Identification of a weight loss-associated causal eQTL in MTIF3 and the effects of MTIF3 deficiency on human adipocyte function

  1. Mi Huang
  2. Daniel Coral
  3. Hamidreza Ardalani
  4. Peter Spegel
  5. Alham Saadat
  6. Melina Claussnitzer
  7. Hindrik Mulder
  8. Paul W Franks
  9. Sebastian Kalamajski

  • Curated by eLife

    eLife logo

    eLife assessment

    The study provides a fundamental framework for linking human genome variation to targetable mechanisms of disease. The authors provide compelling evidence that a strong candidate locus associates with body weight in humans acts through adipocyte MTIF3. Thus, the generalized approaches taken in this study have the potential to inform genetic association studies in general and lay a foundation for future functional genomics studies.

Reviewed by

Read the full article See related articles

Listed in

Log in to save this article

Abstract

Genetic variation at the MTIF3 (Mitochondrial Translational Initiation Factor 3) locus has been robustly associated with obesity in humans, but the functional basis behind this association is not known. Here, we applied luciferase reporter assay to map potential functional variants in the haplotype block tagged by rs1885988 and used CRISPR-Cas9 to edit the potential functional variants to confirm the regulatory effects on MTIF3 expression. We further conducted functional studies on MTIF3-deficient differentiated human white adipocyte cell line (hWAs-iCas9), generated through inducible expression of CRISPR-Cas9 combined with delivery of synthetic MTIF3 -targeting guide RNA. We demonstrate that rs67785913-centered DNA fragment (in LD with rs1885988, r 2 > 0.8) enhances transcription in a luciferase reporter assay, and CRISPR-Cas9-edited rs67785913 CTCT cells show significantly higher MTIF3 expression than rs67785913 CT cells. Perturbed MTIF3 expression led to reduced mitochondrial respiration and endogenous fatty acid oxidation, as well as altered expression of mitochondrial DNA-encoded genes and proteins, and disturbed mitochondrial OXPHOS complex assembly. Furthermore, after glucose restriction, the MTIF3 knockout cells retained more triglycerides than control cells. This study demonstrates an adipocyte function-specific role of MTIF3 , which originates in the maintenance of mitochondrial function, providing potential explanations for why MTIF3 genetic variation at rs67785913 is associated with body corpulence and response to weight loss interventions.

Article activity feed

  1. Version published to 10.7554/elife.84168 on eLife
  2. eLife

    Author Response

    Reviewer #1 (Public Review):

    The article "Identification of a weight loss-associated causal eQTL in MTIF3 and the effects of MTIF3 deficiency on human adipocyte function" explored the functional roles of MTIF3 during adipocyte differentiation. In persons living with obesity, genetic variation at the MTIF3 locus associates with body mass index and responses to weight loss interventions. MTIF3 regulates mitochondrial protein expression and gene knockouts cause cardiomyopathy in mice. This paper provides insight into the impacts of MTIF3 knockout on adipocyte differentiation and the expression effects of the eQTL on MTIF3 levels. The authors implement a CRISPR/Cas9 gene editing approach coupled with an in vitro platform to detect influences of MTIF3 on adipocyte glucose metabolism and gene expression. This method may serve as a platform to explore knockouts in human cell lines, so it may allow the discovery of new gene x environment influences on in vitro outcomes related to differentiation, growth, and metabolism.

    The conclusions of this paper are mostly well supported by data, but some experimental conditions and data analysis needs to be clarified and extended.

    1. The authors use CRISPR/Cas9 to generate the rs1885988 variant in the human white adipocyte cell line and performed a comprehensive validation analysis of gene editing (Figure 1). qPCR analysis showed reduced MTIF3 expression during human adipocyte differentiation (Figure 1E, F). To expand the importance of the rs1885988 variant, the authors should have provided target gene measurements to verify the canonical differentiation profile (e.g., FABP4, ADIPOQ) and help readers understand the overall impact of gene editing at the MTIF3 locus.

    Thank you for your suggestions. As you requested, we have quantified several adipocyte differentiation markers in the allele-edited cells after 12 days of adipogenic differentiation. The data (Figure 1-figure supplement 1) shows no significant difference between cells with the different genotypes. We have added more information about this in lines 100-101, and also in another context in lines 105-116.

    Notably, the intra-group variation of the marker gene expression is large (Figure 1-figure supplement 1), which makes it difficult to clearly state how much the allele editing, as opposed to random variation resulting from single cell cloning, contributes to the differentiation outcome. However, if we also consider MTIF3 knockout cells (that do not need to be single-cell cloned), their differentiation marker expression also appears unaffected (Figure 3-figure supplement 1). Taken together then, it is unlikely the allele editing with the consequent effect on MTIF3 expression affects adipogenic differentiation in our experiments. We mention the absence of effect of MTIF3 knockout on differentiation in the paragraph starting on line 137.

    1. The direct mechanistic influences of MTIF3 on adipocyte function remain unclear. MTIF3 regulates the translation initiation of mitochondrial protein synthesis. Western blots of OXPHOS proteins do not per se underscore supercomplex formation, which is also a process mediated by MTIF3. Blue native gel electrophoresis may prove a better method to establish the effects of MTIF3 loss-of-function on supercomplex formation.

    As suggested, we have run blue native gel electrophoresis to detect the formation of OXPHOS respiration complexes. In the revised manuscript (lines: 158-168 and Figure 4 E,F), we show how MTIF3 knockout indeed interferes with the complex formation, with lower abundance of complexes V/III2+IV1, III2/IV2 and IV1. Additionally, although the blot signal for complex I+III2+IVn is diffuse, it appears higher in scrambled control cells than in MTIF3 knockout cells. Interestingly, complex II content is slightly higher in MTIF3 knockouts, which may result from a compensatory regulation mechanism, as none of the subunits of complex II is encoded by mitochondrial DNA. We also found several faster-migrating (“undefined bands” in the figure) in the MTIF3 knockout samples, although it is hard to determine whether those are single chain proteins, or degradation or mistranslation products. Overall though, the native gel blots show impaired OXPHOS complex assembly in MTIF3 knockout samples.

    In addition, we performed western blots for other mitochondrial proteins, including COX II (subunit of OXPHOS complex IV), ND2 (subunit of OXPHOS complex I), ATP8 (subunit of OXPHOS complex V), and CYTB (subunit of OXPHOS complex III). The data (Figure 4 A,B), show decreased ND2 and COX II, trending decrease of CYTB, and unaffected ATP8 content in MTIF3 knockout adipocytes.

    The methods (paragraph starting at line 479), results (paragraph starting at line 145), and discussion (lines: 261-263, 274-277) were incorporated in the revised manuscript.

    1. Based on the findings, the authors argue that MTIF3 knockout alters the function of adipocytes. However, many of the experiments show fairly small effect sizes (Figure 5A, Figure 6A). How does the MTIF3 knockout explicitly perform functions related to body weight regulation? Gene editing in vivo would have helped to substantiate the authors' conclusions.

    In the paper we are looking at the consequences of MTIF3 deficiency in one cell type, over short time, in vitro. The outcome of body weight regulation, e.g. during weight loss, would result from long-term effects of MTIF3-altered metabolism in more than one tissue. We envisage that small changes in energy metabolism in not only fat, but also in e.g. muscle, would make a substantial difference over time in vivo (this, we cannot capture in in vitro models). We have added this discussion to lines 294-311.

    As for in vivo genomic editing, the alleles of interest are specific to the human genome. Ideally, a genotype-based recall study in humans would be appropriate, but due to time and resource limitation, we are not able to conduct such a study at the moment (although we certainly hope to perform such a study in the future). As for modeling the MTIF3 deficiency in mice – the MTIF3 knockout mice are not viable [1], and certainly other options (e.g. overexpression, tissue-specific knockouts) are possible and tempting to investigate. This, however, would require considerable additional work which we could only perform in a future project.

    1. In several instances, the authors refer to 'feeding' cells with glucose (line 206, line 171). Feeding experiments often imply complex nutrient interventions in animal models and people, which cannot be easily recapitulated in cell culture. The in vitro experiments simply alter levels of glucose and more precise language would state the specific challenges accurately.

    In the revised manuscript, we have substituted “feeding” for exact glucose concentration, or “glucose concentration” where appropriate. (paragraph starting at line 215, and lines 577-578, 597, 873-879)

    Reviewer #2 (Public Review):

    Huang Mi, et al. investigated the role of MTIF3, the mitochondrial translation initiation factor 3, in the function of adipocytes. They first detected the expression of the obesity-related MTIF3 variants based on the GTEx database and found two variants lead to an increase in MTIF3 expression. Then they knockout MTIF3 in differentiated hWAs adipocytes and characterized the mitochondrial function. They found loss of MTIF3 decrease mitochondrial respiration and fatty acid oxidation. They further treated cells with low glucose medium to mimic weight loss intervention and found MTIF3 knockout adipocytes lose fewer triglycerides than control adipocytes. This paper provides new information about MTIF3 in adipocytes and the potential functional role of MTIF3 in mitochondrial function.

    1. The authors provided sufficient data to show those two genetic variants increase MTIF3 expression. Their CRISPR/Cas9 knockin cell line is also convincing. But they didn't show if the genetic variants affect adipogenesis. Adipogenesis is an important process for weight gain and fat deposition. In lines 103-107, the authors mentioned that the "allele-edited cells have some problem in differentiated state, e.g. triglyceride or mitochondrial content", so they used an inducible Cas9 system. However, the issue of differentiated allele-edited cells may be the functional effect of MTIF3 genetic variants, such as interrupting adipogenesis, decreasing triglyceride, or affecting mitochondrial number. The authors should provide that information.

    Thank you for all your suggestions. We think we were not clear regarding this issue. We did not mean that the allele-edited cells have problem in differentiated state, which then definitely could be (as you point out) due to the functional effect of MTIF3 genetic variants. The problem relates to the process of single-cell cloning itself, which inherently introduces random variation. As a consequence, the data on adipogenic differentiation in allele-edited cells has relatively high intra-group variation. We have added more clarifying text in lines 104-116.

    To provide the data on this, per your request, in the revised manuscript we include the results for the rs67785913-edited cells in Figure 1-figure supplement 1. As shown, we observed no differences in the expression of adipogenic markers (ADIPOQ, PPARG, CEBPA, SREBF1 and FABP4) or in mitochondrial content between the two rs67785913 genotypes. Since the intra-group variation is often high, it is hard to conclude how much the rs67785913 eQTL affects the quantified variables. Much of the variation could instead be ascribed to the effects of single cell cloning.

    The cloning per se introduces random variation, but is required to obtain homozygous allele-edited cells. Because of this dilemma, and to clarify how much MTIF3 expression can actually influence adipogenic differentiation, we have, during the revision, also used the hWAs-iCas9 cells to generate MTIF3 knockouts at the preadipocyte stage and then tested their differentiation capacity. As we show in Figure 3-figure supplement 1, we found no apparent differences in adipogenic marker gene expression between scrambled control and MTIF3 knockout cells (we mention that in lines 137-144). Taken together, our results may indicate that the rs67785913 genotype, through affecting MTIF3 expression, is unlikely to regulate adipogenic differentiation.

    1. In Figure 4, the author mentioned that MTIF3 knockout does not affect the expression of adipogenic differentiation markers. They need to provide more evidence to prove their point. Oil-red O staining is a clearer way to quantify adipocyte differentiation in cell culture. In addition, in Fig. 4B western blot, the author should include MTIF3 as a control to show the knockout efficiency. It is not clear the meaning of plus and minus in that panel. The author should also compare the total triglyceride levels in MTIF3 knockout cells and control cells.

    We have now included Oil-red O staining results and total triglyceride levels (Figure 3 F,G), which show no apparent differences between scrambled control and MTIF3 knockout cells (method: lines 427-431; results: lines 137-144). We also added the MTIF3 blots to figure 4A as a control, showing high and consistent MTIF3 knockout efficiency in independent experiments. In the original manuscript, the plus and minus referred to control and knockout, respectively. To clarify that, we have changed the expression to SC and KO in the revised manuscript.

    With regards to Oil-red O vs. quantification of adipogenic markers, we actually prefer the latter method, as it gives more accurate and less variable results than Oil-red O (at least in the cell line we use). We have, however, performed Oil-red O as well to address your question.

    1. MTIF3 is a translation initiation factor in mitochondria and is involved in the protein synthesis of mitochondrial DNA-encoding genes. The authors should check protein levels rather than the mRNA levels of mitochondrial DNA-encoding genes (Fig. 6E). It's interesting to see the increase of mRNA levels of ND1 and ND2, which might be feedback of lower translation. Since ND1 and ND2 are in OXPHOS complex I, the expression levels of complex I in MTIF3 KO cells would be worth checking. Additionally, the author should also check the mitochondria copy number.

    As suggested, we have detected several mitochondrial encoding proteins which are subunits of each mitochondrial OXPHOS complex. As shown in figure 4A, ND2 (subunit of OXPHOS complex I) and COX II (subunit of OXPHOS complex IV) expression were significantly reduced, CYTB (subunit of OXPHOS complex V) expression tended to decrease, and ATP8 expression was not affected in the MTIF3 knockout adipocytes. We also detected the formation of the OXPHOS respiration complex in extracted mitochondrial proteins and found MTIF3 perturbation affect mitochondrial complex assembly. The detailed methods (lines: 479-490), results (lines: 145-169) and discussion (lines: 260-262, 274-277) were incorporated in the revised manuscript.

    We have also added the mitochondrial copy number data (Figure 3A), showing that MTIF3 knockout has lower mitochondrial content (methods: lines 491-500; results: 156-157)

    1. MTIF3 knockout adipocytes retain more triglycerides under glucose restriction is interesting. It may link to the previous result of lower fatty acid oxidation in MTIF3 knockout adipocytes. However, the authors then showed there is no difference in lipolysis. The author should discuss those results in the manuscript.The authors could also check lipolysis in glucose restriction conditions. It's also necessary to include the triglyceride levels of KO cell lines at full medium

    We have now examined the glycerol release in glucose restriction condition, and found no differences between control and MTIF3 knockouts (Figure 6-figure supplement 1). Interestingly, in 1 mM glucose, both genotypes released less glycerol than at 25 mM glucose, and this has been observed before in SGBS cell line [2] According to your suggestion, we have added the total triglyceride content at 25 mM glucose condition (Figure 6C), which also was not different between control and MTIF3 knockout cells. We speculate the higher retention of triglycerides in the knockouts could be due to higher re-esterification of lipolytically released fatty acids, since, as we observed, fatty acid oxidation is impaired in the knockouts. In the revised manuscript, we added that to the discussion (lines: 289-293).

    References

    1. Rudler, D.L., et al., Fidelity of translation initiation is required for coordinated respiratory complex assembly. Sci Adv, 2019. 5(12): p. eaay2118.
    2. Renes, J., et al., Calorie restriction-induced changes in the secretome of human adipocytes, comparison with resveratrol-induced secretome effects. Biochim Biophys Acta, 2014. 1844(9): p. 1511-22.
  3. eLife

    eLife assessment

    The study provides a fundamental framework for linking human genome variation to targetable mechanisms of disease. The authors provide compelling evidence that a strong candidate locus associates with body weight in humans acts through adipocyte MTIF3. Thus, the generalized approaches taken in this study have the potential to inform genetic association studies in general and lay a foundation for future functional genomics studies.

  4. eLife

    Reviewer #1 (Public Review):

    The article "Identification of a weight loss-associated causal eQTL in MTIF3 and the effects of MTIF3 deficiency on human adipocyte function" explored the functional roles of MTIF3 during adipocyte differentiation. In persons living with obesity, genetic variation at the MTIF3 locus associates with body mass index and responses to weight loss interventions. MTIF3 regulates mitochondrial protein expression and gene knockouts cause cardiomyopathy in mice. This paper provides insight into the impacts of MTIF3 knockout on adipocyte differentiation and the expression effects of the eQTL on MTIF3 levels. The authors implement a CRISPR/Cas9 gene editing approach coupled with an in vitro platform to detect influences of MTIF3 on adipocyte glucose metabolism and gene expression. This method may serve as a platform to explore knockouts in human cell lines, so it may allow the discovery of new gene x environment influences on in vitro outcomes related to differentiation, growth, and metabolism.

    The conclusions of this paper are mostly well supported by data, but some experimental conditions and data analysis needs to be clarified and extended.

    1. The authors use CRISPR/Cas9 to generate the rs1885988 variant in the human white adipocyte cell line and performed a comprehensive validation analysis of gene editing (Figure 1). qPCR analysis showed reduced MTIF3 expression during human adipocyte differentiation (Figure 1E, F). To expand the importance of the rs1885988 variant, the authors should have provided target gene measurements to verify the canonical differentiation profile (e.g., FABP4, ADIPOQ) and help readers understand the overall impact of gene editing at the MTIF3 locus.
    2. The direct mechanistic influences of MTIF3 on adipocyte function remain unclear. MTIF3 regulates the translation initiation of mitochondrial protein synthesis. Western blots of OXPHOS proteins do not per se underscore supercomplex formation, which is also a process mediated by MTIF3. Blue native gel electrophoresis may prove a better method to establish the effects of MTIF3 loss-of-function on supercomplex formation.
    3. Based on the findings, the authors argue that MTIF3 knockout alters the function of adipocytes. However, many of the experiments show fairly small effect sizes (Figure 5A, Figure 6A). How does the MTIF3 knockout explicitly perform functions related to body weight regulation? Gene editing in vivo would have helped to substantiate the authors' conclusions.
    4. In several instances, the authors refer to 'feeding' cells with glucose (line 206, line 171). Feeding experiments often imply complex nutrient interventions in animal models and people, which cannot be easily recapitulated in cell culture. The in vitro experiments simply alter levels of glucose and more precise language would state the specific challenges accurately.
  5. eLife

    Reviewer #2 (Public Review):

    Huang Mi, et al. investigated the role of MTIF3, the mitochondrial translation initiation factor 3, in the function of adipocytes. They first detected the expression of the obesity-related MTIF3 variants based on the G5Ex database and found two variants lead to an increase in MTIF3 expression. Then they knockout MTIF3 in differentiated hWAs adipocytes and characterized the mitochondrial function. They found loss of MTIF3 decrease mitochondrial respiration and fatty acid oxidation. They further treated cells with low glucose medium to mimic weight loss intervention and found MTIF3 knockout adipocytes lose fewer triglycerides than control adipocytes. This paper provides new information about MTIF3 in adipocytes and the potential functional role of MTIF3 in mitochondrial function.

    1. The authors provided sufficient data to show those two genetic variants increase MTIF3 expression. Their CRISPR/Cas9 knockin cell line is also convincing. But they didn't show if the genetic variants affect adipogenesis. Adipogenesis is an important process for weight gain and fat deposition. In lines 103-107, the authors mentioned that the "allele-edited cells have some problem in differentiated state, e.g. triglyceride or mitochondrial content", so they used an inducible Cas9 system. However, the issue of differentiated allele-edited cells may be the functional effect of MTIF3 genetic variants, such as interrupting adipogenesis, decreasing triglyceride, or affecting mitochondrial number. The authors should provide that information.

    2. In Figure 4, the author mentioned that MTIF3 knockout does not affect the expression of adipogenic differentiation markers. They need to provide more evidence to prove their point. Oil-red O staining is a clearer way to quantify adipocyte differentiation in cell culture. In addition, in Fig. 4B western blot, the author should include MTIF3 as a control to show the knockout efficiency. It is not clear the meaning of plus and minus in that panel. The author should also compare the total triglyceride levels in MTIF3 knockout cells and control cells.

    3. MTIF3 is a translation initiation factor in mitochondria and is involved in the protein synthesis of mitochondrial DNA-encoding genes. The authors should check protein levels rather than the mRNA levels of mitochondrial DNA-encoding genes (Fig. 6E). It's interesting to see the increase of mRNA levels of ND1 and ND2, which might be feedback of lower translation. Since ND1 and ND2 are in OXPHOS complex I, the expression levels of complex I in MTIF3 KO cells would be worth checking. Additionally, the author should also check the mitochondria copy number.

    4. MTIF3 knockout adipocytes retain more triglycerides under glucose restriction is interesting. It may link to the previous result of lower fatty acid oxidation in MTIF3 knockout adipocytes. However, the authors then showed there is no difference in lipolysis. The author should discuss those results in the manuscript. The authors could also check lipolysis in glucose restriction conditions. It's also necessary to include the triglyceride levels of KO cell lines at full medium.

  6. Version published to 10.1101/2022.10.16.512435v1 on bioRxiv