The impact of intermittent fasting versus ad libitum feeding during Salmonella typhimurium infection was evaluated in terms of duodenum IgA levels, bacterial clearance and intestinal and extra-intestinal infection susceptibility. Mice that were intermittently fasted for 12weeks or fed ad libitum were infected with S. typhimurium and assessed at 7 and 14 days post-infection. Next, we evaluated bacterial load in the faeces, Peyer's patches, spleen and liver by plate counting, as well as total and specific intestinal IgA and plasmatic corticosterone levels (by immunoenzymatic assay) and lamina propria IgA levels in plasma cells (by cytofluorometry). Polymeric immunoglobulin receptor, α- and J-chains, Pax-5 factor, pro-inflammatory cytokine (tumour necrosis factor-α and interferon-γ) and anti-inflammatory cytokine (transforming growth factor-β) mRNA levels were assessed in mucosal and liver samples (by real-time PCR). Compared with the infected ad libitum mice, the intermittently fasted infected animals had (1) lower intestinal and systemic bacterial loads; (2) higher SIgA and IgA plasma cell levels; (3) higher mRNA expression of most intestinal parameters; and (4) increased or decreased corticosterone levels on day 7 and 14 post-infection, respectively. No contribution of liver IgA was observed at the intestinal level. Apparently, the changes following metabolic stress induced by intermittent fasting during food deprivation days increased the resistance to S. typhimurium infection by triggering intestinal IgA production and presumably, pathogen elimination by phagocytic inflammatory cells.
Data from human and animal studies show that the two commonly used dietary restriction regimen models, caloric restriction and intermittent fasting, prolong lifespan by countering the deleterious conditions that accompany natural ageing, including cancer and metabolic, heart and neurodegenerative dysfunctions [1, 2].
Compared with ad libitum food intake, caloric restriction is a dietary regimen that involves a 20–50% reduction in energy input while maintaining adequate nutrient consumption to avoid malnutrition [1-3]. The intermittent fasting protocol involves ad libitum feeding and fasting every other day. There are two benefits of intermittent fasting that are not provided by caloric restriction: (1) body weight maintenance due to a compensatory mechanism of increased caloric intake on the days that food is available and  reduced serum glucose and insulin levels, which are accompanied by an increase in the resistance of neurons to excitotoxic stress .
Salmonella (S.) typhimurium, an invasive intracellular pathogen, causes both intestinal and extra-intestinal infection in mice, which resembles human typhoid fever induced by S. typhi . After oral ingestion, S. typhimurium passes through the stomach, reaches the intestine and colonizes the intestinal epithelium. Salmonella typhimurium eventually crosses this organ, usually via M cells that cover the luminal surfaces of Peyer's patches . At the subepithelial level, S. typhimurium is translocated to the liver and spleen by blood and/or lymphatic pathways [5, 7].
Studies primarily using the murine typhoid fever model have shown that dietary restriction regimens affect intestinal enteropathogenic bacteria (e.g. Salmonella typhimurium) infection outcomes, depending on whether the dietary restriction was applied in the short term or long term. A short-term caloric restriction regimen improved survival in an S. typhimurium infection mouse model , whereas a long-term regimen decreased resistance to S. typhimurium infection . These studies suggest that there is a fine line regarding the energy availability necessary for an optimal immune response to S. typhimurium infection, as was the case with respiratory infections . However, little information regarding the humoral and/or cellular immune response mechanisms involved was reported in these studies. In the absence of infection, a short-term caloric restriction appears to generally enhance the immune response, including increased natural killer cell and IgA levels [11, 12]. In contrast, long-term intermittent fasting downregulates IgA levels and increases or decreases polymeric immunoglobulin receptor (pIgR) levels in the duodenum or ileum, respectively [13, 14]. These data highlight the functional and structural differences that account for the regionalization of IgA and pIgR production, whereas in normal mice, both of these proteins are higher in the duodenum than the ileum .
Secretory IgA and pIgR are key components for mucosal surface protection [16, 17]. Both SIgA and pIgR contribute to protection against infections caused by enteropathogenic bacteria, such as S. typhimurium [18, 19].
To our knowledge, there are no reports regarding the influence of intermittent fasting and infection outcome or intestinal IgA production under S. typhimurium infection conditions. The present study aimed to evaluate the effects of intermittent fasting on intestinal IgA production after S. typhimurium infection in the duodenum and on resistance to sublethal S. typhimurium infections in mice.
Materials and methods
Eight-week-old male Balb/c mice obtained from Harlan Sprague Dawley, Inc., and housed 1 per cage in a room with little noise were fed ad libitum with the NIH-31 diet for the first 18 weeks of life to allow for adaptation to environmental conditions. Throughout the experiment, the animals were kept on a 12:12 h light/dark cycle (lights on at 6 am). Additionally, all handling and assays were carried out between 8 and 11 am to avoid variations in adrenocorticotropic hormone (ACTH) and corticosterone levels due to the circadian cycle. All animals were handled in accordance with the Mexican federal regulations for animal experimentation and care (NOM-062-ZOO-1999, Ministry of Agriculture, Mexico City, Mexico), and the experimental protocol was approved by the Institutional Animal Care and Use Committee and Institutional Ethics Committee.
A group of 24 mice was fed ad libitum until the 18th week of life and then randomly assigned to one of two groups (n = 12) with either ad libitum feeding or intermittent fasting (mice were fed every other day). Intermittent fasting was conducted from the beginning of the 19th to the end of the 30th week of life (for a total of 12 weeks). Food was always removed and replaced at 1 pm. Body weight and food intake were determined weekly on the same day of the week and at the same time of day. Energy intake was calculated based on the fact that the energy content of each gram of NIH-31 was 3 kcal.
On the first day of the 33rd week of life, the animals in both the ad libitum feeding and intermittent fasting groups were infected with a sublethal dose [104 colony forming units (CFU)] of S. typhimurium (ATCC-14028) by an intragastric route using a plastic tube . After infection, all mice were fed ad libitum to prevent mortality. Half of the animals in the ad libitum and intermittent fasting groups (n = 6) were sacrificed (through intraperitoneal administration of a lethal dose of sodium pentobarbital, 90 mg/kg) on day 7. The other half of the mice was sacrificed on day 14 post-infection. For all animals, food was removed at 7:00 pm on the day prior to sacrifice.
The experimental samples included freshly voided faecal pellets (collected and weighed each morning during the 7 or 14 day post-infection periods), as well as blood and organs (obtained following animal sacrifice on day 7 or 14 post-infection). As indicated earlier, the duodenums were analysed given that both IgA secretion and pIgR expression are higher in the duodenum than the ileum, as determined previously in normal mice .
Faecal pellets were used to estimate bacterial elimination using the plate count method . Fresh pellets were suspended in sterile phosphate-buffered saline (PBS) containing 0.05% Triton (Sigma Chemicals, Saint Louis, MO, USA). Additionally, the pellets were vortexed extensively to prepare a uniform faecal suspension.
Bacterial load was also determined using the plate count method in the liver, spleen and Peyer's patches of the duodenum . Briefly, after excising these organs from the sacrificed animals, the samples were weighed and squeezed through a plastic grid to obtain uniform tissue suspension. Aliquots of serial dilutions from each sample prepared in PBS-Triton were cultivated at 37 °C for 48 h on Salmonella Shigella (SS) agar (Difco Laboratories, Sparks, MD, USA), after which time the number of colonies was counted and expressed as CFU/g per sample.
Total and specific intestinal IgA protein determinations by ELISA
Duodenum fluid was analysed using the enzyme-linked immunosorbent assay (ELISA) to determine the levels of total IgA antibodies and the specific Salmonella surface protein IgA antibodies according to a previously reported method . Briefly, 5 ml of PBS containing 0.02% sodium azide was passed through the duodenum to collect the intestinal fluid. The washout material was centrifuged at 10,000 g for 20 min at 4 °C, and then, a protease inhibitor cocktail (Complete mini, Roche Diagnostics, Mannheim, Germany) was added. Intestinal fluid was stored at −70° C for future analyses.
The determination of IgA+ plasma cells
After washing, the duodenum was processed to allow IgA+ plasma cell detection in the lamina propria by immunohistochemistry. Plasma cells containing cytoplasmic IgA (IgA+ plasma cells) were quantified using a previously reported method . Briefly, one-centimetre duodenum sections were frozen and cut with a cryotome in 0.7 μm sections that were fixed with acetone for 20 min. Slides were hydrated with PBS, and the endogenous peroxidase was blocked with a PBS solution containing 0.3% H2O2 and 0.1% NaN3 for 10 min. After washing with 0.05% Tween 20 in PBS (PBS-T), the slides were blocked with 5% bovine serum for 30 min. After washing with PBS-T, the sections were incubated with a goat anti-mouse IgA peroxidase conjugate. After washing, the reaction was visualized following an H2O2 diaminobenzidine reaction. The samples were counterstained with Harris' haematoxylin and then dehydrated and covered with synthetic resin. The negative controls were incubated with a rabbit anti-goat IgG peroxidase conjugated. The IgA+ plasma cells were quantified in random fields at a 400× magnification factor using a 10 × 10 mm2 ocular graticule. The viewing fields were adjusted so that the number of IgA+ plasma cells per 5 mm2 of tissue could be quantified. Two independent observers used single-blinded procedure to analyse the tissue sections.
Q-PCR assays for mRNA expression determination
The duodenum sections were scraped with a coverslip to obtain mucosa samples. Additionally, to ensure that only the mucosa was removed, an immunohistochemistry assay was performed. Liver samples were obtained by surgical resections and were frozen at −70 °C for future RNA extraction. Total RNA extraction and cDNA synthesis were performed as previously described . Real-time quantitative polymerase chain reaction (Q-PCR) was used to quantify mRNA levels. Q-PCR reactions were performed with Taqman probes from the Mouse ProbeLibrary (Exiqon Roche, Mannheim, Germany) and primers that were specific for amplification of the α-chain, J-chain, pIgR, transforming growth factor (TGF)-β, interferon (IFN)-γ, tumor necrosis factor (TNF)-α and the Pax-5 transcriptional factor genes. The primers were designed using ProbeFinder Software, which is available online at the Assay Design Center (http://www.universalprobelibrary.com), to determine the primer sequence for each gene (except Pax-5) as previously reported . The primer sequences for Pax-5 were as follows: pax5-F: ACGCTGACAGGGATGGTG and pax5-R: GGGGAACCTCCAAGAATCAT. The PCR was performed in 20 μl of total volume in capillaries containing a 4 μl mix of LightCycler Taqman Master, 0.2 μl of uracil-DNA glycosylase (2 U/μl), 0.2 μl of primers (0.4 nm each), 4 μl of cDNA (500 ng) and 0.2 μl of probes (labelled with FAM). A water control was systematically included to rule out any possible contamination. The capillaries were run with a LightCycler 2.0 instrument. The initial denaturation and Taq activation was conducted at 95 °C for 10 min. Additionally, the samples were amplified with 45 repeated cycles, in which each cycle consisted of a 15 s denaturing step at 95 °C and a 35 s annealing/extending step at 60 °C. The real-time PCR assay included a standard curve, which included four serial dilution points for each gene. Furthermore, the samples were normalized with the endogenous control, 18S. The data were analysed using the ∆CT method to known relative mRNA levels, as described by Livak and Schmittgen, with LightCycler Software (Version 4.0) (Roche Diagnostics).
The serum corticosterone concentrations (expressed in ng/ml) were determined from blood samples using a commercially available corticosterone assay ELISA kit (Assay Designs, Ann Arbor, MI, USA, Cat. no 901–097) and a standard curve.
Data are presented as the mean ± SD of three independent assays (six mice per assay, 18 animals per group). The corticosterone levels were analysed using Student's t-test. Two-way anovas were performed to analyse differences between the two dietary regimens and the two post-infection periods. If a significant main effect or association was identified, the means of the respective groups were compared using the Holm–Sidak method. For all tests, P ≤ 0.05 was considered significant. All analyses were performed using the SigmaPlot statistical software for Windows (version 2.03, SPSS, Inc., San Jose, CA, USA).
Body weight, food and caloric intake assessments
No significant differences were found between the 19th and 30th weeks in terms of body weight (g/week), food consumption (g/week) or caloric intake (kcal/week) between the two dietary regimens (data not shown).
Intermittent fasting decreased intestinal and systemic bacterial loads in the infected mice
The bacterial count in faecal pellets was significantly lower (P < 0.001) in the intermittent fasting group than the ad libitum group (except on the first and 11th days, when the lower level in the former group was not significant; Fig. 1A). Colony forming S. typhimurium cells were not found in pellets of the corresponding uninfected paired-matched controls (data not shown).
Lower bacterial counts were also found in the Peyer's patches, livers and spleens of the intermittent fasting group compared with the ad libitum group on days 7 and 14 post-infection (Fig. 1B; P < 0.001). Additionally, in both dietary regimen groups, a notably higher colonization level was observed on day 14 compared with day 7 post-infection (P < 0.001) in all organs.
As expected, the bacterial count in these three organs was significantly higher in the infected mice compared with the uninfected mice on both measurement days and for both dietary regimens (P < 0.001). In the intermittent fasting group livers, no significant difference in bacterial count was observed between the infected (whether on day 7 or 14 post-infection) and uninfected animals.
Intermittent fasting modulated intestinal IgA and IgA+ plasma cell levels in the infected mice
The total intestinal IgA levels were increased at day 14 post-infection in the ad libitum group (Fig. 2A). In the intermittent fasting group, however, these levels were increased at 7 days post-infection period (P < 0.001) and higher in the infected intermittent fasting group than in the infected ad libitum group (P < 0.001).
The post-infection IgA levels that were specific to Salmonella followed a similar pattern as that observed for total IgA. The specific IgA levels were higher in the intermittent fasting group compared with the ad libitum group at the two post-infection days (Fig. 2B; P < 0.001) and higher in the infected animals than in the uninfected animals. As expected, unlike the total IgA results, the anti-Salmonella IgA levels were extremely low in the uninfected mice under both dietary regimens. Thus, in the ad libitum-fed mice, significantly higher anti-Salmonella IgA levels were observed on day 7 in the infected group than the uninfected group (P < 0.001).
Regarding the IgA+ plasma cells, although significantly higher levels were observed in the intermittent fasting group compared with the ad libitum group on day 7 post-infection (P < 0.001), the IgA+ plasma cell value for the intermittent fasting group was not significantly different from the ad libitum group on day 14 post-infection due to the relatively high variability within the ad libitum group on this day.
Additionally, the number of IgA+ plasma cells was significantly higher for the uninfected intermittent fasting and ad libitum groups than their pair-matched group on day 7 post-infection (P < 0.05 for intermittent fasting; P < 0.001 for ad libitum). The intermittent fasting group displayed a significantly higher IgA+ plasma cell count on day 14 post-infection compared with the uninfected group (Fig. 2C; P < 0.001).
A representative picture of the IgA plasma cell staining from the uninfected and infected mice of each feeding regimen is depicted in Fig. 2D.
Intermittent fasting modulates intestinal and hepatic IgA-associated protein mRNA expression in infected mice
In addition to the previous evaluation of the IgA protein levels, IgA-associated protein mRNA expression was also assessed. In the intestine, α-chain, J-chain and pIgR mRNA expression was higher in the intermittent fasting group than the ad libitum group among the uninfected (P < 0.05 for α-chain; P < 0.001 for J-chain and pIgR) and infected animals (P < 0.05 for α-chain [7 day] and pIgR; P < 0.001 for J-chain) (Fig. 3).
Within the ad libitum group, intestinal α-chain mRNA expression was significantly higher in the infected compared with the uninfected animals (P < 0.05 on day 7; P < 0.01 on day 14).
Within the ad libitum group, intestinal J-chain mRNA expression was significantly higher in the infected compared with the uninfected animals only on day 14 post-infection (P < 0.001). Within the intermittent fasting group, intestinal J-chain mRNA expression was significantly higher in the infected than the uninfected animals on days 7 and 14 post-infection (P < 0.001). Additionally, for both dietary regimens, intestinal J-chain mRNA expression was significantly higher on day 14 than on day 7 post-infection (P < 0.01 for the intermittent fasting group; P < 0.001 for the ad libitum feeding group; Fig. 3B).
Within the ad libitum group, intestinal pIgR mRNA expression was higher in the infected than uninfected animals (P < 0.05 on day 7; P < 0.01 on day 14). Within the intermittent fasting group, similar intestinal pIgR mRNA expression levels were observed between the infected and uninfected animals (Fig. 3C).
The hepatobiliary transfer of IgA to the duodenum was also investigated by determining the α-chain, J-chain and pIgR mRNA expression levels in the liver (Fig. 3D). In contrast to the levels observed in the intestine, the mRNA expression of these parameters was higher in the liver in almost all cases for the ad libitum group compared with the intermittent fasting group (for significance values, see the Fig. 3D legend).
Within the ad libitum group, α-chain, J-chain and pIgR mRNA levels were lower on day 7 post-infection compared with the uninfected animals (P < 0.001 for α-chain and J-chain; P < 0.05 for pIgR). Compared with the uninfected animals, on day 14 post-infection, mRNA expression in the infected mice was higher for α-chain (P < 0.001), lower for J-chain (P < 0.01) and did not significantly change for pIgR.
In contrast, no significant differences were found for any of these three parameters between the uninfected and infected animals in the intermittent fasting group (on both day 7 and 14).
Intermittent fasting modulates IgA- and pIgR-associated cytokine mRNA expression
mRNA expression of TGF-β and Pax-5 was assessed in association with IgA production. TGF-β mRNA is essential for class switch recombination in early B cell precursors committed to IgA synthesis . Pax-5 mRNA is only expressed in early B cells and encodes the B cell-specific activator protein (BSAP) transcriptional factor, which is crucial for B cell lineage commitment and development .
Intestinal TGF-β mRNA expression levels were significantly higher in the intermittent fasting groups than the ad libitum groups in both the uninfected (P < 0.001) and infected animals (P < 0.001). Within the intermittent fasting group, similar intestinal TGF-β mRNA levels were found in the uninfected and infected mice. Within the ad libitum group, the infected mice had higher intestinal TGF-β mRNA levels than the uninfected mice (P < 0.05 for day 7; P < 0.01 for day 14; Fig. 4A).
Intestinal Pax-5 mRNA expression levels were significantly higher in the intermittent fasting groups than the ad libitum groups in both the uninfected (P < 0.05) and infected animals (P < 0.001). Within the intermittent fasting group, higher intestinal Pax-5 mRNA expression levels were observed in the infected mice than the uninfected mice (P < 0.001 for day 7; P < 0.05 for day 14). Within the ad libitum group, intestinal Pax-5 mRNA expression level was lower in the infected animals than the uninfected animals (P < 0.001 for day 7 and 14; Fig. 4B).
In the liver (Fig. 4C), TGF-β and Pax-5 mRNA expression levels were higher in the ad libitum feeding group than in the intermittent fasting group in the uninfected mice (P < 0.001 for TGF-β; P < 0.01 for Pax-5). For the infected animals, the only significant difference between the two dietary regimens was higher TGF-β mRNA expression in the ad libitum group compared with the intermittent fasting group on day 14 post-infection (P < 0.001). Within the ad libitum group, significantly higher mRNA expression was observed for both TGF-β and Pax-5 compared with the uninfected animals on day 14 post-infection (P < 0.001 for TGF-β; P < 0.05 for Pax-5). Within the ad libitum group, higher Pax-5 mRNA expression levels were observed on day 14 compared with day 7 post-infection (P < 0.01).
We assessed IFN-γ and TNF-α transcriptional expression involved in pIgR expression . Intestinal IFN-γ mRNA expression was significantly higher in the intermittent fasting group than the ad libitum group for the uninfected (P < 0.01) and infected animals (P < 0.05, Fig. 5A) on day 7. Within the ad libitum group, intestinal IFN-γ mRNA expression levels were significantly higher in the infected mice than the uninfected mice (P < 0.001) on day 14. Within the intermittent fasting group, no differences were observed regarding intestinal IFN-γ mRNA expression levels between the uninfected and infected mice.
Intestinal TNF-α mRNA expression was significantly higher in the intermittent fasting group than the ad libitum group (P < 0.05) on day 14 post-infection. Within the intermittent fasting group, higher intestinal TNF-α mRNA expression levels were observed in the infected mice compared with the uninfected mice (P < 0.05; Fig. 5B) on day 14 post-infection.
In the liver, IFN-γ mRNA expression levels were higher in the ad libitum group than the intermittent fasting group (only on day 7 post-infection, P < 0.001) and in the uninfected intermittent fasting group compared with the ad libitum group (P < 0.001; Fig. 5C). Within the intermittent fasting group, liver mRNA IFN-γ levels were higher in the uninfected mice than the infected mice (P < 0.001). In contrast, within the ad libitum group, the infected mice exhibited higher liver IFN-γ mRNA levels than the uninfected mice (P < 0.001).
In the liver, TNF-α mRNA expression was higher in the ad libitum group than the intermittent fasting group (only on day 14 post-infection, P < 0.001) and in the uninfected intermittent fasting group compared with the ad libitum group (P < 0.001). Within the intermittent fasting group, liver TNF-α mRNA expression levels were higher in the uninfected mice than the infected mice (P < 0.001). Additionally, within the ad libitum group, liver TNF-α mRNA expression was higher in the uninfected mice (P < 0.001) on day 14 post-infection.
Under stress conditions that induce adrenal hormone production, IgA and pIgR levels can be modulated [21, 25]. Intermittent fasting, which is a type of stress, is known to trigger corticosterone release (the glucocorticoid adrenal hormone) into the bloodstream . Thus, serum corticosterone levels were assessed to explore the possible stress response elicited by intermittent fasting. In the uninfected animals, no significant differences were observed in serum corticosterone levels between the two dietary regimens (P = 0.08; Fig. 6). Corticosterone levels were significantly higher in the intermittent fasting group than the ad libitum group on day 7 post-infection (P < 0.001) and in the ad libitum group compared with the intermittent fasting group on day 14 post-infection (P < 0.05). Within the ad libitum group, significantly higher corticosterone levels were observed in the infected mice than the uninfected mice (P < 0.001) on day 14 post-infection. Within the intermittent fasting group, higher corticosterone levels were observed on day 7 post-infection than day 14 post-infection (P < 0.01).
An intermittent fasting protocol is known to lengthen lifespan in laboratory animals and provide health benefits [2, 3]. Unlike caloric restriction, the influence of intermittent fasting on S. typhimurium infection outcome has not been investigated. Previous studies have documented that the weight loss in mice caused by a prolonged caloric restriction regime contributes to decreased S. typhimurium infection resistance [8, 9]. In this study, prolonged intermittent fasting positively affected resistance against S. typhimurium infection. Intermittent fasting mice did not lose weight during the 12 weeks of this dietary regimen (prior to infection), which likely contributed to their bacterial infection resistance. The maintenance of body weight, as well as food and caloric intake, appears to be an effect of gorging when food is available . These results suggest that the positive effects of prolonged intermittent fasting on S. typhimurium infection resistance were mediated by periodic food deprivation stress.
The underlying mechanisms of the protective effects of intermittent fasting against S. typhimurium infection are likely related to the metabolic stress elicited by this dietary regimen. The beneficial effects of the stress caused by intermittent fasting could be mediated in part by a resetting of the circadian clock that established synchrony between metabolism and physiology . During the periods of food deprivation, intermittent fasting may ‘reprogram’ the release of neuropeptides, neurotransmitters and adrenal hormones, which in turn are involved in metabolic and physiological pathways that develop and traffic immunocompetent cells that may control S. typhimurium infection at intestinal and systemic levels. Alternatively, in addition to the beneficial effects of intermittent fasting on bacterial resistance, an IgA response may have been caused by stress signals that contribute to the bacterial flora composition . Indeed, stress is capable of modifying the normal intestinal flora composition, which in turn is able to modulate intestinal IgA production [29, 30].
In this study, we evaluated the effect of intermittent fasting on intestinal IgA production and resistance to intestinal and extra-intestinal S. typhimurium infection in mice. Intermittent fasting was associated with reduced bacterial load in the faeces and Peyer's patches and increased total and specific IgA antibody levels.
Intermittent fasting under infection was associated with the upregulation of IgA-associated parameters (e.g. α-chain, J-chain, TGF-β and Pax-5) at the transcriptional level. While α-chain mRNA encodes for IgA, J-chain transcripts are required for the J-chain synthesis necessary for the production of dimeric or polymeric IgA and for IgA transcytosis via pIgR [31, 32]. TGF-β mRNA is essential for class switch recombination in early B cell precursors committed to IgA synthesis . Pax-5 mRNA encodes for BSAP, which is expressed only in IgA+ B cell precursors and inhibits the transcription of α-chain and J-chain. In IgA plasma cells, Pax-mRNA transcription is inhibited, allowing α-chain and J-chain mRNA expression and thus IgA synthesis . In this study, co-expression of Pax-5 mRNA with α-chain and J-chain mRNA may have occurred because the samples analysed were whole mucosal samples that contained IgA+B lymphocytes and IgA plasma cells. Under S. typhimurium infection conditions, intermittent fasting also promoted pIgR transcription, which is necessary for IgA transcytosis, and IFN-γ and TNF-α mRNA expression, which are involved in pIgR transcriptional upregulation [23, 33].
Unlike with ad libitum feeding, intermittent fasting under normal conditions negatively influenced the IgA and plasma cells but upregulated α-chain and pIgR intestinal transcription, as has been reported previously . The up-modulatory effects of intermittent fasting in the absence of infection were also found in terms of J-chain, TGF-β, Pax-5 and IFN-γ transcription. The mechanisms that account for the downregulatory influence of intermittent fasting on IgA and IgA+ plasma cells are unknown but may be related to reduced endothelial homing molecule expression, such as mucosal addressin cellular adhesion molecule-1 (MadCAM-1), which is involved in lymphocyte migration from Peyer's patches to gut lamina propria. Previous studies have reported that the fasting of mice for 1 or 2 days downregulated MadCAM-1 mRNA expression in Peyer's patches, apparently due to a lack of signals triggered by enteral stimulation .
Transcriptional mRNA expression analysis of these parameters was also performed in the liver because, in mice, most proximal intestine IgA is derived from the hepatobiliary transport of IgA from the blood to the bile via pIgR transcytosis . The presumed contribution of this pathway to increased intestinal IgA levels during intermittent fasting was not observed because mRNA expression of the molecules involved in the generation of IgA and pIgR were similar or lower than the levels found with ad libitum feeding.
Intermittent fasting under infection decreased the hepatic mRNA expression of pro-inflammatory mediators that are essential for infection clearance (e.g. IFN-γ and TNF-α on day 14). This decrease occurred despite the lower and identical bacterial load in the spleens and livers, respectively, observed in the intermittent fasting mice compared with those fed ad libitum. The protective role of intermittent fasting may be linked with increased intestinal SIgA production. SIgA is known to promote bacterial clearance and concomitantly decrease bacterial translocation to systemic organs [18, 19]. Additionally, the pro-inflammatory effects of intermittent fasting on TNF-α and IFN-γ transcription at the intestinal level elicits S. typhimurium killing by phagocytes involved in the inflammatory response that may control bacterial translocation at the systemic level [36, 37].
In a previous study, intermittent fasting elicited a hypothalamus pituitary adrenal (HPA)-axis stress response, causing corticosterone release (the glucocorticoid adrenal hormone) into the blood stream . Similar to other reports, in this study, intermittent fasting did not affect corticosterone levels in the absence of infection . However, under S. typhimurium infection conditions, intermittent fasting either positively or negatively modulated corticosterone generation via the HPA axis. These results may suggest that corticosterone did not influence SIgA generation in terms of bacterial clearance. Alternatively, increased corticosterone levels may negatively or positively influence SIgA generation, as demonstrated in mice under repeated restraint stress [21, 25]. The underlying mechanism for the presumed dual regulation of corticosterone on intestinal IgA levels is not fully known; however, downregulation may be the result of intestinal immunocompetent cell apoptosis . Additionally, induced intestinal IgA may be attributed to enhanced pIgR transcription for IgA transcytosis .
The limitation of the present study is the inability to investigate the role of passive IgA transference to the intestine in the presence of a lethal S. typhimurium dose, which was not feasible due to the high susceptibility of Balb/c mice to this pathogen . Another limitation of the present study is that some parameters were only analysed at the transcriptional (not protein) level and not in the spleen, one of the main systemic targets of S. typhimurium.
In conclusion, metabolic stress caused by intermittent fasting during food deprivation favoured bacterial clearance at the intestinal level and was associated with increased intestinal IgA and bacterial resistance at the systemic level, presumably due to the increased IgA and inflammatory responses. Intermittent fasting up- and downregulated corticosterone levels, which may influence the production of IgA necessary for bacterial eradication.
We thank Teresita Rocío Cruz-Hernández for technical assistance. This work was supported by SIP-IPN and COFAA-IPN. R. Campos-Rodríguez, E. Lara-Padilla, J. Pacheco-Yepez and A. Jarillo-Luna were COFAA and EDI-IPN fellows.