A New Cre Driver Mouse Line, Tcf21/Pod1-Cre, Targets Metanephric Mesenchyme

Conditional gene targeting in mice has provided great insight into the role of gene function in kidney development and disease. Although a number of Cre-driver mouse strains already exist for the kidney, development of additional strains with unique expression patterns is needed. Here we report the generation and validation of a Tcf21/Pod1-Cre driver strain that expresses Cre recombinase throughout the condensing and stromal mesenchyme of developing kidneys and in their derivatives including epithelial components of the nephron and interstitial cells. To test the efficiency of this line, we crossed it to mice transgenic for either loss or gain of function β-catenin conditional alleles. Mice with deletion of β-catenin from Tcf21-expressing cells are born with hypoplastic kidneys, hydroureters and hydronephrosis. By contrast, Tcf21-Cre driven gain of function for β-catenin in mice results in fused midline kidneys and hypoplastic kidneys. Finally, we report the first renal mesenchymal deletion of Patched1 (Ptch1), the receptor for sonic hedgehog (Shh), which results in renal cysts demonstrating a functional role of Shh signaling pathway in renal cystogensis. In summary, we report the generation and validation of a new Cre driver strain that provides robust excision in metanephric mesenchyme.


Introduction
The introduction of gene targeting in mouse embryonic stem cells in 1985 provided an efficient means to determine the function of gene(s) in mammalian tissues in vivo. More recent advances have permitted cell specific and temporal regulation of gene deletion and overexpression, providing investigators more refined tools to study the role of genes and molecular pathways in specific organs and tissues. The Cre-loxP system is most broadly used, largely due to the availability of a wide variety of floxed mouse lines in public consortia, pharmaceutical companies and academic labs [1].
In order to take full advantage of available floxed lines, renal researchers must be able to choose from a wide variety of complementary, robust and unique Cre-driver strains. Metanephric mammalian kidneys derive from two cell lineages: metanephric mesenchyme and ureteric bud that both arise from the intermediate mesoderm [2]. The metanephric mesenchyme is comprised of subpopulations including the self-renewing cap mesenchyme and stromal mesenchyme. The cap mesenchyme gives rise to all of the epithelial cell derivatives of the nephron from the glomerulus to the distal nephron, whereas the stromal mesenchyme is believed to give rise to interstitial cells such as myofibroblasts and pericytes, although the precise molecular steps have not been determined [3] [4]. A number of Cre driver strains exist for early kidney studies; e.g. HoxB7 [5] for the ureteric bud lineage, Six2 and Pax3 for the cap mesenhcyme [4,6,7], Foxd1 for the stromal population [8] and Wnt4 for the renal vesicles [9]. In addition, a number of Cre driver strains exist for fully differentiated cells in the nephron (reviewed in [2]).
Tcf21/Pod1 is a basic helix loop helix transcription factor with a unique expression pattern in both cap and stromal mesenchyme, providing an opportunity to develop a Cre-driver strain capable of broad deletion in metanephric mesenchymal populations [10,11,12]. Here we report the generation of a Cre-driver line using the endogenous Tcf21 promoter and describe robust renal phenotypes when this Cre driver strain is crossed to mice carrying a floxed bcatenin allele causing loss of function, an allele resulting in b-catenin gain of function and a floxed Ptch1 receptor allele. The Tcf21-Cre line provides a robust and valuable alternative genetic tool for the developmental renal biologist.

Ethics Statement
All mouse experiments were approved by the animal ethics committee at the Toronto Center for Phenogenomics, and were performed in accordance with 'Canadian Council of Animal Care' regulations.

Construction of a Tcf21-Cre targeting vector
To make the Tcf21-Cre targeting vector, a 3.1 kb fragment including intron 1, exon 2 and the 39 UTR of the Tcf21 gene (accession number AF296764) was amplified from murine genomic DNA by PCR and subcloned into the KpnI site of the pKO (knockout) vector (a kind gift from Dr. J. Rossant, The Hospital for Sick Children, Toronto, Canada). For the 59 homology arm, a 3.5 kb fragment was amplified from the 59 flanking region of exon 1, subcloned into pBluescript KS+, and excised using SmaI and SalI. The Cre recombinase cassette and bactin polyA fragment were excised from the NLS-Cre plasmid (kind gift from Dr. Brian Sauer, Oklahoma Medical Research Foundation, Oklahoma City, OK) and Nephrin promoter NLS-Cre plasmid [13] using EcoRI/SalI sites and EcoRI/XbaI sites, respectively. The Tcf21-PKO vector was digested with XbaI and PmlI, and the four fragments were subjected to ligation to construct the final Tcf21-Cre targeting vector.

ES cell culture and generation of chimeras
The Tcf21-Cre vector was linearized by Pme1 and electroporated into murine R1 ES cells derived from male blastocyst, hybrid of two 129 substrains (129X1/SvJ and 129S1/SV-+ p + Tyr-c Kitl Sl-J /+) [14]. After selection with G418 and FIAU (5-Iodo-29-fluoroarauracil), resistant clones were selected and subjected to Southern analysis using a 39 Tcf21 genomic 500 bp probe outside the region of homology that recognized a 5.3 kb and a 9.1 kb HindIII fragment for the mutant and wild-type alleles, respectively. Embryo manipulation and aggregation of the ES cell clones were carried out. One ES cell line generated chimeras that gave germline transmission.

Breedings of mouse lines
The Tcf21-Cre heterozygous founder mice were crossed with a Z/EG reporter murine line (kind gift from Dr. A Nagy, the Samuel Lunenfeld Research Institute). The Z/EG reporter construct is comprised of a systemic pCAGGS promoter which directs the loxP-flanked b-Geo (lacZ/neomycin-resistance and STOP signal) fusion gene followed by a green fluorescent protein (GFP) cassette [15]. Site-specific Cre recombinase expression leads to removal of the STOP signal resulting in GFP expression. Tcf21-Cre mice were then bred to b-catenin conditional loss of function (LOF) mice (Ctnnb1 fl/fl ) [16], b-catenin conditional gain of function (GOF) mice (Ctnnb1 ex3/+ ) [17], and conditional loss of function Ptch1 mice (Ptch1 fl/fl ) [18] to examine their phenotypes. Six2-Cre mouse line was a kind gift from Dr. Andrew P. McMahon (Harvard University, Cambridge), and crossed with Z/EG mouse line.

Phenotypic analysis
Murine embryos were dissected at specific time points described, and macroscopic phenotypes of each organ were photographed using a Leica dissecting microscope (Leica, MZ6). Whole embryos or organs were fixed in 10% formalin in phosphate-buffered saline (PBS) and subjected to histological analysis.

Immunostaining
For immunohistochemistry, tissues were dissected and fixed in 10% formalin/PBS and embedded in paraffin. 5 mm-thick sections were then rehydrated, boiled in citrate buffer for antigen retrieval, and endogeneous peroxidases were quenched with 3% hydrogen peroxide. For immunofluorescence, samples were fixed overnight in 4% paraformaldehyde at 4uC, cryoprotected in 30% sucrose overnight, embedded in Tissue-Tek OCT 4583 compound (Sakura Finetek USA Inc., Torrance, CA, USA) and 10 mm sections were cut by cryostat (Leica CM3050S). After blocking (10% goat serum, 3% albumin and 0.3% TritonX100 in PBS), samples were incubated with primary antibodies at 4uC overnight, washed and incubated with secondary antibodies for 1 h at room temperature. Primary antibodies used for this study included anti-GFP (A11122, Invitrogen, Camarillo, CA, USA), anti-alpha-smooth muscle actin (a-Sma, ab5694, Abcam, Cambridge, MA, USA) and anti-Tamm Horsfall protein (BT-590, Biomed Technologies, Inc. Stoughton, MA, USA). Immunohistochemical staining was carried out using Vectastain ABC kit (Vector Laboratories). Diaminobenzidine (Vector Laboratories) was used for the color reaction.
Two lectins were utilized to stain specific renal segments. Biotinylated Dolichos biflorus agglutinin (DBA, B-1035, Vector Laboratories, Burlingame, CA, USA) was used as a specific marker of the collecting duct, and visualized by FITC labeled avidin (Vector Laboratories). FITC labeled Lotus tetragolonobus lectin (LTL, FL-1321, Vector Laboratories) is a specific marker of the proximal tubule and was visualized directly with a fluorescent microscope.

In situ hybridization
The samples were fixed overnight in DEPC-treated 4% paraformaldehyde at 4uC, cryoprotected in 30% sucrose overnight at 4uC, embedded in Tissue-Tek OCT 4583 compound and snap frozen. 10 mm sections were cut on a cryostat and transferred to Superfrost microscope slides (Fisher Scientific Co., Pittsburgh, Pennsylvania, USA). Digoxigenin-labeled probes were prepared according to the manufacturer's instruction (Roche Molecular Biochemicals, Mannheim, Germany). Probes used for in situ analysis were Bmp4, Fgf8 (a kind gift from Dr. Janet Rossant, the Hospital for Sick Children, Toronto), Wnt4 (a kind gift from Dr. Andy McMahon at Harvard University), and Gli1 (a kind gift from Dr. Alexandra Joyner, Sloan Kettering Institute, New York). Details of the in situ analysis protocol may be obtained upon request.

Generation of a Tcf21-Cre founder mouse line
The targeting construct for Tcf21-Cre was designed to replace the first exon of the Tcf21 gene with Cre recombinase and a neomycin cassette (Fig. 1A). After electroporation, murine ES cell clones were screened by Southern blot analysis using a HindIII digest and 39 probe outside the region of homology. One correctly targeted ES cell clone was used to generate chimeric mice that gave germline transmission (Fig. 1B).
Tcf21-Cre driver mouse line excises genes in the metanephric mesenchyme and its derivatives Tcf21-Cre mice were bred with Z/EG reporter mice. In embryonic day 10.5 metanephros, expression of the GFP reporter (denoting successful Cre-mediated excision) was observed in the condensing metanephric mesenchyme surrounding the invading ureteric bud (Fig. 1C). At this stage, GFP staining appeared to be thin and mosaic. At E11.5, GFP expression in metanephric mesenchyme was evident (Fig. 1D). By E13.5, gene deletion activity was shown in broad areas in the mesenchyme including condensing mesenchyme, developing nephrons such as pretubular aggregates and comma-shaped bodies, interstitial stromal cells, and the mesenchyme surrounding the stalk of ureteric bud (Fig. 1E). At E16.5, in addition to the GFP expression observed at E13.5, GFP expression was also appreciated in S-shaped bodies, immature podocytes in capillary loop glomeruli, and immature tubular cells (Fig. 1F). We compared the domain of Tcf21-driven Cre excision to that of the Six2-Cre mice. As seen in Fig. 1G, Six2-Cre activity is restricted to cap mesenchyme and its derivatives. At postnatal day 0, in Tcf21-Cre;Z/EG mice, GFP was observed in podocytes, tubules and cells of Bowman's capsule, though it appeared mosaic in the tubules (Fig. 1H). Double immunostaining for GFP and pancytokeratin showed no overlap, demonstrating that Tcf21-Cre is not active in collecting duct cells that are derived from the ureteric bud lineage (Fig. 1I). Taken together, Tcf21-Cre has the capacity for broad gene deletion in the mesenchyme and its derivatives, including both interstitial cells and epithelial components of the nephron. Thus, Tcf21-Cre is a valuable resource permitting the use of a single mouse Cre-driver strain to gain insight into multiple metanephric populations.

Tcf21-Cre excises genes in mesenchymes in multiple organs
To explore the full potential of the Tcf21-Cre line, extrarenal Cre activity was assessed. At E10.5, GFP expression was observed in the 1 st to 3 rd branchial arches (Fig. 2 AB), the heart, and in metanephric mesenchyme surrounding the ureteric bud ( Fig. 2C and 1B). E11.5 embryos showed a similar pattern (data not shown). At E13.5 and E16.5, GFP expression was prominent in epicardium, lung mesenchyme, kidney mesenchyme, lamina propria and smooth muscle layer of the gastrointestinal tract, pancreas, adrenal gland, gonads, proximal part of aorta, facial and sublingual muscles, and in the choroid plexus of the ventricles ( Fig. 2D-Q). Ctnnb1 fl/fl ;Tcf21-Cre mice develop hydroureter and/or hypoplastic rudimentary kidneys Wnt-b-catenin signaling plays a key role in the interaction between the metanephric mesenchyme and the ureteric bud during renal development [19,20]. To validate the Tcf21-Cre line, we bred Tcf21-Cre mice with a b-catenin conditional loss of function (LOF) mouse (Ctnnb1 fl/fl ;Tcf21-Cre, referred to as b-catenin LOF mutant, hereafter). The b-catenin LOF mutants were born with the expected Mendelian ratio, but died within a few hours after birth. The deletion of b-catenin gene was confirmed by PCR, and was observed in the kidney and extrarenal tissues where Tcf21 is expressed including heart, lung, pancreas and gastrointestinal tract (Fig. 3A). Examination at E18.5 and in P0 pups demonstrated hydroureter with hypoplasia in 73.5% of the mutants, whereas 20.5% of mutants showed hypoplastic rudimental kidneys (Fig. 3B-D). Most of the mutant kidneys showed various degrees of hypoplasia. Kidneys with hydroureter demonstrated dilated tubules (Fig. 3F) and expansion of capillary loops, but some glomeruli appeared to be intact (Fig. 3I). The rudimentary kidneys were cystic, had only a few immature glomeruli, and showed random, disorganized nephrogenic aggregates at the periphery (Fig. 3GJ). As expected from this phenotype, expression of nephrogenic induction markers, Wnt4 and Fgf8, were markedly reduced and only detectable in the disorganized nephrogenic area (Fig. 3K-N). Neither tubular structures nor interstitial space were present.
In order to highlight the unique characteristics of Tcf21-Cre, we compared the phenotype to that of previously published mice with b-catenin deletion in the cap mesenchyme (Six2-Cre) or stromal cell (FoxD1-Cre) populations. Six2-Cre;Ctnnb1 fl/fl mice were deficient in nephrogenesis in the nephrogenic zone, whereas FoxD1-Cre;Ctnnb1 fl/fl mice failed to develop the interstitial space [20,21]. Compared to these compartmentalized phenotypes, the Tcf21-Cre phenotype extends from the nephrogenic zone, the interstitial space to the ureter, and also provides a range of severity of phenotypes among littermates due to a variable degree of mosaicism (Table 1). Together, these results demonstrate that the use of Tcf21-Cre allows the assessment of gene function in both the stromal and cap mesenchymes.
In addition to the previously reported phenotypes, we observed a high incidence of hydroureter in b-catenin LOF mutants, likely a result of deletion in the periureteric bud mesenchyme that gives rise to smooth muscle cells. In keeping with this possibility, we examined the expression of a-Sma and Bmp4 , which is known to play a key role in the development of the organized smooth muscle layer in the ureter [22,23]. Although a-Sma positive cells were present in the dilated ureter wall, the smooth muscle layer was extremely thin (Fig. 3OP). Notably, Bmp4 could not be detected in the mesenchyme surrounding the ureteric bud at E13.5 (Fig.  3QR), or in the dilated ureter at E18.5 (data not shown).

Ctnnb1 ex3/+ ;Tcf21-Cre mice develop fusion kidney or hypoplastic rudimentary kidney
Several recent papers have demonstrated that tight regulation of b-catenin expression is required for tissue homeostasis [24,25]. Thus, as a second test to validate the Tcf21-Cre driver strain, we bred it with the conditional gain of function b-catenin mice (Ctnnb1 ex3/+ ;Tcf21-Cre, referred to as b-catenin GOF mutant). Exon 3 of the mutant allele is flanked by loxP sequences. Upon Cre-mediated excision, the serine and threonine residues in exon 3 are removed, preventing phosphorylation and resulting in a stabilized b-catenin molecule that translocates and accumulates in the nucleus escaping ubiquitination and degradation [17].
b-catenin GOF mutants die during embryonic life between E12.5 and P0 (data not shown). At E18.5, 66% of mutants demonstrated fusion kidneys at midline and 33% had severely hypoplastic/dysplastic kidneys with a few disorganized glomeruli (Fig. 4A-E). These phenotypes were striking and differed from the previously reported Six2-Cre;b-catenin gain of function phenotype, which is renal agenesis at birth [20]. Rudimentary kidneys of Tcf21-Cre mutants and agenesis observed in Six2-Cre mutants may both be due to failure of nephrogenesis in the cap mesenchyme and its derivatives, but midline fusion appears to arise as a result of the broader gene deletion of Tcf21-Cre.
Strikingly, 100% of the b-catenin GOF mutants also exhibited embryonic tumor that engulfed the kidney, gut, heart, and lungs, and appeared to arise from multiple mesenchymal tissues (Fig. S1). Importantly, this phenotype did not obscure the renal phenotype but as discussed below, provided new insight into possible genetic interactions between the b-catenin and Shh pathways.

Ptch1 fl/fl ;Tcf21-Cre kidneys develop multiple renal cysts
As a final test of Tcf21-Cre line, we deleted Ptch1, the receptor for Shh. The Shh signaling pathway is crucial for the development of virtually all organs, and regulates cell fate determination, proliferation and tissue patterning [26]. After birth, Shh signaling is associated with certain types of malignancy [27,28,29]. Shh null mice show renal aplasia [30]. Shh and Ptch1 deletion from ureteric bud result in hydroureter and hypoplastic kidney, respectively [31,32]. However, the deletion of Ptch1 in the metanephric mesenchyme has not been previously reported. Ptch1 fl/fl ;Tcf21-Cre mice (referred to as Ptch1 mutants) die at varying time points between E12.5 to E18.5 due to an aggressive embryonic tumor with striking similarities to the tumors observed in the b-catenin GOF mutants described above (Fig. S1). Ptch1 mutant kidneys were of a similar size to those of Ptch1 fl/fl littermates (Fig. 5A). Renal invasion of the tumor appeared to be minimal (Fig. 5C). Histological examination of E18.5 mutant kidneys demonstrated dilation of the pelvic area and multiple cysts  (Fig. 5CD). In some cysts, glomeruli were observed, whereas other cysts arose in the renal tubules (Fig. 5D).
Stainings with Lotus tetragolonobus lectin (LTL), Tamm Horsfall protein (Thp), and Dolichos biflorus agglutinin (DBA) which marks the proximal tubules, loop of Henle, and collecting ducts respectively, showed that the cysts arose along the entire nephron (Fig. 5E-J). Mutant collecting ducts were dilated, distorted and winding (Fig. 5J). Additionally, immunostaining for Ki67 showed increased cell proliferation in the cystic wall, suggesting a possible mechanism for cyst formation (Fig. 5L).
We examined the expression of Gli1, a well-established marker of Shh activation. As Ptch1 suppresses the Shh pathway, genetic deletion of the Ptch1 gene should result in upregulation of Gli1 expression. In keeping with this, in situ hybridization showed increased expression of Gli1 in the condensing mesenchyme and interstitial cells at E16.5 (Fig. 5N). However, surprisingly, Gli1 expression was not seen in developing nephrons, such as pretubular aggregates and S-shaped bodies.

Discussion
Here we report the generation of a novel Cre transgenic mouse line, Tcf21-Cre, that allows gene excision in the metanephric mesenchyme and its derivatives, including interstitial cells and all epithelial components of the nephron from podocytes to distal tubules. It also results in gene deletion throughout the mesenchyme of other organs including lung, heart, gastrointestinal tract, pancreas, gonad and adrenal gland. Gene deletion begins at E10.5 in the condensing metanephric mesenchyme, and is consistently observed in its derivatives throughout renal development.
Currently, Six2-Cre and Pax3-Cre are most widely used to investigate the role of genes in the developing kidney and metanephric mesenchyme [4,6,7]. Is there any advantage or need for Tcf21-Cre? In contrast to Tcf21, Six2-Cre expression is restricted to the cap mesenchyme. While this is useful to determine gene function in this specific compartment, Tcf21-Cre provides a broader mesenchymal excision from both stromal and cap compartments as well as mesenchymal cells that give rise to periureteric bud smooth muscle cells. Similar to Six2-Cre, Tcf21-Cre expression remains active in epithelial nephric derivatives but unlike Six2, Tcf21 is also expressed in interstitial lineages. As a first step in validation, we compared and contrasted phenotypes in mice lacking b-catenin following Six2 versus Tcf21-Cre deletion. As expected, we observed many similarities but also important differences. Given the broader expression domain of Tcf21, we observed hydroureters and a smooth muscle cell phenotype with down-regulation of Bmp4. Similarly, gain of function studies of bcatenin provided similarities but also some differences, with Tcf21-Cre causing midline fusion of the kidneys in b-catenin GOF mutants.
How does Tcf21-Cre compare to the Pax3-Cre? Similar to Tcf21-Cre, Pax3-Cre reportedly excises in both epithelial and interstitial lineages that derive from cap and stromal metanephric mesenchyme. However, the extrarenal sites of Pax3-Cre activation are largely different, and include derivatives of neural crest and somites, such as dorsal root ganglia, skeletal muscle, adrenal medulla, and some subsets of colon epithelium, which do not overlap with the mesenchymal distribution of Tcf21-Cre expression [6,7].
An additional benefit of Tcf21-Cre includes the method of generation of the Cre line through homologous recombination in ES cells. This allows faithful recapitulation of expression pattern ensuring stable expression over subsequent breedings. Randomly inserted or BAC transgenic lines may or may not provide this asset. It is worthwhile to note that Tcf21 is not expressed in intermediate mesoderm [33,34], which suggests that the GFP expression of Tcf21-Cre; Z/EG mice doesn't include an excision prior to the development of metanephros. In addition, we noted mosaicism of the Tcf21-Cre expression and excision. Although this might be viewed as a negative (incomplete excision), we found it to be an asset in our experiments. As demonstrated with our proof of principle studies in this report, a wide spectrum of phenotypes can be observed in each litter ranging from severe to mild, allowing a more complete picture of gene function-i.e. permitting a dose response experiment in a single litter. Finally, the extrarenal expression of Tcf21 in other mesenchymal tissues allows insight from other organs. Although extrarenal expression of a Cre driver strain is often viewed as a weakness, it may provide additional information for the investigator. In this study, although we observed robust extrarenal phenotypes in all of the mutants, they did not preclude analysis of phenotype in kidney. One of the reasons for this is the relatively late onset of Tcf21 expression at mid-gestation. In turn, the development of aggressive embryonic sarcomas in both b-catenin GOF and Ptch1 LOF mutants points  [20].
Relatively normal nephron induction [21].  to a common genetic pathway -providing direction for future studies. Utility of Tcf21-Cre strain is further underscored by the phenotypes observed in the Ptch1 mutants. Although a role for Shh pathway in renal cystic diseases is supported by dysregulation of Gli1 in nephronophthisis , there has not been any direct genetic evidence [35,36]. Here, we show the development of a dramatic cystic renal phenotype in mice lacking the Ptch1 receptor with data to support a role for proliferation. Interestingly, we did not detect Shh pathway activation in the developing nephrons of Ptch1 fl/ fl ;Tcf21-Cre mice. It remains to be determined whether the nephrons found in the mutant kidney are derived from wild type cells that escape Cre excision due to the inability of Ptch1 null cells to form nephrons.
In summary, we report a new Cre driver strain that provides robust excision in the metanephric mesenchyme. It provides researchers with a useful tool to delete genes broadly from multiple metanephric mesenchyme subpopulations, resulting in phenotypes in the stromal cells, the cap mesenchyme and the periureteric bud cells.