FoxO suppresses endoplasmic reticulum stress to inhibit growth of Tsc1-deficient tissues under nutrient restriction

Reviewer #1:

[…] However, I have a few criticisms that should be addressed:

1) The text repeatedly refers genes mis-regulated in the RNA-seq analysis as FoxO targets. However, there is no FoxO binding analysis to prove such an assertion and the text should be modified accordingly.

The reviewer is right. We have modified the text to correctly refer to those genes as genes downregulated in Tsc1 mutant cells upon foxo knockdown.

2) Several of the genes that are classified as being part of the ER stress response have not been characterized in flies (e.g., CG6766). The authors should verify that RNAi-depletion of these candidate genes induce ER stress using either the Cnx99a antibody or by examining expression of spliced Xbp1 transcripts.

We investigated the induction of ER stress response in cells mutant for the top candidate CG6766. We used a mutant allele carrying a piggyBac insertion with a splice-acceptor site and stop codons in all frames in the first intron of the gene (Schuldiner et al., 2008; line 140290 from Kyoto Drosophila Genetic Resource Center). Calnexin staining of homozygous mutant clones showed similar staining pattern compared to the surrounding wild-type tissue and the isogenic control clones (Figure 2—figure supplement 2A). We then examined the expression levels of spliced-Xbp1 and Hsc-70-3 to observe the effect on the upregulation of ER stress response. Control or CG6766 knockdowns were generated in wild-type, weak Tsc1 knockdown and strong Tsc1 knockdown backgrounds. qRT-PCR analysis on eye imaginal discs revealed no change in the expression levels of spliced-Xbp1 or Hsc-70-3 in the wild-type background upon knockdown of CG6766 (Figure 2—figure supplement 2B and B'). By contrast, the expression levels of both ER stress reporters were increased in the Tsc1 knockdown backgrounds, indicating that CG6766 is limiting to maintain ER homeostasis in Tsc1 mutant cells.

The lack of an ER stress response in cells mutant for CG6766 suggests a (partially) redundant function. This may also explain why CG6766 has not been characterized so far.

3) In paragraph two of the Results you speculate about the reason why snoRNAs are up-regulated in the RNAseq data. Since there are no experiments examining why snoRNAs are mis-regulated in the dataset, most of this text should be moved to the Discussion.

We agree with the reviewer. The text relating to the upregulation of snoRNAs has been moved to the Discussion.

Reviewer #2:

[…]

1) Results paragraph one: It's strange that the RNA-seq showed no FoxO knockdown. The conclusion is that the lack of detection of FoxO knockdown is due to biological variability. And the authors show an example of partial mRNA knockdown achieving strong protein knockdown in Kc167 cells. However, the eye disc data show no knockdown at all. Is it possible that the LCM technique may not be accurately just isolating the knockdown clones? Based on the schematic in Figure 1, can the authors use presence/absence of GFP transcripts to clearly delineate that have isolated purely the clonal cells?

We were also puzzled by the unaltered levels of foxo mRNA in the foxo RNAi samples. In addition to the biological variability, we suspect the presence of a transcriptionally-activating feedback upon foxo knockdown as mentioned in the manuscript. An effect on translational capacity could also explain the loss of FoxO protein (Figure 1B and 1B’) with no effect on the transcript levels (Figure 1 —figure supplement 1B).

The reviewer is concerned that contamination with non-clonal tissue could account for the high foxo mRNA levels. We don’t think that contamination is a major issue in our experiments. The most straightforward test for a contamination of the clonal tissues would be to measure the GFP expression levels (as suggested by the reviewer). Unfortunately, a direct comparison is not possible because, in order to maximize the number of biological replicates, we did not sequence GFP-positive clones. However, we compared the GFP expression levels in cells with and without foxo knockdown. As an example, the expression level in Tsc1 mutant clones isolated from normal feeding conditions is shown in Author response image 1. As no major differences were observed, the contamination from GFP-positive cells, if any, is similar in all conditions. Additionally, representative images of isolated Tsc1 mutant clones with and without foxo knockdown are presented to support that the isolated cell populations were negative for GFP (Figure 1—figure supplement 2A). This is further substantiated by the presence of 100% mutant transcripts (Tsc1^Q87X where CAG at position 3R:24,130,984 is substituted by TAG) in the Tsc1 mutant clones (Figure 1—figure supplement 2B). Taken together, it is very unlikely that the isolated clonal samples were contaminated with surrounding non-clonal cells.

Author response image 1

Download asset Open asset

2) The discovery that regulation of ER stress and ERAD genes may explain the effects of FoxO is interesting. These finding would certainly be in line with the known links between Tsc1, FoxO and the control of ER stress. One thing I thought was missing was a little more investigation and explanation of how the ER stress genes are interacting with TSC1 to control cell proliferation – the data in Figures 2, 3 do show changes in eye size, but is it clear that these effects are on cell proliferation (vs growth, death)? And if they are, what might be the mechanism? As it stands the paper reads as an initial (but still very nice) functional validation of RNA-seq data rather than an in-depth mechanistic investigation into how FoxO and Tsc1 function to control tissue growth.

To address if the Tsc1 mutant tissue overgrowth upon induction of ER stress (as depicted in Figures 2 and 3) is due to increased proliferation or cell size, we analyzed the density of ommatidia in adult eyes. Since the loss of Tsc1 does not disrupt ommatidial architecture (Nowak et al., 2018), the density of ommatidia reflects the cellular density in the adult eyes. This density remains unchanged between Tsc1 and Tsc1 foxo knockdown eyes but is slightly reduced upon knockdown of CG6766 and CG8860 (Author response image 2), indicating that at least a part of the enhanced overgrowth of the adult eyes is due to an increase in cell size. However, the small reduction in density (10% for CG6766 and 7% for CG8860, respectively) cannot fully account for the observed overgrowth (Figure 2B, 23% average increase for CG6766 knockdown and 11% average increase for CG8860 knockdown). We therefore conclude that the overgrowth is in part caused by more cells.

Author response image 2

Download asset Open asset

In the Discussion, we speculate about the mechanisms of how ER stress response pathways can regulate proliferation. These are reinforced by the results presented in Figure 4.

We agree with the reviewer that our manuscript represents an initial analysis to describe a link between FoxO and ER stress response. We are confident that it will serve as a strong foundation for future studies aiming at elucidating the exact mechanism of regulation of ER stress by FoxO and the effects on growth.

Reviewer #3:

[…]

1) Title could be shorter. No need to mention TSC.

We are aware that the title is rather long and complicated. However, the identification and analysis of differentially regulated genes were based on clonal populations of Tsc1 mutant cells. The manuscript does not address a generalized function of FoxO in regulating ER stress.

In our opinion, it is more accurate to mention in the title that our findings are specific to Tsc1-deficient tissues.

2) Introduction ought to more carefully and fully explain the TOR negative feedback loop on Akt that means full FOXO activation upon loss of TSC.

The text in the Introduction has been revised to explicitly state the negative feedback to insulin signaling in hyperactive mTORC1 cells and to describe the nuclear localization and activation of FoxO in Tsc1 mutant cells.