https://orcid.org/0000-0003-3101-650X
Abstract
Obesity is associated with increased macrophage and extracellular matrix accumulation in adipose tissue, which can be partially reversed following weight loss by daily caloric restriction. This study examined the effects of 8 weeks of intermittent fasting (IF; 24-hour fast on 3 nonconsecutive days per week) in mice fed a chow or high-fat diet (HFD; 43% fat) on markers of adipose tissue inflammation and fibrosis. We found that IF decreased energy intake, body weight, and fat cell size in HFD-fed mice and decreased fat mass and improved glucose tolerance in chow- and HFD-fed mice. IF decreased mRNA levels of macrophage markers (Lgals3, Itgax, Ccl2, and Ccl3) in inguinal and gonadal fat, as well as adipose tissue macrophage numbers in HFD-fed mice only, and altered genes involved in NLRP3 inflammasome pathway in both diet groups. IF increased mRNA levels of matrix metallopeptidase 9, which is involved in extracellular matrix degradation, and reduced mRNA levels of collagen 6 α-1 and tissue inhibitor of matrix metallopeptidase 1, as well as fibrosis in gonadal fat in HFD-fed mice. In summary, our results show that intermittent fasting improved glucose tolerance in chow- and HFD-fed mice and ameliorated adipose tissue inflammation and fibrosis in HFD-fed mice.
Daily caloric restriction (DR) promotes weight loss, improves health, and extends life span in a variety of species (1). Intermittent fasting (IF) has recently gained attention as a viable alternative to DR. IF extends life span, reduces fat mass, and improves glucose tolerance with a minimal impact on food consumption and body weight in chow-fed mice (2–5). The effect of IF on high-fat diet (HFD)–fed mice is less clear, with only four animal studies conducted recently (6–9). These studies have consistently reported that IF resulted in substantial weight loss, but improvements in glucose tolerance are controversial.
White adipose tissue undergoes remodeling during weight gain, including adipocyte hypertrophy and/or hyperplasia, increases in macrophage infiltration, and extracellular matrix (ECM) deposition (10–12). Increased accumulation of macrophages and ECM in adipose tissue has been linked to the development of insulin resistance (11, 13–15). Macrophages may also promote ECM deposition (16–18), which could negatively affect adipocyte expansion, promote ectopic lipid deposition, and have an impact on metabolic health (11, 19). Marked weight loss by DR reduces macrophage infiltration in adipose tissue (20–24) and may promote macrophage phenotype switching from a proinflammatory M1 toward an anti-inflammatory M2 profile (25, 26). There is also some evidence that long-term DR decreases markers of ECM synthesis in mouse adipose tissue (27).
Paradoxically, a beneficial role of inflammation in healthy adipose tissue expansion and function is reported (28). Increases in macrophage infiltration in adipose tissue occur in the early phase of DR and after 24 hours of fasting (23, 29–31), and this increase coincides with elevated circulating nonesterified fatty acids (NEFAs) (29). These studies have suggested that adipose tissue macrophages may play a positive role in buffering lipid released from adipocytes.
We examined the effects of IF on markers of adipose tissue remodeling in mice fed a chow diet or HFD. We hypothesized that IF would improve metabolic phenotype and reduce adipose tissue fibrosis but might increase adipose tissue macrophage infiltration.
Research Design and Methods
Ethical approval
All experimental protocols for the animal study were approved by the animal ethics committee of the South Australian Health and Medical Research Institute and the University of Adelaide, and they were performed in accordance with the Australian Code of Practice for the Care and Use of Animals for Scientific Purposes.
Animals and diets
Forty-eight male C57BL/6J mice (Bioresources; South Australian Health and Medical Research Institute, Adelaide, Australia) were housed four per cage under a 12:12-hour light/dark cycle, with lights on at 7:00 am [zeitgeber time (ZT) 0]. At 10 weeks old, mice were fed either a lard-based HFD comprising 43%, 21%, and 36% of energy from fat, protein, and carbohydrate, respectively (SF04-001; Specialty Feeds, Glen Forrest, WA, Australia), or a standard chow diet (chow) comprising 18%, 24%, and 58% of energy from fat, protein, and carbohydrate, respectively (2018SX; Envigo, Madison, WI) for 8 weeks. Mice on each diet were then randomized into ad libitum (AL) feeding (n = 8) or IF (n = 16) for another 8 weeks. IF was initiated at ZT11 (1 hour prior to lights off) for 24 hours for 3 nonconsecutive days/week (Fig. 1A). Food access was controlled by transferring mice daily between cages with or without food. AL-fed mice were also transferred between feeding cages at the same time to standardize handling. All mice had free access to water throughout the study. Body weight and food intake were monitored at ZT11 weekly before IF was introduced and daily after IF was implemented. At 28 weeks old, all mice were euthanized at ZT7 to ZT9 with mice in the IF group culled in the fed state or following a 22-hour fast. Whole blood was collected via cardiac puncture with isoflurane anesthesia. Following cervical dislocation, inguinal and gonadal adipose tissues were collected and weighed. Another group of 10-week-old male C57BL/6J mice (n = 24) were fed chow or HFD AL for 8 weeks and culled at ZT7 to ZT9 in the fed state or after a 22-hour fast.
IF differentially affects energy intake and adiposity in chow-fed and HFD mice. (A) Schematic outline of the IF regimen used in this study. (B, C) Cumulative energy intake during diet-induced obesity and IF. Results were calculated based on four mice/cage. (D, E) Body weight during diet-induced obesity and IF. (F, G) Gonadal and inguinal fat mass. (L, M) Gonadal and inguinal fat cell size. (H–K) Representative hematoxylin and eosin staining of gonadal fat. Scale bar, 100 μm. Mean ± SEM, n = 8 in AL and 16 in IF per diet in (A)–(D). n = 7 to 8 per group in (D)–(G). n = 5 to 6 in (H) and (I). Two-way ANOVA with Bonferroni post hoc test. Post hoc test: *P < 0.05, **P < 0.01, and ***P < 0.001; ^P < 0.01 vs chow-IF-fed and #P < 0.01 vs chow-IF-fast.
Glucose tolerance test
At 24 weeks of age and after 6 weeks of IF or AL feeding, mice were fasted from ZT1 for 6 hours and then challenged with an oral gavaging of glucose (2 g/kg body weight). Glucose was assessed at 0, 15, 30, 60, 90, and 120 minutes via tail vein bleeding by a glucometer (AccuChek Performa Monitor; Roche Diagnostics, Risch-Rotkreuz, Switzerland), and insulin was assessed at 0, 15, 30, and 60 minutes.
Plasma analysis
Insulin was measured using an ultra-sensitive ELISA kit (catalog no. 10-1249-01; Mercodia, Uppsala, Sweden) and NEFAs by enzymatic colorimetric assay [NEFA-HR (2); Wako Diagnostics, Osaka, Japan] on a VersaMax ELISA Microplate Reader (Molecular Devices LLC, Sunnyvale, CA).
Quantitative real-time PCR
As described previously (32), total RNA were extracted from gonadal and inguinal adipose tissue (100 to 150 mg) using TRI Reagent (T9424; Sigma, St. Louis, MO) following the manufacturer’s instructions. The concentration and purity of RNA were assessed by a NanoDrop Lite Spectrophotometer (Thermo Fisher Scientific, CA). cDNA synthesis was conducted using a T100 Thermal Cycler (Bio-Rad, Hercules, CA) with 1000 ng of each RNA sample using the QuantiTect reverse transcription kit (catalog no. 205313; Qiagen, Venlo, Netherlands) according to kit instructions. Standard control (25 ng/μL) samples were pooled from each sample. Quantitative real-time PCR was performed using TaqMan primers and Fast Universal PCR Master Mix (Applied Biosystems, Foster City, CA). Genes for pan macrophage [Adgre1 (Mm00802529_m1) and Lgals3 (Mm00802901_m1)], inflammatory M1 macrophage [Itgax (Mm00498701_m1) and Cd38 (Mm01220906_m1)], anti-inflammatory M2 macrophage [Arg1 (Mm00475988_m1) and Mrc1 (Mm01329362_m1)], macrophage recruitment [Ccl2 (Mm00441242_m1) and Ccl3 (Mm00441259_g1)], NLRP3 inflammasome pathway [Nlrp3 (Mm00840904_m1), Il-18 (Mm00434225_m1), and Il-1β (Mm00434228_m1)], ECM synthesis [Col3a1 (Mm01254476_m1) and Col6a1 (Mm00487160_m1)], and ECM degradation [Mmp2 (Mm00439498_m1), Mmp9 (Mm00442991_m1), and Timp1 (Mm01341360_g1)] were assessed using TaqMan primers. The samples were run in duplicate on an ABI 7500 sequence detection system (Applied Biosystems) with internal negative controls and a standard curve. Six reference genes, including Rn18s (Mm03928990_g1), Actb (Mm00607939_s1), Gapdh (Mm99999915_g1), Hprt (Mm01545399_m1), Ppia (Mm02342430_g1), and B2m (Mm00437762_m1), were examined, and the combination of Actb and B2m was determined as the best housekeeper using the NormFinder program as described previously (33). The relative gene expression was determined using the 2−ΔCT method, where ΔCT = (CTtarget gene – CTreference gene).
Histological analysis and immunohistochemical staining
Briefly, adipose tissue samples were fixed with Bouin solution (HT10132; Sigma), dehydrated, paraffin embedded, and sectioned at 5 μm. Hematoxylin and eosin staining was performed using a standard protocol and Masson trichrome staining with a commercial kit (HT15; Sigma). For immunohistochemistry, deparaffinized and rehydrated slides were incubated with ELOXALL solution (SP-600; Vector, Olean, NY) for 10 minutes at room temperature to eliminate endogenous peroxidase and alkaline phosphatase. Antigen retrieval was achieved using modified citrate-based buffer (S1700; Dako, Santa Clara, CA) and incubation in a 95°C water bath for 20 minutes. Slides were incubated with a rabbit anti-F4/80 (1:400; RRID: AB_1140040, catalog no. Ab6640; Abcam, Cambridge, United Kingdom) (34) overnight and then goat anti-rabbit secondary antibody (1:500; RRID: AB_955447; catalog no. Ab6721; Abcam) for 1 hour at room temperature (35). Immunohistochemical detection was performed using 3,3′-diaminobenzidine (SK-4105; Vector), and slides were counterstained with Mayer hematoxylin, followed by dehydrating and mounting. All slides were randomly assigned numeric codes by a research officer to blind the investigator (B.L.) quantifying outcomes. Slides were scanned using the Pannoramic 250 Flash II scanner (3DHISTECH, Budapest, Hungary). At least 1000 adipocytes were analyzed for adipocyte size and collagen content using ImageJ built-in macros (National Institutes of Health, Bethesda, MD). Crown-like structures and F4/80-positive cells were counted in 10 randomly chosen areas at ×40 magnification and adjusted by per 100 adipocytes as described previously (36).
Data analysis
All data are expressed as mean ± SEM. Data were analyzed statistically with SPSS 24 (SPSS, Inc., Chicago, IL) and log-transformed for analysis if not normally distributed. For insulin in the fasted mice, undetectable samples (three mice on each diet following 8 weeks of IF and four mice on chow diet following one acute fast) were input with the minimum values. Area under the curve (AUC) for glucose and insulin was calculated as mentioned previously (37, 38). Single comparisons were performed using two-way ANOVA with diet (chow and HFD) and schedule (AL and IF) as between-group factors, and Bonferroni post hoc tests were performed when diet by schedule effects were present. P < 0.05 was considered as statistically significant.
Results
IF differentially affects energy intake and adiposity in chow- and HFD-fed mice
Cumulative energy intake and final body weight were not different between chow-IF vs chow-AL groups (Fig. 1B–1E). In contrast, HFD-IF displayed reduced energy intake (−28.0%) vs the HFD-AL group and significant weight loss (−20.9%, both P < 0.001). Weight loss plateaued after 5 weeks of IF in the HFD-IF group, and the final body weight was not different from the chow-fed groups. Gonadal and inguinal fat mass was reduced as a result of IF in chow-fed and HFD groups, but fat pad weights remained higher in HFD-IF vs chow-IF (all P < 0.05, Fig. 1F and 1G). Fat cell size was decreased by IF in HFD-fed mice only (all P < 0.001, Fig. 1H–1M).
IF improved glucose tolerance in chow- and HFD-fed mice
IF improved glucose tolerance as assessed by glucose AUC in both diet groups (all P < 0.05, Fig. 2A and 2B). Insulin AUC was reduced by IF in HFD-fed mice only (P < 0.001, Fig. 2Dand 2E). We also measured glucose tolerance in the HFD-IF group after a 20-hour fast, but this was not different from the HFD-IF group that fasted for 6 hours (Fig. 2C). In the fed state, IF reduced terminal blood glucose in both diet groups (schedule effect, P < 0.01, Fig. 2F) but reduced insulin and homeostatic model assessment of insulin resistance (HOMA-IR) in HFD-fed mice only (both P < 0.05, Fig. 2G and 2H). Fasting glucose, insulin, and HOMA-IR were also reduced in the fasted vs fed state in IF groups (all P < 0.01, Fig. 2F–2H).
IF improved glucose tolerance in chow- and HFD-fed mice. (A–C) Glucose and area of glucose under the curve during oral glucose tolerance test (OGTT). (D, E) Plasma insulin and area of insulin under the curve during OGTT. (F, G) Terminal glucose and insulin. (H) Homeostatic model assessment of insulin resistance (HOMA-IR) calculated from terminal glucose and insulin. Mean ± SEM, n = 7 to 8 per group. Two-way ANOVA with Bonferroni post hoc test. Diet effect: ^^P < 0.01. Schedule effect: $P < 0.01 vs AL and &P < 0.001 vs IF-fed. Post hoc test: *P < 0.05, **P < 0.01, and ***P < 0.001; #P < 0.05 vs chow-IF.
IF reduced adipose tissue inflammation in HFD-fed mice
Markers of adipose tissue inflammation were increased in both gonadal and inguinal fat in the HFD-AL vs chow-AL group (Fig. 3A–3G). IF decreased mRNA levels of Lgals3, Itgax, and Ccl2 in both gonadal and inguinal fat, as well as decreased crown-like structure and pan-macrophage numbers in gonadal fat in HFD-fed mice (all P < 0.05, Fig. 3B, 3D, 3E, 3G, 3M, 3O, 3S, and 3T). IF did not alter any markers of inflammation in chow-fed mice, except for decreased Lgals3 mRNA levels in gonadal fat (P < 0.05, Fig. 3B).
IF reduced adipose tissue inflammation in HFD-fed mice. (A–D) mRNA levels of pan-macrophage markers Adgre1 and Lgals3 in gonadal and inguinal fat. (E–H) mRNA levels of M1 macrophage markers Itgax and Cd38 in gonadal and inguinal fat. (I–L) mRNA levels of M2 macrophage markers Mrc1 and Arg1 in gonadal and inguinal fat. (M, N) mRNA levels of macrophage recruitment markers Ccl2 and Ccl3 in gonadal fat. (O) mRNA levels of macrophage recruitment marker Ccl2 in inguinal fat; Ccl3 was undetectable in inguinal fat. (P) Plasma NEFAs. (Q, R) Representative images for immunohistochemical staining of F4/80 (pan-macrophage) in gonadal fat in (Q) chow and (R) HFD AL-fed mice. (S, T) Quantification of crown-like structures and F4/80-positive cells in gonadal fat; results were adjusted by per 100 adipocytes. Scale bar, 100 μm. Mean ± SEM, n = 7 to 8 per group for (A)–(P); n = 5 per group for (S) and (T). Two-way ANOVA with Bonferroni post hoc test. Diet effect: ^P < 0.05, ^^P < 0.01, and ^^^P < 0.001 vs chow. Schedule effect: $P < 0.05 vs AL and #P < 0.05 vs IF-fed. Post hoc test: *P < 0.05, **P < 0.01, and ***P < 0.001; &P < 0.05 vs chow-IF-fed.
Adgre1, Mrc1, Arg1, and Ccl2 mRNA levels were lower in fasted vs fed states in gonadal fat in IF groups (P < 0.01, Fig. 3A, 3I, 3J, and 3M). Crown-like structure and pan-macrophage numbers were not different between fed vs fasted states in both diets (Fig. 3Sand 3T). NEFA levels were increased in the fasted vs fed state in both diets (both P < 0.05, Fig. 3P), but there was no relationship between the changes in NEFAs and any marker of inflammation.
IF altered the NLRP3 inflammasome pathway in gonadal fat
The mRNA level of Il-1β was decreased, and Il-18 was increased by IF in both diet groups (schedule effect, both P ≤ 0.01, Fig. 4B and 4C). A diet-by-schedule effect was observed for Nlrp3 mRNA levels; these levels were reduced by IF in chow-fed mice but increased in HFD-fed mice (both P < 0.01, Fig. 4A).
IF altered the NLRP3 inflammasome pathway in gonadal fat. (A–C) mRNA levels of Nlrp3, Il-18, and Il-1β in gonadal fat. Mean ± SEM, n = 7 to 8 per group. Two-way ANOVA with Bonferroni post hoc test. Schedule effect: $P ≤ 0.01 vs AL. Post hoc test: **P < 0.01.
IF reduced fibrosis in gonadal adipose tissue in HFD-fed mice
IF increased Col3a1 mRNA levels in gonadal and inguinal fat (schedule effect, both P < 0.05, Fig. 5A and 5C) but decreased Col6a1 mRNA levels in gonadal fat (schedule effect, P < 0.01, Fig. 5B). IF increased Mmp9 mRNA levels and decreased Timp1 mRNA levels in gonadal fat in HFD-fed mice only (all P < 0.001, Fig. 5F and 5I). Collagen deposition assessed by histology was reduced in gonadal fat in the HFD-IF vs HFD-AL group (P = 0.05, Fig. 5K and 5L) but was increased in inguinal fat in the chow-IF vs chow-AL group (P < 0.001, Fig. 5M and 5N).
IF reduced fibrosis in gonadal fat in HFD-fed mice. (A–D) mRNA levels of ECM synthesis markers Col3a1 and Col6a1 in gonadal and inguinal fat. (E–J) mRNA levels of ECM degradation markers Mmp2, Mmp9, and Timp1 in gonadal and inguinal fat. (K, M) Representative images of Masson trichrome staining for gonadal and inguinal fat. (L, N) Quantification of collagen-integrated density in gonadal and inguinal fat. Scale bar, 100 μm. Mean ± SEM, n = 7 to 8 per group for (A)–(J) and n = 5 to 6 per group for (L) and (N). Two-way ANOVA with Bonferroni post hoc test. Diet effect: ^P < 0.05, ^^P < 0.01 vs chow. Schedule effect: $P < 0.05, $$P < 0.01, and $$$P < 0.001 vs AL; #P < 0.05, ##P < 0.01, and ###P < 0.001 vs IF-fed. Post hoc test: ***P < 0.001.
Decreased Col3a1mRNA levels but increased Col6a1 and Mmp2 mRNA levels in gonadal fat were observed in the fasted vs the fed state in both IF groups (schedule effect, all P < 0.05, Fig. 5A, 5B, and 5E).
Responses to a single acute fast in chow- and HFD-fed mice
Recent studies have shown that acute fasting for 24 hours increased markers of inflammation in adipose tissue (29–31). We did not detect this in response to IF, but habituation to fasting may have occurred. Thus, a group of mice was examined after the first exposure to an IF diet. In response to one acute 22-hour fast, blood levels of NEFAs were increased and insulin levels were decreased in both diet groups (schedule effect, all P < 0.001, Fig. 6A and 6B). Fasting did not increase the mRNA levels of any inflammation markers examined in gonadal fat or inguinal fat (Fig. 6C–6J, inguinal fat data not shown). These results were confirmed by histology (Fig. 6K and 6L). Col6a1 and Timp1 mRNA levels were decreased after one acute fast (schedule effect, all P < 0.05, Fig. 6N and 6Q).
Responses to a single acute fast in chow- and HFD-fed mice. (A, B) Plasma insulin and NEFA levels. (C, D) mRNA levels of pan-macrophage markers Adgre1 and Lgals3. (E, F) mRNA levels of M1 macrophage markers Itgax and Cd38. (G, H) mRNA levels of M2 macrophage markers Mrc1 and Arg1. (I, J) mRNA levels of macrophage recruitment markers Ccl2 and Ccl3. (K, L) Crown-like structures and F4/80-positive cells; results adjusted per 100 adipocytes. (M, N) mRNA levels of ECM synthesis markers Col3a1 and Col6a1. (O–Q) mRNA levels of ECM degradation markers Mmp2, Mmp9, and Timp1. Mean ± SEM, n = 6 per group. Two-way ANOVA with Bonferroni post hoc test. Diet effect: ^P < 0.05, ^^P < 0.01, and ^^^P < 0.001 vs chow. Schedule effect: $P < 0.05 and $$$P < 0.001 vs AL. Post hoc test: *P < 0.05 and ***P < 0.001.
Discussion
Inflammation and fibrosis of adipose tissue occur in obese animals and in humans, and they are associated with impaired glucose tolerance and insulin resistance (11, 13, 14, 25, 39). Caloric restriction improves this phenotype (3, 21, 27, 40). IF promotes weight loss and may improve glucose tolerance in diet-induced obese mice (7–9). However, the effects of IF on adipose tissue remodeling are unclear. Our results suggest that IF improves glucose tolerance in both lean and obese mice, as well as reduced adipose tissue inflammation and fibrosis in obese mice.
In this study, chow-fed mice were able to consume sufficient energy during feeding days to compensate for the fasting days and maintain body weight. However, mice that were fed a HFD were unable to compensate entirely, resulting in weight loss. This is consistent with recent studies (7, 9). In our study, HFD-IF mice presented with greater inguinal and gonadal fat mass compared with chow-IF mice. This is also in agreement with previous studies, which showed that formerly diet-induced obese mice retained greater adiposity compared with lean controls (41, 42). Gonadal and inguinal fat pad weight was also lower in chow-IF vs chow-AL mice, although these mice did not display overall weight differences. This could indicate increased lean mass after IF. Indeed, increased lean mass was reported in diet-induced obese mice subjected to a chow diet with IF (6). Preservation of lean mass was also observed in humans who underwent a modified IF intervention, where they were allowed to consume 25% of energy requirements on each fasting day (43), but was not observed in a recent study by our group who consumed ∼30% of requirements at breakfast prior to initiating a 24-hour fast (44).
In this study, we observed that oral glucose tolerance was improved by IF in both diet groups. In diet-induced obese mice, this finding is controversial. Gotthardt et al. (6) observed no change in oral glucose tolerance using the same dose applied in this study. However, glucose tolerance was improved in diet-induced obese mice when given at 1 g/kg orally or by intraperitoneal injection (7, 9, 45). None of the studies have included pair-fed groups, and thus the effects of IF cannot be distinguished from weight loss per se. In our hands, measurement of glucose tolerance after the prolonged fasting period did not alter oral glucose tolerance in mice that were fed an HFD. However, Joslin et al. (7) observed that glucose tolerance was impaired on fasted vs fed days. We have also shown transient insulin resistance by clamp following a fasting day after 8 weeks of IF in women with obesity (44). To our knowledge, no studies have examined measures of insulin sensitivity by tracers, or hyperinsulinemic-euglycemic clamp in response to IF in mice. However, insulin sensitivity by an insulin tolerance test was improved (6, 45) or unchanged (9).
Obesity is a low-grade inflammatory state with increased macrophage accumulation in white adipose tissue (13, 14). A large body of evidence has suggested that weight loss reduces adipose tissue macrophages and improves insulin sensitivity in mice (20–22, 24). Weight loss may also promote the phenotype switching of macrophages from an inflammatory M1 to an anti-inflammatory M2 profile (25, 26, 36). In this study, HFD increased markers of adipose tissue macrophages and inflammation, which was rescued by IF. This is in agreement with a recent study by Kim et al. (45), which showed that a modified IF regimen (which comprised 2 feeding days followed by 1 fasting day) reduced inflammation-related genes in gonadal fat in HFD-fed mice (45). However, we did not observe M2 polarization of macrophages following IF as reported by Kim et al. (45).
Increased macrophage accumulation in adipose tissue has also been reported during the early stage of weight loss by calorie restriction and after a 24-hour fast (29–31, 46). Further studies show that adipose tissue macrophages take up and store lipids (29, 47). Contrary to studies that have reported increased macrophage infiltration in response to an acute 24-hour fast (29–31), there was no change in adipose tissue macrophages in either IF group, after the fasting day, despite marked elevation in NEFAs. We theorized that this could be due to habituation to IF and therefore examined the response to the first day of IF in a separate group of mice fed chow or HFD. Under these conditions, fasting did not alter the mRNA levels of any inflammatory genes or macrophage numbers as assessed by histology. This discrepancy between studies could be due to differences in mouse strains, ages or diet composition, or the clock time during which the tissues were collected. In one study, tissue was collected from fasted mice at ZT2 (29), which was 7 to 8 hours earlier than the time tissue was collected in our study. This could be of importance because macrophage and cytokine levels are under circadian control (48). We should also consider that the “fed mice” in our study may not have eaten since lights on at ZT0, potentially elevating the baseline comparison level. However, the single reference gene used to normalize gene expression results in previous studies (29, 31) was reduced by acute fasting in our hands. In addition, that study adjusted the macrophage count by total nuclei, or white blood cells, in adipose tissue. Fasting reduces white blood cells, as well as monocyte counts in blood in mice and humans (49–52), which, if occurring in adipose tissue, could artificially elevate the number of macrophages detected.
Despite improvements in glucose tolerance, there was no change in markers of macrophages in IF mice that were fed a chow diet. This suggests that alternative mechanisms must contribute to the improved metabolic phenotype. In a recent study conducted by Li et al. (8), IF increased the length of small intestine and reshaped the gut microbiota composition. This was linked with the browning of white adipose tissue and contributed to the improved health in mice (8). Fasting also increases circulating ketone levels, including β -hydroxybutyrate (53), which blocks NLRP3 inflammasome-mediated inflammation (54). Thus, we also investigated whether IF reduced macrophage-independent inflammation in adipose tissue. IF decreased the mRNA expression of Nlrp3 and Il-1β, which has been shown to contribute to obesity-associated insulin resistance (55, 56), and increased the mRNA levels of Il-18 in adipose tissue, which protects mice from diet-induced obesity and insulin resistance (57). Thus, reductions in the NLRP3 inflammasome pathway could have contributed to improved glucose tolerance observed in response to IF in chow-fed mice.
Increased fibrosis in adipose tissue is linked with inflammation and insulin resistance (11, 58). The homeostasis of ECM is a balance between collagen synthesis and degradation by matrix metalloproteinases and tissue inhibitor of metalloproteinases that negatively regulate the enzyme activity of matrix metalloproteinase to avoid excessive degradation of the ECM (11, 12, 59). Our results show that IF reduced ECM synthesis and promoted ECM degradation in gonadal fat in mice that were fed an HFD. This result was supported by histology. In contrast, mRNA levels of collagen and collagen content were increased in inguinal adipose tissue by IF. This could be due to the different nature of collagen in gonadal and inguinal fat. In gonadal fat, collagen presents a periadipocyte property, surrounding adipocytes and dominantly locating in crown-like structures (11, 21). In inguinal fat, however, large fiber bundles are frequently presented in adipose tissue through which subcutaneous fat pads attach to the skin. We speculate that IF promotes fat pad loss but may have less impact on large fiber bundles than pericellular collagens. This highlights that not only the “quantity” but also the “structure or quality” of collagen may be important when assessing tissue fibrosis. Further studies are required to investigate the structure or stiffness of adipose tissue following IF.
Active crosstalk exists between adipose tissue macrophages and the ECM (12). In vitro, coculture of rodent primary macrophages with fibroblasts increases a number of proinflammatory cytokines and CC chemokines compared with macrophage monoculture (60). Human dermal fibroblasts are also able to polarize macrophages through an alternatively activated pathway ex vivo (61). In vivo, a time course study of weight gain in mouse nicely demonstrated that the fibrotic process in adipose tissue occurs prior to the detection of increased macrophage infiltration (62), suggesting that the ECM contributes to the recruitment of macrophages in adipose tissue in the development of obesity. This is also supported by a short-term overfeeding study in humans (63) in which 10% weight gain significantly increased the mRNA expression of COL1A1, COL3A1, SPARC, and TGFβ without altering macrophage markers in subcutaneous adipose tissue.
One limitation of this study is that only male mice were included and thus results cannot be extrapolated to females. To our knowledge, one study examined the phenotype of both male and female mice fed a chow diet or HFD following an alternate-day fasting protocol. These C57BL/6J female mice were resistant to HFD-induced obesity and did not display any weight loss following 6 weeks of alternate-day fasting, and their final body weight was not different from HFD-AL–fed female mice (9).
In conclusion, IF promoted fat mass loss and improved glucose tolerance in mice fed a chow diet or HFD. Adipose tissue inflammation and fibrosis were also improved as a result of IF in mice fed an HFD.
Abbreviations:
- AL
ad libitum
- AUC
area under the curve
- DR
daily caloric restriction
- ECM
extracellular matrix
- HFD
high-fat diet
- HOMA-IR
homeostatic model assessment of insulin resistance
- IF
intermittent fasting
- NEFA
nonesterified fatty acid
- ZT
zeitgeber time
Acknowledgments
Parts of this study were presented as an oral presentation at the Joint Scientific Meeting of The Australian and New Zealand Obesity Society and the Obesity Surgery Society of Australia and New Zealand, Adelaide, Australia, 2017.
Financial Support: L.K.H. was supported by an Australian Research Council Future Fellowship (FT120100027). B.L. was supported by an Australian Government Research Training Program Scholarship.
Author Contributions: B.L. performed the study, acquired and analyzed the data, and wrote the manuscript. A.J.P. supervised the study and collected and interpreted the data. G.H. and M.C. performed the study and acquired the data. G.A.W. interpreted the data and helped draft the manuscript. L.K.H. conceived, designed, and supervised the study and interpreted the results. All authors reviewed and approved the final manuscript. L.K.H. is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Disclosure Summary: The authors have nothing to disclose.
