Introduction
It has been hypothesized that due to limited availability of food throughout the majority of human evolution, the body was more adapted towards intermittent feeding rather than to regular meal intervals as currently practiced in the developed world [
1]. Regular access to high calorie diets has contributed to an increase in obesity and associated increases in morbidity and mortality [
2]. Studies of obesity and its antithesis, caloric restriction (CR), in humans and animals have provided insight into the cellular and molecular mechanisms underlying normal aging and chronic diseases including type 2 diabetes, cardiovascular disease, cancers and neurodegenerative disorders [
3‐
5]. The multi-system pleiotropic effects of dietary restriction also extend to the immune system. Many studies suggest that long-term CR improves several components of immune function including responses of T cells to mitogens, natural kill cell (NK) activity, cytotoxic T lymphocyte (CTL) activity and the ability of mononuclear cells to produce pro-inflammatory cytokines [
6‐
8]. CR attenuated the age-associated increase in ratio of memory to naïve T cells in monkeys, and this was associated with a reduction in the pro-inflammatory cytokines, TNF-α and IL-6 [
9]. It has been suggested that a prominent immune-enhancing effect of CR on NK cells and CTL mediates, in part, the reduced incidence of tumors in mice maintained on CR diets [
10,
11].
Data from controlled studies in rodents suggest that intermittent fasting (IF) can protect against age-related diseases and can extend lifespan, and that at least some of the beneficial effects of IF may be independent of calorie intake [
1,
4]. For example, alternate day fasting protected neurons in the brains of mice against dysfunction and degeneration in models of Parkinson's and Alzheimer's diseases and stroke [
12‐
14]. IF resulted in improved glucose regulation and cardiovascular function [
15,
16] and protected the heart against ischemia reperfusion injury [
17]. The latter study provided evidence that the cardioprotective effect of IF is associated with an attenuation of tissue inflammation. Increasing evidence suggests that the signaling mechanisms that regulate energy metabolism and immune function are tightly coupled to each other [
18,
19]. For example, fasting can significantly attenuate inflammation and the development of autoimmune encephalomyelitis [
20]. In addition, the orexigenic hormone ghrelin can act on various immune cell subsets and inhibit pro-inflammatory cytokine production [
21]. Furthermore, genomic profiling studies in rodents revealed that CR can reverse the increased inflammation associated with aging [
22] and inhibit the release of proinflammatory mediators from macrophages [
23].
Recent findings from the Comprehensive Assessment of Long-Term Effects of Reducing Intake of Energy (CALERIE) study suggest that CR has effects on energy metabolism and disease risk in humans that are similar to those seen in rodents [
24]. In humans, long-term CR was reported to be highly effective in reducing the risk for atherosclerosis and associated pro-inflammatory markers [
25], and moderate CR improved cell-mediated immunity [
26]. In contrast to the increasing literature describing effects of CR on the immune system, there have been no reports of studies of how reduced meal frequency/IF affects immune function. It was recently reported that an alternate day calorie restriction IF dietary regimen resulted in a marked improvement in the symptoms of asthma patients, and an associated reduction in serum markers of oxidative stress and inflammation [
27]. However, the IF diet included a large reduction in calorie intake such that the relative contributions of CR and fasting to the outcomes is unknown.
We have previously reported on a human meal frequency study in which the daily calories were held constant between two diets that differed only in meal frequency (3 smaller meals versus one large meal). In this study, a large number of physiological variables were measured, including heart rate, body temperature and blood chemicals and many of these were unaffected by altering meal frequency [
28]. However, when on 1 meal per day, subjects did exhibit a significant reduction of fat mass and significant increases in levels of total, low-density lipoprotein, and high density lipoprotein cholesterol. Moreover, in this same study, Carlson and coworkers [
29] demonstrated that the morning glucose tolerance was found to be impaired in subjects consuming 1 meal per day compared with 3 meals per day. Fasting (morning) plasma glucose levels were also significantly elevated in subjects when they were consuming 1 meal per day (OMD) compared with 3 meals per day (TMD). This OMD diet effect on glucose tolerance was rapidly reversed upon return to the TMD diet, indicating that the diet had no long-lasting effect on glucose metabolism. Interestingly, there were no significant effects of meal frequency on plasma levels of ghrelin, adiponectin, resistin or BDNF.
In a follow-up to these studies, we have here examined the impact of different meal frequencies (without a difference in calorie intake) on plasma inflammatory markers (CRP, sgp130, visfatin) and activation-induced PBMC cytokine expression in normal weight human male and female subjects. Our data suggest that a change in diet causes a transient increase in TCR- and TLR4-mediated pro-inflammatory cytokine production by peripheral blood mononuclear cells (PBMCs), and that the magnitude of these alterations is less when subjects consume OMD vs. TMD.
Subjects and Methods
Subjects, Study Design and Diets
Details of the subject population, selection criteria and study design have been reported [
28]. Briefly, subjects were healthy 40-50 year-old males and females with a body mass index (BMI) between 18 and 25 kg/m
2 with a usual eating pattern of TMD. The experimental protocol was approved by the Johns Hopkins University Committee on Human Research and the MedStar Research Institute Institutional Review Board, and all subjects gave their informed consent. As this is the first study of its kind, there is no historical data for comparison in design and to determine wash out periods. This study was designed based on animal studies in which we found that many physiological variables (heart rate, blood pressure, insulin levels) returned to baseline levels within 2-4 weeks of wash-out [
1,
12‐
14]. Thus, the subjects in this study were divided into two controlled diet groups, a TMD diet and an OMD diet, in a washout and crossover design - 2 months on diets, 2 months off diets, crossover 2 months on diet - with the study lasting 6 months. During both 2-month controlled diet periods, each subject consumed dinner at the Human Study Facility under the supervision of a registered dietitian. Only foods provided by the Human Study Facility were allowed to be consumed during the study. Subjects were allowed unlimited amounts of caloric-free liquids and foods. Prior to initiation of the experimental diets, the energy requirements for weight maintenance were calculated for each subject using the Harris- Benedict formula, which estimates basal energy expenditure, and multiplied by an activity factor of 1.3-1.5. This formula has proven successful in estimating weight-maintenance energy requirements at our facility. For the entire study the average daily calorie intakes were 2364 kcal in the 1 meal/d diet and 2429 kcal in the 3 meals/d diet. More details on the diet composition and methods used to evaluate compliance with the diets are reported elsewhere [
28,
29].
As for the population of subjects examined in this study, the number of subjects (n) examined for each stage of the study are as follows: Pre-treatment/Baseline (n = 15), 1 meal/1 month (n = 8), 1 meal/2 month (n = 12), Off-diet (n = 12), 3 meals/1 month (n = 12) and 3 meal/2 month (n = 12). Complete data were analyzed and are presented for 15 subjects. In the TMD diet arm, 1 subject withdrew because of food dislikes. During the OMD, 5 subjects withdrew because of scheduling conflicts and health problems unrelated to the study. Only 1 of the 5 subjects withdrew specifically because of an unwillingness to consume the 1 meal/d diet.
Separation and stimulation of peripheral blood mononuclear cells (PBMCs)
The PBMCs were separated from fresh heparinized blood of healthy adult donors using Ficoll density gradient centrifugation, followed by extensive washing in phosphate-buffered saline (PBS). The erythrocytes were removed by hypotonic shock (ACK lysis buffer, Quality Biological, Bethesda, MD). The PBMCs were subsequently cultured in serum-free medium (AIM-V) and stimulated with either plate-bound anti-CD3 mAb (200 ng/ml) or
E. coli LPS (10 μg/ml) for 24 hours as described previously [
21].
Cytokine analysis
Serum and cell culture supernatants were analyzed for cytokines using Bio-Plex Cytokine 17-Plex Panel according to manufacturer's instructions (Biorad Laboratories, Hercules, CA).
Real Time PCR analysis
The PBMCs were lysed in RNA lysis buffer and total RNA was extracted from control and stimulated cells using a QIAshredder kit (QIAgen). RNA (500 μg) and oligo-dT primers were used to synthesize single-stranded cDNA. PCR was then performed using SYBR green Master Mix (Applied Biosystems, Foster City, California, USA), 1 μl cDNA, and exon spanning gene-specific primers. Thermal cycling was performed using the Applied Biosystems GeneAmp 7700 Sequence Detector.
Statistical Analysis
Data are presented as the mean and SEM. An analysis of variance appropriate for a 2 period crossover study with repeated measures within period was used to evaluate meal frequency effects on outcome variables. The Student-Newman-Keuls test was employed to test the significance of difference observed in the two study groups.
Discussion
Although there is ample anecdotal evidence of health benefits of fasting in the healthy adults, only recently have such dietary interventions been rigorously studied in laboratory animals and human subjects. Emerging evidence suggests that mice and rats maintained on repeated cycles of 24 hours with no food followed by 24 hours with free access to food (IF) lived up to 30% longer [
37,
38] compared to
ad libitum fed controls. In addition, the IF animals displayed improved insulin sensitivity, reduced cancers and increased resistance of neurons and cardiac cells to oxidative and metabolic stress [
17,
39]. However, the effects of meal frequency and intermittent fasting on immune function in humans are unknown. Given the highly robust beneficial effects of IF in experimental models, we designed the present study in humans to test the hypothesis that a change from a usual TMD diet to an OMD weight maintenance diet would alter proinflammatory cytokine expression in circulating lymphocytes. Analysis of serum levels of the proinflammatory markers, CRP, ICAM-1 and soluble gp130 revealed no significant differences between the OMD and TMD diet groups. We also studied serum visfatin, a novel adipocytokine, which is synthesized mainly by visceral adipocytes and serves as an insulin mimetic and proinflammatory cytokine [
31‐
33]. Similar to other proinflammmatory markers, we observed that OMD and TMD diets have no impact on serum visfatin levels in these subjects. Thus, meal frequency did not significantly affect levels of circulating pro-inflammatory markers, suggesting that a change in meal frequency does not alter the basal inflammatory state.
Because the subjects in our study were healthy and of normal weight, and so would not be expected to display elevated pro-inflammatory markers, we studied the impact of fasting on regulation of T-helper cytokines and chemokines from PBMCs isolated from the subjects at designated time points throughout the study. The isolated PBMCs were activated ex-vivo with LPS and TCR ligation. Compared to the baseline pretreatment values, we observed a large unexpected increase in cytokine secretion from stimulated cells one month after initiation of either the OMD or TMD diets. Although the precise mechanism responsible for this elevated cytokine release is unknown, it is quite feasible that departure from pre-study regular eating patterns and adherence to OMD and TMD controlled diets in both study groups made the PBMCs hyper-responsive to stimulation. As the study progressed into the second month of the controlled TMD and OMD diet periods, the stimulated cytokine levels in the subjects returned closer to the baseline suggesting a habituation and adaptation to the dietary intervention. A stress-based mechanism for the enhanced responsiveness of PBMC when subjects were on the TMD and OMD controlled diet would be consistent with previous studies have provided evidence that mild stress can enhance PBMC activation. For example, PBMCs isolated from subjects following exposure to psychosocial stress [
40] or 5 min of vigorous exercise [
41] exhibited increased activation of NF-κB, a transcription factor know to induce the production of various cytokines. In addition, chemotaxis and expression of cell adhesion molecules were increased in PBMCs isolated from subjects immediately after acute psychological stress [
42].
This meal frequency study was the first human study in which daily calories were held constant between two diets that differed only in meal frequency (3 smaller meals versus one large meal) [
28]. A large number of physiological variables were measured, including heart rate, body temperature and blood chemicals and many of these were unaffected by altering meal frequency. However, when on 1 meal per day, subjects did exhibit a significant reduction of fat mass and significant increases in levels of total, low-density lipoprotein, and high density lipoprotein cholesterol [
28]. In addition, the morning glucose tolerance was found to be impaired when subjects were consuming 1 meal per day compared with 3 meals per day [
29]. Fasting (morning) plasma glucose levels were also significantly elevated in subjects when they were consuming 1 meal per day compared with 3 meals per day and this 1-meal-per-day diet effect on glucose tolerance was rapidly reversed upon return to the 3-meals-per-day diet, indicating that the diet had no long-lasting effect on glucose metabolism. Besides these effects on glucose metabolism, there were no significant effects on the expression of a number of metabolic variables including plasma levels of ghrelin, adiponectin, resistin or BDNF. In the current study, we have further examined these subjects for alterations in inflammatory and immune parameters with changes in diet and noted that meal frequency changes do cause transient increases in TCR- and TLR4 (LPS)-mediated expression of several cytokines and that the magnitude of these alterations is less when subjects consume OMD versus TMD. Interesting patterns of expression were revealed where there appears to be more of a stress response at the initiation of the diets at the one month time interval. More pronounced cytokine expression changes were noted in the 1 month time period of the TMD subjects versus the same subjects given OMD supporting our conclusions that reduced meal frequency can have an impact on PBMC-derived cytokine expression between OMD and TMD subjects. Despite the observed changes in metabolic parameters reported in this study [
28,
29], we failed to note any significant correlations or associations between the observed cytokine changes in expression with diet and BMI or circulating levels of glucose, insulin, leptin, ghrelin, adiponectin, resistin or BDNF.
In the first month after the initiation of the diets, we observed a robust increase in IFN- γ and TNF-α release from PBMCs in subjects fed TMD that was significantly greater than when the subjects were on the OMD diet. In a smaller study, the immune cells derived from fed healthy subjects and stimulated ex vivo by TCR ligation produced significantly higher levels of IFN- γ with lower IL-4 levels compared to overnight fasted individuals [
43] suggesting that fasting promotes Th2 responses. In our study no significant differences were observed in Th2 cytokines from OMD and TMD groups; however similar to the previous study [
43], we observed lower levels of Th1 cytokines in subjects when they were on the OMD diet. We have also recently reported that alternate day calorie restriction in overweight adults with asthma results in marked decline in circulating TNF-α levels with improvement in pulmonary functions and measures of quality of life [
27]. Recent studies suggest that in addition to Th1 and Th2 there is a subset of IL-17 producing T helper cells (Th17) that are involved in various autoimmune inflammatory disorders [
36,
44]. Compared to the TMD diet, we observed significantly reduced IL-17 secretion from stimulated T cells derived from subjects during the OMD diet. It has been suggested that IL-23 is a key regulator of IL-17 production in T cells [
35,
36]. We observed that PBMCs from subjects on TMD for 2 months expressed 6-fold higher levels of IL-23 mRNA compared to the baseline level, with no difference in IL-17 receptor expression. Recent studies suggest that fasting induced down-regulation of leptin protects against autoimmune encephalomyelitis [
45,
46]. Our data suggest for the first time that fasting can also down regulate the IL-17 pathway in human T cells and hence could modify autoimmune processes. Interestingly, compared to the TMD diet group, we also observed a significant reduction in MCP-1 and MIP-1β production from stimulated PBMCs derived from OMD fed subjects suggesting a tight coupling of metabolic and immune systems.
In conclusion, our study demonstrates that upon specific challenges ex vivo, leukocytes cells derived from subjects on an OMD diet respond with lower pro-inflammatory cytokine production. The immune compartment appears to be exquisitely sensitive to the behavioral and metabolic cues, and the application of intermittent fasting as an approach for modifying immune function to improve health warrants further study.
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
All authors read and approved the final manuscript. VDD, HY, KSS, RSP performed the PBMC isolation, cultures, multiplex and ELISA assays and the PCR analysis in this manuscript. KSS, WVR, DJB, DLL, MPM and DDT played a role in the planning, design and performance of the clinical trial and in the evaluation of the data. VDD, MPM and DDT played a role in the writing and editing of the manuscript.