Analysis of Genes Involved in Body Weight Regulation by Targeted Re-Sequencing

Abstract

Introduction

Genes involved in body weight regulation that were previously investigated in genome-wide association studies (GWAS) and in animal models were target-enriched followed by massive parallel next generation sequencing.

Methods

We enriched and re-sequenced continuous genomic regions comprising FTO, MC4R, TMEM18, SDCCAG8, TKNS, MSRA and TBC1D1 in a screening sample of 196 extremely obese children and adolescents with age and sex specific body mass index (BMI) ≥ 99^th percentile and 176 lean adults (BMI ≤ 15^th percentile). 22 variants were confirmed by Sanger sequencing. Genotyping was performed in up to 705 independent obesity trios (extremely obese child and both parents), 243 extremely obese cases and 261 lean adults.

Results and Conclusion

We detected 20 different non-synonymous variants, one frame shift and one nonsense mutation in the 7 continuous genomic regions in study groups of different weight extremes. For SNP Arg695Cys (rs58983546) in TBC1D1 we detected nominal association with obesity (p_TDT = 0.03 in 705 trios). Eleven of the variants were rare, thus were only detected heterozygously in up to ten individual(s) of the complete screening sample of 372 individuals. Two of them (in FTO and MSRA) were found in lean individuals, nine in extremely obese. In silico analyses of the 11 variants did not reveal functional implications for the mutations. Concordant with our hypothesis we detected a rare variant that potentially leads to loss of FTO function in a lean individual. For TBC1D1, in contrary to our hypothesis, the loss of function variant (Arg443Stop) was found in an obese individual. Functional in vitro studies are warranted.

About the Authors:

Anna-Lena Volckmar

Contributed equally to this work with: Anna-Lena Volckmar, Chung Ting Han, Sascha Sauer, Anke Hinney

Affiliation: Department of Child and Adolescent Psychiatry, Psychosomatics and Psychotherapy, University Hospital Essen, Essen, Germany

Chung Ting Han

Contributed equally to this work with: Anna-Lena Volckmar, Chung Ting Han, Sascha Sauer, Anke Hinney

Affiliations Nutrigenomics and Gene Regulation, Max-Planck-Institute for Molecular Genetics, Berlin, Germany, Genomics, Core Facilities and Technology, Institute of Molecular Biology (IMB), Mainz, Germany

Carolin Pütter

Affiliation: Institute for Medical Informatics, Biometry and Epidemiology (IMIBE), University Hospital Essen, Essen, Germany

Stefan Haas

Affiliation: Nutrigenomics and Gene Regulation, Max-Planck-Institute for Molecular Genetics, Berlin, Germany

Carla I. G. Vogel

Affiliation: Department of Animal and Food Production, Santa Catarina State University, Lages, Brazil

Nadja Knoll

Affiliation: Department of Child and Adolescent Psychiatry, Psychosomatics and Psychotherapy, University Hospital Essen, Essen, Germany

Christoph Struve

Affiliation: Department of Child and Adolescent Psychiatry, Psychosomatics and Psychotherapy, University Hospital Essen, Essen, Germany

Maria Göbel

Affiliation: Department of Child and Adolescent Psychiatry, Psychosomatics and Psychotherapy, University Hospital Essen, Essen, Germany

Katharina Haas

Affiliation: Department of Child and Adolescent Psychiatry, Psychosomatics and Psychotherapy, University Hospital Essen, Essen, Germany

Nikolas Herrfurth

Affiliation: Department of Child and Adolescent Psychiatry, Psychosomatics and Psychotherapy, University Hospital Essen, Essen, Germany

Ivonne Jarick

Affiliation: Institute of Medical Biometry and Epidemiology, Philipps-University of Marburg, Marburg, Germany

Harald Grallert

Affiliation: Institute of Epidemiology, Helmholtz-Zentrum Munich, Munich, Germany

Annette Schürmann

Affiliations Institute of Experimental Diabetology, German Institute of Human Nutrition Potsdam-Rehbruecke, Nuthetal, Germany, German Center for Diabetes Research (DZD), München-Neuherberg, Germany

Hadi Al-Hasani

Affiliations German Center for Diabetes Research (DZD), München-Neuherberg, Germany, Institute of Pathobiochemistry, German Diabetes Center and German Center for Diabetes Research, Düsseldorf, Germany

Johannes Hebebrand

Affiliation: Department of Child and Adolescent Psychiatry, Psychosomatics and Psychotherapy, University Hospital Essen, Essen, Germany

Sascha Sauer

Contributed equally to this work with: Anna-Lena Volckmar, Chung Ting Han, Sascha Sauer, Anke Hinney

Affiliations Nutrigenomics and Gene Regulation, Max-Planck-Institute for Molecular Genetics, Berlin, Germany, CU Systems Medicine, University of Wuerzburg, Wuerzburg, Germany

Anke Hinney

Contributed equally to this work with: Anna-Lena Volckmar, Chung Ting Han, Sascha Sauer, Anke Hinney

* E-mail: [email protected]

Affiliation: Department of Child and Adolescent Psychiatry, Psychosomatics and Psychotherapy, University Hospital Essen, Essen, Germany

Introduction

Obesity is one of the major health problems, which is associated with increased mortality and morbidity [1]. To date more than 100 body mass index (BMI) associated loci have been published from GWAS [2, 3, 4]. On the other hand, murine models have shown other relevant genes for weight regulation which were not detected by GWAS (e.g. Tbc1d1 [5]). The aim of this study is to identify functionally relevant mutations in genes involved in body weight regulation derived from either GWAS or murine models. The rationale for the chosen genes is briefly delineated in the following:

TBC1D1

In SJL mice a specific mutation in the Tbc1d1 gene (fsAla4047*4119) results in a truncated protein lacking the TBC Rab-GTPase-activating protein domain. The mutation led to resistance to diet-induced obesity [5] and its causality for the phenotype was confirmed in Tbc1d1 knockout mice [6, 7]. In mouse skeletal muscle cells, knockdown of Tbc1d1 increased fatty acid uptake and oxidation whereas overexpression of Tbc1d1 had the opposite effect [5]. Mutations in the human TBC1D1 gene are associated with increased risk for familial obesity [8, 9]. Polygenic effects on BMI and waist circumference in humans were also described for some TBC1D1 SNP alleles [10].

FTO

The fat mass and obesity associated gene (FTO) harbors GWAS derived polygenic variants with the largest effect size on BMI (FTO; [3, 11]). The body weight of carriers of one risk allele is increased by approximately 1.5 kgs. The effect of the risk alleles in intron 1 of FTO on body weight had been replicated in most analyzed study groups, either across the whole life span [3, 11–14] or in all analyzed ethnic groups [11, 15, 16]. FTO is highly expressed in the hypothalamus which is the key region for control of food intake [17]. FTO belongs to a gene family that is involved in post-translational modifications, DNA repair and fatty acid metabolism [18–20]. Fto deficient mice are lean as a consequence of increased energy expenditure [21] and show an improvement in metabolic syndrome in comparison to leptin-deficient mice, wild type or heterozygotes for Fto [22]. Mice over expressing Fto showed a dose-dependent and diet-independent increase in body weight and fat mass [23]. This is consistent with the finding that obesity associated SNPs in intron 1 increase FTO expression in humans [24]. The same SNP in human is associated with increased protein and lower overall energy intake [25].

TMEM18

The GWAS-derived gene TMEM18 is associated with increased BMI [11, 13, 26]. Besides obesity, variants in TMEM18 are associated with earlier onset of menarche [27], metabolic syndrome [28], and increased waist to hip ratio [26]. The expression of Tmem18 in the prefrontal cortex is strongly and positively correlated with body weight in rats [26]. In humans, a methylation dependent elevated expression of TMEM18 in subcutaneous adipose tissue was associated with BMI and metabolic traits [29]. The transmembrane protein TMEM18 binds DNA with its positively charged C-terminus that contains also a nuclear localization signal and is likely involved in moving the chromatin close to the nuclear membrane [30].

SDCCAG8, MSRA and TNKS

We [13] previously described variants associated with early-onset extreme obesity in children and adolescents. The loci are located near the serologically defined colon cancer antigen 8 gene (SDCCAG8) and between the genes coding for tankyrase (TNKS) and for methionine sulfoxide reductase A (MSRA). For these three genes, the role in the regulation of weight and energy homeostasis is yet unknown. SDCCAG8 is involved in Bardet-Biedl syndrome; which is a syndromal form of obesity [31, 32]. Variants in SDCCAG8 led to reduced weight loss in a group of 401 overweight and obese children and adolescents undergoing a lifestyle intervention [13] but had no effect on weight regain after the intervention [33].

TNKS regulates the centrosome function [34] and telomere conservation [35]. Variants in this gene are associated with type 2 diabetes [36] and several cancers (e.g. gastric cancer [37]; breast cancer [38]; colon cancer [39]; lung cancer [40]). MSRA catalyzes the enzymatic reduction of methionine sulfoxide to methionine. The latter is needed to repair oxidative damage of proteins. Obesity is associated with oxidative stress caused by reactive oxygen species (ROS) in the mitochondrion, with chronic excess of ROS that leads to mitochondrial dysfunction in liver and skeletal muscle which contributes to insulin resistance [41]. Mice without Msra show high-fat-diet-induced insulin resistance, most likely due to increased oxidative stress [42]. Variants in MSRA were associated with increased waist circumference and waist-to-hip ratio in a group of 3494 US Hispanic women [43]. Variants in both TNKS and MSRA did not affect weight loss in the aforementioned weight loss group of 401 overweight and obese children and adolescents [44] and had no effect on weight regain after the intervention [33].

Although heritability of body weight is high [45], GWAS could only detect a small fraction of the assumed heritability. The 97 lead SNPs detected by the currently largest meta-analysis of GWAS for BMI explain only approximately 5% of the total heritability of the variance of BMI [3].

Here, we analyzed the coding region of genes relevant for body weight regulation derived from GWAS (FTO, MC4R, SDCCAG8, TNKS/MSRA, and TMEM18) or animal models (TBC1D1) via targeted re-sequencing for potentially causal variants involved in body weight regulation. We used extreme phenotypes (196 individuals with the highest and 176 with the lowest BMI of our previously described case control GWAS dataset; [12, 46]).

Material and Methods

Study group

The screening group consisted of individuals with the highest or lowest BMI, respectively, of our previously described case control GWAS dataset [12, 13]. 196 extremely obese cases (43.9% male, mean age 15.0 years, mean BMI 38.45 kg/m², mean BMI SDS 3.11) and 176 lean controls (52.0% male, mean age 24.74 years, mean BMI 18.15 kg/m², mean BMI SDS -2.31) were analyzed (Fig 1). All individuals were previously analyzed by GWAS (Affymetrix Genome-Wide Human SNP Array 6.0) and were screened by dHPLC for mutations in the MC4R [47].

[Figure omitted. See PDF.]

Fig 1. Flowchart of the experimental setup.

https://doi.org/10.1371/journal.pone.0147904.g001

Confirmation groups.

(A) Family-based: A total of 705 obesity trios were used for confirmation. All detected non-synonymous, nonsense or frameshift variants were genotyped in an independent study group of 355 obesity trios (index patients: 45.37% male, mean age 13.72 ± 3.12 years, mean BMI 31.75 ± 6.06 kg/m², mean BMI SDS 4.13 ± 2.06; and both biological parents; subgroup of a study group previously described [48]). The variants rs2279027 (TBC1D1 [Ser14Pro])), rs58983546 (TBC1D1 [Arg695Cys]), and rs34790717 (TNKS [Gly237Ala]) were additionally genotyped in 350 independent obesity trios (total of 705 trios; see [48]). (B) Case-control: The potentially functionally relevant variant MSRA [R6Efs*88] was additionally analyzed in 243 young extremely obese cases and 261 lean adult controls by TaqMan assays ([12, 13], Fig 1).

Written informed consent was given by all participants and in case of minors by their parents. The studies were approved by the Ethics Committees of the Universities of Marburg and Duisburg-Essen and were performed in accordance with the Declaration of Helsinki.

Targeted Re-sequencing by Next-Generation Sequencing (NGS)

5 μg of genomic DNA was sheared to 200 bp by Covaris S2 instrument. 2 μg of sheared DNA was subjected to paired-end library construction, namely end-repair, a-tailing, and amplification with slight modifications [49]. Libraries were individually indexed and pooled prior to targeted enrichment by Agilent SureSelect In-Solution Target Enrichment System. The complete coding regions, including introns, of the genes FTO, MC4R, TMEM18, SDCCAG8, TKNS, MSRA and TBC1D1 with 10kb 5’ and 3’ flanks to include potential regulatory regions were targeted (S1 Table).

After target enrichment, the libraries were sequenced for 100 bp, using pair-end reads, on an Illumina Genome Analyzer IIx. Read mapping on hg18 and variant calling was performed using the same processing steps as described by Hu et al [50]. In order to reduce false-positive calls only variants with a genotype score >20 were used for subsequent analyses. The raw data were analyzed with pibase (http://www.ikmb.uni-kiel.de/pibase/index.html). For further analyses, the complete exon regions of all protein forming splice variants (http://www.ensembl.org/index.html, hg18) was extracted from the sequencing results. The analysis of 20 lean individuals failed during this process, so that the control group only consisted of 176 individuals.

The MC4R was included in the analysis despite being previously screened with an independent method (dHPLC, Sanger re-sequencing, [47]) to serve as a control for the used NGS protocol. Prior to analysis of the MC4R variants, the validity threshold for non-synonymous variants was set to a genotype score of ‘20’ ([50], which was a conservative threshold to rather include false positive variants, than to miss an existing mutation. With this threshold, a total of 19 variants were detected in MC4R (S2 Table). The NGS results were not corrected for failed samples or low coverage. These 19 variants detected by NGS were compared to the initial results of the dHPLC/Sanger re-sequencing approach. Deviant results were validated with additional Sanger re-sequencing. Of the 19 initial variants (threshold 20), only 5 could be confirmed by the independent methods. All variants that could not be reproduced by Sanger sequencing had a score below ‘100’. Hence, we decided to use a score of ‘100’ as a threshold for analysis of the NGS data for all analyzed genes.

Variant verification

All detected non-synonymous variants (Table 1) with minor allele frequencies below one percent were uni-directionally Sanger re-sequenced. At least two experienced individuals independently assigned the genotypes; discrepancies were resolved by reaching consensus or by re-sequencing. The primers for the PCRs can be obtained from the authors.

[Figure omitted. See PDF.]

Table 1. Number of detected variants per gene (FTO, TMEM18, SDCCAG8, TKNS, MSRA and TBC1D1) in 196 extremely obese children and adolescents and 176 lean adults.

https://doi.org/10.1371/journal.pone.0147904.t001

In silico functional analysis

All non-synonymous variants were in silico analyzed with MutationTaster (http://www.mutationtaster.org/ [51]), PMUT (http://mmb.pcb.ub.es/PMut/ [52]), PolyPhen-2 (http://genetics.bwh.harvard.edu/pph2/ [53]), and SNAP (https://www.rostlab.org/services/SNAP/ [54]) for potential functional impact on the protein level.

Genotyping for association analysis

All detected non-synonymous variants were genotyped in 355 obesity trios via MALDI TOF. The variants rs2279027 (TBC1D1 [Ser14Pro]) and rs58983546 (TBC1D1 [Arg695Cys]) were analyzed by TaqMan Assay (C___1673024_1_ and C__89662374_10, respectively) and the variant rs34790717 (TNKS [Gly237Ala]) by PCR-RFLP (Primers F 3’-GGCAAACGTAAATGCAAAGG-5’ and R 3’-CCTCACCAGAAGACTGGAGG-5’; digest of the C-allele with Mwo1) in 350 independent obesity trios. The variant TMP_ESP_8_9912039 (MSRA [R6Efs*88]) was additionally genotyped by a custom TaqMan SNP Genotyping Assay (AHZAE78) in 243 extremely obese children and adolescents and 261 lean adult controls.

Statistical analysis

At the start of the screening we decided to genotype the detected non-synonymous, frameshift and nonsense mutations in 355 independent obesity trios to identify potential transmission disequilibrium [55] for the analyzed variants. At that time, we did not know the allele frequencies of the variants to be detected, hence power calculations could not be performed. The TDT procedure was previously very successful in our hands for infrequent variants (1–3% allele frequency). For instance, we detected significantly reduced transmission of the infrequent allele at the MC4R Val103Ile polymorphism to obese children in a relatively small number of obesity trios (n = 520 [56]). We were the first group to describe the polygenic weight lowering effect of the MC4R 103Ile allele; previous case control association studies did not identify the small effect. Hence we regard the trio approach as especially powerful for variants with a low frequency, as stratification effects are mainly eliminated [55]. SNPs with tentatively low p-values in the 355 trios were genotyped in an additional 350 further obesity trios, so that a total of 705 obesity trios was screened for promising SNPs.

Association to obesity was analyzed by standard statistical programs (case-control: Fisher’s exact test from Excel, obesity trios: transmission-disequilibrium test in Excel [55]. All reported p-values are nominal, uncorrected and two-tailed.

Results

Screening procedure and quality control

We aimed to identify new mutations by targeted re-sequencing in the continuous genomic regions of seven genes involved in weight regulation. We used a screening sample of 196 obese and 176 lean individuals of different weight extremes; detected variants were confirmed in 355 independent obesity trios, SNPs with a nominal p-value below 0.05 were genotyped in a total of 705 obesity trios (355 plus 350 trios). Additionally, for the variant MSRA [R6Efs*88] which likely confers a functional effect, we analyzed 243 young extremely obese cases and 261 lean adult controls ([12, 13], Fig 1) on top of the 705 trios.

MC4R was used as a reference gene to define a validity score for the initial discovery of variants derived from re-sequencing of enriched genomic regions (S2 Table). We hence used the genotype score ‘100’ [50] for the remaining genes TBC1D1, FTO, TMEM18, SDCCAG8, TKNS, and MSRA (Table 1).

Mutational analysis of TBC1D1, FTO, TMEM18, SDCCAG8, TKNS, and MSRA

A total of 49 variants (48 non-synonymous and 1 frameshift, S2 Table) were detected in the sample of 196 extremely obese children and adolescents and 176 lean adults within the 6 genes (FTO, TMEM18, SDCCAG8, TKNS, MSRA and TBC1D1). Of these, 32 variants were known SNPs as listed in the dbSNP. The others were yet not described in the accessible databases (data access for all: November 9^th 2015: EVS: http://evs.gs.washington.edu/EVS/, 1000Genomes: http://browser.1000genomes.org/index.html, dbSNP: http://www.ncbi.nlm.nih.gov/snp/, and ExAC: http://exac.broadinstitute.org/variant/). The previously unknown variants had a minor allele frequency below 0.001 and were detected heterozygously in all cases (Table 2). The 22 non-synonymous and frameshift variants detected in the sample were Sanger re-sequenced; only one could not be confirmed (TNKS Asn1103Lys).

[Figure omitted. See PDF.]

Table 2. Non-synonymous, stop and frameshift variants¹ detected by NGS of the genes FTO, TMEM18, SDCCAG8, TKNS, MSRA and TBC1D1 in 196 extremely obese children and adolescents and 176 lean adults.

https://doi.org/10.1371/journal.pone.0147904.t002

Confirmation and genotyping in large independent study groups

Genotyping was performed in up to 705 obesity trios plus 243 young extremely obese cases and 261 lean adult controls [12, 13] independent of the screening sample. All detected non-synonymous, nonsense and frameshift variants were genotyped by MALDI-TOF in 355 independent family-based obesity trios. As most of these variants were rare, nine of the 21 could not be detected in the additional sample. None of the variants showed transmission disequilibrium for obesity in the 355 obesity trios (Table 3).

[Figure omitted. See PDF.]

Table 3. Transmission disequilibrium test of the 22 variants detected by NGS in 355 German obesity trios.

https://doi.org/10.1371/journal.pone.0147904.t003

Only a few of the detected non-synonymous variants were in silico predicted to behave like the wild type protein (FTO [Cys9Tyr], MSRA [Asp142Tyr] and MSRA [Gly187Ser]). For all others, an altered protein function was predicted (Table 4; Mutation Taster).

[Figure omitted. See PDF.]

Table 4. Phenotypes of heterozygous carriers of rare non-synonymous variants in the screening sample and extended in silico analyses for the respective variants.

https://doi.org/10.1371/journal.pone.0147904.t004

TBC1D1.

Of the 17 variants detected in TBC1D1, eleven were non-synonymous. The SNP Arg695Cys (rs58983546) was not associated with obesity in the initial sample (p = 1), but nominally associated in the 355 obesity trios (nominal p = 0.05). We added 350 independent trios and confirmed the association of the minor allele with obesity (nominal p for the complete set of 705 trios = 0.03). The variant is located in a domain of unknown function (DUF3350; PFAM) of TBC1D1; minor allele frequency is 1.25% in CEU (“Utah residents with ancestry from northern and western Europe”, dbSNP).

The stop gain variant TBC1D1 Arg443* was only detected once in an extremely obese heterozygous carrier (Table 4) out of a total of 551 obese and 419 lean individuals. A mutation leading to a stop codon would result in a loss of the Rab-GTPase-TBC domain which would result in a complete loss of function.

FTO.

Of the six detected variants in FTO, three were rare and non-synonymous (Table 2). These were found in either obese or lean individuals. Interestingly one mutation (Ala163Thr) was detected in two independent obese individuals, implying cryptic relatedness between these mutation carriers [57]. The variant was previously described by Meyre et al. [58] in both obese and lean subjects. The variant is located in a surface loop of FTO but was not tested in vitro for functional effects. In silico predictions for this variant were variable (Table 4). The other novel FTO variant Val83Leu was detected heterozygously in a lean female. It is located in the substrate recognition lid of the protein and in silico methods predict a highly likely functional outcome (Table 4).

TMEM18.

None of the variants in TMEM18 altered the amino acid sequence.

SDCCAG8.

Of the six variants detected in SDCCAG8, two were non-synonymous. Both are previously known SNPs (rs2275155 [Glu378Asp], rs79435766 [Thr398Met]). The latter variant was found in an extremely obese male adolescent (Table 4). The variant is located in a coiled coil domain. In silico functional prediction for this variant was variable (Table 4).

TNKS.

In TNKS, only 2 of 14 detected variants were non-synonymous. One of them was a previously known SNP (rs34790717 [Gly237Ala]), while the other one was previously unknown (TNKS [Pro275Ala]). The rare variant TNKS [Pro275Ala] was detected in an extremely obese female. However, in silico predictions for this variant mainly implied no functional effect on protein level (Table 4).

MSRA.

In MSRA we identified a frameshift mutation (Arg6Glufs*88; TMP_ESP_8_9912039) in an extremely obese female (age 13.6 years, BMI 42.67 kg/m², 100^th age and sex specific BMI percentile, Table 4). Within the 355 obesity trios the variant was detected in an additional unrelated father. This obese carrier (male, age 45 years, BMI 35 kg/m²) did not transmit the variant to his extremely obese offspring. We also screened for the frameshift mutation in an independent sample of 243 extremely obese children and adolescents and 261 lean adult controls. Here, the variant was detected in a normal weight adult female (age 26.8 years, BMI 21.3 kg/m²) and a lean adult male (age 24.3 years, BMI 16.5 kg/m²). Hence, in sum a strong relevance of this variant for body weight regulation is unlikely. Additionally, three non-synonymous variants were detected (MSRA rs6601444 [Thr88Met], [Asp142Tyr], and rs201155438 [Gly187Ser]). One of these [187Ser] was found in an extremely obese male adolescent; the other [142Tyr] in a lean participant (Table 4). Both mutations are located in the MSRA domain of the protein. In silico predictions for both variants range from neutral to highly functional (Table 4).

Discussion

We aimed to identify new mutations with relevance for body weight regulation by targeted re-sequencing (NGS) in the coding regions of six genes.

The gene TBC1D1 which was originally derived from animal studies contained most variants of all analyzed genes (17 total, of these 11 non synonymous, Table 1). SNP Arg695Cys (rs58983546) was not associated with obesity in the initial sample (p = 1), but showed nominal over-transmission of the 695Cys allele in the 355 obesity trios (p = 0.05; Table 3) and an increased sample size of 705 trios (p = 0.03). In the latest GWAS meta-analysis for BMI, the SNP was not associated with obesity (p = 0.38, [3]). We could not confirm the previously described obesity association [8, 9] of variant Arg125Trp (p = 0.66) in 355 obesity trios. However, the allele frequency for the risk allele was higher in our index patients than in a population based CEU sample (MAF in obese 0.17 vs. 0.07 in the ESP “NHLBI Exome Sequencing Project” cohort, dbSNP), or in previous studies (MAF 0.02 in n = 940 [8]; MAF 0.09 in n = 1,109 [9]). Both reported associations referred only to familial extreme obesity; in fact, Meyre et al. [9] were not able to confirm their finding in a population based sample for mild obesity and overweight (n = 4,634). Although the initial positive association was based mainly on females [9], we did not perform sex-stratified analyses as the allele frequency of the variant is too low for sufficiently powered analyses in females only. SNP Ser14Pro (rs2279027) was not associated with obesity in our study groups (Table 3).

All of the rare TBC1D1 variants (Arg443*, Gln619Arg, Leu838Val, Arg1091His) were detected in obese cases only (see Table 4). The mutation Arg443* was only detected heterozygously in one extremely obese individual. The mutation leads to a loss of the Rab-GTPase-TBC domain and thus presumably to a loss of function. Although Cheng et al. [59] described binding of the Tbc1d1 protein to the adaptor protein APPL1 via N-terminal PTB domains, for signal transduction to partners downstream of TBC1D1, a full Rab-GTPase-TBC domain is necessary. For the other three variants, in silico prediction was rather mixed and conservation was not very strong (Gln619Arg: 63% conservation, Leu838Val: 64% conservation, Arg1019His: 71% conservation for 67 analyzed species; http://www.ensembl.org/).

In short, Arg443* is the functionally most relevant variant we detected in TBC1D1. Contrary to our hypothesis, this variant was detected in an obese individual. Functional analyses are therefore warranted. TBC1D1 SNP Arg695Cys (rs58983546) could contribute to the obesity association of the gene. Hence, although the gene was initially identified in mice with diet induced obesity resistance [5] our findings imply that loss of function and function-reducing variants in humans might be relevant for obesity development.

For the most relevant obesity polygene FTO, our data are comparable to previous studies [58, 60] as we also detected non-synonymous variants in both obesity and control study groups. The obesity GWAS signals are localized to the first intron of FTO [61, 62] and apparently do not affect the amino acid structure of the protein [58] but rather the expression of FTO [24]. A recent study indicates that the obesity association of the first intron does not imply FTO as the causal gene, but IRX3 located downstream on the same chromosome which is regulated by the first intron of FTO [63]. The FTO variant Val83Leu, detected in a lean adult male, has the highest potential for a functional impact as the in silico predictions for this variant all indicate an altered function which would be concordant with the mouse model (lean phenotype in a knock out mouse). However, as the prediction methods can only analyze change of function, this could also hint at an increased function. Hence, in vitro analysis of this variant is highly warranted.

For SDCCAG8, previous mutation screens focused on obesity in patients with Bardet-Biedl syndrome [32, 64] and revealed frameshift-, nonsense and loss of splicing enhancer mutations. For the genes SDCCAG8, TNKS and MSRA, we expected to find variants leading to a reduced protein function. However, the variants we detected were too infrequent for meaningful obesity association analysis (SDCCAG8 [Thr398Met], TNKS [Pro275Ala], MSRA [Arg6GlufsX88], [Asp142Tyr], [Gly187Ser]) or showed no obesity association (SDCCAG8 [Glu378Asp], TNKS [Gly237Ala]; MSRA [Thr88Met],). In detail:

(a) The frameshift variant in MSRA [Arg6Glufs*88] was initially detected in an extremely obese individual. The functional implications are presumably strong (in silico analyses). Only two lean heterozygous mutation carriers were identified in an independent study group (243 extremely obese children and adolescents and 261 lean adult controls). An obese father of the 705 obesity trios did not transmit the variant his extremely obese offspring. Hence, an impact of the mutation on body weight regulation is rather unlikely. (b) The other rare MSRA variant (Gly187Ser) detected in an obese male was mainly predicted to be functionally irrelevant (Table 4). It is located at a position which is not conserved (3% conservation among 68 species, http://www.ensembl.org/). (c) The rare variant Asp142Tyr in MSRA was detected in a lean female (Table 4). It is also located in the MSRA domain of the protein. The amino acid position 142 was conserved between 68 species (conservation 71%, http://www.ensembl.org/). Taken together, our findings hint that the detected MSRA mutations are unlikely to be relevant for weight regulation.

The two genes TNKS and SDCCAG8, derived from GWAS meta-analyses in extremely obese children and adolescent, both harbored rare mutations only in the obese cases. The conservative variant Pro275Ala is located in a region of TNKS that is highly conserved (81% in 72 species, http://www.ensembl.org/). Nonetheless in silico prediction is mainly neutral (Table 4). Conservation of amino acid position Thr398 is low (41% in 68 species, http://www.ensembl.org/). In sum a strong impact of the detected mutations in TNKS and SDCCAG8 on obesity cannot be derived from our data.

Limitations of our study include a small sample size, thus low power and the lack of functional studies pertaining to the rare non-synonymous and frame shift mutations we discovered. The respective in vitro functional studies are highly warranted as the predictive power of in silico programs is limited [65]. We did not correct for multiple testing as none of the variants was reproducibly nominally associated with obesity.

In summary, we screened the coding regions of seven genes for variants leading to monogenic forms of obesity. For FTO, we detected, concordant with our hypothesis one rare variant leading to a loss of FTO function in a lean individual. In TBC1D1 discordant to our hypothesis the loss of function variant (Arg443*) was found in an obese individual.

Supporting Information

[Figure omitted. See PDF.]

S1 Table. Genetic regions (hg19/GRCh37) for NGS including the genes (largest transcript) of interest with additional 10kb flank.

All analyzed genes (FTO, TMEM18, SDCCAG8, TKNS, MC4R, MSRA and TBC1D1) with regions covered.

https://doi.org/10.1371/journal.pone.0147904.s001

(DOCX)

S2 Table. Comparison of the results of next generation sequencing and Sanger sequencing for the MC4R coding region in 196 extremely obese children and adolescents and 176 lean adults.

All detected variants in MC4R and the respective scores for quality validation.

https://doi.org/10.1371/journal.pone.0147904.s002

(DOCX)

S3 Table. List off all variants detected with Score > 100.

All variants in the exonic regions of the screened genes FTO, TMEM18, SDCCAG8, TKNS, MC4R, MSRA and TBC1D1 in 196 extremely obese children and adolescents and 176 lean adults. Every deviant call from wild type is listed in one line including the probability of heterozygousity (column “Zygosity”) and the score for the overall validity of the variant (column “Score”).

https://doi.org/10.1371/journal.pone.0147904.s003

(DOCX)

Acknowledgments

We thank all the probands and their families for their participation. We are indebted to S. Düerkop and J. Andrä for technical support. The ascertainment of study groups belonging to different weight extremes was supported by the Deutsche Forschungsgemeinschaft (DFG). All other analyzes were supported by the DFG (HI865/2-1) and the BMBF (01KU0903; 0315082/01EA1303; NGFNplus: 01GS0820) and the IFORES program of the University of Duisburg-Essen.

Author Contributions

Conceived and designed the experiments: JH SS AH AS HA-H CIGV. Performed the experiments: ALV CTH NK SH HG CGIV. Analyzed the data: ALV NK CTH HG NH CS MG KH CP IJ. Contributed reagents/materials/analysis tools: AS HG SH CTH SS JH AH CP IJ. Wrote the paper: ALV CTH HA-H SS AH.

Citation: Volckmar A-L, Han CT, Pütter C, Haas S, Vogel CIG, Knoll N, et al. (2016) Analysis of Genes Involved in Body Weight Regulation by Targeted Re-Sequencing. PLoS ONE 11(2): e0147904. https://doi.org/10.1371/journal.pone.0147904

References

1. WHO. Obesity and Overweight. Fact sheet Number 311. 2011. http://www.who.int/mediacentre/factsheets/fs311/en/

2. Fall T, Ingelsson E. Genome-wide association studies of obesity and metabolic syndrome. Mol Cell Endocrinol. 2014;382: 740–757. pmid:22963884

3. Locke AE, Kahali B, Berndt SI, Justice AE, Pers TH, Day FR, et al. Genetic studies of body mass index yield new insights for obesity biology. Nature. 2015;518: 197–206. pmid:25673413

4. Bradfield JP, Taal HR, Timpson NJ, Scherag A, Lecoeur C, Warrington NM, et al. A genome-wide association meta-analysis identifies new childhood obesity loci. Nat Genet. 2012;44: 526–531. pmid:22484627

5. Chadt A, Leicht K, Deshmukh A, Jiang LQ, Scherneck S, Bernhardt U, et al. Tbc1d1 mutation in lean mouse strain confers leanness and protects from diet-induced obesity. Nat Genet. 2008;40: 1354–1359. pmid:18931681

6. Dokas J, Chadt A, Nolden T, Himmelbauer H, Zierath JR, Joost HG, et al. Conventional knockout of Tbc1d1 in mice impairs insulin- and AICAR-stimulated glucose uptake in skeletal muscle. Endocrinology. 2013;154: 3502–3514. pmid:23892475

7. Chadt A, Immisch A, de Wendt C, Springer C, Zhou Z, Stermann T, et al. Deletion of both Rab-GTPase–activating proteins TBC1D1 and TBC1D4 in mice eliminates insulin- and AICAR-stimulated glucose transport [corrected]. Diabetes. 2015;64: 746–759. pmid:25249576

8. Stone S, Abkevich V, Russell DL, Riley R, Timms K, Tran T, et al. TBC1D1 is a candidate for a severe obesity gene and evidence for a gene/gene interaction in obesity predisposition. Hum Mol Genet. 2006;15: 2709–2720. pmid:16893906

9. Meyre D, Farge M, Lecoeur C, Proenca C, Durand E, Allegaert F, et al. R125W coding variant in TBC1D1 confers risk for familial obesity and contributes to linkage on chromosome 4p14 in the French population. Hum Mol Genet. 2008;17: 1798–1802. pmid:18325908

10. Knüppel S, Rohde K, Meidtner K, Drogan D, Holzhütter HG, Boeing H, et al. Evaluation of 41 candidate gene variants for obesity in the EPIC-potsdam cohort by multi-locus stepwise regression. PLoS One. 2013;8: e68941. pmid:23874820

11. Speliotes EK, Willer CJ, Berndt SI, Monda KL, Thorleifsson G, Jackson AU, et al. Association analyses of 249,796 individuals reveal 18 new loci associated with body mass index. Nat Genet. 2010;42: 937–948. pmid:20935630

12. Hinney A, Nguyen TT, Scherag A, Friedel S, Brönner G, Müller TD, et al. Genome wide association (GWA) study for early onset extreme obesity supports the role of fat mass and obesity associated gene (FTO) variants. PLoS One. 2007;2: e1361. pmid:18159244

13. Scherag A, Dina C, Hinney A, Vatin V, Scherag S, Vogel CI, et al. Two new Loci for body-weight regulation identified in a joint analysis of genome-wide association studies for early-onset extreme obesity in French and german study groups. PLoS Genet. 2010;6: e1000916. pmid:20421936

14. Bradfield JP, Taal HR, Timpson NJ, Scherag A, Lecoeur C, Warrington NM, et al. A genome-wide association meta-analysis identifies new childhood obesity loci. Nat Genet. 2012;44: 526–531. pmid:22484627

15. Wen W, Cho YS, Zheng W, Dorajoo R, Kato N, Qi L, et al. Meta-analysis identifies common variants associated with body mass index in east Asians. Nat Genet. 2012;44: 307–311. pmid:22344219

16. Monda KL, Chen GK, Taylor KC, Palmer C, Edwards TL, Lange LA, et al. A meta-analysis identifies new loci associated with body mass index in individuals of African ancestry. Nat Genet. 2013;45: 690–696. pmid:23583978

17. Gerken T, Girard CA, Tung YC, Webby CJ, Saudek V, Hewitson KS, et al. The obesity-associated FTO gene encodes a 2-oxoglutarate-dependent nucleic acid demethylase. Science. 2007;318: 1469–1472. pmid:17991826

18. Sanchez-Pulido L, Andrade-Navarro MA. The FTO (fat mass and obesity associated) gene codes for a novel member of the non-heme dioxygenase superfamily. BMC Biochem. 2007;8: 23. pmid:17996046

19. Wåhlén K, Sjölin E, Hoffstedt J. The common rs9939609 gene variant of the fat mass- and obesity-associated gene FTO is related to fat cell lipolysis. J Lipid Res. 2008;49: 607–611. pmid:18048838

20. Jia G, Yang CG, Yang S, Jian X, Yi C, Zhou Z, et al. Oxidative demethylation of 3-methylthymine and 3-methyluracil in single-stranded DNA and RNA by mouse and human FTO. FEBS Lett. 2008;582: 3313–3319. pmid:18775698

21. Fischer J, Koch L, Emmerling C, Vierkotten J, Peters T, Brüning JC, et al. Inactivation of the Fto gene protects from obesity. Nature. 2009;458: 894–898. pmid:19234441

22. Ikels K, Kuschel S, Fischer J, Kaisers W, Eberhard D, Rüther U. FTO is a relevant factor for the development of the metabolic syndrome in mice. PLoS One. 2014;9: e105349. pmid:25144618

23. Church C, Moir L, McMurray F, Girard C, Banks GT, Teboul L, et al. Overexpression of Fto leads to increased food intake and results in obesity. Nat Genet. 2010;42: 1086–1092. pmid:21076408

24. Berulava T, Horsthemke B. The obesity-associated SNPs in intron 1 of the FTO gene affect primary transcript levels. Eur J Hum Genet. 2010;18: 1054–1056. pmid:20512162

25. Livingstone KM, Celis-Morales C, Lara J, Ashor AW, Lovegrove JA, Martinez JA, et al. Associations between FTO genotype and total energy and macronutrient intake in adults: a systematic review and meta-analysis. Obes Rev. 2015;16:666–678. pmid:26016642

26. Rask-Andersen M, Jacobsson JA, Moschonis G, Chavan RA, Sikder MA, Allzén E, et al. Association of TMEM18 variants with BMI and waist circumference in children and correlation of mRNA expression in the PFC with body weight in rats. Eur J Hum Genet. 2012;20: 192–197. pmid:21952719

27. Mumby HS, Elks CE, Li S, Sharp SJ, Khaw KT, Luben RN, et al. Mendelian Randomisation Study of Childhood BMI and Early Menarche. J Obes. 2011; 2011: 180729. pmid:21773002

28. Dušátková L, Zamrazilová H, Sedláčková B, Včelák J, Hlavatý P, Aldhoon Hainerová I, et al. Association of obesity susceptibility gene variants with metabolic syndrome and related traits in 1,443 czech adolescents. Folia Biol (Praha). 2013;59: 123–133.

29. Rohde K, Keller M, Klös M, Schleinitz D, Dietrich A, Schön MR, et al. Adipose tissue depot specific promoter methylation of TMEM18. J Mol Med (Berl). 2014;92: 881–888.

30. Jurvansuu JM, Goldman A. Obesity risk gene TMEM18 encodes a sequence-specific DNA-binding protein. PLoS One. 2011;6: e25317. pmid:21980424

31. Schaefer E, Zaloszyc A, Lauer J, Durand M, Stutzmann F, Perdomo-Trujillo Y, et al. Mutations in SDCCAG8/NPHP10 Cause Bardet-Biedl Syndrome and Are Associated with Penetrant Renal Disease and Absent Polydactyly. Mol Syndromol. 2011;1: 273–281. pmid:22190896

32. Billingsley G, Vincent A, Deveault C, Héon E. Mutational analysis of SDCCAG8 in Bardet-Biedl syndrome patients with renal involvement and absent polydactyly. Ophthalmic Genet. 2012;33: 150–154. pmid:22626039

33. Hinney A, Wolters B, Pütter C, Grallert H, Illig T, Hebebrand J, et al. No impact of obesity susceptibility loci on weight regain after a lifestyle intervention in overweight children. J Pediatr Endocrinol Metab. 2013;26: 1209–1213. pmid:23843577

34. Kim MK, Dudognon C, Smith S. Tankyrase 1 regulates centrosome function by controlling CPAP stability. EMBO Rep. 2012;13: 724–732. pmid:22699936

35. Chang S. The telomere protein tankyrase 1 regulates DNA damage responses at telomeres. Aging (Albany NY). 2010;2:639–642.

36. Zee RY, Ridker PM, Chasman DI. Genetic variants of 11 telomere-pathway gene loci and the risk of incident type 2 diabetes mellitus: the Women's Genome Health Study. Atherosclerosis. 2011;218: 144–146. pmid:21665207

37. Gao J, Zhang J, Long Y, Tian Y, Lu X. Expression of tankyrase 1 in gastric cancer and its correlation with telomerase activity. Pathol Oncol Res. 2011;17: 685–690. pmid:21455637

38. Pellatt AJ, Wolff RK, Torres-Mejia G, John EM, Herrick JS, Lundgreen A, et al. Telomere length, telomere-related genes, and breast cancer risk: the breast cancer health disparities study. Genes Chromosomes Cancer. 2013;52: 595–609. pmid:23629941

39. Pellatt AJ, Wolff RK, Lundgreen A, Cawthon R, Slattery ML. Genetic and lifestyle influence on telomere length and subsequent risk of colon cancer in a case control study. Int J Mol Epidemiol Genet. 2012;3: 184–194. pmid:23050049

40. Busch AM, Johnson KC, Stan RV, Sanglikar A, Ahmed Y, Dmitrovsky E, et al. Evidence for tankyrases as antineoplastic targets in lung cancer. BMC Cancer. 2013;13: 211. pmid:23621985

41. de Ferranti S, Mozaffarian D. The perfect storm: obesity, adipocyte dysfunction, and metabolic consequences. Clin Chem. 2008;54: 945–955. pmid:18436717

42. Styskal J, Nwagwu FA, Watkins YN, Liang H, Richardson A, Musi N, et al. Methionine sulfoxide reductase A affects insulin resistance by protecting insulin receptor function. Free Radic Biol Med. 2013;56: 123–132. pmid:23089224

43. Graff M, Ngwa JS, Workalemahu T, Homuth G, Schipf S, Teumer A, et al. Genome-wide analysis of BMI in adolescents and young adults reveals additional insight into the effects of genetic loci over the life course. Hum Mol Genet. 2013;22: 3597–3607. pmid:23669352

44. Scherag A, Kleber M, Boes T, Kolbe AL, Ruth A, Grallert H, et al. SDCCAG8 obesity alleles and reduced weight loss after a lifestyle intervention in overweight children and adolescents. Obesity (Silver Spring). 2012;20: 466–470.

45. Hebebrand J, Hinney A, Knoll N, Volckmar AL, Scherag A. Molecular genetic aspects of weight regulation. Dtsch Arztebl Int. 2013;110: 338–344. pmid:23762205

46. Pütter C, Pechlivanis S, Nöthen MM, Jöckel KH, Wichmann HE, Scherag A. Missing Heritability in the Tails of Quantitative Traits? A Simulation Study on the Impact of Slightly Altered True Genetic Models. Hum Hered. 2011;72: 173–181. pmid:22041814

47. Hinney A, Hohmann S, Geller F, Vogel C, Hess C, Wermter AK, et al. Melanocortin-4 receptor gene: case-control study and transmission disequilibrium test confirm that functionally relevant mutations are compatible with a major gene effect for extreme obesity. J Clin Endocrinol Metab. 2003;88: 4258–4267. pmid:12970296

48. Jarick I, Vogel CI, Scherag S, Schäfer H, Hebebrand J, Hinney A, et al. Novel common copy number variation for early onset extreme obesity on chromosome 11q11 identified by a genome-wide analysis. Hum Mol Genet. 2011;20: 840–852. pmid:21131291

49. Quail MA, Swerdlow H, Turner DJ. Improved protocols for the illumina genome analyzer sequencing system. Curr Protoc Hum Genet. 2009; 18: Unit 18.2.

50. Hu H, Haas SA, Chelly J, Van Esch H, Raynaud M, de Brouwer AP, et al. X-exome sequencing of 405 unresolved families identifies seven novel intellectual disability genes. Mol Psychiatry. 2015 Feb 3. [Epub ahead of print]

51. Schwarz JM, Rödelsperger C, Schuelke M, Seelow D. MutationTaster evaluates disease-causing potential of sequence alterations. Nat Methods 2010; 7:575–576. pmid:20676075

52. Ferrer-Costa C, Orozco M, de la Cruz X. Sequence-based prediction of pathological mutations. Proteins 2004; 57:811–819. pmid:15390262

53. Ramensky V, Bork P, Sunyaev S. Human non-synonymous SNPs: server and survey. Nucleic Acids Res 2002;30: 3894–3900. pmid:12202775

54. Bromberg Y, Rost B. SNAP: predict effect of non-synonymous polymorphisms on function. Nucleic Acids Res 2007; 35:3823–3835. pmid:17526529

55. Spielman RS, McGinnis RE, Ewens WJ. Transmission test for linkage disequilibrium: the insulin gene region and insulin-dependent diabetes mellitus (IDDM). Am J Hum Genet. 1993;52:506–516. pmid:8447318

56. Geller F, Reichwald K, Dempfle A, Illig T, Vollmert C, Herpertz S, et al. Melanocortin-4 receptor gene variant I103 is negatively associated with obesity. Am J Hum Genet. 2004;74:572–581. pmid:14973783

57. Mühlhaus J, Pütter C, Brumm H, Grallert H, Illig T, Scherag S, et al. Do common variants separate between obese melanocortin-4 receptor gene mutation carriers and non-carriers? The impact of cryptic relatedness. Horm Res Paediatr. 2012;77: 358–368. pmid:22688572

58. Meyre D, Proulx K, Kawagoe-Takaki H, Vatin V, Gutiérrez-Aguilar R, Lyon D, et al. Prevalence of loss-of-function FTO mutations in lean and obese individuals. Diabetes. 2010;59: 311–318. pmid:19833892

59. Cheng KK, Zhu W, Chen B, Wang Y, Wu D, Sweeney G, et al. The adaptor protein APPL2 inhibits insulin-stimulated glucose uptake by interacting with TBC1D1 in skeletal muscle. Diabetes. 2014;63: 3748–3758. pmid:24879834

60. Zheng Z, Hong L, Huang X, Yang P, Li J, Ding Y, et al. Screening for coding variants in FTO and SH2B1 genes in Chinese patients with obesity. PLoS One. 2013;8: e67039. pmid:23825611

61. Frayling TM, Timpson NJ, Weedon MN, Zeggini E, Freathy RM, Lindgren CM, et al. A common variant in the FTO gene is associated with body mass index and predisposes to childhood and adult obesity. Science. 2007;316: 889–894. pmid:17434869

62. Peters U, North KE, Sethupathy P, Buyske S, Haessler J, Jiao S, et al. A systematic mapping approach of 16q12.2/FTO and BMI in more than 20,000 African Americans narrows in on the underlying functional variation: results from the Population Architecture using Genomics and Epidemiology (PAGE) study. PLoS Genet. 2013;9: e1003171. pmid:23341774

63. Smemo S, Tena JJ, Kim KH, Gamazon ER, Sakabe NJ, Gómez-Marín C, et al. Obesity-associated variants within FTO form long-range functional connections with IRX3. Nature. 2014;507: 371–375. pmid:24646999

64. Schaefer E, Zaloszyc A, Lauer J, Durand M, Stutzmann F, Perdomo-Trujillo Y, et al. Mutations in SDCCAG8/NPHP10 Cause Bardet-Biedl Syndrome and Are Associated with Penetrant Renal Disease and Absent Polydactyly. Mol Syndromol. 2011;1: 273–281. pmid:22190896

65. Bonnefond A, Clément N, Fawcett K, Yengo L, Vaillant E, Guillaume JL, et al. Rare MTNR1B variants impairing melatonin receptor 1B function contribute to type 2 diabetes. Nat Genet. 2012;44: 297–301. pmid:22286214