Some tubular epithelia, such as nephrons of the kidney, and the oviduct, possess functional polarities along their length, in which distinct domains of cells execute specialised physiological activities. The order of these domains can be important for the emergent properties of the system, for example the sequential processing of primary urine by functionally distinct domains along the length of the nephron. For this reason, the correct and precise establishment of cells with specialised physiological activities along the length of the tube is an important aspect for how the system will ultimately function. Developmental mechanisms that establish and pattern these axes act secondary to those that the establish and pattern the main body axes of the animal. Important insights have been gained into the patterning of epithelial tubules such as kidney nephrons, and the zebrafish pronephros. These insights include transcription factors involved in segment identity (Reggiani et al., 2007; Wingert et al., 2007; Cheng and Wingert, 2015), and signalling molecules such as retinoic acid, BMP, Notch and Wnt necessary for patterning (Wingert et al., 2007; Lindstrom et al., 2015; Schneider et al., 2015; Deacon et al., 2019). There are indications that some of these may act as graded morphogens (Wingert et al., 2007; Lindstrom et al., 2015). However in general these patterning mechanisms remain poorly understood.

Insect Malpighian, or renal tubules (MpTs) are tubular epithelia with a distinct functional axis along their proximo-distal (P-D) length including distinct regions for calcium homeostasis, fluid secretion and reabsorption (Wessing and Eichelberg, 1978; Dow et al., 1994; O’donnell and Maddrell, 1995; Sozen et al., 1997). During embryogenesis, clusters of MpT primordial cells bud out from the wall of the gut, before extending to form each of the MpTs (Skaer, 1993). Functional polarity along the P-D axis is established during this period. It occurs several hours after the patterning of anterior-posterior and dorso-ventral axes of embryo and therefore presents an interesting system to explore how such secondary axes are established and patterned during development. This provides an alternative and therefore informative comparison, for example with P-D patterning of insect appendages which are well-studied examples of secondary axis formation. The MpTs also provide an example of how developmental patterning occurs in the context of a tubular epithelium.

In our previous paper we discovered a mechanism by which the proximal region of the MpT P-D axis is patterned in Drosophila. We determined that the signalling ligand Wingless/Wnt (Wg), produced in gut cells abutting the end of the proximal tubule, patterns the proximal part of the tubule. The ability of the Wg ligand to disperse from its source is essential for this function (Beaven and Denholm, 2018).

The mechanism(s) that establish distal identity is not known, and we therefore turned our attention to this problem here. We considered two putative mechanisms by which patterning of the distal segments of the MpTs could be achieved (Figure 1). In the first, an asymmetric signal from one or other end of the MpT, would play an instructive role in defining distal tubule identity. The signal could either be a distal identity-inducing signal emanating from the distal end of the developing MpT (Figure 1A’), or a distal identity-repressing signal emanating from the gut at the proximal MpT end (Figure 1A’’). We have previously shown that cells derived from the tip cell lineage are a signalling hub that establish a gradient of EGF signalling activity in the distal tubule (Saxena et al., 2014), and thus might be an inductive signal for distal identity. As mentioned above, Wg emanating from the gut at the proximal tubule end induces proximal identity by regulating the expression of genes required for the differentiation of proximal cells (Beaven and Denholm, 2018). It is conceivable that Wg also acts to restrict distal identity by inhibiting the expression of genes required for the differentiation of distal cells.

The second mechanism we considered is that MpT primordia are already pre-patterned at a stage before they bud from the gut wall and that these domain identities are retained as the MpTs bud out, proliferate and elongate as a tubular structure (Figure 1B). This would be conceptually similar to the way in which the Drosophila legs are patterned, with positional information defined in the leg disc which endures as the disc telescopes out to form the mature leg (Meinhardt, 1983b; Basler and Struhl, 1994; Diaz-Benjumea et al., 1994; Lecuit and Cohen, 1997; Giorgianni and Mann, 2011; refs. Within Ruiz-Losada et al., 2018).

Previous studies suggested this as a plausible mechanism, and indicated that Wg could also be involved in this earlier process. There is a proneural cluster of cells within the MpT primordium, one of which gives rise to the tip-cell and sibling-cell (Hoch et al., 1994; Wan et al., 2000). Wg is expressed in the hindgut, as well as the MpT primordium from the start of tubule development (Skaer and Martinez Arias, 1992), and is required for expression of Achaete in the proneural cluster. Wg is also required for expression of Seven up and PointedP2 (Wan et al., 2000; Sudarsan et al., 2002), two distal transcription factors considered to play roles in promoting cell proliferation (Kerber et al., 1998; Sudarsan et al., 2002).

Here we set out to test these alternative models of how the distal MpT is patterned during development, to gain insights into the principles by which secondary axes are generated and patterned.

We have found evidence of a mechanism in which epithelial tubules, the Malpighian tubules of Drosophila, achieve patterning of their P-D axis by early specification of a subgroup of cells which expand during development, giving rise to the distal domain (Figures 9A–D). Wg signalling during early MpT development plays a critical role in specifying their distal identity. This mechanism of early patterning followed by expansion does not rely on dispersal of the Wg ligand. This is in contrast to our previous findings for the role of Wg in patterning the proximal tubule, which does require release and spread of Wg.

In the context of proximal patterning, Wg activates Odd skipped (Odd) expression to a distance of ∼30 μm (Beaven and Denholm, 2018; Figure 9C). Using Wg-GFP, we find that in late stage embryos (∼stage 15–16) Wg in the proximal tubule clearly forms a gradient, which extends ∼25 μm into the MpT (Figures 8F,G). It also appears to extend ∼40 μm into the midgut (Figures 8F,G). This fits well with the range of Wg dispersal in the wing imaginal disc, which is in the order of ∼30 μm (Chaudhary et al., 2019), as well as the range of the Wg-GFP gradient observed in the adult gut (Tian et al., 2019). These figures may constitute the upper limit of Drosophila Wg dispersal. Considering that the anterior tubules extend to over 300 μm by late embryogenesis, and that the domains of distal transcription factors such as Dac and Hth extend to about half this length, this would make the final size of distal expression domains much greater than the effective range of Wg. It seems that the maximum range of Wg is too small to allow Wg scaling to pattern the distal MpT in the manner of a classical morphogen (as, for example, Decapentaplegic (Dpp) does in the wing disc, scaling to 100 μm as the disc reaches full size (Teleman and Cohen, 2000; Wartlick et al., 2011)). This could explain why Wg defines the distal tubule domain in the tubule primordium whilst it is a small cluster of cells, and these cells then maintain their distal identity as they proliferate and the tubule elongates.

Our findings that Wg acts very early in MpT development to define distal cell identity also provoked us to contemplate whether the Wg signal is the earliest symmetry breaking step by which the axis is established. Intuitively one might expect the distal cells to arise from the centre of the primordium, as these would naturally become distal as the tubule telescopes out during development. Such a mechanism is responsible for P-D patterning of the Drosophila legs; here the discs are patterned as concentric rings, with the centre of the disc ultimately becoming the distal end of the leg (Estella et al., 2012). This patterning is achieved by first specifying the central point of the disc which becomes the distal pole of the P-D axis. This is achieved by Dpp and Wg which are expressed dorsally and ventrally at the anterior-posterior compartment boundary, under regulation of hedgehog from the posterior compartment. The intersection of the Dpp and Wg signals activates Distal-less expression at the centre of the leg disc (Basler and Struhl, 1994; Campbell et al., 1993; Diaz-Benjumea et al., 1994). However, our observations in the MpTs are different, in that the cells that ultimately give rise to the distal MpT originate on one side of the MpT primordium—the side which abuts the hindgut and expresses Wg (Figure 8A, Figure 9A), rather than at the centre of the primordium. It is conceivable that the established axis of the developing gut acts to break symmetry in the MpT primordium and thus establish the P-D axis of the tubules. In this context the expression of Wg on one side of the MpT primordium is likely to be regulated by the processes concomitantly patterning the developing gut, such as the division of the hind- and midgut. Further processes define a proneural cluster of cells in the developing MpT, and specify the tip-cell lineage within this cluster. These processes are perturbed by depletion of Wg in early MpT development (Wan et al., 2000), which may be because only the distal lineage of MpT cells are competent to give rise to the proneural cluster. The emergence of the tip/sibling-cells as a signalling centre could then orchestrate the morphological changes underlying tubule elongation, in a way which results in their surrounding cells coming to have a distal position in the final tubule. The EGF signal which emanates from the tip/sibling-cells drives cell proliferation (Baumann and Skaer, 1993; Kerber et al., 1998; Sudarsan et al., 2002; Figure 9C), as well as defining planar cell polarity in surrounding cells to orchestrate their intercalation and the convergent extension of the distal MpT (Saxena et al., 2014), so orchestrating the morphological changes which propel the tip/sibling-cells, and their neighbouring cells with distal identity, to finally reside at the distal end of the extended MpT. Together these processes could convert an already established asymmetry in the gut, into a P-D axis in the MpTs which grow out perpendicular to the gut axis.

The formation of this secondary axis is conceptually similar to the formation of the body’s main axes, in that both are underpinned by the breaking of symmetry under influence from external source(s). It also shows parallels to formation of other studied secondary axes such as Drosophila legs, where establishment of the P-D axis requires an initial breaking of symmetry which is derived from patterning of the main body axes. In diverse examples, including all the Drosophila imaginal discs, as well as vertebrate limbs, primordia develop straddling compartment boundaries formed along the primary body axes. As the cells of the primordia straddle boundaries, they are conferred a developmental asymmetry from the outset, allowing them to establish secondary axes (Meinhardt, 1983a; Meinhardt, 1983b; Basler and Struhl, 1994; Tabata et al., 1995). This may explain why the MpT primordia arise precisely at the boundary between the developing hindgut and midgut.

The mechanism we have uncovered for patterning the distal tubules, i.e. early patterning and expansion of a distal domain, must require some form of cellular memory. Distal cells appear to begin expressing transcription factors such as Dac and Hth some time after their exposure to the Wg signal, and their expression is maintained into the adult, in the absence of Wg. Therefore, there must be a mechanism for cells to retain their identity in the absence of the initial signal. One means to achieve such cellular memory is transcriptional memory as has been characterised for Hox genes, for example in conferring positional identity along the anterior-posterior axis, as well as for hedgehog expression in the Drosophila wing discs. Transcriptional memory of Hox genes and hedgehog involves binding of Trithorax- and Polycomb-group proteins to Trithorax response elements (TREs) and Polycomb response elements (PREs) which are regulatory chromosomal regions (Francis and Kingston, 2001; Maurange and Paro, 2002; Wang et al., 2009; Pérez et al., 2011).

In the case of patterning mechanisms by Wg, it has been shown in wing discs that the frizzled 2 receptor can function to maintain Wg signalling in cells which are no longer exposed to the Wg ligand (Chaudhary et al., 2019). We do not think that this mechanism is the one used in the MpTs, as it relies on perpetuating activity of the Wg signalling pathway (which we found not to be necessary in our experiment driving dTCF∆N, Figure 3D, Supplementary Figure S2D). It has also been shown in gastruloids that cells retain memory of their exposure to Wnt via an unknown mechanism, which controls their responsiveness to activin (Yoney et al., 2018).

To further understand the mechanisms by which Wg could regulate distal target genes, we looked at the activity of dac enhancers containing dTCF/Pan binding sites, but found that these enhancers, which regulate expression of dac in the leg disc, are not responsible for dac expression in the distal tubule (Supplementary Figure S5). This provides evidence to suggest regulation by Wg is indirect (at least for dac), indicating that Wg might trigger a signalling cascade which activates and maintains expression of distal identity genes. Testing the regulation of Hth may also shed light on this mechanism. A genomic fragment of hth was previously found which contains the enhancer region responsible for its expression in the distal MpT (Zohar-Stoopel et al., 2014). There is a dTCF/Pan binding site within this region (Junion et al., 2012), which provides a putative site for direct Wg regulation of hth expression in the MpT.

We have found that Wg, acting at different times during development, plays roles in patterning both the distal and proximal ends of the MpT, and is required for the expression of different target genes in these two contexts. This highlights how context dependent the outcomes of Wg signalling are, even within a single tissue. We had previously found that constitutively activating Wg signalling throughout the MpT leads to an expansion of the proximal Odd expression domain, but that the majority of the MpT including all of the distal region still lacked Odd expression (Beaven and Denholm, 2018). In contrast, we report here that activating Wg signalling throughout the MpT from the very beginning of its development does not cause expansion of the distal domain, as indicated by Dac expression (Figures 4B,G,H). The different contexts may be underpinned by differences in the signals which proximal and distal MpT cells have encountered during their development, apart from Wg.

We first observed Wg expression in the half of the tubule primordium which abuts the hindgut at stage 10 (Figure 8A, Figure 9A). At stage 11 we observed Wg expression on one side of the MpT (Figures 8B,C). This expression is consistent with a previous report of wg mRNA and Wg protein expression at this stage, and may be a continuation of expression seen at stage 10. At stage 11 Wg drives cell division in this cell population (Skaer and Martinez Arias, 1992; Sudarsan et al., 2002; Figure 9B), as well as acting to specify the tip-cell lineage (Wan et al., 2000). Secondly, from late stage 11, throughout the rest of embryogenesis, we observed another distinct region of Wg expression in the developing gut abutting the proximal tubule (Figures 8C–F, Figures 9C,D). This is the region of Wg that specifies the identity of a proximal tubule domain, being required for expression of Odd (Beaven and Denholm, 2018). This expression of Wg at the midgut-hindgut boundary persists stably through larval and adult stages, where it regulates cellular behaviours of the gut and MpT (Takashima et al., 2008; Fang et al., 2016; Tian et al., 2016; Xu et al., 2018; Tian et al., 2019).

Context dependence would allow Wg to play a complex and rapidly changing set of roles in MpT development, in specifying distal identity and the tip-cell lineage in the primordium, driving cell proliferation on one side of the tubule and in specifying proximal identity in the later MpT. These multiple, sequential phases of Wg activity are reminiscent of findings from the wing discs. Here the first phase of Wg function is to subdivide the identity of the disc, by contributing to specification of wing fate as opposed to wing hinge and notum fate. This occurs during the larval second instar, during which Wg is expressed in a wedge of cells in the ventral/anterior compartment (Couso et al., 1993; Williams et al., 1993; Ng et al., 1996; Klein and Arias, 1998). Wg is expressed in the entire wing primordium until early third instar larvae where it contributes to growth and patterning of the disc, but is not expressed in the rest of the wing disc. By the late third larval instar Wg is only expressed in a narrow line of cells at the presumptive wing margin, in a double ring around the distal region of the wing and in part of the notal region (Couso et al., 1993; Williams et al., 1993; Couso et al., 1994; Ng et al., 1996; Martinez Arias, 2003; Alexandre et al., 2014). The roles of Wg at the presumptive wing margin include patterning the wing (Swarup and Verheyen, 2012 - references within; Chaudhary et al., 2019).

The findings here provide mechanistic insights into how a secondary axis is established and patterned in an epithelial tubule. Epithelial tubules are a fundamental component of many organs, and are frequently patterned along their P-D axis. The oviduct or fallopian tube is one example of this, and recent evidence shows that distinct distal and proximal populations of oviduct cells are specified early in development before the oviduct has extended by cell proliferation. These 2 cell populations form distinct stable lineages (Ford et al., 2021). It may therefore be that early patterning followed by tissue expansion could be a widespread means to pattern the P-D axis of epithelial tubules. Such a mechanism may be important in patterning the P-D axis of developing epithelial tubules of the lungs for which there are indications of a role for Wnt signalling, as well as nephrons in the kidney, which are also known to require β-catenin, most likely acting in the Wnt signalling pathway (Mucenski et al., 2003; Lindstrom et al., 2015; Deacon et al., 2019). In the case of kidney nephrons, there is evidence that a persistent gradient of β-catenin signalling is required for correct patterning (Lindstrom et al., 2015), and this may be more akin to what we observe in later stages of MpT development, where a Wingless gradient patterns the proximal tubule (Beaven and Denholm, 2018; Figures 8F,G, Figure 9C). However in common with the mechanisms we report here, the P-D domains that are defined in the developing nephron, appear to be subsequently expanded by cell proliferation (Lindstrom et al., 2015). Furthermore, β-catenin signalling appears to act reiteratively during nephron development, for example it is required for conferring distal identity, and also for the final differentiation of the proximal tubule domain (Lindstrom et al., 2015; Deacon et al., 2019). This is similar to the highly dynamic, context dependent roles of Wingless signalling which we have characterised in the MpTs. How widespread early patterning and subsequent tissue growth are utilised for patterning P-D axes in other epithelial tubules remains an open question, requiring further investigation.