Characterization of a Novel Fibroblast Growth Factor 10 (Fgf10) Knock-In Mouse Line to Target Mesenchymal Progenitors during Embryonic Development

Fibroblast growth factor 10 (Fgf10) is a key regulator of diverse organogenetic programs during mouse development, particularly branching morphogenesis. Fgf10-null mice suffer from lung and limb agenesis as well as cecal and colonic atresia and are thus not viable. To date, the Mlcv1v-nLacZ-24 transgenic mouse strain (referred to as Fgf10LacZ), which carries a LacZ insertion 114 kb upstream of exon 1 of Fgf10 gene, has been the only strain to allow transient lineage tracing of Fgf10-positive cells. Here, we describe a novel Fgf10Cre-ERT2 knock-in line (Fgf10iCre) in which a Cre-ERT2-IRES-YFP cassette has been introduced in frame with the ATG of exon 1 of Fgf10 gene. Our studies show that Cre-ERT2 insertion disrupts Fgf10 function. However, administration of tamoxifen to Fgf10iCre; Tomatoflox double transgenic embryos or adult mice results in specific labeling of Fgf10-positive cells, which can be lineage-traced temporally and spatially. Moreover, we show that the Fgf10iCre line can be used for conditional gene inactivation in an inducible fashion during early developmental stages. We also provide evidence that transcription factors located in the first intron of Fgf10 gene are critical for maintaining Fgf10 expression over time. Thus, the Fgf10iCre line should serve as a powerful tool to explore the functions of Fgf10 in a controlled and stage-specific manner.


Introduction
Fibroblast growth factor 10 (Fgf10) signals through its epithelial receptor Fgfr2b and is critical for branching morphogenesis in the mouse embryo. Loss of function of either Fgf10 or Fgfr2b by genetic deletion leads to total failure of lung and limb formation as well as cecal and colonic atresia [1,2,3,4,5]. In addition to its vital role during development, Fgf10 over-expression attenuates bleomycininduced pulmonary fibrosis in mice [6].
Using enhancer-trap strategies, the Mlcv1v-nLacZ-24 transgenic mouse line (hereby referred to as Fgf10 LacZ ) was characterized. The LacZ cassette was mapped and found to be integrated within regulatory elements lying 114 kb upstream of the endogenous Fgf10 coding sequence [7]. LacZ expression pattern observed in these mice closely mimics Fgf10 expression in the heart, lung, gut, mammary glands and brain [4,5,7,8,9,10,11]. The Fgf10 LacZ mouse reports Fgf10 expression and also results in a Fgf10 hypomorph as demonstrated by the phenotype of Fgf10 LacZ/2 (one copy of Fgf10 null and one copy of Fgf10 LacZ ) embryos. These embryos display decreased lung branching [12] as well as mammary gland [10] and limb abnormalities (Hajihosseini and Bellusci, unpublished results) consistent with decreased Fgf10 expression.
Until recently, the Fgf10 LacZ transgenic line has been the only available tool that can be used to transiently trace Fgf10-positive cells in mice. Here, we describe a novel Fgf10 iCre knock-in line where a tamoxifen-inducible Cre recombinase (Cre-ERT2-IRES-YFP) has been inserted in frame with the start codon of exon 1 of the endogenous Fgf10 gene (Fig.1). Our validation studies show that Fgf10 iCre is a loss of function allele for Fgf10. At the heterozygous state, this line is viable and fertile whereas the homozygous state is lethal. By crossing this line with a Tomato flox reporter line, we show that Fgf10-positive cells are specifically labeled upon tamoxifen administration in many tissues during embryonic and postnatal stages. Moreover, we demonstrate that crossing Fgf10 iCre driver line with responder lines carrying 'floxed' genes allows inducible conditional gene inactivation, but to a limited extent. This is because the expression of Cre from the recombined Fgf10 locus reveals mosaicism in the labeled tissues and tends to decline with developmental progression. This is probably due to disruption of regulatory elements downstream of exon 1 where the cassette has been initially inserted. The mosaic recombination pattern observed in this genetic model is a valuable tool for clonal analysis and lineage tracing of Fgf10positive cells.
59 and 39 homology regions (10.3 kb) for Fgf10 were amplified by PCR using DNA from a Fgf10-specific BAC clone derived from a BAC library of 129Sv/AB2.2 mouse strain (Wellcome Trust Sanger Institute). The isolated sequences were used to design the targeting construct. The linearized targeting construct was transfected into TVB2 mouse ES cells (following the electroporation procedure: 5610 6 ES cells in presence of 40 mg of linearized plasmid, 260 Volt, 500 mF). Positive selection was started 48 hours after electroporation by addition of 200 mg/mL of G418. G418-resistant colonies were selected based on cell growth and morphology. One electroporation session was performed and a total of 209 clones were isolated and amplified in 96-well plates. Duplicates of 96-well-plates were made; one copy was frozen down and the other copy was amplified on gelatine to be used for genomic DNA preparation. Clones were screened by PCR for homologous recombination and verified by Southern blotting. Altogether, PCR, sequencing and Southern blot screening allowed the characterization of 10 recombined clones, 7 of which were suitable for blastocyst injection. Recipient blastocysts were isolated from pregnant C57BL/6J females (Health status VAF -Virus Antibody Free). Based on screening results and morphological criteria, ES cell clones were injected into C57BL/6J blastocysts. Injected blastocysts were then re-implanted into OF1 pseudo-pregnant females and allowed to develop to term. Four of the highly chimeric males generated were selected to breed with C57BL/6J Flp-deleter mice (health status SOPF -Specific and Opportunist Pathogen Free) to allow the germline excision of the neomycin selection cassette.

Mice and Tamoxifen Administration
Tomato flox/flox reporter mice (B6;129S6-Gt(ROSA)26Sor tm9(CAG-tdTomato)Hze /J) [13] were purchased from Jackson lab and Fgf10 flox/flox mice [14,15] were a kind gift from Professor Suzanne L. Mansour (University of Utah, USA). Embryonic day 0.5 (E0.5) was assigned to the day when a vaginal plug was detected. Mice were housed in an SPF environment. Animal experiments were approved by the Regierungspraesidium Giessen (Project ID763, approval number RP GI/20-Nr38/2011, 548_GP) and by the Animal Research Committee at Children's Hospital Los Angeles (Protocol number 31-11) in strict accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals of the National Institute of Health. The approval identification for Children's Hospital Los Angeles is AAALAC A3276-01.
Tamoxifen stock solution was prepared by dissolving tamoxifen powder (T5648, Sigma, Schnelldorf, Germany) in corn oil at a concentration of 20 mg/mL at room temperature. Pregnant females carrying Fgf10 iCre/+ ; Tomato flox/+ embryos received a single intra-peritoneal (IP) injection of 0.075 mg tamoxifen per gram of body weight. For continuous tamoxifen exposure, pregnant females were fed tamoxifen-containing food (0.4 g of tamoxifen per kg of food) (Altromin, Lage, Germany). Dissected embryos were examined using Leica M205 FA fluorescent stereoscope (Leica, Wetzlar, Germany) and images were acquired using Leica DFC360 FX camera. Cecum lengths were measured using Leica's LAS AF software and digit peripheries were measured using Image J software (NIH, USA). Figures were assembled in Adobe Photoshop CS5.

X-Gal Staining
X-Gal staining was performed as previously described [16]. Briefly, Fgf10 LacZ/+ lungs were dissected out in Hank's solution (HBSS) from E18.5 embryos, shortly fixed in 4% PFA, washed in PBS and incubated with LacZ buffer solution for 10 min. Then, they were incubated with LacZ buffer solution containing 40 mg/ mL X-Gal (Sigma) at 37uC overnight.

Quantitative Real-time PCR and Statistical Analysis
Freshly isolated embryos and lungs were lysed and RNA was purified using RNeasy kit (Qiagen, Hilden, Germany). E11.5 and E13.5 embryos were homogenized using QiaShredder columns (Qiagen) whereas lungs from older tissues were homogenized using Bullet Blender Blue (Next Advance, NY, USA). 1 mg of RNA was used for cDNA synthesis using Quantitect Reverse Transcription kit (Qiagen). Primers and probes for Fgf10, Cre and b-actin were designed using Universal ProbeLibrary Assay Design center (Roche Applied Science, available online at https://www.rocheapplied-science.com/sis/rtpcr/upl/index.jsp?id = UP030000). More details about the used primers and probes can be found in Table S1. Quantitative real-time PCR (qPCR) was performed using LightCycler 480 real-time PCR machine (Roche Applied Science). Samples were run in triplicates using b-actin as a reference gene and the DDCT method was used for relative quantification. Data were assembled using GraphPad Prism software (GraphPad Software, USA) and statistical analyses were performed using Student's t-test (for comparing two groups) or One-way ANOVA (for comparing three or more groups). Data were considered significant if P,0.05.

Results
Generation of Fgf10 iCre (C57BL/6-Fgf10Cre-ERT2-YFP/J) Driver Line 129Sv ES cells were electroporated with a targeting vector containing the 59 end of exon 1 of Fgf10 open reading frame (Fig.1A, B). Immediately downstream of the start codon is the coding sequence of a tamoxifen-inducible form of Cre recombinase (Cre-ERT2) [17], followed by IRES sequence, enhanced YFP and Neomycin-resistance gene (Neo) respectively. Resistant ES cell clones were selected, screened by PCR and then verified by Southern blotting. Selected ES clones were injected into C57BL/ 6J blastocysts to generate chimeric pups (Fig. 1C). Chimeras were then crossed with C57BL/6J mice ubiquitously expressing Flp recombinase to generate heterozygous Fgf10 iCre knock-in mice where the Neo cassette was totally excised (Fig. 1D).

Insertion of the Cre-ERT2 Cassette in Exon1 Results in Fgf10 Loss of Function
To determine whether the insertion of Cre-ERT2 in the endogenous Fgf10 locus led to loss of function of Fgf10, Fgf10 iCre heterozygous animals were crossed together and embryos were harvested at E12.5. Fgf10 iCre homozygous embryos suffered from lung and limb agenesis in addition to cecal and colonic atresia, consistent with complete loss of function of Fgf10 ( Fig

Inducible Conditional Gene Inactivation
To test the potential use of this line for gene inactivation studies, Fgf10 iCre/+ ; Tomato flox/flox mice were crossed with mice carrying a 'floxed' version of Fgf10 (Fgf10 flox/flox ) [14]. Cre activity was induced by tamoxifen food from E8.5 to E14.5. By using the Tomato flox reporter as readout of Cre activity, recombination was observed in the limbs, lungs and ceca (Fig. 5C, D, G, H, L; n = 3; penetrance = 100%). Fgf10 iCre/flox ; Tomato flox/+ embryos suffered from diverse developmental abnormalities (n = 3; penetrance = 100%). A formation defect characterized by webbed digits was observed at the level of the forelimbs as compared to their Fgf10 +/flox ; Tomato flox/+ control littermates. The phenotype was quantified by measuring the periphery of the digits (Fig. 5H vs. E, M; n = 3; P,0.0001). Hindlimbs did not show any obvious abnormalities. The lung showed an abnormal shape and suffered from branching simplification as illustrated by the reduced number of terminal buds in the accessory lobe (Fig. 5C vs. B, G vs. F, J; n = 3; P,0.05). The cecum, on the other hand, was shorter
The 3 kb region downstream of Fgf10 transcriptional start site, which was deleted upon homologous recombination, was analyzed for putative regulatory elements by comparing transcription factor binding sites conserved between mouse and human (using rVista, http://rvista.dcode.org). The analysis revealed the presence of several conserved transcription factor binding site-dense regions (Fig. 7A). Interestingly, the 716 bp stretch at the 39 end of the deleted region contains putative binding sites for Smad4, Nkx2.5, Tbx5 and Isl1. Furthermore, the deleted region was analyzed for lung-related histone modifications using UCSC Genome Browser (http://genome.ucsc.edu). The analysis revealed the presence of a dense H3K4me3 (histone H3 trimethylated at Lys4) modification site overlapping exon 1-intron 1 boundary (Fig. 7B).

Discussion
Fgf10 is a morphogen that is crucial for the embryonic development of many organs including the lung, duodenum, cecum, colon, pancreas, limb and mammary placodes. Fgf10 knockout mice suffer from agenesis of the corresponding organs, thus limiting any real use of these mice in studying the behavior or lineage commitment of Fgf10-positive cells. On the other hand, due to the stability of b-Galactosidase, the Fgf10 LacZ transgenic line allows non-inducible transient tracing of Fgf10-positive cells in many organs. However, the gold standard for an optimal lineage tracing tool relies on the inducibility of constitutive cell labeling via the controlled activation of Cre recombinase. Then, short-and long-term lineage commitment can be investigated. Therefore, the data obtained with the Fgf10 LacZ line should always be validated using the proper Cre inducible line. Nevertheless, this line has helped identify the progeny of Fgf10-positive cells in the heart, lung, gut, mammary glands, brain [4,5,7,8,9,10,11] and limbs [Hajihosseini and Bellusci, unpublished results]. Thus, the Fgf10 iCre knock-in line characterized in this paper represents a novel tool that bypasses the pitfalls associated with the Fgf10 LacZ transgenic line. This is because Fgf10-positive cells are exclusively labeled  upon tamoxifen injection, after which labeled cells can be traced and characterized. A previous report has shown that intraperitoneally administered tamoxifen reaches effective levels in the embryonic circulation within 12 hours, after which it is cleared from the pregnant female within 12 hours [18]. This 24-hour time frame allows accurate labeling of Fgf10-positive cells.
In this study, we induced recombination in pregnant females carrying Fgf10 iCre/+ ; Tomato flox/+ embryos at E15.5. At E18.5, lineage-labeled cells were found in the ears of the embryos (Fig. 3A'') and this observation is consistent with previously published reports that describe the role of Fgf10 in inner ear formation [19]. Moreover, labeled cells were observed in the skin (Fig. 3B'') where Fgf10 is known to be expressed in the dermal papillae [20]. In the lung, labeled cells were dispersed throughout the mesenchyme (Fig. 4A'). However, the signal was more pronounced in interlobular septae (Fig. 4B'). Further studies need to be performed in order to explain this observation. As for the trachea, labeled cells were arranged in discrete ring-like structures (Fig. 4C'). This pattern is consistent with previously published reports about the role of Fgf10 in tracheal cartilage ring formation in mice [21,22]. We also used Fgf10 LacZ/+ lungs from the same stage (E18.5) to validate our data. As expected, X-Gal staining of Fgf10 LacZ/+ lungs revealed similar sites of Fgf10 expression (Fig. 4A''-C'').
For functional inactivation studies, we generated E14.5 Fgf10 iCre/flox ; Tomato flox/+ embryos in which Cre had been activated since E8.5. By virtue of the embedded Tomato flox reporter, we were able to locate recombination sites and thus, we restricted our diagnostic to those specific sites. An abnormal phenotype, characterized by webbing of the digits, was observed in the forelimbs (Fig. 5H vs. E, M). This finding agrees with previous reports about the involvement of Fgf10 misregulation in digit abnormalities such as syndactyly [23], [Al Alam and Bellusci, in revision]. Moreover, Fgf10 iCre/flox ; Tomato flox/+ lungs showed clear developmental abnormalities characterized by a deformed shape as well as branching simplification (Fig. 5C vs. B, G vs. F, J). This phenotype is similar to that obtained by Abler et al. who used a general mesenchymal driver line (Dermo1 Cre ) to perform conditional gene inactivation of Fgf10 [14]. Last but not least, our research group has previously demonstrated that the gut is a major site of Fgf10 expression and that hypomorphic Fgf10 embryos (Fgf10 LacZ/2 ) suffered from cecal atresia [4]. Our data agrees with this finding as Fgf10 iCre/flox ; Tomato flox/+ embryos suffered from a hypomorph-like phenotype at the level of the cecum upon partial loss of function of Fgf10 (Fig. 5L vs. I, K).
One drawback of the Fgf10 iCre line is the mismatch between Cre and Fgf10 expression levels as examined at different developmental stages by qPCR (Fig. 6A). To understand the inefficient expression of Cre from the Fgf10 iCre locus, we performed a bioinformatic analysis for the 3 kb deleted sequence downstream of ATG codon of exon 1 of Fgf10 gene using online tools (Fig. 7). The analysis revealed the presence of H3K4me3 site overlapping exon 1-intron 1 boundary. This histone modification is known to be associated with the 59 end of actively transcribed genes [24]. On the other hand, several conserved binding sites for lung-related transcription factors were detected, including binding sites for Smad4, Nkx2.5, Tbx5 and Isl1. Tbx5 is thought to directly control Fgf10 expression in the lung mesenchyme [25]. Moreover, a recent report has shown that ISL1 regulates FGF10 transcription within the second heart field by binding to an enhancer element in intron 1 of FGF10 gene [26]. Altogether, these data suggest the presence of key regulatory elements in intron 1 critical for the maintenance of Fgf10 expression over time. Nonetheless, this line allows to target a subset of Fgf10-positive cells after birth allowing to study the lineage commitment of these cells during homeostasis and during the repair process after injury.
Alternative strategies that would bypass potential consequences on Cre expression levels include fusion of Cre-ERT2 in frame with the ATG codon of Fgf10 without deleting intronic sequences, thus preserving putative regulatory elements and matching Cre expression levels to endogenous Fgf10. Another approach would be the insertion of IRES-Cre-ERT2 in the 3' UTR downstream of Fgf10 stop codon. In this case, Fgf10 would still be expressed at physiological levels but Cre expression levels would depend greatly on the activity of the IRES element.
In conclusion, the Fgf10 iCre knock-in line described in this paper is a novel tool that allows labeling and tracking of Fgf10-positive cells. The mosaic recombination pattern, as illustrated by the Tomato flox reporter, makes this line convenient for clonal analysis, especially when a partial loss of function approach is desired. This tool can potentially bring insight on the role of Fgf10-positive progenitor cells in development and repair after injury.

Supporting Information
Table S1 Primers and probes used for quantitative real-time PCR and designed using Roche's Universal ProbeLibrary Assay Design center. (DOCX)