ppargc1a controls nephron segmentation during zebrafish embryonic kidney ontogeny

Essential revisions:

Both reviewers are positive about the work, but both suggest revisions, some of which are overlapping.

1) Please use qRT-PCR to quantify the expression changes detailed in the manuscript (a suggestion of both reviewers).

As suggested by both of the reviewers, we have used qRT-PCR to quantify ppargc1a transcripts in response to treatment with bezafibrate, which is provided in revised Figure 2G.

2) Please revise the arguments about epistasis so that they refer to the comparison of double and single mutant phenotypes and not expression analysis in a single mutant, as suggested by reviewer 1.

The arguments about epistasis have been revised so that they refer to comparison of double and single mutant phenotypes.

3) Provide more data on the initial screen (reviewer 2).

The manuscript has been revised to provide more data on the initial screen, which is now located in Figure 1—figure supplement 1.

4) Include a more comprehensive study of cell proliferation and cell death (reviewer 2).

As suggested we performed comprehensive studies of cell proliferation and death at the 20 ss in the PST and DL, which are the two affected segments in ppargc1a deficient embryos, using combined fluorescent in situ hybridization and immunofluorescence techniques. We found that there was no significant difference in proliferation or death in either segment. This is consistent with the conclusion that changes in cellular turnover do not cause the PST and DL segment boundary changes that occur in ppargc1a deficient embryos.

5) Test for cross-repressive interactions by RNA overexpression (reviewer 2).

We performed RNA overexpression studies to assess the effect of sim1a RNA on the domain of ppargc1a expression in the kidney, and vice versa, and to evaluate the coincident development of the PST segment (Figure 5). Overexpression of sim1a was associated with a statistically significant reduction in the ppargc1a domain (Figure 5A, 5D) and an enlarged PST segment (Figure 5C, 5F). Overexpression of ppargc1a led to a statistically significant reduction in the sim1a domain (Figure 5B, 5E) at 20 ss and a shortened PST segment (Figure 5C, 5F) in addition to a decrease in the number of sim1a⁺ cells at 28 ss (Figure 5—figure supplement 2). These results with RNA overexpression provide further support for the conclusion that there are cross-repressive interactions between sim1a and ppargc1a during pronephros ontogeny.

For further information, the full reviews are appended below.

Reviewer #1:

[…]

– The conclusion that bezafibrate treatment affects ppargc1a expression is not compelling and requires further experimental analysis. Currently, expression of ppargc1a is analyzed in the drug treated embryos during late segmentation (28S – the stage is not provided on the Figure 2 or in the text) and reduced expression domain is observed. However, by this stage as shown in Figure 1, ppargc1a expression is limited to the cells that express the distal segment markers. Therefore, that ppargc1a expression domain is reduced could simply reflect that DL is reduced what is the key finding of the manuscript. The authors also note that expression level of ppargc1a RNA appears reduced, but WISH is not a quantitative method. Therefore, it would be important to analyze ppargc1a expression at 15 somite stage when the authors show it is expressed in the entire domain of cells expressing pax2a, or intermediate mesoderm. At this stage qRT PCR should also be performed.

We performed further experimental analyses to assess how bezafibrate treatment affects ppargc1a expression. First, we performed studies to investigate the time window of bezafibrate treatment in order to better define the correlation between activation of PPAR signaling and ppargc1a expression. Bezafibrate was added at the 4, 5, 6, 8, 9, and 10 hpf stage as well as the 5 ss and 10 ss, and embryos were incubated until the 28 ss, at which time they were fixed and ppargc1a expression was assessed by WISH (Figure 2—figure supplement 5). Bezafibrate addition at the 4, 5, 6, 8, or 9 hpf time points led to statistically significant decreases in the ppargc1a domain length compared to control embryos; further the effect of bezafibrate addition at these time points was not statistically different when compared by ANOVA (Figure 2—figure supplement 5A, B). Thus, we next added bezafibrate at the 5 hpf stage and fixed embryos at the 15 ss to analyze ppargc1a expression. Interestingly, we found no difference in ppargc1a expression between bezafibrate treated and control embryos (Figure 2—figure supplement 5C). Finally, we conducted qRT-PCR analysis on pools of embryos treated with bezafibrate or DMSO vehicle control from the 5 hpf time point until the 28 ss. Although this time window of bezafibrate exposure was associated with a reduced absolute domain length of ppargc1a expression in the pronephros (Figure 2E, 2F), the qRT-PCR analysis revealed that the relative mRNA levels were not significantly different between the groups (Figure 2E). Furthermore, we found that ppargc1a RNA overexpression was not sufficient to rescue DL development in bezafibrate treated embryos (Figure 2—figure supplement 5D, 5E). Taken together, these results are consistent with the notion that the ppargc1a expression domain is reduced in bezafibrate treated embryos because the DL is reduced, and not specifically due to the loss of ppargc1a activity. We have revised the manuscript to reflect this conclusion.

Additionally, we have addressed the reviewer’s comment about missing information in the legend to Figure 2, to clarify that the embryonic stage of the representative images is the 28 ss.

– A conclusion is reached that ppargc1a is epistatic to tbx2b based on the observations that in ppargc1a deficient embryos expression of tbx2b is impaired but expression of ppargc1a is not altered in tbx2b deficient embryos. Epistatic relationships are based on combining and comparing contrasting loss or gain of function phenotypes. Whereas the two LOF phenotypes are similar, gain of tbx2b function/expression masks the loss of ppargc1a, supporting the notion that tbx2b is epistatic to ppargc1a, but ppargc1a acts upstream of tbx2b.

We have revised the manuscript to correct the conclusions about epistatic relationships.

– Likewise, that domain of ppargc1a expression in renal progenitors was significantly increased in length, is interpreted that "sim1a was probably epistatic to ppargc1a". Again these are not experiments to test epistasis. The authors do perform such experiment by creating a compound loss of function between these two genes, which have contrasting LOF phenotypes, with PST fates being lost in sim1a deficient embryos and expanded in ppargc1a mutants. And because ppargc1a;sim1a doubly deficient embryos show normal PST segment, it is ppargc1a that is epistatic to sim1a. Even though epistasis is not complete, consistent not only with pparg1a negatively regulating sim1a expression but also vice versa. The conclusions about epistatic relationships need to be revised both in the Results and Discussion sections.

We have revised the manuscript to correct the conclusions about epistatic relationships.

Reviewer #2:

[…]

Subsection “Bioactive small molecule chemical genetic screen reveals that alteration of PPAR signaling leads to changes in embryonic nephron segmentation”: More data from the initial screen concerning 'compounds that are known to alter activity of PPAR signaling' should be provided. These data might shed some light on the somehow contradictory and currently poorly explained observations that PPAR agonist treatment and loss of ppargc1 cause very similar phenotypes.

In the PPAR literature, other laboratories have reported both increases and decreases of ppargc1a after treatment with PPAR agonists and antagonists (Pardo et al., 2011; Liao et al., 2010; Sanoudou et al., 2010; Goto et al., 2017; Wang and Moraes, 2011). This suggests to us that different cells respond differently to PPAR pathway modulation with respect to ppargc1a/PGC1a expression.

We provide additional data from the initial screen in the revised Figure 1—figure supplement 1. Along with bezafibrate, a PPAR alpha agonist, there were 2 PPAR gamma antagonists identified as screen hits, compounds known as BADGE and GW-9662. Both PPAR gamma antagonists led to an expanded DE, and GW-9662 also led to an expanded PCT based on double WISH (for the DE marker slc12a1 and the PCT marker slc20a1a). However, the screen did not directly assess either the PST or DL segments.

Subsection “ppargc1a is necessary for proper formation of proximal and distal segment boundaries”, last paragraph: Benzafibrate treatments restricted to the time window of ppargc1a expression (>8 somite stage) rather than starting at 5hp should be used to better define the correlation between activation of PPAR signaling and ppargc1a expression/function. In case of a direct role of benzafibrate in repressing ppargc1 expression it should also be possible to rescue the phenotype by ppargc1 RNA-injection. Alternatively, the distal shift of ppargc1 expression in benzafibrate treated embryos may be caused by an earlier patterning defect that cannot be rescued by RNA injection.

Please see our response to reviewer #1, comment 1.

“Further, the ppargc1a signal intensity appeared…”: Quantitative analyses such as RT-qPCR should be used to confirm the suggested general reduction.

As noted above, qRT-PCR was performed to measure ppargc1a signal in bezafibrate treated embryos. We have revised the manuscript accordingly.

“…whether either of these changes were associated with regional fluctuations in cell birth or death.”: To answer this relevant and interesting question, it would have been necessary to study proliferation and cell death in relation to the segment/region specific markers trpm7 and slc12a3 rather than using the pan-nephric marker cdh17?

As suggested we performed cell proliferation and death immunofluorescence studies in combination with fluorescent in situ hybridization for the segment-specific markers trpm7 (PST) and slc12a3 (DL), which is now presented in revised Figure 3. Wild-type and ppargc1a deficient embryos had no statistically significant difference in the numbers of proliferating or dying cells in either the PST or the DL.

Subsection “ppargc1a regulates PST boundary formation through a reciprocally antagonistic relationship with the sim1a transcription factor”: The genetic cross-repressive interaction between ppargc1a and sim1a is interesting. However, to better understand the underlying molecular mechanisms it would be important to further explore whether cross-repressive interactions can be seen when using overexpression of ppargc1a and sim1a RNA. In this context the authors should also explain why embryo-wide overexpression of their RNAs appears to cause very restricted phenotypes.

Please see our response for Essential revisions, point #5.

Embryo-wide overexpression of these RNAs may cause restricted phenotypes because only certain tissues are competent to respond to altered levels of these transcripts. It is also possible that embryo-wide effects would be discernible at higher mRNA dosages, but the present work has not exhaustively chronicled the dosages and resultant phenotypes of each RNA through overexpression studies.