZ-tagged cellsby -galactosidase immunostaining. After a single percutaneousinjection of 30,000 FACS-selected LacZ/T/GFP+ cells into theleft kidneys of immunocompetent CD1 mouse pups (2 to 5 d old,n = 33), groups of mice were killed and analyzed by -galactosidaseimmunohistochemistry from 1 d through 7 mo after injection (Figure 3).Twenty-four hours after injection, the majority of -galactosidaseimmunoreaction product (red staining) was seen in the interstitium(Figures 3A and 4A), whereas by 15 d, it was exclusively localizedto some cells incorporated into cortical proximal tubules (Figures 3,B, C, G, and H, and 4B). Similar patterns of proximal tubuleincorporation were seen using anti-GFP antibodies (SupplementaryFigure 3). Reaction product was similarly detected exclusivelyin proximal tubules of the cortex and outer medulla 90 d (Figures 3,D and I, and 4C) and 6 to 7 mo (Figures 3E and 4D) after injection.Lack of staining in uninjected kidneys confirmed that the stainingwas not due to endogenous renal -galactosidase (Figure 3F).This confirmed that the monospecific antibody recognized onlythe E. coli-galactosidase, without cross-reactivity, thus allowingdiscrimination between endogenous and engineered cells. In additionto -galactosidase immunohistochemistry, real-time PCR of LacZexpression was used to quantify incorporation in >80% ofinjected kidneys (Supplementary Figure 4).
Figure 3. Distribution of T/GFP+ cells after injection into kidneys of live newborn mice. Newborn CD1 mouse pups were lightly anesthetized, and 30,000 FACS-selected LacZ/T/GFP+ cells were percutaneously injected directly into the left kidney. Avidin-biotin -galactosidase immunohistochemistry of sectioned kidneys, AEC substrate, red reaction product. (A) One day after injection, -galactosidase–positive cells (arrows) were present in the interstitium. (B) Fifteen days after injection, -galactosidase–positive cells were present in cortical proximal tubules. (C) High-power image of cytoplasmic distribution of -galactosidase–positive proximal tubule cells (P) of the outer cortex. (D) Ninety days after injection, -galactosidase–positive cells were present in cortical proximal tubules. (E) Six months after injection, -galactosidase–positive cells were present in corticomedullary proximal tubules. (F) No staining was seen in an uninjected kidney. (G) High-power image of -galactosidase–positive proximal convoluted tubule, 15 d after injection. (H) High-power image of -galactosidase–positive cortical proximal tubules (S1 and S2 segments), 15 d after injection. (I) High-power image of -galactosidase positive proximal straight tubule (S3 segment), 7 mo after injection. Magnifications: x10 in A and F (bright-field illumination); x20 in B (bright-field illumination); x60 in C (Nomarski illumination); x40 in D and E (bright-field illumination); x100 in G through I (Nomarski illumination).
Figure 4. Co- distribution of LacZ/T/GFP+ cells with alkaline phosphatase after injection into kidneys of live newborn mice. Injection conditions are as in legend to Figure 3. Avidin-biotin -galactosidase immunohistochemistry of sectioned kidneys, AEC substrate, red reaction product. 4-Chloro-3-indolyl phosphate/nitro-blue tetrazolium enzyme reaction for alkaline phosphatase: Blue reaction product. (A) One day after injection, -galactosidase–positive cells (red) were present in the interstitium. The majority did not co-localize with proximal tubules that express alkaline phosphatase (blue). (B) Fifteen days after injection, some -galactosidase–positive cells (red) were present in proximal tubules, stained blue at the brush border. No red staining was seen in glomeruli (G), cortical collecting tubules (C), or thick ascending limbs of Henle's loop (T). Arrow denotes the normal connection between the glomerulus and proximal convoluted tubule (C) Ninety days after injection, -galactosidase–positive (red) cytoplasm was present in some proximal tubules in the renal cortex stained blue for alkaline phosphatase at the brush border. Collecting tubules (C) were unstained. (D) Seven months after injection, -galactosidase–positive cells (red) were present in the outer medulla in longitudinal profiles of some proximal straight tubules stained blue for alkaline phosphatase. (E) High-power image of junction of proximal convoluted tubule, 15 d after injection, containing -galactosidase–positive cells (red) and its glomerulus (arrow). Note that glomerulus is unstained. A collecting tubule (C) is also unstained. (F) High-power image of two proximal convoluted tubules, 15 d after injection, containing many -galactosidase–positive cells (red) and apical membrane alkaline phosphatase (blue). Magnifications: x10 in A (bright-field illumination); x20 in B (bright-field illumination); x40 in C (bright-field illumination) and D (Nomarski illumination); x60 in E (bright-field illumination); x100 in F (bright-field illumination).
To further confirm that the site of incorporation of the selectedLacZ/T/GFP+ cells was indeed proximal tubules, we conductedco-localization studies. Figure 4 confirms a similar time scaleof incorporation in that after 1 d, no co-localization withalkaline phosphatase was seen, consistent with initial distributionof LacZ/T/GFP+ cells into the interstitium (Figure 4A). However,after 15 d, all incorporated -galactosidase–positive (red)cells were seen in alkaline phosphatase–positive (blue)proximal tubule profiles (Figure 4B). This was also seen after90 d (Figure 4C) and 7 mo (Figure 4D). It was noted, however,that the numbers of -galactosidase–positive cells withinan individual proximal tubule profile ranged from 0 to many(Figures 3 through 5). No LacZ/T/GFP+ cells were seen in glomeruli,collecting tubules, or thick ascending limbs of Henle. Importantly,the morphology of the proximal tubules with and without integrated-galactosidase–positive cells seemed to have normal connectionswith glomeruli (Figure 4E, arrow) or narrow or closed lumens,which are characteristic of normal proximal tubules with morphologicallyintact brush borders and exclusively apically localized alkalinephosphatase reaction product (Figure 4F). Exclusive integrationof FACS-selected LacZ/T/GFP+ cells into normally differentiatedstable proximal tubules was further substantiated by paralleldouble-localization studies with the proximal tubule water channelaquaporin-1 and the proximal tubule lineage marker CD133 (Figure 5).
Figure 5. Co-distribution of LacZ/T/GFP+ cells with aquaporin-1 after injection into kidneys of live newborn mice. Injection conditions are as in legend to Figure 3. Avidin-biotin -galactosidase immunohistochemistry: Red reaction product. Aquaporin-1 immunohistochemistry: Blue reaction product. (A) One day after injection, -galactosidase–positive cells were present in the interstitium. No co-localization with aquaporin-1–positive proximal tubules or thin descending limbs of Henle's loop stained blue. (B) Ninety days after injection, a -galactosidase–positive (red) cell (arrow) is seen in a proximal tubule with apical and basal membrane staining of aquaporin-1 (blue). Note that aquaporin-1–negative tubules (bottom left) do not stain for -galactosidase. (C) Co-localization of -galactosidase (red) and CD133 (blue) in proximal tubules 7 mo after injection. Magnifications: x10 in A (bright-field illumination); x60 in B (bright-field illumination); x40 in C (Nomarski illumination).
LacZ/T/GFP– Progenitors Integrate into Ureteric Bud–Derived Collecting Tubules
A similar set of studies were carried out in newborn mice toanalyze the fate of optimized EB-derived LacZ/T/GFP– cellsafter a single injection into the kidneys of live newborn mice(Figure 6). Specific -galactosidase immunolocalization indicatedLacZ/T/GFP– integration into distal tubular structuresof the cortex (Figure 6A), into collecting tubule epitheliaof the outer medulla (Figure 6B), and especially at the tipof the papilla (Figure 6C). The urothelium was also positive(Figure 6C). No glomerular staining was seen (Figure 6A).
Figure 6. Distribution of LacZ/T/GFP– cells after injection into kidneys of live newborn mice. Newborn CD1 mouse pups were lightly anesthetized, and 30,000 FACS-selected LacZ/T/GFP+ cells were percutaneously injected directly into the left kidney. Avidin-biotin -galactosidase immunohistochemistry of sectioned kidneys, AEC substrate, red reaction product. (A) -Galactosidase–positive cells were seen in cortical distal convoluted tubules. No staining was seen in glomeruli (G). (B) -Galactosidase–positive cells are seen in medullary collecting ducts (C) -Galactosidase–positive cells are seen in collecting ducts at the tip of the papilla and in the urothelium. (D) No staining is seen in contralateral uninjected kidneys. Magnification, x20 (Nomarski illumination).
No Teratoma Formation 7 Months after Injection of Selected LacZ/T/GFP+ Proximal Tubule Progenitors
Even though large numbers (n = 33) of immunocompetent CD-1 micewere administered an injection of LacZ/T/GFP+ EB-derived cellsduring a highly proliferative phase of renal development, noteratomas or inflammatory reactions were seen (Table 2). Thiscontrasted with the results that were seen after injection ofthe less homogeneous LacZ/T/GFP– EB-derived cell populations,in which four (12.5%) of 32 mice, from day 15 to 7 mo afterinjection, developed teratomas (Table 2 and Figure 7A) thatwere -galactosidase positive (Figure 7B). The injection of undifferentiatedLacZ/ES cells resulted in even higher levels of teratoma formationin five (20%) of 20 newborn mice as early as 9 d after injection(Table 2 and Figure 7, C and D).
Figure 7. Hematoxylin and eosin (H&E) staining and -galactosidase immunohistochemistry of teratomas. Avidin-biotin -galactosidase immunohistochemistry of sectioned kidneys, AEC substrate, red reaction product. (A) H&E staining of teratoma 20 d after injection of LacZ/T/GFP– cells. (B) -Galactosidase–positive (red) cells in a teratoma 20 d after injection of LacZ/T/GFP– cells (C) H&E staining of teratoma 9 d after injection of undifferentiated LacZ embryonic stem (ES) cells. (D) -Galactosidase–positive (red) cells in a teratoma 9 d after injection of undifferentiated LacZ ES cells. Magnification, x20 (bright-field illumination).
Because kidney failure is a major cause of morbidity and mortalityand demand far outstrips the supply of kidneys that are availablefor transplant, an important goal in renal medicine is to developcell replacement therapies. ES cells offer a strong potential,but safety concerns have been raised because of their high propensityto develop proliferative abnormalities, including tumors. Here,we reasoned that the generation and use of committed renal progenitorcells could offer a safer option than multipotent stem cells.To that end, we used ES cells that expressed a selectable mesodermalmarker to facilitate selection and enrichment of specific renalprogenitor cell populations after optimized limited differentiationinto EB with specific growth factors.
T/GFP was chosen as the selectable marker because it specificallymarks the mesoderm at gastrulation and is expressed during mesodermalspecification, including intermediate mesoderm from which thekidney is derived. Members of the TGF- superfamily BMP-4 andActivin are known to play important roles in mesoderm development(2,30) and to be temporally associated with the gradual downregulationof T and upregulation of Pax-2 during intermediate mesodermpatterning. This reciprocal pattern of high T and low Pax-2expression is seen in unselected EB that are cultured for 4d (Figure 1C). Because Kubo et al. (16) had shown that specificconcentrations of Activin-A can drive mouse EB differentiationdown specific cell lineages, we carried out Activin-A dosageresponse analysis coupled with RT-PCR gene expression analysisof characteristic renal progenitor markers. Although there isno single specific marker of a renal progenitor cell, a combinationof specific markers can be used, including the mesenchymal-specificcadherin-11 (4), WT-1 (5), and Pax-2 (6), all of which are expressedin the uninduced metanephric mesenchyme, as well as Wnt-4 (7),which is expressed in the induced mesenchyme. It is noteworthythat Pax-2 is also highly expressed in the developing uretericbud as well as induced metanephric condensates, and WT-1 isexpressed in differentiating glomerular podocytes.
It is important to note in this study that all ES cell maintenanceand EB cell differentiation was carried out in the absence ofill-defined feeder or serum factors. This probably contributedsignificantly to the high reproducibility of the isolationsand enrichments observed. Optimization of renal progenitor generationwas obtained by the simple manipulations of the length of timein culture (4 d) and addition of a specific concentration (10ng/ml) of Activin-A. Such a highly reproducible and simple systemwill be essential for application for therapeutic purposes.Most significantly, the combination of 4 d of EB culture in10 ng/ml Activin-A with FACS selection based on the expressionof the T/GFP selectable marker enabled the enrichment of renalprogenitors and elimination of Pax-6–expressing neuroectodermalcell populations (Figure 1D). The sorting of the LacZ/T/GFP–FACS-selected fraction enriched for cells with high levels ofPax-2 expression, consistent with ureteric bud progenitors,although their level of purity remains to be determined. Takentogether with the observations that injection of either thisLacZ/T/GFP– population or undifferentiated LacZ/ES cellsled to teratoma formation, this suggests that purification anddifferentiation are essential for the generation of normal renalprogenitors.
Selection based on T expression allowed for enrichment of distinctrenal progenitor populations from EB. LacZ/T/GFP+ cells werehighly enriched for proximal tubule progenitors and LacZ/T/GFP–cells for ureteric bud/collecting tubule progenitors. This wasdetermined by lineage tracing and distribution analysis of LacZ-encoded-galactosidase. LacZ/T/GFP+ populations were shown to be proximaltubule progenitors, incorporating into the nephrogenic mesenchymeof fetal kidneys during the early, highly proliferative stagesof nephron induction and into stable proximal tubules when injectedinto newborn live mouse kidneys in vivo, as demonstrated byco-localization with the proximal tubule markers alkaline phosphataseand aquaporin-1 (31). It is of note that the structural profilesof proximal tubules that integrated the -galactosidase–positiveEB-derived LacZ/T/GFP+ progenitors showed normal connectionswith glomeruli, closed tubule lumens, and elaborate brush borders(Figure 4). In addition, the normal apical membrane polarizationof alkaline phosphatase and apical and basal membrane localizationof aquaporin-1 further suggested normal differentiation of thesetubules. Since the BCIP/NBT method that was used to localizealkaline phosphatase shows activity because it is dependenton enzymatic breakdown of its substrate, this further suggeststhat these proximal tubules exhibited some normal functions.
Injection of renal progenitor cells into live newborn mousekidneys offers a powerful new method for analysis of integrationduring development and persistence into adulthood. The specificityand the kinetics of incorporation were highly reproducible.Initially, (4 h to 1 d) injected cells were seen in the interstitiumclose to the site of injection. During the next week, apparentmigration along peritubular tracks (Supplementary Figure 5)toward the nephrogenic zone and integration into PCNA-positiveproliferating progenitor cell pools took place, so that by 15d after injection, cells were seen at their final proximal tubule(T/GFP+) or collecting tubule (T/GFP–) destinations, integratedinto maturing tubular structures. This pattern was always seenin successfully injected kidneys: LacZ/T/GFP+ cells were seenonly in proximal tubules, never in ureteric bud–derivedstructures, whereas LacZ/T/GFP– cells were never seenin proximal tubules.
The apparent rapid migration of newly injected LacZ/T/GFP+ progenitorcells from their site of injection through the interstitiumto the cortex allows the speculation that growth factor gradientsand/or extracellular matrix niche environments may contributeto the homing of a specific progenitor cell to a particularlocale (32–35). In that regard, it is of interest thatmany LacZ/T/GFP– progenitors incorporated into collectingtubules at the tip of the papilla, which has been identifiedas a niche stem cell in the adult kidney (36). Although immunofluorescencecannot be used for analysis of fixed kidney tissues and immunohistochemistrycan mask immunoreactivity, the observation that in the majorityof cases -galactosidase staining was present in all cells ofa tubular cross-section suggests clonal expansion of incorporatedEB-derived progenitors. It is encouraging that integration wasstill seen in kidneys at the termination of the studies, 7 moafter a single injection.
Our studies suggest that it is possible to derive different,monopotent renal progenitor populations from EB. By a combinationof culture condition optimization and selection, we have generateda relatively pure population of renal proximal tubule progenitorsthat are capable of long-term integration into normal nephronsand do not lead to teratoma formation when followed for 7 mo.Ureteric bud progenitors have also been detected, although theconditions for purification have not yet been attained. Thelack of integration of either progenitor cell population intoglomeruli suggests that different conditions will be necessaryto direct differentiation along a glomerular lineage. Our studiesalso confirm the notion that it is preferable to predifferentiateES cells in vitro into the desired lineage to reduce the riskfor teratoma formation and provide the first step to regenerativecell therapy in response to genetic, ischemic, or toxic insultsthat target the renal proximal tubule.
This work was supported by a fellowship from the French Societyof Nephrology (C.V.) and by National Institutes of Health grantsP01 DK62345 (P.W.), RO1 HL71800 and HL48834 (G.K.), and RO1AG19366 (G.S.). O.W. and H.J.F. were supported by Sonderforschungsbereich497-Projekt A7.
The work was reported in part in abstract form at the InternationalSociety for Stem Cell Research; June 29 to July 1, 2006; Toronto,Ontario, Canada.
We thank Prof. Jérome Rossert, Prof. Jean-Daniel Sraer,Dr. Stuart Fraser, Dr. Joan Isern, Dr.Valerie Gouon-Evans, MarionKennedy, Macarena Kattman, and all members of Prof. Keller'slaboratory for discussion, guidance, and patience. We also thankHaojiang Zhang for technical assistance in microdissection.Flow cytometry and fluorescence microscopy was carried out withthe assistance of the Mount Sinai shared facilities.
Footnotes
Published online ahead of print. Publication date availableat www.jasn.org.
Wilkinson DG, Bhatt S, Herrmann BG: Expression pattern of the mouse T gene and its role in mesoderm formation.
Nature 343
: 657
–659, 1990[CrossRef][Medline]
James RG, Schultheiss TM: Bmp signaling promotes intermediate mesoderm gene expression in a dose-dependent, cell-autonomous and translation-dependent manner.
Dev Biol 288
: 113
–125, 2005[CrossRef][Medline]
Birchmeier C, Birchmeier W: Molecular aspects of mesenchymal-epithelial interactions.
Annu Rev Cell Biol 9
: 511
–540, 1993[CrossRef][Medline]
Challen GA, Martinez G, Davis MJ, Taylor DF, Crowe M, Teasdale RD, Grimmond SM, Little MH: Identifying the molecular phenotype of renal progenitor cells.
J Am Soc Nephrol 15
: 2344
–2357, 2004[Abstract/Free Full Text]
Kreidberg JA, Sariola H, Loring JM, Maeda M, Pelletier J, Housman D, Jaenisch R: WT-1 is required for early kidney development.
Cell 74
: 679
–691, 1993[CrossRef][Medline]
Dressler GR, Douglass EC: Pax-2 is a DNA-binding protein expressed in embryonic kidney and Wilms tumor.
Proc Natl Acad Sci U S A 89
: 1179
–1183, 1992[Abstract/Free Full Text]
Kispert A, Vainio S, McMahon AP: Wnt-4 is a mesenchymal signal for epithelial transformation of metanephric mesenchyme in the developing kidney.
Development 125
: 4225
–4234, 1998[Abstract]
Brophy PD, Ostrom L, Lang KM, Dressler GR: Regulation of ureteric bud outgrowth by Pax2-dependent activation of the glial derived neurotrophic factor gene.
Development 128
: 4747
–4756, 2001[Abstract/Free Full Text]
Ribes D, Fischer E, Calmont A, Rossert J: Transcriptional control of epithelial differentiation during kidney development.
J Am Soc Nephrol 14[Suppl 1]
: S9
–S15, 2003
Menke AL, Schedl A: WT1 and glomerular function.
Semin Cell Dev Biol 14
: 233
–240, 2003[CrossRef][Medline]
Evans MJ, Kaufman MH: Establishment in culture of pluripotential cells from mouse embryos.
Nature 292
: 154
–156, 1981[CrossRef][Medline]
Smith AG, Heath JK, Donaldson DD, Wong GG, Moreau J, Stahl M, Rogers D: Inhibition of pluripotential embryonic stem cell differentiation by purified polypeptides.
Nature 336
: 688
–690, 1988[CrossRef][Medline]
Daley GQ: From embryos to embryoid bodies: Generating blood from embryonic stem cells.
Ann N Y Acad Sci 996
: 122
–131, 2003[Medline]
Colman A: Making new beta cells from stem cells.
Semin Cell Dev Biol 15
: 337
–345, 2004[CrossRef][Medline]
Soria B, Roche E, Berna G, Leon-Quinto T, Reig JA, Martin F: Insulin-secreting cells derived from embryonic stem cells normalize glycemia in streptozotocin-induced diabetic mice.
Diabetes 49
: 157
–162, 2000[Abstract]
Kubo A, Shinozaki K, Shannon JM, Kouskoff V, Kennedy M, Woo S, Fehling HJ, Keller G: Development of definitive endoderm from embryonic stem cells in culture.
Development 131
: 1651
–1662, 2004[Abstract/Free Full Text]
Hamazaki T, Iiboshi Y, Oka M, Papst PJ, Meacham AM, Zon LI, Terada N: Hepatic maturation in differentiating embryonic stem cells in vitro.
FEBS Lett 497
: 15
–19, 2001[CrossRef][Medline]
Okabe S, Forsberg-Nilsson K, Spiro AC, Segal M, McKay RD: Development of neuronal precursor cells and functional postmitotic neurons from embryonic stem cells in vitro.
Mech Dev 59
: 89
–102, 1996[CrossRef][Medline]
Keller G: Embryonic stem cell differentiation: Emergence of a new era in biology and medicine.
Genes Dev 19
: 1129
–1155, 2005[Abstract/Free Full Text]
Burrow CR, Wilson PD: A putative Wilms tumor-secreted growth factor activity required for primary culture of human nephroblasts.
Proc Natl Acad Sci U S A 90
: 6066
–6070, 1993[Abstract/Free Full Text]
Rookmaaker MB, Verhaar MC, van Zonneveld AJ, Rabelink TJ: Progenitor cells in the kidney: Biology and therapeutic perspectives.
Kidney Int 66
: 518
–522, 2004[CrossRef][Medline]
Kim D, Dressler GR: Nephrogenic factors promote differentiation of mouse embryonic stem cells into renal epithelia.
J Am Soc Nephrol 16
: 3527
–3534, 2005[Abstract/Free Full Text]
Steenhard BM, Isom KS, Cazcarro P, Dunmore JH, Godwin AR, St John PL, Abrahamson DR: Integration of embryonic stem cells in metanephric kidney organ culture.
J Am Soc Nephrol 16
: 1623
–1631, 2005[Abstract/Free Full Text]
Fehling HJ, Lacaud G, Kubo A, Kennedy M, Robertson S, Keller G, Kouskoff V: Tracking mesoderm induction and its specification to the hemangioblast during embryonic stem cell differentiation.
Development 130
: 4217
–4227, 2003[Abstract/Free Full Text]
Zambrowicz BP, Imamoto A, Fiering S, Herzenberg LA, Kerr WG, Soriano P: Disruption of overlapping transcripts in the ROSA beta geo 26 gene trap strain leads to widespread expression of beta-galactosidase in mouse embryos and hematopoietic cells.
Proc Natl Acad Sci U S A 94
: 3789
–3794, 1997[Abstract/Free Full Text]
Soriano P: Generalized lacZ expression with the ROSA26 Cre reporter strain.
Nat Genet 21
: 70
–71, 1999[CrossRef][Medline]
Avner ED, Piesco NP, Sweeney WE Jr, Ellis D: Renal epithelial development in organotypic culture.
Pediatr Nephrol 2
: 92
–99, 1988[CrossRef][Medline]
Polgar K, Burrow CR, Hyink DP, Fernandez H, Thornton K, Li X, Gusella GL, Wilson PD: Disruption of polycystin-1 function interferes with branching morphogenesis of the ureteric bud in developing mouse kidneys.
Dev Biol 286
: 16
–30, 2005[CrossRef][Medline]
Maeshima A, Yamashita S, Maeshima K, Kojima I, Nojima Y: Activin a produced by ureteric bud is a differentiation factor for metanephric mesenchyme.
J Am Soc Nephrol 14
: 1523
–1534, 2003[Abstract/Free Full Text]
Wilson PD: In vitro methods in renal research. In:
Pediatric Nephrology, 5th Ed., edited by Avner E, Harmon W, Niaudet P, Philadelphia, Lippincott Williams & Wilkins, 2004
, pp 317
–333
Scadden DT: The stem-cell niche as an entity of action.
Nature 441
: 1075
–1079, 2006[CrossRef][Medline]
Moore KA: Recent advances in defining the hematopoietic stem cell niche.
Curr Opin Hematol 11
: 107
–111, 2004[CrossRef][Medline]
Nagamatsu G, Ohmura M, Mizukami T, Hamaguchi I, Hirabayashi S, Yoshida S, Hata Y, Suda T, Ohbo K: CTX family cell adhesion molecule, JAM4, expresses in stem cell and progenitor cell populations of both male germ cell and hematopoietic cell lineages.
Mol Cell Biol 26
: 8498
–8506, 2006[Abstract/Free Full Text]
Campos LS, Decker L, Taylor V, Skarnes W: Notch, epidermal growth factor receptor, and beta1-integrin pathways are coordinated in neural stem cells.
J Biol Chem 281
: 5300
–5309, 2006[Abstract/Free Full Text]
Oliver JA, Maarouf O, Cheema FH, Martens TP, Al-Awqati Q: The renal papilla is a niche for adult kidney stem cells.
J Clin Invest 114
: 795
–804, 2004[CrossRef][Medline]
Top
Abstract
Introduction
-galactosidase immunohistochemistry showed incorporation into blastemal cells of the nephrogenic zone. After a single injection into developing live newborn mouse kidneys, co-localization studies showed that the LacZ/T/GFP+ cells were stably integrated into proximal tubules with normal morphology and normal polarization of alkaline phosphatase and aquaporin-1 for 7 mo, without teratoma formation. It is concluded that defined differentiation of ES cells into embryoid bodies with Activin-A and selection for T expression provides a means to isolate and purify renal proximal tubular progenitor cells with the potential for safe use in regenerative therapies. 
85% of T/GFP+ cells were Pax-2/WT-1/Wnt-4 positive. These results suggested that FACS selection of T+ cells from EB cells that were cultured for 4 d in the presence of 10 ng/ml Actvin-A effected a significant enrichment of a renal progenitor cell population. By contrast, FACS-selected T cells that were cultured under the same conditions, although not homogeneous, were enriched for renal progenitors that contained high levels of Pax-2 expression. These isolation and enrichment procedures were highly reproducible, showing consistent results in multiple experiments. 
