Predictive Analysis of the Leptin-Melanocortin and Adiponectin Signaling Pathways in Obesity through In Silico Techniques

doi:10.21203/rs.3.rs-4018264/v1

Genetic and epigenetic alterations have been reported to significantly influence the global burden of obesity. Single nucleotide polymorphisms (SNPs) including both coding and non-coding amino acid changes are the key regulators of the protein structural and functional modifications. The current computational study utilizing in silico techniques focused on the screening and identification of the most pathogenic missense SNPs of the selected candidate genes of the leptin-melanocortin and adiponectin signaling pathways provoking obesity. A total of 2424 SNPs from 9 candidate genes were extracted from the NCBI database followed by pathogenicity prediction using seven servers, SIFT, PANTHER, Meta-SNP, PhD-SNP, PredictSNP, PolyPhen-2, and SNAP2. The shortlisted variants (n = 7) were analyzed for structural stability using DynaMut, iMutant, INPS3D, MuPro, and iStable followed by the functional stability analysis (n = 3) using Mut-Pred2, Project HOPE, and I-TASSER. Gene-network analysis of the finally screened SNPs (n = 3) was created using the STRING database. Two SNPs of ADIPOR1 (rs1419320091 and rs1654109863) and one variant of MC4R (rs1159323398) were predicted in the study to be the most pathogenic resulting in altered protein functionality. Therapeutic approaches designed based on early pathogenicity predictions using in silico analysis techniques would be a new horizon for the effective control of disease prevalence.

Biological sciences/Computational biology and bioinformatics

Biological sciences/Molecular biology

Health sciences/Diseases

Health sciences/Risk factors

Obesity

in silico

SNP

pathogenicity

protein stability

The global obesity prevalence accounts for the increased adiposity in 50 million girls, 74 million boys, 390 million women, and 281 million men worldwide [1]. Factors like rapid urbanization, global trade liberalization, and economic growth fundamentally drive the obesity epidemic. The increased consumption of animal products, refined food grains, and sugar-rich food with decreased overall food prices contributes to the epidemic [2]. Low-income and middle-income countries like Sub-Saharan Africa, India, China, and parts of Southeast Asia and South America currently experience the major disease burden of obesity-related metabolic syndrome (OMS) [3]. OMS including cardiovascular diseases, type 2 diabetes mellitus, polycystic ovarian syndrome, etc. may even be a major risk factor for cancer predisposition. A recent study on academic professionals with sedentary lifestyles has reported higher ectopic fat deposition and increased risk of OMS [4]. Environmental factors like physical activity, diet, pollutants, stress, addictions, low family income, lack of education, screen time, family history, etc. contribute to the development of OMS which can be controlled greatly through lifestyle modifications [5].

Obesity is regulated by both genetic and epigenetic alterations [5]. The leptin-melanocortin pathway is one of the significant pathways associated with obesity. The mature adipocytes release leptin hormone encrypted by lep which in turn binds to the leptin receptor located on the pro-opiomelanocortin (POMC) neurons promoting the synthesis of α-melanocyte stimulating hormone (α-MSH) thereby accelerating the decrease in energy intake [6]. Lepr, POMC, and PRKAA2 are the genes encoding leptin receptor protein, α-MSH, and the catalytic regulatory subunit of AMP-activated kinase (AMPK) respectively, playing a pivotal role in this pathway [7]. Obesity regulation involves brain AMPK through coordinating feeding behavior, brown adipose tissue thermogenesis, insulin sensitivity, and browning of white adipose tissue [8]. The hypothalamic AMPK activity was observed to be downregulated by the administration of leptin leading to a decrease in feeding behavior [9]. Leptin, alternatively, binds with the leptin receptor to activate a different pathway involving the reduced secretion of Agouti-related protein (AgRP) and neuropeptide Y (NPY), a complex regulating hunger. During the starvation period, this AgRP/NPY complex gets up-regulated by the minimized leptin circulation in blood complemented by the increased ghrelin circulation [10]. AgRP in collaboration with POMC-generated hormones like α-MSH, β-MSH, γ-MSH, adrenocorticotrophin (ACTH), and their five G protein-coupled melanocortin receptors (MC1R-MC5R) has been reported to regulate obesity through energy homeostasis [11]. Loss of function of most MC4R variants leads to the development of obesity [12, 13]. Leptin receptors (LEPR) located in the arcuate nucleus of the hypothalamus once activated by leptin modulate the melanocortin pathway, stimulating the anorectic effects [14]. Lipid accumulation was observed in the intestines, plasma, and liver due to LEPR deficiency [11]. The circulating adiponectin concentration is inversely proportional to adiposity, insulin resistance, and hypertension [15]. Adiponectin acts through two different receptors, ADIPOR1 and ADIPOR2. In the liver, PPARα and AMPK get activated by the adiponectin bound to ADIPOR2 thus maintaining cellular energy homeostasis. In the skeletal muscle cells, ADIPOR1 signaling stimulates fatty acid oxidation and glucose metabolism [16]. Increased use of MUFA (monounsaturated fatty acid) instead of SFA (saturated fatty acid) in the diet significantly increased the availability of PPARγ-binding ligands and thereby increased the serum adiponectin concentration. The ADIPOQ polymorphism − 10066GG homozygotes in the Reading Imperial Surrey Cambridge King’s study supported the mentioned observation [17]. PPARγ interacting with AgRP and ACTH was also reported to regulate adipocyte differentiation [18]. The P12A variant of PPARγ when present alongside the V162L variant of PPARα in a Caucasian study population indicated a significant reduction in sdLDL (small dense low-density lipoprotein) cholesterol post-highMUFA-containing diet [17].

The single-base deletion, transversion, transition, or insertion in the genome that results in alteration of the encoded amino acid and present in more than 1% of the population is usually referred to as single nucleotide polymorphism (SNP). Transition and transversion are the most predominant modes of SNP generation [19]. The location of the SNPs on the genome classifies them into coding and non-coding SNPs. The non-coding SNPs account for the majority and include intergenic (iSNPs) and perigenic SNPs (pSNPs) [20]. The non-coding variants mostly provide a regulatory function often affecting the gene expression [21]. The coding SNPs or cSNPs may be either synonymous or non-synonymous. The non-synonymous cSNPs comprise non-sense SNPs coding the same amino acid and missense SNPs that result in the alteration of the encoded amino acid [22]. The missense cSNPs are responsible for the functional alteration of the corresponding protein and the development of genetic disease risk [23]. Advancements in research have introduced the prevalence of intronic SNPs in regulating gene expression through mis-splicing, alternative splicing, and several other mechanisms [24]. Intronic variants of ADIPOR1 have been studied and reported not to be associated with Type 2 Diabetes Mellitus (T2DM) in French Caucasian and African-American populations [25]. In contrast, insulin resistance was observed to be associated with intronic SNP rs75172865 in the Korean population [26]. In a study on the Chinese population, an intronic variant of ADIPOR1, rs3737884 was observed to be correlated with coronary artery disease besides T2DM [27]. In silico analyses conducted have identified variants in both non-coding and coding regions. ADIPOR1intronic variant rs7539542 has been reported to express lipid reduction in patients with colorectal cancer [28]. Non-synonymous SNPs including rs765425383, rs752071352, rs759555652, rs200326086, and rs766267373 have been reported to be deleterious although no structural alteration of the ADIPOR1 protein was caused by these polymorphic changes [29]. A missense variant, rs766665118, in MC4R, was detected from the Brazilian database as a potential pathogenic variant for obesity [30]. Multiple other non-synonymous SNPs of MC4R (E308K, P299L, D298H, C271F, C271R, P260L, T246N, G243R, C196Y, W174C, Y157S, D126Y, and D90G) were also observed to be causative agents of childhood obesity [31]. The potential non-synonymous missense SNPs associated with the development of obesity have been identified in this study using bioinformatics techniques like sequence-based analysis, machine learning, molecular modeling, and molecular dynamics. The in silico approach facilitates the early prediction of the genetic predisposition of the obesity-associated comorbidities and emerges as a pioneer for designing future therapeutic interventions.

The initiation of the study involved the accumulation of all the reported missense variants of our genes of interest. Subsequently, stepwise analyses included pathogenicity prediction analysis, and structural stability checking followed by the protein functionality analysis. In each step, the variations predicted to be tolerated or non-damaging were eliminated and the SNPs predicted to be deleterious or damaging were considered for further analyses (Fig. 1). The methodology has been partially adopted from [32] and further discussed in the following subsections.

2.1 Selection of the candidate genes and their associated variations

The study of the published articles available is guided through the procedure of candidate gene selection in the research design. Relevant search terms like ‘pathways for obesity’, ‘energy homeostasis’, ‘genes in adiposity’ etc. were used for proper screening of the research articles. The National Centre for Biotechnology Information (NCBI) [33] dbSNP database (version 156) was searched using the target gene symbol (provided by the Human Genome Organization [HUGO] gene nomenclature committee). The inclusion criteria for our study included human-specific genes and their missense variants. The non-coding transcript variants, synonymous, and intronic SNPs were excluded from the study.

2.2 Pathogenicity analysis

The pathogenicity prediction servers, SIFT (= 1) [34], PANTHER (≥ 0.85) [35], Meta-SNP (≥ 0.8) [36], PhD-SNP (≥ 0.9) [37], PredictSNP (≥ 85%) [38], PolyPhen-2 (= 1) [39], and SNAP2 (≥ 75%) [40] were run for the initial screening of the missense SNPs. The variants that were unanimously predicted as “pathogenic” (highest consensus agreement) by all the servers were then further checked for stability analysis. In the case of PolyPhen-2, HumVar dataset data was recorded.

SIFT is a popularly used sequence homology-based tool that predicts whether an amino acid substitution in a protein will modify its folding pattern. The change resulting in no significant functional alteration is scored as ‘tolerated’ or else it is ‘not tolerated’. Similarly, PANTHER (Protein Analysis Through Evolutionary Relationship), a web application for the classification of genes based on their functions, has a vast database of gene families. Analyzing multiple sequence alignments within the gene families, phylogenetic trees were constructed, and the impact of amino acid substitutions on the protein function was derived. PANTHER-derived results are generally indicated as ‘possibly damaging’, ‘probably damaging’, or ‘benign’ based on the position-specific evolutionary preservation (PSEP) method where the longer the variation is preserved, the more damaging effect it displays. The third tool, Polyphen 2 makes predictions based on machine learning from eight sequence-based (multiple sequence alignment or MSA) and three structure-based predictors. A scale from 0 to 1 is used to determine the damaging role of the variant. SNAP2 is a classifier working on neural networks. This tool predicts the impact (effect) of single amino acid substitutions on protein function. It automatically generates an MSA from the user-input FASTA sequence of a protein. It distinguishes between effect and neutral variants/non-synonymous SNPs by taking a variety of sequence and variant features into account. Structural features and annotations are also considered before predicting a change as either ‘neutral’ or ‘effect’. Of the remaining tools, PhD-SNP is an amino acid sequence-based SVM (support vector model) predictor, and MetaSNP is a random forest-based classifier building on PANTHER, PhD-SNP, SIFT, and SNAP. At the same time, PredictSNP is a neural network-based consensus classifier utilizing MAPP, PhD-SNP, PolyPhen-1, PolyPhen-2, SIFT, and SNAP. The amino acid sequence obtained from the Uniprot database [41] was used as the template in each server.

2.3 Structural stability analysis

The variants shortlisted based on their deleterious effects denoted by all the seven pathogenicity predicting tools were further analyzed for their structural stabilities. Change in free energy (δδG) of the protein structure can be of two types: destabilizing (-δδG) or stabilizing (-δδG). Depending on this free energy status, the folding pattern of the protein is determined.

DynaMut [42] measures the effect of mutation on the dynamics and stability of the protein conformation through normal mode analysis, along with graph-based signatures, and records the value of the free energy change. iMutant [43] uses SVM (support vector model) for predicting the same. In INPS3D [44], along with sequence descriptors being mapped to δδG change, two 3D structure-based descriptors (solvent accessibility and local energy change between native and mutated structures) are used to predict the stability. MuPro [45] uses both SVM and neural networks for prediction. iStable [46], on the other hand, is an integrated predictor, combining sequence data and outputs from other prediction software. This integrated information is then run through an SVM to generate the final prediction.

The protein configuration in the protein databank (PDB) format was initially extracted from the RCSB PDB [47]. SWISS-Model, a homology-modeling server, was also used to generate the protein models matching at least 90% of the sequence homology in case the structures were unavailable in RCSB PDB. For some proteins where the native structures were unavailable through RCSB PDB or could not be created through SWISS-Model, models were generated using the Robetta web server. Robetta uses the Rosetta macromolecular modeling suite and predicts 3D protein structures based on the amino acid sequence. It scans UniProt and Uniclust [48] (sequence databases) and the PDB for existing homologs. Robetta also uses deep learning to simultaneously analyze patterns in protein sequences, amino acid interactions, and a protein’s possible three-dimensional structures before making predictions. The most structurally destabilized variants disqualified in all five software used proceeded to the next step, the functionality analysis.

2.4 Functional stability analysis

Any structural change in a protein corresponds to the modified (gain in or loss of) functioning of said protein, as ‘function follows form’. Our study involved three functionality analyzing tools for the final screening of the risk variants.

Project HOPE [49] analyses the mutational effect on protein structure by collecting structural information from various sources, including calculations on the 3D protein structure, sequence annotations in UniProt, and predictions from DAS servers. The results are presented in a detailed, easy-to-interpret format that even non-bioinformaticians can understand.

Mut-Pred2 [50] is a machine learning-based method employing neural networks. Mut-Pred2 is a machine learning-based method employing neural networks. After analyzingthe molecular changes resulting from the amino acid substitutions, it gives a Posterior probability and a ‘P’ value for the probability. These indicators signify the impact of substitution on different structural and/or functional properties of the protein such as gain or loss of helical structures, intrinsic disorders, phosphorylation, and allosteric and catalytic sites. The threshold value of Mut-Pred2 was set as 0.90. Substitutions meeting the threshold value were considered as ‘most harmful’. These variants were further validated using the I-TASSER server [51].

I-TASSER (Iterative Threading Assembly Refinement) [52] (accessed on April- December2023) predicts 3D models and biological functions of proteins from input amino acid sequence. The 3D models are based on PDB templates and ab initio modeling while functions are predicted after matching with established libraries of known proteins with enzyme classification (EC) number, gene ontology (GO) vocabulary, and ligand-binding sites.

2.5 Gene network analysis

Protein-protein interactions including physical and functional associations were considered for the creation of the gene network. STRING [53] is a database that functions based on five aspects viz. genomic content, data extracted from laboratory experiments, conserved co-expressions, data mining, and data accumulation of other relevant databases. The final loci derived were used as input in the STRING for the gene network modeling.

3.1 Retrieval of SNP

Based on previous literature, a total of 9 genes having intricate involvement in the hypothalamic signaling to adipocyte differentiation process, were selected for the present study. Figure 2 depicts the number of SNPs in each gene corresponding to their functional consequences. Amongst, a total of 2,424 missense SNPs were taken for screening. The prediction statistics of each gene can be shown in Table 1.

Table 1

Summary of prediction outcomes of all tools used in this study
Genes	No. of nsSNPsanalysed	Pathogenicity Analysis (2424)							Protein Structure Stability Analysis (7)					Protein Function Analysis (3)
Genes	No. of nsSNPsanalysed	SIFT (NT) = 1	PANTHER (Dam.) ≥ 0.85	Meta SNP (Dis.) ≥ 0.8	PHD SNP (Dis.) ≥0.9	Polyphen 2 (Dam.) = 1	Predict SNP (Del.) ≥85%	SNAP2 (Eff.) ≥ 75%	Dyna Mut (Des.)	iMutant (Dec.)	INPS	MuPro	iStable	Project HOPE	MutPred	iTasser
ADIPOR1	209	48	27	20	17	16	16	15	2	2	2	2	2	2	2	2
ADIPOR2	279	67	19	28	19	0	28	22	0	0	0	0	0	0	0	0
AGRP	123	67	0	3	6	19	13	13	0	0	0	0	0	0	0	0
BDNF	241	58	7	2	1	31	16	22	0	0	0	0	0	0	0	0
MC4R	427	179	20	67	42	84	93	41	1	4	1	4	4	1	1	1
NPY	106	39	0	0	0	0	11	11	0	0	0	0	0	0	0	0
POMC	403	89	0	18	1	29	93	28	0	0	0	0	0	0	0	0
PRKAA2	433	65	46	22	16	93	40	37	1	1	0	1	1	0	0	0
TNF-α	203	18	6	0	0	16	21	5	0	0	0	0	0	0	0	0

3.2 Pathogenicity analysis

Pathogenicity screening was carried out on all 2,424 SNPs retrieved. After passing through 7 filters (without any cut-off), 362 (14.94%) variants were sorted out as “Pathogenic” in nature. However, only 7 (0.29%) of the total missense SNPs were listed as “highly pathogenic” when these filters were set to certain cut-off limits. Looking into the details, SIFT annotated 630 (25.99%) SNPs as ‘Not tolerated’, Panther annotated 125 (5.15%) as ‘Damaging’, MetaSNP annotated 160 (6.60%) as ‘Disease causing’, PhDSNP annotated 102 (4.20%) as ‘Disease’, Polyphen2 annotated 288 (11.88%) as ‘Damaging’, PredictSNP annotated 331 (13.65%) as ‘Deleterious’, SNAP2 annotated 194 (8%) as ‘Effect’.

3.3 Protein structure stability analysis

Protein thermal stability was performed on 7 SNPs that were considered highly pathogenic by nature. Only those SNPs that had destabilized the protein structures by all 5 filters were taken for further analysis to predict the functional effect of these SNPs on the proteins. Results showed that DYNAMUT identified 4 (~ 57%) SNPs as ‘Destabilizing’, iMutant, MuPro, and iStable predicted 7 (100%) mutations having ‘Decreased stability’, and INPS 3D identified 3 (~ 42%) SNPs having ‘Decrease Stability’. Overall, only 3 out of 7 SNPs were structurally ‘highly destabilized’ in nature. Henceforth, 2 SNPs from the ADIPOR1 gene (rs1419320091 and rs1654109863 ) and 1 SNP from the MC4R gene (rs1159323398) were allowed for protein functionality analysis.

3.4 Protein Functionality Analysis

Protein functionality was checked for selected 3 SNPs using 3 different most popular tools, i.e., HOPE, MutPred, and I-tasser (Table 2).

Table 2: Protein functionality analysis of 3 most pathogenic variants

SNP_ID

Protein details

Project Hope

I-tasser

MutPred2

Site

WT

MT

Size

Hydro-phobicity

Charge

Conservation

Location

LB

site

Enzyme active site

WT

MT

WT

MT

ADIPOR1 (rs1419320091)

107

N

S

MT<WT

MT>WT

No change

Very conserved

Near highly conserved

Adipor/Haemolysin-III-related domain core (can disturb the core structure of the domain)

NC

AlteredMetal binding (Pr = 0.28 | P = 6.0e-03);

Lossof Relative solvent accessibility (Pr = 0.28 | P = 0.02); AlteredTransmembrane protein (Pr = 0.12 | P = 0.03);

Gainof Sulfation at Y109 (Pr = 0.05 | P = 7.4e-03)

ADIPOR1 (rs1654109863)

333

W

C

MT<WT

Not reported

Very conserved

Near highly conserved

Adipor/Haemolysin-III-related domain surface (can affect molecules or other domains in contact with the surface)

NC

Lossof Strand (Pr = 0.26 | P = 0.04);

AlteredMetal binding (Pr = 0.25 | P = 0.03);

Lossof Catalytic site at S336 (Pr = 0.23 | P = 7.5e-03);

Gainof Allosteric site at H337 (Pr = 0.21 | P = 0.03)

MC4R (rs1159323398)

174

W

C

MT<WT

Not reported

G Protein-Coupled Receptor, Rhodopsin-like (can affect the correct function of the protein by altering the interaction with other domains)

Alter

NC

Gainof Helix (Pr = 0.35 | P = 5.3e-04);

AlteredOrdered interface (Pr = 0.32 | P = 3.4e-03);

Lossof Strand (Pr = 0.31 | P = 1.1e-03);

AlteredTransmembrane protein (Pr = 0.19 | P = 6.3e-03)

WT: Wild Type; MT: Mutant Type; LB: Ligand Binding, NC: No Change

3.5 Gene-network Analysis

All the 9 genes were used for protein-protein interaction analysis, which further added 100 proteins. The protein-protein enrichment of 109 proteins was highly significant (Fig. 3). In this network, 51 cellular components, 79 molecular functions, and 733 biological processes were enriched based on the Gene Ontology (GO) of the whole genome. Apart from these, this network also predicted 10 genes (MC3R, NGFR, NFKB1, TRAF2, TNFRSF1A, PRKAB2, RIPK1, NTRK2, APPL1, and PRKAG1) that are functionally associated with 9 input genes (Supplementary Table 1).

Systematic analysis of the SNP density and distribution in the human genome is yet to be reported. The variants in the protein-coding sequence particularly the non-synonymous missense polymorphisms are more likely to be damaging [54]. Studies on obesity-related SNPs have not considered multiple gene networking. The structural and functional alterations of the proteins encoded by the candidate genes selected in the current study and their impact on disease predisposition have been our specific focus. In our previous study [32], we considered the missense variants of a few candidate genes in the energy homeostasis pathway including LEP, LEPR, ADIPOQ, and PPARG, etc.The pathogenicity prediction and structural and functional alteration screening identified the variant rs757012788 (LEPR) among the others, to be a probable cause of disturbing the domain and resulting in a loss of function of the protein. In the current study, the genes were targeted as an extension of the relevant pathways.

The STRING analysis in our study provided a gene-network pattern (Fig. 3) where our selected genes of interest were observed to be associated further with a significant number of genes of associated pathways (Supplementary Table 1). The adiponectin signaling cascade (Fig. 4) [55] occurs through the binding of APPL1, an adaptor protein, to ADIPOR1 and ADIPOR2 present in the skeletal muscle and hepatic cells [56]. The pathway regulated through ADIPOR1 involves increased AMPK activity [56] while the one activated through ADIPOR2 leads to PPARα proliferation and increased insulin sensitivity [57] The α and β subunits of the AMPK, the heterotrimeric enzyme is encoded by PRKAA1/2 and PRKAB1/2 respectively, whereas, the γ subunit is encoded by PRKAG1/2/3. The involvement of PRKAB2 along with PRKAA2 has gained significance in lipid and glucose metabolism [58]. PRKAG2 has been widely associated with cardiomyopathy [59]. PRKAG1, on the other hand, is distributed throughout the organs and its association with obesity is yet to be concluded. Adiponectin, as observed in certain in vitro studies, also inhibits the TNF-α induced NF-κβ activation, in turn, leading to reduced inflammatory responses of the endothelial cells [60]. The subunit 1 of the NF-κβ protein complex is encoded by NFKB1 [61]. The transcription factor of the NF-κβ is known to be activated by receptor-interacting serine/threonine-protein kinase 1 (RIPK1) promoting cellular apoptosis [62]. Genetic variants in the RIPK1 locus have been reported to be associated with higher expression levels of RIPK1 in human adipose tissue [63]. Moreover, the expression of tumor necrosis factor (TNF) receptor superfamily member 1A (TNFRSF1A) has been downregulated with uncontrolled diet and increased obesity [64]. Interleukin-6 (IL-6), a pro-inflammatory cytokine, gets upregulated as a consequence of reduced adiponectin synthesis and NF-κβ activation [65]. Tumor necrosis factor receptor–associated factor (TRAF2) has been reported to be a key mediator of the pro-inflammatory cytokine signaling and also regulating hepatic glucose metabolism [66]. In 2020, the U.S. Food and Drug Administration identified MC4R agonist, an activated moiety of POMC as responsible for LEPR or POMC deficiency weight management [67]. Most of the studies have considered MC4R as the major receptor of melanocortin playing a significant role in energy homeostasis. In contrast, MC3R also acts as the co-player in controlling obesity [68]. AGRP was shown to be expressed in the hypothalamus as a natural antagonist of MC4R and MC3R. The balance of the orexigenic AGRP and the anorexigenic melanocortin signaling along with the regulation of the energy expenditure is sensed by the MC4R neurons [69]. Mutations in the genes encoding BDNF and its receptor TrkB (encoded by NTRK2) lead to severe obesity in humans [70]. BDNF signaling through TrkB, MAPK, and PI3K promotes the synaptic plasticity of the neurons regulating the feeding and satiety [71]. Apart from the gene cascade selected for the current study, a significant cluster of related genes was observed through a 100-interactor network generated. Some of these genes are involved in neurodegenerative and cardiometabolic pathways. The specific functions of these genes in energy homeostasis are yet to be extensively studied.

The inclusion criteria in our study involved the analysis of only non-synonymous missense variants. Protein coding was observed to be directly influenced by our screened novel variants, rs1419320091, and rs1654109863. The structural modification of the protein due to rs1419320091 predicted the altered core of the protein leading to loss of relative solvent accessibility and altered transmembrane protein with differential metal binding. In the case of rs1654109863, the polymorphism resulted in the altered surface structure of the protein involving loss of strand and catalytic site with gain of allosteric site. Thereby, the surface binding of the ADIPOR1 protein gets modified due to this pathogenic variant. The polymorphism observed in MC4R, rs1159323398 as also reported by [72, 31], leads to variable interactions with other domains, coupled with a gain of helix, loss of strand, and altered transmembrane protein structure. In both ADIPOR1 and MC4R, the protein functions get altered with the observed variants.

Validation of our findings in population-based studies is one of the significant limitations of our study. The heterozygosities of the observed SNPs are reported to be negligible in the Southeast Asian population and, therefore can be considered as ‘rare variants’ [rs1419320091: T = 1, C = 0; rs1654109863: C = 1, G = 0; rs1159323398: C = 1, G = 0]. Another important drawback of the study is the inclusion of only the non-synonymous missense variants. The overall scenario would have been more prominent with the consideration of the intronic and variants in the promoter regions. Moreover, the drug interaction patterns of the altered domains were also not analyzed in the current study.

The pathogenic probability of the identified variants would be considered as a future tool for the early prediction of the disease manifestation. The designing of a therapeutic approachbased on the predicted structural changes of the protein domains is one of the major concerns for future research. The research prospects, hence, involve the probability of disease prediction through the genomic analysis of the proband and effective therapeuticstrategies.

Author’s contribution: SG designed the study, composed the figures, and drafted the manuscript. SD composed the tables and contributed to result analysis and manuscript writing. UR contributed to data extraction and manuscript writing. SM contributed to data extraction. PB supervised the work and also gave necessary inputs. All authors have contributed to the article.

Data availability: The datasets generated during and/or analyzed during the current study are available from the corresponding author upon reasonable request.

Competing Interest: The authors declare that they have no competing interests.

Ethics Approval Statement: The study did not involve any ethics approval requirements.

Funding Statement: Not applicable.

Emily, A. C. Current status and response to the global obesity pandemic: proceedings of a workshop in National Academies of Sciences, Engineering, and Medicine. ISBN: 978-0-309-48505-0; 10.17226/25273 (2019).
Kakkar, A. K., & Dahiya, N. Drug treatment of obesity: current status and future prospects. European Journal of Internal Medicine, 26(2), 89-94 (2015).
Malik, V. S., Willett, W. C., & Hu, F. B. Global obesity: trends, risk factors and policy implications. Nature reviews endocrinology, 9(1), 13-27 (2013).
Ghosh, S., Paul, M., Mondal, K. K., Bhattacharjee, S., & Bhattacharjee, P. Sedentary lifestyle with increased risk of obesity in urban adult academic professionals: an epidemiological study in West Bengal, India. Scientific Reports, 13(1), 4895 (2023).
Ghosh, S., Dhar, S., Bhattacharjee, S., & Bhattacharjee, P. Contribution of environmental, genetic and epigenetic factors to obesity-related metabolic syndrome. The Nucleus, 1-23 (2023).
Andermann, M. L., & Lowell, B. B. Toward a wiring diagram understanding of appetite control. Neuron, 95(4), 757-778 (2017).
Srour, N., Caron, A., & Michael, N. J. Do POMC neurons have a sweet tooth for leptin? Special issue: Role of nutrients in nervous control of energy balance. Biochimie; 10.1016/j.biochi.2022.09.006 (2022).
Lopez, M., Nogueiras, R., Tena-Sempere, M., & Dieguez, C. Hypothalamic AMPK: a canonical regulator of whole-body energy balance. Nature Reviews Endocrinology, 12(7), 421-432 (2016).
Andersson, U. et al. AMP-activated protein kinase plays a role in the control of food intake. Journal of Biological Chemistry, 279(13), 12005-12008 (2004).
Baldini, G., & Phelan, K. D. The melanocortin pathway and control of appetite-progress and therapeutic implications. Journal of Endocrinology, 241(1), R1-R33 (2019).
Wen, X. et al. Signaling pathways in obesity: mechanisms and therapeutic interventions. Signal Transduction and Targeted Therapy, 7(1), 298 (2022).
Farooqi, I. S. et al. Clinical spectrum of obesity and mutations in the melanocortin 4 receptor gene. New England Journal of Medicine, 348(12), 1085-1095 (2003).
Lotta, L. A. et al. S. Human gain-of-function MC4R variants show signaling bias and protect against obesity. Cell, 177(3), 597-607 (2019).
Cowley, M. A. et al. Leptin activates anorexigenic POMC neurons through a neural network in the arcuate nucleus. Nature, 411(6836), 480-484 (2001).
Brochu-Gaudreau, K. et al. Adiponectin action from head to toe. Endocrine, 37, 11-32 (2010).
Kadowaki, T., & Yamauchi, T. Adiponectin and adiponectin receptors. Endocrine reviews, 26(3), 439-451 (2005).
AlSaleh, A., Sanders, T. A., & O'Dell, S. D. Effect of interaction between PPARG, PPARA and ADIPOQ gene variants and dietary fatty acids on plasma lipid profile and adiponectin concentration in a large intervention study. Proceedings of the Nutrition Society, 71(1), 141-153 (2012).
Mynatt, R. L., & Stephens, J. M. Regulation of PPARγ and obesity by agouti/melanocortin signaling in adipocytes. Annals of the New York Academy of Sciences, 994(1), 141-146 (2003).
Butler, J. M. Single nucleotide polymorphisms and applications. Advanced Topics in Forensic DNA Typing, 347-369 (2012).
Vallejos-Vidal, E. et al. Single-nucleotide polymorphisms (SNP) mining and their effect on the tridimensional protein structure prediction in a set of immunity-related expressed sequence tags (EST) in Atlantic salmon (Salmo salar). Frontiers in genetics, 10, 1406 (2020).
Zhang, F., &Lupski, J. R. Non-coding genetic variants in human disease. Human molecular genetics, 24(R1), R102-R110 (2015).
Yates, C. M., & Sternberg, M. J. The effects of non-synonymous single nucleotide polymorphisms (nsSNPs) on protein–protein interactions. Journal of molecular biology, 425(21), 3949-3963 (2013).
Yang, S., Gill, R. A., Zaman, Q. U., Ulhassan, Z., & Zhou, W. Insights on SNP types, detection methods and their utilization in Brassica species: Recent progress and future perspectives. Journal of Biotechnology, 324, 11-20 (2020).
Salih, M. H., Al-Azzawie, A. F., & Al-Assie, A. H. A. Intronic SNPs and Genetic Diseases: A Review. International Journal for Research in Applied Sciences and Biotechnology, 8(2), 267-274 (2021).
Vaxillaire, M. et al. Genetic analysis of ADIPOR1 and ADIPOR2 candidate polymorphisms for type 2 diabetes in the Caucasian population. Diabetes, 55(3), 856-861 (2006).
Kim, J. T. et al. Polymorphisms of ADIPOR1 and ADIPOR2 are associated with phenotypes of type 2 diabetes in Koreans. Clinical endocrinology, 70(1), 66-74 (2009).
Jin, Z. et al. Identification of susceptibility variants in ADIPOR1 gene associated with type 2 diabetes, coronary artery disease and the comorbidity of type 2 diabetes and coronary artery disease. PLoS One, 9(6), e100339 (2014).
Mihajlović, M. et al. Association of adiponectin receptors with metabolic and immune homeostasis parameters in colorectal cancer: In silico analysis and observational findings. International Journal of Environmental Research and Public Health, 19(22), 14995 (2022).
Saleh, M. A. et al. Impacts of nonsynonymous single nucleotide polymorphisms of adiponectin receptor 1 gene on corresponding protein stability: a computational approach. BioMed research international, 2016, 1-12 (2016).
Bratti, L. D. O. S., Nunes, B. F., Gorges, D. M., &Filippin-Monteiro, F. B. In silico approach to identify non-synonymous missense variants in human obesity-related genes: Comprehensive analyses in variants reported in Brazilian databases. Human Gene, 36, 201174 (2023).
Douiyeh, I. et al. Computational analysis of missense variants of human MC4R and childhood obesity. Cellular and Molecular Biology, 69(10), 30-42 (2023).
Dhar, S., Mridha, S., & Bhattacharjee, P. Mutational landscape screening through comprehensive in silico analysis for polycystic ovarian syndrome–related genes. Reproductive Sciences, 29(2), 480-496 (2022).
https://www.ncbi.nlm.nih.gov/snp/
Ng, P. C., &Henikoff, S. SIFT: Predicting amino acid changes that affect protein function. Nucleic acids research, 31(13), 3812-3814 (2003).
Tang, H., & Thomas, P. D. PANTHER-PSEP: predicting disease-causing genetic variants using position-specific evolutionary preservation. Bioinformatics, 32(14), 2230-2232 (2016).
Capriotti, E., Altman, R. B., & Bromberg, Y. Collective judgment predicts disease-associated single nucleotide variants. BMC genomics, 14, 1-9 (2013).
Calabrese, R., Capriotti, E., &Casadio, R. PhD-SNP: a web server for the prediction of human genetic diseases associated to missense single nucleotide polymorphisms. In EMBNET 08 78-78 (2008).
Bendl, J. et al. PredictSNP: robust and accurate consensus classifier for prediction of disease-related mutations. PLoS computational biology, 10(1), e1003440 (2014).
Adzhubei, I. A., Schmidt, S., Peshkin, L., Ramensky, V. E., Gerasimova, A., Bork, P., ... &Sunyaev, S. R. A method and server for predicting damaging missense mutations. Nature methods, 7(4), 248-249 (2010).
Hecht, M., Bromberg, Y., & Rost, B. Better prediction of functional effects for sequence variants. BMC genomics, 16(8), 1-12 (2015).
https://www.uniprot.org/
Rodrigues, C. H., Pires, D. E., & Ascher, D. B. DynaMut: predicting the impact of mutations on protein conformation, flexibility and stability. Nucleic acids research, 46(W1), W350-W355 (2018).
Capriotti, E., Fariselli, P., &Casadio, R. I-Mutant2. 0: predicting stability changes upon mutation from the protein sequence or structure. Nucleic acids research, 33(suppl_2), W306-W310 (2005).
Savojardo, C., Fariselli, P., Martelli, P. L., &Casadio, R. INPS-MD: a web server to predict stability of protein variants from sequence and structure. Bioinformatics, 32(16), 2542-2544 (2016).
Cheng, J., Randall, A., & Baldi, P. Prediction of protein stability changes for single‐site mutations using support vector machines. Proteins: Structure, Function, and Bioinformatics, 62(4), 1125-1132 (2006).
Chen, C. W., Lin, M. H., Liao, C. C., Chang, H. P., & Chu, Y. W. iStable 2.0: predicting protein thermal stability changes by integrating various characteristic modules. Computational and structural biotechnology journal, 18, 622-630 (2020).
https://www.rcsb.org/
https://uniclust.mmseqs.com/
Dunlavy, D. M., O'leary, D. P., Klimov, D., & Thirumalai, D. HOPE: A homotopy optimization method for protein structure prediction. Journal of Computational Biology, 12(10), 1275-1288 (2005).
Pejaver, V. et al. Inferring the molecular and phenotypic impact of amino acid variants with MutPred2. Nature communications, 11(1), 5918 (2020).
Roy, A., Kucukural, A., & Zhang, Y. I-TASSER: a unified platform for automated protein structure and function prediction. Nature protocols, 5(4), 725-738 (2010).
https://zhanglab.ccmb.med.umich.edu/I-TASSER/
https://string-db.org/
Zhao, Z., Fu, Y. X., Hewett-Emmett, D., & Boerwinkle, E. Investigating single nucleotide polymorphism (SNP) density in the human genome and its implications for molecular evolution. Gene, 312, 207-213 (2003).
https://www.biorender.com/
Luo, N. et al. Enhanced adiponectin actions by overexpression of adiponectin receptor 1 in macrophages. Atherosclerosis, 228(1), 124-135 (2013).
Yamauchi, T., & Kadowaki, T. Adiponectin receptor as a key player in healthy longevity and obesity-related diseases. Cell metabolism, 17(2), 185-196 (2013).
Randrianarisoa, E. et al. AMPK subunits harbor largely nonoverlapping genetic determinants for body fat mass, glucose metabolism, and cholesterol metabolism. The Journal of Clinical Endocrinology & Metabolism, 105(1), 14-25 (2020).
Komurcu-Bayrak, E., Kalkan, M. A., Coban, N., Ozsait-Selcuk, B., & Bayrak, F. Identification of the pathogenic effects of missense variants causing PRKAG2 cardiomyopathy. Archives of Biochemistry and Biophysics, 727, 109340 (2022).
de Oliveira Leal, V., & Mafra, D. Adipokines in obesity. ClinicaChimica Acta, 419, 87-94 (2013).
Tessier, F., Fontaine-Bisson, B., Lefebvre, J. F., El-Sohemy, A., & Roy-Gagnon, M. H. Investigating Gene–Gene and Gene–Environment Interactions in the Association Between Overnutrition and Obesity-Related Phenotypes. Frontiers in genetics, 10, 151 (2019).
Silke, J., Rickard, J. A., &Gerlic, M. The diverse role of RIP kinases in necroptosis and inflammation. Nature immunology, 16(7), 689-697 (2015).
Karunakaran, D. et al. RIPK1 gene variants associate with obesity in humans and can be therapeutically silenced to reduce obesity in mice. Nature metabolism, 2(10), 1113-1125 (2020).
Ulven, S. M. et al. An isocaloric nordic diet modulates RELA and TNFRSF1A gene expression in peripheral blood mononuclear cells in individuals with metabolic syndrome—a SYSDIET sub-study. Nutrients, 11(12), 2932 (2019).
Bray, G. A., Clearfield, M. B., Fintel, D. J., &Nelinson, D. S. Overweight and obesity: the pathogenesis of cardiometabolic risk. Clinical cornerstone, 9(4), 30-42 (2009).
Chen, Z. et al. Hepatic TRAF2 regulates glucose metabolism through enhancing glucagon responses. Diabetes, 61(3), 566-573 (2012).
Clément, K. et al. Efficacy and safety of setmelanotide, an MC4R agonist, in individuals with severe obesity due to LEPR or POMC deficiency: single-arm, open-label, multicentre, phase 3 trials. The lancet Diabetes & endocrinology, 8(12), 960-970 (2020).
Yang, Z., & Tao, Y. X. Mutations in melanocortin-3 receptor gene and human obesity. Progress in molecular biology and translational science, 140, 97-129 (2016).
Yeo, G. S. et al. The melanocortin pathway and energy homeostasis: From discovery to obesity therapy. Molecular metabolism, 48, 101206 (2021).
Xu, B., & Xie, X. Neurotrophic factor control of satiety and body weight. Nature Reviews Neuroscience, 17(5), 282-292 (2016).
Cordeira, J., & Rios, M. Weighing in the role of BDNF in the central control of eating behavior. Molecular neurobiology, 44, 441-448 (2011).
Buono, P. et al. Six novel mutations in the proopiomelanocortin and melanocortin receptor 4 genes in severely obese adults living in southern Italy. Clinical chemistry, 51(8), 1358-1364 (2005).

No competing interests reported.

SupplementaryMaterial.pdf

Predictive Analysis of the Leptin-Melanocortin and Adiponectin Signaling Pathways in Obesity through In Silico Techniques

Status:

Version 1

Abstract

Figures

1. Introduction

2. Methodology

2.1 Selection of the candidate genes and their associated variations

2.2 Pathogenicity analysis

2.3 Structural stability analysis

2.4 Functional stability analysis

2.5 Gene network analysis

3. Results

3.1 Retrieval of SNP

3.2 Pathogenicity analysis

3.3 Protein structure stability analysis

3.4 Protein Functionality Analysis

3.5 Gene-network Analysis

Discussion

Declarations

References

Additional Declarations

Supplementary Files

Status:

Version 1