Dear Editor,
Obesity is a major health issue with global prevalence and increases the risk of many metabolic diseases. Of particular concern is the increasing incidence of type 2 diabetes, the primary causes of which are obesity-driven insulin resistance in white adipose tissue (WAT), skeletal muscle, and liver, and decreased insulin secretion by pancreatic β-cells [
1]. Abnormal adipose expansion is accompanied by adverse consequences such as chronic inflammation and fibrosis, dysregulated adipokine secretion, and mitochondrial dysfunction [
2]. Although many molecules are known to regulate adipose function, factors that are highly specific for WAT function and metabolic status remain underexplored.
Calsyntenin 3b (
Clstn3b) is a novel gene that promotes brown fat thermogenesis in mice [
3]. It is an abundant adipocyte-selective product of the mouse
Clstn3 locus, and has three exons: the first exon is large and unique and the last two are shared with the known
Clstn3 gene. We previously reported that the transcript form of
Clstn3/
CLSTN3 locus in WAT is discordant between mice and humans.
CLSTN3 is routinely expressed in human white fat, while the form of
CLSTN3B appears to be rare [
4]. A recent study identified that
Clstn3b enforces adipocyte multilocularity to facilitate lipid utilization in mice, and
CLSTN3B marks human thermogenic fat [
5]. However, the relevance of
CLSTN3B in human white fat remains unknown. We thus aimed to examine its transcript profile in human WAT and associate it with metabolic status in individuals.
We included a total of 210 participants with body mass index (BMI) ranging from 18.0 to 65.4 kg/m2 in the present study. Anthropometric and biochemical parameters were measured, and paired samples of abdominal subcutaneous adipose tissue (SAT) and visceral adipose tissue (VAT) were obtained from patients who underwent cholecystectomy or bariatric surgery. We first determined the transcript profile of the CLSTN3B gene with quantitative polymerase chain reaction (PCR) and then performed gene co-expression analysis combined with RNA sequencing data of these samples. The clinical relevance of novel variants occurred at the first unique exon of the CLSTN3B gene was lastly screened in these individuals.
Firstly, we analyzed the distribution of RNA sequencing reads at the CLSTN3B locus, and revealed that this novel gene exhibited a distinctive expression pattern in WAT between individuals: it was highly expressed in some WAT samples and extremely low in others (Supplementary Fig. S1a). Genotype-Tissue Expression database showed that CLSTN3B is exclusively expressed in human adipose tissue, in comparison to other metabolic tissues (Supplementary Fig. S1b). We performed quantitative PCR analysis of paired SAT and VAT samples from 210 participants who were lean (n = 48), overweight (OW, n = 20), and obese (n = 142) to determine the transcript profile of the CLSTN3B gene. The clinical characteristics of study participants are shown in Supplementary Table S1. CLSTN3B mRNA expression in SAT and VAT samples was inversely correlated with BMI (SAT, r = −0.312, P < 0.001; VAT, r = −0.473, P < 0.001; Fig. 1a), and CLSTN3B expression significantly decreased in both fat depots from the lean group to the obese group (Fig. 1b). By contrast, transcript level of CLSTN3 in SAT samples was positively correlated with BMI (r = 0.189, P = 0.006), and this association was also observed in VAT samples (r = 0.280, P < 0.001; Fig. 1c). The CLSTN3 expression was significantly higher in SAT and VAT samples from obese participants than those from lean participants (Fig. 1d). Furthermore, CLSTN3B expression was inversely correlated with CLSTN3 expression both in SAT and VAT samples (SAT, r = −0.254, P < 0.001; VAT, r = −0.248, P < 0.001; Supplementary Fig. S1c and d).
Next, we examined whether CLSTN3B expression in WAT relates to obesity-associated metabolic status. Spearman correlation analysis was performed between adipose tissue CLSTN3B expression and metabolic variables in these individuals. CLSTN3B expression in SAT and VAT samples was inversely correlated with several metabolic characteristics of obesity and adipose tissue, including body fat%, waist-to-hip ratio, circulating levels of triglyceride (TG), adipokine leptin, and inflammatory cytokines [tumor necrosis factor-α (TNF-α), interleukin-6 (IL-6), and C-reactive protein (CRP)], and it was positively associated with circulating concentrations of high-density lipoprotein cholesterol (HDL-C) and adiponectin. In addition, CLSTN3B expression was inversely correlated with certain parameters of whole-body and adipose glucose metabolism, including fasting plasma glucose (FPG), fasting insulin (FINS), glycosylated hemoglobin A1c (HbA1c), adipose tissue insulin resistance (Adipo-IR), and homeostasis model assessment of insulin resistance (HOMA-IR). Interestingly, the correlation of CLSTN3B expression with metabolic variables was stronger in VAT than in SAT samples. Meanwhile, CLSTN3 expression in SAT and VAT samples was significantly correlated with some metabolic characteristics, which was mostly opposite to that of the CLSTN3B gene (Fig. 1e and Supplementary Table S2). Moreover, when adjusted for BMI, the correlation of CLSTN3B expression with FPG, FINS, HbA1c, HOMA-IR, Adipo-IR, TG, and LDL-C retained statistically significant only in VAT samples (Table 1).
Further, we performed CLSTN3B co-expression analysis to predict its functional processes in SAT and VAT samples. Since CLSTN3B has not been updated in the human hg38 genome, here we combined the quantitative PCR and transcriptional data of these samples to analyze the correlation between the expressions of CLSTN3B and other genes. Gene Set Enrichment Analysis (GSEA) was performed to identify CLSTN3B-associated gene sets and pathways enriched in the ranked gene lists. Genes co-expressed with CLSTN3B in SAT samples showed a significant enrichment in pathways related to ribosome and vascular endothelial growth factor signaling pathway, and genes anti-expressed with CLSTN3B were enriched in pathways related to lysosome and so on (Fig. 1f). However, gene sets co-expressed with CLSTN3B expression in VAT samples were enriched in mitochondrial energy metabolism, such as oxidative phosphorylation, branched-chain amino acid degradation, tricarboxylic acid cycle (TCA cycle), pyruvate metabolism, fatty acid degradation, and arginine biosynthesis, and gene sets negatively anti-expressed with CLSTN3B expression were enriched in inflammatory activities (Fig. 1g). Additionally, as shown in the enrichment plots and heatmap of the leading gene sets, CLSTN3B-related enhancement of mitochondrial fundamental activities occurred, and expression levels of genes involved in oxidative phosphorylation, TCA cycle, and fatty acid degradation were upregulated with the increase of CLSTN3B expression in human VAT (Supplementary Fig. S2a, Fig. 1h). Moreover, as shown in the scatter plot, expression levels of beige/brown adipocyte marker genes (PPARGC1A, CIDEA, and UCP1) were positively correlated with CLSTN3B expression in VAT samples (Supplementary Fig. S2b).
Given that obesity status substantially influenced gene expressions in human VAT, we restrictively explored gene signatures associated with CLSTN3B in lean individuals. There was no significant correlation between VAT CLSTN3B expression and BMI (r = 0.207, P = 0.158; Supplementary Fig. S2c). Differential expression analysis revealed 132 differentially expressed genes (DEGs), including 118 upregulated and 14 downregulated genes, in CLSTN3B-high VAT samples (n = 16) compared with the CLSTN3B-low samples (n = 32). In the functional enrichment analysis, DEGs were mainly enriched in TCA cycle and respiratory electron transport, mitochondrial fatty acid oxidation, and several other metabolic processes (Fig. 1i). Indeed, CLSTN3B-high VAT samples showed higher expression levels of mitochondrial genes, such as the most enriched biological process, TCA cycle and respiratory electron transport, than the CLSTN3B-low VAT samples (Fig. 1j).
Finally, we sequenced the unique protein-coding region of
CLSTN3B in all participants to identify its genetic variation and associated the variants with metabolic traits. Three variants were found to be located within this region, including rs7296261 (g.398G>A, p.R133Q), rs1868799, and rs1868800 (it undergoes a missense mutation p.R224S if the allele of rs1868799 and rs1868800 is A and C, respectively). Genotyping for rs7296261 resulted in 45 GG, 122 GA, and 43 AA genotype carriers. Compared with GG genotype carriers, AA genotype carriers had significantly higher BMI (31.6 ± 8.3 vs. 35.5 ± 8.1,
P < 0.05; Fig. 1k), which was in line with our previous work [
4]. In addition, mean HOMA-IR was significantly higher in AA genotype carriers than in GG genotype carriers (6.4 ± 5.3 vs. 9.3 ± 8.3,
P < 0.05; Fig. 1l). We further found that changes in BMI and HOMA-IR did not differ between the two genotypes for rs1868799 (Supplementary Fig. S3a and b). Genotyping for rs1868800 divided participants into three groups, namely GG (
n = 43), GC (
n = 115), and CC (
n = 52) genotypes. Compared with GG genotype carriers, CC genotype carriers showed an increasing trend in BMI (31.1 ± 8.2 vs. 34.6 ± 9.0,
P = 0.056; Fig. 1m), as well as in HOMA-IR (6.3 ± 5.4 vs. 8.6 ± 8.0,
P = 0.108; Fig. 1n).
To our knowledge, this study is the first to report that
CLSTN3B expression in human WAT correlates with BMI and various metabolic parameters.
CLSTN3B co-expression analysis in VAT samples shows that the transcript profile of this gene also associates with a mitochondrial gene program. Our data provide further insights into the potential function of the
CLSTN3B gene in human adipose tissue and obesity in addition to its roles in thermogenic fat in mice. The mouse adipocyte-specific
Clstn3b gene encodes a protein, CLSTN3β, localized at endoplasmic reticulum (ER), and it partly drives the secretion of the neurotrophic factor S100B to maintain innervation of thermogenic adipose tissue [
3]. Another independent study showed that CLSTN3β localizes to ER-lipid droplet contact sites through a conserved hairpin-like domain; it associates with the cell death-inducing DNA fragmentation factor 45 (DFF45)-like effector (CIDE) proteins to enforce adipocyte multilocularity and facilitate lipid utilization of thermogenic adipocytes in mice;
CLSTN3B is a selective marker of thermogenic fat in humans [
5]. Thermogenic (brown and beige) adipocytes have a remarkable ability to oxidize fatty acids, whereas white adipocytes residing in WAT are specialized for energy storage and mobilization [
6]. In the present study, the strong association between
CLSTN3B expression and mitochondrial gene program in white fat challenges underlying mechanisms. ER-mitochondria contacts are involved in the regulation of lipid metabolism and mitochondrial dynamics [
7]. Recently, Combot
et al. demonstrated that seipin is localized to ER-lipid droplet contact sites under lipid loading, while seipin association with ER-mitochondria contact sites is stimulated under a fasting state and thus promotes mitochondria oxidation in adipocytes [
8]. Therefore, more studies are needed to decipher the potential role of CLSTN3β on mitochondrial function.
Outstanding aspects of WAT are its anatomical location and function. SAT depot comprises the majority of body fat, while small VAT depot is often considered more pernicious, and VAT hypertrophy is associated with insulin resistance and metabolic dysfunction [
9]. Our present work shows that the
CLSTN3B transcript in VAT samples is more relevant to clinical parameters than that in SAT samples. The correlation between
CLSTN3B expression and key metabolic parameters adjusting for BMI retains statistically significant only in VAT samples.
CLSTN3B co-expression analysis in VAT samples reveals that this gene may function in adipose homeostasis by influencing mitochondrial activities. Mechanistic investigations are warranted to verify this novel role of
CLSTN3B in mitochondrial signature. Previous study has shown that VAT depot has stronger mitochondrial activity and higher thermogenic gene expression than SAT [
10]. Furthermore, a single-cell sequencing study identified a potentially visceral thermogenic adipocyte subpopulation (hAd6) characterized by higher mitochondrial intensity and increased expression of mitochondrial and thermogenic genes [
11]. High expression level of
CLSTN3B in some VAT samples and the strong association with expression levels of mitochondrial genes suggest that VAT depot may predispose to have a highly active metabolic characteristic under specific cues. Adipose tissue harbors heterogeneous cell populations, including progenitor cells, adipocytes, and immune cells, and cell types with high
CLSTN3B expression in human fat need to be further elucidated at the single-cell resolution. Given the adipose-specific expression of the
CLSTN3B gene and its potential role in regulating mitochondrial homeostasis, adipose-targeted gene therapy may be a novel approach to treat obesity and associated metabolic disorders.
Many genetic variants have been associated with obesity and type 2 diabetes, and the majority are located in non-coding regions. Only a minority of genetic variants are located within protein-coding regions, which marks phenotypic impact by affecting protein function [
12]. We have previously shown that an intronic variant rs7296261 at the
CLSTN3 locus associated with obesity risk and high
CLSTN3 expression in human adipose tissue [
4]. This variant is simultaneously localized to the first exon of the
CLSTN3B gene, and here, we screen 210 individuals and observe three common variants within this unique region. Among these loci, rs7296261 affecting the R133Q mutant is the only variant that is linked to metabolic phenotypes, namely, AA genotype carriers have higher BMI and accelerated insulin resistance than GG genotype carriers. This mutant is located within the hairpin-like domain of the CLSTN3β protein that mediates its localization to other organelles, and it may have an impact on protein function. The variant rs1868800 affecting the R224S mutant has a highly homologous genetic variation with rs7296261 although they are spatially distant, and this mutant is located within a disordered region linking the hairpin domain to the transmembrane domain of the CLSTN3β protein. In this study, we have a better understanding of the genetic basis of the
CLSTN3B gene, which may bring implications for the functional impact of protein mutations. More interestingly, the transcript level of
CLSTN3 in human WAT is negatively correlated with
CLSTN3B expression. In addition, the two genes are inversely correlated with some metabolic parameters. We believe that underlying cues may be involved in the expression pattern and biological function of the two genes, and the variant rs7296261 may be one of the regulators.
Our study has several strengths. First, this study is the first to analyze the transcript profile of the CLSTN3B gene in human WAT and associate its level with metabolic status in individuals. Second, the strong correlation between CLSTN3B transcript and expression levels of mitochondrial genes highlights its potential function in human VAT, and this brings a great implication for following mechanistic studies. Finally, the protein-coding variant of CLSTN3B is linked to metabolic phenotypes and may provide cues for exploring underlying mechanisms. However, this study also has a few limitations. This study is unable to clarify the causal relationship between the CLSTN3B transcript profile in human WAT and the progression of obesity. The exact effect of the CLSTN3B gene on obesity warrants further confirmation in large-scale prospective studies and mechanistic studies.
In conclusion, we provide a novel observation on the transcript profile of CLSTN3B in human adipose tissue and highlight its associations with mitochondrial gene program and metabolic status. These findings may contribute to the development of adipose-specific CLSTN3B upregulation to improve metabolic dysfunction.
The Author(s) 2023. Published by Oxford University Press on behalf of Higher Education Press.
PDF
(8227KB)
