Loss of VSTM2A promotes adipocyte hypertrophy and disrupts metabolic homeostasis

Abstract

Objective

Adipose tissue expands through hyperplasia and hypertrophy to store excess lipids, a process that is essential for the maintenance of metabolic homeostasis. The mechanisms regulating adipocyte recruitment from progenitors remain unclear. We have previously identified V‐set and transmembrane domain‐containing protein 2A (VSTM2A) as a factor promoting fat cell development in vitro. Whether VSTM2A impacts adipose tissue and systemic metabolism in vivo is still unknown.

Methods

We generated VSTM2A knockout mice (Vstm2a ^−/− ) using clustered regularly interspaced short palindromic repeats (CRISPR)/CRISPR‐associated protein 9 (Cas9) and fed them either a chow or high‐fat diet. These mice were evaluated for body weight, adiposity, blood parameters, and glucose homeostasis.

Results

Vstm2a ^−/− mice were viable and showed no body weight differences. Although adipose mass was similar, Vstm2a ^−/− mice had larger adipocytes, an effect linked to inflammation, ectopic lipid deposition, and impaired glucose and lipid metabolism. Transcriptomic analysis revealed that VSTM2A loss affects the expression of several genes in adipose tissue, including some related to the lysosome. Interestingly, acute lysosomal inhibition early in life is sufficient to cause adipocyte hypertrophy in adults.

Conclusions

VSTM2A is dispensable for adipose tissue formation, but its loss causes adipocyte hypertrophy and impairs glucose and lipid homeostasis. Our study also underscores a critical role of the lysosome in initiating adipogenesis.

Study Importance.

What is already known?

• V‐set and transmembrane domain‐containing 2A (VSTM2A) is a protein that is highly expressed and secreted by preadipocytes early in adipogenesis both in vitro and in vivo.

• Elevated VSTM2A expression is observed in embryonic adipose tissue, with higher levels in adipose precursors compared with mature adipocytes.

• VSTM2A overexpression enhances adipogenesis in vitro, whereas its deletion impairs this process, suggesting a role in promoting adipogenesis.

What does this study add?

• VSTM2A‐null mice are viable, with no major developmental defects. Loss of VSTM2A leads to adipocyte hypertrophy, inflammation, ectopic lipid storage, and impaired glucose and lipid homeostasis.

• VSTM2A loss impairs gene expression in developing adipose tissue, including lysosomal genes. Additionally, early inhibition of lysosomal function impairs fat cell development in both in vitro and in vivo models.

How might these results change the direction of research or the focus of clinical practice?

• The identification and characterization of novel regulators of adipogenesis are crucial to understand the mechanisms controlling adipose tissue development. A better comprehension of adipose tissue is key for the development of new tools to fight obesity.

• Studying VSTM2A loss in vivo provides insights into its physiological roles and helps develop strategies to prevent adipose tissue dysfunction, a main characteristic of obesity, potentially improving approaches to manage obesity.

INTRODUCTION

White adipose tissue (WAT) is the primary site for energy storage in mammals. The white adipocytes that compose WAT accumulate triglycerides in times of energy surplus and break these molecules to release fatty acids and glycerol during fasting. Beyond energy metabolism, WAT secretes several factors that deeply impact the function of distant tissues. The endocrine function of WAT has been well established, and this tissue is considered as a central player in the control of metabolic homeostasis [1].

WAT develops early in life and rapidly expands after birth [2]. In adults, WAT remains dynamic, expanding in response to overnutrition [3, 4]. The obesogenic expansion of WAT is linked to a rise in adipocyte size (i.e., hypertrophy) that can be followed by an increase in cell number (i.e., hyperplasia) [5]. Strong associations between adipocyte hypertrophy and metabolic diseases have been established in both mice and humans [6]. Hypertrophic and dysfunctional adipocytes usually fail to store excessive lipids, which causes ectopic lipid storage, local inflammation, and insulin resistance. Conversely, hyperplastic WAT is linked to better metabolic outcomes [6].

The molecular mechanisms that control the recruitment of adipocytes from progenitors are still incompletely characterized. Early studies have suggested that white adipocytes originate from a “primitive organ,” a cluster of blood vessels observed during development. This close association between vasculature and white adipocyte development has been well documented [7, 8]. WAT progenitors, isolated from the stromal vascular fraction (SVF) of adipose tissue, can be differentiated in vitro using an adipogenic cocktail [9, 10]. Single‐cell RNA sequencing experiments performed from SVF have revealed that adipose progenitors are heterogeneous in nature, express an array of different cell surface markers, and have diverse impacts on WAT development and function [11].

We previously identified V‐set and transmembrane domain‐containing 2A (VSTM2A) as a glycoprotein that is highly expressed and secreted by preadipocytes [12, 13]. Our studies showed that VSTM2A is expressed during early adipogenesis in various cell lines and in the initial stages of white and brown adipocyte development. We found that VSTM2A is enriched in adipose precursors when compared with mature adipocytes and that VSTM2A‐expressing cells are localized close to the vasculature. We showed that overexpressing VSTM2A enhances adipogenesis in vitro, whereas its deletion impairs this process [12]. We proposed that VSTM2A preserves and amplifies the adipogenic potential of adipose precursors.

Despite the impact of VSTM2A in promoting adipogenic commitment in vitro, its contribution to adipose tissue formation in vivo has never been tested. Here, we report the generation of a whole‐body VSTM2A‐null (VSTM2A knockout, i.e., Vstm2a ^−/−) mouse model. These mice exhibited no developmental defects and gained similar weight compared with wild‐type animals on both chow and high‐fat (HF) diets. Nevertheless, despite similar adipose tissue masses, VSTM2A‐null mice showed hypertrophic adipocytes, a phenotype associated with local inflammation, ectopic lipid storage, and impaired glucose and lipid metabolism. Transcriptomics analyses performed in WAT of young mice revealed that VSTM2A loss altered the expression of numerous genes in developing WAT, including some lysosomal genes. Here, we show that acute inhibition of lysosomal function early in life is sufficient to affect WAT development and cause adipocyte hypertrophy in mice. Overall, our work shows that VSTM2A is dispensable for the formation of adipose tissue in mice but that its loss causes adipocyte hypertrophy and impairs systemic metabolic homeostasis. Our study also highlights a key role of the lysosome in the initiation of adipogenesis.

METHODS

Animal care

All procedures were approved by Université Laval's Animal Ethics Committee (CPAUL) and followed Canadian Council on Animal Care guidelines. C57BL/6J mice (#000664, Jackson Laboratory) were housed individually in ventilated cages under a 12:12‐h light‐dark cycle at a temperature of 23 ± 1°C. Mice were fed an HF diet (60% of kilocalories from fat, D12492, Research Diets, Inc.) for 2 months. For lysosomal inhibition studies, C57BL/6J pups were injected subcutaneously with vehicle (dimethyl sulfoxide [DMSO]) or bafilomycin A1 (BafA1) (0.1 mg/kg) from postnatal day 1 (P1) to postnatal day 4 (P4) and then aged to adulthood for analysis.

Generation of VSTM2A‐null mice

The guidelines of the Canadian Council of Animal Care were followed for all experiments and were approved by McGill University's Animal Care Committee. Single guide RNA (sgRNA) sequences were designed using the Massachusetts Institute of Technology (MIT) clustered regularly interspaced short palindromic repeats (CRISPR) tool and synthesized at the McGill Integrated Core for Animal Modeling (MICAM) with a T7 transcription kit (AM1354, Thermo Fisher Scientific). CRISPR‐associated protein 9 (Cas9) mRNA was obtained from Sigma‐Aldrich. The sgRNA and Cas9 mRNA (20 ng/μL each) were microinjected into C57BL/6N zygotes and cultured to blastocysts (E3.5), and DNA was extracted for polymerase chain reaction (PCR) and sequencing. Selected sgRNA was then used for microinjection into C57BL/6N mouse zygotes, which were transferred to pseudo‐pregnant CD‐1 females, generating founder mice. Founders were next backcrossed three times in the C57BL/6J background. Following these crosses, Vstm2a ^+/− mice (87.5% C57BL/6J; 12.5% C57BL/6N) were crossed together to produce the wild‐type and knockout littermates used in the study.

Body composition

Body composition was measured in mice by dual‐energy x‐ray absorptiometry using the PIXIMUS densitometry apparatus (Lunar Corp.). This measurement was performed under isoflurane.

Blood metabolite measurement

Blood glucose levels were measured using a Roche Accu‐Chek Performa glucometer. Plasma was collected in EDTA‐treated syringes and stored at −80°C for metabolite analysis. The following commercial kits were used according to the manufacturer's instructions: insulin (SRI‐13K, MilliporeSigma); cholesterol (CH200, Randox Laboratories); triglycerides (TR22421, Thermo Fisher Scientific); nonesterified fatty acids (NEFA; 999‐34691, 995‐34791, 991‐34891, 993‐35191, and 276‐76491, Wake Chemicals); and adiponectin (80569, Crystal Chem Inc.).

Tissue processing and cell size measurement

Tissues were fixed in 10% formalin at 4°C for 48 h and then dehydrated, paraffin‐embedded, and sectioned into 10‐μm slices for hematoxylin & eosin staining. Images were captured with an Olympus BX60 microscope (Olympus Life Science). White adipocyte size was measured using ImageJ 1.53c on blinded samples, with 25 adipocytes randomly selected and counted per mouse (250–300 cells per genotype). For brown adipose tissue (BAT), nuclei within a defined area were counted, and the inverse of the nuclei to area ratio was used to estimate cell size.

Liver triglyceride extraction and quantification

Lipids were extracted from tissues as described [14]. The extracted triglycerides were suspended in isopropanol and quantified using the Infinity Triglyceride Liquid Stable Reagent (TR22421, Thermo Fisher Scientific).

Western blotting

Cells and tissues were lysed in a Triton‐X 100 buffer (50mM HEPES, pH 7.4; 2mM EDTA; 10mM sodium pyrophosphate; 10mM sodium glycerophosphate; 40mM sodium chloride; 50mM sodium fluoride; 2mM sodium orthovanadate; 1% Triton‐X 100; 0.1% sodium lauryl sulfate; 1% sodium deoxycholate; EDTA‐free protease inhibitors). After a 10‐min rotation at 4°C and centrifugation, protein concentrations were measured using Bradford reagent and analyzed via Western blotting on 10% Tris‐glycine gels (Life Technologies). Proteins were transferred to polyvinylidene fluoride membranes, blocked, and incubated with the following primary antibodies: phospho‐SMAD1/5/9 (13820S); β‐actin (4967); phospho‐LRP6 (2568S); SQSTM1/p62 (5114S); β‐tubulin (2128S); and Akt (4691S), all from Cell Signaling Technology. Secondary antibodies (7076S and 7074S, Cell Signaling Technology) were diluted 1:5,000, and blots were visualized using Amersham ECL Western Blotting Detection Reagent (RPN2106).

Quantitative real‐time PCR

mRNA was extracted from tissues using the RNeasy Lipid Tissue Mini Kit (74104, QIAGEN N.V.), and concentrations were measured by absorbance at 260 nm. Complementary DNA (cDNA) synthesis was conducted with the iScript Advanced cDNA Synthesis Kit for RT‐qPCR (Bio‐Rad Laboratories, Inc.). Both mRNA extraction and cDNA synthesis followed the manufacturer's instructions. The synthesized cDNA was diluted in DNase‐free water (1:15). mRNA transcript levels were measured in duplicates using a CFX96 or CFX384 Touch Real‐Time PCR Detection System (Bio‐Rad Laboratories) with SYBR Green detection. Gene expression was normalized to reference genes. The primers used are listed in Table S1.

Glucose tolerance test and insulin tolerance test

Mice were fasted for 6 h before being injected intraperitoneally with 1 g/kg of D‐glucose for the glucose tolerance test (GTT) or 0.75‐U/kg of insulin (Humulin, Lilly) for the insulin tolerance test (ITT). Blood glucose levels were measured from tail vein samples using a glucometer.

Induction of adipogenesis in vitro

3T3‐L1 cells were cultured to confluence in DMEM with 10% fetal bovine serum (FBS). Two days after reaching confluence (day 0), cells were exposed to an adipogenic cocktail (500μM 3‐isobutyl‐1‐methylxanthine [IBMX], 830nM insulin, 1μM dexamethasone). The medium was replaced 2 days later (day 2) with DMEM with 10% FBS containing 830nM insulin. Media was changed every 2 days until the end of the differentiation process. In order to assess lysosomal function, cells were treated with 10nM BafA1 (11038, Cayman Chemical) from day 0 to day 2, day 2 to day 4, or day 4 to day 6 of differentiation.

Microarrays

Whole‐genome gene expression was performed using the Affymetrix Mouse Gene 2.0 ST Array (Applied Biosystems). Briefly, RNA was labeled and hybridized using a standard Affymetrix protocol. The quality of arrays was judged using standard quality control parameters. Expression values were extracted using the Robust Multichip Average method [15, 16] implemented in the oligo package in R (R Project for Statistical Computing) [17].

Statistics

Statistical analyses were performed using Prism version 9.0.2 (GraphPad Software). In all figure panels, data represent mean ± SEM. When comparing two groups, significance was determined by two‐tailed, unpaired t test. When comparing more than two conditions, significance was determined by two‐way ANOVA.

RESULTS

Generation of a whole‐body VSTM2A knockout mouse model

In order to assess the impact of VSTM2A loss in vivo, we created a whole‐body VSTM2A knockout model using CRISPR/Cas9. An sgRNA‐targeting exon 2 of Vstm2a was microinjected into mouse zygotes along with Cas9 nuclease, resulting in a mutant allele with a 137‐base pair (bp) deletion that disrupts the Vstm2a coding sequence (Figure 1A–C). In order to determine whether VSTM2A is required for normal mouse development and growth, heterozygote mice carrying a VSTMA2‐null allele (Vstm2a ^+/−) were crossed to produce VSTMA2‐null mice (Vstm2a ^−/−). This approach revealed that VSTM2A knockout mice are viable and born at the expected Mendelian ratio (Figure 1D). These knockout mice developed normally, with no significant changes in body length and weight (Figure 1E,F). Moreover, we observed no change in food intake (Figure S1A). This indicates that VSTM2A loss does not cause noticeable developmental defects in mice.

FIGURE 1.

Generation of a whole‐body V‐set and transmembrane domain‐containing 2A (VSTM2A) knockout mouse model. (A) Overview of the strategy used to generate VSTM2A‐null mice using the clustered regularly interspaced short palindromic repeats (CRISPR)/CRISPR‐associated protein 9 (Cas9) method. The single guide RNA used led to the deletion of 137 bp of genomic DNA, including 47 bp of the exon 2 of Vstm2a. (B) Genotyping analysis performed from tails collected from wild‐type (Vstm2a ^+/+), heterozygote (Vstm2a ^+/−), and knockout (Vstm2a ^−/−) mice. (C) Quantitative PCR performed from tissues collected from wild‐type (Vstm2a ^+/+; n = 11) and knockout (Vstm2a ^−/−; n = 9) mice. Tissues were collected at postnatal day 21. (D) Mendelian ratios calculated from crossing heterozygote (Vstm2a ^+/−) mice. (E) Growth curves and body length measurements of male Vstm2a ^+/+ (n = 11) and Vstm2a ^−/− (n = 10) mice. (F) Growth curves and body length measurements of female Vstm2a ^+/+ (n = 12) and Vstm2a ^−/− (n = 11) mice. In panels E and F, data represent mean ± SEM. Significance was determined by two‐tailed, unpaired t test; ^**** p < 0.0001. [Color figure can be viewed at wileyonlinelibrary.com]

VSTM2A loss does not prevent adipose tissue formation and expansion but causes adipocyte hypertrophy

In order to investigate the role of VSTM2A in adipose tissue development and expansion, we compared Vstm2a wild‐type and VSTMA2‐null mice fed either a standard chow or HF diet for 2 months. As shown in Figure 2A,C, loss of VSTM2A did not affect body weight in both male and female mice fed either a chow or HF diet. Vstm2a deletion did not affect lean or fat mass (Figure S1B,C). Also, the weight of various organs did not differ consistently between Vstm2a ^+/+ and Vstm2a ^−/− mice (Figure 2B,D). Although VSTM2A  depletion has been previously shown to impair adipogenesis in vitro [12], we found no consistent effect of its deletion on the weight of various WAT and BAT depots (Figure 2B,D). These observations were consistent in both male and female mice. Altogether, these results indicate that VSTM2A loss does not prevent the formation and the obesogenic expansion of adipose tissue in mice.

FIGURE 2.

V‐set and transmembrane domain‐containing 2A (VSTM2A) loss does not prevent adipose tissue development and expansion. (A) Body weight of wild‐type and knockout male mice fed either a chow diet (Vstm2a ^+/+ [n = 11] and Vstm2a ^−/− [n = 10]) or a high‐fat (HF) diet (Vstm2a ^+/+ [n = 12] and Vstm2a ^−/− [n = 12]) for 9 weeks. (B) Weight of the tissues collected from the mice described in panel A. (C) Body weight of wild‐type and knockout female mice fed either a chow diet (Vstm2a ^+/+ [n = 12] and Vstm2a ^−/− [n = 11]) or an HF diet (Vstm2a ^+/+ [n = 12] and Vstm2a ^−/− [n = 12]) for 9 weeks. (D) Weight of the tissues collected from the mice described in panel C. In all panels, data represent mean ± SEM. Significance was determined by two‐tailed, unpaired t test; *p < 0.05. [Color figure can be viewed at wileyonlinelibrary.com]

In order to further characterize the impact of VSTM2A loss on adipose tissue physiology, we next measured the size of adipocytes in various depots of Vstm2a ^+/+ and Vstm2a ^−/− mice fed either a chow or HF diet. Even though WAT mass was similar between Vstm2a ^+/+ and Vstm2a ^−/− mice, we found that adipocytes from VSTM2A‐null mice were significantly enlarged compared with control mice (Figure 3A–C). This effect was more pronounced in retroperitoneal WAT (rWAT) compared with inguinal WAT (iWAT) and epididymal WAT (eWAT). Adipocyte hypertrophy was also observed in BAT of VSTM2A‐null animals (Figure 3D). Together, these observations indicate that VSTM2A loss causes adipocyte hypertrophy in mice. Although we did not directly measure the total number of adipocytes per fat depot, these results also suggest that VSTM2A loss reduces fat cell number in mice.

FIGURE 3.

V‐set and transmembrane domain‐containing 2A (VSTM2A) loss induces adipocyte hypertrophy. Histological analyses of (A) retroperitoneal white adipose tissue (rWAT), (B) inguinal WAT (iWAT), (C) epididymal WAT (eWAT), or (D) brown adipose tissue (BAT) collected from wild‐type and knockout male mice fed either a chow diet (Vstm2a ^+/+ [n = 11] and Vstm2a ^−/− [n = 10]) or a high‐fat (HF) diet (Vstm2a ^+/+ [n = 12] and Vstm2a ^−/− [n = 12]) for 9 weeks. Representative pictures are shown. In panels A, B, and C, the distribution of adipocyte population is presented for each group (25 adipocytes/mouse, n = 250–300 adipocytes). In panel D, the number of nuclei per area was counted, and this relationship (nuclei/area)⁻¹ was calculated to estimate brown fat cell size. In all panels, data represent mean ± SEM. Significance was determined by two‐tailed, unpaired t test; *p < 0.05, **p < 0.01, and ^**** p < 0.0001. [Color figure can be viewed at wileyonlinelibrary.com]

VSTM2A loss promotes adipose and liver inflammation

The increase in adipocyte size due to defective adipogenesis can drive inflammation and impair metabolism [6]. Because the effect of VSTM2A loss on adipocyte hypertrophy was more important in rWAT and BAT, we next tested whether this was associated with adipose dysfunction by measuring the expression of genes coding for proteins controlling both inflammation and lipid metabolism. In rWAT, VSTM2A deficiency increased inflammatory markers (e.g., arginase 1 [Arg1], heme oxygenase 1 [Hmox1], monocyte chemoattractant protein 1 [Mcp1; also known as C‐C motif chemokine ligand 2, Ccl2], secreted phosphoprotein 1 [Spp1]) and decreased expression of genes related to lipid uptake and adipogenesis (e.g., acyl‐CoA synthetase long‐chain family member 1 [Acsl1], lipoprotein lipase [Lpl], peroxisome proliferator‐activated receptor gamma 2 [Pparg2]; Figure 4A). However, lipid oxidation and thermogenesis genes were unaffected (Figure 4A). We next measured the same panel of genes in BAT of Vstm2a ^+/+ and Vstm2a ^−/− mice. As observed in rWAT, loss of VSTM2A led to a significant increase in the expression of many inflammatory markers, including Hmox1, the macrophage cell surface receptors Cd68 and F4/80 (also known as adhesion G protein‐coupled receptor E1 [Adgre1]), and several proinflammatory cytokines, including interleukin‐1b (Il1b), Mcp1, and tumor necrosis factor α (Tnfa; Figure 4B). Quantification of CD68 by immunohistochemistry confirmed the rise in macrophage infiltration in adipose tissue of VSTM2A‐null mice fed an HF diet (Figure S2A,B). Contrary to rWAT, the expression of lipogenic and adipogenic genes was not reduced in BAT of Vstm2a ^−/− mice. In fact, we found a slight but significant rise in the expression of Lpl and Acsl1 in BAT of mice fed an HF diet. Finally, Vstm2a deletion did not impact the expression of classical thermogenic genes in BAT (Figure 4B). Overall, these findings indicate that VSTM2A loss promotes markers of inflammation and alters markers of lipid metabolism in both WAT and BAT.

FIGURE 4.

V‐set and transmembrane domain‐containing 2A (VSTM2A) loss increases local inflammation and ectopic lipid deposition. (A) Quantitative PCR (qPCR) analysis of retroperitoneal white adipose tissue (rWAT) collected from wild‐type and knockout male mice fed either a chow diet (Vstm2a ^+/+ [n = 11] and Vstm2a ^−/− [n = 10]) or a high‐fat (HF) diet (Vstm2a ^+/+ [n = 12] and Vstm2a ^−/− [n = 12]) for 9 weeks. (B) qPCR analysis of brown adipose tissue (BAT) collected from wild‐type and knockout male mice fed either a chow diet (Vstm2a ^+/+ [n = 11] and Vstm2a ^−/− [n = 10]) or an HF diet (Vstm2a ^+/+ [n = 12] and Vstm2a ^−/− [n = 12]) for 9 weeks. (C) Histological analysis and triglyceride content of liver samples collected from wild‐type and knockout male mice fed either a chow diet (Vstm2a ^+/+ [n = 11] and Vstm2a ^−/− [n = 10]) or an HF diet (Vstm2a ^+/+ [n = 12] and Vstm2a ^−/− [n = 12]) for 9 weeks. Representative pictures are shown. The triglyceride content is presented in the right section of each panel. (D) qPCR analysis of liver collected from wild‐type and knockout male mice fed either a chow diet (Vstm2a ^+/+ [n = 11] and Vstm2a ^−/− [n = 10]) or an HF diet (Vstm2a ^+/+ [n = 12] and Vstm2a ^−/− [n = 12]) for 9 weeks. In all panels, data represent mean ± SEM. Significance was determined by two‐tailed, unpaired t test; *p < 0.05, **p < 0.01, and ***p < 0.001. [Color figure can be viewed at wileyonlinelibrary.com]

Adipocyte hypertrophy and adipose tissue inflammation are often linked to ectopic lipid deposition in peripheral tissues, especially the liver. In order to test whether adipocyte hypertrophy and inflammation present in adipose tissue of Vstm2a ^−/− mice were associated with  defects in liver metabolism, hepatic triglycerides were quantified. As shown in Figure 4C, Vstm2a ^−/− mice fed a chow diet showed increased lipid deposition in their liver compared with wild‐type mice. This difference was not observed in mice fed an HF diet (Figure 4C). Gene expression analyses next revealed a proinflammatory profile in the liver of VSTM2A‐null mice. For instance, we measured an increase in the expression of hepatic Arg1, Hmox1, IL12 subunit p40 (Il12p40), Mcp1, and Tnfa in Vstm2a ^−/− mice fed a chow diet (Figure 4D). A similar gene expression profile was found in mice fed an HF diet. These results indicate that the loss of VSTM2A induces ectopic lipid deposition and inflammation in the liver, which are common characteristics of an unhealthy metabolic state.

VSTM2A loss impairs glucose and lipid homeostasis

In order to define whether adipose tissue hypertrophy and the proinflammatory profile linked to VSTM2A loss translated into alterations in systemic metabolic homeostasis, we next measured several metabolites and hormones in the plasma of Vstm2a ^+/+ and Vstm2a ^−/− mice. A significant rise in fasting glucose was observed in Vstm2a ^−/− mice fed a chow diet (Table 1). This effect occurred despite similar circulating insulin levels, indicating a reduction in insulin sensitivity in this model. Cholesterol levels were significantly higher in chow‐fed Vstm2a ^−/− mice compared with controls, but we found no changes in circulating triglycerides and NEFA between the genotypes (Table 1). Finally, adiponectin levels were not different between Vstm2a ^+/+ and Vstm2a ^−/− mice. Overall, these results indicate that VSTM2A loss impacts glucose and lipid homeostasis even in lean mice fed a normal chow diet. We next looked at the same blood metabolites in HF diet‐fed mice. Fasting glucose was not different between Vstm2a ^+/+ and Vstm2a ^−/− mice. However, we measured an important rise in insulin levels in VSTM2A‐null animals, further supporting the presence of insulin resistance in this model. Cholesterol levels were similar between genotypes, but we found a significant rise in circulating triglycerides and NEFA in Vstm2a ^−/− mice (Table 1). Overall, these findings indicate that Vstm2a deletion negatively impacts systemic glucose and lipid homeostasis in mice. These observations also support a key role of diet in affecting metabolism in VSTM2A‐null mice.

TABLE 1.

Impact of VSTM2A loss on blood metabolites.

	Chow diet		HF diet
	Vstm2a+/+	Vstm2a−/−	Vstm2a+/+	Vstm2a−/−
Glucose, mmol/L	7.35 ± 0.12	8.25 ± 0.30*	9.60 ± 0.47	10.38 ± 0.34
Insulin, ng/mL	0.77 ± 0.22	0.65 ± 0.14	1.42 ± 0.26	2.40 ± 0.35*
Cholesterol, mmol/L	2.69 ± 0.20	3.49 ± 0.10**	4.82 ± 0.12	4.98 ± 0.25
Triglycerides, mmol/L	0.34 ± 0.03	0.43 ± 0.05	0.42 ± 0.02	0.51 ± 0.04*
NEFA, mEq/L	0.24 ± 0.03	0.19 ± 0.01	0.25 ± 0.02	0.34 ± 0.03*
Adiponectin, μg/mL	5.44 ± 0.31	6.11 ± 0.47	6.81 ± 0.21	7.36 ± 0.50

Note: Blood metabolite analyses performed from plasma collected from wild‐type and knockout male mice fed either a chow diet (Vstm2a ^+/+ [n = 11] and Vstm2a ^−/− [n = 10]) or a high‐fat (HF) diet (Vstm2a ^+/+ [n = 12] and Vstm2a ^−/− [n= 12]) for 9 weeks. Mice were fasted from 6 h prior to these analyses. Data represent mean ± SEM. Significance was determined by two‐tailed, unpaired t test.

Abbreviations: HF, high‐fat; NEFA, nonesterified fatty acids; VSTM2A, V‐set and transmembrane domain‐containing 2A.

^* p < 0.05.

^** p < 0.01.

In order to further define the impact of VSTM2A loss on glucose metabolism, we next performed the GTT and ITT. In chow‐fed mice, we found that glucose levels remained slightly higher in Vstm2a ^−/− mice compared with wild‐type mice, but this effect was not significant. Interestingly, glucose intolerance was more pronounced in VSTM2A‐null mice fed an HF diet (Figure 5A). As shown in Figure 5A, Vstm2a ^−/− mice failed to normalize circulating glucose levels to the levels of wild‐type animals. Consistently, we observed that insulin sensitivity was also impaired in these mice (Figure 5B). These results confirm that VSTM2A loss impairs systemic glucose homeostasis in mice.

FIGURE 5.

Loss of V‐set and transmembrane domain‐containing 2A (VSTM2A) impairs glucose homeostasis. (A) Glucose tolerance test (GTT) performed in wild‐type and knockout male mice fed either a chow diet (Vstm2a ^+/+ [n = 11] and Vstm2a ^−/− [n = 10]) or a high‐fat (HF) diet (Vstm2a ^+/+ [n = 12] and Vstm2a ^−/− [n = 12]) for 9 weeks. Data represent mean ± SEM. Significance was determined by two‐way ANOVA. The area under the curve (AUC) is also presented in these panels. For the AUC, significance was determined by two‐tailed, unpaired t test. (B) Insulin tolerance test (ITT) performed in wild‐type and knockout male mice fed either a chow diet (Vstm2a ^+/+ [n = 11] and Vstm2a ^−/− [n = 10]) or an HF diet (Vstm2a ^+/+ [n = 12] and Vstm2a ^−/− [n = 12]) for 9 weeks. Data represent mean ± SEM. Significance was determined by two‐way ANOVA. *p < 0.05 and **p < 0.01. [Color figure can be viewed at wileyonlinelibrary.com]

VSTM2A loss impacts gene expression in developing adipose tissue

The results presented earlier indicate that, although VSTM2A is dispensable for the formation of adipose tissue in mice, its deletion translates into adipocyte hypertrophy, local inflammation, ectopic lipid deposition, and impaired metabolic homeostasis. We have previously shown that VSTM2A is highly expressed in the early phases of WAT development in mice and humans [12, 13]. Previous work from our group showed that VSTM2A controls preadipocyte determination in vitro, at least in part by amplifying bone morphogenetic protein (BMP) signaling [12], a key pathway regulating the early steps in adipogenesis [18, 19, 20]. In order to define the early effects of VSTM2A loss on adipose tissue development in an in vivo, physiological context, eWAT was collected from mice at a young age (i.e., postnatal day 21 [P21]), and phosphorylation of SMAD family member 1/5/9 (mothers against decapentaplegic homolog; SMAD1/5/9) was measured as a marker of BMP activation. eWAT was selected because this tissue develops in mice after birth, allowing us to catch the early effects of VSTM2A loss in mice. Although we found a tendency for lower SMAD1/5/9 phosphorylation in eWAT of VSTM2A‐null mice, this effect did not reach significance (Figure S3A,B). VSTM2A was also reported to impact proto‐oncogene Wnt signaling in cancer cells by binding and repressing low‐density lipoprotein receptor‐related protein 6 (LRP6) [21]. Because Wnt activation represses adipogenesis [22, 23, 24], we next tested whether VSTM2A could affect this pathway in vivo. As presented in Figure S3A,B, we found a tendency for a rise in LRP6 phosphorylation that was not significant. Together, these results indicate that Vstm2a deletion does not translate into significant changes in BMP and Wnt signaling in adipose tissue.

In order to further investigate the early effects of VSTM2A loss on adipose tissue development, transcriptomics analyses were next performed in eWAT collected from P21 mice (Figure 6A). As depicted in Figure 6B, we found that VSTM2A loss induced significant transcriptional changes in eWAT. Comprehensive analysis of the genes that were differentially expressed (p < 0.01) with Metascape revealed an enrichment for several pathways and cellular processes, including “lysosome,” “fatty acid metabolic process,” “regulation of lipid metabolic processes,” and “signaling by receptor tyrosine kinase” (Figure 6C) [25]. Quantitative PCR (qPCR) analyses confirmed the reduction in the expression of lysosomal genes in eWAT of P21 mice (Figure 6D). In addition to these genes, we also observed a reduction in the expression of fibroblast growth factor 1 (Fgf1), insulin receptor substrate 1 (Irs1), and Irs2 in response to VSTM2A loss. These genes, coding for key proteins regulating receptor tyrosine kinase (RTK) activation, have been previously shown to play key roles in fat cell development [25, 26, 27, 28]. Interestingly, most of the alterations in gene expression observed in developing eWAT were no longer observed in eWAT of adult mice (Figure S4A), indicating that VSTM2A loss specifically affected gene expression during the developmental phase of WAT. As shown in Figure S4A, only Irs1 and cathepsin O (Ctso) expression was still significantly reduced in eWAT of adult VSTM2A‐null mice.

FIGURE 6.

V‐set and transmembrane domain‐containing 2A (VSTM2A) loss reduces lysosomal gene expression in developing white adipose tissue (WAT). (A) Schematic representation of the experiments designed to test the transcriptional impact of VSTM2A loss in developing adipose tissue. WAT was collected from Vstm2a ^+/+ and Vstm2a ^−/− littermates at postnatal day 21 (P21), and transcriptomics analyses were performed. (B) Heat map presenting the expression profile of the 249 genes differentially expressed (p < 0.01) between Vstm2a ^+/+ and Vstm2a ^−/− mice. (C) Gene ontology analysis performed with Metascape on the list of genes differentially expressed (p < 0.01) between WAT of Vstm2a ^+/+ and Vstm2a ^−/− mice. (D) Quantitative PCR analysis of epididymal WAT (eWAT) collected from wild‐type and knockout male mice at P21 (Vstm2a ^+/+ [n = 11] and Vstm2a ^−/− [n = 9]). In all panels, data represent mean ± SEM. In panel D, significance was determined by two‐tailed, unpaired t test; *p < 0.05. [Color figure can be viewed at wileyonlinelibrary.com]

We next performed studies to test whether the alterations in gene expression and changes in adipocyte development brought by VSTM2A loss in vivo could be reproduced in vitro. Specifically, we collected SVF from eWAT of wild‐type and VSTM2A‐null mice at P21. Differentiation was induced in these cells, and expression of classic adipogenic markers was measured. We also measured the expression of lysosomal genes and other genes differentially expressed in eWAT of VSTM2A‐null mice. As presented in Figure S5A, loss of VSTM2A did not impair adipogenesis in these models, and we found no alterations in the expression of lysosomal genes and other genes reported to be reduced in WAT of VSTM2A‐null mice. This indicates that this in vitro model cannot mimic the in vivo phenotype and the complexity of adipogenesis in mice.

Repressing lysosomal function early during adipogenesis blocks adipocyte development in vitro and cause adipocyte hypertrophy in vivo

As shown in Figure 6C,D, loss of VSTM2A was associated with the reduction in the expression of lysosomal genes in developing WAT. In order to investigate the importance of the lysosome in regulating early steps in fat cell development, preadipocytes were exposed to BafA1, a macrolide antibiotic that interacts with V‐ATPase to block lysosomal acidification. Specifically, 3T3‐L1 cells were plated to confluence, and adipogenesis was induced following a standardized approach. BafA1 was added at a different time following the induction of adipogenesis to test the impact of lysosomal inhibition at various stages of adipocyte differentiation (Figure 7A). We first validated the efficiency of BafA1 to impair lysosomal function in cells by measuring sequestosome‐1 (SQSTM1/p62) levels in cells. As expected, a clear rise in this marker was observed in response to BafA1 (Figure 7B). Strikingly, we found that impairing lysosomal function with BafA1 severely impaired adipogenesis, especially when the compound was added in the early phases of adipogenesis. As shown in Figure 7C, cells treated with BafA1 from day 0 to day 2 (d0➔d2) of the adipogenic protocol were incapable of differentiating and accumulating lipids over the course of the experiment. The expression of key adipogenic regulators such as Pparg, CCAAT/enhancer binding protein α (Cebpa), CCAAT/enhancer binding protein β (Cebpb), and zinc finger protein 423 (Zfp423) was severely reduced in these cells (Figure 7D). Consistently, we also measured an important decrease in the expression of terminal markers of differentiation, including fatty acid binding protein 4 (Fabp4), Acsl, Lpl, and sterol regulatory element binding transcription factor 1 (Srebf1). Interestingly, cells treated with BafA1 from day 2 to day 4 (d2➔d4) also showed adipogenic defects. However, the impairment in adipogenesis was less profound, and we were able to observe differentiated adipocytes in this condition. The expression of adipogenic genes was also less severely impaired compared with the d0➔d2 treatment. Supporting the idea that lysosomal function plays a specific role in the early development of adipocytes, we found that cells treated in the late phase of differentiation (i.e., d4➔d6) expressed normal levels of most adipogenic regulators and terminal markers of adipogenesis (Figure 7D). Visually, the amount of mature fat cells was still reduced compared with control cells, but the effect of BafA1 from d4➔d6 on adipogenesis was clearly less profound compared with the cells treated from d2➔d4 and d0➔d2.

FIGURE 7.

Repressing lysosomal function early during adipogenesis blocks adipocyte development and causes adipocyte hypertrophy in vivo. (A) Overview of the experiment designed to test the impact of lysosomal inhibition on adipogenesis. Briefly, 3T3‐L1 cells were plated to confluence and treated with bafilomycin A1 (BafA1) from day 0 to day 2 (d0➔d2), from day 2 to day 4 (d2➔d4), or from day 4 to day 6 (d4 ➔d6) of the adipogenic protocol. (B) 3T3‐L1 cells were treated with vehicle (DMSO; n = 4) or BafA1 (n = 4) for 6 h. Proteins were extracted, and Western blots were performed for the indicated proteins. (C) 3T3‐L1 cells were treated as described in panel A, and pictures were taken after 6 days of adipogenic differentiation. Representative pictures are shown. (D) RNA was extracted from the 3T3‐L1 treated as described in panel A, and gene expression was measured by quantitative PCR (n = 4/condition). (E) Schematic representation of the experiments designed to test the impact of lysosomal inhibition early in life on adipose tissue development. Briefly, pups were injected with BafA1 (0.1 mg/kg of body weight) during 4 consecutive days from postnatal day 1 (P1) to P4. Mice were then allowed to grow and were sacrificed between the age of 16 to 18 weeks. (F) Male and female pups (P1) were injected with vehicle (DMSO; n = 3) or BafA1 (n = 3), as described in panel E, and were sacrificed at P4, 2 h after the last injection. Proteins were extracted from the liver and Western blots were performed against the indicated proteins. (G) Body weight and (H) organ weight of female adult mice treated with vehicle (DMSO; n = 13) or BafA1 (n = 7), as described in panel E. (I) Histological analysis of inguinal WAT (iWAT) collected from adult female mice treated with vehicle (DMSO; n = 13) or BafA1 (n = 7), as described in panel E. Representative pictures are shown. On the right side of the panel, the distribution of adipocyte population and the size of fat cells are presented for each group (25 adipocytes/mouse, n = 325 adipocytes for mice treated with vehicle [DMSO] and n = 175 adipocytes for BafA1‐treated mice). (J) Histological analysis of retroperitoneal WAT (rWAT) collected from adult female mice treated with vehicle (DMSO; n = 13) or BafA1 (n = 6), as described in panel E. Representative pictures are shown. On the right side of the panel, the distribution of adipocyte population and the size of fat cells are for each group (25 adipocytes/mouse, n = 325 adipocytes for mice treated with vehicle [DMSO] and n = 150 adipocytes for BafA1‐treated mice). In all panels, data represent mean ± SEM. In panels D, G, H, I, and J, significance was determined by two‐tailed, unpaired t test; **p < 0.01, ***p < 0.001, and ^**** p < 0.0001. [Color figure can be viewed at wileyonlinelibrary.com]

In order to test whether inhibition of the lysosome early during adipose tissue development could be sufficient to impact WAT structure and organization in adult mice, pups at P1 were injected with vehicle or BafA1 for 4 consecutive days. The timing of injection was chosen based on previous studies that have shown that key adipogenic events take place during this time frame in mice [29]. Following injections, mice were allowed to grow normally until adulthood (16–18 weeks), and adipose tissue was collected to define whether early impairment in lysosome function can cause adipogenic lesions that persist in time in adult mice. An overview of the experimental scheme is presented in Figure 7E. As presented in Figure 7F, BafA1 was efficient in rising SQSTM1/p62 in tissues of female pups. Unexpectedly, BafA1 injections failed in producing the same effect in males (Figure 7F). Based on these results, only female mice were used for the full study. Compared with controls, mice injected with BafA1 did not show obvious physiological differences when they reached adulthood. We found no effect of BafA1 on body weight (Figure 7G). Also, the weight of individual tissues did not differ between the groups (Figure 7H). We next performed histological analyses of adipose tissue to evaluate the impact of lysosomal inhibition on adipocyte size. Strikingly, we observed that mice injected with BafA1 postnatally showed a significant increase in fat cell size later in adult life (Figure 7I,J). Importantly, this effect was observed in both iWAT and rWAT. No difference in adipocyte cell size was found in BAT (Figure S6A), a fat depot that is already developed and fully functional at birth [30]. Although we observed a trend toward lower glucose tolerance in BafA1 mice fed a chow diet, this effect was not significant (Figure S6B). The ITT, fasting metabolites, liver triglycerides, and inflammation markers were not different versus control mice (Figure S6C‐F), indicating that adipocyte hypertrophy in this model was not sufficient to cause metabolic disturbances. Altogether, these results indicate that acute impairment in lysosomal function early in life is sufficient to impact the normal development of WAT and to cause adipocyte hypertrophy in adults.

DISCUSSION

Although the role of VSTM2A in adipogenesis in vitro is established [12], its impact on adipose tissue development in vivo has never been tested, to our knowledge. Here, we report that VSTM2A loss does not prevent adipose tissue formation but causes adipocyte hypertrophy in mice. This phenotype is linked to inflammation, ectopic lipid storage, and impaired glucose and lipid metabolism. These results indicate that VSTM2A plays a role in normal adipose tissue formation and in the maintenance of systemic metabolic homeostasis.

We previously reported that VSTM2A is highly expressed during early stages of WAT development [12, 13]. In mice, elevated VSTM2A levels were observed in late embryonic stages (i.e., E16.5–E20) in iWAT and early postnatal days (i.e., P4–P7) in eWAT [12, 13]. In humans, higher VSTM2A levels were found in WAT between gestational weeks 17 and 21 [12]. These timings coincide with key events in preadipocyte recruitment and maturation [2, 31, 32]. VSTM2A expression decreases as adipogenesis progresses and is low in fully developed fat pads [12, 13], suggesting that VSTM2A might play a specific role in controlling the embryonic and/or postnatal development of adipose tissue. Supporting this possibility, we show herein that VSTM2A loss impacted gene expression in adipose tissue of young mice before weaning but that these defects were not found in adipose tissue of adult animals. Interestingly, an emerging body of evidence shows that important differences exist between developmental and adult adipogenesis [4, 33, 34]. These studies have shown that early versus adult adipogenesis involves distinct adipose progenitors and effectors and is controlled by different extracellular cues. Based on its expression pattern, regulation during adipogenesis, and impact on gene expression, we propose that VSTM2A might preferentially affect the embryonic/postnatal development of adipose tissue.

It is important to point out that adipogenesis remained possible in the absence of VSTM2A. As shown here, VSTM2A‐null mice developed WAT and BAT depots and gained weight when fed an obesogenic diet. Moreover, SVF isolated from WAT of VSTM2A‐null mice did not show clear adipogenic defects. These observations are somehow surprising considering that we previously showed that VSTM2A depletion completely repressed adipogenesis in immortalized white and brown preadipocytes [12]. VSTM2A is highly expressed and secreted by committed preadipocytes in vitro [12]. In mice, single‐cell RNA sequencing analyses revealed that VSTM2A is expressed by a poorly characterized cell population present in SVF [10]. Other studies have indicated that, even within a single adipose depot, there appear to be multiple subpopulations of adipose progenitors and adipocytes [11]. Although we do not know the identity of the cells that are sensitive to VSTM2A, it is tempting to propose that the local production of VSTM2A could preferentially impact the function of some cells versus others. The presence of VSTM2A in the microenvironment might also impact the communication among the various cell populations that compose the organ through a non‐cell autonomous mode, which could be important for the normal development of adipose tissue. The absence of VSTM2A might thus impact only a portion of adipose progenitors, an effect that could prevent the healthy development of adipose tissue without blocking its general growth. Finally, we cannot exclude the possibility that VSTM2A loss could be partially compensated by the expression of other pro‐adipogenic factors. For instance, our transcriptomics experiments revealed that the loss of VSTM2A was associated with a consistent rise in the expression of angiopoietin‐like 8 (Angptl8; Table S2 and Figure S7A‐B), a secreted glycoprotein that has been previously reported to promote adipogenesis in various experimental contexts [35, 36]. This rise in Angptl8 expression was observed in WAT and BAT of young, but not adult, mice. Why exactly VSTM2A loss triggers Angptl8 expression in developing adipocytes is unclear, but this relationship warrants further investigations.

How VSTM2A impacts adipocyte development is an important but still unresolved question. In previous work, we have shown that VSTM2A controls preadipocyte determination in vitro by amplifying BMP signaling. Interestingly, the activation of BMP signaling by VSTM2A was observed specifically in cell lines with established adipogenic potential, but not in nonadipogenic cell lines, suggesting that adipose progenitors might be particularly sensitive to VSTM2A [12]. Here, in young mice, a trend toward lower BMP signaling was noted in VSTM2A‐null WAT, although it was not statistically significant. Because adipose tissue contains only a small fraction of preadipocytes and because these signaling pathways are highly dynamic during development, measuring changes specifically in this cell population remains technically challenging. VSTM2A was also reported to impact Wnt signaling in colorectal cancer cells [21]. In this study, it was shown that VSTM2A antagonizes the Wnt pathway by directly binding to LRP6. Because Wnt signaling is a strong inhibitor of adipogenesis [22, 23, 24], this suggests that VSTM2A secretion might amplify fat cell development by inhibiting this pathway. Here, we found no significant effect of VSTM2A loss on LRP6 phosphorylation in adipose tissue of VSTM2A‐null mice. Why VSMT2A loss impacts the Wnt pathway in cancer cells, but not in adipose cells, is an interesting question that warrants additional studies.

In order to investigate the early effects of VSTM2A loss on adipose tissue development, we conducted transcriptomics analyses of developing WAT. This revealed that Vstm2a deletion affects the expression of several genes involved in growth factor signaling and adipogenesis, such as Irs1, Irs2, and Fgf1 [25, 27, 28]. Interestingly, there was also a decrease in lysosomal gene expression in WAT of VSTM2A‐null mice. The lysosome is an organelle that plays important roles in regulating the degradation of macromolecules, including proteins, nucleic acids, lipids, and carbohydrates [37]. Lysosomes are also essential for the recycling of damaged organelles, a process known as autophagy [38]. Several reports have shown that autophagy is highly active and essential in the early steps of adipocyte development [39, 40, 41]. Inhibition of autophagy early during adipocyte development was reported to impair fat cell development by blocking mitotic clonal expansion and mitochondrial remodeling [41]. Chaperone‐mediated autophagy, a selective form of lysosomal protein degradation, has also been shown to control early preadipocyte development by regulating proliferation, metabolism, and transcription [42]. In line with these findings, lysosome‐mediated ferritinophagy was recently shown to control the initiation of adipocyte differentiation [43]. Here, we confirmed the role of the lysosome in regulating the early development of adipocytes. We show that blocking the lysosome during the early phases of adipogenesis severely impairs fat cell development in vitro. Importantly, we observed that acute injections of BafA1 to pups following birth (P1–P4) was sufficient to cause adipocyte hypertrophy in adult mice several weeks later. These results indicate that alterations in lysosomal function early in life can have a significant impact on WAT architecture in adults. How precisely lysosomal inhibition impacted early WAT development in our study still remains to be determined. Also, although our work reveals an intriguing association between VSTM2A and the lysosome, it is extremely important to point out that the precise nature of this relationship remains to be determined. Because VSTM2A loss impacts the expression of genes involved in several other processes, it is clear that the effects of VSTM2A loss on adipose tissue cannot be solely attributed to its impact on lysosomal gene expression. Additional studies are needed to clarify how VSTM2A impacts adipogenesis in vivo.

Adipocyte hypertrophy, a key feature of unhealthy adipose tissue, is linked to metabolic complications such as dyslipidemia, ectopic fat deposition, glucose intolerance, and insulin resistance [6]. This condition is associated with immune cell infiltration and local inflammation in hypertrophic WAT. This relationship was also reported to occur in BAT [44, 45, 46]. Our study found that VSTM2A loss leads to adipocyte hypertrophy and inflammation in both tissues. This unhealthy adipose tissue profile is coupled with impaired glucose and lipid metabolism. The metabolic disturbances linked to VSTM2A loss were observed in lean mice, and most of them were amplified in response to HF diet feeding, indicating that VSTM2A deficiency impairs the metabolic adaptation to nutrient excess. Because our research used a whole‐body VSTM2A‐null mouse model, we cannot exclude the possibility that VSTM2A might affect systemic metabolism through other metabolic tissues such as skeletal muscle or the liver. The generation of additional mouse models allowing the conditional deletion of Vstm2a  in specific tissues and cell types will be needed to better define its role in regulating systemic metabolism and fat cell development.

In conclusion, our work indicates that VSTM2A loss causes adipocyte hypertrophy and impairs metabolic homeostasis in mice. We show that Vstm2a deletion is associated with alterations in the expression of several genes in developing WAT, and we provide evidence that maintaining lysosomal function is crucial for the normal development of adipose tissue.

CONFLICT OF INTEREST STATEMENT

The authors declared no conflicts of interest.

Supporting information

Figure S1. VSTM2A loss does not affect food intake and body composition. (A) Food intake in Vstm2a ^+/+ and Vstm2a ^−/− mice fed either chow or a HF diet. (B) Lean mass in Vstm2a ^+/+ and Vstm2a ^−/− mice fed either chow or a HF diet. (C) Fat mass in Vstm2a ^+/+ and Vstm2a ^−/− female mice fed either chow or a HF diet. Body composition was measured in mice by dual‐energy X‐ray absorptiometry (DEXA) using the PIXIMUS mouse densitometry apparatus (Lunar Corp.). The following groups of mice were used in these studies: Male mice fed chow diet (Vstm2a ^+/+ [n = 11] and Vstm2a ^−/− [n = 10]); male mice fed HF diet (Vstm2a ^+/+ [n = 12] and Vstm2a ^−/− [n = 12]); female mice fed chow diet (Vstm2a ^+/+ [n = 12] and Vstm2a ^−/− [n = 11]); female mice fed a HF diet (Vstm2a ^+/+ [n = 12] and Vstm2a ^−/− [n = 12]). Data represent the mean ± SEM. Significance was determined by two‐tailed, unpaired t test, comparing the genotype effects (*p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001).

Figure S2. VSTM2A loss increases local inflammation in adipose tissues. Immunohistochemistry (IHC) staining of the macrophage cell surface receptor CD68 (black arrows) in (A) rWAT from Vstm2a ^+/+ (n = 11) and Vstm2a ^−/− (n = 12) mice fed HF diet for 9 weeks. The left panel shows representative pictures and the right panel shows CD68 staining quantification. (B) CD68 staining in BAT from Vstm2a ^+/+ (n = 12) and Vstm2a ^−/− (n = 12) mice fed HF diet for 9 weeks. The left panel shows representative pictures and the right panel shows CD68 staining quantification. Data represent the mean ± SEM. Significance was determined by two‐tailed, unpaired t test (*p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001).

Figure S3. VSTM2A does not significantly impact BMP and WNT signaling in vivo. (A) Proteins were extracted from eWAT of Vstm2a+/+ [n = 5] and Vstm2a‐/‐ [n = 4] mice at post‐natal day 21 [P21] and western blots were performed for the indicated proteins. (B) Quantification of the blots presented in A. The results are normalized to ACTIN levels.

Figure S4. The expression of lysosomal genes is minimally affected in adipose tissue of VSTM2A null adult mice. A) Quantitative PCR (qPCR) analysis of eWAT collected from wild‐type and knockout male mice fed a chow diet (Vstm2a+/+ [n = 11] and Vstm2a‐/‐ [n = 11]) for 9 weeks. Data represent the mean ± SEM. Significance was determined by two‐tailed, unpaired t test (*p ⟨0.05, **p ⟨0.01, ***p ⟨0.001, ****p ⟨0.0001).

Figure S5. VSTM2A loss does not impair adipogenesis in SVF. Stromal vascular fraction (SVF) was collected from eWAT of P21 mice and differentiation was induced. Five days later, mRNA was extracted and qPCR was performed for the indicated genes (Vstm2a+/+ [n = 6] and Vstm2a‐/‐ [n = 6]). Data represent the mean ± SEM. Significance was determined by two‐tailed, un‐paired t test (*p ⟨0.05, **p ⟨0.01, ***p ⟨0.001, ****p ⟨0.0001).

Figure S6. Repressing lysosomal function early during adipogenesis does not affect BAT morphology and systemic metabolic homeostasis. Pups were injected with BafA1 (0.1 mg/kg of body weight) during 4 consecutive days from P1 to P4. Mice were then allowed to grow and were sacrificed at the age of 16–18 weeks. (A) Histological analysis of BAT collected from adult female mice treated with vehicle (DMSO) (n = 13) or BafA1 (n = 7). Representative pictures are shown. On the right side of the panel, adipocyte size is presented for each group. The number of nuclei per area was counted to evaluate cell size. (B) Glucose tolerance test (GTT) and (C) Insulin tolerance test (ITT) performed in mice treated with vehicle (DMSO) (n = 13) or BafA1 (n = 7). (D) Blood metabolites and (E) liver triglycerides were measured from mice treated with vehicle (DMSO) (n = 13) or BafA1 (n = 7). Mice were fasted for 6 h prior to these analyses. (F) Quantitative PCR analysis of inflammatory genes from rWAT (left panel) and BAT (right panel) collected from mice treated with vehicle (DMSO) (n = 13) or BafA1 (n = 7). Data represent the mean ± SEM. Significance was determined by two‐tailed, unpaired t test (*p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001).

Figure S7. Angptl8 expression is increased in adipose tissue of young VSTM2A‐null mice. (A) Quantitative PCR analysis of Angptl8 expression in adipose tissues collected from wild‐type and knockout male mice at postnatal day 21 (P21). The number of samples per group is declined as followed. iWAT: Vstm2a ^+/+ (n = 11) and Vstm2a ^−/− (n = 9); rWAT: Vstm2a ^+/+ (n = 12) and Vstm2a ^−/− (n = 15); eWAT: Vstm2a ^+/+ (n = 10) and Vstm2a ^−/− (n = 8); BAT: Vstm2a ^+/+ (n = 13) and Vstm2a ^−/− (n = 15). (B) Quantitative PCR analysis of Angptl8 expression in adipose tissues collected from adult wild‐type and knockout male mice. The number of samples per group is declined as followed. iWAT: Vstm2a ^+/+ (n = 9) and Vstm2a ^−/− (n = 10); rWAT: Vstm2a ^+/+ (n = 11) and Vstm2a ^−/− (n = 9); eWAT: Vstm2a ^+/+ (n = 11) and Vstm2a ^−/− (n = 11); BAT: Vstm2a ^+/+ (n = 11) and Vstm2a ^−/− (n = 10). In all panels, data represent the mean ± SEM. Significance was determined by two‐tailed, unpaired t test (*p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001).

Table S1: Sequence of the qPCR primers used in this study.

Table S2: Overview of the genes differentially expressed in WAT of wild‐type and VSTM2A‐null mice at P21. The genes with a p value <0.01 are presented.

Al Dow M, Secco B, Mouchiroud M, et al. Loss of VSTM2A promotes adipocyte hypertrophy and disrupts metabolic homeostasis. Obesity (Silver Spring). 2025;33(3):522‐536. doi: 10.1002/oby.24224