A role for PI K controlling ENPs

A role for PI3K controlling ENPs fits well with current ideas of how FGF, BMP, β-catenin, and FAT4 regulate ENP self-renewal and differentiation (Figure 4C). The ENPs all receive the differentiation promoting β-catenin signal from the ureteric bud, but only the DMH-1 that receive a stromal signal differentiate (Das et al., 2013). The FAT4 signal from the stromal cells triggers phosphorylation and removal of YAP from the nucleus, thereby altering the transcriptional output of β-catenin (Das et al., 2013). As in other cell types, YAP signaling was shown to result in low levels of miR-29, increased PTEN, and inhibition of PI3K (Tumaneng et al., 2012). A model can be envisioned in which also this part of ENP control is fine-tuned by PI3K; however, this needs additional experimental verification. In embryonic stem cells, PI3K inhibition has been shown to reduce β-catenin phosphorylation (Paling et al., 2004). In line with these findings, in our hands, inhibition of PI3K resulted in β-catenin signaling and simultaneous activation of β-catenin, and inhibition of PI3K had an additive effect on β-catenin activity. BMP signalling has previously been shown to be necessary for ENPs to respond to β-catenin activation in older kidneys, whereas cells from younger kidneys do not need this (Brown et al., 2013). In E12.5 kidneys, as used here, inhibition of BMP signaling did not on its own lead to ectopic nephron formation nor activation of β-catenin signaling. However, it was still possible to drive β-catenin signaling by inhibiting GSK3β, confirming that BMP signaling is not necessary for the ENPs to be able to signal via β-catenin during early kidney development. The dynamics of the ENP response to this dual inhibition/activation of BMPR and GSK3β was clearly different from that seen when β-catenin was activated on its own with the GSK3β inhibitor. This suggests that BMP signaling must still be controlling the induction process at this time point in kidney development, but perhaps via a different mechanism. Likewise, the dynamics of inhibiting PI3K signaling together with GSK3β was different from just altering GSK3β on its own or simultaneously blocking BMP and GSK3β. This again confirms that PI3K is not solely acting as a regulator of β-catenin signaling. Although the different pathways that we investigated possibly converge to regulate PI3K signaling, they clearly also control separate processes. BMPR inhibition on its own did not trigger ectopic ENP differentiation, but when either BMPR or PI3K inhibitors were applied together with the GSK3β inhibitor, they both produced additive but still distinguishable effects. It recently has been shown that BMP and FGF signaling could be interacting in an antagonistic balance, where FGF promotes ENP survival and BMP/SMAD signaling controls apoptosis, and WT1 regulates both protein pathways (Motamedi et al., 2014). Untangling the very complex interactions among these pathways will require a significant effort in the future, and our study exemplifies the need to gently modify, rather than obliterate, signaling pathways, as can be achieved through careful use of inhibitors instead of full gene knockouts (Davies, 2009). Unexpectedly, we found expression of nephron segmentation and epithelialization markers prior to the formation of CDH+ adherence junctions and the formation of a structurally distinct epithelium in PI3K-inhibited kidneys. We found that tight junctions began to form (ZO-1+) and β-catenin+ foci started to assemble, indicative of the formation of adherens junctions. Several other cadherins are expressed by the ENPs and could explain how β-catenin+ foci assembled without CDH1 (Goto et al., 1998; Klein et al., 1988). Further evidence of the ENPs beginning to epithelialize comes from the cells depositing a basement membrane (β-laminin+). Although, admittedly, inhibition of PI3K and simultaneous activation of β-catenin signaling does not necessarily reflect the normal situation, it does emphasize that shifts between the mesenchymal and epithelial states are fluid and dynamic transitions rather than sudden shifts. Indeed, while CDH1 expression can be detected only after the initial aggregation during nephrogenesis (Vestweber et al., 1985), proximal nephron marker and adhesion protein CDH6 can be detected in mesenchymal cells before CDH1 is detected (Cho et al., 1998). We have suggested previously that nephron segmentation starting before the formation of a rigid epithelium could explain the patterning defects in nephrons with reduced Rho-kinase activity (Lindström et al., 2013). ENPs can express genes associated with differentiation and segmentation prior to nephron formation (Brunskill et al., 2014). A better description of the dynamics of the renal MET is clearly necessary, and not just for semantic reasons. Understanding when markers are first expressed will help the phenotypic description of kidney development. Moreover, every step in the MET process is a potential moment when phenotypes can arise under experimental conditions, as shown here, or in disease situations.