NHR-14 loss of function couples intestinal iron uptake with innate immunity in C. elegans through PQM-1 signaling

[Editors’ note: a previous version of this study was rejected after peer review, but the authors submitted for reconsideration. The first decision letter after peer review is shown below.]

Thank you for submitting your work entitled "NHR-14 integrates iron uptake and innate immunity in Caenorhabditis elegans through PQM-1 signaling" for consideration by eLife. Your article has been reviewed by three peer reviewers, and the evaluation has been overseen by a Reviewing Editor and a Senior Editor.

Our decision has been reached after consultation between the reviewers. Based on these discussions and the individual reviews below, we regret to inform you that your work will not be considered further for publication in eLife.

All the reviewers found the manuscript to be very interesting, but the lack of rigor or in-depth analysis in many areas makes it unsuitable for publication in its current form. The manuscript could be improved by going more in depth on the metal biology or by strengthening the innate immunity aspects. The diversity of the reviewer's suggestions indicate that the revision will likely take more than two months, the time limit in eLife for revision. This is why the paper was rejected. Nevertheless, the authors could consider re-submitting their article to eLife as a new submission once the reviewer's comments have been addressed in full.

Reviewer #1:

The authors had previously discovered a role for the C. elegans HIF1a homolog in regulating intestinal iron homeostasis during iron deficiency by activating and inhibiting genes involved in iron uptake and storage. Taking advantage of a developmental delay phenotype of hif-1(ia4) mutants on low iron media, the authors performed a suppressor screen and uncovered two mutations in a nuclear hormone receptor gene, nhr-14, that rescued the delay. NHR-14 transcript and protein levels were unaffected by iron levels. RNA-seq anaylsis comparing gene expression profiles between wildtype animals and nhr-14 mutants revealed that nhr-14 regulates the iron transport gene smf-3, and innate immunity genes. Analysis of the smf-3 promoter and mining modENCODE ChIP-seq data suggested that PQM-1 acts downstream of NHR-14 to regulate smf-3. The authors explored the role of NHR-14 in innate immunity, demonstrating that in addition to regulating immune genes, nhr-14 mutants were resistant to P. aeruginosa and S. enterica infection, and that bacterial infection downregulated NHR-14 protein levels. Finally, the authors showed that knockdown of the nhr-14 ortholog, HNF4a, in human cells resulted in a similar upregulation of the smf-3 homolog (DMT1). This is an interesting study looking at how nuclear hormone receptors coordinate iron availability and innate immunity, but the manuscript suffers from a lack of details in the genetic analyses, and a lack of exploration of the roles of NHR-14 in innate immunity required to show that the gene expression signature is not simply a consequence of disrupting a core cellular process.

Substantive concerns

1) Elements of the manuscript were difficult to evaluate, as many details about the screen and the genetics were lacking. For the screen, how many genomes were screened? How many mutants were recovered? Was the screen saturating? As this was the first report of a screen, these details should be provided. Additionally, were the suppressors, and the nhr-14(tm1473) pqm-1(ok485) mutants backcrossed? If so, how many times?

2) There is a complete lack of characterization of the nhr-14 mutant strains that hampers evaluation of many elements of the manuscript. Do these strains have normal broodsizes? Do they develop at a normal rate? Do they have a normal lifespan? These questions are particularly important for interpreting the pathogen data. What is the nhr-14(tm1473) allele? Is it a null? Are the three nhr-14 mutants phenotypically equal? If so, that would be particularly interesting as the insertion in the DNA binding domain would be predicted to disrupt DNA binding, but not the activation domains or ligand binding domain, and would support that direct gene regulation by NHR-14 contributes to the suppression of the hif-1a arrest on low iron media.

3) The innate immunity sections are problematic. Melo and Ruvkun (2012) demonstrated that RNAi disruption of core cellular processes induces expression of detoxification and innate immune effectors, even in the absence of toxins or pathogens. Similar to point 1, much more information about the nhr-14 strains needs to be provided to convince the reader that this is specifically an immune phenotype and not due to affecting a core process. The authors claim that NHR-14 is the first C. elegans NHR to be involved in innate immunity (and iron metabolism), but Ward et al., 2014, demonstrated that inactivation of nhr-25 causes sensitivity to a range of stresses and pathogens. Several experiments would help strengthen the immunity sections: i) lifespan on OP50 as a baseline for comparing the pathogen sensitivity data; ii) NHR-14 overexpression, which would be predicted to cause increased sensitivity to pathogens. Kirienko et al., 2013, reported a liquid killing assay for PA14 that involves different pathogenesis mechanisms than plate-based assays and is particularly dependent on iron. This assay would be very useful to evaluate the importance of NHR-14-mediated iron regulation in the context of immunity.

4) A major weakness of the paper is lack of any mechanism of how NHR-14 is regulating PQM-1 and in turn SMF-3, or how pathogenic stress regulates NHR-14 levels. An INS-7 overexpression experiment would be very straightforward.

Reviewer #2:

Elucidating mechanisms that organisms use to mediate iron homeostasis through the regulation of transcription factors and transport proteins is an important objective in the field of metal biology. This manuscript details the interaction of two transcription factors that link iron homeostasis with the response of C. elegans to infection by pathogenic bacteria. The authors performed an innovative forward genetic screen leading to the identification of two mutations in the nhr-14 gene. They used rigorous methods to demonstrate that mutations in nhr-14 cause the phenotype. They characterized the phenotype of nhr-14 mutant animals, including the rescue of the hypersensitivity to low iron phenotype of the hif-1 mutant, an increase in metal content for iron and manganese, an increase in the expression of many genes including smf-3 and multiple genes linked to innate immunity, and an alteration in the localization of the transcription factor PQM-1, and resistance to bacterial infections. They also demonstrate that the loss of the human homolog of NHR-14 causes an increase in the expression of another iron transport protein in vertebrate cells, suggesting a conservation of function of the identified nuclear receptor. The impact and quality of the work found within this manuscript would make it a strong candidate for eLife if the authors address the following concerns:

Major Issues:

1) The authors identified mutations in nhr-14 as suppressors of the low iron hypersensitivity of hif-1, which is an innovative discovery technique. They used powerful positional cloning approaches to clearly demonstrate that the mutations in nhr-14 are the cause of the enhanced growth in low iron. The major question that arises is how does mutation of nhr-14 rescue the hif-1 growth defect, and a key approach is to analyze the phenotype of nhr-14 mutations in an otherwise wild type background. This manuscript would benefit from experiments that compare the growth of nhr-14 mutants to wild type animals in two or three concentrations of BP and two or three concentrations of FAC. If the nhr-14 mutations do indeed act by enhancing iron uptake, then it is predicted that they will enhance growth in BP compared to wild-type animals, and they may be hypersensitive to FAC compared to wild type animals. If these effects are not observed, it would call into question the model that the nhr-14 mutation acts by enhancing iron uptake.

2) The authors argue that the cellular site of action of nhr-14 is intestinal cells, as shown in Figure 9. This conclusion appears to be based on the localization of a NHR-14:GFP fusion protein, as shown in Figure 2D. However, the fusion protein is also expressed in other cell types – at a minimum, the red arrows highlight expression in head neurons. It is not possible to judge from the single image how many other cell types may also express NHR-14. Thus, if the conclusion about cellular site of action is based solely on expression data, then an alternative possibility is that the key site of action is the head neurons, or other cell types that may not be visible in this single image. If the authors want to draw a conclusion about the cellular site of action, then they should do experiments to express NHR-14 in a specific cell type using a heterologous promoter, such as an intestinal specific promoter. I do not think determining the cellular site of action is essential for the main conclusions of the paper, so another approach is for the authors to deemphasize the conclusion that the site of action is intestinal cells.

3) The authors investigated the exciting hypothesis that nhr-14 is regulated by dietary iron. However, the results shown in Figure 2 are negative and support the conclusion that both the transcriptional and translational levels of NHR-14 are unaffected by dietary iron. However, a possible unexplored explanation for the function of NHR-14 in iron homeostasis is that the localization of NHR-14 is altered in response to changes in supplemental iron, since some nuclear receptors are regulated by nuclear localization. The functional characterization of NHR-14 would be more complete with an examination of the localization change when animals expressing the tagged version of NHR-14 are exposed to excess and limited iron conditions.

4) The conclusion that nhr-14 acts through regulation of smf-3 to increase the growth of hif-1 mutants should be considered more rigorously. The authors clearly demonstrate that nhr-14 mutants have higher levels of smf-3 transcript, and SMF-3 is a known iron import protein, leading to the hypothesis that this is the critical target gene for the hif-1 rescue. I acknowledge this is a reasonable and even plausible hypothesis. However, the transcriptional analysis of nhr-14 mutant animals demonstrates that hundreds of transcripts are altered, so in principle any one of these changes, or many of these changes in combination might mediate the hif-1 rescue. The critical issue is what additional evidence can test the smf-3 hypothesis. The authors create a smf-3; nhr-14; hif-1 triple mutant and the rescue is lost, in that these triple mutants do not grow in low iron. My concern is that the smf-3 single mutants may not be able to grow in low iron, in which case this experiment is not really informative. For example, if the authors create a triple mutant with nhr-14, hif-1 and any mutation that causes L1 arrest, they will get this result. To interpret this experiment, the authors should analyze the smf-3 single mutant in BP. If the smf-3 single mutant displays arrested growth, then I think there is no clear interpretation of the double mutant of smf-3 with hif-1 and the triple mutant of smf-1, hif-1 and nhr-14. These data would therefore not support the conclusion that NHR-14 functions in iron homeostasis through smf-3.

5) In Figure 3B, there is a significant increase in the amount of manganese in animals with a mutation in nhr-14. This raises the possibility that the change in manganese levels causes any or all of the nhr-14 mutant phenotypes. It would be helpful if the authors addressed this possibility, certainly in the text and possibly experimentally. Have the authors examined the phenotypes of nhr-14 mutants with respect to excess and limited manganese?

6) The model that NHR-14 works in combination with PQM-1 to reduce the iron in the intestinal lumen through SMF-3 is speculative. For example, there are no measurements of the level of iron in the intestinal lumen. Alternative models are that it is changes in iron in the body of the worm that influence pathogen resistance, or that there is a non-iron based mechanism of pathogen resistance. One possible test of the model could be examining the effect of adding supplemental iron to the pathogenicity of the tested bacteria. This would help to test the prediction that decreasing the iron in the lumen is protective to these animals. In a similar vein, the model suggests that nhr-14 mutant animals are resistant because there is an increase in the expression of smf-3 in these animals. What are the levels of smf-3 expression in mutant animals exposed to pathogenic bacteria? This could be included as part of Figure 7D and 7F. These data would also provide support for the current model.

7) Looking at the pathogen resistance data, it was surprising that the authors decided to pursue further phenotype analysis of the nhr-14 mutant using PA14 rather than S. enterica as the pathogenic bacteria tested. The lifespan phenotype associated with S. enterica was much stronger. The data would be more compelling if Figure 7I was performed with S. enterica. Also in Figure 7I, the intermediate phenotype of the double mutant is hard to interpret and does not support the idea that NHR-14 acts upstream of smf-3. If the loss of NHR-14 makes animals resistant through SMF-3, then the double mutant of nhr-14 and smf-3 would be expected to be as hypersensitive as the single mutant of smf-3.

Reviewer #3:

The paper is well written, the methods are very diverse and appropriately used, and the approach is sound and thorough. The topic is trendy, the interpretations and conclusions are fair and the scientific contribution is significant.

The authors describe a new role for the transcription factor NHR-14, attracting attention to the somewhat understudied extended family of 269 C. elegans HNF4-derived transcription factors. They find a key role for NHR-14 in gut infection handling through iron status modulation (intake by the DMT1/NRAMP2 orthologue SMF-3 and sequestration by FTN-1) and transcriptional regulation of a set of innate immunity-associated genes. Notably, they also identify PQM-1 has a downstream effector that shuttles between the cytosol and the nucleus of enterocytes to directly regulates smf-3 expression.

This study hence connects PQM-1, which is gaining recognition as a major transcription factor in the modulation of C. elegans fat handling, iron metabolism and ageing, to innate immunity. Although expected and supported by two other recent studies, it is an important step toward an understanding of C. elegans innate immunity regulation. Moreover, the upstream role of NHR-14 could not have been inferred from the current literature and it contributes significantly to piecing together the complex network that regulates trade-offs between energy handling and immune responses.

There is a great potential for further studies trying to delineate the exact role of NHR-14 in resistance to infection (it seems specific to certain infections), its upstream regulators (there are hints from mammalian studies, notably specific fatty acids) and how it modulates PQM-1 shuttling, and indirectly DAF-16 shuttling. Connections between immunity, fat metabolism, lipid signalling and aging are being noticed more and more and this study contributes to solving the puzzle.

On a more technical note, I liked the fact that the authors made the effort to accumulate corroborative evidence without relying abusively on a single technical approach. The genetic work is also appreciated (double and triple mutants will make excellent tools for the community). I also liked a couple of points of detail that helped convince me of the interpretations. For instance, the qRT-PCR results give fold increases that match very well the fold increases seen by RNAseq. Also, in Figure 7I, nhr-14 seems to be able to increase smf-3 resistance to P. aeruginosa infection, which is consistent with the idea that NHR-14 increases infection resistance through other genes than smf-3 and iron sequestration alone.

General concerns.

A couple of side issues may have been neglected to ensure that the picture presented is not blown out of proportions (see specific concerns).

The raw and analysed data from the RNAseq and lists of overlapping genes between different datasets should be available. If not, the claims of the authors cannot be verified. It is also common practice to make these datasets available upon publication. It would help the peer-review process if at least the analysed data used in the manuscript were made available to the reviewers.

Although the methods are generally sound, some should be expanded upon to allow for replication of the results. Numbers of animals and/or replicates are low for some experiments and would require additional repeats to meet desirable standards. When it comes to biological replicates, their definition is unclear. One would expect that each biological replicate would be a different population on a different day as interday variability is far greater for C. elegans populations compared to intraday variability, but the low numbers might indicate that biological replicates came from the same day. Clarification is required on this issue.

Specific concerns.

First, the title is slightly misleading. Although the authors provided evidence that a NHR-14 – PQM-1 axis does integrate elements of iron metabolism through SMF-3, the case is less strong when it comes to innate immunity. Indeed, no clear immunity pathway has been defined as it is mostly based on observed enrichment in innate immunity-associated genes in RNAseq analysis without hierarchy. Moreover, direct regulation of innate immunity genes by PQM-1 has not been verified (using biochemical assays, qRT-PCR under infection conditions, or transcriptional reporters in pqm-1 mutant, and/or WT animals under infection conditions). "Integrate" is a too strong word as it suggests pivotal role that is not demonstrated.

Second, SMF-1 and SMF-2 are two other DMT1 homologues in C. elegans and SMF-1 is expressed in the gut just as SMF-3 is. Yet there is no comment on SMF-1 potential role here and its possible regulation by nhr-14, which should have been checked and reported.

The specificity of NHR-14 action on PQM-1 is not clearly established especially in the absence of pathway between the two. Several other high-ranking intestinal transcription factors (ELT-2, DAF-16, SKN-1, ATF-7, NHR-49) have been involved in gut immunity in worms and could have been looked at more carefully, especially when transgenic reporters are available to look at nuclear shuttling (SKN-1::GFP and DAF-16::GFP for instance).

The role of pqm-1 in infection resistance is not clearly established, while nhr-14 and smf-3 are.

HIF-1 is mentioned initially because nhr-14 was found in a suppressor screen for the hif-1 growth defect under iron deprivation, but it is not discussed later, nor is it connected to the findings.

Suggested additional experiments.

– Infections assays with P. aeruginosa and S. enterica on pqm-1 mutants are necessary to demonstrate a functional role of PQM-1 in innate immunity (high priority).

– To corroborate this the PQM-1::GFP strain should be exposed to P. aeruginosa and nuclear shuttling should be observed (high priority).

– Make sure there are at least 3 independent experiments (different days) with a minimum of 30 worms per strain for the infection assays. A good idea would be to simply do the additional experiments required with nhr-14, WT and pqm-1 mutants in parallel (high priority).

– C. elegans gut bacterial infections are performed at adult stage because larvae usually do not take up bacteria. It is essential that the PQM-1::GFP nuclear localisation shift seen in L3 larvae on nhr-14 RNAi is replicated in adult worms to convince that PQM-1 can act upon gut bacterial infection to regulate innate immunity genes (high priority).

– qRT-PCR of smf-1, -2 and -3 in nhr-14 and in pqm-1 at least (double mutants nhr-14;pqm-1 welcome) to demonstrate the specificity of SMF-3 regulation (if time and money are an issue, it is not the highest priority experiment)

– qRT-PCR of pqm-1 in nhr-14 mutant vs WT to confirm the RNAseq data (low priority, up to the good will of the authors)

– Look at DAF-16::GFP vs PQM-1 shuttling in control, nhr-14, and pqm-1 RNAi to confirm the antagonistic shuttling between DAF-16 and PQM-1 and the specificity of the effect of nhr-14 on PQM-1 vs DAF-16 or ELT-2 (medium priority, depending on which other experiments are done)

– P. luminescens, E. faecalis, S. marscens, S. aureus infections would have made sense in Figure 7 given that the authors compare their RNAseq data to datasets on worms infected by these pathogens. This seems particularly valid as Table 2 seems to point toward a more specific role of nhr-14 downregulation in resistance against certain pathogens and not others (low priority, up to the good will of the authors).

– Inclusion of the elt-2 microarray dataset published by Block DH et al., 2015 would make sense in Figure 6 and Table 2, considering that the present article did briefly investigate elt-2. (medium priority).

[Editors’ note: what now follows is the decision letter after the authors submitted for further consideration.]

Thank you for submitting your article "NHR-14 loss of function couples intestinal iron uptake with innate immunity in C. elegans through PQM-1 signaling" for consideration by eLife. Your article has been reviewed by three peer reviewers, and the evaluation has been overseen by a Reviewing Editor and Tadatsugu Taniguchi as the Senior Editor. The reviewers have opted to remain anonymous.

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:

Elucidating mechanisms that organisms use to mediate iron homeostasis through the regulation of transcription factors and transport proteins is an important objective in the field of metal biology. This manuscript details the interaction of two transcription factors that link iron homeostasis with the response of C. elegans to infection by pathogenic bacteria. The authors performed an innovative forward genetic screen leading to the identification of two mutations in the nhr-14 gene. They used rigorous methods to demonstrate that mutations in nhr-14 cause the phenotype. They showed NHR-14 is expressed in intestinal cells and other cells in the head and does not appear to be regulated by iron. They characterized the phenotype of nhr-14 mutant animals, including the rescue of the hypersensitivity to low iron phenotype of the hif-1 mutant, an increase in metal content for iron, an increase in the expression of many genes including smf-3 and multiple genes linked to innate immunity, nuclear localization of PQM-1, and resistance to bacterial infections. PQM-1 appears to interact directly with the smf-3 promoter. The findings lead to the proposal that pathogenic infections down regulate nhr-14, leading to increased expression of SMF-3, which imports iron into intestinal cells, thereby depriving the pathogen of iron in the lumen of the intestine and limiting its capacity for proliferation. While largely conclusions are supported by the results, some additional experiments are needed to reinforce the major findings.

Essential revisions:

1) Transcriptome analysis. A main concern is with the analysis presented in Figure 5. Apart from the fact that DAVID has not been updated (https://david.ncifcrf.gov/helps/update.html), it contains far less information than C. elegans-specific tools such as WormExp. Running their list through WormExp reveals several very significant overlaps with other datasets, including nhr-8 mutants that merit at least discussion if not investigation.

Further, the authors write, "Among the 261 downregulated genes, there was enrichment in genes involving cellular organization and body morphogenesis, including sqt-1, noah-1 and sym-1 (Figure 5D and Figure 3—source data 1)".

But what is really striking is the number of genes that normally undergo a sharp drop of expression at the L4 to adult transition, many of them expressed in the epidermis. As can be visualized at https://elegans.mdc-berlin.de/cel_ex.html, this applies, for example, to sqt-1, col-17, rol-8 (the top 3 in Figure 5D), as well as to the heterochronic gene lin-42. There is no comment on this very remarkable pattern, nor the simple experiments that might address the question of how it might be related to the continued nuclear localization of PQM-1 into adulthood.

On the same line, the authors write, "Interestingly, we observed less overlap between nhr-14 upregulated genes with Serratia marscens [marcescens] (Wong et al., 2007). The basis for this difference is unknown, but is likely a result of a highly specialized response to S. marscens [marcescens]". If they had used WormExp, they would have seen a very significant overlap with Serratia-induced genes in independent study (Engelmann et al., 2011). And reading that study, there is even the following, "In the current analysis, the results of our previous study of the response to S. marcescens using oligo-arrays [22] stood out. As this was not the case for the results for the response to 2 other bacterial pathogens, and given the underrepresentation in the S. marcescens data set of 'common response genes' [22], this presumably reflects an experimental difference in the strength of the infection for the samples prepared for analysis using oligo-arrays". In other words, the authors' conclusion is unlikely to be correct. And that makes more sense since the screens that used C. elegans to identify S. marcescens virulence factors highlighted a role for iron (see for example the Abstract of PMID: 12660152, and 25070509).

2) The authors clearly demonstrate that nhr-14 mutants have higher levels of smf-3 transcript, and SMF-3 is a known iron import protein, leading to the hypothesis that this is the critical target gene for the hif-1 rescue. I acknowledge this is a reasonable and even plausible hypothesis. However, the transcriptional analysis of nhr-14 mutant animals demonstrates that hundreds of transcripts are altered, so in principle any one of these changes, or many of these changes in combination might mediate the hif-1 rescue. The critical issue is what additional evidence can test the smf-3 hypothesis. The authors create a smf-3; nhr-14; hif-1 triple mutant and the rescue is lost, in that these triple mutants do not grow in low iron. My concern is that the smf-3 single mutants also do not grow in low iron. So, yes nhr-14 requires smf-3 to rescue hif-1, but this is somewhat different from proving that nhr-14 rescues by increasing smf-3 activity. The authors should acknowledge that because the smf-3 single mutant cannot grow in low iron, there are multiple interpretations for the triple mutant.

3) Additional controls:

3A) Figure 3. Results shown in panels B, C, and D lack essential genotype that has to be tested, namely nhr-14,smf-3 double mutant. This genotype is essential to prove that the phenotypes of nhr-14 mutants are mostly mediated by the elevated level of smf-3.

3B) Figure 5E and D. It is nice to see enrichment in PQM-1 binding motif in the promoters of the upregulated genes, but experimentally do these genes really require PQM-1 for their expression? What will be their level of induction in nhr-14, pqm-1 double mutant? I think this experiment is important to prove that pqm-1 regulates these genes downstream of nhr-14.

3C) Figure 6A. This experiment lacks nhr-14, pqm-1 double mutant. Again, this is essential to prove epistatic relations between the two genes.