Generation and characterization of a Dkk4-Cre knock-in mouse line

Ectodermal appendages in mammals, such as teeth, mammary glands, sweat glands and hair follicles, are generated during embryogenesis through a series of mesenchymal-epithelial interactions. Canonical Wnt signaling and its inhibitors are implicated in the early steps of ectodermal appendage development and patterning. To study the activation dynamics of the Wnt target and inhibitor Dickkopf4 (Dkk4) in ectodermal appendages, we used CRSIPR/Cas9 to generate a Dkk4-Cre knock-in mouse (Mus musculus) line, where the Cre recombinase cDNA replaces the expression of endogenous Dkk4. Using Cre reporters, the Dkk4-Cre activity was evident at the prospective sites of ectodermal appendages, overlapping with the Dkk4 mRNA expression. Unexpectedly, a predominantly mesenchymal cell population in the embryo posterior also showed Dkk4-Cre activity. Lineage-tracing suggested that these cells are likely derived from a few Dkk4-Cre-expressing cells in the epiblast at early gastrulation. Finally, our analyses of Dkk4-Cre-expressing cells in developing hair follicle epithelial placodes revealed intra- and inter-placodal cellular heterogeneity, supporting emerging data on the positional and transcriptional variability in placodes. Collectively, we propose the new Dkk4-Cre knock-in mouse line as a suitable model to study Wnt and DKK4 inhibitor dynamics in early mouse development and ectodermal appendage morphogenesis.

of β-catenin, which together with the lymphoid enhancer factor/T-cell factor (Lef/Tcf), initiate the transcription of Wnt target genes (Nusse & Clevers, 2017).To prevent the constitutive accumulation of β-catenin, Wnt is antagonized by a set of inhibitors, including the secreted proteins from the Dickkopf (DKK) family 1-4 that bind to the LRP5/6 co-receptors, leading to the assembly of the β-catenin destruction complex (Mao et al., 2001;Zorn, 2001).Together with the transmembrane protein Kremen (KREM), DKKs and LRP6 form a ternary complex, which ensures the removal of LRP6 from the cell surface, thereby inhibiting Wnt activity (Mao et al., 2002).
In this context, DKK4 has been shown to synergistically act with KREM to exert its Wnt inhibitory functions (Patel et al., 2018).Dkk4 is also a canonical Wnt target and proposed to act in a negative feed-back loop with Wnt signaling during ectodermal appendage morphogenesis (Bazzi et al., 2007;Sick et al., 2006).Moreover, stratified epitheliaspecific overexpression of Dkk4, whose mRNA is normally restricted to ectodermal placodes (Bazzi et al., 2007;Sick et al., 2006), affects secondary HF development (Cui et al., 2010;Sima et al., 2016), and decreases the number of sweat glands (Cui et al., 2014).Studying this interplay between Wnt and DKK4 during development is important to understand how negative feedback loops pattern tissues and organs.
At the initiation of gastrulation, the posterior part of the epiblast undergoes an epithelial-to-mesenchymal transition at the primitive streak, whose location is determined by Wnt activity (Mohamed et al., 2004).
The gastrulation movements organize the germ layers, where the epiblast cells that do not ingress through the primitive streak adopt an ectodermal fate, and the pre-somitic mesoderm develops in the posterior region (Bardot & Hadjantonakis, 2020).A hallmark of organogenesis is the formation of the somites that derive from the pre-somitic mesoderm and develop sequentially to give rise to the future bones, muscles, connective tissue and mesenchyme including a large part of the skin dermis (Tam, 1981).However, compared to Wnt signaling, which has been extensively studied during development and in other contexts, the regulation of Dkk4 expression in early mouse (Mus musculus) development warrants further investigation and requires the generation of new models and tools.
In this study, we describe the generation of a Dkk4-Cre knock-in mouse line to investigate Dkk4 transcriptional regulation in response to Wnt signaling activity during early development and in ectodermal appendage formation, particularly in HFs.We use lineage-tracing of the Dkk4-Cre-reporters to show that the reporter is active in ectodermal appendages.Our data also reveal that a handful of cells are induced in the epiblast around the initiation of gastrulation and likely give rise to posterior somites and mesenchymal cells in the posterior region of the embryo.Notably, our data also suggest cellular heterogeneity within and among the HF placodes.

Dkk4-Cre-expressing cells
To study the Wnt-DKK4 signaling interplay in ectodermal appendage formation, we used CRISPR/Cas9 to engineer a Cre-recombinase knock-in mouse line that utilizes the Dkk4 endogenous promoter.The Cre cDNA was successfully inserted into the start codon of exon1 in the Dkk4-locus and is expected to disrupt and replace the expression of the Dkk4 open reading frame (ORF) (Figure 1a,b and Figure S1).To assess the Cre-recombinase expression, we crossed heterozygote Dkk4-Cre mice with a dual-reporter line that expresses tdTomato or EGFP at the cell membrane (mTmG) (Muzumdar et al., 2007), or in the nucleus (nTnG) (Prigge et al., 2013).Upon induction of the Cre recombinase in Dkk4-expressing cells, the tdTomato cDNA, which is flanked by two loxP sites, is excised and EGFP is expressed instead (Figure 1c).Thus, the cells that are derived from Dkk4-Cre-expressing progenitors will maintain the EGFP signal and their lineage can be traced over time.To test the specificity of the Dkk4-Cre line, we investigated the expression of EGFP at ectodermal appendages, where Dkk4 mRNA has been shown to be expressed using in situ hybridization (Figure 2) (Bazzi et al., 2007;Sick et al., 2006).EGFP-positive cells localized at the prospective sites of the primordia of the dental field (embryonic day [E]10.5),vibrissae follicles (E12.5),mammary glands (E13.5),pelage or back-skin HFs (E14.5) and eccrine glands on the palmo-plantar skin (E18.5),but not cranial placodes (data not shown), confirming that the activity of the Cre recapitulated the Dkk4 mRNA expression (Figure 2a,b and Table 1).At E18.5, we also observed EGFP-positive signals in the heart (Figure S2a), kidney (white arrow, Figure S2b), adrenal gland (yellow arrow, Figure S2b) and other organs including muscles (data not shown), indicating that Dkk4-expressing cells contribute to other tissues and organs besides ectodermal appendages.

| A subset of Dkk4-Cre-expressing cells arises around gastrulation
In addition to the expression of the Dkk4-Cre in ectodermal appendages, lineage-tracing in Dkk4-Cre reporter embryos suggested that a developmentally earlier cell population expressed Dkk4-Cre (Figure 3).Specifically, 90% of dissected Dkk4-Cre KI/wt -reporter embryos showed evident EGFP-positive cells in the posterior region between E6.5 and E15.5 (Figure 3a,b), while the remaining 10% exhibited a rather widespread distribution of EGFP-positive cells (Figure S3a).Because of the mostly consistent pattern of the EGFP signals and given that Dkk4 in situ hybridization did not show Dkk4 mRNA expression in the same region in early embryos (E8.5 and beyond, data not shown), we reasoned that the progenitors of these posterior cells arose earlier during development.In agreement, lineage-tracing revealed a few EGFPpositive cells in the epiblast of early gastrulating mouse embryos at E6.5 and E7.5, some of which co-localized with the primitive streak and mesoderm marker Brachyury antibody signal (Figure 3b and Figure S3b) (Wilkinson et al., 1990), and correlated with increased EGFP-positive cells in the pre-somitic mesoderm at E8.5 (Figure 3b, right), as well as in the posterior somites between E9.5-E10.5 (Figure 3a,right).Sagittal sections of Dkk4-Cre:nTnG embryos at E14.5, with predominantly posterior expression, indicated that in the anterior region, EGFP-positive cells were mainly found in the epithelial skin layer and mostly localized to the HF placodes (Figure 3c).
However, in the posterior region, the EGFP-positive cells were also found in the dermis (Figure 3c), likely representing the same lineage of the earlier mesenchymal population.In addition, a section through the skin and subcutis in the posterior region at E18.5, revealed EGFPpositive cells in the connective tissue layers, muscle as well as in the dermis and epidermis (Figure 3d and Table 1).

|
The hair follicle epithelium is not solely derived from Dkk4-Cre-expressing cells Given the wide-spread expression of the Dkk4-Cre lineage in the posterior region of most embryos, we focused on the anterior skin region, which showed more restricted epithelial expression that appeared to arise later in development.In particular, we assessed Dkk4-Cre reporter expression in developing HFs, as a representative of ectodermal appendages.Interestingly, EGFP-positive cells were detected in the basal layer of E12.5 and E13.5 epidermis, even before any morphological signs of HF placode formation (Figure 4a).
As expected at E14.5 and E15.5, the EGFP-positive cells were found in HF placodes (yellow arrows in Figure 4a), but also occasionally in patches of the inter-follicular epidermis (white arrows in Figure 4a).Intriguingly, a closer inspection of the Dkk4-Cre reporter expression in HF placodes revealed that not all placodal cells were EGFPpositive, suggesting that a subset of cells in the placode did not F I G U R E 1 Cre insertion into the Dkk4-locus using CRISPR/Cas9.(a) A schematic representation of single-stranded (ss) DNA template of $1.6 Kb encoding Cre cDNA, a nuclear localization signal (NLS) of the Simian Virus (SV) 40T antigen and a bovine growth hormone polyadenylation signal (bGH Poly(A)).Cre was inserted into exon1 of Dkk4 via homology-directed repair with arms of homology of 150 bp.The primers used for genotyping and sequencing are depicted with arrows (fwd and rev).The 5 0 untranslated region, UTR, and Intron1, In1, are also shown.(b) PCR analyses of the wild-type (wt) and knock-in (KI) alleles in some founder animals (1-10).The 368 bp PCR-product from (fwd1 + rev1) corresponds to the wt band (top band in founder 3, for example).The Cre insertion was verified at the 5 0 end by the 215 bp KI-band (fwd1 + rev2, bottom band in founder 3 on the left), and at the 3 0 end by the 231 bp KI-band (fwd2 + rev1, bottom band in founder 3 on the right), as represented in (a), and the products were confirmed by sequencing (Figure S1).Founder 3, red box, was used for further breeding and experiments.M is for DNA marker or ladder.(c) Scheme of Dkk4-Cre KI/wt male mice crossed to mTmG tg/tg or nTnG tg/tg females.The mTmG/nTnG alleles are at the ROSA26 (R26) locus and driven by the chicken beta-actin core promoter (ACTB) (Muzumdar et al., 2007;Prigge et al., 2013).The tdTomato cDNA is flanked by LoxP sites and excised in cells that express Cre.The embryonic progeny with the genotype Dkk4-Cre KI/wt ;mTmG tg/wt or Dkk4-Cre KI/wt ;nTnG tg/wt and EGFP-positive expression were used for lineage-tracing analyses.
descend from Dkk4-Cre-expressing cells.Co-staining with the placode markers LHX2 or SOX9 showed that, on average, around half of the LHX2-or SOX9-expressing placodal cells were EGFP-negative (arrows in Figure 4b,c).In fact, the wide distribution of the double positive cells indicated intra-and inter-placodal heterogeneity (Figure 4c).In confirmation of these findings, in growing HFs at post-natal day (P) 8, only a fraction of cells in the matrix and HF layers seemed to be derived from Dkk4-Cre reporter-expressing cells (Figure 4d).The data suggest a heterogeneous origin of the HF epithelium that includes and goes beyond the Dkk4-Cre reporterexpressing cells in the placode.

| DISCUSSION
We describe the generation of a Dkk4-Cre knock-in mouse line that reproduces the expression of Dkk4 mRNA in ectodermal appendages (Bazzi et al., 2007), using Cre-reporter lines.Unexpectedly, the lineage-tracing strategy uncovered a rare population of individual cells around early gastrulation (E6.5-E7.5)that expresses Dkk4-Cre.The progeny of these cells mostly occupies the pre-somitic mesoderm (E8.5) and subsequently the posterior somites (E9.5), likely giving rise to the mesoderm-derived tissues in the posterior region of developing embryos from E10.5-E18.5.In addition, the data suggest that these cells contribute mostly to mesodermally-derived organs such as the heart and kidney.Interestingly, Wnt is active in pre-gastrulating E6.5 embryos (Maretto et al., 2003;Mohamed et al., 2004), covering the Dkk4-Cre expression domain.Dkk4 is a direct target gene of canonical Wnt signaling and is hypothesized to act in a negative feedback loop to fine tune Wnt activity in ectodermal appendages (Bazzi et al., 2007;Sick et al., 2006).The site of the primitive streak and the induction of the mesoderm is largely dictated and coordinated by the canonical Wnt pathway (Mohamed et al., 2004).Therefore, it is highly likely that the early Dkk4-Cre-expressing cells at E6.5 are induced by high Wnt activity in these cells.One caveat of our knockin/knockout strategy is that reducing the dose of DKK4 in heterozygous Dkk4-Cre embryos is expected to reduce the inhibition on canonical Wnt signaling, likely potentiating the Wnt pathway and enhancing the expression of Dkk4-Cre, which itself is a target of Wnt, resulting in ectopic Cre expression in this rare cell population and perhaps others in later development (see below).In addition, $10% of the embryos at different stages showed a more wide-spread and mosaic pattern of Dkk4-Cre reporter expression, we currently do not have a definitive explanation for this observation.One possibility is that the mixed genetic background might impact on the Dkk4-Cre expression around

Adrenal glands Unknown Yes
Note: The Dkk4 mRNA expression was assessed using colorimetric-based whole-mount in situ hybridization, and the Dkk4-Cre was analyzed using nTnG or mTmG reporters.
E6 producing more or less recombined cells depending on the contribution of FVB and C57BL/6.
The ectoderm and its lineages are derived from the anterior epiblast, which do not express Dkk4-Cre around the beginnings of gastrulation (E6.5-E7.5).This finding suggest that the vast majority of ectodermal Dkk4-Cre-driven reporter expression arises later in embryogenesis coinciding with the Dkk4 mRNA and is predominantly in the placodes of developing ectodermal appendages, including HFs (Bazzi et al., 2007;Sick et al., 2006).The Dkk4-Cre-driven reporter expression in the inter-follicular epidermis between E12.5 and E14.5 suggests that the Dkk4 expression in the skin epithelium is not restricted to the placode as initially suggested by colorimetric in situ hybridization, a technique with insufficient resolution to definitively ascertain single-cell expression.We also observed that not all cells in the HF placode expressed Dkk4-Cre.The basal cells of the placode are labeled by LHX2 expression (Rhee et al., 2006), while the SOX9-positive cells mostly occupy the suprabasal part and periphery of the placode (Morita et al., 2021;Nowak et al., 2008;Ouspenskaia et al., 2016).In vibrissae HFs, Morita et al. suggested the existence of four genetically distinct concentric layers that form the HF placode and give rise to the different compartments of the mature HF (Morita et al., 2021).The Dkk4 mRNA was mainly detected in the center of vibrissae HF placodes, while Sox9 localized at the periphery, similar to our observations for the back-skin (Morita et al., 2021).The heterogeneity of the Dkk4-Credriven reporter expression in the HF placode is also reflected in later stages of HF morphogenesis, with cells in the matrix and HF layers labeled in a mosaic manner.Our data suggest heterogeneity in the placodal cell fates that give rise to the different parts of the HF epithelium, as has been recently shown by time-lapse imaging and single-cell-RNA-sequencing (Morita et al., 2021;Sulic et al., 2023).
Collectively, the Dkk4-Cre mouse line provides a new tool to study Wnt signaling induction of the DKK4 inhibitor in early gastrulation and ectodermal appendage morphogenesis, using lineage-tracing and conditional knockouts.primers rev3 (5 0 -CGAGTGATGAGGTTCGCAAG-3 0 ) and fwd3 (5 0 -CGCTGGAGTTTCAATACCGG-3 0 ) were used for sequencing.The founder whose sequence is shown in Figure S1 was mainly used in this study and backcrossed to WT FVB/NRj for two to three generations.

| Embryo dissection
The  (Wilkinson, 1998).The embryos were photographed using a DFC 450C camera fitted onto an M165 FC stereomicroscope (Leica Microsystems, Wetzlar, Germany).Germany) and image acquisition was performed using LAS X (Leica Microsystems) or Zen lite (Carl Zeiss AG) software, respectively.

| Whole-embryo clearing
Embryos from E12.5 to E15.5 were cleared using ethyl cinnamate (Sigma-Aldrich) as previously described (Klingberg et al., 2017).Fixed embryos were washed three times for 10 min in 1Â PBS and dehydrated in an ethanol series (50%, 70%, then twice in 100%).Every dehydration step was for 1 h at room temperature with gentle horizontal rotation.After dehydration, the embryos were transferred to ethyl cinnamate and imaged, when they became completely transparent, on an SP8 confocal microscope (Leica Microsystems).

| Reproducibility
Each experiment was repeated independently at least three times to ensure reproducibility.

F
I G U R E 3 Dkk4-Cre-expressing cells are induced in the early epiblast and spread to the mesenchymal cells of the posterior embryo region.(a) Confocal images of E9.5-E15.5 (right to left) Dkk4-Cre KI/wt :mTmG tg/wt embryos with evident EGFP-expression in the posterior body region and hair follicles (scale bar: 1000 μm).Anterior (A) is up, posterior (P) is down, ventral (V) is to the left and dorsal (D) is to the right.(b) Confocal images of E6.5 -E8.5 Dkk4-Cre KI/wt :nTnG tg/wt embryos that were immune-stained with Brachyury antibody (scale bar: 200 μm).Dashed squares: single confocal Z-planes show co-localization of some EGFP-positive cells (yellow arrowheads) and Brachyury antibody signal likely in the mesoderm (M) (scale bar: 30 μm).In E6.5 and E7.5 embryos, anterior (A) is to the left, while the posterior (P) region, where the primitive streak (PS) is localized, is to the right.The proximal (Pr) region is up and marked by the extraembryonic region (Ex), whereas the distal (Di) region is down.A ventral view of an E8.5 embryo is shown along the anterior (A)-posterior (P) axis, with the notochordal plate (NP), the pre-somitic mesoderm (PSM) and somites labeled.(c) Sagittal section of an E14.5 Dkk4-Cre KI/wt ;nTnG tg/wt embryo (scale bar: 1000 μm) with anterior or posterior dorsal skin regions enlarged and rotated 90 counter-clockwise in the insets (scale bar: 20 μm).(d) Sagittal section of an E18.5 Dkk4-Cre KI/wt ;mTmG tg/wt embryo showing the different skin layers expressing EGFP at the posterior embryo region (scale bar: 50 μm).A dotted line demarcating the basement membrane between the epidermis and dermis is shown in (c) and (d).
embryos were dissected in cold 1Â phosphate buffer saline (1Â PBS; VWR, Radnor, PA, USA) with 0.4% bovine serum albumin (BSA; Sigma-Aldrich, St. Louis, MO, USA) (E6.5-E11.5)or in 1Â PBS (E12.5-E15.5),and fixed in 4% paraformaldehyde (PFA; Carl Roth, Karlsruhe, Germany) overnight at 4 C. Embryos were then washed in 1Â PBS, and either dehydrated in a methanol series to be used for in situ hybridization, or stored in 1Â PBS at 4 C until further processing.4.3 | Whole mount in situ hybridization cDNA from E12.5 embryos was used as a template to amplify and generate the Dkk4 in situ hybridization probe of $0.7 Kb and covering the entire coding region (Dkk4_ISH_forward: 5 0 -CTCCGAGAGACCA-GAGTGAC-3 0 and Dkk4_ISH_reverse: 5 0 -CACAACAACAAGTCCCGTGT-3 0 ).The PCR products were cloned into the pCR TM II dual promoter (T7 and SP6) vector (Invitrogen, Waltham, MA, USA) and the DIGlabeled probes were generated according to the manufacturer's recommendations (Roche Applied Science, Indianapolis, IN, USA).In situ hybridization was performed on embryos from different stages per detailed published protocols AppliChem, VWR) for 1 h at RT and mounted with Prolong™ Gold Antifade reagent (Cell Signaling Technology, Danvers, MA, USA).For whole-mount staining, embryos from E6.5 to E11.5 were fixed in 4% PFA for 2 h-overnight at 4 C, washed in 1Â PBS, then in buffer containing 0.1%-0.3%Triton™ X-100 and blocked with blocking buffer made of washing buffer with 10% serum.Primary antibody (Brachyury 1:200, #AF2085 R&D Systems, Minneapolis, MN, USA) and secondary antibody stainings were performed overnight at 4 C on 2 successive days.Embryos were washed three times for 30 min with washing buffer after each step and then mounted on 35 mm glass bottom dishes (Ibidi, Munich, Germany) using Vectashield ® mounting medium (Vectorlabs, Newark, CA, USA) and embedded in 1% low melting agarose (Lonza, Basel, Switzerland).Images were obtained using a SP8 confocal microscope (Leica Microsystems) or a Meta 710 confocal microscope (Carl Zeiss AG, Oberkochen, For sections, embryos were washed with 1Â PBS and cryoprotected in 30% sucrose (Sigma-Aldrich) in 1Â PBS overnight at 4 C.After embedding in Tissue-Tek ® (Optimal Cutting Temperature Compound,