Background
Epidemiological or cross-sectional studies in human populations support an inverse relationship between dietary calcium (Ca) and dairy food consumption with obesity. Several clinical weight loss intervention studies that included high-Ca or dairy foods indicated a positive effect on body fat loss [
1,
2], but this has not been universally observed [
3], suggesting that these effects are context-specific (e.g., dependent upon the specific population, dietary Ca status, type of dairy food, or provision of elemental Ca vs. Ca in a dairy matrix). In contrast, fat loss or reduced weight gain in response to high-Ca plus dairy protein in rodent obesity models has been consistently reported [
4‐
7]. The mechanisms underlying the metabolic effects of Ca and dairy foods to reduce adiposity remain to be elucidated. Increased fecal fat loss due to the formation of indigestible Ca soaps in the gastrointestinal tract has been proposed as a possible mechanism by which high dietary Ca reduces adiposity [
8]. Research from one group has provided evidence
in vitro and in the aP2-agouti transgenic mouse model of diet-induced obesity for a potential role of the calcitrophic hormone 1,25-dihydroxyvitamin D (calcitriol) in promoting adipocyte lipogenesis and inhibiting lipolysis, thus encouraging adipocyte lipid accumulation [
9‐
11]. According to this model, increasing dietary Ca would suppress circulating calcitriol, discouraging energy storage and white adipose tissue (WAT) expansion.
Dietary Ca may also impact inflammatory phenotypes associated with obesity. Chronic inflammation of WAT, especially in visceral depots, contributes to the pathogenesis of obesity-related insulin resistance [
12‐
14], and WAT macrophages are a prominent source of pro-inflammatory cytokine production [
15‐
18]. Calcitriol increased pro-inflammatory cytokine gene expression and secretion from cultured adipocytes and macrophages, and this response was dependent on intracellular Ca-provoked reactive oxygen species (ROS) generation; co-culture of both cell types increased gene expression and protein secretion of pro-inflammatory cytokines, which was further enhanced by calcitriol [
19].
In vivo evidence for anti-inflammatory properties of high dietary Ca was demonstrated in the aP2-agouti transgenic mouse obesity model in which increased dietary Ca was associated with reduced plasma calcitriol [
20,
21]. Despite a potential mechanism for direct anti-inflammatory properties of dietary Ca through suppression of circulating calcitriol, reduced inflammation could not be dissociated from lower body weight and adiposity in studies of the aP2-agouti transgenic mice. Furthermore, one cannot exclude a confounding effect of adipose agouti protein overexpression on these phenotypes, or possible effects of agouti expression in macrophages since the aP2 promoter can be activated in both macrophages and fat cells. Thus, further research is necessary to confirm potential anti-inflammatory properties of high dietary Ca and/or dairy in polygenic obesity models while controlling for body weight and adiposity changes typically observed in animals fed these diets.
Additional bioactive components of dairy foods (e.g. branched chain amino acids, peptide fragments with angiotensin converting enzyme (ACE)-inhibitory properties) might contribute to anti-obesity/anti-inflammatory outcomes, since high Ca in a dairy matrix yielded greater effects vs. elemental Ca alone in aP2-agouti mice [
22]. Studies demonstrating an anti-obesity effect of increased dietary Ca in polygenic obesity models have been only in the context of nonfat dry milk (NFDM) or whey providing all dietary protein [
4,
6,
7]. Thus, comparing high Ca diet in matrices containing differing protein and carbohydrate sources (dairy and non-dairy) is of interest to understand if non-Ca factors in dairy foods could play a role in regulation of metabolism and obesity-associated inflammation.
To address whether dietary Ca can impact obesity phenotypes and to understand if putative anti-inflammatory effects of high-Ca or dairy-based diets are yoked to body weight, we investigated the potential anti-obesity and anti-inflammatory effects of high dietary Ca in a well-characterized polygenic mouse model of diet-induced obesity (DIO): male C57BL/6J mice display susceptibility to diet-induced obesity and glucose intolerance when fed a high fat diet [
23] and we have previously described obesity with associated inflammation and impaired glucose homeostasis when feeding these mice a diet providing 45% energy as fat (with soy- and sucrose/cornstarch-based protein and carbohydrate sources, respectively) for up to 12 wk [
24]. We hypothesized that independent of differences in body weight gain, DIO mice fed high-Ca would have reduced body weight, decreased adiposity, and improved metabolic and inflammatory phenotypes compared to control mice fed a lower Ca diet. Our study design also enabled a comparison of metabolic and inflammatory outcomes in high-Ca DIO mice relative to animals fed high-Ca in a dairy matrix (NFDM as the sole protein source and containing primarily dairy-based carbohydrates).
Discussion
Consistent with previous reports [
4,
6,
7,
9], we showed reduced body weight and adiposity upon feeding a NFDM diet containing high Ca to high fat fed polygenic DIO rodents. Reductions in body weight and adiposity were accompanied by lower WAT inflammation and liver triglycerides as well as improved glucose tolerance. However, the data do not support the idea that high dietary Ca in isolation provides protection against obesity and associated inflammation since DIO mice fed a soy protein-based high-Ca diet actually became more obese and displayed higher WAT inflammation, compared to lower-Ca controls or mice fed high-Ca in a NFDM matrix. This differs from observations in high-Ca fed aP2-agouti transgenic mice fed an obesity-promoting diet where increased dietary Ca attenuated weight gain, adiposity, and WAT inflammation [
5,
20,
21,
34]. Experimental variables that may have contributed to these discrepancies include diet composition, energy consumption, genetic variation, and study duration.
In the present study, body weight gain was not significantly different between control mice and mice fed high-Ca when controlling for energy consumption as a covariate. Thus, increased energy consumption can explain differences in body weight gain between these diet treatment groups, whereas daily food intake was not different in aP2-agouti transgenic mice fed an obesigenic diet containing high Ca [
5,
20,
21,
34]. Similar to our results, 21 wk food consumption was increased in DIO male C57BL/6J mice fed high Ca in the context of a high fat diet (60% kcal as fat) containing whey, and a modest non-significant increase was observed in DIO mice fed a high Ca, casein-based diet [
6]. Mice fed a 43% fat, NFDM-based high-Ca diet did not display a difference in energy intake relative to low Ca, casein controls [
7]. Thus, results in polygenic DIO mouse models are consistent in that animals fed a high-Ca diet containing NFDM or whey protein display reduced body weight gain and adiposity despite increased or unchanged energy intakes. Protein source, and not Ca level, primarily contributed to body composition differences in DIO rats fed differing levels of Ca in the context of different dairy protein sources (NFDM, whey, or casein) [
35]. These observations indicate that non-Ca components derived from dairy mainly drive adiposity reduction in polygenic DIO rodents.
Suppression of circulating calcitriol via increased Ca intake has been proposed as a primary mechanism for the anti-obesity and anti-inflammatory properties of high dietary Ca [
22]. There is evidence that mice lacking the nuclear vitamin D receptor or enzyme required to generate calcitriol have reduced adiposity, lower serum leptin, and increased food intake [
36]. Reduced adiposity in aP2-agouti transgenic mice fed obesigenic high-Ca diets was associated with reductions in blood calcitriol [
20,
21]. However, we observed that adiposity and body weight outcomes in DIO mice fed high-Ca in isolation and in high-Ca + NFDM showed no relation to differences in plasma calcitriol, indicating that changes in calcitriol are not sufficient to explain differences in body weight and adiposity in this model.
Mice that consumed high-Ca in a soy protein-based diet demonstrated increased feed efficiency, in contrast to a lower feed efficiency in mice fed high-Ca + NFDM. These results could emanate from diet-related differences in gut energy uptake and/or tissue thermogenesis. There was in fact increased net uptake of energy across the gut in mice fed the soy-based high-Ca diet. An observed increase in fecal energy loss in high-Ca + NFDM fed mice cannot explain their reduced feed efficiency, since increased energy intake in this group negated any difference in metabolizable energy compared to controls. With respect to thermogenesis, reduced feed efficiency in mice fed high-Ca + NFDM was not accompanied by increased BAT UCP1 protein expression. Since UCP1 expression is a surrogate for BAT thermogenic activation, one interpretation is that in the high-Ca + NFDM mice enhanced BAT thermogenesis did not contribute to the anti-obesity effects of this diet. Alternatively, BAT UCP1 expression was maintained at control levels despite lower body weight in the high-Ca + NFDM mice, so another interpretation is that this diet resulted in relative retention of BAT thermogenic potential. The increased BAT size and UCP1 protein expression in obese mice fed high-Ca was reminiscent of that seen in cafeteria fed rats [
37], and this might be explained teleologically as a physiological adaptation to overnutrition. Additional studies of thermogenesis-related pathways in muscle and other tissues are warranted to clarify the mechanisms underpinning reduced feed efficiency in high-Ca + NFDM fed mice. The influence of calcitriol on WAT and muscle expression of IL-15 has been proposed to alter energy metabolism, with expression increased upon feeding high-Ca [
20,
26]. Increased IL-15 secretion has also been hypothesized to discourage lipid accumulation [
20,
38]. In our studies, WAT IL-15 gene expression was actually decreased in mice fed high-Ca compared to controls (a relationship held constant even after controlling body weight gain), and there were no significant differences in plasma IL-15 across diets. Thus, in polygenic DIO mice, differences in adiposity and feed efficiency were not associated with diet-related changes in WAT IL-15 expression or circulating IL-15 concentrations.
Non-Ca bioactive components of dairy thought to contribute to the anti-obesity effects of NFDM may be derived primarily from the whey protein fraction [
22]. Supplementation with whey protein isolate reduced adiposity and increased energy expenditure in high fat fed mice [
39]. Whey protein is a rich dietary source of branched chain amino acids (BCAA), which could in theory contribute to activation of thermogenesis. Leucine or whey protein isolate supplementation of drinking water increased energy expenditure and reduced diet induced obesity in mice without reducing energy intake [
39,
40]. Leucine when provided to muscle cell cultures increased energy consumption and indices of mitochondrial biogenesis (e.g., PGC-1 expression, mitochondrial mass), hypothesized to result from activation of mammalian target of rapamycin complex 1 (mTORC1) [
41,
42] or indirectly through anaplerosis [
43]. Whether BCAA derived from the NFDM diet contributes to metabolic outcomes requires further experimental validation. Notably, milk proteins also possess bioactive peptides that have ACE-inhibitory properties [
44], and the ACE product angiotensin II enhances adipocyte lipid storage, promotes inflammation via activation of the NFκB pathway, and stimulates ROS-mediated insulin resistance [
45].
Chronic inflammation of WAT, hallmarked by macrophage accumulation, contributes to obesity-associated morbidities. Since calcitriol evokes a pro-inflammatory response in both adipocytes and macrophages, and enhances pro-inflammatory mediators in co-cultures of both cell types
in vitro [
19], high dietary Ca suppression of plasma calcitriol may have anti-inflammatory properties [
20]. High dietary Ca both with and without dairy reduced WAT and systemic inflammation compared to low dietary Ca in aP2-agouti transgenic obese mice [
21]. However, interpretations of these observations are confounded by reduced adiposity resulting from high dietary Ca intake in that model. We found that WAT expression of the pan-macrophage surface antigen CD68 correlated positively with body weight regardless of diet. WAT mRNA abundance for CD11d (an immune cell-specific cell adhesion molecule expressed on a subset of macrophages [
46]) also correlated with body weight, but for any given weight was lower in NDFM-fed mice compared to controls or high-Ca fed mice. Diet-associated differences in most WAT inflammatory gene markers were lost when controlling for body weight. These results conclusively demonstrate that dietary Ca and/or NFDM effects on WAT inflammation in DIO mice are primarily explained by their impact on weight. In DIO mice fed different levels of fat (10%, 45%, and 60%), WAT macrophage marker expression correlated with body weight and adiposity in lean-to-moderately obese animals [
24], similar to our current observations. These findings support the hypothesis that WAT macrophage infiltration (at least for some macrophage sub-types) is closely yoked to expansion of energy storage capacity during modest caloric overnutrition, possibly as a mechanism for macrophage-mediated WAT extracellular matrix and vascular remodeling to accommodate normal physiologic adipocyte hypertrophy. Significant correlations between WAT inflammatory markers and body weight was reported when looking across disparate mouse obesity models [
15], and positive correlations were observed between %CD14+ cells in isolated stromal vascular cells from WAT and BMI in normal weight to obese humans [
47,
48]. Shaul et al. [
49] have also considered a remodeling function based on dynamic M2-like macrophage phenotypes in rodents. The observations that WAT macrophage marker expression closely tracks body weight (at least in lean to moderate obese states) highlights a need to understand weight-specific immune signals and macrophage functions during normal, non-pathological WAT expansion and contraction.
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
All authors edited the manuscript, and read and approved the final version of the paper. APT and TND were jointly responsible for conducting the studies and sample assays. APT and SHA were primarily responsible for writing, JBD and PJO assisted with sample processing and data analysis.