Short-Term Caloric Restriction Attenuates Obesity-Induced Pro-inflammatory Response in Male Rhesus Macaques

White adipose tissue (WAT) hypertrophy is an essential hallmark of obesity and is associated with the activation of resident immune cells. While the benefits of caloric restriction (CR) on health span are generally accepted, its effects on WAT physiology are not well understood. We previously demonstrated that short-term CR reverses obesity in male rhesus macaques exposed to a high-fat Western-style diet (WSD). Here, we analyzed subcutaneous WAT biopsies collected from this cohort of animals before and after WSD and following CR. This analysis showed that WSD induced adipocyte hypertrophy and inhibited β-adrenergic-simulated lipolysis. CR reversed adipocyte hypertrophy, but WAT remained insensitive to β-adrenergic agonist stimulation. Whole-genome transcriptional analysis revealed that β3-adrenergic receptor and de novo lipogenesis genes were downregulated by WSD and remained downregulated after CR. In contrast, WSD-induced pro-inflammatory gene expression was effectively reversed by CR. Furthermore, peripheral blood monocytes isolated during the CR period exhibited a significant reduction in the production of pro-inflammatory cytokines compared to those obtained after WSD. Collectively, this study demonstrates that short-term CR eliminates an obesity-induced pro-inflammatory response in WAT and peripheral monocytes.


Introduction
Consumption of a high-fat/calorie-dense Western-style diet (WSD) and physical inactivity are the main risk factors contributing to obesity characterized by a chronic low-grade pro-inflammatory state and insulin resistance [1]. A widely used approach for weight management is caloric restriction (CR). Studies in several organisms, including worms, flies, mice and nonhuman primates (NHPs), have shown that CR promotes survival [2] in part through the extension of health span [3,4]. In humans, CR decreases the risk of cardiovascular and metabolic diseases [5][6][7][8][9] and reduces obesity in older adults [10]. Furthermore, CR has been shown to decrease body weight and adiposity in postmenopausal women [11][12][13]. However, a significant fraction of these women regained weight soon after the termination of dietary intervention [14]. Moreover, we have shown that long-term CR initiated during early adulthood delayed T-cell senescence [15,16]. Specifically, CR initiated during early adulthood preserved circulating naïve CD8 and CD4 T-cells, maintained T-cell receptor diversity, and reduced T-cell production of pro-inflammatory cytokines.

Cell-Based Assays
Free fatty acid (FFA) uptake, adipocyte size and lipolysis was conducted as previously described [17,[20][21][22]. For the lipolysis assay,~200-mg SC-WAT explants were placed in M199 media (General Electric Company, Boston, MA, USA) at room temperature and transported to the laboratory. 50-mg SC-WAT explants (three basal and three isoproterenol-stimulated replicates) were placed in a 48-well plate containing 0.2 mL incubation medium (Hank's Salt (HBSS), 0.2% BSA (Sigma-Aldrich, St, Louis, MO, USA) and 5 mM glucose), and incubated at 37 • C for 2 h with or without 10 µM isoproterenol (Sigma-Aldrich, St. Louis, MO, USA). Glycerol release was determined using a glycerol detection kit (Sigma-Aldrich, St. Louis, MO, USA). Glycerol concentrations were normalized to wet tissue weight. For FFA uptake, 100 mg of SC-WAT were collected in M199 media at room temperature and separated into smaller explants. WAT explants were incubated floating in a 48-well plate filled with 0.4 mL incubation medium (M199 medium, 0.1% fatty acid-free BSA (Sigma-Aldrich, St. Louis, MO, USA), 20 mM HEPES (pH 7.4)). Fluorescently-labeled fatty acid BODIPY-500/510 C 1 , C 12 (BODIPY-C12; Life Technologies, Waltham, MA, USA) was prepared by diluting a 2.5 mM methanol stock solution in incubation medium to a final concentration of 10 µM and incubated for 15 min in a 37 • C water bath. One hundred µL of diluted BODIPY-C12 and 2 µL of live-cell staining dye Calcein Red-Orange AM (Life Technologies, Waltham, MA, USA) were added to each well and WAT explants were incubated for 15 min at 37 • C. Media was removed and explants were washed three times with incubation medium. WAT explants were fixed at room temperature in 4% paraformaldehyde (Sigma-Aldrich, St. Louis, MO, USA)/PBS for 20 min, washed with PBS and analyzed by confocal microscopy. WAT explants were placed into a 35-mm glass-bottom imaging culture dish (MatTek, Ashland, MA, USA) in PBS. Confocal microscopy was performed using a Leica SP5 AOBS spectral confocal system supplied with a ×20 objective. Adipocyte area and BODIPY fluorescence were measured using Fiji imaging software. The statistical analysis of adipocyte parameters was conducted using repeated measure one-way ANOVA followed by the Tukey's multiple comparison test using GraphPad Prism 8.2.0.

RNA-Seq Analysis
200 mg of frozen SC-WAT samples were homogenized with a TissueLyzer-II (QIAGEN, Hilden, Germany). RNA was isolated using the AllPrep DNA/RNA purification kit (QIAGEN). High-quality RNA (RIN >8) was used for library construction. RNA-Seq libraries were prepared using the TruSeq protocol (Illumina, San Diego, CA, USA) as described in [18]. Poly (A) + RNA was purified using oligo-dT coated magnetic beads and chemically fragmented followed by cDNA generation using random hexamer primers. The cDNA ends were repaired and ligated to library adaptors. Following purification with AMPure XP beads (Beckman Coulter Inc., Brea, CA, USA), libraries were amplified using 11 PCR cycles. Amplified libraries were cleaned using AMPure XP beads. Libraries were analyzed on a Bioanalyzer (Agilent, Santa Clara, CA, USA) and quantified using qRT-PCR (Kapa Biosystems, Wilmington, MA, USA) on a StepOnePlus qRT-PCR workstation (Life Technologies, Carlsbad, CA, USA). Libraries were multiplexed and final concentrations were determined by qRT-PCR. Multiplexed libraries were diluted to 1 nM for denaturation and then diluted to deliver optimal clustering on the flow cell. Flow cells were prepared on a cBot (Illumina, San Diego, CA, USA). Libraries were sequenced on a HiSeq 2000 (Illumina, San Diego, CA, USA). Data was assembled into standard fastq files using Bcl2Fastq (Illumina, San Diego, CA, USA). The quality of the raw reads was verified using FastQC (version 0.11.3). Low quality bases as well as any remaining Illumina adapters were trimmed. Reads with less than 25 bases remaining were discarded. The remaining reads were aligned to the rhesus macaque genome (genome assembly: Macaca mulatta 1.0) from ENSEMBL using the splice aware short read aligner suite Bowtie2/TopHat2 [23] in a strand-specific fashion allowing up to 5% mismatches. Reads were assigned to features using the featureCounts function of the Rsubread package [24] in R 3.4, using default parameters except for setting allowMultiOverlaps to TRUE. Transcripts were normalized and analyzed using the glm functionality of edgeR [25]; the Nutrients 2020, 12, 511 4 of 17 candidate differentially expressed genes (DEGs) were identified as those genes with a fold change (FC) ≥ 2 and a false discovery rate (FDR) ≤ 0.05.

Bioinformatic Analysis
Differential gene expression results (fold change, p-value, and FDR) were imported into the Ingenuity Pathway Analysis program (IPA, QIAGEN, Hilden, Germany). To facilitate this, ENSEMBL rhesus gene IDs were first mapped to human gene IDs using Ensembl's BioMart. Only high-confidence matches were used, and only the top match was chosen when one rhesus ID mapped to many human gene IDs. If no human gene ID was found, the rhesus gene ID was retained. Once imported into IPA, a fold-change cutoff of 1.5 and an FDR cutoff of 0.05 was applied to all comparisons to select differentially expressed genes. The STRING protein-protein interaction analysis (https://string-db.org/) was performed by pasting ENSEMBL IDs of DEGs into the STRING input window. The evidence-based STRING analysis was performed using the medium confidence value of 0.4.

WSD-Induced Adipocyte Hypertrophy but not β-Adrenergic Resistance is Reversed by CR
In our previous studies, we demonstrated that adult male rhesus macaques exposed to six months of WSD exhibited a significant increase in fat mass. Furthermore, fat mass was decreased significantly following the subsequent CR period. We also determined that the consumption of a WSD induced systemic insulin resistance, while CR restored normal insulin sensitivity in rhesus macaques [17,18]. The analysis of SC-WAT biopsies collected during each dietary period showed that WSD led to a significant increase in the mean area of SC adipocytes, while subsequent CR reduced adipocyte area to the baseline levels ( Figure 1B). To assess the effect of diet on lipolysis, SC-WAT explants derived from three longitudinal Nutrients 2020, 12, 511 5 of 17 biopsies were incubated ex vivo in basal media or in the presence of the β-adrenergic receptor agonist isoproterenol, and glycerol release in incubation media was quantified. SC-WAT collected during the WSD period exhibited a blunted β-adrenergic response compared to a chow diet, while the basal lipolysis remained unchanged. Interestingly, CR failed to restore a normal β-adrenergic response in SC-WAT despite a significant reduction in adipocyte area observed at the end of the CR period ( Figure 1C). To assess the uptake of FFA in adipocytes, SC-WAT explants were incubated with the metabolizable fluorescent FFA analogue BODIPY-C12 [21,22,[26][27][28] and single-cell fluorescence was quantified by confocal microscopy. SC adipocytes from WSD biopsies exhibited a trend to higher FFA uptake compared to the chow control and the CR conditions ( Figure 1D,E). Collectively, this study shows that WSD induces SC adipocyte hypertrophy, which was efficiently reduced to a pre-obese level by CR. In contrast, WSD led to the inhibition of the β-adrenergic lipolytic response in SC adipocytes and this effect persisted even after adipocyte size, body weight and insulin sensitivity [17,18] were normalized by CR.
Nutrients 2020, 12, x FOR PEER REVIEW 5 of 17 was quantified. SC-WAT collected during the WSD period exhibited a blunted β-adrenergic response compared to a chow diet, while the basal lipolysis remained unchanged. Interestingly, CR failed to restore a normal β-adrenergic response in SC-WAT despite a significant reduction in adipocyte area observed at the end of the CR period ( Figure 1C). To assess the uptake of FFA in adipocytes, SC-WAT explants were incubated with the metabolizable fluorescent FFA analogue BODIPY-C12 [21,22,[26][27][28] and single-cell fluorescence was quantified by confocal microscopy. SC adipocytes from WSD biopsies exhibited a trend to higher FFA uptake compared to the chow control and the CR conditions ( Figure 1D, E). Collectively, this study shows that WSD induces SC adipocyte hypertrophy, which was efficiently reduced to a pre-obese level by CR. In contrast, WSD led to the inhibition of the βadrenergic lipolytic response in SC adipocytes and this effect persisted even after adipocyte size, body weight and insulin sensitivity [17,18] were normalized by CR.

WSD-Induced Pro-Inflammatory Gene Expression in SC-WAT Is Reversed by CR
Three SC-WAT biopsies collected at the end of each dietary period were subjected to whole-genome RNA-Seq transcriptional analysis. DEGs were identified using three comparisons: WSD/CHOW, CR/WSD, and CR/CHOW (Supplementary Table S1). This experimental design allowed us to identify diet-specific DEGs, as well as genes whose expression did not change after the transition from WSD to CR. Functional enrichment was performed on WSD/CHOW, CR/WSD, and CR/CHOW subsets using IPA. Additionally, we identified 301 DEGs whose expression was significantly upregulated by WSD, then subsequently significantly downregulated by CR (Figure 2A and Supplementary Table  S2). DEGs of this category with the highest fold change encoded pro-inflammatory factors such as S100A7/8, CXCL1, CCL3, CCL5 and MMD; chemokine receptors e.g., CCR1 and CXCR2, as well as proteins involved in the heme biosynthesis pathway, such as ALAS2, HBB, HBA1 and CYPE (Table 1). WSD induced the upregulation of genes involved in granulocyte (p-value 4.2 × 10 −12 ) and agranulocyte (p-value 8.1 × 10 −10 ) migration and cell adhesion. The top upstream regulators identified by IPA included the pro-inflammatory cytokines TNFα (p-value 4.5 × 10 −21 ) and interleukin-1β (IL-1β, p-value 4.8 × 10 −19 ). To identify the functional interaction networks between proteins encoded by DEGs, we used the STRING Functional Protein Network Analysis. This analysis revealed the largest node of functional connectivity associated with immune response (n = 85; FDR 3.7 × 10 −23 ) and neutrophil activation (n = 34; FDR 3.3 × 10 −11 ) ( Figure 2C, blue and yellow). Furthermore, we identified three functional nodes representing oxygen transport processes (n = 9; FDR 1.3 × 10 −8 ), the histone H3 network (n = 4, FDR 0.01), and lipoprotein metabolism (n = 23; FDR 0.02) ( Figure 2C and Table 2, "LDLR"). Intriguingly, STRING identified a well-segregated node of functionally interacting proteins implicated in the regulation of cell cycle progression, DNA replication and intracellular signaling ( Figure 2C, "Cell Cycle"). Collectively, this analysis suggests that WSD induces a pro-inflammatory transcriptional response in SC-WAT that is effectively reversed by subsequent CR.

CR Diminishes a Pro-Inflammatory Response in Circulating Monocytes
Obesity is associated with the development of a systemic pro-inflammatory state and the production of pro-inflammatory cytokines. In contrast, CR typically exerts the opposite effect on the immune system, leading to a reduced pro-inflammatory response [34], although the later has not been directly demonstrated in a NHP model of CR. To examine these dietary effects further, PBMCs isolated after exposure to WSD and at the end of CR were treated with the toll-like receptor-4 (TLR4) agonist LPS and the frequency of TNFα-producing monocytes was quantified using intracellular staining and flow cytometry. Approximately 10-40% of the total monocyte population isolated during the WSD period were TNFα-positive following LPS stimulation. In contrast, very few monocytes isolated at the end of the CR period expressed TNFα following LPS stimulation, suggesting that fasting supported the loss of obesity-induced pro-inflammatory phenotype in circulating monocytes ( Figure 4A). Similar results were obtained using another pro-inflammatory cytokine IL-6, which was also downregulated in monocytes by CR ( Figure 4B). In contrast to TNFα and IL-6, the percent of IFNγ + monocytes dramatically increased following a transition from WSD to CR ( Figure 4C), suggesting that WSD suppressed and CR supported the production of a cytokine that is critical for antiviral and antibacterial immunity. Interestingly, a transition from WSD to CR did not cause significant changes in the proportion of total circulating monocytes or the ratio of classical (CD16−) to non-classical (CD16+) monocyte subtypes in peripheral blood ( Figure 4D-F), suggesting that CR may modulate the functional capacity of the cells rather than cause major shifts in monocyte population frequencies.

CR Diminishes a Pro-Inflammatory Response in Circulating Monocytes
Obesity is associated with the development of a systemic pro-inflammatory state and the production of pro-inflammatory cytokines. In contrast, CR typically exerts the opposite effect on the immune system, leading to a reduced pro-inflammatory response [34], although the later has not been directly demonstrated in a NHP model of CR. To examine these dietary effects further, PBMCs isolated after exposure to WSD and at the end of CR were treated with the toll-like receptor-4 (TLR4) agonist LPS and the frequency of TNFα-producing monocytes was quantified using intracellular staining and flow cytometry. Approximately 10-40% of the total monocyte population isolated during the WSD period were TNFα-positive following LPS stimulation. In contrast, very few monocytes isolated at the end of the CR period expressed TNFα following LPS stimulation, suggesting that fasting supported the loss of obesity-induced pro-inflammatory phenotype in circulating monocytes ( Figure 4A). Similar results were obtained using another pro-inflammatory cytokine IL-6, which was also downregulated in monocytes by CR ( Figure 4B). In contrast to TNFα and IL-6, the percent of IFNγ + monocytes dramatically increased following a transition from WSD to CR ( Figure 4C), suggesting that WSD suppressed and CR supported the production of a cytokine that is critical for antiviral and antibacterial immunity. Interestingly, a transition from WSD to CR did not cause significant changes in the proportion of total circulating monocytes or the ratio of classical (CD16−) to non-classical (CD16+) monocyte subtypes in peripheral blood ( Figure 4D-F), suggesting that CR may modulate the functional capacity of the cells rather than cause major shifts in monocyte population frequencies.

The Effects of WSD and CR on a Pro-Inflammatory Response
The main finding of the present report is that short-term CR efficiently reversed an obesity-induced pro-inflammatory transcriptional response in SC-WAT and reduced the production of pro-inflammatory cytokines in LPS-treated peripheral monocytes. The latter finding is consistent with the reported reduction in the cytokine levels in LPS-treated T-cell populations isolated from long-term calorically restricted rhesus macaques [16,34], suggesting that fasting can have a beneficial effect on both adaptive and innate immune systems. It is important to emphasize that previous NHP studies were conducted using standard monkey chow low in saturated fats [16,34], while we used WSD matching the composition of a typical diet consumed in a modern society. Since WAT inflammation is strongly associated with insulin resistance [35], the present study suggests that a relatively short-term energy restriction can efficiently mitigate obesity-induced insulin resistance and reverse a pro-inflammatory phenotype in both SC-WAT and peripheral monocytes. Studies shows that the key pro-inflammatory cytokines driving insulin resistance are TNFα and IL1β secreted by local macrophages [36]. Furthermore, the suppression of a pro-inflammatory response by CR has been attributed to the production of ketone bodies and acetoacetate in the liver as alternative energy sources during nutrient deficiency [37]. The anti-inflammatory effect of ketogenic bodies has been linked to the direct inhibition of the NLRP3 inflammasome in human monocytes stimulated with LPS [38], which is consistent with the present and previous reports [16,34] that CR causes a significant reduction in the production of pro-inflammatory cytokines in rhesus monocytes and T-cells, respectively. The induction of erythroid-specific genes in SC-WAT by WSD reported here is a new phenomenon. Although the expression of HBB in mouse macrophages has been previously reported [39], the present study shows that WSD induces and CR reverses the expression of multiple erythroid-specific genes, including embryonic hemoglobin genes HBZ and HBE1. To our knowledge, this is the first report that shows a diet-specific regulation of erythroid genes in peripheral WAT. This may represent an increased production and/or bone marrow output of the erythroid progenitors in the presence of WSD. This hypothesis is consistent with our previous study demonstrating that WSD induced the expression of erythroid genes in the skeletal muscle of rhesus macaques [18].

The Effects of WSD and CR on Lipolysis
The present study demonstrates that male rhesus macaques transitioned from a low-fat diet to a WSD for six months exhibited SC adipocyte hypertrophy, a blunted β-adrenergic lipolytic response, and the downregulation of the ADRB3 gene in SC-WAT. While CR efficiently reduced SC adipocyte size, it did not normalize the β-agonist lipolytic response and failed to restore ADRB3 expression in SC-WAT. Several studies have shown that obese patients undergoing weight loss interventions experience weight regain [40][41][42], which coincides with functional changes in WAT [43]. Consistent with the present report, obesity has been shown to be associated with the development of lipolytic resistance of WAT to catecholamines [44]. A number of studies, however, have also demonstrated that CR increases the rate of catecholamine response and the density of β-adrenoceptors in WAT of obese patients [45][46][47], while others reported no significant effects of CR on lipolysis [48][49][50][51][52]. Interestingly, studies have shown that, during the weight-maintenance phase after the bariatric surgery, there was a marked decrease in the levels of hormone-sensitive lipase (HSL) in SC-WAT of women but not in men [53]. Similarly to the present study, it has been shown that the WAT lipolytic response of obese patients three years after initiation of a weight-reducing program was lower compared to the pre-diet conditions [54]. These studies demonstrate that the initial improvement of the lipolytic response after weight loss persists during the dynamic phase of energy restriction, but returns to lower values after adaptation to a new energy balance during the weight-maintenance phase. We also observed the sustained downregulation of GPR39 in SC-WAT by WSD. GPR39 deficiency has been reportedly associated with obesity and reduced lipolysis in adipocytes [32,33].

The Effects of WSD and CR on De Novo Lipogenesis Pathway
The present study demonstrates that DNL and ChREBP genes, whose expression in SC-WAT was suppressed by WSD, remained downregulated after CR. Previously, it has been shown that the WAT-specific isoform of the transcription factor ChREBP, ChREBPβ, is the central regulator of glucose-mediated activation of lipogenic genes, while its expression in WAT correlates with increased insulin sensitivity [29]. ChREBP activation in WAT is altered in obesity, resulting in the downregulation of ChREBPβ, ACLY, ACACA, FASN, and ELOVL6 [55]. Activation of DNL in WAT has a positive effect both on direct glucose disposal [56,57] and metabolic homeostasis through the production of lipokines [58,59]. While the adipose-specific DNL pathway significantly contributes to body lipid reserves and thus directly regulates glucose homeostasis in humans [58], new studies have shown that it is also involved in the synthesis of biologically active branched FFAs that participate in the regulation of systemic insulin sensitivity and have anti-inflammatory properties [60,61]. The metabolic role of DNL-derived lipokines, in particular palmitoleate (C16: 1n−7), has been extensively described in a recent review [59]. The lipid biosynthesis pathway is dysregulated in obesity, which has been associated with the remodeling of lipid composition of adipocyte membranes [62] and the reduced presence of polyunsaturated fatty acids in high-BMI individuals [63]. Thus, we propose that WSD exerts a long-lasting effect on the lipolytic and DNL pathways in adipocytes, possibly through epigenetic modifications. Consistent with this hypothesis, we recently conducted the whole-genome gene expression and epigenetic analysis in visceral WAT of female rhesus macaques exposed to a WSD and excess androgens in combination or individually. This study demonstrated that testosterone could modulate the effects of WSD on transcription of lipid metabolism genes and DNA methylation in females WAT [64].

Conclusions
In summary, the present study advances our understanding of the functional and transcriptional response in SC-WAT associated with dynamic changes in fat mass in response to diet-induced obesity and weight loss and expands our previous studies in NHPs [17,18,20,[64][65][66]. Our study has several limitations. In the present report, we used the RNA-Seq analysis of whole WAT. Further studies are needed to understand the immunophenotypic changes in WAT (in both SC and visceral depots) in the content of resident immune cells. While several previous reports addressed the metabolic adaptation of WAT to weight loss through energy restriction [67][68][69], sex-specific mechanisms that control WAT function in health, during the development of obesity and after weight loss are also not completely understood [31,70] and need to be further explored in a NHP model. Further studies are needed to understand the role of epigenetic factors, including DNA methylation and histone modifications, in programing WAT transcriptome in males and females.
Supplementary Materials: The following are available online at http://www.mdpi.com/2072-6643/12/2/511/s1, Table S1: Title: Diet-specific changes in gene expression, Table S2: Title: Top 200 WSD-induced genes whose expression was reversed by CR, and Table S3: Title: WSD-suppressed genes whose expression was not reversed by CR. Funding: This study was supported by NIH grants R21AG047543-01 to OV, and P51 OD01192 for operation of the Oregon National Primate Research Center.