CTRP6 rapidly responds to acute nutritional changes, regulating adipose tissue expansion and inflammation in mice

Abstract

In chronic obesity, activated adipose tissue proinflammatory cascades are tightly linked to metabolic dysfunction. Yet, close temporal analyses of the responses to obesogenic environment such as high-fat feeding (HFF) in susceptible mouse strains question the causal relationship between inflammation and metabolic dysfunction, and/or raises the possibility that certain inflammatory cascades play adaptive/homeostatic, rather than pathogenic roles. Here, we hypothesized that CTRP6, a C1QTNF family member, may constitute an early responder to acute nutritional changes in adipose tissue, with potential physiological roles. Both 3-days high-fat feeding (3dHFF) and acute obesity reversal [2-wk switch to low-fat diet after 8-wk HFF (8wHFF)] already induced marked changes in whole body fuel utilization. Although adipose tissue expression of classical proinflammatory cytokines (Tnf-α, Ccl2, and Il1b) exhibited no, or only minor, change, C1qtnf6 uniquely increased, and decreased, in response to 3dHFF and acute obesity reversal, respectively. CTRP6 knockout (KO) mouse embryonic fibroblasts (MEFs) exhibited increased adipogenic gene expression (Pparg, Fabp4, and Adipoq) and markedly reduced inflammatory genes (Tnf-α, Ccl2, and Il6) compared with wild-type MEFs, and recombinant CTRP6 induced the opposite gene expression signature, as assessed by RNA sequencing. Consistently, 3dHFF of CTRP6-KO mice induced a greater whole body and adipose tissue weight gain compared with wild-type littermates. Collectively, we propose CTRP6 as a gene that rapidly responds to acute changes in caloric intake, acting in acute overnutrition to induce a “physiological inflammatory response” that limits adipose tissue expansion.

NEW & NOTEWORTHY CTRP6 (C1qTNF6), a member of adiponectin gene family, regulates inflammation and metabolism in established obesity. Here, short-term high-fat feeding in mice is shown to increase adipose tissue expression of CTRP6 before changes in the expression of classical inflammatory genes occur. Conversely, CTRP6 expression in adipose tissue decreases early in the course of obesity reversal. Gain- and loss-of-function models suggest CTRP6 as a positive regulator of inflammatory cascades, and a negative regulator of adipogenesis and adipose tissue expansion.

INTRODUCTION

Most organisms are frequently confronted with both food scarcity and overabundance, requiring homeostatic mechanisms capable of adaptively dealing with wide fluctuations in caloric intake. Yet, modern sedentary lifestyles present for many a chronic condition of caloric overabundance resulting in obesity (1, 2). Not surprisingly, a major focus of metabolic research in recent years has centered on understanding the impact of chronic overnutrition on tissue insulin sensitivity and health of major organ compartments (e.g., heart, liver, pancreas, skeletal muscle, and the immune system). Although chronic consumption of a high-caloric diet undoubtedly results in excessive weight gain, obesity requires a significant amount of time (typically months or years) to manifest as adiposity beyond the healthy level; this observation, often understated, suggests the presence of homeostatic mechanisms that, at least initially, resist excessive weight gain, until they are eventually overwhelmed by the chronicity of the insult. Indeed, adipose tissue, the primary depot for storing excess calories, plays an especially important role in maintaining systemic energy balance in response to nutritional overload or reduced food availability (3). The fat-derived satiety hormone, leptin, is an established long-term regulator of adiposity, which signals to the hypothalamus to resist excessive weight gain (4). However, much less is known about other secreted regulators within the fat depot that can rapidly respond to acute changes in nutritional status to appropriately maintain fat mass.

Recently, pathways and mechanisms generally thought to be pathogenic in the context of chronic overnutrition and obesity have been proposed to serve beneficial roles in the early stages of overnutrition, aiming to maintain (or restore) homeostasis (5, 6). This is best exemplified by adipose tissue inflammatory responses to caloric surplus (7). Their complex and potentially opposite roles in response to short-term versus chronic energy imbalance also provide insights into why targeting inflammatory pathways in obesity has been clinically met with only limited success thus far (8–10).

Our recent efforts to better understand hormonal control of metabolism have led to the identification and characterization of a novel family of secreted proteins, the C1q/TNF-related proteins (CTRP1-15) (11, 12). Most of the CTRPs are expressed by adipose tissue at varying levels, and their expression is modulated by the nutritional state (13). Genetic mouse models have highlighted the salutary physiological roles of multiple CTRPs in regulating various aspects of sugar and fat metabolism, local and systemic insulin sensitivity, and adipose tissue inflammation (14–20). We hypothesize that one or more of the CTRP family members, expressed within the fat depot, could potentially serve as a rapid responder to acute nutritional challenges. Of the CTRPs, one in particular (CTRP6/C1qtnf6) is a likely candidate based on our prior observations in models of chronic high-fat feeding (21): 1) its expression in both visceral and subcutaneous fat depots is significantly elevated in diet-induced obese mice; 2) recombinant CTRP6 treatment impairs insulin action in cultured adipocytes; and 3) knockout mice lacking CTRP6 have improved insulin sensitivity and reduced adipose tissue inflammation, and conversely, mice overexpressing CTRP6 showed the opposite effects. Whether CTRP6, or other related family members, also play an adaptive role in short-term/acute changes in nutritional state is unknown and is the focus of the present study.

To specifically test our hypothesis, we metabolically characterized two short-term nutritional challenge models: 1) an acute 3-day high-fat feeding to induce a sudden energy surplus and adiposity, and 2) an obesity early reversal model to induce rapid energy deficit and weight loss. Adipose tissue expression of selected CTRPs and “classical” adipose tissue inflammatory genes were assessed, uncovering putative unique responsiveness of CTRP6 to alterations in caloric/fat intake. Further uncovering pathways regulated by CTRP6 in mouse embryonic fibroblasts (MEFs) and challenging CTRP6-KO mice with 3dHFF complement our previous findings (21), and suggest CTRP6 as a novel secreted regulator that acts within a short timescale to modulate adipose tissue inflammation and limit fat mass expansion.

MATERIALS AND METHODS

Materials

Tissue culture media, fetal bovine serum (FBS), l-glutamine, penicillin-streptomycin-amphotericin B, and PBS were purchased from Biological Industries (Biological Industries, Israel). Recombinant mouse CTRP6 (with a C-terminal FLAG epitope tag) was produced and purified from the conditioned medium of transfected mammalian HEK293T cells as previously described (21).

Animals and Experimental Protocols

The study was approved by Ben-Gurion University Institutional Animal Care and Use Committee and was conducted according to the Israeli Animal Welfare Act following the guidelines of the Guide for Care and Use of Laboratory Animals (National Research Council, 1996). Five-weeks-old male C57BL/6 mice were purchased from ENVIGO RMS Ltd (Jerusalem, Israel). Animals were acclimatized to the animal facility with free access to control, sucrose-matched, low-fat diet (LFD, D12450J, Research Diets, NJ), and water for 2 wk, two mice per cage, with 12-h light-dark photocycle. At age of 7 wk, two acute dietary intervention models were initiated by high-fat feeding (HFF) consisting of 60% calories derived from fat, 20% from carbohydrate, and 20% from protein (D12492, Research Diets, NJ): Acute overnutrition model was achieved by 3 days of high-fat feeding (3dHFF), whereas obesity reversal model was performed by 8 wk of HFF and a 2-wk switch to LFD, in comparison with 10 wk either LFD (lean) or HFF (obese) mice. At the end of the experiments, mice were fasted for 12 h and terminated using isoflurane. Gonadal (visceral) and mesenteric fat depot and liver were collected and weighed. CTRP6 knockout (KO; ^−/−) mice were generated as previously described (21). Wild-type (WT; ^+/+) littermates and CTRP6-KO male mice were obtained by crossing CTRP6 heterozygous (^+/−) mice. All mice were housed in polycarbonate cages on a 12-h light-dark photocycle with ad libitum access to water and food. At termination of the study, mice were fasted for 2 h and euthanized. Visceral (gonadal) and subcutaneous (inguinal) fat depots were collected and weighed. All mouse protocols were approved by the Institutional Animal Care and Use Committee of the Johns Hopkins University School of Medicine.

Multi-Parameter Metabolic Assessment (Metabolic Cages)

Mice were housed in one/cage in a Promethion High-Definition Behavioral Phenotyping System (Sable Instruments, Inc., Las Vegas, NV) for 3 days. Data collection started after 24 h, exactly as recently described in a study by Bhandarkar et al. (22). Respiratory exchange rate (RER), total carbohydrate, and fat oxidation (g/h) were calculated as follows: RER = V̇co₂/V̇o₂; Carbohydrate oxidation = 4.585 × V̇co₂ – 3.23 × V̇o₂; Fat oxidation = 1.69 × V̇o₂ – 1.69 × V̇co₂ (23). Energy intake (Kcal/h) was calculated as follows: Food intake × calories per gram. Energy expenditure (EE) was calculated as V̇o₂ × (3.815 + 1.232 × RER). Parameters of mice in the obesity reversal model, except for the RER, were normalized to effective body weight, as previously described (24).

Cell Culture

Mouse embryonic fibroblasts were derived from CTRP6-KO and WT littermates. Embryos were harvested from pregnant female mice (at 13.5 days postcoitum). The head, heart, and liver were removed and discarded. The remaining embryo part was minced into 1–2 mm pieces in the presence of trypsin and incubated for 10 min in a 37°C tissue culture incubator. The embryo pieces were pipetted up and down multiple times and incubated for another 10 min at 37°C. The cell suspension was transferred to a 50 mL tube and a 20 mL MEF culture medium (DMEM containing 10% fetal bovine serum and penicillin-streptomycin) was added to inactivate the trypsin. The cell suspension was allowed to sit for 5 min at room temperature to allow larger embryo fragments to sink to the bottom of the tube. The supernatant containing single cells and cell clusters was transferred to a 150 cm plate. Medium was changed once the following day and MEF cells were allowed to grow until confluent. Low-passage MEFs were frozen down and store in liquid nitrogen. MEFs were cultured in DMEM media containing 4.5 mM glucose, 20% FBS, 2 mM l-glutamine, and 100 U/mL penicillin-streptomycin-amphotericin B. For quantitative real-time PCR experiments: 1) MEFs from WT and CTRP6^−/− were seeded in 6-well plates. Seventy-two hours after seeding (at 90% confluence), cells were washed twice with ice-cold PBS and subjected to quantitative real-time PCR analysis. 2) MEFs from CTRP6-KO mice were seeded in 6-well plates. Forty-eight hours after seeding, media was changed to DMEM containing 0.5% BSA (with 2 mM l-glutamine, and 100 U/mL penicillin-streptomycin-amphotericin B) and supplemented with 10 µg/mL recombinant CTRP6 (rCTRP6). After 24 h incubation, cells were washed twice with ice-cold PBS and subjected to both quantitative Real-time PCR and RNA-Seq. MEF-derived adipocyte-like cells (MEF-adip) were generated exactly as described previously (25).

Lipolysis Assay

WT and CTRP6^−/− MEF-adip were incubated without (basal) or with (stimulated) TNFα (10 ng/mL) + IL-1β (10 ng/mL) + INFγ (100 ng/mL) in DMEM containing 0.5% BSA for 24 h. Next, cells were washed twice with PBS and incubated with Krebs–Ringer buffer (KRBH; 135 mM NaCl, 5 mM KCl, 1 mM MgSO₄, 0.4 mM K₂HPO₄, 1 mM CaCl₂, 20 mM HEPES, pH 7.4) for 1 h. Free glycerol was measured using Sigma Aldrich’s free glycerol reagent (F6428), and results were normalized to milligram protein cell lysate.

Leptin Secretion

Leptin secretion was measured in DMEM + 0.5% BSA cultured media (24 h) by ELISA (R&D Systems Inc., MOB00) following the manufacturer’s protocol.

RNA Extraction and Quantitative Real-Time PCR

Total RNA from MEFs was extracted using the RNeasy Lipid Tissue Mini Kit (QIAGEN, Germantown, MD) and quantified with nanodrop. Then, 2 μg of RNA were reverse transcribed with High-Capacity cDNA Reverse Transcriptase Kit (Applied Biosystems, Foster City, CA). TaqMan system (Applied Biosystems, Foster City, CA) was used for real-time PCR amplification. Relative gene expression was obtained after normalization to murine endogenous controls, Hprt and Rplpo for epididymal adipose tissue; Hprt and Rn18s for MEFs (Table 1) using the formula 2^−ΔΔCt.

Table 1.

TaqMan primers used for PCR assays

Reference Genes
Hprt	Mm00446968_m1
Rplpo	Mm00725448_s1
Rn18s	Hs0392899_g1

Detected Genes
Tnf	Mm00443258_m1
Il1b	Mm00434228_m1
Il6	Mm00446190_m1
Il1rn	Mm00446186_m1
Ccl2	Mm00441242_m1
C1qtnf1	Mm00480204_m1
C1qtnf2	Mm07298677_m1
C1qtnf3	Mm00473047_m1
C1qtnf6	Mm00511605_m1
C1qtnf9	Mm02392597_m1
Pparg	Mm00440940_m1
Cebpa	Mm00514283_s1
Fabp4	Mm00445878_m1
Slc2a4	Mm00436615_m1
Adipoq	Mm00456425_m1

RNA-Seq Data Analysis

RNA-Seq of 12 samples of CTRP6-KO MEFs was performed using NextSeq 500 Illumina sequencing technology. The experiment was done in two batches (six samples in each) and within each experiment three biological replicates of the treatment group (rCTRP6) and three biological replicates of the control (con). Initial analysis of the raw sequence reads was carried out using the NeatSeq-Flow platform. The sequences were quality trimmed and filtered using Trim Galore (v0.4.5) and cutadapt (v1.15). Alignment of the reads to the mouse genome (GRCm38) was done using STAR (v2.5.2a). The number of reads per gene per sample was counted using RSEM (v1.2.28). Quality assessment (QA) of the process was carried out using FASTQC (v0.11.8) and MultiQC (v1.0.dev0). After trimming, each sample had an average of 79 M reads with an average sequence length of 74.4 bp. Statistical testing for identification of differentially expressed genes, batch correction, gene annotation, clustering, and enrichment analysis was performed with the DESeq2 module within the NeatSeq-Flow platform. Batch correction was done using the SVA/Combat R package. Gene annotation was done using the “AnnotationHub” R package (snapshot: 2020-04-27). After examining the PCA of the normalized count it was clear that one of the samples was very different relative to all other samples and therefore was removed from downstream analysis. The statistical analysis was done using the DESeq2 [5] R package. For comparison (contrast) between treatment group (rCTRP6) and control (con), the statistical model considered two effects: the treatment group and the batch (experiment). The analysis produced P value, FDR-adjusted P value, and fold of change (FC) per gene. Genes with FDR adjusted P value < 0.05 were considered differentially expressed (DE). KEGG enrichment analysis was done using the clusterProfiler (v3.16.0) R package.

Statistical Analysis

Analyses were performed using GraphPad Prism 9.1.0. Statistical significance was examined using nonparametric unpaired Mann–Whitney test or nonparametric ANOVA test (Kruskal–Wallis) when comparing two or three groups, respectively.

RESULTS

To determine if CTRP6 responds rapidly to nutrition-related metabolic changes, we used two nutritional paradigms using high-fat feeding (HFF): 1) 3dHFF, to model acute caloric/fat surplus; and 2) an obesity reversal model (REV), in which 8-wk HFF mice were switched back to low-fat diet for 2 wk, to investigate the early adaptation to decreased dietary calorie/fat intake in obesity. Basic biochemical characteristics of mice in both experimental models are presented in Table 2 and Table 3, respectively. Notably, 3dHFF was sufficient to induce a significant increase in fasting glucose levels, total and HDL cholesterol. Insulin resistance was induced, as assessed by HOMA-IR, with P = 0.054 between the groups. In the early obesity reversal model, 10wHFF mice exhibited significantly elevated glycemic and lipid parameters compared with mice on low-fat diet (LFD). Switching mice in the last 2 wk of dietary intervention (REV) significantly, though partially, reversed glycemic parameters, whereas lipid parameters were near fully reversed to the level of mice continuously fed LFD (Table 3). 3dHFF was sufficient to induce a significant increase in total body weight, liver weight, and epididymal, but not mesenteric fat weight, compared with mice on LFD (Fig. 1, A and B). We used metabolic cages to determine if these interventions induced significant shifts in whole body metabolism. Respiratory exchange rate (RER) was lower throughout the day in the 3dHFF mice and exhibited a lower increase between the inactive (day) and active (night) phases, compared with LFD (Fig. 1C). Correspondingly, calculated carbohydrate and fat oxidation rates (Fig. 1, D, and E, respectively) revealed lower reliance on carbohydrates and higher on fat as a selected fuel source for oxidation in 3dHFF mice, with blunting of the night-to-day switch in fuel selection.

Table 2.

Biochemical parameters of the short-term high-fat diet model

	LFDn = 7–10	3dHFDn = 7–10	P Value
Body weight, g	20.2 ± 0.2	22.2 ± 0.3	<0.001
Fasting glucose, mg/dL	72.2 ± 3.3	93.1 ± 3.5	0.002
Fasting insulin, µIU/mL	6.5 ± 0.7	8.8 ± 1.4	0.336
HOMA-IR	1.1 ± 0.1	2.0 ± 0.3	0.054
Triglycerides, mg/dL	103.3 ± 14.2	105.9 ± 16.8	0.999
Total cholesterol, mg/dL	70.4 ± 9.2	103.9 ± 9.4	0.042
HDL-c, mg/dL	53.5 ± 6.7	83.2 ± 7.2	0.024
Non-HDL-c, mg/dL	12.1 ± 1.1	15.5 ± 1.7	0.115
AST, IU/L	114.8 ± 21.4	87.9 ± 13.5	0.246
ALT, IU/L	50.8 ± 9.6	47.2 ± 14.0	0.525

ALT, alanine aminotransferase; AST, aspartate aminotransferase; HFD, high-fat diet; LFD, low-fat diet.

Table 3.

Biochemical parameters of the acute obesity reversal model

	LFD n = 13–21	10wHFD n = 13–22	REV n = 14–26	P Value of ANOVA
Body weight, g	29.1 ± 0.8	38.4 ± 0.7****	32.2 ± 0.8##	<0.001
Fasting glucose, mg/dL	71.1 ± 3.7	110.2 ± 8.8***	96.1 ± 3.7**	<0.001
Fasting insulin, uIU/mL	14.4 ± 2.7	82.9 ± 12.1****	29.7 ± 4.5#	<0.001
HOMA-IR	2.8 ± 0.5	22.9 ± 3.6****	7.3 ± 1.2#	<0.001
Triglycerides, mg/dL	91.6 ± 4.6	68.4 ± 3.8**	92.1 ± 6.4##	0.003
Total cholesterol, mg/dL	75.3 ± 2.1	106.1 ± 6.9**	79.1 ± 2.4#	0.003
HDL-c, mg/dL	55.8 ± 1.7	76.5 ± 5.8*	55.5 ± 3.9	0.016
Non-HDL-c, mg/dL	19.8 ± 1.1	29.5 ± 3.5**	23.9 ± 3.2	0.012
AST, IU/L	117.4 ± 20.7	141.6 ± 27.8	126.8 ± 29.1	0.447

AST, aspartate aminotransferase; LFD, low-fat diet; REV, obesity reversal model; 10wHFD, 10-wk high-fat diet. *, **, ***, ****P < 0.05, 0.01, 0.001, 0.0001 compared to LFD, respectively; #, ##P < 0.05, 0.01 compared to 10wHFD, respectively.

Figure 1.

Whole body weight and metabolic effects of 3d high-fat feeding (3dHFF) and early obesity reversal (REV) in mice. A–E: three-days high-fat feeding (3dHFF) was used to model short-term excess calorie/fat, before establishment of overt obesity. A and B were from n = 10 mice per group in two independent experiments. C–E were performed on different mice, n = 8 per group, in two independent experiments. F and G: the effect of early obesity reversal (REV) was modeled by feeding mice for 8 wk on high-fat diet, followed by switching back to low-fat diet (LFD). REV mice were compared with lean mice constantly fed LFD and to nutritionally obese mice fed HFF for 10 wk (10wHFF). Data were from n = 13 or 14 mice per group in two independent experiments. For metabolic cages (H–J), data are derived from n = 7–9 per group, in three independent experiments. A and F: whole body weight; B and G: organ weights. C and H: the respiratory exchange ratio (RER) was measured by indirect calorimetry using metabolic cages [Promethion High-Definition Behavioral Phenotyping System (Sable Instruments, Inc.)], as detailed in methods. D and I: calculated carbohydrate oxidation rates, and E and J: calculated fatty acid oxidation rates based on O₂ consumption and CO₂ production rates. Given body weight and composition differences between the intervention groups in the obesity reversal model, carbohydrate and fat oxidation (I, J) were controlled for effective body mass (see methods for detail). *, **, ***, ****P < 0.05, 0.01, 0.001, 0.0001, respectively, compared with LFD. #, ##P < 0.05, 0.01, respectively, compared with 10wHFF.

Early adaptation to reversal of nutritional obesity was also apparent, with REV mice losing ∼60% of their excess total body weight (Fig. 1F). Epididymal and mesenteric fat weights also significantly differed from 10wHFF mice (Fig. 1G). REV mice fully regained the daily pattern of RER characteristic of the LFD control group (Fig. 1H). Interestingly, carbohydrate oxidation tended to be lower than LFD mice, and fat oxidation was “hypernormalized,” i.e., was lower than in LFD mice (Fig. 1, I and J, respectively), consistent with diminished energy expenditure during the phase of negative energy balance (not shown). Jointly, 3dHFF and REV mice represent models of early adaptation to an increase or decrease, respectively, in dietary calorie/fat content, with significant metabolic impact, but before obesity had been fully established or reversed.

We next assessed changes in adipose tissue expression of “classical” proinflammatory cytokines known to be upregulated in established murine and human obesity, as well as C1qtnf (CTRP) family members proposed to participate in immunometabolic regulation. 3dHFF did not significantly alter the expression of Tnf-α, Il-1β, Il-1rn, Il-6, or Ccl2 in epididymal fat (Fig. 2, A–E) or in mesenteric fat (not shown). Among the CTRPs tested (Fig. 2, F–J), only CTRP6 exhibited a marginally significant (P = 0.05) 1.5-fold increase in epididymal fat (Fig. 2I). This putative uniqueness of CTRP6 to “respond” to early changes in caloric/fat intake was also apparent in the obesity reversal model: The expression of classical inflammatory cytokines was elevated in 10wHFF mice compared with those fed LFD and exhibited variable reversal in epididymal fat pads of REV mice (Fig. 3, A–D). Among the CTRPs (Fig. 3, E–I), C1qTNF6 in epididymal fat of 10wHFF mice was >2.2-fold higher than in LFD mice, and this difference was near fully (76%) reversed in REV mice (Fig. 3H), being statistically not different from LFD controls. Thus, CTRP6 stands out as a potential early responder to changes—both increase and decrease—in dietary calorie/fat content, preceding or exceeding dynamic changes in classical inflammatory genes activated in adipose tissue in obesity.

Figure 2.

Changes in adipose tissue expression of classical inflammatory cytokines and selected C1qTNF (CTRP) family members by 3-days high-fat diet. Epididymal adipose tissue of mice fed either a low-fat diet (LFD) or high-fat diet for 3 days (3dHFF) was subjected to mRNA extraction, and expression of selected classical proinflammatory cytokines (A–E) and CTRP family members (F–J) were measured by real-time PCR. Outliers (maximum of a single data point/group) were defined by the “identify outlier” function in GraphPad and excluded, in A, D, and G. Data are from two independent experiments.

Figure 3.

Changes in adipose tissue expression of classical inflammatory cytokines and selected C1qTNF (CTRP) family members by early obesity reversal. Epididymal adipose tissue of mice fed either low-fat diet (LFD) or high-fat diet for 10 wk (10wHFF), or fed 8-wk high-fat diet and then switched to LFD for 2 wk (REV) was subjected to mRNA extraction, and expression of selected classical proinflammatory cytokines (A–D) and CTRP family members (E–I) were measured by real-time PCR. Outliers were defined by the “identify outlier” function in GraphPad and excluded, in A, C, D, and F. *, **, ***, ****P < 0.05, 0.01, 0.001, 0.0001, respectively, compared with LFD. #P < 0.05 compared with 10wHFF. Data are from two independent experiments.

To gain insights on pathways that may be directly activated by CTRP6, recombinant CTRP6 was generated as previously described (21) and used to stimulate cultured mouse embryonic fibroblasts (MEFs) generated from CTRP6-KO mice (to ensure the absence of CTRP6 autostimulation in cells not stimulated with exogenous CTRP6). Twenty-four-hour stimulation with 10 µg/mL recombinant CTRP6 (rCTRP6) induced a significant (FDR-adjusted P value < 0.05) change in 304 upregulated and 658 downregulated genes, as measured via RNA sequencing (Fig. 4A). Among the 10 most highly overexpressed genes in response to CTRP6 stimulation, the first is Steap4 (Fig. 4B), a metalloreductase involved in adipocyte development and metabolism, and downregulated in adipose tissue in established obesity (26, 27). Indeed, adipose tissue in 3dHFF mice expressed >twofold higher levels of Steap4 compared with LFD, whereas in established obesity (10wHFF) Steap4 was downregulated (Fig. 4C). Five of the 10 most highly overexpressed genes belong to the chemokine family (Fig. 4B), and indeed by real-time PCR, Ccl2 was 2.5-fold elevated in CTRP6-stimulated MEFs (Fig. 4D), validating the 1.6-fold increase observed in the RNA-Seq data (Fig. 4B). Intriguingly, Pparg was significantly downregulated in MEFs stimulated by rCTRP6 (Fig. 4D), whereas another adipogenic transcription factor – Cebpa, exhibited a mild, nonsignificant reduction. KEGG pathway enrichment analysis of the RNA-Seq data identified several CTRP6-regulated pathways in MEFs, including TNF-α, IL-17, and chemokine/cytokine-related pathways, which include genes that were mostly upregulated (Figs. 4E and 5). Conversely, pathways related to adipocyte differentiation and function (e.g., adipocyte lipolysis) seemed to include genes that were mainly downregulated at the gene expression level by CTRP6 (Figs. 4E and 6), consistent with the decrease in Pparg (Fig. 4D). It is noteworthy that the effects of CTRP6 were observed with embryonic fibroblasts that have not been stimulated to exhibit a particular cell-type characteristic, including those of adipocytes.

Figure 4.

Changes in gene expression induced by recombinant CTRP6 in mouse embryonic fibroblasts of CTRP6-knockout mice. Mouse embryonic fibroblasts (MEFs) were obtained from CTRP6-KO mice and stimulated with 10 µg/mL recombinant CTRP6 (rCTRP6) for 24 h. Cells from two independent batches were subjected to RNA sequencing, as detailed in methods. A: differentially expressed genes between rCTRP6-stimulated and control cells exhibit 304 upregulated and 658 downregulated genes (FDR-adjusted P < 0.05). B: list of 10 most highly upregulated (left) or downregulated (right) genes in rCTRP6-stimulated CTRP6-KO-MEFs. Fold change is presented in linear scale, and the minus sign indicates downregulation. C: relative gene expression of (by real-time qPCR) Steap4 in epididymal adipose tissue in two nutritional models: 3dHFF (left) and early obesity reversal (right). Number of mice in each group is depicted in the figure. Data are from two and one independent experiment(s), respectively. D: in an independent cell batch of MEFs, the expression of selected inflammatory genes (left) and adipogenic genes (right) was measured by real-time qPCR. Results represent two independent experiments performed with biological duplicates, *P < 0.05 compared with control (nonstimulated) cells. E: KEGG pathway enrichment analysis among differentially expressed genes between rCTRP6-stimulated and control MEFs. CTRP, C1q/TNF-related protein; 3dHFF, 3-days high-fat diet.

Figure 5.

Expression level changes in genes of KEGG “Cytokine-cytokine receptor interaction” pathway induced by recombinant CTRP6 in CTRP6-KO MEFs. KEGG pathway enrichment analysis was performed on RNA-Seq data derived from CTRP6-KO MEFs treated with recombinant CTRP6 versus nontreated (shown in Fig. 4). Shown are genes of the “Cytokine-cytokine receptor interaction” KEGG pathway. Marked in green are genes whose expression levels did not change, red was increased and blue was decreased, in response to rCTRP6 stimulation. Results demonstrate CTRP6-mediated increased expression in several genes of the CC and CXC chemokine families, and receptors of the IL6/IL12-like, IL-1-like, and TNF families. CTRP, C1q/TNF-related protein; KO, knockout; MEF, mouse embryonic fibroblast.

Figure 6.

Expression level changes in genes of KEGG “Regulation of lipolysis in adipocytes” pathway induced by recombinant CTRP6 in CTRP6-KO MEFs. KEGG pathway enrichment analysis was performed on RNA-Seq data derived from CTRP6-KO MEFs treated with recombinant CTRP6 versus nontreated (shown in Fig. 4). Shown are genes of the “Regulation of lipolysis in adipocytes” KEGG pathway. Marked in green are genes whose expression levels did not change, red was increased and blue was decreased, in response to rCTRP6 stimulation. CTRP, C1q/TNF-related protein; KO, knockout; MEF, mouse embryonic fibroblast.

To substantiate the possible CTRP6-mediated regulation of inflammatory pathways (upregulation) and adipocyte-related pathways (downregulation), we assessed key genes related to these pathways in Ctrp6-WT versus Ctrp6-KO MEFs (the latter confirmed to express no C1qtnf6 mRNA, Fig. 7A). Ctrp6-KO-MEF expressed significantly higher levels of the adipogenic transcription factors Pparg (∼30-fold) and Cebpa, and the adipocytic genes Fabp4, and adiponectin (Adipoq, >20-fold) when compared with WT-MEFs (Fig. 7B). Conversely, the expression of the proinflammatory genes Ccl2, Tnf-α, and Il-6 was lower in Ctrp6-KO-MEFs compared with WT (Fig. 7C). To gain insights on CTRP6 function in cells more closely reflecting mature adipocytes, we subjected MEFs to an adipogenic differentiation protocol (25), generating MEF-derived adipocyte-like cells (MEF-Adip). We did not observe decreased capacity of Ctrp6-KO-MEF-adip to exhibit morphological feature of adipocytes (round cells with lipid droplets, Fig. 7D). Both WT and Ctrp6-KO-MEF-adip, equally greatly upregulated the expression of Pparg or Cepba compared with wild-type MEF upon adipogenic stimulation (Fig. 7E). Yet, Ctrp6-KO-MEF-adip exhibited a mild, but statistically significant, lower expression of Fabp4 and Adipoq compared with WT-MEF-adip, opposite of the effect of CTRP6 absence in nondifferentiated MEFs. The effect of CTRP6 deficiency on key inflammatory gene expression in MEF-adip was also variable compared with MEFs, with no significant effect (compared with wild-type MEFs) on Ccl2, a marked decrease in Tnf-α and increase in Il-6 gene expression (Fig. 7F). Together, our results from MEFs strongly suggest that CTRP6 is a regulator of adipogenesis and inflammation, inhibiting the former and promoting the latter, whereas the effect is more variable in cells with a greater commitment to adipocyte lineage.

Figure 7.

The in vitro effect of CTRP6-KO. A: C1qtnf6 (CTRP6) expression in mouse embryonic fibroblasts (MEFs) of wild-type (WT) and CTRP6-KO (CTRP6^−/−) mice. B and C: relative gene expression in MEFs of CTRP6-KO compared with WT of adipogenic/adipocyte-specific genes (B), or inflammatory genes (C). D: WT or CTRP6^−/− MEFs were differentiated into MEF-derived adipocyte-like cells (MEF-adip) as detailed in methods. Representative light microscopy images of MEF-adip of WT and CTRP6^−/− are shown. Scale bar = 400 µm. E and F: relative gene expression in MEF-adip (compared with WT-MEFs) of CTRP6-KO compared with WT, adipogenic/adipocyte-specific genes (E), or inflammatory genes (F). Experiments were performed in three to four independent batches of cells, each in triplicates. *,***P < 0.05, 0.001, respectively, compared with WT. CTRP, C1q/TNF-related protein; KO, knockout.

We next investigated the possible functional impact of the lack of CTRP6 gene by examining how CTRP6 deletion would affect the response to 3dHFF. Compared with WT mice, those lacking CTRP6 gained more total body weight and adipose tissues’ weight (Fig. 8, A–C). These data suggest that upregulated Ctrp6 upon 3dHFF (Fig. 2I) may act to limit adipose tissue expansion, and weight gain, an effect lacking in CTRP6-KO mice. To determine whether CTRP6 functional impact on adipocytes-like cells may contribute to the observed in vivo effect, we measured leptin release and lipolysis in MEF-adip cells. Ctrp6-KO-MEF-adip released less leptin into the medium (Fig. 8D), suggesting that at least in response to short-term HFF, adipocytes of Ctrp6-KO mice may signal a lower satiety response. Lipolysis was not affected basally by the lack of CTRP6; however, stimulated lipolysis (induced by a mix of inflammatory cytokines) was significantly lower in Ctrp6-KO-MEF-adip compared with WT-MEF-adip (Fig. 8E). These two effects of CTRP6 absence in adipocyte-like cells may provide putative mechanisms for increased weight gain in response to 3dHFF in Ctrp6-KO compared with WT mice, possibly inducing overeating and increased fat storage.

Figure 8.

The in vivo/functional effect of CTRP6-KO. A–C: whole body, gonadal and inguinal white adipose tissue (gWAT, iWAT, respectively) weights in WT and CTRP6-KO littermate mice after 3-days high-fat feeding (3dHFF). Number of mice in each group is depicted in the figures. D: leptin secretion to the media from MEF-adip. E: MEF-adip from WT or CTRP6^−/− mice were either left unstimulated, or treated for 24 h with a cytokine mix (IL-1β 10 ng/mL + TNF-α 10 ng/mL + INFγ 100 ng/mL), washed, and glycerol released to the medium was collected for 1 h. Results are from three to four independent batches of cells, each in triplicates. *,**,***P < 0.05, 0.01, 0.001, respectively, compared with WT. CTRP, C1q/TNF-related protein; KO, knockout; WT, wild-type

DISCUSSION

In the present study, we hypothesized that adipose tissue CTRP6 is a rapid responder to acute metabolic alterations, and attempted to uncover whether it may also exert homeostatic and adaptive roles. Using two opposite nutritional challenge paradigms, we showed that the expression of CTRP6 within adipose tissue is dynamic and rapidly regulated. The rapid changes in CTRP6 expression occurred while several classical adipose tissue inflammatory genes have not responded (yet) to the acute change in caloric/fat intake. We further showed that these changes are likely to be associated with, and linked to, an altered inflammatory response within fat depots: using recombinant protein, we provided evidence for the direct action of CTRP6, as an isolated factor, in eliciting a proinflammatory gene signature—most notably upregulating the expression of chemokine/chemokine receptor genes. Although in nonadipocytes, genes related to adipose-specific function (lipolysis) were mostly downregulated. Consistently, we observed a proadipogenic gene program and a marked dampening of inflammatory gene expression in MEFs that lack CTRP6. In MEF-adip, the lack of CTRP6 was associated with lower leptin secretion and stimulated lipolysis, consistent with findings in CTRP6-deficient mice that exhibited significantly greater weight gain and adiposity in response to 3-day high-fat feeding. These observations suggest a role for CTRP6 in limiting fat mass expansion in response to acute caloric excess, possibly by activating a “homeostatic/physiological” inflammatory program, which may manifest uniquely in different cell types that comprise adipose tissue. Together, these data suggest CTRP6 to be a secreted factor that can rapidly respond to energy surplus or deficit to modulate inflammation and adiposity within the fat compartment.

Several limitations of the study are noted here. In vivo, the dynamic changes we observed for CTRP6 in adipose tissue are at the level of the transcript. We presumed that the corresponding protein level of CTRP6 is likewise altered by 3-day HFF and obesity reversal. The lack of a validated ELISA assay hampers our ability to accurately quantify the level of CTRP6 protein in tissues and in circulation. In addition, we did not determine the cellular source of increased adipose tissue CTRP6 expression in response to 3dHFF. In vitro, it remains to be determined whether the dose of recombinant CTRP6 used (10 μg/mL) approaches that seen in the physiological setting. Again, technical limitations preclude us from knowing what the local concentration of CTRP6 within the adipose compartment is versus its concentration in the plasma. The local concentration of a secreted protein is generally believed to be significantly higher in the vicinity milieu where it is made when compared with plasma. Like all C1q family members, the basic structural unit of CTRP6 is that of a trimer, and the trimer can be further assembled into hexamer and high molecular weight oligomer via intermolecular disulfide bonds involving the conserved N-terminal Cys residues (12). Although it remains to be confirmed, the effective concentration of CTRP6 used (after accounting for its higher-order oligomeric structures) in our in vitro system is likely within the physiological range encountered in vivo.

Established obesity has been repeatedly shown to be associated with adipose tissue inflammation, characterized by increased abundance of proinflammatory immune cell types, proinflammatory “polarization” of macrophages (and possible other leucocytes), and consequently—increased pro/anti-inflammatory cytokine balance (7, 28). Beyond associations, a significant amount of evidence, particularly from mouse models and in vitro studies, causally tied various facets of this complex proinflammatory response to the development of adipose tissue dysfunction and whole body insulin resistance. However, some reports have also questioned the notion that any inflammatory response is necessarily a detrimental (pathogenic) response, consistent with Metchnikoff view of inflammation as a process activated when other homeostatic mechanisms have insufficiently dealt with a perturbation, acting as a backup mechanism to restore homeostasis (recently reviewed in Ref. 29). The temporal coordinates for the development of specific inflammatory responses and metabolic dysfunction calls for a broader role of inflammation in adipose tissue’s response to overnutrition (7). Even the well-established connection between adipose tissue macrophage accumulation and metabolic dysfunction can be questioned when a careful time course is assessed: insulin resistance develops within days of initiating high-fat feeding, but increased abundance of macrophages in the adipose tissue is documented when the high-fat stimulus persists for more than 3 wk (30). Conversely, when mice are switched after 8 wk of high-fat feeding to low-fat diet, insulin sensitivity is restored before macrophage abundance in adipose tissue declined (31). Indeed, temporality between a suspected causal factor and its presumed consequence is one of Hill’s criteria for causality (32), which calls for a careful assessment of the sequence of events—inflammatory and metabolic disturbances, in the acute responses to nutritional interventions before obesity has been established or fully resolved. In other words, examining the early response to obesogenic diet can untangle associations that are obscured by the complexity of mechanisms participating in chronic obesity.

In search of putative early inflammatory and metabolic modulatory factors in adipose tissue that rapidly responds to acute nutritional changes, we hereby suggest CTRP6 as a candidate. In established obesity (chronic high-fat feeding), mice lacking CTRP6 have reduced adipose inflammation and improved systemic insulin sensitivity; conversely, mice with elevated expression of CTRP6 have impaired insulin action (21), which can be attributed to the autocrine and paracrine actions of CTRP6 on macrophages and adipocytes. Here, an acute 3-day HFF resulted in greater weight gain and adiposity in CtrpP6-KO mice relative to WT littermates. CTRP6 expression in adipose tissue increased in an acute manner, unlike other CTRPs and classical proinflammatory cytokines. Consistent with the genes and pathways regulated by CTRP6, it appears that this protein participates in a physiological inflammatory response that acts to limit adipose tissue expansion. Interestingly, chronic HFF over a 12-wk period did not lead to a significant difference in body weight between CtrpP6-KO and WT controls, despite significant reduction in tissue and systemic inflammatory profiles (21). Although the reason for this is unclear, we speculate that when obesity is induced by chronic, long-term consumption of high-fat diet, many positive and negative pathways that normally operate within adipose tissue are eventually disrupted. This can be exemplified by CTRP6-mediated upregulation of Steap4 in MEFs and in adipose tissue of 3dHFF, whereas in established obesity Steap4 is downregulated (27). The net result is that proinflammatory effects of CTRP6 persist, whereas its antiobesity (adipose tissue expansion) effect is lost or overwhelmed. Future studies are clearly needed to address this possibility.

In obesity reversal, many inflammatory cascades in adipose tissue are slow to reverse, whereas whole body metabolic dysfunction quickly recovers (31). This suggests the involvement of “nonclassical” proinflammatory mediators that rapidly respond to cessation of the obesogenic challenge. Here, again, CTRP6 stands out as a possible candidate, exhibiting a significant decrease toward the level of LFD-fed mice at a time that classical proinflammatory cytokines like TNF-α still remain highly expressed in adipose tissue. Indeed, we have previously demonstrated using this model that although macrophage abundance remained high in adipose tissue of mice 2 wk after initiating obesity reversal, adipose tissue macrophages’ lipid content decreased by 50% (31). This demonstrates that while still present in the tissue, adipose tissue macrophages do exhibit metabolic adaptation early in the course of obesity reversal, a change that may also manifest by lower adipose tissue expression of CTRP6, as we have shown here. Indeed, infiltrated adipose tissue macrophages are a major source of CTRP6 in chronic HFF (21).

In summary, our data highlight the utility of both acute nutritional challenge and obesity reversal models in uncovering novel and relevant mediators governing the dynamic response of adipose compartment to acute metabolic perturbations. Our results suggest that CTRP6 is one such mediator with putative homeostatic function within the fat pad, limiting or moderating its expansion in response to acute energy surplus. In a broader context, our findings support the growing notion that obesity-associated morbidity is related to its temporality (i.e., chronicity), and that “physiological inflammation” is likely a normal acute adaptive response to short-term changes in energy balance, which become pathogenic and maladaptive when the metabolic imbalance becomes chronic.

GRANTS

This study was supported by the US-Israel Binational Science Foundation Grant 2017027 (to A.R. and G.W.W), as well as by the National Institutes of Health Grant DK084171 (to G.W.W.) and the Israel Science Foundation Grant ISF-2176/19 (to A.R.).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

R.L., Y.H., G.W.W., and A.R. conceived and designed research; R.L., Y.H., N.B., L.L., V.C.-C., D.S., A.S., N.M., B.K., and A.R. performed experiments; R.L., Y.H., N.B., L.L., V.C.-C., D.S., A.S., B.K., and A.R. analyzed data; R.L., Y.H., N.B., L.L., V.C.-C., N.M., G.W.W., and A.R. interpreted results of experiments; R.L., Y.H., N.B., L.L., and V.C.-C. prepared figures; R.L., Y.H., N.B., L.L., G.W.W., and A.R. drafted manuscript; R.L., Y.H., N.B., L.L., V.C.-C., D.S., A.S., N.M. B.K., G.W.W., and A.R. edited and revised manuscript; R.L., Y.H., G.W.W., and A.R. approved final version of manuscript.