Wt1-expressing cells contribute to mesoderm-derived tissues in intestine and mesentery in two distinct phases during murine embryonic development

The mesoderm gives rise to diverse tissues including the mesothelium, as well as the visceral and vascular smooth muscle. Previously, genetic lineage tracing based on the mesothelial marker Wt1, appeared to show that visceral mesothelial cells are the direct progenitors of vascular smooth muscle in the intestine. However, the timing and underlying developmental mechanisms regulating this lineage were not fully understood. Here, using a temporally-controlled Wt1-based genetic lineage tracing approach, we demonstrate that (i) the adult visceral mesothelium of the intestine only maintains itself and fails to contribute to other visceral tissues; (ii) the vascular and visceral smooth muscle in the developing intestine and mesentery arise earlier than and independently from the visceral mesothelium. Our results suggest that Wt1 is switched on and remains expressed in the visceral mesothelium from around E9.5 onwards throughout life, while earlier transient expression of Wt1 in the nascent mesoderm specifies the future vascular and visceral smooth muscle of intestine and mesentery.


Introduction
During embryonic development, vascular smooth muscle cells (VSMCs) and pericytes arise from different sources, including the neural crest, the somatic and the lateral plate mesoderm (Majesky, 2018, Roostalu and Wong, 2018, Wang et al., 2015. Hence, VSMCs / pericytes appear to form a mosaic with different developmental origins (Bentzon and Majesky, 2018). Recently, the molecular mechanisms by which VSMCs or pericytes differentiate from induced pluripotent stem (iPS) cells in vitro have been described (Kumar et al., 2017), however, it remains unclear how VSMCs or pericytes are specified or recruited in vivo to the developing vasculature.
The epicardium, the mesothelium of the heart, contributes VSMCs to the developing coronary vessels (Red-Horse et al., 2010, Zhou et al., 2008. Based on this observation, we previously tested the hypothesis that a similar developmental relationship exists between the visceral mesothelium and the vasculature of the intestine. The visceral mesothelium of the peritoneal cavity arises from the lateral plate mesoderm, where it forms a simple squamous epithelium covering all organs within the cavity (Wilm et al., 2005, Winters et al., 2012. The Wilms' tumour protein 1 (Wt1) is a key marker of mesothelia, and a transgenic mouse line expressing Cre recombinase under control of regulatory elements of the human WT1 gene (Tg(WT1-cre)AG11Dbdr, in short: Wt1-Cre) had been previously used to track mesothelial cells in mice (Wilm et al., 2005). In adult Wt1-Cre; Rosa26 LacZ/LacZ compound mutant mice, XGal-labelled vascular smooth muscle cells were found in the vasculature of the mesentery and intestine, as well as the heart and lungs (Que et al., 2008, Wilm et al., 2005. These findings led us to conclude that cells of the visceral mesothelium give rise directly to VSMCs in the intestine and mesentery (Wilm et al., 2005). Studies using a different Wt1-Cre line (Tg(Wt1-cre)#Jbeb) revealed a similar contribution of Wt1-expressing cells to the mesothelium and the visceral and vascular smooth muscle of the developing intestine (Carmona et al., 2013).
However, these genetic lineage tracing systems were unable to distinguish the time window during embryonic development or postnatal stages when cells had expressed Wt1, as once tagged, the cells were consequently irreversibly labelled. Therefore, it is unclear whether Wt1-expressing mesothelial cells give rise to VSMCs continuously throughout life, or only during a specific time window in embryonic development. Furthermore, it is not clear whether postnatal or even adult mesothelial cells in the peritoneal cavity generally contribute to the maintenance of the intestinal and parietal body wall. This is pertinent because of reports suggesting that mesothelial cells have the plasticity to differentiate into cells of different mesodermal lineages (Colunga et al., 2019, Karki et al., 2014, Mutsaers et al., 2015, Dauleh et al., 2016. To address these questions, we determined the spatio-temporal contribution of Wt1-expressing cells to the adult, postnatal and embryonic intestine and parietal peritoneum, as well as the VSMCs in the mesentery and intestinal wall. For this purpose, we used a temporally controlled and Wt1-driven Cre system (Wt1 tm2(cre/ERT2)Wtp ) which relies on tamoxifen administration to activate recombination and thus lineage tracing Munoz-Chapuli, 2016, Zhou et al., 2008). Our analysis revealed that Wt1-lineage traced mesothelial cells in neonatal or adult intestine and peritoneal body wall only contributed to maintenance of cells of the serosa (and mesenteric fat) and not to vascular cells or intestinal wall tissue. This was different in the heart, where coronary vasculature was labeled, indicating postnatal epicardial contribution. By contrast, Wt1-lineage tracing in the embryo indicated two stages of contribution: (i) Wt1-expressing cells solely gave rise to the visceral mesothelium of intestine and mesentery from around E9.5 onwards; and (ii) Wt1-expressing cells labeled at between E7.5 and E8.5 contributed to vascular and visceral smooth muscle cells in the intestine and mesentery in spatially restricted segments. Thus, our analysis describes two separate phases of lateral mesoderm differentiation which contribute to different lineages and indicates that Wt1 plays separate roles within each. Our data for the first time delineate the temporal events that lead to vascular and visceral smooth muscle cell formation and separately to visceral mesothelium development in the intestine and mesentery.

Adult Wt1-derived cells contribute to coronary but not intestinal vascular cells and maintain the visceral and parietal mesothelium
We examined whether mesothelial cells (MCs) contribute to the formation of vascular cells and other tissue components in the adult intestine. After a 2-4 week chase period in Wt1 CreERT2/+ ; Rosa26 LacZ/LacZ and Wt1 CreERT2/+ ; Rosa26 mTmG/+ mice we found lineage-traced cells in a patchy pattern in the intestinal serosa ( Figure 1A, B). In addition, we observed a recently described Wt1expressing submesothelial mesenchymal population (Buechler et al., 2019) in the serosa of the mesentery and body wall, and in the omentum (Supplementary Figure S1). We could not detect any contribution of labelled Wt1-expressing mesothelial cells to vascular smooth muscle in the mesentery or the intestine, or other cell types of body or intestinal wall, except for the previously reported contribution to mesenteric fat, which was confirmed by histological analysis ( Figure 1E, F, Supplementary Figure S2A, B) (Chau et al., 2014, Chau et al., 2011. Both the Wt1 CreERT2/+ ; Rosa26 LacZ/LacZ and Wt1 CreERT2/+ ; Rosa26 mTmG/+ reporter systems gave very similar results, as revealed by direct comparison of GFP and XGal staining in the intestinal mesothelium in Wt1 CreERT2/+ ; Rosa26 LacZ/mTmG mice (Supplementary Figure S3).
To determine whether the previously reported contribution of MCs to the intestinal vascular smooth muscle compartment is controlled by a slow process of tissue homeostasis, we performed a longterm chase experiment of two and six months, respectively. Our data revealed that even six months after the tamoxifen pulse, coverage of the visceral mesothelium with labelled cells was still patchy, and no contribution to intestinal vascular smooth muscle cells could be detected (Supplementary Figure 4). Furthermore, using the Wt1 CreERT2/+ ; Rosa26 Confetti/+ reporter system, we observed a limited clonal expansion of labelled visceral MCs in adult mice over time (Supplementary Figure S5). Therefore, our observations indicate that MCs in healthy adult mice were restricted to maintaining homeostasis of the peritoneum at a low turnover rate (Mutsaers, 2004).
To test this observation further we ablated Wt1 in adult mesothelial cells by using Wt1 CreERT2/co mice, similar to the approach described by Chau and colleagues (Chau et al., 2011). We observed severely deteriorating health in all animals from approximately day 10 after the start of tamoxifen administration onwards, as described (Chau et al., 2011). This precluded a detailed analysis of the impact of loss of Wt1 on the homeostasis of the serosa of the peritoneum and in particular the intestine. However, our histological analysis of various regions of the intestine at day 10 after tamoxifen administration showed no effect on the presence and appearance of the mesothelium, or the overall intestinal morphology (Supplementary Figure S6).
We also included kidneys in our analysis because adult podocytes express Wt1, thus allowing assessment of successful tamoxifen administration. Adult kidneys showed the expected labelling of glomeruli two weeks after tamoxifen (Supplementary Figure

Wt1-derived cells in newborn lineage tracing reveal wider contribution to the heart and kidneys, but not to the serosal mesothelium
Since the adult serosa of the peritoneum appeared to be restricted to its maintenance, we tested whether mesothelial cells may still have some degree of plasticity either in juvenile mice directly after weaning, or within the first few days after birth (Boulland et al., 2013, Hartman et al., 2007, Porrello and Olson, 2014, Seely, 2017, and thus would show capacity to contribute to intestinal tissue homeostasis or to differentiate into VSMCs. After tamoxifen administration in four weeks old juvenile Wt1 CreERT2/+ ; Rosa26 LacZ/+ mice and chase periods between 7 and 17 weeks, restricted contributions of Wt1-derived XGal-stained cells were found in similar locations to those in adult mice (data not shown). Correspondingly, after initiating a 7-weeks chase period in newborn Wt1 CreERT2/+ ; Rosa26 LacZ/+ mice, we detected coverage of XGal-stained cells in the visceral mesothelium comparable to adult mice (Figure 2A), and no contribution to the vasculature of the mesentery or intestine, or within the intestinal wall ( Figure 2B, C). By contrast, lineage-traced cells were found as expected in the epicardial layer of the heart, and also contributing to its coronary and micro-vasculature ( Figure 2D-F). These results indicate that while in the newborn, Wt1-expressing epicardial cells continued to give rise to coronary vessels, visceral mesothelial cells failed to contribute to the intestinal vasculature (Porrello and Olson, 2014, Cao and Poss, 2018, Porrello et al., 2011, Quijada et al., 2020. In the kidneys, neonate Wt1-expressing cells gave rise to XGal-stained cells in the glomeruli and in nephron tubules ( Figure 2G-J). In contrast to lineage tracing in adult kidneys, the contribution of XGal-labelled cells to the nephron tubules was abundant, indicating that Wt1-expressing cells in the neonatal kidneys still had the capacity to give rise to entire nephron structures (Hartman et al., 2007).
Taken together, our lineage tracing analysis of newborn, juvenile and adult mice showed that Wt1expressing cells of the peritoneum failed to contribute to the vasculature, or other components of the intestinal or body wall besides mesenteric fat. This indicates that in healthy postnatal mice peritoneal mesothelial cells are mostly restricted to self-renewal.

From E9.5 onwards, Wt1-expressing cells give rise to visceral mesothelium
Next, we aimed to determine at which time point of prenatal development Wt1-expressing cells gave rise to intestinal vascular smooth muscle. We administered tamoxifen once to time-mated pregnant females at stages between E14.5 and E7.5, followed by analysis just before birth (Figure 3, Supplementary Table 1). Whole mount analysis for LacZ staining or GFP fluorescence in intestine and mesentery revealed that after tamoxifen administration at stages between E10.5 and E14.5, Wt1-expressing cells contributed solely to the mesothelium ( Figure 3A-C, Supplementary Figure   S8B, E, H, J, K and data not shown). We noted that the coverage of the mesothelium of the intestine and mesentery with LacZ-positive lineage-traced cells was almost complete when tamoxifen had been administered at stages E13.5 and E12.5 ( Figure 3A, B), while tamoxifen given at stages E14.5, E11.5, E10.5 or E9.5 led to a more patchy distribution of XGal-stained or GFP-labelled cells in the mesenteric and intestinal mesothelium ( Figure 3C, D, Supplementary Figure S8B, E, H, H', J, J', K, and data not shown). We also observed that activation of the Wt1-based lineage tracing in the mesothelium at these stages resulted in the presence of labeled cells along the entire length of the small intestine (Supplementary Figure S8B). Immunostaining of tissue sections of Wt1 CreERT2/+ ; Rosa26 mTmG/+ embryos after tamoxifen administration at E11.5 and analysis at E19.5 confirmed that the presence of the GFP-labelled cells was limited to the visceral mesothelium, where these cells co-expressed cytokeratin and Wt1 ( Figure 4A, A', B, B'). By contrast, we failed to observe any coexpression of GFP with the endothelial marker CD31 or the vascular smooth muscle marker SMA within the intestine or mesentery at these stages ( Figure 4C, C', D, D').
Between E7.5 and E8.5, Wt1-expressing cells give rise to visceral and vascular smooth muscle of the intestine When we administered tamoxifen at E7.5 and analysed the embryonic intestines just before birth, labelled cells were found as vascular smooth muscle cells in short segments of the mesenteric vasculature ( Figure 3I). Within these regions, labelled cells were also scattered within the mesentery ( Figure 3G). Similarly, after tamoxifen administration at E8.5, we found that XGalpositive cells contributed consistently to vascular smooth muscle cells in short segments of the mesenteric vasculature, but also to small patches of the mesothelium over the mesentery ( Figure 3E Table 1 and data not shown), suggesting a very narrow time window of deciding the potential developmental fate.
Immunostaining of tissue sections from Wt1 CreERT2/+ ; Rosa26 mTmG/+ embryos after tamoxifen at E8.5 and analysis at E17.5 showed that GFP-positive cells co-expressed SMA but not CD31 in mesenteric blood vessels ( Figure  These results suggest that Wt1-expressing cells gave rise to some of the visceral mesothelium covering the mesentery and intestine from E8.5 onwards. However, already in E7.5 embryos, before mesothelium formation, Wt1-expressing cells were the source of vascular and visceral smooth muscle of the intestine and mesentery, indicating that the visceral and vascular smooth muscle cells arise independently of mesothelium formation. Of note, our data suggest that there is a developmental overlap between these separate lineages at around E8.5 to E9.5. We performed whole mount in situ hybridisation for Wt1 of mouse embryos between E7.5 and E9.5, to determine which expression domain would correspond to the activation of Cre recombinase and resultant lineage tracing at the observed stages (Supplementary Figure 9). Our analysis showed that there was relatively high unspecific background staining in embryos between E7.5 and E8.5, with faint signals in the posterior of the embryos near the primitive streak (Supplementary Figure   9A, C). By contrast, at E9.0-E9.5 the previously described domains in the epicardium and the urogenital ridge were clearly detectable (Supplementary Figure S9E, F) (Armstrong et al., 1993).

Tracking lineage-labelled cells during embryonic stages reveals their fate
In order to gain further understanding of the characteristics and fate of the lineage labeled cells, we administered tamoxifen at E7.5 or E8.5 and analysed the embryos at E9.5. Tamoxifen administration at E7.5 revealed GFP expression in a discreet population of around 7-9 epicardial cells in the E9.5 embryo ( Figure  We next administered tamoxifen to embryos at E8.5 and performed analysis at later stages, which revealed a dynamic pattern of GFP-expressing cells (Supplementary Figure S12). Whole mount analysis at E10.5 (after tamoxifen administered at E8.5), showed prominent GFP-expressing cells in the heart and urogenital ridge, but also in an isolated speckled appearance in other regions of the embryo (Supplementary Figure S12A). In sections through levels below the heart, we detected a number of GFP-expressing cells in the urogenital ridge, but also in the region between the dorsal aorta and the mesentery (Supplementary Figure S12B). Analysis at E12.5 showed GFP-expressing cells in the mesentery of the developing intestine (Supplementary Figure S12C Taken together, our results indicate that cells expressing Wt1 predominantly between E7.5 and E8.5 contributed to vascular and visceral smooth muscle formation in the developing small intestine and mesentery. This process was clearly independent of the establishment of the mesothelium of the intestine and parietal peritoneum from around E9.5 onwards, by an as yet unknown developmental mechanism. Further studies are needed to elucidate the molecular mechanisms that drive the development of Wt1-expressing cells in the gastrulating embryo towards vascular and visceral smooth muscle fate, and the role of Wt1 during this process.

Discussion
Here, we have attempted to dissect the relationship between Wt1 expression and the mesothelial lineage with particular focus on the embryonic development of the visceral and vascular smooth muscle in the intestine and mesentery. By utilising temporally controlled lineage tracing of Wt1expressing cells, our results have revealed that the formation of the peritoneal mesothelium and the visceral and vascular smooth muscle of intestine and mesentery, are linked to Wt1 expression, but in two distinct phases. Wt1-based lineage tracing in the embryo and in postnatal stages has shown that once the mesothelium of intestine and mesentery has formed from around E9.5 onwards, this tissue maintains itself, and fails to contribute to other tissues except visceral fat. By contrast, in an earlier phase, Wt1-expressing cells present in the embryo predominantly between E7.5 and E8.5 and not later than E9.5, give rise to the visceral and vascular smooth muscle of the intestine and mesentery. Therefore, our findings presented here indicate two phases of contribution of Wt1expressing cells to mesodermal tissues in the peritoneal cavity.
The mesothelium is a continuous sheet covering the organs housed within the three body cavities, pleural, pericardial and peritoneal, and lining the walls of the cavities. Our previous study, using a continuously active Cre only controlled by the human WT1 promoter, had shown that Wt1expressing cells contributed to the vasculature of the mesentery and intestine (Wilm et al., 2005).
Based on this finding we concluded that the visceral mesothelium gives rise to these vascular structures during embryonic development. Further, our results led us to hypothesise that there may be a role for the visceral mesothelium in maintaining the vasculature, and possibly other intestinal structures during adult life. The work presented here, using a temporally-controlled tamoxifeninducible reporter system driven from the endogenous Wt1 locus, demonstrates that the relationship between Wt1 expression and mesothelial lineage is more complex. In particular, we have shown here that the adult serosa of the peritoneal cavity, only maintains itself, besides giving rise to visceral fat.
Previous pulse-chase studies using the same tamoxifen-inducible Wt1-driven reporter system as used in this report, had shown that the postnatal lung mesothelium makes no contribution to other cells within the lungs (von Gise et al., 2016). By contrast, embryonic lineage tracing experiments revealed that the Wt1-expressing lung mesothelium at E10.5 gives rise to bronchial and vascular smooth muscle, as well as PDGFRβ-expressing pericytes and PDGFRα-expressing fibroblasts (von Gise et al., 2016).
In the liver, the visceral mesothelium arises from the septum transversum at around E9.0. Wt1 is expressed in the septum transversum, and later in the liver mesothelium from E11.5 onwards and throughout postnatal stages (Asahina et al., 2011). Using the tamoxifen-inducible Wt1-based lineage tracing system during embryonic development of the liver mesothelium demonstrated that mesothelial cells give rise to submesothelial cells, hepatic stellate cells as well as perivascular mesenchymal cells, while lineage tracing in the adult liver revealed that only the visceral Wt1expressing mesothelium is labeled (Asahina et al., 2011, Lua et al., 2015.

Previously, Lua and colleagues had demonstrated that parietal mesothelial cells have a different
developmental origin compared to the liver mesothelium, using a Mesp1-Cre-based lineage system (Lua et al., 2015, Asahina et al., 2009). However, the developmental origin of the parietal mesothelium has not been further assessed. Lineage tracing of the adult parietal mesothelium of the peritoneal cavity using the tamoxifen-inducible Wt1-based lineage system revealed selfmaintenance of the mesothelium (Lua et al., 2015), while a minor contribution to collagen 1a1expressing submesothelial cells was observed by Chen and colleagues (Chen et al., 2014). A recent study confirmed by cytometric sorting the presence of a small population of submesothelial fibroblastic cells that expresses Wt1 together with the mesothelial marker podoplanin and the fibroblast marker PDGFRα, with a possible role in maintaining Gata6 expression in large cavity macrophages (Buechler et al., 2019). Here, we have visualized these cells using the lineage tracing system.
Our own findings suggest that during embryonic development, the visceral mesothelium of the intestine is different to that of the liver and lung since after its emergence it fails to contribute to the stroma of the intestinal wall. However, the postnatal visceral mesothelium appears to behave similarly between the lung, liver and intestine in that there is no contribution to stromal compartments, other than the progenitor niche of the visceral fat.
During embryonic development of the heart, Wt1-expressing epicardial cells have been shown to contribute to the mural cells of the coronary vessels as well as fibroblasts (Rudat and Kispert, 2012, Sereti et al., 2018, Tian et al., 2013. In the adult heart, Wt1 expression is downregulated (Smart et al., 2011), and lineage tracing studies using the tamoxifen-inducible Wt1-driven lineage system in adult mice have revealed no contribution to cells other than the epicardium (Quijada et al., 2020. Our data presented here suggest that in the adult as well as in newborns, the tamoxifen-inducible Wt1-driven lineage system allows the detection of Wt1-derived cells in the coronary endothelial and mural cells. This is particularly striking since the visceral mesothelium of the postnatal intestine failed to provide this contribution.
Of note, in contrast to the normal, healthy mesothelium, lineage tracing studies after injury in the lungs, liver, heart and peritoneum have shown that adult mesothelial cells can be activated and undergo a range of physiological changes including epithelial-mesenchymal transition (EMT) into smooth muscle cells and myofibroblasts, and subsequent contribution to scar formation (Chen et al., 2014, Karki et al., 2014, Lua et al., 2015, Namvar et al., 2018, Smart et al., 2011, Kendall et al., 2019.
In the kidneys, we observed differences in the contribution to renal tissue between adult and newborn lineage tracing, indicating that there is a larger degree of plasticity present in the newborn kidney, where Wt1-expressing cells contribute to nephron tubules in addition to the glomeruli and parietal epithelial cells of the Bowman's capsule.
It is important to point out that the efficiency of cell labelling in lineage tracing systems using the Wt1 CreERT2/+ ; Rosa26 mTmG/+ mouse line has been reported to be between 14.5% and 80% in different laboratories (Chen et al., 2014, Li et al., 2013, suggesting that rare lineage labelling events may not be detected using this approach (Rudat and Kispert, 2012). Therefore, in the current study, we can conclude that XGal-or GFP-positive cells have expressed Wt1 at the time of tamoxifen administration, but there may be some cells that have expressed Wt1 which are not labelled and evade the lineage tracing system. The reasons for this variation could be inefficiency of the recombination system or insufficiency of the tamoxifen distribution inside the animals.
One of the unexpected findings in this study was the observation that Wt1-derived cells contribute to the vascular and visceral smooth muscle in the intestinal wall and mesentery before the emergence of the mesothelium. In our previous study, we had used a transgenic mouse line (Tg(WT1-cre)AG11Dbdr; Gt(ROSA)26Sor/J; in short Wt1-Cre; Rosa26 LacZ ) composed of a Cre reporter system driven by human Wilms Tumour protein 1 (WT1) regulatory elements (Wilm et al., 2005), which had been shown to faithfully recapitulate the Wt1 expression domains in mice (Moore et al., 1998). In the Wt1-Cre; Rosa26 LacZ mice, XGal staining had been prominent in the vascular smooth muscle surrounding the veins and arteries in the mesentery and those inserting into the intestinal wall. We concluded that the labelled vascular smooth muscle cells, which had expressed Wt1 at some undetermined time point in the life of the mouse, must have arisen from the visceral mesothelium as the only known tissue in the peritoneal cavity to express Wt1 in the embryo and throughout life (Wilm et al., 2005). What remained unknown was the time point at which the vascular smooth muscle fate was induced in the mesothelial cells. To address this question, we performed the temporally controlled lineage study presented here. Our new data suggest that the vascular and visceral smooth muscle cells labelled in the mesentery and intestine, arise from so far undetermined Wt1-expressing cells in embryos at around E7.5 to E8.5, but not later than E9.5. Very few visceral mesothelial cells in intestine and mesentery were found to be GFP-positive after tamoxifen administration in this early developmental time window. It is important to note that there is a time lag between tamoxifen administration and onset of Cre activation and subsequent recombinatory activity in the nucleus of about 12-24 hours (Nakamura et al., 2006), suggesting that Cre activity may be targeting Wt1-expressing cells from around E8.0 onwards when tamoxifen is administered at E7.5.
In the primitive streak mesoderm, Mesp1 has been shown to be active during gastrulation stages (E6.5-E8.0) (Saga et al., 1996). Interestingly, Asahina and colleagues found that embryonic Mesp1-Cre driven lineage tracing, via the septum transversum mesenchyme, gave rise to liver mesothelial cells as well as cells within the liver parenchyma, but not to parietal mesothelial cells (Asahina et al., 2009). Furthermore, the group demonstrated that in Wt1 CreERT2/+ ; Rosa26 mTmG E11.5 embryos the entire liver mesothelium contains GFP-positive cells when tamoxifen is administered at both E7.5 and E8.5 (Asahina et al., 2011), similar to our own observation of abundant GFP-positive cells over and in the liver at E12.5 after tamoxifen at E8.5. Together with our own observations presented here, this would suggest that there are differences in the developmental mechanisms by which the visceral mesothelium of the liver and the intestine as well as the parietal mesothelium are formed.
In some of the E9.5 embryos after tamoxifen administration at E7.5, we observed the presence of GFP-positive cells within a very restricted domain of the neural tube. This is in contrast to previous reports that Wt1 expression in the neural tube starts between E11 and E12 in the mouse embryo (Armstrong et al., 1993, Haque et al., 2018. The restricted GFP domain in the neural tube suggested that these cells were Cre-recombined after tamoxifen administration while transiently expressing Wt1 and undergoing a short specific phase of gastrulation, possibly as epiblast cells.
Whether Wt1 is indeed expressed in the epiblast during early gastrulation stages needs to be confirmed in future studies. However, the observation that the GFP-positive visceral and vascular smooth muscle cells within intestine and mesentery after tamoxifen administration at E7.5 or E8.5 were restricted to specific segments of the small intestine supports the interpretation that the cells originally expressed Wt1 in a short and transient phase during embryonic development.
While our findings here shed new light on the temporal processes of the development of vascular and visceral smooth muscle in the intestine and mesentery, future studies are needed to elucidate the molecular mechanisms, in particular the role of Wt1, that drive these steps in the gastrulating embryo.
For ablation of Wt1 in adults, animals were dosed with tamoxifen (100 µg/g body weight) via oral gavage on 5 consecutive days. Animals were monitored for their well-being, and typically culled at day 10 after the start of the tamoxifen regime.

XGal staining and histology
Tissues and embryos were fixed in 2% paraformaldehyde (PFA)/0.25% glutaraldehyde in phosphate buffered saline (PBS) for between 1 and 1.5 hours, whole-mount XGal staining performed overnight according to standard protocols, followed by overnight post-fixation in 4% PFA (PBS) at 4°C. Histological analysis was performed on post-fixed XGal-stained specimen after dehydration into isopropanol and paraffin embedding. Serial sections (7 mm) were counterstained with Eosin and images taken on a Leica DMRB upright microscope with a digital DFC450 C camera supported by LAS.

Immunofluorescence on frozen sections
Embryos or tissues were fixed in 4% PFA for between 30 to 90 mins, protected in 30% sucrose overnight, placed in Cryomatrix (Thermo Scientific) and snap frozen. Frozen sections were generated at 7 mm on a Thermo Scientific HM525 NX Cryostat. Immunofluorescence analysis was performed following standard protocols (Wilm et al., 2005). A bleaching step of 10 min in 3% H 2 O 2 /MetOH was included for embryos or tissues from Wt1 CreERT2 ; Rosa26 mTmG/+ mice in order to remove the tdTomato fluorescence (Lua et al., 2015). The following primary antibodies were used: anti-Wt1 rabbit polyclonal ( (1:5000, ab6556 or ab6673, Abcam). The anti-SMA antibody was directly labeled using Zenon direct labeling kit (Invitrogen/ThermoScientific) according to manufacturer's instructions.
Secondary antibodies were Alexa fluorophore-coupled (Invitrogen/ThermoScientific) and were used at a dilution of 1:1000. Sections were counterstained with DAPI (D9542, SigmaAldrich) at 1:1000, coverslipped with Fluoro-Gel (with Tris buffer; Electron Microscopy Sciences, USA), and imaged on a Leica DM 2500 upright microscope with a Leica DFC350 FX digital camera and LAS.

Whole mount in situ hybridization
Mouse embryos from CD1 time matings (Charles River, Harlow, UK) were dissected at between E7.5 and E10.5. The Wt1 full-length cDNA probe was a gift from the Kreidberg lab (Gao et al., 2005, Pelletier et al., 1991. In situ hybridization was performed following published protocols (Wilm et al., 2004, Hogan et al., 1994. In short, the linearized Wt1 probe was labeled using the DIG-RNA labelling mix (Roche), and hybridization performed under RNAse-free conditions using 50% Formamide, 5x SSC at 70°C.

Light sheet microscopy
The trunk area of E9. Germany) by using one 5x detection objective and two 5x illumination objectives. The 488 nm channel was used to detect Wt1-derived GFP+ cells, while the 561 nm channel was used to visualize the membrane-bound dTomato, to support with orientation and tissue context.

Image analysis
Whole mount imaging with and without fluorescence: Imaging of embryos and tissues was performed using a Leica MZ 16F dissecting microscope equipped with a Leica DFC420 C digital camera supported by the Leica Application Suite software package (LAS, version 3 or 4; Leica Microsystems, Germany/Switzerland), and Leica EL6000 fluorescence light source.
Due to uneven tissue geometry, images were taken at different focal levels and subsequently assembled to multilayer composites according to highest focal sharpness.
Confetti imaging: Tissues were imaged in form of multilayer Z-stacks with a 3i spinning disk confocal microscope system (Intelligent Imaging Innovations Ltd.) and images subsequently rendered to Z-projection composites using Fiji ImageJ software. For the short-and long-term chase experiments two groups of three animals each were analysed. The small intestine was dissected, cut into 2-3 cm long segments and sliced flat. Four 2 cm long thin tissue segments were randomly chosen and cleaned from feces by multiple PBS washes. Slides were prepared by gluing two layers of 2x22mm coverslips on a standard slide to form an inner rectangular area for tissue placement to prevent leakage during subsequent inverted confocal microscopy. Tissue samples were placed into the space in the correct orientation, PBS added, covered with a standard 22x40mm coverslip and sealed with clear nail polish. Due to the uneven tissue geometry Z-stack images were taken at random where RFP, YFP and CFP cell labelling was identified in close proximity, in some cases any two of the three possible markers. Cells were scored and counted for all three markers in all Zstacks according to either being a single cell with or without direct contact to a cell of different marker or being in direct contact with another cell(s) of the same marker (designated clones).
Statistical analysis was performed by unpaired multiple t-tests with Holm-Šídák multiple comparisons correction using Graphpad Prism 8.4.2.
3D Image analysis: 3D images and movies were generated by using IMARIS x64 software (version 9.5.1 Bitplane). IMARIS tools used were Rendering, Slices, Orthoslicer, Snapshot images, Animations, Rotations, Surfaces around the samples, and Spots to show the location of the cells in the movies. The images were cropped and assembled by using Photoshop 2020.