1. Developmental Biology and Stem Cells
Download icon

Prostaglandin signaling regulates nephron segment patterning of renal progenitors during zebrafish kidney development

Research Article
Cited
1
Views
604
Comments
0
Cite as: eLife 2016;5:e17551 doi: 10.7554/eLife.17551

Abstract

Kidney formation involves patterning events that induce renal progenitors to form nephrons with an intricate composition of multiple segments. Here, we performed a chemical genetic screen using zebrafish and discovered that prostaglandins, lipid mediators involved in many physiological functions, influenced pronephros segmentation. Modulating levels of prostaglandin E2 (PGE2) or PGB2 restricted distal segment formation and expanded a proximal segment lineage. Perturbation of prostaglandin synthesis by manipulating Cox1 or Cox2 activity altered distal segment formation and was rescued by exogenous PGE2. Disruption of the PGE2 receptors Ptger2a and Ptger4a similarly affected the distal segments. Further, changes in Cox activity or PGE2 levels affected expression of the transcription factors irx3b and sim1a that mitigate pronephros segment patterning. These findings show for the first time that PGE2 is a regulator of nephron formation in the zebrafish embryonic kidney, thus revealing that prostaglandin signaling may have implications for renal birth defects and other diseases.

https://doi.org/10.7554/eLife.17551.001

Introduction

The kidney serves central functions in metabolic waste excretion, osmoregulation, and electrolyte homeostasis. Vertebrate kidney organogenesis is a dynamic process involving the generation of up to three distinct structures that originate from the intermediate mesoderm (IM) (Saxen, 1987). In mammals, a pronephros, mesonephros, and metanephros develop in succession. Of these structures, the pronephros and mesonephros both eventually disintegrate, leaving the metanephros as the adult kidney. In contrast, lower vertebrates such as fish and amphibians only form a pronephros and mesonephros, which are active during embryogenesis and larval stages, respectively, and the mesonephros subsequently serves as the adult organ (Dressler, 2006).

During the progression of vertebrate kidney ontogeny, composition of the basic renal functional unit, termed the nephron, remains largely similar across species (Desgrange and Cereghini, 2015). Nephrons contain a renal corpuscle that filters the blood, a tubule that modifies the filtrate solution, and a collecting duct (Romagnani et al., 2013). The tubule portion of the nephron is configured along its proximo-distal axis with specific groupings of cells, termed segments, which possess unique physiological roles in solute reabsorption and secretion. While the organization of proximal and distal nephron segments is broadly conserved (Romagnani et al., 2013), the genetic and molecular mechanisms that regulate formation of each segment lineage have yet to be fully described (Costantini and Kopan, 2010).

The zebrafish embryonic pronephros is a useful model to delineate the processes that regulate vertebrate nephron segmentation because it is anatomically simple, being comprised of only two nephrons (Gerlach and Wingert, 2013). Further, the transparent nature of zebrafish embryos, their ex utero development, and the ease at which large numbers can be obtained and managed, are all features that readily facilitate renal development and disease studies (Pickart and Klee, 2014; Poureetezadi and Wingert, 2016). The zebrafish pronephric tubule has four discrete tubule segments: a proximal convoluted tubule (PCT), proximal straight tubule (PST), distal early (DE), and distal late (DL) (Wingert et al., 2007) (Figure 1A). The proximal segments are homologous to the PCT and PST in mammals, while the distal segments are homologous to the mammalian thick ascending limb (TAL) and distal convoluted tubule (DCT), respectively (Wingert et al., 2007; Wingert and Davidson, 2008).

A novel small molecule screen reveals that prostaglandins alter nephron patterning.

(A) A diagram detailing the segmentation of the pronephros in relation to somites within the zebrafish embryo. Arrows indicate the blood filter, duct, and cloaca. (B) A schematic of the chemical genetic screen used for evaluating small molecules. Embryos were arrayed in 96-well plates and then exposed to drugs diluted in E3 medium from 60% epiboly to 24 hpf, where the embryos were then fixed and underwent WISH using a riboprobe cocktail to detect the P (wt1b), PCT (slc20a1a), and DE (slc12a1). Black and blue bars are used to illustrate changes between the WT embryo and an embryo with a patterning phenotype, respectively. (C) A pie graph and table denoting the number and percentage of small molecules hits from the chemical screen that expanded or restricted the P (blue and teal), PCT (green and purple) or DE (red and yellow). (D) WISH in 24 hpf stage embryos to detect the P (wt1b), PCT (slc20a1a), and DE (slc12a1) in WTs and those treated with 4-HPR, PGD2, PGA1, PGJ2, and PGB2. A black or blue bar was used to notate the segment change between the WT and drug treated embryos, respectively. Red scale bar, 70 µm. (E) Schematic showing example components of prostaglandin production and signaling. The precursor arachidonic acid (AA) interacts with either the Ptgs1 or Ptgs2a enzyme to generate an intermediate moiety, with the example here being PGH2. The intermediate interacts with a subsequent enzyme to produce the bioactive prostanoid molecule. Here, we depict the prostaglandin E synthase, Ptges, creating the bioactive prostaglandin PGE2 that can transduce signals through binding several G-protein coupled receptors such as Ptger2a and Ptger4a. Other receptors work with other bioactive prostaglandins. Indomethacin is a nonselective Cox (Ptgs1/Ptgs2a) inhibitor that prevents prostaglandin biosynthesis.

https://doi.org/10.7554/eLife.17551.002

During zebrafish kidney development, renal progenitors arise rapidly from the IM and undergo a mesenchymal to epithelial transition (MET) to engender the tubule by 24 hr post fertilization (hpf) (McKee et al., 2014; Gerlach and Wingert, 2014). Prior to this, the renal progenitors undergo complex segment lineage patterning events, beginning with their segregation into rostral and caudal subdomains, a process that is orchestrated by the morphogen retinoic acid (RA) which is locally secreted by the adjacent paraxial mesoderm (PM) (Wingert et al., 2007; Wingert and Davidson, 2011). Modulating levels of RA affects the specification of renal progenitors, inducing proximal segment lineage formation over distal, which can be accentuated by the addition of exogenous all-trans RA, while distal fates are induced over proximal by inhibiting endogenous production of RA through the application of the biosynthesis inhibitor N,N-diethlyaminobenzaldehyde (DEAB) (Wingert et al., 2007; Wingert and Davidson, 2011). Through expression profiling and subsequent functional studies, several transcription factors have been mapped as acting downstream of RA signaling to regulate pronephros segmentation and epithelial fate choice, including hepatocyte nuclear factor-1 beta (paralogues hnf1ba and hnf1bb), iroquois homeobox 3b (irx3b), mds1/evi1 complex (mecom), single minded family bHLH transcription factor 1a (sim1a), and t-box 2 (paralogues tbx2a and tbx2b), among others (Wingert and Davidson, 2011; Naylor et al., 2013; Li et al., 2014; Kroeger and Wingert, 2014; Cheng and Wingert, 2015; Marra and Wingert, 2016Marra et al., 2016; Drummond et al., 2017). Despite these advances, the identity of the other essential signals that control renal progenitor fate decisions has remained elusive (Cheng et al., 2015).

Historically, prostaglandins have been defined as functionally diverse molecules that regulate an array of biological tasks, including inflammation and vasoregulation (Funk, 2001; Tootle, 2013). With regard to the adult kidney, prostaglandins regulate many aspects of renal physiology, ranging from tubular transport processes to hemodynamics (Nasrallah et al., 2007). Prostaglandins are lipid mediators produced by the sequential actions of a series of enzymes, and exert their effects by paracrine or autocrine signaling through distinct G-protein coupled receptors (Funk, 2001; Tootle, 2013). More specifically, there are five major prostaglandins produced from the precursor arachidonic acid (AA) by the enzymes Prostaglandin-endoperoxide synthase one or Prostaglandin-endoperoxide synthase 2a (Ptgs1 and Ptgs2a in zebrafish, also known as cyclooxygenases COX-1 and COX-2 in mammals) followed by subsequent processing by particular synthases (Funk, 2001; Tootle, 2013). Each bioactive prostanoid interacts with one or more G-protein coupled membrane receptors (Funk, 2001; Tootle, 2013). For example, COX activity on AA can generate the intermediate PGH2, from which the PGE2 bioactive can be produced by the prostaglandin E synthase (Ptges) (Figure 1E). PGE2 will signal by subsequent interactions with Ptger G-protein coupled receptors including EP1, EP2, EP3 and EP4 (known as Ptger1-4 in zebrafish) on receiving cells (Figure 1E) (Funk, 2001; Tootle, 2013; Yang et al., 2013).

While prostaglandin biosynthesis and signal transduction have been extensively studied in both healthy and diseased adult tissues (Matsuoka and Narumiya, 2007; Smyth et al., 2009), knowledge of their roles in development have been more challenging to ascertain for several reasons. Firstly, although it is thought that various factors that produce prostaglandins are broadly expressed during ontogeny, precise knowledge about the spatiotemporal progression of particular pathway components is incomplete. Secondly, there is a substantial void in our understanding due to the results of murine loss of function studies where genetic disruptions of components within the prostaglandin pathway was associated with observably normal development. This led to the hypothesis that maternal prostaglandin sources had rescued embryogenesis, thereby complicating the use of mammalian models to study prostaglandin requirements during ontogeny. The importance of reevaluating prostaglandin signaling in kidney formation has been emphasized by a recent report that COX-2 dosage is critical for murine metanephros development, though it is presently enigmatic whether there are requirement(s) for discrete stages of nephrogenesis (Slattery et al., 2016).

In lieu of the challenges of using mammalian systems to delineate the roles of prostaglandin signaling during development, considerable insights in vertebrates have nevertheless been achieved recently through research using the zebrafish model. Most notably, there have been transformative revelations regarding the conserved roles of prostaglandin signaling during definitive blood formation, where PGE2 was found to regulate hematopoietic stem cell (HSC) development and function (North et al., 2007). A chemical genetic screen in zebrafish also identified the prostaglandin pathway as a modifier of endoderm organogenesis, where in subsequent work it was found that PGE2 activity controls opposing cell fate decisions in the developing pancreas and liver through the ep4a receptor (also known as ptger4a), which derive from a bipotential endoderm progenitor (Garnaas et al., 2012; Nissim et al., 2014). Other than these studies, there is little known about how prostaglandin signaling may affect cell fate decisions during the emergence of other vertebrate tissues.

Here, we report the discovery that PGE2 signaling has potent effects in regulating proximal and distal segment formation during nephrogenesis in the developing zebrafish kidney. Using the zebrafish embryo for gain and loss of function studies, in addition to whole mount in situ hybridization (WISH) to profile gene expression, we uncovered that the Cox enzymes Ptgs1 and Ptgs2a, as well as the PGE2 receptors Ptger2a and Ptger4a, are necessary to properly establish distal nephron segment boundaries during pronephros genesis. Further, we found that addition of PGE2 was sufficient to rescue distal segmentation in Ptgs1 and Ptgs2a deficient zebrafish. Interestingly, treatment with exogenous PGE2 or PGB2 during nephrogenesis induced a striking expansion of a proximal tubule segment lineage in a dosage-dependent manner. Taken together, this work reveals for the first time that alterations in PGE2 signaling, and possibly other prostaglandins as well, has important consequences for the developing nephron.

Results

Chemical genetic screen reveals that prostaglandin levels affect nephron development

To date, much remains unknown concerning the factors that control nephron segment development and cell fate decisions. The zebrafish pronephros is an experimentally tractable system to interrogate the genetic factors that regulate nephrogenesis because of its simple, conserved tubule structure, with two proximal segments and two distal segments (Figure 1A) (Ebarasi et al., 2011; Drummond and Wingert, 2016). The nephrons share a blood filter comprised of podocyte cells (P), followed by a neck (N) segment that transports fluid into the tubule, and finally a pronephric duct (PD) that drains caudally at the cloaca (C), a common exit for the kidney and gut in the embryo (Figure 1A, middle panel). Nephron segment fates are established by the 24 hpf stage, based on the expression of unique solute transporters, and each segment has been mapped to a precise axial location relative to the somites that comprise the embryonic trunk (Figure 1A, bottom panel), which facilitates the analysis of pattern formation within the renal progenitor field (Wingert et al., 2007).

Chemical genetics is a powerful approach to study developmental events in the context of the whole organism, and the application of chemical genetics in the zebrafish has led to a number of valuable discoveries about the mechanisms of organogenesis in diverse tissues, including derivatives of the mesoderm (Lessman, 2011; Poureetezadi and Wingert, 2013). Therefore, we hypothesized that a chemical genetic screen could provide new insights about the identity of nephrogenesis regulators. To this end, we performed a chemical genetic screen using the Screen-Well ICCB Known Bioactives Library (Enzo Life Sciences), a collection that includes 480 compounds with known biological activities. Zebrafish embryos were collected from timed matings of wild-type (WT) adults, and then arrayed in 96-well plates for control (dimethyl sulfoxide, DMSO) or experimental treatment between 4 and 24 hpf (Figure 1B). At the 24 hpf stage, embryos were fixed for multiplex WISH analysis, during which they were assessed for expression of a set of genetic markers that distinguished alternating nephron segments within the pronephros, namely wt1b to directly label the P, slc20a1a to label the PCT, and slc12a1 to label the DE (Figure 1B). Because these riboprobes stain alternating nephron segments, they enabled precise scoring as to whether exposure to each chemical led to an expansion or restriction of these distinct cell types (Figure 1C, Figure 1—source data 1).

In total, 16.25% (78/480) of ICCB bioactives were associated with nephron phenotypes (Figure 1C, Figure 1—source data 1). The effect of each compound was annotated as to whether the experimental dosage was associated with WT development, an expansion in segment(s) (P+, PCT+, DE+) or a restriction in segment(s) (P-, PCT-, DE-) (Figure 1CFigure 1—source data 1). The compounds that led to alterations in nephrogenesis included numerous RA pathway agonists and antagonists, such as 4-hydroxyphenylretinamide (4-HPR), a synthetic analog of all-trans RA (Figure 1D) (Poureetezadi et al., 2014). Compared to WTs, exposure to 1 mM 4-HPR led to an expansion of the PCT, caudal shift of the DE, and a dramatic expansion of the interval between these segments where the PST normally emerges, suggestive of an expanded PST segment (Figure 1D) (Poureetezadi et al., 2014). The observation that molecules which impact the RA pathway were flagged as hits in the screen provided an important positive control for our experimental system, given the well-established effects of RA levels on renal progenitors (Wingert et al., 2007; Wingert and Davidson, 2011; Li et al., 2014; Cheng and Wingert, 2015; Marra and Wingert, 2016; Drummond et al., 2017).

In further surveying the identities and respective classifications of the small molecules that impacted nephrogenesis, we noted a striking trend with regard to prostaglandin pathway agonists and tubule segmentation. Among the screen hits, a series of prostaglandin cytokine moieties were independently flagged as modifiers of tubule segment formation, including PGD2, PGA1, PGJ2, and PGB2 (Figure 1D, Figure 1—source data 1). Exposure to these bioactive prostaglandins was associated with changes in the pronephros whereby there was a reduced PCT segment length and a posterior shift in the position of the DE, such that there was a noticeably longer domain between these segment regions compared to WT control embryos (Figure 1D). The discovery that exposure to exogenous prostaglandins was linked with several segmentation changes was particularly fascinating to us because PGE2 signaling has been associated recently with the development of several tissues, including HSCs and fate choice in endoderm derivatives between the liver and pancreas (North et al., 2007; Nissim et al., 2014). Therefore, we next sought to further explore how elevated prostaglandin levels, including PGE2, affected nephron segment development.

Elevated PGE2 or PGB2 levels induce an expansion of the PST segment and DL reduction

Prostaglandins typically have a short half-life, and have been characterized as secreted molecules that activate receptors close to their site of production, thus inducing local effects in a paracrine or autocrine fashion (Smyth et al., 2009). Prostaglandins have also been shown to elicit dosage-specific effects, leading to their description as morphogens (Nissim et al., 2014). Therefore, to validate and further explore the screening results, we selected two prostanoids: one was a hit from our screen, PGB2, and the other was 16,16-dimethyl-PGE2 (dmPGE2), a long-acting derivative of PGE2 which has been extensively used to study the effects of PGE2 in zebrafish (North et al., 2007; Goessling et al., 2009; Nissim et al., 2014). WT embryos were collected and incubated in varying concentrations of drug (30 μM, 50 μM, 80 μM or 100 μM) from the 4 hpf stage to the 24 hpf stage. Double WISH was then performed to determine the resultant nephron segments alongside the trunk somites, and the absolute lengths of nephron segment domains were also measured (Figure 2, Figure 2—figure supplements 1 and 2).

Figure 2 with 4 supplements see all
Exogenous prostaglandin activity promotes proximal straight tubule identity.

(A) Embryos were exposed to 100 µM dmPGE2 between 4 hpf and 24 hpf. WISH was used to stain for the PCT (slc20a1a), PST (trpm7), DE (slc12a1), and DL (slc12a3) (purple) and the somites (smyhc1) (red) at the 24 hpf stage. Black bars indicate segment gene expression domain. Red scale bar, 70 µm. (B,C) The PST and DL segments were measured in microns after incubation in 100 μM of dmPGE2 (n = 5 for each control and experimental group). (D) Summary depicting the nephron segments after exogenous dmPGE2 treatment. Data are represented as ± SD, significant by t test comparing each drug treatment to the DMSO control, *p<0.05, ***p<0.0005.

https://doi.org/10.7554/eLife.17551.004

Exposure to dmPGE2 or PGB2 resulted in a dose-dependent increase in the domain length of the PST segment compared to WT embryos, visualized by WISH with the marker trpm7 (Figure 2A,B, Figure 2—figure supplements 1,2). In conjunction with this change, the DL segment was significantly reduced in length, as visualized by WISH with the marker slc12a3 (Figure 2A,C, Figure 2—figure supplements 1,2). Additionally, the rostral domain of the PCT was reduced in a dosage-dependent fashion, based on expression of slc20a1a (Figure 2A,D, Figure 2—figure supplements 1,2). Further, there was a caudal shift in the position of the DE segment though its absolute length was unchanged, based on expression of the DE-specific marker slc12a1, which resides between the domains of the PST and DL segments (Figure 2A,D, Figure 2—figure supplements 1,2). Overall, these results recapitulated the phenotypes observed following treatment with various bioactive prostaglandins during the chemical screen (Figure 1D). To determine if embryo dimensions were a factor in pronephros segment domain changes, we measured control and treated embryos from tip to tail as well as their pronephric domain (somite 3 to somite 18). We found no statistical differences in the body axis length or pronephric domain between WT controls and dmPGE2 treated embryos (Figure 2—figure supplement 3). To further gauge the possible side effects of dmPGE2 treatment on surrounding tissues, we assessed development of specific tissues using WISH. We noted no significant changes in the vascular marker flk1 or primitive blood precursors using the marker gata1 between WT controls and 100 μM dmPGE2 treated embryos (Figure 2—figure supplement 4A,B). Furthermore, we performed o-dianisidine staining, which labels hemoglobinized erythrocytes and thereby provides a sensitive assessment of defects in circulation or vascular integrity that can be undetected by live imaging with stereomicroscopy. o-dianisidine staining showed that blood flow in WTs and 100 μM dmPGE2 treated embryos was equivalent through the 48–55 hpf stage, as we did not observe compromised vessel integrity or hematomas (e.g. bleeding, blood pooling) (Figure 2—figure supplement 4C). This suggests that PGE2 exposure did not cause major aberrations in tissues surrounding the pronephros. In sum, these observations confirmed the finding from the chemical screen that exogenous PGB2 had profound effects on nephron segment formation, and revealed that alterations in PGE2 had similar consequences.

Expression of Ptges enzymes is required for normal distal pronephros segment development

Next, we determined whether endogenous prostaglandin biosynthesis mediated by the Ptges (e.g. Cox1, Cox2) enzymes was necessary for normal nephron segmentation. To test this, we incubated WT embryos with the compound indomethacin, a nonselective Cox1 and Cox2 enzyme inhibitor, which inhibits the first stage of prostanoid biosynthesis, and has been shown to suppress PGE2 production in zebrafish by mass spectrometry (Figure 1E, Figure 3) (Grosser et al., 2002; Cha et al., 2005; North et al., 2007). Exposure of WT embryos to 30 μM indomethacin was associated with normal proximal segment locations along the embryonic trunk (Figure 3A, Figure 3—figure supplement 1). However, the balance of distal segments was disrupted after indomethacin treatment, such that the majority of embryos developed an slc12a1-expressing DE segment that was significantly expanded in length and an slc12a3-expressing DL segment that was significantly reduced in length compared to wild-type controls (Figure 3A,B and C). Absolute segment length measurements of the proximal domains in indomethacin treated embryos compared to wild-types confirmed there was no significant change in the lengths of these segments (Figure 3D, Figure 3—figure supplement 1). As with dmPGE2 treated embryos, we assessed the effect of indomethacin exposure at this dosage with various morphological dimensions and the formation of surrounding tissues such as the vasculature, and observed no differences compared to WT controls (Figure 2—figure supplements 3,4).

Figure 3 with 1 supplement see all
Prostaglandin production is required for normal distal cell fate specification.

(A) Embryos were treated with 0.3% DMSO or the nonselective Cox inhibitor indomethacin at 30 μM from 4 hpf to 24 hpf. WISH was used to stain for the PCT (slc20a1a), PST (trpm7), DE (slc12a1), and DL (slc12a3) (purple) and the somites (smyhc1) (red) at the 24 hpf stage. Black bars indicate segment gene expression domain. Red scale bar, 70 µm. (B,C) The DE and DL segments were measured in microns (n = 5 for each control and experimental group). (D) A summary depicting the nephron segments after indomethacin treatment. Data are represented as ± SD, significant by t test comparing each drug treatment to the DMSO control group, **p<0.005.

https://doi.org/10.7554/eLife.17551.009

To further explore these results, we examined the effect of other small molecules that have been validated to interfere with Cox enzyme activity. Treatment with the Ptgs1 (Cox1) selective inhibitor SC-560 or the Ptgs2a (Cox2) selective inhibitor NS-398 (Grosser et al., 2002; Cha et al., 2005; North et al., 2007) induced an expansion of the DE segment and a restriction of the DL compared to wild-type embryos, while having no discernible effect on proximal segment development (Figure 4, Figure 4—figure supplements 1 and 2). The DE and DL segment domain phenotypes following SC-560 or NS-398 treatment were statistically significant based on absolute length analysis compared to wild-type controls (Figure 4—figure supplement 9).

Figure 4 with 10 supplements see all
Prostaglandin signal inhibition results in an expansion of the distal early domain, which is rescued by the addition of exogenous dmPGE2.

Embryos were treated with a (A) 1% DMSO control, the (B) Ptgs1 selective inhibitor SC-560 at 50 μM from 4 hpf to 24 hpf, microinjected with the ptgs1 MO alone, or microinjected with the ptgs1 MO and treated with dmPGE2 from 4 hpf to 24 hpf. Embryos were treated with the (C) Ptgs2a selective inhibitor NS-398 at 50 μM from 4 hpf to 24 hpf, microinjected with the ptgs2a MO or microinjected with the ptgs2a MO and then treated with dmPGE2 at 50 μM from 4 hpf to 24 hpf. (D) Embryos were microinjected with a combination of the ptgs1 MO and ptgs2a MO. Embryos were microinjected with (E) ptger2a MO1, ptger2a MO2, (F) ptger4a MO1 and ptger4a MO2. (AE) WISH was used to stain for the DE (slc12a1), DL (slc12a3) (purple), and the somites (smyhc1) (red) at the 24 hpf stage. Black bars indicate segment gene expression domain. Red scale bar, 70 µm. (G) Summary depicting the nephron segments after inducing deficiency of prostaglandin synthesis or receptor activity.

https://doi.org/10.7554/eLife.17551.011

To corroborate the effects we observed on zebrafish distal nephron segment development from Cox1/2 enzyme inhibition, we generated ptgs knockdowns through microinjection of previously described translation blocking morpholinos to target ptgs1, ptgs2a, or both ptgs1/2a into 1-cell stage WT embryos (Grosser et al., 2002; North et al., 2007). Nephron segmentation was analyzed at the 24 hpf stage by WISH with the panel of specific markers to delineate the domains of the PCT, PST, DE and DL (slc20a1a, trpm7, slc12a1 and slc12a3, respectively) as well as the somites (smyhc1). Single and double ptgs1/2a morphants developed a larger DE domain that was increased in length compared to WT control embryos (Figure 4B–4D, Figure 4—figure supplements 1, 2 and 9). In addition, single and double ptgs1/2a morphants had a shortened DL segment compared to the DL in WT control embryos (Figure 4B–4D, Figure 4—figure supplements 1, 2 and 9). These DE and DL segment domain phenotypes were all statistically significant based on absolute length analysis compared to WT controls (Figure 4—figure supplement 9). In contrast, PCT and PST segment development was normal in ptgs1, ptgs2a, or ptgs1/2a deficient embryos at the 24 hpf stage (Figure 4—figure supplements 1 and 2).

To further validate these findings, we conducted independent analysis of ptgs1 or ptgs2a morpholinos that were confirmed to interfere with mRNA splicing, by which inclusion of an intron was found to generate transcripts encoding prematurely truncated proteins (Figure 4—figure supplement 10). Compared to wild-type controls, both of these ptgs1 and ptgs2a splice morpholinos similarly affected pronephros development by causing a DE expansion and a DL reduction, with no perceivable consequence to the proximal segments (Figure 4—figure supplement 3). Again, the segment domain findings were statistically significant based on absolute segment length analysis with WT controls (Figure 4—figure supplement 9). Taken together, these data provide independent validation that expression of both ptgs1 and ptgs2a are critical for normal formation of the DE and DL segments in the zebrafish pronephros.

PGE2 rescues nephron development in Ptgs-deficient zebrafish embryos

Next, we tested whether distal nephron segmentation in ptgs1 and ptgs2a deficient embryos could be rescued by provision of a bioactive prostanoid. For these experiments, we selected dmPGE2 treatment, in part to test the hypothesis that PGE2 signaling is required for pronephros development. WT embryos were injected with ptgs1 and ptgs2a morpholinos and then subsequently treated with 50 μM dmPGE2 between the 4 hpf stage and the 24 hpf stage. Nephron segments were then assessed by WISH using our panel of segment-specific riboprobes. We observed that the alterations in DE and DL segments in ptgs1 and ptgs2a deficient embryos were indeed rescued by exposure to dmPGE2, a treatment combination that was again not associated with altered proximal segment domains (Figure 4B,C, Figure 4—figure supplements 1, 2 and 9). Notably, the ptgs1 and ptgs2a morphants treated with dmPGE2 exhibited statistically similar DE and DL segment lengths compared to WT controls, in contrast to the longer DE and shortened DL in ptgs1 and ptgs2a morphants treated with DMSO vehicle (Figure 4—figure supplement 9). These data indicate that the expansion of the DE segment and the restriction of the DL in ptgs1 and ptgs2a deficient embryos were caused specifically by diminished prostaglandin activity, and implicate PGE2 as the essential bioactive prostanoid because dmPGE2 was sufficient to rescue pronephros segmentation in the context of either Cox1 or Cox2 knockdown.

PGE2 regulates nephron segmentation via the ptger2a and ptger4a receptors

PGE2 is known to signal to its target cells by binding with the G-protein coupled receptor Prostaglandin E receptor 2a or the Prostaglandin E receptor 4a (Ptger2a, Ptger4a; also known as EP2 and EP4, respectively) (Cha et al., 2006). PGE2 signaling in zebrafish acts through both Ptger2a and Ptger4a to modulate HSC formation (North et al., 2007), and during endoderm specification, wherein the differential expression of these receptors mediates tissue development at discrete stages (Nissim et al., 2014). There has been some spatiotemporal expression analysis of these genes during zebrafish embryogenesis as well, which revealed that ptger2a transcripts were expressed at the six somite stage (ss) within bipotential endoderm progenitors and that ptger4a transcripts were expressed at 72 hpf within liver precursors (Nissim et al., 2014). However, further characterization of these genes’ expression in relation to other organs, such as the kidney, has not been addressed.

Based on this, we examined if Ptger2a and/or Ptger4a expression was associated with any stages of pronephros development. We utilized WISH to determine the spatiotemporal expression of ptger2a and ptger4a transcripts between the tailbud stage and 24 hpf to determine if they localized to the areas occupied by the nephron progenitors (Figure 4—figure supplement 4). Interestingly, we found that ptger2a transcripts were expressed in a continuous stretch of IM renal progenitors between the 12 ss and 24 ss based on their location in bilateral stripes of cells situated adjacent to the paraxial mesoderm (Figure 4—figure supplement 4A). ptger4a transcripts were similarly expressed in the IM renal fields between the 12 ss and 15 ss, where cells expressed varying levels of signal, suggesting patches of somewhat variable expression (Figure 4—figure supplement 4B). ptger4a transcripts showed low levels of ubiquitous mesoderm expression at the 20 ss, and then were localized to the cloaca region at the 24 ss (Figure 4—figure supplement 4B). We next performed double WISH in WT embryos at the 14 ss and 18 ss to label cadherin17 (cdh17) expressing renal progenitors along with either ptger2a or ptger4a (Figure 4—figure supplement 4C). As expected, ptger2a expressing cells in the IM fully occupied the bilateral stripes of cdh17 expressing cells, as did ptger4a, though again we noted the slight variability of ptger4a transcript staining in cells residing within the cdh17 pronephros fields (Figure 4—figure supplement 4C). These data were consistent with the notion that Ptger2a and/or Ptger4a may operate in renal progenitors to modulate their development.

To test the hypothesis that Ptger2a and/or Ptger4a function was necessary for pronephros segmentation, we next utilized previously published 5’UTR or start site targeting morpholinos to abrogate expression of either ptger2a or ptger4a transcripts during embryogenesis (Cha et al., 2006; North et al., 2007), as well as independent pharmacological treatments with two different Ptger2a receptor antagonists to block its activity (Figure 4E,F, Figure 4—figure supplements 2 and 59). Deficiency of ptger2a or ptger4a resulted in a statistically significant expansion of the DE segment and a reduction of the DL (Figure 4E,F, Figure 4—figure supplements 2 and 59). These data recapitulate the phenotypic effects that resulted from treatment with the Ptgs1/2a small molecule inhibitors, as well as deficiency of Ptgs1, Ptgs2a, and the combination of Ptgs1/2a (Figure 4E). We also specifically evaluated whether ptger2a or ptger4a knockdown could be rescued by dmPGE2, and found that dmPGE2 was not sufficient to rescue either the DE segment expansion or DL reduction (Figure 4—figure supplement 8). These findings are consistent with the notion that PGE2 acts specifically via Ptger2a and Ptger4a to mitigate DE-DL formation during pronephros ontogeny (Figure 4—figure supplement 8).

To further validate these observations and conclusions, we next examined pronephros segment development following morpholino-mediated knockdowns that were confirmed to alter the normal splicing of either ptger2a or ptger4a transcripts, and were consequently predicted to disrupt normal protein expression (Figure 4—figure supplement 10). We observed a statistically significant decrease in the length of the DL using these morpholinos, consistent with our previous observations with other knockdown reagents and pharmacological inhibitions, and thus lending further credence to the conclusion that DE and DL segmentation is reliant on Ptger2a or Ptger4a expression (Figure 4—figure supplement 10). Taken together, these data suggest that Ptger2a and Ptger4a have developmental roles in renal progenitors where they interact with PGE2 to regulate distal nephron segment formation.

Alterations in Ptgs activity or PGE2 levels influences nephron segmentation after gastrulation

Since we found that ptger2a and ptger4a transcripts were expressed within renal progenitors beginning as early as the 12 ss, we hypothesized that prostaglandin signaling may begin to operate at that time period to influence pronephros segmentation. To test this, we treated WT embryos with either the nonselective Cox1/2 antagonist indomethacin (30 µM) to block Ptgs activity or the agonist dmPGE2 (100 µM) from the 12 ss through to the 24 hpf time point, and then performed WISH to assess the pronephros segments. Indomethacin treatment during this time window elicited an expansion of the DE domain and a reduction of the DL similar to that seen from indomethacin treatments from 4 hpf to 24 hpf (Figure 5A). Further, dmPGE2 treatment from the 12 ss to 24 hpf was sufficient to induce an expansion of the PST and a reduction of the DL (Figure 5B). Absolute measurements of these segments changes in indomethacin and dmPGE2 treated embryos revealed that they were significant compared to controls and were similar to pharmacological exposures performed between the 4 hpf and 24 hpf time period (Figure 5C). Interestingly, we also found that treatments with either of two different Ptger2 small molecule antagonists, PF04418948 or AH6809, from the 12 ss to 24 hpf was likewise sufficient to induce a statistically significant expansion of the DE segment and reduction of the DL segment compared to WT controls (Figure 5—figure supplement 1). These data identify the 12 ss through the 24 hpf time period as the critical interval when pronephros progenitors require PGE2 signaling for normal segment development.

Figure 5 with 1 supplement see all
Prostaglandin signaling influences nephron patterning after gastrulation.

Embryos were treated with 0.3% DMSO or 1% DMSO vehicle control, (A) 30 μM indomethacin, or (B) 100 μM dmPGE2 from the 12 ss to the 28 ss and stained for the PST (trpm7), DE (slc12a1), and DL (slc12a3) (purple) using WISH at the 28 ss stage. Black bars indicate segment gene expression domain. Red scale bar, 70 µm. (C) The PST, DE, and DL domains were measured in microns (n = 5 for each control and experimental group). Data are represented as ± SD, ANOVA used to compare samples, *p<0.01, **p<0.05, where n.s. indicates not significant.

https://doi.org/10.7554/eLife.17551.022

Cox expression and PGE2 levels affect renal progenitor transcription factor domains

Several transcription factors are known to be critical for proper tubule segment patterning during zebrafish pronephros development. Some of these include: sim1a, which is expressed throughout both the PCT and PST, and is essential for PST fate (Cheng and Wingert, 2015); irx3b, which is expressed throughout both the PST and DE, and is essential for DE segment fate (Wingert and Davidson, 2011Morales and Wingert, 2014); and mecom, which is expressed dynamically along the renal progenitor field, ultimately becoming restricted to the DL domain where it is essential for normal formation of this segment (Li et al., 2014). Given these genes’ respective roles in segment ontogeny, we hypothesized that PGE2 signaling affects renal progenitor fate by modulating the expression domains of one or more of these crucial factors. To investigate this, WT embryos were treated with either a control vehicle DMSO, dmPGE2, or indomethacin between the 4 hpf stage and the 20 ss, and then the spatial distribution of sim1a, irx3b, or mecom transcripts in the IM renal progenitors was assessed by WISH (Figure 6, Figure 6—figure supplement 1).

Figure 6 with 1 supplement see all
Prostaglandin signaling modulates key nephrogenesis transcription factors.

(A) Embryos were treated with a 1% DMSO control, dmPGE2 at 100 μM, or indomethacin at 30 μM from 4 hpf to the 20 ss. WISH was used to stain for the transcription factors sim1a, mecom, and irx3b (purple) and the somites (red) at the 20 ss stage. Black bars indicate segment gene expression domain. Red scale bar, 70 µm. (B) The sim1a, mecom and irx3b domains were measured in microns (n = 5 for each control and experimental group). Data are represented as ± SD, ANOVA used to compare samples, *p<0.01, where n.s. indicates not significant.

https://doi.org/10.7554/eLife.17551.024

Interestingly, we found that dmPGE2 treatment led to a significant expansion of the sim1a and irx3b domains, along with a significant reduction of the mecom domain, in the majority of embryos (Figure 6A,B). These alterations are consistent with the observation that dmPGE2 expands the PST segment and restricts the DL (Figure 2). Further, we found that indomethacin treatment led to no significant change in the sim1a domain, while the irx3b domain was significantly increased in absolute length and the mecom domain was significantly reduced in length (Figure 6A,B). These alterations are in keeping with the prior observations that nonselective Cox inhibition with indomethacin, selective Cox1 or Cox2 inhibitors, as well as knockdown or inhibition of Ptger2a and Ptger4a expanded the DE segment and reduced the DL (Figures 3,4). In summary, these data suggest that PGE2 signaling influences segment programs in part by affecting the expression of sim1a, irx3b, and mecom, either directly or indirectly, to mediate nephron segmentation.

Cox activity and PGE2 act upstream of irx3b and sim1a

Next, we tested the epistatic relationships between prostaglandin signaling and the essential transcription factors irx3b, mecom, and sim1a. As knockdown of irx3b results in loss of the DE segment (Wingert and Davidson, 2011), we exposed irx3b morphants to the Ptgs1 (Cox1) selective inhibitor SC-560 to test how the combined deficiency of irx3b and prostaglandin synthesis would impact the process of DE development during nephron segmentation. Embryos that were treated with SC-560 concomitant with irx3b deficiency failed to form a DE segment, similar to irx3b deficiency alone (Figure 7A,B). Taken together with the observation that indomethacin treatment was sufficient to expand the irx3b expression domain, this result is consistent with the conclusion that prostaglandin signaling occurs upstream of irx3b to regulate DE segment development.

Figure 7 with 2 supplements see all
Prostaglandin signaling acts upstream of the transcription factors sim1a and irx3b.

(A) Embryos were microinjected with the irx3b MO and treated with 0.5% DMSO or 50 µM SC-560 from 4 hpf to 24 hpf and stained for the DE using WISH. WT control embryos were also treated with 0.5% DMSO and 50 µM SC-560. (B) The DE (slc12a1) (purple) was stained for using WISH and quantified in triplicate, with at least 20 embryos per control and experimental group, by observing for the presence of the DE segment at the 24 hpf stage. Black bars indicate segment gene expression domain. Red scale bar, 70 µm. (C) Embryos were microinjected with the sim1a MO then treated with either 1% DMSO or 100 µM dmPGE2 from 4 hpf to 24 hpf. 1% DMSO and 100 µM dmPGE2 was also applied to WT control embryos from 4 hpf to 24 hpf. The PST (purple) and somites (red) were then stained for using WISH. (D) Quantification was undertaken in triplicate based on the presence of a PST or lack thereof, with at least 20 embryos per control and experimental group. Data are represented as ± SD, with t tests comparing each drug treatment to the corresponding DMSO control group, where n.s. indicates not significant.

https://doi.org/10.7554/eLife.17551.026

As we observed that the domain of mecom expression in renal progenitors is restricted when prostaglandin production was blocked with indomethacin (Figure 6), and that mecom deficiency in turn is associated with a reduced DL, we next explored whether restoration of mecom would be sufficient to rescue DL segment development in the absence of normal prostaglandin synthesis. To test this, mecom capped mRNA was synthesized and microinjected into 1-cell stage embryos, which were then treated with indomethacin or DMSO control as previously described (Figure 7—figure supplement 1). Overexpression of mecom at dosages ranging from 14–70 pg, in either the experimental or vehicle control group, was not sufficient to alter DL segment fate (Figure 7—figure supplement 1, data not shown). At higher dosages of mecom cRNA, embryos were dysmorphic, which has been reported previously (Li et al., 2014), precluding further examination with this approach. Despite these negative results, the alterations in the mecom expression domain in indomethacin treated embryos suggest that prostaglandin synthesis likely acts upstream of mecom to influence DL segment development.

Finally, we explored the relationship between sim1a and PGE2 signaling. Previous work from our laboratory has demonstrated that sim1a overexpression is sufficient to expand the PST segment (Cheng and Wingert, 2015). In light of this, along with our present finding that exogenous PGE2 treatment induces both an expanded PST segment and expanded sim1a domain, we hypothesized that the gain of function prostaglandin phenotype was reliant on sim1a for the alteration in PST fate. To test this, embryos were microinjected at the 1-cell stage with a sim1a (Cheng and Wingert, 2015), then incubated in either DMSO control or treated with dmPGE2 between 4 and 24 hpf. We found that knockdown of sim1a concomitant with dmPGE2 treatment led to an abrogation of the PST segment, similar to sim1a deficiency alone (Figure 7C,D, Figure 7—figure supplement 2). These results are consistent with the conclusion that sim1a acts downstream of PGE2 signaling in the context of exogenous treatment to drive expansion of the PST segment.

Discussion

Understanding the genetic factors necessary to generate different cell types is an important aspect of developmental biology. Knowledge of these, along with an appreciation of the modulators that can impact the genesis of cell lineages, including their related morphogenetic processes, provides powerful insights relevant to congenital defects, disease pathology, regeneration and in vitro reprogramming (Morales and Wingert, 2014). Relevant to the present report, the kidney organ has many associated congenital diseases, and there is an escalating incidence of acute and chronic renal diseases for which a deeper understanding of mesodermal developmental processes has many possible applications (Nakanishi and Yoshikawa, 2003).

In this study, we uncovered evidence of a role for prostaglandin signaling in nephron segment formation during embryonic zebrafish development. Prostaglandins have diverse and potent biological actions (Funk, 2001), however, their effects on developing tissues including stem cells have only recently begun to be appreciated (North et al., 2007; Goessling et al., 2009; Nissim et al., 2014). By conducting a chemical genetic screen in zebrafish embryos to identify factors that affect nephron development, we found that several prostaglandin moieties were capable of modulating proximal-distal segmentation, first validating PGB2 and then subsequently identifying PGE2 as well (Figure 8). Specifically, we demonstrated that the addition of exogenous PGE2 or PGB2 was sufficient to increase PST segment size in a dosage-dependent manner. Using several chemical and genetic approaches, we demonstrated that abrogated prostaglandin activity alters formation of the DE and DL nephron tubule segments, where deficiencies in ptgs1, ptgs2a, or the PGE2 receptors encoded by ptger2a and ptger4a led to an expanded DE and reduced DL segment, and that PGE2 could specifically rescue the loss of Ptgs1 or Ptgs2a. We also determined that changes in Cox-mediated prostaglandin synthesis or PGE2 correlated with alterations in the expression domains of essential segmentation transcription factors in renal progenitors, suggesting some mechanisms by which prostaglandin signaling acts to influence segmentation (Figure 8). Understanding how these changes relate with morphogenesis involving cellular dynamics (e.g. proliferation, turnover) or even migration of the renal progenitors will be important aspects for future investigations.

The role of prostaglandin signaling during zebrafish nephrogenesis.

(A) PGE2 is generated by the Ptgs1 and Ptgs2a enzymes, which interact with the Ptger2a and Ptger4a receptors to define the boundaries of the DE by negatively regulating the domain of irx3b expression and promoting DL identity. (B) Exogenous prostaglandin treatment promotes PST identity, partly by regulating sim1a expression, and these alterations occur at the expense of the DL segment. Not depicted: exogenous PGE2 is also associated with an increased irx3b domain, which correlates with the shifted posterior location of the DE in the context of PST expansion. Further, in both (A) and (B), the mecom expression domain is reduced in length, which is likely part of the mechanism underlying the DL segment size reduction in these cases.

https://doi.org/10.7554/eLife.17551.029

Previous studies have shown that PGE2 is among the major prostanoids produced during the first day of zebrafish embryogenesis (Cha et al., 2005, 2006). Based on this and our ability to rescue Cox1/2 deficiency with dmPGE2 treatment alone, we theorize that PGE2 is the central endogenous signaling component that affects pronephros development. Further, the localization of ptger2a and ptger4a expression in renal progenitors suggests that these cells directly receive and respond to prostaglandin signals. Since prostaglandins are known to act in an autocrine or paracrine fashion with short half-lives, we speculate that the IM renal precursors themselves or nearby tissues like the paraxial mesoderm are the local source(s) of prostanoids. Based on the results of several epistasis experiments, we currently hypothesize that PGE2 signaling restricts the DE fate in part by negatively regulating irx3b, and that elevated PGE2 levels expand the PST segment through positive regulation of sim1a, though future studies are needed to delineate if these interactions are direct or indirect (Figure 8). Finally, our data repository of small molecules that affect pronephros ontogeny will provide a useful starting point for future studies by our lab and others to study nephrogenesis further in zebrafish or in other vertebrates (Figure 1—source data 1).

Mechanisms of prostaglandin signaling during nephrogenesis

Early genetic studies that interrogated the effect of disrupting prostaglandin synthesis in homozygous mutant COX-1 or COX-2 mice did not report observing overt abnormalities at birth (Langenbach et al., 1995; Morham et al., 1995; Mahler et al., 1996). Similarly, knockouts of other biosynthesis enzymes and the prostaglandin receptors had normal neonatal phenotypes (Sugimoto et al., 2000; Kobayashi and Narumiya, 2002a2002b). Intriguingly, however, mice lacking COX-2 exhibit postnatal kidney pathologies associated with neonatal fatality, including nephron hypoplasia and atrophy, impaired cortical growth, and even cyst formation in multiple nephron segments—phenotypes suggestive of significant disruptions in renal ontogeny (Dinchuk et al., 1995; Morham et al., 1995; Mahler et al., 1996). These observations were not further explored until recently, however, where COX-2 gene dosage and pharmacological inhibition were linked to renal defects in glomerular size, such that COX-2+/- mice were found to exhibit kidney insufficiency (Slattery et al., 2016). In alignment with these data, exposure to COX inhibitors like indomethacin during human development is associated with renal failure, where the histological aspects include small glomeruli and microcystic lesions among other defects (Gloor et al., 1993; van der Heijden et al., 1994; Kaplan et al., 1994).

In light of these recent observations along with our findings, we propose that prostaglandin signaling, likely through PGE2, has critical roles in nephrogenesis, which warrant further investigation. However, there are significant challenges of studying nephron formation in mammalian models due to their in utero development, and while metanephric organ culture has been informative for branding morphogenesis studies, it is not conducive to studying nephrogenesis. Therefore, the zebrafish pronephric model provides an alternative for continued genetic dissection of the cellular and molecular effects in nephron development due to alterations in prostaglandin levels. Based on the requirement for PGE2 during pronephros development, it will be interesting to explore its effects on renal progenitors in other stages of zebrafish kidney ontogeny, and during new nephron formation and epithelial regeneration events in adults as well (McCampbell et al., 2015). In the near future, emergent organoid technologies will likely provide a complementary in vitro experimental system to probe the mechanisms of prostaglandin signaling in mammalian nephrogenesis (Chambers et al., 2016).

Further, while our report documents an essential role for PGE2 signaling during nephron formation, more work needs to be done to further understand the genetic networks that affect segment fate. For example, we have previously shown that RA acts as a morphogen in the zebrafish, and that a gradient of RA induces proximalization of the impending pronephros (Wingert et al., 2007; Wingert and Davidson, 2011). It is intriguing to speculate that a prostaglandin signaling gradient, where PGE2 acts as a morphogen, as recently proposed (Nissim et al., 2014), may articulate with RA to balance proximo-distal specification of the renal progenitors, though further studies are needed to interrogate this possibility. In addition, prostaglandins have been shown to initiate transcription through either interacting with cognate Ptger (EP) receptors or alternatively, passing through the cell membrane and binding with peroxisome proliferator-activated receptors (PPARs) (Guan and Breyer, 2001; Berger and Moller, 2002). Interestingly, PPARs can heterodimerize with Retinoid X Receptors (RXRs), a nuclear receptor for RA (Guan and Breyer, 2001; Berger and Moller, 2002). Also, it has been shown that different prostaglandins can interact interchangeably, at the right concentration, with the various Ptger receptors (Tootle, 2013). This might explain why different prostaglandins could induce an expansion of the PST.

Understanding the broader roles of PGE2 signaling in development

Prostaglandins have only just started to become recognized as important factors and determinants of cell fate decisions and growth during development. These new roles for prostaglandins have been revealed in part through a study showing that PGE2 activity has a conserved function to expand the domain of HSCs in development and enhance their ability to home to the bone marrow in transplants (North et al., 2007). PGE2 is currently in phase two clinical trials for increasing the efficiency of bone marrow transplants (Hagedorn et al., 2014). Furthermore, PGE2 was shown to be a regulator of bipotential endoderm cell fate decisions in development for zebrafish and mouse endodermal cells (Nissim et al., 2014). Curiously, it was also shown that PGE2 activity later in organ formation induced proliferation of both the liver and pancreas buds. This change in the function of PGE2 is explained by the spatial and temporal expression of ptger2a and ptger4a, where blocking ptger2a promoted liver versus pancreas specification and blocking ptger4a promoted the outgrowth of the liver and pancreas buds. These cornerstone studies, along with the present report, give credence to the notion that PGE2 is a key regulator of progenitor populations during embryogenesis and set the stage for further inquiry into how prostaglandin signaling affects developing cell populations. As more knowledge comes to light about how PGE2 and other prostaglandins influence ontogeny, they are likely to become an increasingly intriguing option for clinical therapeutic applications.

Materials and methods

Zebrafish husbandry

Zebrafish were cared for and maintained in the Center for Zebrafish Research at the University of Notre Dame using experiments approved under protocol 16–025. Adult Tübingen strain fish were used for these studies, and their offspring were staged as described (Kimmel et al., 1995).

Chemical genetic screen and other chemical treatments

Zebrafish wild-type (WT) embryos were arrayed and treated with small molecules using the ICCB Known Bioactives Library as described (Poureetezadi et al., 2014). Zebrafish embryos were staged at 2 hpf, then at least 30 fertilized embryos were arrayed into the chambers of 24-well plates and incubated at 28°С in E3 media until just prior to 4 hpf. Working stocks of small molecules were stored at −80°С, then dissolved in 100% DMSO to make 10 mM concentrations (Lengerke et al., 2011). For drug exposure, the E3 media was completely drawn off the embryos using a glass transfer pipet and the appropriate solution of DMSO, PGB2, dmPGE2, indomethacin, SC-560, NS-398, AH6809, or PF04418948 was applied at a discrete development time point (eg 4 hpf or 12 ss) (American Bioanalytical, Enzo Life Sciences, Santa Cruz) (North et al., 2007; Eisinger et al., 2007; Jin et al., 2014). Embryos were raised to the 20 ss or 24 hpf, washed three times with E3, then fixed in 4% paraformaldehyde. For rescue of prostaglandin activity, ptgs1 or ptgs2a deficient embryos were generated, and cohorts of approximately 30 embryos were arrayed in 24-well plates with E3, then placed in a 28°С incubator until 4 hpf. E3 was then completely drawn off the wells using a glass transfer pipet and a 50 µM concentration of dmPGE2 was applied. The embryos were placed into a 28°С incubator, raised until 24 hpf, washed three times with E3, and fixed as previously described.

Embryo staining for WISH, o-dianisidine, and image acquisition

WISH was conducted as described (Cheng et al., 2014). RNA probes were digoxigenin or fluorescein labeled and generated by in vitro transcription using plasmid templates as described (Wingert et al., 2007; Wingert and Davidson, 2011; Lengerke et al., 2011; Li et al., 2014; Cheng and Wingert, 2015). For o-dianisidine staining, embryos were treated with 1% DMSO, 50 µM dmPGE2, or 30 µM, were allowed to develop until 48 hpf and o-dianisidine staining was performed (Wingert et al., 2004). Images were taken using a Nikon Eclipse Ni with a DS-Fi2 camera. Figures were assembled using Adobe Photoshop CS5.

Morpholino knockdown and RT-PCR

Antisense morpholino oligonucleotides (MOs) were purchased from Gene Tools, LLC. MOs were solubilized in DNase/RNase free water to create 4 mM stocks and stored at −20°С. WT embryos were collected after fertilization, injected with approximately 1 nl of diluted MO at the 1-cell stage and then placed in a 28°С incubator until the desired stage. MO sequences and dosages used were: irx3b 5'-ATAGCCTAGCTGCGGGAGAGACATG-3', 1 ng (Wingert and Davidson, 2011); ptger2a MO1 5'-GATGTTGGCATGTTTGAGAGCATGC-3', 3 ng (North et al., 2007); ptger2a MO2 5'-ACTGTCAATACAGGTCCCATTTTC-3', 1.6 ng (North et al., 2007); ptger2a MO3 splice 5'-CAATAAATCTTACTATTAACGGCAG-3', 3 ng; ptger2a MO4 splice 5'-ATGTACACACGGATCTG-AAGAGAAG-3', 3 ng; ptger4a MO1 5'-CGCGCTGGAGGTCTGGAG-ATCGCGC-3', 3 ng (North et al., 2007); ptger4a MO2 5'-CACGGTGGGCTCCATGCTGCTGCTG-3', 3 ng (Cha et al., 2006); ptger4a MO3 splice 5'-CCTGGAACTTACAACAAGCGGGATT-3', 3 ng; ptger4a MO4 splice 5'-TGAGAAACA-CCTGGACCTGCCAGAA-3', 3 ng; ptgs1 MO 5'-TCAGCAAAAAGTTACACTCTCTCAT-3’, 3 ng (North et al., 2007); ptgs2a MO 5'-AACCAGTTTATTCATTCCAGAAGTG-3', 3 ng (Grosser et al., 2002); ptgs1 MO splice 5'-AACTTTCATTGCTC-ACCTCTCATTG-3', 2 ng; ptgs2a MO splice 5'-ATTCAACTTA-CACAACAGGATATAG-3', 2 ng (Yeh et al., 2009), sim1a MO 5'-TCGACTTCTCCTTCATGCTCTACGG-3', 1 ng (Cheng and Wingert, 2015). To assess knockdown efficacy, RT-PCR and sequence analysis was performed as previously described (Marra and Wingert, 2016) and using the following primers, where uppercase letters indicate location in an exon and lowercase indicates primer location in an intron: ptgs1-F1 5'-TTTATTTATTTGCAGCTTTTTCTT-3'; ptgs1-R1 5'-CAGTGTTTGATGAAGTCGGGCTTTC-3'; ptgs2a-F1 5'-CTGAGCTTCTCACACGCATCAAAT-3'; ptgs2a-R1 5'-GGCGAAGAAAGCAAACATGAGACT-3'; ptger2a-F1 5'-AGACCGAGCGTATGCCAATGT- 3'; ptger2a-R4 5'-caggagggctaataattcagactt-3'; ptger2a-F3 5'-ctgtttcagtgatcagtttgt-3'; ptger2a-R7 5'-CCGCAGAGCTATGAGATCAGTC-3'; ptger2a-R8 5'-GCTGAGGATGATGAACACCAAG-3'; ptger4a-F3 5'-ATGGTCATCCTGTTGATCGCC-3'; ptger4a-R2 5'-aatgagagtcctggaacttac-3'; ptger4a-F5 5'-gggtgtagtcatttatgttgagca-3'; ptger4a-R5 5'-CAGGACCGCTTTACGCAGTAAG-3'.

Quantification of phenotypes, segment measurements, imaging and statistical analysis

Gene domains were assessed with respect to somite boundaries to assess pattern formation (Wingert et al., 2007). Segment domains were analyzed and counted in triplicate with at least 15 embryos per replicate. To measure absolute segment lengths, images were taken of at least five representative embryos. Images were collected using a Nikon Eclipse Ni with a DS-Fi2 camera and measurements performed with Nikon Elements Software. The average was generated and standard deviation (± SD) calculated, and unpaired (student) t-tests were performed to compare experimental groups with the corresponding wild-type control group. In cases where there were several percentage categories of phenotypes, the statistical comparison was performed between like categories, e.g. percentage increased were compared between control and each experimental treatment. In addition, ANOVA tests were performed to assess statistical significance in the context of comparing multiple samples.

References

  1. 1
  2. 2
  3. 3
  4. 4
  5. 5
  6. 6
  7. 7
  8. 8
  9. 9
  10. 10
  11. 11
  12. 12
  13. 13
  14. 14
  15. 15
  16. 16
  17. 17
  18. 18
  19. 19
  20. 20
    Prenatal maternal indomethacin use resulting in prolonged neonatal renal insufficiency
    1. JM Gloor
    2. DG Muchant
    3. LL Norling
    (1993)
    Journal of Perinatology  13:425–427.
  21. 21
  22. 22
  23. 23
  24. 24
  25. 25
  26. 26
  27. 27
  28. 28
  29. 29
  30. 30
  31. 31
  32. 32
  33. 33
  34. 34
  35. 35
  36. 36
  37. 37
  38. 38
  39. 39
  40. 40
  41. 41
  42. 42
  43. 43
  44. 44
  45. 45
  46. 46
  47. 47
  48. 48
  49. 49
  50. 50
  51. 51
  52. 52
  53. 53
  54. 54
    Organogenesis of the Kidney
    1. L Saxen
    (1987)
    Cambridge: Cambridge University Press.
  55. 55
  56. 56
  57. 57
  58. 58
  59. 59
  60. 60
  61. 61
  62. 62
    Zebrafish nephrogenesis involves dynamic spatiotemporal expression changes in renal progenitors and essential signals from retinoic acid and irx3b
    1. RA Wingert
    2. AJ Davidson
    (2011)
    Developmental Dynamics, 240, 10.1002/dvdy.22691, 21761484.
  63. 63
  64. 64
  65. 65

Decision letter

  1. Tanya T Whitfield
    Reviewing Editor; University of Sheffield, United Kingdom

In the interests of transparency, eLife includes the editorial decision letter and accompanying author responses. A lightly edited version of the letter sent to the authors after peer review is shown, indicating the most substantive concerns; minor comments are not usually included.

Thank you for submitting your article "Prostaglandin signaling regulates nephron segment patterning of renal progenitors during vertebrate kidney development" for consideration by eLife. Your article has been reviewed by three peer reviewers, and the evaluation has been overseen by a Reviewing Editor (Tanya Whitfield) and Sean Morrison as the Senior Editor. The following individuals involved in review of your submission have agreed to reveal their identity: Dirk Meyer (Reviewer #1); Lilianna Solnica-Krezel (Reviewer #3).

The reviewers have discussed the reviews with one another and the Reviewing Editor has drafted this decision to help you prepare a revised submission.

Summary:

This is an interesting study on that addresses the factors that influence cell fate decisions in the developing vertebrate kidney, using the embryonic zebrafish pronephros as a model system. The authors have used a chemical genetic approach to identify prostaglandins as mediators that influence cell fate choice during development of the proximal and distal nephron segments. Results from the pharmacological screen have been validated through the use of morpholino-based gene knockdown to inhibit prostaglandin synthesis. The authors propose that the prostaglandin pathway acts upstream of the previously-identified transcriptional regulators Irx3b and Sim1a in the developing pronephros. The reviewers and editors agreed that the work was interesting and potentially significant. However, they also had some concerns that should be addressed to strengthen the manuscript, as outlined below.

Essential revisions:

1) Title: as suggested by reviewer 1, please change the word 'vertebrate' to 'zebrafish'. (Please note, however, that it is not essential to extend the work to other vertebrate species or different stages of kidney development.)

2) Given the known interactions between prostaglandin and other pathways such as Wnt signalling, there are concerns over the relevance of the phenotypes resulting from long-term treatment (4-22 hpf) with PG agonists for a nephron-specific patterning role of prostaglandins. These concerns are exacerbated by the rather mild nephric phenotypes described for prostaglandin pathway morphants. The option of indirect phenotypes therefore needs further discussion. In addition, experiments with later onset of agonist treatment are also required to exclude interference with general early embryonic patterning, for example by Wnt signalling. See comments from reviewer 1.

3) Following discussion between the reviewers, it was felt that a full analysis of any connection with Wnt signalling was not essential, but might be an interesting extension of the work in the future.

4) Measurements. There were concerns over the measurements and quantitation of the data, in particular, in the way that the extent of expression domains was measured (by a horizontal bar in terms of somite length, vs the actual length of the curved domain of gene expression shown). It was felt that more accurate measurements should be provided. In addition, n numbers should always accompany percentage values, wherever shown.

5) Effects on surrounding tissues. Please provide some assessment of the extent to which surrounding tissues are affected by the treatments shown.

6) The text needs tightening up in places, especially with reference to the distinction between PGs in general and PGE2 (see comments from reviewer 2).

7) For the experiments using morpholinos, it is important that these were done as rigorously as possible. Further supporting evidence is needed here (see comments from reviewer 3).

Reviewer #1:

In their manuscript Poureetezadi et al. use the zebrafish pronephros as a model to study functions of the prostaglandin pathway on nephron formation and patterning. Prostaglandins and in particular PGE2 have well documented functions in postnatal mammalian kidney development and renal physiology. In contrast, very little is known about activities in early kidney formation, mainly as genetic studies in mouse provided no major evidence for such requirements. In a screen for small compounds affecting zebrafish pronephron formation Poureetezadi et al. identified several prostaglandin agonists as agents causing caudally shifted or expanded tubule segments upon treatment of embryos between 4-24hpf. Detailed pharmacological analyses confirmed dose-dependent activities of PGE2 and the PGE2-derived PGB2 in tubule segmentation, and provided hints for a requirement of prostaglandin synthesizing enzymes COX1/2 and the PGE2 receptors Ptger2a/4a in the precise positioning of caudal segments. The pharmacological loss-of-function data were further confirmed by knockdown studies using morpholinos that supposedly prevent expression of COX1/2 and Ptger2a/4a proteins. Further, the authors show that nephric phenotypes seen after early embryonic interference with prostaglandin signaling correlate with changed expression of key transcriptional regulators of tubule segmentation and they provided evidence for a genetic role of prostaglandin signaling upstream of these factors. This interesting work is the first one demonstrating prostaglandin activities in nephric tubule patterning.

Unfortunately, the authors had addressed neither the underlying molecular mechanism nor the relevance for other vertebrate model systems or the metanephric kidney.

As pointed out by the authors, prostaglandins only recently became recognized as developmentally import fate determinants. In this context interference with canonical wnt-signaling had been identified as one of the most critical prostaglandin activities. Considering the importance of Wnt-signaling in kidney formation and patterning it is surprising that this connection had not been further analyzed.

The word 'vertebrate' in the title is not justified as the zebrafish is the only vertebrate that had been analyzed. Slattery et al. (2016) recently suggested a role of COX2 not only in postnatal but also in embryonic metanephric kidney development. While these mouse data support a possibly conserved requirement for prostaglandins, it should be noted that the phenotypic analyses of COX2 mutants was restricted to the glomerulus and no tubule data were provided.

The authors suggest a direct connection between the proposed nephron specific expression of ptger2a/4a (Figure 4—figure supplement.4) and the nephric loss and gain of function phenotypes. Expression of prostaglandin pathways components is not well documented and the experimental settings leave space for various alternative explanations. The images shown in Figure 4—figure supplement.4 suggest a restricted expression of ptger2a/4a in proximal tubule of 12-24ss embryos. In case of a direct function, pharmacological treatments starting at 12ss rather than at 4hpf should be sufficient to induce relevant gain and loss of function phenotypes. Additional data should be provided to confirm nephron specificity and to give details on the posterior-anterior extent of ptger2a/4a expression (for example sections and co-stains with nephric markers). Further, studies should be performed to determine the critical time-window of prostaglandin responsiveness.

Reviewer #2:

In the manuscript entitled "Prostaglandin signaling regulates nephron segment patterning of renal progenitors during vertebrate kidney development" by Poureetezadi et al., stemming from a chemical screening approach, the authors uncover a role for prostaglandin activity which impacts distal vs. proximal segment development of the pronephric tubules in the kidney. Specifically they found that exposure to prostaglandin agonists expanded the proximal segment (slc20a1a+) lineage and inhibited distal fate (slc12a1a+). In contrast inhibition of PG synthesis altered the distribution of specific distal fates. The authors also identified downstream consequences on key segment associated genes (irx3b and sim1a) that appeared to be influenced, directly or indirectly by PG modulation. The data presented throughout the paper strongly support the authors claims, however, descriptions of prostaglandin synthesis and signaling cascades are over-simplified, and general claims are made for "prostaglandins" when only PGE2-specific components are thoroughly investigated. As several types of prostaglandins were isolated in the screen, which can act in methods different from that investigated here (thromboxane receptor and PPARg), more precise language and a rationale for selecting PGE2 (not hit in the screen) would aid the reader. Similarly, while the data is nicely presented and logically organized to support the authors conclusions, the paper would benefit from addition of quantitative evaluations to aid appreciation by those outside of the field.

1) The authors over-simplify prostaglandin synthesis and signaling pathways, applying terminology and biological function relevant to PGE2 under the broad headline of "prostaglandins". This occurs as early as Figure 1, where they identify PGD2, PGA2, PGB2, and PGJ2 in the screen, which are known to stimulate thromboxane receptors (PGA2, B2) and PPARg (PGJ2), yet follow PGE2 (not a hit) and its relevant machinery. Similarly, Ptges is the enzyme for prostaglandin E2 synthesis (hence the "e" in its name, not general secondary processing), and Ptgers are the receptors for PGE2, not all of the PGs. While PTGS inhibition (Cox enzymes), could indeed broadly impact PG synthesis, the rest is quite specific to PGE2 and should not be generalized in the text, particularly when the screen hit several other potential modifiers as relevant to kidney biogenesis. This is not to say the data for PGE2 modification is wrong, it is just over generalized to "prostaglandins" and as such could cause issues for future investigations.

2) Given the independent signaling cascades associated with PGB2 and PGE2, it is surprising that exposure to both enzymes elicited the same biological effect. While receptor modulation is examined for 2 of the 4 PGE2 receptors, similar studies should be done in the context of exogenous PGE2 and PGB2 addition, as well as with the thromboxane receptors to confirm that addition of exogenous levels of each prostanoid doesn't cause errant signaling.

3) The doses utilized for dmPGE2 and Indomethacin are substantially higher than that found in other zebrafish or mouse papers, and well above physiological concentrations in humans. While the embryos appear grossly normal in whole mount images, it is important to document that alterations in kidney associated expression patterns are not simply due to gross morphological development issues or off target effects. Assessment of vascular markers by in situ or with a reporter line (Grosser et al) should be a quick way to test toxicity; receptor blocking analysis should confirm specificity.

4) While the overall conclusions seem well supported by the data shown, to make the paper more accessible to a general audience, the authors should use a more precise way to quantify the alterations in expression observed, or at least confirm the relevant ones, using ImageJ and/or a reporter line. The assays are currently very observational (with straight bars correlated with expression drawn on curvy embryos) and it is unclear how definitive statements like "50% reduction" are made when no quantification is shown.

Reviewer #3:

Starting with a chemical screen in developing zebrafish embryos, the Wingert and collaborators, identify components of prostaglandin signaling pathway as regulators of proximo-distal patterning of developing kidney. Using treatments with the pathway agonists (dmPGE2, PGB2) or inhibitors, and antisense morpholino oligonucleotides to impair expression of the PG synthetizing enzymes and their receptors (Ptger2a, Ptger4a), the authors provide evidence that prostaglandin signaling limits distal segment formation while promoting the proximal pronephric fates, acting upstream of Irx3b and Sim1a transcription factors, which have been previously implicated in this process. This is a very interesting work that expands our appreciation of the roles of prostaglandin signaling in cell fate specification during vertebrate embryogenesis. These roles were difficult to discern in the murine model, but recent studies in mouse and human point to PG involvement in kidney development and function. The authors leverage the experimental strengths of the zebrafish model, in which prostaglandin signaling can be easily manipulated during kidney development and the resulting effects can be monitored.

Whereas the proposed conclusions are significant and would be of interest to the developmental biology and renal research communities, they are not sufficiently supported by the presented data. Addressing the following questions and concerns would significantly strengthen the manuscript and make it suitable for publication.

One of the major concerns about the current manuscript is that there are many quantifications presented but without sufficient experimental detail to evaluate their statistical or biological significance. For several experiments (e.g. treatment of antagonists), there is insufficient experimental detail to evaluate the results and their interpretation. The quantifications of the phenotypes (reduction or expansion) of segments in various experimental regimens is given throughout the manuscript as% of embryos with altered expression (expanded or reduced). This shows that the fraction of affected embryos increases in a dose dependent manner but does not address whether the expressivity (degree of reduction or expansion) of the phenotype is also dose dependent. This is important given the overall conclusion that prostaglandins regulate segment patterning.

Were the measurements carried in a blinded fashion?

The number of embryos in individual experiments showing% of phenotype should be provided.

For Indomethacin treated embryos, the authors conclude that they "developed a 50% larger DE segment domain and a 20% smaller DL segment domain" referring to Figure 3 and its supplement. It is not clear where the data are presented on which this conclusion has been reached given that Figure 3 shows fractions of affected embryos and not dimensions of pronephric segments in control and treated embryos. Such conclusions need to be supported by clearly presented morphometric data, with the method of measurement clearly described, as well as the numbers of analyzed embryos.

The authors conclude that the changes in the dimensions of the pronephros proximo-distal segments represent the specific effects of prostaglandin signaling. However, the possibility of broader effects of prostaglandin signaling on embryonic dimensions and thus indirectly on pronephros is not sufficiently addressed. In the Methods the authors state "Gene domains were measured with respect to somite boundaries to assess pattern formation". However, the somitic staining that should provide such a reference is not clear in most of the figures. Given that earlier reports implicated prostaglandin signaling in morphogenetic movements of gastrulation (Gasser et al., PNAS, 2002; Cha et al., Genes & Dev, 2006; Speirs et al., Development, 2011), this needs to be carefully analyzed. Anteroposterior dimensions of the treated embryos should be measured as well as the relative size of pronephric segments to the neighboring somites for a reviewer/reader to evaluate.

In experiments in which PG signaling was inhibited with various agents, it is not clear at what developmental stages was this done? Given the previous work implicating PG signaling in gastrulation movements the timing of the treatment is important. It would be also of interest to identify when during embryogenesis PG signaling is required for segment patterning and when it can affect it.

Injection into 1 cell stage embryos of translation and splicing morpholinos targeting ptgs1 (cox1) and ptgs2a (cox2) enzymes is reported to affect pronephros patterning. However, it has been previously reported that ptgs1 morpholino injections caused severe epiboly defects (Gasser et al., 2002). In this work the morpholino experiments are insufficiently described. There is no information on the dosage of morpholinos per embryos (only stock concentration is given in methods), no dose response, no experiments describing the degree of loss of function. Yet, the authors conclude "these data indicate that endogenous levels of prostaglandins are critical for specification of the DE and DL segments". This conclusion is given before dmPGE2 rescue experiments are described.

Similar lack of detail and controls is a concern for the experiments in which two genes encoding the prostaglandin receptors, ptger2a and ptger4a are targeted using morpholinos. There are no data on the effectiveness of these reagents to downregulate expression of these genes and encoded proteins, or to lower the levels of prostaglandins. Since in recent years, significant concerns have been raised about the specificity of morpholinos (e.g. Kok et al., Dev Cell, 2015), the sparse details and controls for the experiments presented raise concerns.

In epistasis experiments presented in Figure 6A, B it would be important to see the effect of SC-580 alone. From the perspective of the model of PG – irx3b pathway proposed in Figure 7A, the effect of dmPGE2 treatment that affect PST and DL but do not affect DE are confusing.

[Editors' note: further revisions were requested prior to acceptance, as described below.]

Thank you for resubmitting your work entitled "Prostaglandin signaling regulates nephron segment patterning of renal progenitors during zebrafish kidney development" for further consideration at eLife. Your revised article has been favorably evaluated by Sean Morrison (Senior editor) and Tanya Whitfield (Reviewing editor).

The manuscript has been substantially improved and all reviewers' comments have been addressed well. The additional experimental data help to strengthen and support the findings, and the writing is now clear and precise. Overall, this is a very interesting study and should be of wide interest.

One concern remains: the quantification has been improved but there was very little information about the statistical tests used for analysis of the data. t-tests are mentioned in the Materials and methods section, but these will not be suitable for analysis of some of the datasets, where there are multiple conditions and comparisons. Here, a test such as ANOVA, with appropriate post-test correction for multiple samples, would be preferred. Alternatively, the authors should clarify and justify the tests they have used. The tests that are used should be stated in each of the figure legends where appropriate, in addition to the statement in the Materials and methods.

https://doi.org/10.7554/eLife.17551.030

Author response

Essential revisions:

1) Title: as suggested by reviewer 1, please change the word 'vertebrate' to 'zebrafish'. (Please note, however, that it is not essential to extend the work to other vertebrate species or different stages of kidney development.)

The authors agree and the recommended title change has been made.

2) Given the known interactions between prostaglandin and other pathways such as Wnt signalling, there are concerns over the relevance of the phenotypes resulting from long-term treatment (4-22 hpf) with PG agonists for a nephron-specific patterning role of prostaglandins. These concerns are exacerbated by the rather mild nephric phenotypes described for prostaglandin pathway morphants. The option of indirect phenotypes therefore needs further discussion. In addition, experiments with later onset of agonist treatment are also required to exclude interference with general early embryonic patterning, for example by Wnt signalling. See comments from reviewer 1.

We performed additional studies to address the time window when changes in PG levels affect nephron segment development. First, we treated embryos with either the antagonist indomethacin or agonist dmPGE2 starting at 12 ss with exposure lasting through to the 24 hpf time point. This indomethacin treatment elicited an expansion of the DE domain and reduction of the DL similar to that seen after the treatment lasting from 4 hpf to 24 hpf (new Figure 5A). dmPGE2 treatment between 12 ss and 24 hpf induced an expansion of the PST and a DL reduction similar to the long dmPGE2 exposure time (new Figure 5B). To quantify and further analyze these changes, we determined the absolute segment lengths and found the phenotypes following each PG alteration to be significant compared to controls, while the segment changes were indistinguishable from the corresponding drug exposure over the longer interval (new Figure 5C). Further, we found that treatment with Ptger2 antagonists from the 12 ss to 24 hpf was also sufficient to induce a statistically significant DE expansion and DL reduction (new Figure 5—figure supplement 1), similar to the 4-24 hpf exposure (Figure 4—figure supplement 6; new Figure 4—figure supplement 9).

3) Following discussion between the reviewers, it was felt that a full analysis of any connection with Wnt signalling was not essential, but might be an interesting extension of the work in the future.

We have not pursued the connection to Wnt signaling in the current revision but agree that it will be an interesting extension for our future research.

4) Measurements. There were concerns over the measurements and quantitation of the data, in particular, in the way that the extent of expression domains was measured (by a horizontal bar in terms of somite length, vs the actual length of the curved domain of gene expression shown). It was felt that more accurate measurements should be provided. In addition, n numbers should always accompany percentage values, wherever shown.

We now provide absolute length measurements for all nephron segment phenotypes (Please see Figure 2B; Figure 2—figure supplement 2E, F; Figure 3B, C; Figure 3—figure supplement 1E, F; Figure 4—figure supplement 9). Data involving phenotype penetrance and dosage responses have been moved to supplemental figures. We provide n values in either the graph or legend for all quantitations.

5) Effects on surrounding tissues. Please provide some assessment of the extent to which surrounding tissues are affected by the treatments shown.

We assessed surrounding tissues following PG agonist (dmPGE2, 100 µM) and PG antagonist (indomethacin, 30 µM) exposure by whole mount in situ hybridization (as suggested by reviewer #2), and also examined circulation within the vasculature. We did not observe differences flk1 or gata1 transcript expression at 24 hpf in embryos exposed to DMSO, dmPGE2 or indomethacin between 4 hpf and 24 hpf, indicating normal vascular and primitive blood development, respectively (new Figure 2—figure supplement 4A, B). This is consistent with a published study that reported effects of indomethacin at 40 µM, but normal development of the vasculature following indomethacin treatment at 25 µM (Cha, et al., 2005).

Next, we performed o-dianisidine (benzidene) staining in embryos treated with DMSO, dmPGE2 or indomethacin between 4 hpf and 24 hpf, which labels hemoglobinized erythrocytes and provides a sensitive assessment of defects in circulation or vascular integrity that can be undetected when using only live imaging with stereomicroscopy. Blood flow in these drug treated embryos was equivalent to wild-type control embryos through the 48-55 hpf stage, and we did not observe compromised vessel integrity or hematomas (e.g. bleeding, blood pooling) with o-dianisidine staining (new Figure 2—figure supplement 4C).

6) The text needs tightening up in places, especially with reference to the distinction between PGs in general and PGE2 (see comments from reviewer 2).

We have revised the manuscript extensively to address this weakness and to be more specific throughout.

7) For the experiments using morpholinos, it is important that these were done as rigorously as possible. Further supporting evidence is needed here (see comments from reviewer 3).

In the revision, we provide the data that morpholinos used to interfere with mRNA splicing of ptgs1, ptgs2a, ptger2a and ptger4a caused retention of intronic sequences that disrupt the coding sequence of each gene (new Figure 4—figure supplement 10). A full description can be found in the corresponding text and figure legend.

Reviewer #1:

[…] As pointed out by the authors, prostaglandins only recently became recognized as developmentally import fate determinants. In this context interference with canonical wnt-signaling had been identified as one of the most critical prostaglandin activities. Considering the importance of Wnt-signaling in kidney formation and patterning it is surprising that this connection had not been further analyzed.

We agree that it would be interesting to delineate the relationship between prostaglandin signaling and canonical Wnt-signaling in relation to kidney development. However, we agree with the collective decision of the reviewers that this interrogation is beyond the scope of our present study.

The word 'vertebrate' in the title is not justified as the zebrafish is the only vertebrate that had been analyzed. Slattery et al. (2016) recently suggested a role of COX2 not only in postnatal but also in embryonic metanephric kidney development. While these mouse data support a possibly conserved requirement for prostaglandins, it should be noted that the phenotypic analyses of COX2 mutants was restricted to the glomerulus and no tubule data were provided.

We have changed the title to "Prostaglandin signaling regulates nephron segment patterning of renal progenitors during zebrafish kidney development" to more specifically describe the context of the study.

The authors suggest a direct connection between the proposed nephron specific expression of ptger2a/4a (Figure 4—figure supplement 4) and the nephric loss and gain of function phenotypes. Expression of prostaglandin pathways components is not well documented and the experimental settings leave space for various alternative explanations. The images shown in Figure 4—figure supplement 4 suggest a restricted expression of ptger2a/4a in proximal tubule of 12-24ss embryos. In case of a direct function, pharmacological treatments starting at 12ss rather than at 4hpf should be sufficient to induce relevant gain and loss of function phenotypes.

Pharmacological treatments have been performed between the 12 ss to 24 hpf for dmPgE2, indomethacin, and the Ptger2 antagonists. Treatment with either indomethacin or the Ptger antagonists between 12 ss and 24 hpf was sufficient to cause an expanded DE and reduced DL, as seen in treatments from 4 hpf to 24 hpf. dmPgE2 treatment between 12 ss and 24 hpf cause an equivalent expansion of the PST and a restriction of the DL as treatment between the 4 hpf to 24 hpf time. These findings are consistent with the notion that Ptger2a/4a act in zebrafish embryo renal progenitors between the 12 ss and 24 hpf to influence segment patterning. Descriptions of these studies have been added to the revised Results section, and the new data are located in new Figure 5 and new Figure 5—figure supplement 1.

Additional data should be provided to confirm nephron specificity and to give details on the posterior-anterior extent of ptger2a/4a expression (for example sections and co-stains with nephric markers). Further, studies should be performed to determine the critical time-window of prostaglandin responsiveness.

We revised Figure 4—figure supplement 4 to provide details of the anterior-posterior domain of ptger2a and ptger4a expression in the intermediate mesoderm with respect to the somites. Transcripts for each gene are initially detected in putative renal progenitors situated adjacent to somites 5 through 10 at the 12 ss stage, and as somitogenesis progresses they are both detected in a domain situated adjacent to somites 5 through 13 until approximately the 24 ss.

We did perform exhaustive testing of various ptger2a/4a probes for fluorescent WISH detection but were not able to determine the appropriate conditions for successful fluorescent labeling, which would have facilitated the most precise co-localization studies and confocal imaging. However, using standard WISH we were able to validate that ptger2a and ptger4a transcripts are expressed in the stripes of pronephros progenitors within the intermediate mesoderm due to their colocalization with cadherin-17 (cdh17) transcripts (Figure 4—figure supplement 4C).

Reviewer #2:

[…] 1) The authors over-simplify prostaglandin synthesis and signaling pathways, applying terminology and biological function relevant to PGE2 under the broad headline of "prostaglandins". This occurs as early as Figure 1, where they identify PGD2, PGA2, PGB2, and PGJ2 in the screen, which are known to stimulate thromboxane receptors (PGA2, B2) and PPARg (PGJ2), yet follow PGE2 (not a hit) and its relevant machinery. Similarly, Ptges is the enzyme for prostaglandin E2 synthesis (hence the "e" in its name, not general secondary processing), and Ptgers are the receptors for PGE2, not all of the PGs. While PTGS inhibition (Cox enzymes), could indeed broadly impact PG synthesis, the rest is quite specific to PGE2 and should not be generalized in the text, particularly when the screen hit several other potential modifiers as relevant to kidney biogenesis. This is not to say the data for PGE2 modification is wrong, it is just over generalized to "prostaglandins" and as such could cause issues for future investigations.

We provided the simplified prostaglandin synthesis and signaling pathway in Figure 1, focusing on PGE2 relevant molecules, as an example to help orient readers. The accompanying text for this section was partly incorrect due to poor editing, and has now been revised for accuracy.

In the revision, we have endeavored to be more specific throughout. For example, the Abstract has been modified to articulate more precisely the manipulations and phenotypes. The Introduction has been edited to provide a better description of prostanoid production and to clearly designate that Figure 1E is an example of PGE2. Throughout the results, we revised our language to be specific when discussing PGE2, particular enzymes and receptors.

2) Given the independent signaling cascades associated with PGB2 and PGE2, it is surprising that exposure to both enzymes elicited the same biological effect.

We agree. However, it has been shown that different prostaglandins after certain thresholds of concentrations will freely interact with other prostaglandin receptors (Kiriyama et al., 1997, Tootle, 2013). This may explain why other prostaglandin compounds, such as PGB2, were capable of eliciting a similar effect as PGE2. One interpretation of the dosage-sensitive segment phenotypes is that they are related to this phenomenon. However, further studies are needed to truly distinguish specific effects of particular prostaglandins, such as by testing loss of function of particular prostaglandin synthases and the appropriate prostaglandin receptors during pronephros development.

While receptor modulation is examined for 2 of the 4 PGE2 receptors, similar studies should be done in the context of exogenous PGE2 and PGB2 addition, as well as with the thromboxane receptors to confirm that addition of exogenous levels of each prostanoid doesn't cause errant signaling.

We performed the experiments to test how addition of PGE2 affected the nephron phenotype in ptger2a and ptger4a deficient zebrafish (new Figure 4—figure supplement 8). Addition of PGE2 did not alter either the phenotype of ptger2a or ptger4a knockdowns, where we observed similar DE expansions and DL reductions as DMSO treated controls (new Figure 4—figure supplement 8A, B). This was confirmed by absolute segment length measurements and statistical analysis (new Figure 4—figure supplement 8C). While examination of thromboxane receptors would be another interesting area to explore, we believe it is beyond the scope of the present work.

3) The doses utilized for dmPGE2 and Indomethacin are substantially higher than that found in other zebrafish or mouse papers, and well above physiological concentrations in humans. While the embryos appear grossly normal in whole mount images, it is important to document that alterations in kidney associated expression patterns are not simply due to gross morphological development issues or off target effects. Assessment of vascular markers by in situ or with a reporter line (Grosser et al) should be a quick way to test toxicity; receptor blocking analysis should confirm specificity.

We assessed surrounding tissues following PG agonist (dmPGE2, 100 µM) and PG antagonist (indomethacin, 30 µM) exposure by whole mount in situ hybridization, and also examined vascular integrity at later stages. We did not observe differences flk1 or gata1 transcript expression at 24 hpf in embryos exposed to DMSO, dmPGE2 or indomethacin between 4 hpf and 24 hpf, indicating normal vascular and primitive blood development, respectively (new Figure 2—figure supplement 4A, B). This is consistent with a published study that reported effects of indomethacin at 40 µM, but normal development of the vasculature following indomethacin treatment at 25 µM (Cha, et al., 2005). We also performed o-dianisidine (benzidene) staining in embryos treated with DMSO, dmPGE2 or indomethacin between 4 hpf and 24 hpf, which labels hemoglobinized erythrocytes and provides a sensitive assessment of defects in circulation or vascular integrity that can be undetected when using only live imaging with stereomicroscopy. With o-dianisidine staining, we observed that the drug treated embryos had equivalent vascular integrity compared to wild-type control embryos through the 48-55 hpf stage, such that we did not observe blockages or hematomas (e.g. bleeding, blood pooling) (new Figure 2—figure supplement 4C).

4) While the overall conclusions seem well supported by the data shown, to make the paper more accessible to a general audience, the authors should use a more precise way to quantify the alterations in expression observed, or at least confirm the relevant ones, using ImageJ and/or a reporter line. The assays are currently very observational (with straight bars correlated with expression drawn on curvy embryos) and it is unclear how definitive statements like "50% reduction" are made when no quantification is shown.

To make the paper more accessible, we now provide absolute segment length values in addition to our evaluation of segment domains relative to the somite number. We measured segment lengths in microns for the pertinent segments that change in drug treatments, morphants, and mutants using Nikon Software.

Reviewer #3:

[…]Whereas the proposed conclusions are significant and would be of interest to the developmental biology and renal research communities, they are not sufficiently supported by the presented data. Addressing the following questions and concerns would significantly strengthen the manuscript and make it suitable for publication.

One of the major concerns about the current manuscript is that there are many quantifications presented but without sufficient experimental detail to evaluate their statistical or biological significance. For several experiments (e.g. treatment of antagonists), there is insufficient experimental detail to evaluate the results and their interpretation. The quantifications of the phenotypes (reduction or expansion) of segments in various experimental regimens is given throughout the manuscript as% of embryos with altered expression (expanded or reduced). This shows that the fraction of affected embryos increases in a dose dependent manner but does not address whether the expressivity (degree of reduction or expansion) of the phenotype is also dose dependent. This is important given the overall conclusion that prostaglandins regulate segment patterning.

Within Figure 2—figure supplement 2 – different colored bars represent the degree of expansion or reduction as indicated by the legend and suggest that indeed, as prostaglandin concentration is increased there is an expansion of the PST up to 6 somites in domain length from that of the WT state of 3 corresponding somites. Additionally, we performed measurements of domain lengths of segments throughout the revised manuscript, which we believe provides the most clear description of the segment phenotypes.

Were the measurements carried in a blinded fashion?

The number of embryos in individual experiments showing% of phenotype should be provided.

All experiments showing% phenotype consisted of a minimum of at least 15 embryos per group in triplicate, which we have notated in the figure legends for the most complex data sets.

For Indomethacin treated embryos, the authors conclude that they "developed a 50% larger DE segment domain and a 20% smaller DL segment domain" referring to Figure 3 and its supplement. It is not clear where the data are presented on which this conclusion has been reached given that Figure 3 shows fractions of affected embryos and not dimensions of pronephric segments in control and treated embryos. Such conclusions need to be supported by clearly presented morphometric data, with the method of measurement clearly described, as well as the numbers of analyzed embryos.

In the revision we now provide absolute length measurements for all nephron segment phenotypes (Please see Figure 2B; Figure 2—figure supplement 2E, F; Figure 3B, C; Figure 3—figure supplement 1E, F; Figure 4—figure supplement 9). Data involving phenotype penetrance and dosage responses have been moved to supplemental figures. We provide n values in either the graph or legend for all quantitations.

The authors conclude that the changes in the dimensions of the pronephros proximo-distal segments represent the specific effects of prostaglandin signaling. However, the possibility of broader effects of prostaglandin signaling on embryonic dimensions and thus indirectly on pronephros is not sufficiently addressed. In the Methods the authors state "Gene domains were measured with respect to somite boundaries to assess pattern formation". However, the somitic staining that should provide such a reference is not clear in most of the figures. Given that earlier reports implicated prostaglandin signaling in morphogenetic movements of gastrulation (Gasser et al., PNAS, 2002; Cha et al., Genes & Dev, 2006; Speirs et al., Development, 2011), this needs to be carefully analyzed. Anteroposterior dimensions of the treated embryos should be measured as well as the relative size of pronephric segments to the neighboring somites for a reviewer/reader to evaluate.

At the dosages for chemical treatments and various knockdowns, we did not observe noticeable morphological changes. To actually document this and provide quantification as well, we performed measurements to acquire absolute values in microns using Nikon Software for the pronephric domain (where the PCT begins and the DL ends), as well as for the anterior posterior axis of the embryos which we have denoted as "tip to tail" for dmPgE2, indomethacin, and 1% DMSO treated embryos. In both cases, we found no statistical differences between the groups (new Figure 2—figure supplement 3). These data support the conclusion that the segment domains changes are not simply a product of embryo dimension changes, but rather a modulation of segment specific identities.

In experiments in which PG signaling was inhibited with various agents, it is not clear at what developmental stages was this done? Given the previous work implicating PG signaling in gastrulation movements the timing of the treatment is important. It would be also of interest to identify when during embryogenesis PG signaling is required for segment patterning and when it can affect it.

We have revised the manuscript to ensure that the time window for chemical treatments was made more apparent in the results rather than having this stated in the methods. Next, we agreed that the perturbation of the prostaglandin pathway at later time windows was critical to assess. To address this, we treated embryos with dmPgE2, indomethacin, and 1% DMSO between the 12 ss to 24 hpf. In all cases, we observed that these treatments phenocopied their corresponding treatments from 4 hpf to 24 hpf (new Figure 5). Additionally, we treated embryos with the ptger2a antagonists (AH6809 and PF04418948) from the 12ss to 24hpf and saw a similar increase of the DE and decrease of the DL as seen in treatments from 4hpf to 24hpf (new Figure 5—figure supplement 1).

Injection into 1 cell stage embryos of translation and splicing morpholinos targeting ptgs1 (cox1) and ptgs2a (cox2) enzymes is reported to affect pronephros patterning. However, it has been previously reported that ptgs1 morpholino injections caused severe epiboly defects (Gasser et al., 2002). In this work the morpholino experiments are insufficiently described. There is no information on the dosage of morpholinos per embryos (only stock concentration is given in methods), no dose response, no experiments describing the degree of loss of function. Yet, the authors conclude "these data indicate that endogenous levels of prostaglandins are critical for specification of the DE and DL segments". This conclusion is given before dmPGE2 rescue experiments are described.

Similar lack of detail and controls is a concern for the experiments in which two genes encoding the prostaglandin receptors, ptger2a and ptger4a are targeted using morpholinos. There are no data on the effectiveness of these reagents to downregulate expression of these genes and encoded proteins, or to lower the levels of prostaglandins. Since in recent years, significant concerns have been raised about the specificity of morpholinos (e.g. Kok et al., Dev Cell, 2015), the sparse details and controls for the experiments presented raise concerns.

We have modified our Methods and figures/figure legends to more clearly describe our morpholino related experiments. While most of morpholinos described in the present work have been used in a number of other published studies (Grosser et al., 2002, Cha et al., 2006, North et al., 2008, Nissim et al., 2014), we do recognize the need for morpholino work to be verified. Indeed, part of our rationale for extensive parallel studies with chemical genetics and use of multiple knockdown reagents was to study these factors with independent tools. Unfortunately, we are unable at the present time to assess the consequential protein levels of the genes targeted by morpholinos that interact with 5’UTR or translational start site due to the lack of appropriate specific antibodies. However, in the revision, we provide a thorough transcriptional analysis to document the effects of splice morpholinos used for knockdown of ptgs1, ptgs2a, ptger2a, and ptger4a using RT-PCR (new Figure 4—figure supplement 10). In all cases, we were able to confirm that incorrect splicing occurred as a product of each splice morpholino or combination (new Figure 4—figure supplement 10). In fact, for each gene, we found that the morpholino microinjection eliminated the presence of wild-type transcripts. In the case of ptger2a and ptger4a, we needed to obtained splice targeting morpholinos and we have validated their phenotypic effects on the pronephros (new Figure 4—figure supplement 10, E, F).

In epistasis experiments presented in Figure 6A, B it would be important to see the effect of SC-580 alone.

We provide data on the effect of the selective Cox1 inhibitor in Figure 4, with the absolute segment length changes documented in Figure 4—figure supplement 9. In Figure 6A we provide a representative image to remind readers of the DE phenotype from exposure to this compound.

From the perspective of the model of PG – irx3b pathway proposed in Figure 7A, the effect of dmPGE2 treatment that affect PST and DL but do not affect DE are confusing.

The working models depict the effects of PGE2 on segment size, where (A) PGE2 biosynthesis is required to restrict the DE by inhibiting the domain of irx3b expression and to promote the DL, and (B) PGE2 applied exogenous is capable of positively modulating the size PST segment by promoting sim1a expression, and meanwhile inhibiting the DL. In the case of (B), the DE segment develops at a normal size but the position is altered along the anterior-posterior axis, likely due to the PST expansion and the shifted domain of irx3b expression, and thus we thought that drawing an arrow between PGE2 and the DE in this case would be misleading compared to the other arrows and phenotypes because the DE segment size in this case is normal. We edited the figure legend to remind readers that we correlated a change in the irx3b domain to the shift in the DE location, and that changes in the mecom expression domain correlate with the DL reduction in both (A) and (B).

[Editors' note: further revisions were requested prior to acceptance, as described below.]

[…] The manuscript has been substantially improved and all reviewers' comments have been addressed well. The additional experimental data help to strengthen and support the findings, and the writing is now clear and precise. Overall, this is a very interesting study and should be of wide interest.

One concern remains: the quantification has been improved but there was very little information about the statistical tests used for analysis of the data. t-tests are mentioned in the Materials and methods section, but these will not be suitable for analysis of some of the datasets, where there are multiple conditions and comparisons. Here, a test such as ANOVA, with appropriate post-test correction for multiple samples, would be preferred. Alternatively, the authors should clarify and justify the tests they have used. The tests that are used should be stated in each of the figure legends where appropriate, in addition to the statement in the Materials and methods.

We have revised the Materials and methods as requested, and we have edited each figure legend to state which statistical analysis was used.

With regard to our statistical tests: In many figures, we utilized unpaired t-tests to compare nephron segment lengths between two samples. However, in cases of 3 or more samples, we now report the outcome of ANOVA tests rather than t-tests, as ANOVA is preferred for comparisons across multiple samples. We have now revised the following figures with ANOVA results: Figure 2—figure supplement 3A, B; Figure 4—figure supplement 8C; Figure 4—figure supplement 10F; Figure 5C; Figure 5—figure supplement 1B, C; and Figure 6B. These revisions provide a more appropriate statistical analysis for numerous aspects of the work.

In contrast, however, in the graphs reporting the percentage phenotype following drug or morpholino treatments, we focused on the key comparison of each individual phenotype to the corresponding wild-type control. For this, we utilized unpaired t-tests, and have revised the figure legends to clearly indicate our analysis. While it might be interesting to look across groups in some cases, our primary goal of these statistical analyses was to determine whether PGE2 treatment or alterations of endogenous prostaglandin biosynthesis significantly altered nephron segments compared to wild-type control embryos, as opposed to delving into the cross-treatment comparisons.

https://doi.org/10.7554/eLife.17551.031

Article and author information

Author details

  1. Shahram Jevin Poureetezadi

    1. Department of Biological Sciences, University of Notre Dame, Notre Dame, United States
    2. Center for Stem Cells and Regenerative Medicine, Center for Zebrafish Research, University of Notre Dame, Notre Dame, United States
    Contribution
    SJP, Conception and design, Acquisition of data, Analysis and interpretation of data, Drafting or revising the article
    Competing interests
    The authors declare that no competing interests exist.
  2. Christina N Cheng

    1. Department of Biological Sciences, University of Notre Dame, Notre Dame, United States
    2. Center for Stem Cells and Regenerative Medicine, Center for Zebrafish Research, University of Notre Dame, Notre Dame, United States
    Contribution
    CNC, Conception and design, Acquisition of data, Analysis and interpretation of data, Drafting or revising the article
    Competing interests
    The authors declare that no competing interests exist.
  3. Joseph M Chambers

    1. Department of Biological Sciences, University of Notre Dame, Notre Dame, United States
    2. Center for Stem Cells and Regenerative Medicine, Center for Zebrafish Research, University of Notre Dame, Notre Dame, United States
    Contribution
    JMC, Acquisition of data, Analysis and interpretation of data
    Competing interests
    The authors declare that no competing interests exist.
  4. Bridgette E Drummond

    1. Department of Biological Sciences, University of Notre Dame, Notre Dame, United States
    2. Center for Stem Cells and Regenerative Medicine, Center for Zebrafish Research, University of Notre Dame, Notre Dame, United States
    Contribution
    BED, Acquisition of data, Analysis and interpretation of data
    Competing interests
    The authors declare that no competing interests exist.
  5. Rebecca A Wingert

    1. Department of Biological Sciences, University of Notre Dame, Notre Dame, United States
    2. Center for Stem Cells and Regenerative Medicine, Center for Zebrafish Research, University of Notre Dame, Notre Dame, United States
    Contribution
    RAW, Conception and design, Acquisition of data, Analysis and interpretation of data, Drafting or revising the article
    For correspondence
    rwingert@nd.edu
    Competing interests
    The authors declare that no competing interests exist.
    ORCID icon 0000-0003-3133-7549

Funding

National Science Foundation (Graduate Research Fellowship - DGE-1313583)

  • Bridgette E Drummond

National Institutes of Health (R01DK100237)

  • Rebecca A Wingert

The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication.

Acknowledgements

NIH Grant R01DK100237 to RAW supported this work. BED was also supported by a National Science Foundation Graduate Research Fellowship DGE-1313583. We are grateful to Elizabeth and Michael Gallagher for a generous gift to the University of Notre Dame on behalf of their family for the support of stem cell research. The funders had no role in the study design, data collection and analysis, decision to publish, or manuscript preparation. We thank the staffs of the Department of Biological Sciences and the Center for Zebrafish Research at the University of Notre Dame for their dedication and care of our zebrafish aquarium. Finally, we thank the members of our lab for their support, discussions, and insights about this work.

Ethics

Animal experimentation: Zebrafish were cared for and maintained in the Center for Zebrafish Research at the University of Notre Dame using experiments approved under protocol 16-025.

Reviewing Editor

  1. Tanya T Whitfield, Reviewing Editor, University of Sheffield, United Kingdom

Publication history

  1. Received: June 9, 2016
  2. Accepted: December 1, 2016
  3. Version of Record published: December 20, 2016 (version 1)

Copyright

© 2016, Poureetezadi et al.

This article is distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use and redistribution provided that the original author and source are credited.

Metrics

  • 604
    Page views
  • 143
    Downloads
  • 1
    Citations

Article citation count generated by polling the highest count across the following sources: PubMed Central, Scopus, Crossref.

Comments

Download links

A two-part list of links to download the article, or parts of the article, in various formats.

Downloads (link to download the article as PDF)

Download citations (links to download the citations from this article in formats compatible with various reference manager tools)

Open citations (links to open the citations from this article in various online reference manager services)

Further reading

    1. Neuroscience
    Lin Cong et al.
    Tools and Resources