Potential renal stem/progenitor cells identified by in vivo lineage tracing

Abstract

Mammalian kidneys consist of more than 30 different types of cells. A challenging task is to identify and characterize the stem/progenitor subpopulations that establish the lineage relationships among these cellular elements during nephrogenesis in the embryonic and neonate kidneys and during tissue homeostasis and/or injury repair in the mature kidney. Moreover, the potential clinical utility of stem/progenitor cells holds promise for the development of new regenerative medicine approaches for the treatment of renal diseases. Stem cells are defined by unlimited self-renewal capacity and pluripotentiality. Progenitor cells have pluripotentiality but no or limited self-renewal potential. Cre-LoxP-based in vivo genetic lineage tracing is a powerful tool to identify stem/progenitor cells in their native environment. Hypothetically, this technique enables investigators to accurately track the progeny of a single cell or a group of cells. The Cre/LoxP system has been widely used to uncover the function of genes in various mammalian tissues and to identify stem/progenitor cells through in vivo lineage tracing analyses. In this review, we summarize the recent advances in the development and characterization of various Cre drivers and their use in identifying potential renal stem/progenitor cells in both developing and mature mouse kidneys.

INTRODUCTION

The differentiation, growth, and morphogenesis of various cell types are coordinated to generate structurally and functionally complex organs including the kidney. The kidney is composed of over 30 different cell types. Extensive efforts have been made to identify the stem/progenitor elements that establish the lineage relationships among various populations of renal cells during embryogenesis, organogenesis, maintenance, and injury repair. Stem cells possess the unique capacity for limitless self-renewal and are capable of differentiating into one or more cellular phenotypes. A progenitor cell, in contrast, is an undifferentiated cell that lacks or has limited self-renewal capability and can only differentiate into one or several tissue-specific cell types. Self-renewal is the potential for stem/progenitor cells to asymmetrically or symmetrically divide to produce one or two daughter cells that maintain the developmental ability similar to the respective mother cell, empowering continued renewal capacity (1). The application of stem/progenitor cells holds promise for the development of new and powerful strategies for future regenerative medicine.

Multiple approaches are adaptable to the study of stem/progenitor cells including in vitro clonogenicity and multilineage differentiation analyses. These methods have allowed for the identification of potential stem/progenitor cells in various organs and tissue types, including the kidney (2, 3), adipose (4), bone marrow (5), liver (6), and skin (7). Alternatively, stem cells can be identified in vivo through long-term label retention assay since they seldom divide (8). These methods are considered classic approaches to assess stemness. Recently, Cre-LoxP-based in vivo genetic lineage tracing has become a powerful approach to monitor stem/progenitor cell populations in their native environment, making it possible to characterize cellular hierarchies. Theoretically, this technique can unambiguously track the offspring of a single cell or a group of cells. This Cre/LoxP system is widely used to ablate numerous genes of interest in efforts to identify their tissue-specific function in various mammalian tissues as well as to perform in vivo lineage tracing studies to identify the stem/progenitor cohort. The technical aspects of Cre-LoxP-based lineage tracing and stem cell products as a potential therapy for kidney disease has been previously reviewed (9–12). This review focuses on recent advances in the identification of renal stem/progenitor cells through Cre-LoxP-based lineage tracing in both developing and mature mouse kidneys.

CONSTITUTIVE AND INDUCIBLE CRE DRIVERS

The Cre/LoxP technique uses cyclization recombinase (Cre) of the bacteriophage P1, which targets specific 34-bp LoxP sites and deletes intervening DNA through recombination (13). Thus, a DNA fragment of interest, such as a key exon of a gene or a STOP cassette preventing transcription of a reporter gene, can be flanked by two LoxP sites and removed by Cre recombinase (14). Spatial specificity is acquired by forcing the expression of Cre under the control of specific promoters that are activated only in the cell type of interest. A mouse carrying a constitutively active Cre allele driven by a constitutively active promoter is referred to as a constitutive Cre driver.

It is, of course, essential to temporally control Cre-mediated recombination. This is usually achieved via the use of Cre fusions containing a modified variant of the estrogen receptor (ER^T2). The Cre fusions are retained in the cytoplasm until ER^T2 binds to its ligand tamoxifen, leading to nuclear localization, where they mediate recombination (15–18) There are now multiple Cre fusions including CreER^T2 (18), GFPCreER^T2 (GCE) (19), and ER^T2CreER^T2 (ECE) (20), that have variable expression and recombination frequency that are dependent on tamoxifen for induction. Temporal control of Cre-mediated recombination can also be accomplished through constitutively active Cre regulated by drug-inducible promoters, such as tetracycline-regulated systems (21). However, these systems require combining three, rather than two, different transgenes through a complicated and costly breeding strategy. In parallel, a mouse harboring an inducible Cre-mediated recombination system is an inducible Cre driver.

IDEAL CRE DRIVERS FOR LINEAGE TRACING

Lineage tracing relies exclusively on reporter expression that faithfully replicates the expression of the endogenous gene whose promoter drives Cre expression. Therefore, Cre expression and Cre-mediated recombination should only be detected in cell types where the promoter of the endogenous gene is active (complete fidelity). For lineage tracing using inducible Cre drivers, the inducible driver should ideally satisfy four criteria (20): 1) cell specific, 2) no background recombination in the absence of induction, 3) high recombination rate after induction, and 4) complete fidelity in cell specificity as in the constitutive drivers.

A growing list of Cre driver mouse strains have been generated through a variety of approaches involving either inclusion of random integration of Cre transgenes containing short promoter fragments or much larger bacterial artificial chromosomes (BACs). The former may cause promiscuous expression patterns, complicating data interpretation. The latter is more likely to harbor all unknown regulatory sequences and is thus more predisposed to faithfully replicate the expression profile of the endogenous gene of interest. Alternatively, a Cre transgene is inserted at the position of the translation initiation codon of the gene of interest. Such knockin approaches are also more inclined to recapitulate the endogenous expression pattern, because all adjacent DNA regulatory elements are intact at the locus of interest (for a review, see Ref. 9). Nevertheless, it cannot be assumed with absolute certainty that the expression pattern of any integrated transgene recapitulates the expression pattern of the endogenous gene product (22). Characterization of the faithfulness (or fidelity) of each Cre driver and lack of leakiness of each inducible driver in the tissue of interest is, thus, crucial for meaningful data analyses. Importantly, online databases that archive the characterization of Cre activities are emerging (9) and are critical resources for designing lineage tracing studies. All Cre drivers discussed in this review are shown in Table 1.

Table 1.

Cre drivers

Driver	Allele	Transgene	Induction	Fidelity	Lack of Leakiness	Reference(s)
Ren1c	Ren1cCre	BAC	Constitutive	Yes	NA	(23–25)
Ren1d	Ren1d-cre	Knockin	Constitutive	Yes	NA	(24, 26, 27)
Aqp2	Aqp2Cre	BAC	Constitutive	Yes	NA	(28–30)
Foxd1	Foxd1-eGFPCre (Foxd1-GC)	Knockin	Constitutive	Yes		(31,32)
Six2	Six2-GFPCre (Six2-GC)	Knockin	Constitutive			(19, 32)
Actin	ActinCreER	a	Tamoxifen	NA		(33)
Axin	Axin2CreER	Knockin	Tamoxifen		Yes, in the mammary gland	(33, 34)
Aqp2	Aqp2ERT2CreERT (Aqp2ECE)	Knockin	Tamoxifen	Yes	Yes	(20, 29)
Cited 1	Cited1-CreERT2-IRESeGFP	BAC	Tamoxifen	Yes	Yes	(35)
Foxd1	Foxd1-eGFPCreERT2 (Foxd1-GCE)	Knockin		Yes	Yes	(31)
Foxd1	Foxd1-CreERT2 (Foxd1-CE)	Knockin				(31)
Six2	Tet-off-eGFPCre (Six2-TGCtg/+) b	BAC	Doxycycline	Yes		(19)
Six2	Tet-off-eGFPCre (Six2TGC/+) c	Knockin	Doxycycline	Yes		(19)
Six2	CreERT2 (Six2CE/+)	Knockin	Tamoxifen	Yes		(19)
Six2	eGFPCreERT2 (Six2GCE/+)	Knockin	Tamoxifen	Yes	Yes	(19)
Lgr5	Lgr5-EGFP-ires-CreERT2	Knockin	Tamoxifen	Yes		(36)
Pax2	Pax2.rtTA;TetO.Cre	BAC	Doxycycline	Yes	Yes	(37,38)
Pax8	Pax8.rtTA;TetO.Cre	Fragmentd	Doxycycline		Yes	(38, 39)
PODXL1	PEC-rtTA|LC1	Fragmente	Doxycycline	No	Yes	(22, 40–42)
Renin	Ren-rtTA2/LC1	BAC	Tetracycline	Yes		(27)
Ren1c	Ren1cCreER	BAC	Tamoxifen	Yes	Yes	(23, 43)
Slc34a1	SLC34a1-GFPCreERT2 (SLC34a1GCE)	Knockin	Tamoxifen	Yes	Yes	(44)
Sox9	Sox9CreERT2	BAC	Tamoxifen			(45, 46)
Sox9	Sox9IRES-CreERT2	Knockin	Tamoxifen		Yes	(47, 48)
Zfyve27	Zfyve27-CreERT2		Tamoxifen	No	Yes	(49)

NA, not applicable.

^a No details given; ^b,cdoxycycline addition did not silence Cre activity in the TGC bacterial artificial chromosome (BAC) transgenic allele but disabled Cre activity in most cells in the TGC knockin allele; ^d4.3 kb of the upstream regulatory sequence, complete exon 1 and intron 1, part of exon 2, and 0.8 kb of intron 2 of the murine Pax8 gene; ^e3 kb of the human podocalyxin (hPODXL1) 5′-flanking region and 0.3 kb of the rabbit Podxl1 5′-untranslated region.

STRATEGY FOR THE IDENTIFICATION OF STEM/PROGENITOR CELLS THROUGH LINEAGE TRACING

In a lineage tracing experiment, a Cre driver is combined with a reporter gene. Various reporter lines have been previously reviewed (9). By selecting a cell-specific promoter driving Cre expression and pairing it with a reporter gene, investigators can potentially mark any cell population in which that promoter is active. If the promoter is stem/progenitor cell specific, the stem/progenitor cell, which is marked by the expression of the reporter, divides to generate daughter cells that either preserve the reservoir of stem/progenitor cells (self-renewal) or differentiates into various phenotypes (multipotentiality) to form a single cell-derived multiple-cell clone (clonogenicity). Since all progeny inherit expression of the reporter, all cells in the clone display the same color if the reporter is a fluorescent protein.

POTENTIAL RENAL STEM/PROGENITOR CELLS IDENTIFIED BY IN VIVO GENETIC LINEAGE TRACING

Several potential renal progenitor cells have been identified through in vivo genetic lineage tracing. These include progenitor cells marked by Cbp/P300 interacting transactivator with Glu/Asp-rich carboxy-terminal domain 1 (Cited1) (35), forkhead box D1 (Foxd1) (31), SIX homeobox 2 (Six2) (19), leucine-rich repeat containing G protein-coupled receptor 5 (Lgr5) (36), Sox9 (26), Protrudin (49), paired box transcription factor (Pax2) (37, 38), renin (23–27, 43), and aquaporin 2 (Aqp2) (28, 29), which are briefly discussed below in the following two sections: EMBRYONIC AND NEONATE RENAL PROGENITOR CELLS and ADULT RENAL PROGENITOR CELLS.

EMBRYONIC AND NEONATE RENAL PROGENITOR CELLS

Cited1-CreER^T2-IRESeGFP Transgene-Based Lineage Tracing Revealed a Population of Self-Renewing Epithelial Progenitor Cells in the Cap Mesenchyme

Cited1 encodes a transcriptional regulator that is uniquely expressed in the cap mesenchyme. Boyle et al. (35) generated and characterized a BAC Cited1-CreER^T2-IRESeGFP transgene and performed lineage tracing in mice. Expression of enhanced green fluorescent protein (EGFP) precisely overlaps with expression of endogenous Cited1 protein in the cap mesenchyme of Cited1-CreER^T2 mice. Accordingly, the authors concluded that the transgene is faithful. The dependence of induction on tamoxifen was also validated by the lack of recombination of the R26R^LacZ conditional reporter in the absence of tamoxifen. Analyses of the location and cellular characteristics of LacZ⁺ cells at different time points following tamoxifen injection revealed that the cap mesenchyme does not contribute to collecting duct (CD) cells in the adult. In addition, by injecting tamoxifen at different time points, the authors showed that the nephronic epithelium arising at different stages of nephrogenesis has specific spatial distribution in the adult kidney, demonstrating that the cap mesenchyme contains a population of self-renewing epithelial progenitor cells (35).

Six2 Marks a Multipotent Nephron Progenitor Population within the Cap Mesenchyme

During nephrogenesis, a pool of mesenchymal progenitor cells repeatedly generates nephrons, the basic functional units of the kidney. Six2, a homeodomain transcriptional regulator, is expressed in a subset of metanephric mesenchyme throughout kidney development (50), although its expression is undetectable in adult mouse kidneys (32). To genetically identify the progenitor cells within this population of the cap mesenchyme in vivo, Kobayashi et al. (19) developed four Six2-Cre alleles in the mouse: a BAC transgenic allele with a Tet-off-eGFPCre (Six2-TGC^tg/+) cassette and knockin alleles with TGC (Six2^TGC/+), CreER^T2 (Six2^CE/+), and eGFPCreER^T2 (Six2^GCE/+) cassettes inserted into the Six2 locus at the position of the Six2 translation initiation codon. The knockin alleles disrupt Six2 function. The TGC alleles permit doxycycline-regulated control, and the CE and GCE alleles provide for tamoxifen-dependent induction of Cre activity.

In mice carrying TGC and GCE alleles, GFP was detected exclusively in the cap mesenchyme from the onset of metanephric development. GFP⁺ cells also expressed Six2, and no GFP⁺ cells were Six2⁻. However, some Six2⁺ cells lacked detectable GFP expression in Six2-TGC^tg/+ and Six2^TGC/+ kidneys. Such slight mosaicism is often observed in other Cre-LoxP systems (29, 51). In brief, GFP expression from all alleles was confined to most or all Six2⁺ cells of the cap mesenchyme from the onset of nephrogenesis, demonstrating the faithfulness of these transgenes for lineage tracing analyses. By combining TGC or GCE alleles with a R26R-lacZ reporter allele (R26R^lacZ/+), the authors documented that Six2 marks a multipotent nephron progenitor population within the cap mesenchyme and that Six2⁺ cells generate all cell types in the main body of the nephron throughout nephrogenesis. Six2⁺ cells repopulate this progenitor pool over the course of nephrogenesis by self-renewal, and at least a subset of Six2⁺ cells is multipotent, giving rise to multiple regions within a single nephron harboring podocytes and proximal and distal tubule structures. It should be noted that no reporter expression was observed in oil-injected GCE controls, demonstrating the lack of background recombination and the absolute dependence of GCE-mediated recombination on tamoxifen addition (19).

Lgr5^+ve Stem/Progenitor Cells Contribute to the Thick Ascending Limb of Henle’s Loop and Distal Convoluted Tubule

Lgr5, a Wnt target gene, is a marker of epithelial stem cells in several adult organs including the colon, small intestine, stomach, and hair follicles, exhibits a high rate of Wnt-dependent self-renewal (52–56). To determine if Lgr5^+ve populations behave as renal stem cells in vivo in the developing mouse kidney, lineage tracing was performed with Lgr5-EGFP-ires-CreER^T2 reporter mice (36). These rodents carry a knockin EGFP-ires-CreER^T2 transgene at the Lgr5 locus. The tight correlation of Lgr5-EGFP expression with endogenous highly sensitive Lgr5-FISH signals validated the Lgr5-EGFP-ires-CreER^T2 model as an accurate reporter of endogenous Lgr5 expression in the developing mouse kidney. R26RLacZ and R26-4color reporters were then separately introduced into Lgr5-EGFP-ires-CreER^T2 mice to assess the contribution of Lgr5^+ve cells to the formation of various nephron segments. LacZ^+ve renal structures generated by neonate Lgr5^+ve cells significantly overlapped with Tamm-Horsfall protein (a marker of the thick ascending limb of Henle’s loop) and calbindin [a marker of the distal convoluted tubule (DCT)]. Occasionally in the cortex, but not in the medulla, LacZ costained with Aqp2. This was interpreted as evidence of a contribution to the Aqp2^+ve region connecting the convoluted tubule to CDs (36). In contrast, coexpression of LacZ and the proximal convoluted tubule marker megalin was never observed, thus confirming that Lgr5^+ve cells are distinct from the truly multipotent Six2⁺ population (19). The clonogenicity and contribution of individual Lgr5^+ve cells to these nephron segments were illustrated through multicolor in vivo lineage tracing using validated Rosa26-4color reporter mice. Taken together, these data revealed that Lgr5 is in fact a marker of a stem/progenitor cell population contributing to the formation of nephron segments that consist of the thick ascending limb of Henle’s loop, DCT, and connecting tubule (CNT), which links the DCT to CDs in neonate kidneys (19).

Foxd1 Progenitor Cells Generate Stromal Tissues of the Interstitium, Mesangium, and Pericytes

The transcription factor Foxd1 is an early marker of the stromal lineage (57). To fate map Foxd1-expressing (Foxd1⁺) population, McMahon and coworkers introduced eGFPCre (Foxd1^GC/+), eGFPCreER^T2 (Foxd1^GCE/+), and CreER^T2 (Foxd1^CE/+) transgenes into the 5′-untranslated region of the endogenous Foxd1 locus in mice. The authors demonstrated the lack of leakiness of Foxd1^GCE/+ and the fidelity of Foxd1^GC/+ and Foxd1^GCE/+ alleles. Lineage tracing analyses with these Foxd1-Cre knockin alleles revealed that the Foxd1⁺ cortical stroma represents a distinct multipotent self-renewing progenitor population that forms stromal tissues of the interstitium, mesangium, and pericytes throughout kidney organogenesis. A small subset of Foxd1⁺ cells contributes to Six2⁺ cells at the early stage of kidney outgrowth, generating a strict nephron and stromal lineage boundary derived from Six2⁺ and Foxd1⁺ progenitor cells, respectively (31).

Renin⁺ Progenitor Cells Are the Source of Multiple Cell Types

Renin-producing cells are responsible for the regulation of fluid-electrolyte homeostasis and blood pressure. To trace the lineage of renin-expressing cells, knockin mice expressing Cre in renin-expressing cells (Ren1^d-cre) were generated (26) and bred with R26R and Z/EG reporter mice. These reporter mice express either β-galactosidase (β-gal) or GFP, respectively, after recombination. Based on the consistent data from newborn kidneys of Ren1^d-Cre;R26R and Ren1^d-Cre Z/EG mice, the authors concluded that renin-expressing cells are precursors for smooth muscle, mesangial, and epithelial cells within the kidney (26). Given that both Ren1^d-Cre is a knockin allele and the pattern of β-gal distribution, or GFP, is consistent with transient expression of renin within the kidney during embryonic and fetal life, it is reasonable that faithfulness could be assumed. It has been previously observed that Renin⁺ cells begin to appear in the undifferentiated metanephric mesenchyme before renal vascularization and before renin is required for blood pressure homeostasis. If Ren1^d-Cre were truly faithful, one would expect that Cre be exclusively expressed in Renin⁺ cells. Since Cre may not be degraded in embryonic Renin⁺ cells before differentiation into Renin⁻ cells, faithfulness would be strictly validated by direct comparison of Cre expression with renin in adult kidneys, as was previously demonstrated in our work for the Aqp2Cre BAC transgene (28, 29).

Aqp2⁺ Progenitor Cells Generate Distal Renal Segments Comprising the Late DCT, CNT1 and CNT2, and CD

The results from our previous study suggest that Aqp2-expressing cells give rise to not only principal cells (PCs) but also intercalated cells (ICs) in the CD (28). PCs are marked by Aqp2 and Aqp3, whereas ICs can be identified by expression of carbonic anhydrase II (CAII) and V-ATPase complex that are composed of at least 13 subunits including B1 and B2 (B1B2). ICs may be further classified into α-ICs and β-ICs, both of which are characterized by the expression of anion exchanger 1 (AE1) and pendrin, respectively. Recently, we also reported that a subset of Aqp2⁺ cells behave as Aqp2⁺ progenitor (AP) cells through in vivo lineage tracing using both constitutive (Aqp2Cre RFP/+) and tamoxifen-inducible (Aqp2^ECE/+ RFP/+, Aqp2^ECE/+ Brainbow/+ and Aqp2^ECE/+ Brainbow/Brainbow) mouse models (29).

Coexpression of Aqp2 and/or red fluorescent protein (RFP) with a variety of segment-specific or cell-specific markers through high-resolution immunofluorescence confocal microscopy to identify the cell types derived from Aqp2⁺ cells in Aqp2Cre RFP/+ mice from embryonic day 14.5 to adult stage was also examined. Our data demonstrated that APs emerged and differentiated into late DCT (DCT2), CNT, and CD cells in Aqp2Cre RFP/+ at embryonic day 15.5 (29). CIdU chase labeling revealed that the percentage of CIdU⁺ APs in total APs was increased 15 and 46 times compared with the percentages of CIdU⁺ PCs and CIdU⁺ ICs, respectively (Fig. 1) (29). In adult Aqp2Cre RFP/+ mice, we confirmed that Cre was exclusively detected in Aqp2⁺ cells. All Cre⁺ cells were Aqp2⁺, and no Cre⁺ cells were Aqp2⁻, demonstrating the faithfulness of Aqp2Cre BAC transgene in recapitulating endogenous Aqp2 expression (29). In addition, the CNT harbors two molecularly distinct segments: CNT1, which lacks continued Aqp2 expression, and CNT2, which preserves Aqp2 expression similar to the CD (Fig. 2) (29). During development, APs give rise to at least five types of cells including PCs, α-ICs, and β-ICs to form the DCT2, CNT1, CNT2, and CD segments. All of these cells are RFP⁺ in Aqp2Cre RFP/+ kidneys (29).

Figure 1.

Embryonic aquaporin 2 (Aqp2)⁺ progenitor (AP) cells preferably incorporate CIdU at embryonic day 19.5 (E19.5). A: E19.5 embryos were pulse chased by injection of CIdU into their mothers and then euthanized at postnatal day 1 (P1). B and C: images of Aqp2 (blue), B1B2 (green), and CIdU (red) in the chased P1 kidneys. The boxed area in B was magnified ×6 in C to highlight a tubule containing multiple CIdU⁺ APs and lacking any CIdU⁺ principal cells (PCs) or CIdU⁺ intercalated cells (ICs). As indicated by the arrowheads, cells were magnified ×2.3 on the right, with split channels, to highlight labeled APs (Aqp2⁺B1B2⁺CIdU⁺) and unlabeled PCs (Aqp2⁺B1B2⁻CIdU⁻) and ICs (Aqp2⁻B1B2⁺CIdU⁻). CIdU staining was nuclear. D: percentage of CIdU⁺ cells in PCs, ICs, and APs, respectively. In all cases, scale bars = 50 µm. [Reproduced from Ref. 29 with permission from the Journal of the American Society of Nephrology.] n = 3 mice in D.

Figure 2.

Embryonic aquaporin-2 (Aqp2)-expressing cells likely generate late distal convoluted tubule (DCT2) and connecting tubule (CNT) segments. A: labeling schematic for DCT/CNT/collecting duct (CD) segments. B and C: images of paraffin sections (B) and microdissected red fluorescent protein (RFP)⁺ tubules (C) stained for various markers in adult Aqp2Cre RFP/+ kidneys. B: DCT2, CNT1, and CNT2 segments revealed by triple staining of RFP (red), Aqp2 (green), and NaCl cotransporter (NCC; blue). C: DCT2, CNT1, and CNT2 segments revealed by microdissected RFP⁺ (red) tubules stained for Aqp2 (green) and NCC (blue). In all cases, scale bars = 50 μm. [Reproduced from Ref. 29 with permission from the Journal of the American Society of Nephrology.]

The inducible models harbor an Aqp2^ECE allele, which was generated by inserting an ER^T2CreER^T2 (ECE) cassette at the position of the Aqp2 initiation codon of the Aqp2 locus to express ECE and to disrupt Aqp2 function (20). This is an ideal system because Aqp2^ECE is highly tamoxifen inducible, cell specific, completely faithful (ECE-mediated recombination occurs only in Aqp2⁺ cells), and absolutely dependent on tamoxifen for induction (there is no background recombination in the absence of tamoxifen) (20).

Induction at postnatal day 1, and subsequent analyses at postnatal days 3 and 42, demonstrated that neonate APs, like their embryonic counterparts, continue to generate various DCT2, CNT, and CD cells (29). Individual APs asymmetrically generated daughters with some inheriting the property of APs (i.e., Aqp2⁺B1B2⁺, self-renewal) and others differentiating into regular PCs, ICs, or α-ICs by losing B1B2 or Aqp2 (multipotency), respectively, to form single cell-derived multiple-cell clones (clonogenicity) through proliferation. This is convincingly demonstrated by the presence of PCs (Aqp2⁺B1B2⁻RFP⁺), ICs (Aqp2⁻B1B2⁺RFP⁺), and APs (Aqp2⁺B1B2⁺RFP⁺) in single cell-derived RFP clones in Aqp2^ECE/+ Brainbow/Brainbow mice at postnatal day 42 (Fig. 3) (29). We concluded that both Aqp2Cre and Aqp2^ECE/+ faithfully indicate the activation of the endogenous Aqp2 promoter for lineage tracing. A unique subset of Aqp2⁺ cells that also express V-ATPase B1B2 behaves as potential APs. Embryonic and neonate APs have the capacity of self-renewal, multipotentiality, and clonogenicity. During development, AP generates five types of cells in a step-wise manner to form the DCT2, CNT1, CNT2, and CD (Fig. 4) (29). APs emerge at embryonic day 15.5 before Aqp2⁺CAII⁺ cells. Inactivation of Foxi1 resulted in a single cell type positive for both Aqp2 and CAII, supporting that Aqp2⁺CAII⁺ cells are the ancestral cells of PCs and ICs (59). Activation of Foxi1 was required for these cells to further develop into more specialized cell types. Foxi1 expression was maintained in the cell population that differentiates into ICs (Foxi1⁺Aqp2⁻CAII⁺) and lost in those differentiating into PCs (Foxi1⁻Aqp2⁺CAII⁻) (59). Anyhow, Aqp2⁺CAII⁺ cells were not detectable until embryonic day 16.5 (29). Differentiation and/or maturation of ICs thus apparently require acquisition of various IC-specific molecules in a step-wise manner. Once the putative APs are committed to IC fate by loosing Aqp2 and maintaining B1B2, these cells sequentially obtain other IC features including CAII at embryonic day 16.5, followed by either AE1 or pendrin at postnatal day 1. Expression of CAII well before AE1 and pendrin in ICs is consistent with the finding that CAII may be important for IC “maturation” and its ablation reduces both α-ICs and β-ICs (60). Future studies are required to determine if APs express Foxi1 during development and renal maintenance.

Figure 3.

Individual neonate aquaporin 2 (Aqp2)⁺ progenitors (APs) exhibit clonogenicity, self-renewal, and multipotential capacity during development. A: Aqp2^ECE/+ Brainbow/Brainbow mice were induced by feeding nursing mothers a tamoxifen-containing diet from postnatal day 1 (P1) to postnatal day 2 (P2) and analyzed at postnatal day 3 (P3) and postnatal day 42 (P42), respectively. B: percentage of XFP⁺ clones containing various numbers of continuous XFP⁺ cells of the same color without interruption by an unlabeled cell or by a cell labeled by a different color at P42. n = 3 mice in B. C and D: images showing two different confocal planes of the same multiple-cell red fluorescent protein (RFP) clone that was presumably derived from an individual red fluorescent protein (RFP)-labeled AP at P42. Arrowheads: cells were magnified ×2.3 on the right, with split channels showing Aqp2 (blue), B1B2 (green), and RFP (red) to highlight the presence of principal cells (PCs; Aqp2⁺B1B2⁻RFP⁺), intercalated cells (ICs; Aqp2⁻B1B2⁺RFP⁺), and APs (Aqp2⁺B1B2⁺RFP⁺). Note that the AP in D was not detected in the same position in C. Each confocal plane apparently contained 5 RFP⁺ cells. E: images showing a five-cell membrane-bound CFP⁺ clone possessing PCs, α-ICs, and transitional cells (TCs). Arrowheads: cells were magnified ×2.3 on the right, with split channels showing Aqp2 (blue), CFP detected by GFP antibody (green) (58), and anion exchanger 1 (AE1; red) to highlight the presence of PCs (Aqp2⁺AE1⁻CFP⁺), α-ICs (Aqp2⁻AE1⁺CFP⁺), and TCs (Aqp2⁻AE1⁻CFP⁺), respectively. In all cases, scale bars = 50 μm. [Reproduced from Ref. 29 with permission from the Journal of the American Society of Nephrology.]

Figure 4.

Aquaporin 2 (Aqp2)⁺ progenitors (APs) generate five types of cells in a step-wise manner to form distal renal segments. Aqp2⁺ B1B2⁺ progenitor cells (AP) at embryonic day 15.5 (E15.5) emerged and differentiated stepwise into more than five types of cells to form molecularly distinct late distal convoluted tubule (DCT2), connecting tubule 1 (CNT1), connecting tubule 2 (CNT2), and collecting duct (CD) segments during development by either losing Aqp2 (−Aqp2) or B1B2 (−B1B2) and acquiring intercalated cells markers [carbonic anhydrase II (CAII), anion exchanger 1 (AE1), and pendrin] or DCT2 marker [NaCl cotransporter (NCC)]. Neonate APs also have clonogenicity, self-renewal, and multipotential capacity to generate DCT2, CNT, and CD cells. E16.5, embryonic day 16.5; P1, postnatal day 1. [Reproduced from Ref. 29 with permission from the Journal of the American Society of Nephrology.]

ADULT RENAL PROGENITOR CELLS

In contrast to embryonic kidneys, it has been debatable whether bona fide progenitor cells exist in the adult mammalian kidney (61, 62). Since mammalian kidneys have a remarkable capacity to regenerate new cells following injury, adult renal progenitor cells are defined as those that not only have 1) self-renewal capacity and 2) multipotentiality but also 3) participate in injury repair (63). However, we prefer to use a more strict definition. An adult renal progenitor cell should also 4) form single cell-derived multiple-cell clones (clonogenicity) and 5) contribute to tissue maintenance. Our reasoning is that kidneys undergo constant cellular renewal as evidenced by accumulated tubular cells in the urine resulting from normal shedding and by the magnitude of tubulogenesis within the adult kidney (equivalent to a 4.6- to 6-fold complete turnover of the renal epithelium over 7 mo) (33). Multipotentiality should also be strictly defined as the capacity of a single cell to generate multiple cell types. An adult mammalian renal progenitor cell that has been demonstrated in vivo to strictly satisfy these five requirements is still lacking. Nevertheless, in vivo lineage tracing resulted in the identification of several populations possessing some properties of renal progenitor cells in adult mouse kidneys.

Fate-Restricted Precursors Serving as Unipotent Progenitors Continuously Maintain and Self-Preserve the Mouse Kidney throughout Life

To genetically lineage trace individual cells of all renal epithelial cell types, including rare and potential multipotent stem/progenitor cells within the adult kidney, Rinkevich et al. (33) generated Actin^CreER R26^Rainbow mice, allowing expression of an inducible Cre-ER fusion protein under the ubiquitous Actin promoter. Adult Actin^CreER R26^Rainbow mice showed negative Rainbow fluorescence in the absence of tamoxifen administration, demonstrating the lack of leakiness. A dose of 1–2 mg tamoxifen was identified as the minimal concentration to induce sparse labeling of ∼1% within all renal epithelial cells. Clonal analyses revealed the clonogenicity of segment-specific progenitor cells as evidenced by the presence of single-colored multiple cell clones that maintained the fate of one single renal lineage and tubule type during renal maintenance. Clones with fates of the proximal tubule, distal tubule, and CD were observed, respectively. Similar results were obtained during ischemia-reperfusion injury (IRI) repair and during fetal development. Consistently, distinct segment-specific fates were maintained in single cell-derived renal spheres in vitro. Analysis of tetrachimeric mice created by transfer of color-labeled embryonic stem cells (ESCs) into colorless blastocysts revealed that nephrons were polyclonal, developing from expansions of singly fated clones (Fig. 5, A and B) (33). Accordingly, the authors concluded that fate-restricted precursors functioning as unipotent progenitors continuously maintain and self-preserve the mouse kidney throughout life (33). Since GFP clones with a CD fate harbored both Aqp3⁺GFP⁺ and Aqp3⁻GFP⁺ cells (Fig. 5, C and D) (33), which might be PCs and ICs, respectively. Triple staining of markers specific for PCs, ICs, and GFP should convincingly answer this question. Similar results were obtained in Pax2⁺ lineage tracing (see below in Pax2fl Progenitors Expanded Clones and Regenerated Podocytes and Necrotic Tubule Segments). Hence, progenitor cells fated to the CD might be multipotent APs as identified by us (29).

Figure 5.

Single-colored collecting duct (CD) clones harbor both aquaporin-3 (Aqp3)⁺ and Aqp3⁻ cells in adult Actin^CreER Rainbow mice during kidney maintenance and in developing tetrachimeric mice. A–C: singly colored clones that emerged 7 mo after tamoxifen induction were entirely retained within the segment-specific domains of label. A: peanut agglutinin (PNA) labeling of a clone with a convoluted distal tubule fate. B and C: Aqp2 and Aqp3 illustrate CD clones. Arrows and arrowheads indicate Aqp3⁺ and Aqp3⁻ cells within the clone, respectively. D: Aqp3 illustrates a singly colored CD clone. Arrows and arrowheads indicate Aqp3⁺ and Aqp3⁻ cells within the clone, respectively. Scale bars = 50 µm. [Revised from Ref. 33 with permission from Elsevier.]

Sox9⁺ Cells Are Major Contributors to Renal Tubule Repair and Regenerate Epithelia of the Proximal Tubule, Loop of Henle, and Distal Tubules Following Injury

McMahon and coworkers (47) reported in vivo lineage tracing of Sox9⁺ cells early in renal IRI using Sox9^{IRES-CreERT2/+} R26RtdT/+ mice. In these animals, IRES-CreER^T2 was inserted into the 3'-untranslated of Sox9 exon 3 (48), mediating expression of td-Tomato (RFP) in Sox9⁺ cells after tamoxifen induction. RFP was not observed in vehicle (corn oil) injection or in noninjected IRI controls, demonstrating the lack of leakiness. To confirm if the reporter faithfully recapitulated Sox9 expression, the authors examined Sox9 and RFP colabeling 48 h after IRI and tamoxifen injection. A subset of Sox9⁺ cells was RFP⁺. Although not clearly stated, the presence of Sox9⁻RFP⁺ cells was implied since the authors stated that “the great sensitivity of the CRE-system detects Sox9 activation at levels not observed by direct antibody detection of Sox9 in other kidney epithelium” (47). Strictly, the possibility that Sox9⁻RFP⁺ cells resulted from the lack of the reporter faithfully recapitulating Sox9 expression could not be completely ruled out. There were rare Sox9⁺ proximal tubular epithelial cells in the uninjured adult kidney. Nevertheless, they are assumed to play a minor role in the repair process. Consistently, significant proliferation was observed at 48–72 h after IRI, but by 48 h Sox9⁺ cells were already markedly increased within damaged proximal tubules, which is unlikely a result of the expansion of the rare resident Sox9⁺ cells. The major contributing factor, the authors claimed, must originate from de novo activation of Sox9 in surviving differentiated proximal tubular epithelial cells (47). It should be noted that this finding does not entirely eliminate a role for the rare Sox9⁺ cells in renal tubule maintenance under normal physiological conditions (47).

Susztak and coworkers (45) identified potential Sox9⁺ progenitor cells via the in vitro limiting dilution method and in vivo lineage tracing with the Sox9CreER^T2 BAC transgene as the inducible Cre driver. In vitro, Sox9⁺ cells showed unlimited proliferation and multilineage differentiation capacity. In vivo, Sox9⁺ cells can generate new epithelial cells in the proximal tubule, loop of Henle, and distal tubule segments, but not within the glomeruli and CDs following injury induced by folic acid (45). The leakiness and faithfulness of the Sox9CreER^T2 BAC transgene in the kidney, however, remains to be established. Cre activity in the absence of induction and direct comparison of Cre and Sox9 expression in adult mouse kidneys are lacking. Moreover, whether adult Sox9⁺ cells possess self-renewal and clonogenicity and contribute to tissue maintenance awaits further study. These two apparently conflicting studies can be reconciled by the hypothesis that both dedifferentiated epithelial cells and segment-specific progenitors contribute to epithelial regeneration (45).

Protrudin⁺ Cells Generated Long Tubular Segments After Severe Injury

Papillary label-retaining cells are, by definition, low cycling, but they divide to produce daughter cells after kidney injury. Transcriptome analyses of isolated live papillary label-retaining cells resulted in the identification of Zfyve27, which encodes protrudin, as a marker of a subpopulation of papillary label-retaining cells (49). Subsequent lineage tracing was conducted using bitransgenic Zfyve27-CreER^T2;R-tdTomato mice (49). No Cre “leakage” was observed in the absence of tamoxifen. Following tamoxifen induction, tdTomato⁺ cells were predominantly found in the upper part of the papilla. These cells cycled at lower frequency than most kidney cells and failed to produce progeny up to 9 mo postinduction. However, only after severe kidney injury induced by artery occlusion did they activate a program of proliferation, migration, and morphogenesis, forming multiple and long tubular segments. These regenerated tubules were found to reside primarily in the medulla (49). Therefore, Protrudin⁺ cells do not appear to contribute to kidney homeostasis but can regenerate medullary tubules following severe injury. These data also suggest that progenitor cell pools and repair of kidney injury is regionally specific within the kidney. It should be stressed that there were multiple tdTomato⁺ Protrudin⁻ cells as early as 1 day after tamoxifen induction (Fig. 1D in Ref. 49). Further investigation is warranted to establish the faithfulness of Zfyve27-CreER^T2 as well as to determine the self-renewal, multipotency, and clonogenicity of individual Protrudin⁺ cells within adult mouse kidneys.

Pax2⁺ Progenitors Expanded Clones and Regenerated Podocytes and Necrotic Tubule Segments

Romagnani and coworkers (2, 64–67) characterized a subset of parietal epithelial cells (PECs) in Bowman’s capsule that coexpressed two previously proposed “stemness” markers of progenitor cells, CD133 and CD24. These potential human renal progenitor cells were capable of self-renewal, formed single cell-derived clones, and generated podocytes or tubular epithelium when cultured in appropriate media. Intravenous injection of PKH26-labeled human renal progenitor cells into an acute tubular necrosis mouse model, or mice with adriamycin-induced nephropathy, ameliorated tissue damage and fibrosis. These changes were associated with improvement of kidney function. The authors attributed these therapeutic effects to the engraftment of renal progenitor cells into injured glomeruli and tubules (2, 12, 64–68). Persistence of renal progenitor cell activity, however, may require recapitulation of their self-renewal, clonogenicity, and multipotency in vivo in recipient mice. Ronconi et al. (65) further studied the capacity of CD133⁺CD24⁺ cells at the clonal level in vitro and in vivo. In particular, they injected clonal populations of CD133⁺CD24⁺ cells labeled with PKH26 into mice with adriamycin-induced nephropathy and showed their capacity to engraft in the kidney and differentiate into podocytes as well as tubular cells. These properties were not shared by CD133⁻CD24⁻ cells, which represented over 90% of all other kidney cells. The authors concluded that CD133⁺CD24⁺ human cells displayed clonogenicity, self-renewal, and multidifferentiative capacity. On the other hand, the authors still acknowledged the possibility that the engraftment observed might at least in part be related to cell fusion and exchange of PKH26 dye between cells (65). If CD133⁺CD24⁺ human cells can truly self-renew in vivo, they should be expected to preserve the expression of CD133 and CD24 after they are injected into recipient mice. This should be easily tested. Nevertheless, this important finding as an independent indicator of self-renewal apparently remains unresolved. In addition, PKH26 labeling was never observed in healthy mice injected with CD133⁺ CD24⁺ renal cells (65), indicating that CD133⁺ CD24⁺ cells play little or no role in renal maintenance.

Although CD133 is a useful marker for the isolation of many different types of adult human tissue stem and progenitor cells, only antibodies (such as the AC133 clone) that specifically bind the epitopes in the second extracellular loop of CD133, and that identify human renal progenitors, are suitable for stem and/or progenitor cell recognition (69, 70).

To overcome these limitations, Lasagni et al. (37) conducted lineage tracing experiments using conditional Pax2.rtTA;TetO.Cre;R26.Confetti(Pax2/Confetti) mice. This strategy was based on coexpression of Pax2 with CD133 in adult human renal progenitor cells (71). Pax2.rTA is a BAC transgene that drives expression of TetO-controlled transgenes upon doxycycline administration in Pax2⁺ cells (37). The group validated Pax2 promoter fidelity and the lack of leakage. Newly generated podocytes within glomeruli were labeled with different colors, suggesting they do not result from clonal division of a single progenitor and are derived from migration and differentiation of multiple different progenitor cells within the same glomerulus (37). Independent lineage tracing with dual reporter PEC-rtTA|LC1|tdTomato|Nphs1-FLPo|FRT-EGFP mice also revealed differentiation of parietal epithelial cells to podocytes during aging (72) and in the setting of experimental focal segmental glomerulosclerosis (FSGS) (22, 40). PEC-rtTA|LC1|R26R mice showed negligible nonspecific Cre recombination in podocytes or other cells in the glomerular tuft (<1 cell/100 glomeruli) and in tubular cells, most notably within the CDs. The lack of leakiness was validated by the absence of Cre recombination in uninduced pPEC-rtTA/LC1/R26R mice (41). Both PEC-rtTA and Nphs1-FLPo lines were generated by random integration of the transgenes. It remains to be solved whether Pax2-rtTA and PEC-rtTA used in these studies marked the same potential progenitor cells.

To challenge the current paradigm that all surviving tubular epithelial cells have a regenerative capacity in the functional recovery of acute kidney injury, Romagnani and coworkers (38) used the same strategy and revealed that a small subset of Pax2⁺ tubular cells also function as progenitor cells to repair renal IRI. These Pax2⁺ cells are more resistant to cell death, more clonogenic, and are responsible for both spontaneous and drug-stimulated regeneration of damaged tubule segments after injury (38). These findings appear to be inconsistent with the widespread proliferation of tubular epithelial cells following IRI that has been reported in other lineage tracing studies (32, 42, 44, 61). The authors attributed these conflicting conclusions to two factors: 1) the different methods used to estimate proliferating cells and 2) the demonstration of two phenomena occurring after IRI: massive tubular cell loss and endoreplication. When calculated over the total number of tubular epithelial cells survived at the end of the lineage tracing period, dividing cell number would be profoundly overestimated because of persistent cell loss, enriching dividing clones. Moreover, markers of cell cycle activation, proliferating cell nuclear antigen and Ki-67, can label endocycling cells that may result in further overestimation. In brief, the authors emphasized that referring to these markers inappropriately as “proliferation markers” may be a major reason for invalid conclusions (38).

It should be noted that Pax2⁺ progenitor cells were also apparently segment specific because each of the single-colored clones was confined to a specific tubular segment (Fig. 6, A–C). Further studies are required to determine if the potential Pax2⁺ progenitor cells can self-renew and whether they generate both PCs and ICs when fated to the CNT/CD in vivo in mouse kidneys. The latter is suggested, but not confirmed, by the presence of the apparent Aqp2⁺ and Aqp2⁻ cells within a single-colored clone (Fig. 6D).

Figure 6.

Single-colored clones of different fates were detected in Pax2/Confetti mice at day 30 after ischemia-reperfusion injury. Representative images of a kidney section show single-colored clones of various fates identified by segment-specific markers (all in white) in Pax2/Confetti mice at day 30 after ischemia-reperfusion injury. A: single-colored Megalin⁺ clones in S1–S2 segments of the proximal tubule. B: single-colored aquaporin (Aqp)1⁺ clones in S3 segments of the proximal tubule. C: single-colored Tamm-Horsfall protein (THP)⁺ clones in the thick ascending limb. D: single-colored Aqp2⁻ and Aqp2⁺ clones. Arrows and arrowheads indicate apparently Aqp2⁺ and Aqp2⁻ cells within a single-colored clone, respectively. Scale bars = 20 µm. [Revised from Ref. 38 with permission from Springer Nature.]

Renin⁺ Progenitor Cells Repopulate Glomerular Mesangial and Epithelial Cells After Injury

Since renin⁺ progenitor cells generate mesangial cells during nephrogenesis (26), Hugo and coworkers (27) evaluated whether adult renin⁺ cells contribute to glomerular regeneration after mesangial injury. To this end, they developed mRen-rtTA2 BAC transgene and combined it with LC1, which expresses Cre recombinase and firefly luciferase under control of the tetracycline response element and the R26R-LacZ reporter. The resulting Tet-on inducible triple-transgenic line (mRen-rtTA2/LC1/R26R-LacZ) permitted selective labeling of renin⁺ cells along renal afferent arterioles of adult mice (27). No intraglomerular LacZ staining was observed in healthy mice. However, about two-thirds of the glomerular tufts expressed LacZ during the regenerative phase following severe mesangial injury induced by administration of an antimesangial cell serum plus LPS. These cells expressed markers for mesangial cells only and not for endothelial, podocyte, or parietal epithelial cells. In contrast with LacZ⁺ cells in the afferent arterioles, LacZ⁺ cells in the glomerular tuft did not express renin. The authors interpreted this as evidence that adult extraglomerular renin⁺ cells are the primary source for regenerating the intraglomerular mesangial cells after injury (27). It is worth pointing out that the triple transgenic mice were induced with doxycycline for 16 days (27). The faithfulness of the mRen-rtTA2 BAC transgene would be unanimously validated by the lack of LacZ⁺ Renin⁻ cells at the earlier time points and/or the lack of Cre⁺ Renin⁻ in the extensively induced mice.

Shankland and coworkers (23) fate mapped kidney cells expressing renin in adult Ren1cCreER × Rs-tdTomato-R, Ren1cCre × Rs-ZsGreen-R, and Ren1dCre × Z/EG reporter mice. Ren1cCreER is a tamoxifen-inducible BAC transgene. After validation of the tamoxifen dependence on recombination and the promoter fidelity in adult Ren1cCreER × Rs-tdTomato-R mice, the authors depleted podocytes in all three cell-specific reporter mice through administration of cytotoxic antipodocyte antibodies. They found that the number of labeled cells of renin lineage (CoRL) was significantly increased after a decrease in podocyte number. These labeled cells were focally distributed within the glomerular tuft and/or along Bowman’s capsule. Marker analyses revealed that these labeled cells were glomerular parietal epithelial cells and podocytes. Renin expression at both mRNA and protein levels was not detectable de novo in diseased glomeruli. The authors concluded that CoRL might function as progenitors to regenerate glomerular epithelial cells following podocyte depletion (23). The same group also performed genetic cell fate mapping in aging Ren1cCre × Rs-ZsGreen reporter mice as well as in Ren1cCre/R26R-ConfettiTG/WT and Ren1dCre/R26R-ConfettiTG/WT mice during experimental FSGS (24, 25). They found that a subset of CoRL coexpressed de novo multiple podocyte proteins (podocin, nephrin, synaptopodin, Wilms’ tumor-1, and p57) and glomerular parietal epithelial cell proteins (claudin-1 and PAX8). Intravital multiphoton microscopy offered direct visual evidence for the migration of CoRL from the juxtaglomerular compartment to the parietal Bowman’s capsule, early proximal tubule, mesangium, and glomerular tuft. The authors concluded that these renin⁺ cells served as adult podocyte stem/progenitor cells and migrated to the glomerulus after podocyte depletion, where they acquired a podocyte-like phenotype (25). Because Ren1cCre is a constitutive Cre driver, an alternative interpretation could not be completely ruled out. That is, the labeled cells in the glomerular compartment might result from de novo transient activation of renin, as proposed by MacMahon and coworkers (47) to interpret the increased Sox9⁺ cells during IRI. To distinguish these possibilities, Shankland and coworkers generated Ren1cCreER/tdTomato/Nphs1-FLPo/FRT-EGFP mice to simultaneously label cells of renin lineage with RFP (tdTomato) after tamoxifen induction and podocytes with GFP in the adult kidney and to genetically test whether CoRL can transdifferentiate to a podocyte fate. Following podocyte depletion by nephrotoxic antibody and subsequent enalapril-enhanced partial replacement, double-labeled CoRL (RFP⁺GFP⁺) cells were observed as yellow-colored cells in a subset of podocin-expressing podocytes. These studies provide strong genetic evidence in vivo that CoRL can transdifferentiate to a podocyte fate and that lost podocytes can be regenerated in part by CoRL (43). Together with the findings reported by Hugo and coworkers (27), it seems that potential Renin⁺ progenitor cells can differentiate into mesangial cells, parietal epithelial cells, or podocytes, depending on which cell type is injured and requiring regeneration.

Two points should be emphasized. First, CoRL did not proliferate in the intraglomerular compartment at the time point examined, as evidenced by the lack of overlap between Confetti reporters and bromodeoxyuridine labeling (Fig. 7A) (24). In the absence of proliferation, it is not clear how single-colored small clones were formed within the intraglomerular compartment in Ren1cCre/R26R-ConfettiTG/WT animals at day 28 of FSGS (Fig. 7B) (24). Second, it remains unknown if the potential renin progenitor cells possess the self-renewal capacity and if single cell-derived clones harbor different types of cells (multipotency) in vivo in mouse kidneys.

Figure 7.

Multicolored reporters of cells of renin lineage (CoRL) were detected in glomerular tufts of Ren1cCre/R26R-ConfettiTG/WT mice with focal segmental glomerulosclerosis (FSGS). A: confocal images showing the lack of colocalization of all four reporters (converted to green color) with bromodeoxyuridine (BrdU; red) at baseline or FSGS day 14 (D14). BrdU⁺ cells were present in some tubules but were not readily detected in the juxtaglomerular compartment (JGC) or glomerular tuft (dashed white circles). B: confocal images showing that four CoRL reporter colors were detected without the use of antibodies: green fluorescent protein (GFP; green), red fluorescent protein (RFP; red), cyan fluorescent protein (CFP; blue), and yellow fluorescent protein (YFP; yellow). All four reporters were restricted to the JGC at baseline and were not detected in the glomerular tuft (dashed white circles). At FSGS D14, all four CoRL reporter colors were found in a subpopulation of cells in the glomerular tufts. The arrowhead indicates a single-colored clone harboring multiple RFP⁺ cells. [Revised from Ref. 24 with permission from PLoS One.]

PERSPECTIVES AND SIGNIFICANCE

In vivo lineage tracing is a powerful genetic tool that is widely available. Numerous Cre driver mice have been developed and used to identify several potential stem/progenitor cells in mouse kidneys, particularly in embryonic and neonate kidneys. Although adult Sox9⁺, Protrudin⁺, Pax2⁺, and renin⁺ cells have been reported to harbor some properties of progenitor cells, their self-renewal, multipotency, or both have not been systematically addressed. In addition, the fidelity and leakage of the Cre drivers used to identify these potential progenitor cells were not vigorously validated in all cases. In brief, an adult renal progenitor cell that meets the strict definition requiring the in vivo demonstration of 1) self-renewal capacity, 2) multipotentiality, 3) clonogenicity, 4) contribution to tissue maintenance, and 5) participation in injury repair still remains elusive. Identification of stem/progenitor cells through in vivo lineage tracing in various mouse models will be crucial for elucidating the contribution of embryonic and neonate renal stem cells to developmental defects and for isolating human renal progenitor cells as a prerequisite to assessing their therapeutic potential. The continued identification and characterization of renal progenitor cells will fundamentally dictate the future direction of renal regenerative medicine.

GRANTS

This work was supported by National Institute of Diabetes and Digestive and Kidney Diseases Grant DK080236 (to W.Z.).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

W.Z. and P.J.H. conceived and designed research; C.G. and A.T. performed experiments; W.Z. and C.G. analyzed data; W.Z. interpreted results of experiments; W.Z. prepared figures; W.Z. and P.J.H. drafted manuscript; W.Z. and P.J.H. edited and revised manuscript; W.Z., C.G., R.S., R.P. and P.J.H. approved final version of manuscript.