Energy metabolism and whole-exome sequencing-based analysis of Sasang constitution: a pilot study

Abstract

Background

Traditional Korean Sasang constitutional (SC) medicine categorizes individuals into four constitutional types [Tae-eum (TE), So-eum (SE), Tae-yang (TY), or So-yang (SY)] based on biological and physiological characteristics. As these characteristics are closely related to the bioenergetics of the human body, we assessed the correlation between SC type and energy metabolism features.

Methods

Forty healthy, young (22.3 ± 1.4 years) males volunteered to participate in this study. Participants answered an SC questionnaire, and their face shape, voice tone, and body shape were assessed using an SC analysis tool. Thirty-one participants (10 TE, 10 SE, 3 TY, and 8 SY) were selected for further analysis. Collected blood samples were subjected to blood composition analysis, mitochondrial function analysis, and whole-exome sequencing.

Results

The SY type showed significantly lower total cholesterol and high-density lipoprotein cholesterol levels than the SE type. Cellular and mitochondrial Adenosine triphosphate (ATP) levels were similar across types. All types showed similar basal mitochondrial oxygen consumption rates, whereas the TE type showed a significantly lower ATP-linked oxygen consumption rate than the other types. Whole-exome sequencing identified several genes variants that were exclusively detected in particular SC types, including 19 for SE, seven for SY, 11 for TE, and six for TY.

Conclusion

SC type-specific differences in mitochondrial function and gene mutations were detected in a small group of healthy, young Korean males. These results are expected to greatly improve the accurate screening and utilization of SC medicine.

1. Introduction

Sasang constitutional medicine (SCM) is a Korean medical tradition that classifies physiological and pathological traits into four constitutional types: Tae-yang (TY), So-yang (SY), Tae-eum (TE), and So-eum (SE). The concept that constitution can be “typed” is the most basic underlying paradigm described in the Donguisusebowon.¹ Unlike Westernized diagnostic tools based on molecular biological evidence, SCM emphasizes integrative and holistic characteristics of the individual.¹ SCM is not only used for the clinical diagnosis of individual constitution but is also widely used in the SC type-specific treatment of disease.² Thus, the SCM tradition can be considered as a complement to the current “personalized medicine” approach.3, 4

SCM considers balances between food intake and waste discharge, energy consumption and storage, and catabolism and anabolism.² TY, SY, TE, and SE types display different hyper- and hypoactive organs. For example, the TE type shows a tendency toward a hyperactive liver and hypoactive lung, whereas the TY type shows the opposite tendency. In the same manner, the SY and SE types also show opposite tendencies toward hyper- and hypoactive organs; the SY type has a hyperactive spleen and hypoactive kidney, whereas the SE type has a hypoactive spleen and hyperactive kidney.² Beyond single organ-specific characteristics, SCM is deeply concerned with various biological processes such as food intake, digestion, waste excretion, and enidney.² Beyond single organ-specific characteristics, SCM is deeply concerned with various biological processes such as food intake, digestion, waste excretion, and energy storage in multiple organs as well as catabolism and anabolism balance at the cellular level. The detailed biological characteristics and bodily features of different SC types have previously been described.² According to the SCM theory, the imbalance of energy metabolism under pathological conditions affects the sensitive hypoactive organ of each type, which can cause disease.² Because SC types are reflected by integrative systemic features and are closely associated with metabolic status, SCM can facilitate the diagnosis and treatment of metabolic syndromes or diseases including obesity, hyperlipidemia, diabetes, and hypertension.5, 6, 7, 8, 9, 10 These previous studies provide strong evidence of correlations between SC types and particular diseases, suggesting the value of SC types as holistic biomarkers for a wide range of diseases.

SC types are determined by professional oriental medical doctors based on traditional diagnostic methods including seeing, listening, questioning, and touching. However, these traditional diagnostic methods are based on subjective observations and are catechetic methods, making them difficult to objectify and quantify.¹ To overcome difficulties in traditional diagnosis and to establish an evidence-based diagnostic tool, we developed the Sasang constitution analysis tool (SCAT),¹¹ a system designed to provide objective information for determining the SC type. In addition to the SCAT, several genomic approaches have been used to identify the genetic loci or pathways responsible for different SC types.12, 13 Won et al¹² found a significant link between the constitution and chromosomes 8q11.22–23 and 11q22.1–3 based on a genome-wide scan of a Korean family. Kim et al¹³ not only performed a genome-wide association analysis but also analyzed the pathways involved in SC. These studies, therefore, suggest the possibility of genomic differences among SC types.

Next-generation sequencing is a novel and powerful genomics tool that can be used to rapidly sequence whole or specific regions of an individual human genome and provide integrative genomic information about the individual.¹⁴ Whole-exome sequencing (WES) is the most widely used targeted sequencing method, as the exome contains the majority of known disease-causing variants. WES enables the selective capture and sequencing of the protein-coding portion of the genome to understand relationships between gene variants and their associated phenotypes.15, 16, 17 Based on this advantage, we aimed to use WES to identify gene variants of SC types and to link this genomic information to the morphological and physiological characteristics of each SC type.¹⁸

In the bodies of mammals, intracellular organelles, called the mitochondria, play an essential role in the production, storage, and transformation of biological energy molecules, known as Adenosine triphosphate (ATP), and the regulation of catabolic and anabolic pathways.19, 20, 21, 22 Impaired mitochondrial function results in an imbalance of energy metabolism and an increase in oxidative stress in major organs, including the heart, liver, kidney, and brain, which cause a wide range of diseases including inflammatory diseases, neurodegenerative diseases, metabolic syndromes, cancers, and cardiovascular diseases.23, 24, 25, 26, 27, 28 Recent studies by Shim et al29, 30 show that the TE type has reduced mitochondrial metabolism and an obesity-prone tendency. Therefore, we hypothesized that variations in mitochondrial energy metabolism are associated with different SC types and phenotypic characteristics. The aim of this study was to identify SC type-specific energy metabolic and/or genetic variants that can be used for the precise diagnosis of SC type. We measured the energy status-related metabolites in blood samples, as well as mitochondrial oxidative phosphorylation, and identified SC type-specific gene variants in 31 healthy individuals.

2. Methods

2.1. Participants

This study was approved by the Institutional Review Board and Ethics Committee of Inje University Paik Hospital, Busan, Korea (15-0287). The study was performed in accordance with previously established experimental protocols and guidelines. Informed consent was obtained from all participants. Forty healthy males were recruited through posters placed on community boards in the College of Medicine, Inje University. The study included 31 participants after excluding nine individuals for the following reasons: (1) diagnosis of diabetes, hypertension, hyperlipidemia, or other chronic disease, (2) drug supplementation within 3 months of the study, (3) results of initial screening interviews, or (4) SC type diagnosis based on SCAT score and the opinion of a medical doctor with expertise in SCM.

2.2. SCAT

The SCAT was used to diagnose the SC type of 40 males at the College of Medicine, Inje University. The diagnosis of constitution was based on face pictures, voice recordings, body measurements, and responses to survey questions.¹¹ First, front- and side-view pictures of the face were taken. Second, an audio recording of the participant, making a specific verbal statement, was made. Third, circumference measurements of the eight parts of the body were obtained. Finally, individual responses to an SC structure classification survey were recorded. SCAT scoring results were confirmed by a medical doctor with expertise in SCM. We only included individuals with a difference ratio of at least 5% between the top- and second-ranked SC types. We included 10 SE, eight SY, 10 TE, and three TY participants (Fig. 1).

Fig. 1.

Flow diagram of SC type diagnosis.

IRB, institutional review board; SCM, Sasang constitutional medicine; SE, So-eum; SY, So-yang; TE, Tae-eum; TY, Tae-yang.

2.3. Blood collection and composition analysis

Blood samples were collected into serum-separating tubes or EDTA-containing anticoagulation tubes under fasting conditions for composition analysis or WES analysis, respectively. Blood samples were allowed to clot at room temperature for 30 minutes and centrifuged at 3000 g for 15 minutes. The separated serum was stored at −80 °C until further analysis. Serum glucose, insulin, total cholesterol (TC), triglyceride (TG), low-density lipoprotein cholesterol (LDL-C), and high-density lipoprotein cholesterol (HDL-C) were determined using enzymatic techniques based on a colorimetric assay as previously described.³¹ Blood samples in EDTA tubes were kept at −4 °C until WES analysis.

2.4. Mitochondrial function analysis

2.4.1. Measurement of cellular and mitochondrial ATP levels

Mitochondrial ATP levels were measured by Mitochondrial ToxGlo assay (Promega, Madison, WI, USA) according to the manufacturer’s protocol. Briefly, isolated human platelets were plated at 1 × 10⁴ cells/well in a white and clear-bottomed 96-well culture plate. Plates were centrifuged at 200 g for 10 minutes to remove the medium. The medium was replaced with 50 μL fresh medium lacking glucose and supplemented with 10 mM glucose (cellular ATP) or 10 mM galactose (mitochondrial ATP). The plate was incubated at 37 °C in a humidified and CO₂-supplemented incubator for 90 minutes. The assay solution (100 μL) was added to the plate, which was incubated at room temperature for 30 minutes. Luminescence was measured using a luminometer (Molecular Device, Sunnyvale, CA, USA).

2.4.2. Oxygen consumption analysis

Oxygen consumption rate (OCR) was measured using an XF24 analyzer (Seahorse Bioscience, Billerica, MA, USA) as previously described.32, 33 Briefly, isolated human platelets were plated at 2 × 10⁴ cells/well in a Corning Cell-TAK cell and tissue adhesive (Corning Inc., Corning, NY, USA)-treated XF24 cell culture plate (Seahorse Bioscience, Billerica, MA, USA). Plates were centrifuged at 200 g for 10 minutes to remove the medium. The medium was replaced with 500 μL XF Assay medium-modified Dulbecco’s modified Eagle’s medium (Seahorse Bioscience, Billerica, MA, USA) and incubated at 37 °C without CO₂ for 1 hour. OCR was measured using an XF24 analyzer and software. We measured basal OCR, ATP-linked OCR, maximal OCR, and spare respiration capacity using specific mitochondrial inhibitors including oligomycin (ATPase inhibitor for ATP-linked OCR) and carbonyl cyanide p-[trifluoromethoxy]-phenyl-hydrazone (FCCP; uncouples the mitochondrial inner membrane and allows maximum electron flux through the electron transport chain). Basal OCR was measured without any mitochondrial inhibitor. ATP-linked OCR was calculated as the difference in OCR prior to and after treatment with oligomycin. Spare respiration capacity was calculated as the difference between maximal and basal OCR.³²

2.5. WES and variant analyses

WES and variant analyses were performed using previously described methods.¹⁵ An Illumina Hiseq2500 machine and Sureselected Exome V5 kit were used as a sequencing platform. On average, 6.25 gigabases of raw sequences were used for the analyses. Quality control was performed using FastQC,³⁴ and raw sequence reads were aligned to a reference genome (NCBI b37) using the Burrows–Wheeler Aligner-MEM algorithm.³⁵ The initial alignment was refined by local realignment and base quality recalibration using GATK tools.³⁶ Variants were called by the GATK HaplotypeCaller and filtered by the GATK Variant Recalibrator walker.¹⁶ These variants were annotated by SnpEff and filtered by SnpSift.³⁷

2.6. Gene set enrichment analyses

We obtained gene-to-function annotation from a gene ontology (GO) database³⁸ (Homo sapiens, “go.obo”, June 2016). Categories used for analyses were “biological process,” “molecular function,” and “cellular component.” Biological pathway information was obtained from the GSEA (Gene Set Enrichment Analysis) database (http://software.broadinstitute.org/gsea), which contains the BIOCRATA, KEGG (Kyoto Encyclopedia of Genes and Genomes), and REACTOME pathways.³⁹ Disease phenotypes were obtained from the Menche et al⁴⁰ dataset containing OMIM (Online Mendelian Inheritance in Man) disease and GWAS (genome-wide association study) information. We estimated the enrichment p values using hypergeometric tests to determine which biological modules were enriched in SC type-specific variants. Hypergeometric p values were obtained by comparing the fraction of genes included in the specific modules within the whole genome and those in the corresponding modules within SC type-specific variants (enriched p values: <0.01).

2.7. Statistical analysis

Data are presented as mean ± standard error of the mean. Statistical analyses were performed using IBM SPSS Statistics for Windows, version 23.0 (IBM Corp., Armonk, NY, USA). One-way analysis of variance and Duncan post hoc test were used to determine differences between the groups. Statistical significance was set at p < 0.05. Because the sample size of the TY type was very small (n = 3), we eliminated TY values from the mean comparison tests.

3. Results

3.1. Participant characteristics

Participant characteristics, including age, height, body weight, and body mass index (BMI), are shown in Table 1. Age and height were not significantly different among the CS types. Body weight and BMI for the TE type were significantly higher than those for the other types, which was consistent with previous studies.41, 42

Table 1.

Participant characteristics.

	SE (n = 10)	SY (n = 8)	TE (n = 10)	TY (n = 3)	F
Age (y)	22.50 ± 0.45	21.62 ± 0.26	22.80 ± 0.64	21.66 ± 0.33	1.345
Height (cm)	172.10 ± 1.50	176.37 ± 1.13	174.70 ± 0.86	176.00 ± 2.64	3.076
Body weight (kg)	60.20 ± 1.91a	70.50 ± 1.08b	82.40 ± 3.23c	68.33 ± 2.02	23.021
BMI (kg/m2)	20.20 ± 0.55a	22.62 ± 0.32a	26.90 ± 1.19b	22.00 ± 0.00a	17.655

Values are expressed as mean ± SEM. Values with different letter superscripts (a, b, and c) are significantly different.

SE, So-eum; SEM, standard error of the mean; SY, So-yang; TE, Ta-eum; TY, Tae-yang.

3.2. SCAT results

The specific SC type of participants was determined by the SCAT and expert opinion of medical doctors in the Korea Institute of Oriental Medicine.¹¹ The average values for face shape, voice tone, and body shape are shown in Table 2, Table 3, Table 4, respectively. Based on the results of this analysis, we confirmed the SC type of 31 participants (10 TE, 10 SE, 3 TY, and 8 SY) and collected their blood samples for further testing.

Table 2.

Average values of face shape variables.

Type	Jaw width (mm)	Angle inside eyes (degree)	Ratio of face width to nose length	Nose aspect ratio
SE	116.11	32.75	0.36	1.94
SY	122.48	34.23	0.34	1.82
TE	130.79	31.08	0.33	1.83
TY	123.18	33.75	0.35	1.81
Overall	123.13	32.78	0.34	1.86

SE, So-eum; SY, So-yang; TE, Ta-eum; TY, Tae-yang.

Table 3.

Average values of voice tone variables.

Type	MFCC1	MFCC3	MFCC8	sF50	sF10
SE	−2.67	1.54	0.82	125.06	1047.29
SY	−2.62	0.86	−5.12	128.59	1094.00
TE	−2.93	−0.40	−3.42	127.77	1082.56
TY	−2.19	2.91	0.42	121.48	321.27
Overall	−2.69	0.87	−2.30	126.62	3545.13

Mel-frequency cepstral coefficients (MFCC), 50th and 10th percentile of average fundamental frequency distribution (sF50 and sF10, respectively).

SE, So-eum; SY, So-yang; TE, Ta-eum; TY, Tae-yang.

Table 4.

Average values of body shape variables.

Type	Rib circumference (cm)	Hip–waist circumference
SE	71.18	1.27
SY	81.65	1.19
TE	88.40	1.15
TY	79.33	1.17
Overall	80.31	1.20

SE, So-eum; SY, So-yang; TE, Ta-eum; TY, Tae-yang.

3.3. Blood composition results

Blood composition analysis revealed significant differences in blood TC and HDL-C levels between the SE and SY types (p < 0.05); however, all other components were comparable among the groups (p > 0.05; Table 5 and Fig. 2). The SY type showed the lowest blood TC and HDL-C levels, which were significantly different from the levels for the SE type (Fig. 2A). However, the ratio of TC/HDL-C was similar across types. Importantly, all measured values were within normal physiological ranges.

Table 5.

Blood composition.

	SE (n = 10)	SY (n = 8)	TE (n = 10)	TY (n = 3)
TC (mg/dL)	182.70 ± 8.04a	150.00 ± 8.93b	179.10 ± 13.54ab	195.00 ± 19.34
TG (mg/dL)	80.30 ± 9.70	86.62 ± 14.53	115.40 ± 20.79	103.33 ± 19.70
HDL-C (mg/dL)	66.00 ± 5.24a	48.75 ± 3.33b	53.00 ± 4.43ab	58.33 ± 6.00
LDL-C (mg/dL)	108.80 ± 5.10	90.75 ± 7.36	112.20 ± 13.46	125.66 ± 15.45
TC/HDL-C	3.08 ± 0.16	2.84 ± 0.17	3.60 ± 0.41	3.36 ± 0.27
LDL/HDL	1.88 ± 0.14	1.70 ± 0.14	2.30 ± 0.36	2.17 ± 0.26
HOMA-IR	2.18 ± 0.31	1.73 ± 0.17	3.06 ± 0.54	1.65 ± 0.29
Glucose (mg/dL)	97.40 ± 1.14	96.25 ± 2.19	91.50 ± 3.22	91.00 ± 1.15
Insulin (nU/mL)	8.00 ± 1.01	8.65 ± 1.29	13.60 ± 2.31	7.40 ± 1.40
Irisin (ng/mL)	750.63 ± 23.56	753.66 ± 25.60	715.99 ± 27.04	806.07 ± 58.78

Values are expressed as mean ± SEM. Values with different letter superscripts (a, ab, and b) are significantly different.

HOMA-IR, homeostatic model assessment insulin resistance; HDL-C, high-density lipoprotein cholesterol; LDL-C, low-density lipoprotein cholesterol; SEM, standard error of the mean; TC, total cholesterol; TG, triglyceride.

Fig. 2.

Blood composition analysis. (A) Comparison of serum cholesterol, triglyceride, HDL-C, LDL-C, and glucose levels between SC types. (B) Comparison of serum insulin levels between SC types. (C) Comparison of serum irisin levels between SC types.

HDL-C, high-density lipoprotein cholesterol; LDL-C, low-density lipoprotein cholesterol; SC, Sasang constitutional; SE, So-eum; SY, So-yang; TE, Tae-eum; TY, Tae-yang.

3.4. Mitochondrial function analysis results

We investigated differences in the mitochondrial function among the four constitution types. To validate the mitochondrial oxidative phosphorylation capacity, we measured the cellular ATP and mitochondrial ATP levels in isolated platelets from each participant (Fig. 3). Both ATP levels were similar across the four SC types. We further analyzed OCR using an XF24 analyzer (Fig. 4). Basal OCR levels were comparable across the SC types (Fig. 4A). ATP-linked OCR was significantly different between the TE and SY types (p < 0.05), with lower ATP-linked OCR in the TE type than in the SY type (Fig. 4B). The TE type also showed a significantly higher maximal OCR than the SE and SY types (p < 0.05; Fig. 4C). Spare respiration capacity was also significantly higher for the TE type than for the SE type (p < 0.05; Fig. 4D).

Fig. 3.

Cellular and mitochondrial ATP analysis. (A) Cellular ATP levels were measured in platelet cells in the presence of 10 mM glucose. (B) Mitochondrial ATP levels were measured in platelet cells in the presence of 10 mM galactose.

ATP, ; SE, So-eum; SY, So-yang; TE, Tae-eum; TY, Tae-yang.

Fig. 4.

Mitochondrial OCR analysis. (A) Basal OCR was measured in platelet cells by an XF24 analyzer without a mitochondrial inhibitor. (B) ATP-linked OCR was calculated as the difference in OCR prior to and after treatment with oligomycin, a mitochondrial ATPase inhibitor. (C) Maximal OCR was measured in the presence of FCCP, a mitochondrial uncoupler. (D) Spare respiration capacity was calculated as the difference between maximal and basal OCR.

ATP, ; FCCP, p-[trifluoromethoxy]-phenyl-hydrazone; OCR, oxygen consumption rate; SE, So-eum; SY, So-yang; TE, Tae-eum; TY, Tae-yang.

3.5. WES and variant analysis results

We investigated which gene variants were SC type-specific by WES analysis. We used GO terms, pathway, and disease information to identify which biological roles were associated with each SC type. Gene variants were defined as genes containing at least one altered amino acid. We investigated the number of gene variants and their associated amino acid residue alterations. We found 3344, 3070, 3291, and 1986 gene variants for the SE, SY, TE, and TY types, respectively. Of these gene variants, 886, 702, 828, and 247 genes were specific to SE, SY, TE, and TY types, respectively (Fig. 5A). We also counted the number of amino acid residue variations for each SC type to measure the amino acid level alterations that might affect protein function. The numbers of specific amino acid variants were 2983, 2348, 2754, and 751 for the SE, SY, TE, and TY types, respectively (Fig. 5B). A total of 1369 gene variants were commonly found across all SC types (Fig. 5A), and these genes contained an average of 1.85 altered residue variants.

Fig. 5.

Venn diagram of gene variants for each SC type. (A) Number of gene variants for each SC type. (B) Number of amino acid residue alterations for each SC type.

SC, Sasang constitutional; SE, So-eum; SY, So-yang; TE, Tae-eum; TY, Tae-yang.

Next, we tested the hypothesis that SC type-specific gene mutations, existing in the exome region, could be used as genetic markers to identify the SC type of an individual. We identified variant genes that were more frequently found in certain SC types and examined their biological roles (Data S1). We detected 19, seven, 11, and six SC type-specific genes for the SE, SY, TE, and TY types, respectively (Fig. 6 and Table 6). We positively selected the genes observed at least two human samples in the group.

Fig. 6.

Selected gene–biological module associations. Gene variants and biological module associations from GO terms, pathways, and diseases. (A) SE type-specific variants. (B) SY type-specific variants. (C) TE type-specific variants. (D) TY type-specific variants.

GO, gene ontology; SE, So-eum; SY, So-yang; TE, Tae-eum; TY, Tae-yang.

Table 6.

SC type-specific gene variants.

SC type	Gene	Accession	Frequency	HGVS.p
SE	HSPBAP1	rs71270423	6	p.Pro456_Gln457insSer
SE	KNTC1	—	3	p.Phe684Tyr
SE	KNTC1	—	3	p.Phe721Tyr
SE	KNTC1	rs186936079	3	p.Leu1511Phe
SE	KNTC1	—	2	p.Lys106Asn
SE	KNTC1	—	2	p.Lys1110Asn
SE	KNTC1	—	2	p.Lys2185Asn
SE	KNTC1	rs75696429	2	p.Val71Gly
SE	KNTC1	rs186936079	3	p.Leu70Phe
SE	KNTC1	rs75696429	2	p.Val1512Gly
SE	VIPR1-AS1	rs413042	5	intron_variant
SE	COBLL1	rs3841875	4	p.Ser1210fs
SE	COBLL1	rs3841875	4	p.Ser1181fs
SE	COBLL1	rs3841875	4	p.Ser1143fs
SE	COBLL1	rs3841875	4	p.Ser1105fs
SE	C6orf163	rs9353479	5	p.Ala72Val
SE	C14orf166B	rs201329545	2	p.Lys449Asn
SE	C14orf166B	rs139864900	3	p.Ile132Thr
SE	C14orf166B	—	2	p.Glu139Lys
SE	C14orf166B	—	2	p.Glu122Lys
SE	ACOT8	rs201025211	4	p.Met1fs
SE	NT5C1B-RDH14	—	2	p.Ser176Trp
SE	NT5C1B-RDH14	rs145698850	3	p.Gly353Arg
SE	NT5C1B-RDH14	—	2	p.Ser118Trp
SE	NT5C1B-RDH14	rs145698850	3	p.Gly411Arg
SE	CC2D1A	rs76113658	3	p.Ala338Thr
SE	CC2D1A	—	2	p.Glu363Lys
SE	CC2D1A	—	2	p.Glu909Lys
SE	CC2D1A	—	2	p.Glu910Lys
SE	SKOR2	rs77291182	2	p.Pro927Leu
SE	SKOR2	—	2	p.Ser899Cys
SE	SKOR2	—	2	p.Pro291del
SE	RNF166	rs117837843	4	p.Arg134Met
SE	DIAPH3	rs36084898	4	p.Asn293Ser
SE	DIAPH3	rs36084898	4	p.Asn352Ser
SE	DIAPH3	rs36084898	4	p.Asn363Ser
SE	DIAPH3	rs36084898	4	p.Asn317Ser
SE	TMEM176A	rs35972858	4	p.Gly63Asp
SE	TMEM176A	rs35972858	4	p.Gly52Asp
SE	TMEM176A	rs35972858	4	p.Gly111Asp
SE	SCN4A	—	2	p.Val1589Ala
SE	SCN4A	rs7218917	3	synonymous codon
SE	FAM132B	rs111241405	3	p.Ala260Ser
SE	FAM132B	rs117194634	3	p.Val185Met
SE	FAM132B	rs117199696	2	p.Ala81Ser
SE	PPIG	rs76754479	4	p.Thr76Met
SE	SPOCK2	rs2306322	4	p.Gly353Ser
SE	DNM2	rs3745674	4	p.Pro263Leu
SE	DNM2	rs3745674	4	p.Pro15Leu
SY	METTL1	rs118007790	5	p.Ser46Phe
SY	SULT1A2	rs138147609	3	p.Glu217*
SY	SULT1A2	—	2	p.Val176Gly
SY	SULT1A2	rs138147609	3	p.Glu184*
SY	TTYH1	rs3745433	2	p.Gly86Arg
SY	TTYH1	rs3745433	2	p.Gly21Arg
SY	TTYH1	rs3745433	2	p.Gly82Arg
SY	TTYH1	rs144026046	3	p.Gln51*
SY	TTYH1	rs3745433	2	p.Gly135Arg
SY	VWA2	rs200171942	2	p.Arg591Trp
SY	VWA2	rs201721125	2	p.Ser378Cys
SY	PSMA6	rs17103147	4	p.Arg5Trp
SY	PACSIN2	rs139155473	4	p.Cys27Phe
SY	LPCAT2	—	2	p.Ser51fs
SY	LPCAT2	—	2	p.Ser194fs
SY	LPCAT2	rs144432562	3	p.Arg167*
TE	CHD1	rs138635992	5	p.Pro1684del
TE	SP4	rs139491266	4	p.Leu241Val
TE	MICAL3	—	2	p.Thr122Ile
TE	MICAL3	rs148674116	2	p.Gly309Ser
TE	MICAL3	—	2	p.Thr1829Ile
TE	MICAL3	rs61739477	2	p.Val319Ile
TE	MICAL3	—	2	p.Thr84Ile
TE	FLNC	rs180834558	3	p.Leu2538Phe
TE	FLNC	—	2	p.Ser589Leu
TE	FLNC	rs180834558	3	p.Leu2505Phe
TE	ARL5C	rs151045610	4	p.Trp171Arg
TE	KLC3	rs186054339	2	p.Ala219Ser
TE	KLC3	rs182912549	2	p.Glu333Lys
TE	KLC3	—	2	p.Arg440_Gly441insGluSerIleArgArg
TE	KLC3	—	2	p.Arg455_Gly456insGluSerIleArgArg
TE	KLC3	—	2	p.Arg441_Gly442insGluSerIleArgArg
TE	KLC3	rs186054339	2	p.Ala189Ser
TE	KLC3	rs182912549	2	p.Glu319Lys
TE	KLC3	rs182912549	2	p.Glu318Lys
TE	KLC3	rs186054339	2	p.Ala205Ser
TE	KLC3	rs186054339	2	p.Ala204Ser
TE	PRG3	rs115707133	4	p.Cys215*
TE	OR4M2	rs140079625	4	p.Ile200Met
TE	TNK2	rs148791867	2	p.Pro402Leu
TE	TNK2	—	2	p.Arg6Leu
TE	TNK2	rs148791867	2	p.Pro434Leu
TE	TNK2	rs200619114	2	p.Leu198Gln
TE	TNK2	rs200619114	2	p.Leu589Gln
TE	TNK2	rs200619114	2	p.Leu667Gln
TE	TNK2	rs148791867	2	p.Pro465Leu
TE	TNK2	rs200619114	2	p.Leu156Gln
TE	TNK2	rs200619114	2	p.Leu621Gln
TE	OR5D13	rs74548274	4	p.Gln198*
TE	MNT	—	2	p.Arg232fs
TE	MNT	rs185455119	3	p.Gly32Arg
TY	RP11-274B21.1	rs77554651	3	splice donor variant
TY	SHKBP1	—	2	p.Gln111Glu
TY	SHKBP1	rs114918214	2	p.Ala207Thr
TY	SHKBP1	rs114918214	2	p.Ala130Thr
TY	SHKBP1	rs114918214	2	p.Ala201Thr
TY	PCDP1	rs182093842	2	p.Met105Ile
TY	PCDP1	rs200947466	2	p.Gly611Arg
TY	PCDP1	rs200947466	2	p.Gly325Arg
TY	PCDP1	rs187802437	2	p.Pro106Thr
TY	PCDP1	rs200947466	2	p.Gly38Arg
TY	PCDP1	rs200947466	2	p.Gly169Arg
TY	FAM211A	rs140374173	3	p.Arg210*
TY	MYCBPAP	rs78242165	3	p.Ala535Val
TY	MYCBPAP	rs78242165	3	p.Ala546Val
TY	MYCBPAP	rs78242165	3	p.Ala572Val
TY	MYCBPAP	rs78242165	3	p.Ala5Val
TY	C1orf222	—	2	p.Thr111Met
TY	C1orf222	—	2	p.Pro45Leu
TY	C1orf222	—	2	p.Pro167Leu

SE, So-eum; SY, So-yang; TE, Ta-eum; TY, Tae-yang.

For SE type-specific gene variants, those associated with musculoskeletal disease (SCN4A and DNM2) were found in six of 10 participants, and those associated with nervous system diseases (CC2D1A, SKOR2, SCN4A, SPOCK2, and DNM2) were found in nine of 10 participants (Fig. 6A). SY type-specific gene variants were confirmed to have biological roles in immune response, ion transport, and cell–cell adhesion (Fig. 6B). TE type-specific variants, associated with GTP-mediated signaling, were found in seven of 10 participants, including ARL5C, which directly binds to GTP, and OR4M2 and OR5D13, which are involved in the G-protein-coupled signaling pathway (Fig. 6C). As the TY type included a small number of participants, only six frequent TY-specific variants were found, which were associated with protein homoligomerization, axoneme formation, and cell differentiation (Fig. 6D).

To assess correlations between specific gene variants and their biological roles, we isolated SC type-specific gene variants and performed module enrichment tests to determine which biological terms were primarily associated with SC type-specific variants (Data S2). The module analysis utilized GO terms, REACTOME, BIOCRATA, and KEGG pathways, and disease phenotypes. Enrichment tests were conducted for module sizes >4 and <100. The SE, SY, TE, and TY type-specific variants were enriched for 299, 286, 219, and 114 modules, respectively. The top five terms based on enrichment significance for SE type-specific variants were “detection of visible light” (GO: 0009584), “cytochrome-b5 reductase activity, acting on NAD(P)H” (GO: 0004128), “negative regulation of actin filament polymerization” (GO: 0030837), “G-protein coupled photoreceptor activity” (GO: 0008020), and “exit from mitosis” (GO: 0010458). SY type-specific variants were involved in “the regulation of protein localization to plasma membrane” (GO: 1903076), “polynucleotide adenylyl transferase activity” (GO: 0004652), “hepatocyte differentiation” (GO: 0070365), “vitamin D metabolic process” (GO: 0042359), and “citrate metabolic process” (GO: 0006101). TE type-specific variants were identified as “RNA transport” (GO: 0050658), “dorsal/ventral neural tube patterning” (GO: 0021904), “S-adenosylhomocysteine metabolic process” (GO: 0046498), and “regulation of fibroblast growth factor receptor signaling pathway” (GO: 0040036). Finally, TY type-specific variants were identified as “mitochondrial fragmentation involved in apoptotic process” (GO: 0043653), “cellular response to arsenic-containing substance” (GO: 0071243), “morphogen activity” (GO: 0016015), and “actin-dependent ATPase activity” (GO: 0030898). A summary of enriched terms for each SC type is shown in Fig. 7.

Fig. 7.

Enriched terms for each SC type. Different colors indicate the enriched biological terms for different SC types. Enrichments were calculated with hypergeometric tests. Values on the x-axis denote the significance of enrichment tests.

SC, Sasang constitutional; SE, So-eum; SY, So-yang; TE, Tae-eum; TY, Tae-yang.

4. Discussion

During the past decade, more than 100 studies have contributed to improving the evidence-based analysis of SCM. A wide range of approaches has successfully been used to accumulate unbiased evidence for SCM. To our knowledge, this is the first study to investigate the mitochondrial function characteristics and WES-based SC type-specific gene variants in healthy participants. Our results reveal the novel finding that aspects of mitochondrial oxygen consumption, including ATP-linked OCR, maximal OCR, and spare respiration capacity, differ between the TE type and other SC types.

Our study included student participants from a single university who were healthy (i.e., nondiseased), a single sex (i.e., male), and within a narrow age range (22.3 ± 1.4 years). A major hurdle in genomic studies is the wide variation among individuals in age, sex, environmental factors, and pathological history. Therefore, the similarity of participants in the present study could serve to minimize such biases among individuals. As DNA mutations accumulate during aging and in pathological conditions, the inclusion of relatively young participants in the current study could serve to minimize acquired DNA mutations and preserve inborn gene variants, thereby strengthening links between WES results and SC types.

Examination of basic physical characteristics showed that the TE type had a significantly greater body weight and BMI and a tendency toward higher insulin levels than the other SC types, which is consistent with previous studies.41, 42 Various studies suggest that the TE type is prone to metabolic disorders and cardiovascular diseases, including obesity, diabetes, and ischemic stroke.10, 18, 29, 30, 43, 44, 45 Although we found no significant pathological abnormalities in the serum indices of any tested sample, the higher BMI and body weight of the TE type could be risk factors for these diseases.

We also found significantly lower serum TC and HDL-C levels for the SY type than for the SE type, with the SY type showing the lowest HDL-C level among the four SC types. A lower HDL-C level is a well-known risk factor for cardiovascular disease.⁴⁶ Importantly, a recent study suggests that the SY type has a higher cardiovascular disease risk ratio than other SC types.⁴⁵ Thus, our finding that the SY type has the lowest HDL-C level could be one reason for their higher cardiovascular disease risk ratio. Irisin is a recently discovered muscle-derived hormone that is implicated in metabolic homeostasis and various metabolic diseases.47, 48, 49 However, we found no differences in serum irisin levels among SC types. Because irisin levels are only altered by high-intensity exercise,⁵⁰ examining exercise response and alterations in blood irisin levels after long-term exercise would be valuable for further investigations in the differences in energy metabolism among the SC types. We note that the agreement between the results of the present study and previous SCM studies reflects the high reliability of our sample group despite its small size.

Mitochondria consume oxygen for oxidative phosphorylation, which converts various substrates (i.e., glucose, free fatty acids, and amino acids) into the high-energy biomolecule ATP. Thus, measurements of ATP level and OCR are essential for assessing the mitochondrial function. Although ATP contents and basal OCR were comparable among the SC types, some specific components, including ATP-linked OCR, maximal OCR, and spare respiration capacity, differed between the TE type and the SE or SY types. ATP-linked OCR is the amount of oxygen required to produce ATP. A higher ATP-linked OCR indicates a greater efficiency of ATP generation, whereas a lower ATP-linked OCR indicates a lower efficiency of substrate and oxygen utilization. Generally, a lower efficiency is found in metabolic disease such as type 2 diabetes.51, 52 Maximal OCR and spare respiration capacity are measured in the presence of FCCP, a hydrogen ion (or proton, H⁺) uncoupler. The proton gradient is a fundamental driving force of ATP production via ATPase of the mitochondrial inner membrane. Under physiological conditions, mitochondria are not fully activated and have spare respiration capacity; thus, maximal OCR can be measured only in ex vivo conditions with FCCP.³² The higher maximal OCR and spare respiration capacity of the TE type might indicate that this constitutional type maintains a low respiration rate for survival under normal conditions but shows full activation in high energy-requiring environments, which might be reflected by a higher obesity rate (i.e., energy-storing condition).

In summary, we found that the specific energy metabolism-related features of the TE type were a higher BMI, lower ATP-linked OCR, and higher maximal OCR without a significant change in the ATP level. These features are most likely characteristic of the TE type under normal conditions. Similar biological features are found in peripheral blood mononuclear cells of type 2 diabetes patients.⁵³ Coincidently, the TE type is known to have a higher risk of metabolic syndromes, including obesity and diabetes.29, 30 However, in the present study, we found that basal OCR rate and ATP level were not significantly modified in TE-type healthy individuals. These results may explain why TE-type individuals have a higher risk of metabolic syndromes.

An important finding of the present study is the identification of gene variants exclusively detected in different SC types with a high frequency (19, 7, 11, and 6 variants in the SE, SY, TE, and TY types, respectively). Although there is a limit to associating gene variants with constitutional features or specific energy metabolism-enriched pathways, we found that the top 10 enriched biological pathways of each type were completely distinct without overlap. Thus, these independent sets of gene variants could serve as genetic markers for SC type diagnosis.

The present and previous41, 42 findings, showing that the TE type has a significantly higher BMI and a greater tendency toward obesity than other SC types, leads us to question whether the TE type-specific gene variants are primarily associated with obesity or other TE type-specific features. However, the BMI of the TE type fell in the overweight (25.0–29.9), but not the obese (>30), range. In addition, the tested blood component indices (e.g., TC, TG, HDL-C, LDL-C, and insulin) were within normal ranges for all types. Importantly, the TE type-specific gene variants were also found in TE participants with normal BMI (18.5–24.9). Therefore, we believe that the specific genetic variants of the TE type are not merely associated with the overweight tendency of the TE type but rather are integrative features of the TE type.

Although the quality of our collected samples was good and nongenetic variance among participants was minimal, the small number of tested samples is a major limitation of the present study. Therefore, validations of mitochondrial function and SC type-specific gene variations within larger controlled populations are needed for the clinical application of these findings in SC-based Korean medicine.

In conclusion, our study demonstrates SC type-specific differences in mitochondrial function and gene mutations. An understanding of the TE type-specific mitochondrial function differences could be helpful for the diagnosis and treatment of metabolic diseases among TE individuals. Our findings suggest that mitochondria-mediated energy metabolism analysis and WES can be effective methods for SCM research.

Conflicts of interest

The authors declare no conflict of interest.

Appendix A. Supplementary data

The following are Supplementary data to this article: