ORIGINAL RESEARCH published: 18 October 2019 doi: 10.3389/fimmu.2019.02464 Influence of Gut Microbiota on Progression to Tuberculosis Generated by High Fat Diet-Induced Obesity in C3HeB/FeJ Mice Lilibeth Arias 1,2,3 , Galo Adrián Goig 4 , Paula Cardona 1,2,3 , Manuela Torres-Puente 4 , Jorge Díaz 1 , Yaiza Rosales 1 , Eric Garcia 1 , Gustavo Tapia 5 , Iñaki Comas 4,6 , Cristina Vilaplana 1,2,3 and Pere-Joan Cardona 1,2,3* 1 Experimental Tuberculosis Unit (UTE), Fundació Institut Germans Trias i Pujol (IGTP), Badalona, Spain, 2 Centro de Investigación Biomédica en Red de Enfermedades Respiratorias (CIBERES), Madrid, Spain, 3 Department of Genetics and Microbiology, Universitat Autònoma de Barcelona (UAB), Badalona, Spain, 4 Tuberculosis Genomics Unit (TGU), Instituto de Biomedicina de Valencia (IBV-CSIC), Valencia, Spain, 5 Pathology Department, Hospital Universitari Germans Trias i Pujol (HUGTIP), Universitat Autònoma de Barcelona (UAB), Badalona, Spain, 6 Centro de Investigación Biomédica en Red de Epidemiologia y Salud Pública (CIBERESP), Madrid, Spain Edited by: Wilson Savino, Oswaldo Cruz Foundation The administration of a high fat content diet is an accelerating factor for metabolic (Fiocruz), Brazil syndrome, impaired glucose tolerance, and early type 2 diabetes. The present study Reviewed by: aims to assess the impact of a high fat diet on tuberculosis progression and microbiota Yaqing Qie, University of Texas MD Anderson composition in an experimental animal model using a C3HeB/FeJ mouse strain Cancer Center, United States submitted to single or multiple consecutive aerosol infections. These models allowed us Zijian Zhang, Baylor College of Medicine, to study the protection induced by Bacillus Calmette-Guérin vaccination as well as by the United States natural immunity induced by chemotherapy after a low dose Mycobacterium tuberculosis *Correspondence: infection. Our results show that a high fat diet is able to trigger a pro-inflammatory Pere-Joan Cardona response, which results in a faster progression toward active tuberculosis and an pj.cardona@gmail.com impaired protective effect of BCG vaccination, which is not the case for natural immunity. Specialty section: This may be related to dysbiosis and a reduction in the Firmicutes/Bacteroidetes ratio in This article was submitted to the gut microbiota caused by a decrease in the abundance of the Porphyromonadaceae Nutritional Immunology, a section of the journal family and, in particular, the Barnesiella genus. It should also be noted that a high fat diet Frontiers in Immunology is also related to an increase in the genera Alistipes, Parasuterella, Mucispirillum, and Received: 11 July 2019 Akkermansia, which have previously been related to dysbiotic processes. As diabetes Accepted: 02 October 2019 mellitus type 2 is a risk factor for developing tuberculosis, these findings may prove useful Published: 18 October 2019 in the search for new prophylactic strategies for this population subset. Citation: Arias L, Goig GA, Cardona P, Keywords: tuberculosis, high fat diet, obesity, comorbidity, BCG, mice, gut microbiota, C3HeB/FeJ Torres-Puente M, Díaz J, Rosales Y, Garcia E, Tapia G, Comas I, Vilaplana C and Cardona P-J (2019) Influence of Gut Microbiota on INTRODUCTION Progression to Tuberculosis Generated by High Fat Diet-Induced Tuberculosis (TB) is still a major challenge facing humanity in the twenty-first century. Since the Obesity in C3HeB/FeJ Mice. declaration by the World Health Organization of a global emergency 25 years ago, the absolute Front. Immunol. 10:2464. number of cases seems to have stabilized. However, the reduction in incidence of around 1.5% is doi: 10.3389/fimmu.2019.02464 far from the value of 25% needed to achieve the 2025 objective (1). Frontiers in Immunology | www.frontiersin.org 1 October 2019 | Volume 10 | Article 2464 Arias et al. Microbiota High Fat Diet and Tuberculosis Several factors, including severe immunosuppression (caused impact of standard treatment (HRZE) in TB patients has also by HIV, immunosuppressive drugs or anti-TNF antibodies), been assessed, demonstrating a long-lasting microbiota dysbiosis alcoholism and tobacco smoking, have been reported to increase with a variation in several species of the Firmicutes and the probability of developing TB (2). One of the most important Bacteroidetes phyla (17). Interestingly, both studies showed an factors is malnutrition (3, 4). A reduction in body mass index increased abundance of the Erysipelotrichaceae family, which, (BMI) to below 18 has been shown to be a strong risk factor in contrast to other Firmicutes, has been linked to a pro- for TB (5). In several infectious diseases, such as community- inflammatory milieu (18). Curiously, a recent study in a human acquired pneumonia, sepsis, Chagas disease, or TB, there are cohort has shown an increase in the gut microbiota diversity in consistent epidemiological data showing the advantages of TB patients. This is an interesting finding given that the BMI of obesity, with negative outcomes being inversely related to BMI. TB patients was significantly lower than for healthy controls (19). This benefit can also be seen for a variety of surgical and non- The first objective of this study was to evaluate the role of surgical conditions (6, 7). This means that subjects with better HFD-induced obesity in the development of TB in an active nutrition, more body fat and highly activated immune functions TB murine model using the C3HeB/FeJ mouse strain (20) by are more likely to survive. Metabolic syndrome intensifies with assessing the bacillary load (BL), pathology and cytokine profiles an increase in adiposity, thereby raising the level of immune in lungs, survival rates, and gut microbiota content. We also protection, which at the same time is counterbalanced by the evaluated the effect of multiple consecutive infections vs. a deleterious effects of dysglycemia. This process is known as the single aerosol infection. This is a better way of mimicking the “obesity paradox” (6). scenario of household contacts, where subjects are at high risk In contrast, it is well-known that diabetes mellitus type 2 of developing TB due to their daily contact with a TB patient (DM2) is also a risk factor for TB (8), although this is also and those persons that live in regions with a high infection risk linked to obesity and, therefore, an increased BMI. This new link (21). In this complex scenario, we tested the influence of Bacillus is specially worrying considering the overwhelming increase in Calmette-Guérin (BCG) vaccination and natural infection on DM2 incidence in high TB burden countries, such as India and disease progression. South Africa (9, 10). Obesity, insulin resistance and DM2 are As there is a lack of animal studies regarding the mechanisms closely associated with a chronic state of low-grade inflammation behind the DM2-TB comorbidity, which are key to better (11). In particular, in obesity there is an over-expression of understanding the pathology and improving the management TNF-α, which attracts M1 macrophages, thereby increasing the of the disease, we believe this research represents a significant levels of inflammatory mediators such as C-reactive protein, contribution to the field by providing an animal model for testing IL-1, IL-6, IL-8, IL-12, and IFN-γ. In this sense, DM2 also TB vaccines in a context of obesity and early DM2, an issue results in an increase in neutrophils and monocyte-secreted closely related to current society. cytokines, including TNF-α, IL-6, and IL-8. In addition, there is an impairment in infection-induced cytokine production and a MATERIALS AND METHODS reduction in phagocytic and antibacterial activity. DM2 decreases the level of regulatory T cells (Tregs) while increasing Th17, thus Experimental Design promoting an exaggerated inflammatory response (12). Two different experiments were planned in order to evaluate: (a) A high fat diet (HFD) has been linked to an increased the effect of HFD-induced obesity on the mouse model of active diversity in the intestinal microbiota, thus providing an anti- TB using the C3HeB/FeJ strain (comorbidity model), as well as inflammatory milieu (13), and a decrease in the Firmicutes to BCG vaccination and two different types of challenge [single Bacteroidetes (F/B) ratio, which has been related to glucose infection or multiple consecutive infections (SI or MCI)]; and intolerance (14). In an experimental model of TB using BALB/c (b) the effect of BCG vaccination compared to natural immunity mice, infection induced an initial decrease in microbiota (NI) in the same comorbidity model. diversity, which tended to recover with chronification. This Table 1 shows the experimental design of both experiments work highlights the significant decrease in Clostridiales (phylum and Table 2 the experimental groups and experimental Firmicutes), as promoters of Tregs, and Bacteroidales (phylum conditions used. Bacteroidetes), as inducers of IL-10-producing T cells and In the first experiment (n = 240), four groups were modulators of the Th1/Th2 balance (15). Recent studies of subcutaneously vaccinated with 106 colony-forming units (CFU) an experimental model in C57BL/6J-CD45a mice infected of BCG Live U.S.P vaccine (SII-ONCO-BCG R ); the four with Mycobacterium tuberculosis (Mtb) demonstrated the rapid remaining groups were sham vaccinated. After 12 weeks, mice dysbiosis caused by conventional TB chemotherapy, with this were challenged with a low dose aerosol to deliver around 50 condition persisting for a long time after cessation (16). The CFU of M. tuberculosis H37Rv Pasteur strain to the lungs using an airborne infection apparatus (Glas-col Inc., Terre Haute, IN, Abbreviations: TB, Tuberculosis; BMI, Body mass index; DM2, Diabetes USA). In the case of MCI, this challenge was done by infecting mellitus type 2; Tregs, Regulatory T cells; HFD, High fat diet; F/B, Firmicutes animals eight times over a period of 5 days. Animals were to Bacteroidetes; Mtb, Mycobacterium tuberculosis; BL, Bacillary load; BCG, euthanized at week 3 (n = 5 per group), week 4 [n = 5 per group, Bacillus Calmette-Guérin; SI, Single infection; MCI, Multiple consecutive infections; NI, Natural immunity; CFU, Colony-forming unit; ND, Normal except for group normal diet (ND) and MCI with n = 4] and week diet; p.i., Post-infection; OTU, Operational taxonomic unit; NMDS, Non-metric 16 (n = 10 per group) post-infection (p.i.). BL and pathology multidimensional scaling; MST, Median survival time. in lungs were analyzed. Lung samples at week 4 were also used Frontiers in Immunology | www.frontiersin.org 2 October 2019 | Volume 10 | Article 2464 Arias et al. Microbiota High Fat Diet and Tuberculosis TABLE 1 | Experimental design of both experiments performed in the study. Experimental design Experiment 1 Time of infection –w12 w0 w3 w4 w16 w16→ Time of diet w0 w12 w15 w16 w28 w28→ BCG Infection Euthanasia (n = 5) Euthanasia (n = 5) Euthanasia (n = 10) Survival (n = 10) Experiment 2 Time of infection –w12 –w10 to w0 w0 w0→ Time of diet w0 w2 to w10 w12 w12→ Single infection Antibiotic treatment Infection Survival (n = 10) Bold values correspond to weeks post and pre-infection. –w12: 12 weeks before challenge; w3, w4, w16: 3, 4, and 16 weeks post-challenge. TABLE 2 | Experimental groups and conditions tested per experiment. Experiment 1 Experiment 2 Group Experimental conditions Group Experimental conditions Obesity BCG Type of infection Obesity Protection Type of infection ND-SI No No SI ND-NI NO NI MCI ND-MCI No No MCI HFD-NI YES NI MCI ND-BCG-SI No Yes SI ND-BCG-MCI No Yes MCI HFD-SI Yes No SI HFD-MCI Yes No MCI HFD-BCG-SI Yes Yes SI HFD-BCG-MCI Yes Yes MCI to evaluate immune responses. Fecal samples (n = 5 per group) USA), and all procedures were conducted in a BSL-3 facility. were obtained at weeks 4 and 16 in order to analyze microbiota Animals were maintained on a 12 h light-dark cycle in a composition. The 10 remaining animals per group were allocated temperature- and humidity-controlled room. Animals from non- to assess impact on survival. obese experimental groups were fed with ND containing 13% In the second experiment, the protective effect of BCG calories from fat (2014S Teklad Global 14% Protein Rodent vaccination in MCI mice obtained from experiment 1 was Maintenance Diet, Envigo). Animals from those experimental compared to NI in MCI animals (n = 30). NI was used groups in which obesity was assessed were fed with an HFD to evaluate the memory-immune response protection after M. containing 60.3% calories from fat [TD.06414 Adjusted Calories tuberculosis challenge. To evaluate NI, animals were challenged Diet (60/Fat), Envigo]. Animal weight was recorded weekly. with a SI of low dose aerosol and animals were left 2 weeks to develop immunity. Then, animals were treated with isoniazid Ethics and rifapentine (25 and 10 mg/kg, respectively) for 10 weeks All procedures were performed according to protocol from week 2 p.i., to sterilize lungs. In order to test whether DMAH6119, which was reviewed by the Animal undetectable BL was achieved, three animals of each group were Experimentation Ethics Committee of the Hospital Universitari euthanized and lung samples were plated. No CFU counts were Germans Trias i Pujol (registered as B9900005) and approved by detected after 21 days of incubation period (limit of detection the Departament d’Agricultura, Ramaderia, Pesca, Alimentació = 10 CFU) (data not shown) as described in previous works i Medi Natural of the Catalan Regional Government, according (22–24). Two weeks after stopping antibiotic treatment, animals to current national and European Union legislation regarding were challenged with MCI. Fecal samples (n = 5 per group) were the protection of experimental animals. Mice were supervised obtained at week 4 in order to analyze microbiota composition. daily following a strict monitoring protocol in order to ensure Differences in protection were assessed by survival (n = 10 animal welfare, and euthanized, if required, with isoflurane per group). (inhalation excess). Animals Oral Glucose Tolerance Test Female C3HeB/FeJ specific-pathogen-free mice (6–8 weeks old) Oral glucose tolerance test was performed 11 weeks after diet were obtained from Jackson Laboratories (Bar Harbor, Maine, feeding. After overnight fasting period, mice were administered Frontiers in Immunology | www.frontiersin.org 3 October 2019 | Volume 10 | Article 2464 Arias et al. Microbiota High Fat Diet and Tuberculosis with 2 g/kg by oral gavage. Blood samples were taken at 0, 15, 0.3 µM primers, 0.5 U KAPA HiFi HotStart DNA Polymerase, 30, 60, and 120 min and glucose levels were measured with a and 15 ng DNA template. The PCR program with Mastercycler glucometer (Accutrend Plus, Roche Diagnostics, Switzerland). pro (Eppendorf, Hamburg, Germany) consisted of the following steps: initial denaturation at 95◦ C for 3 min, 18 cycles at 95◦ C Bacillary Load for 30 s, 55◦ C for 30 s, and 72◦ C for 30 s, and a final extension Left lung samples from each animal were collected, homogenized at 72◦ C for 4 min. The expected product size was ∼550 bp. and several dilutions plated on nutrient Middlebrook 7H11 agar The PCR products were purified using Agencourt AMPure XP (BD Diagnostics, Spark, USA). Visible CFU were counted after magnetic beads (Beckman Coulter, Brea, CA) using a Magnetic incubation for 28 days at 37◦ C. Stand-96 (Thermo Fisher Scientific). The PCR products were Lung Pathology quality controlled with a 1.4% Agarose Gel. The Index PCR Right lower lung lobe samples were fixed in 10% buffered was performed using Nextera XT DNA Library Preparation formalin, embedded in paraffin and 5 µm sections stained with Kit (Illumina) and KAPA HiFi HotStart PCR Kit as follows: hematoxylin-eosin for microscopic observation and analysis of 1X KAPA HiFi Fidelity Buffer, 0.5 mM MgCl2 , 0.3 mM KAPA the damaged area using the NISElements D version 3.0× software dNTP Mix, 0.3 µM primers, 0.5 U KAPA HiFi HotStart DNA package (Nikon Instruments Inc., Tokyo, Japan). Four recuts of Polymerase, and 5 µL of Nextera XT Index 1 Primers and a block containing all group samples were used to determine the Nextera XT Index 2 Primers per sample. The procedure was damaged area as a percentage of total lung area. Moreover, the performed as described in the 16S Metagenomic Sequencing proportion of exudative lesions in the total damaged area was Library Preparation guide. Library products were purified as measured to obtain the relative exudative area. PCR products. The final DNA concentrations of the purified products were measured using a Qubit 3.0 fluorimeter (Thermo Survival Fisher Scientific) and validated with the Agilent 2100 Bioanalyzer The effect on survival was evaluated in 10 animals from each (Agilent Technologies, Santa Clara, CA, USA). The purified group. Animals were euthanized according to the humanized products were diluted to a final concentration of 4 nM and end-point protocol approved by the ethics committee. pooled. The pool was denatured and sample loaded, as per the 16S Metagenomic Sequencing Library Preparation guide, at a Cytokine Profile Analysis final concentration of 7 p.m. The 16S rRNA gene libraries were A cytokine profile study was performed in lung homogenates sequenced with 2 × 300 paired-end reads using the Illumina from right upper and middle lobes. The following cytokines MySeq system (Illumina). were measured by Luminex xMAP R technology: IFN-γ, TNF-α, transforming growth factor beta (TGF-β), IL-2, IL-6, IL-10, IL12, Statistics, Sequence Processing and Bioinformatics IL-17, CXCL1, and CXCL5. Results are expressed as pg per ml Data Analysis of supernatant. The assay was performed with the MILLIPLEX R The non-parametric Mann–Whitney test was used to make MAP kit (EMD Millipore Corporation, Billerica, MA, USA) comparisons between groups using Graphpad Prism (GraphPad following the manufacturer’s instructions and analyzed with software v6.0, La Jolla, California, USA). Survival curves were xPONENT Software (Luminex Corporation, Austin, TX, USA). created using the product limit method of Kaplan and Meier Microbiota Analysis and were compared using the log-rank test. Body weight increase Sample Processing and DNA Extraction slopes were compared using linear regression test. Statistically Cecum from the animals was extracted and stored at −80◦ C significant differences are designated as follows: ∗ p < 0.05; ∗∗ p until DNA extraction. Samples were thawed on ice and stools < 0.01; ∗∗∗ p < 0.001; ∗∗∗∗ p < 0.0001. separated from the intestinal epithelia. DNA was extracted using a QIAamp Fast DNA Stool Mini Kit (QIAGEN GmbH, Germany) Immune Response Analysis according to the manufacturer’s instructions. Three 4 mm glass A three-way ANOVA was used to assess the effect of diet, beads were added in the first step before vortexing to enhance infection and BCG vaccination on the immune response in cell lysis. The total volume of eluate was stored at −20◦ C until lungs. A logarithmic transformation was applied to normalize the concentration measurement. DNA concentrations of the extracts cytokine concentrations. In order to obtain equally sized samples, were measured fluorometrically using the Qubit dsDNA HS assay the missing value in the ND-MCI group was replaced by the kit (Thermo Fisher Scientifics, Waltham, MA, USA). DNA was mean for the group. Holm–Sidak multiple comparisons were stored at −20◦ C until 16s rRNA gene library preparation. performed using the means of groups differing in only one factor. Plotting and statistical analysis were carried out using Graphpad 16S rRNA Gene Sequencing Prism (GraphPad software v7.0, La Jolla, California, USA). Microbial DNA was analyzed using the Illumina 16S The immune response in lungs was also analyzed by Principal Metagenomic Sequencing Library Preparation guide, with Component Analysis (PCA) using RStudio (version 1.1.463). some modifications. The protocol targets the V3–V4 regions of PC1 and PC2 represent the two principal components that the 16S rRNA gene, which were amplified using the KAPA HiFi contribute to most of the variance. The direction of the arrows HotStart PCR Kit (KAPA Byosystems, Wilmington, MA, USA). and their lengths represent the contribution and strength of each The composition of a typical PCR was as follows: 1X KAPA principal component. Cytokine concentration levels were scaled HiFi Fidelity Buffer, 0.5 mM MgCl2 , 0.3 mM KAPA dNTP Mix, to unit variance. Frontiers in Immunology | www.frontiersin.org 4 October 2019 | Volume 10 | Article 2464 Arias et al. Microbiota High Fat Diet and Tuberculosis FIGURE 1 | Weight evolution during the experiment in pre- and post-infection status. Weight evolution of the same experiment is represented by comparing ND and HFD (A–F) and Sham and BCG (G–L). Panels represent: pre-infection and sham (A), SI and sham (B), MCI and sham (C), pre-infection and BCG (D), SI and BCG (E), MCI and BCG (F), pre-infection and ND (G), SI and ND (H), MCI and ND (I), pre-infection and HFD (J), SI and HFD (K), MCI and HFD (L). Linear regression are represented together with p-values. Slopes’ values are included in Table S1. Frontiers in Immunology | www.frontiersin.org 5 October 2019 | Volume 10 | Article 2464 Arias et al. Microbiota High Fat Diet and Tuberculosis Bacterial Diversity Analysis (version 1.0.143). Diversity was quantified using the Bray–Curtis The resulting read files were analyzed using mothur pipeline dissimilarity test and plotted using non-metric multidimensional (version 1.39.5). Paired-end reads were joined and quality scaling (NMDS). Beta diversity (Shannon’s, Simpson’s, and controlled. Sequences were aligned using the SILVA reference Inverse Simpson’s index) was calculated to analyze differences in database. Those alignments that did not fit exactly with our diversity between samples. region of interest were removed. Sequences were pre-clustered and chimeric sequences were identified and removed, then RESULTS taxonomically classified using RDP (ribosomal database project). Sequences classified as chloroplasts, mitochondria, archaea, and HFD Determines the Spectrum of Body eukaryote were eliminated. The distances between sequences Weight Before and After Infection were calculated with a distance cut-off of 0.03. After creating The animals’ weight was recorded weekly throughout the study. the distance matrix, sequences with >97% similarity were HFD was found to significantly increase body weight at 2 weeks clustered into operational taxonomic units (OTUs). These OTUs after starting the diet. Before Mtb challenge there was a clear were filtered and only those that had more than 0.1% relative difference in the slope of the body weight increase between ND abundance and were in at least two samples were conserved. and HFD (Figures 1A–F; Table S1; Figure S1). However, this Each OTU was assigned to a taxonomic classification. OTUs difference in the growth slope disappeared suddenly after the were classified into the last taxonomic rank possible, with challenge for both SI (Figures 1B,C) and MCI (Figures 1E,F), genus being the furthest rank allowed by the method used. with HFD exhibiting a sudden deceleration. BCG vaccination Rarefaction curves, diversity and richness indexes and distances significantly increased body weight in the HFD group just 6 between samples were calculated and plotted using RStudio weeks after starting the diet (Figures 1G–L; Figure S1). Animals FIGURE 2 | Bacillary load progression at different end time points (w3, w4, and w16) are shown as log CFUs/ml. Each panel compares sham and BCG vaccinated groups: ND and SI (A), ND and MCI (B), HFD and SI (C), HFD and MCI (D). Box and whiskers plots show the minimum, first quartile, median, third quartile and maximum values. *p < 0.05, **p < 0.01, ***p < 0.001; Mann–Whitney test. Frontiers in Immunology | www.frontiersin.org 6 October 2019 | Volume 10 | Article 2464 Arias et al. Microbiota High Fat Diet and Tuberculosis also showed glucose intolerance and increased cholesterol and if fed with ND. However, the opposite was observed in HFD- HDL (Figure S2). fed animals. In contrast, under MCI and ND conditions, vaccination resulted in larger damaged areas when compared BCG Generates Lower Protection in Mice to control animals. Under MCI and HFD conditions, although vaccinated mice exhibited more infiltration, this difference was Fed With HFD in Terms of BL in Lung not statistically significant as the sham group also presented a Tissue large damaged area. In the chronic phase, BCG induced less The BL in lungs was evaluated at weeks 3, 4, and 16 p.i. No infiltration than sham-vaccinated mice under all conditions. statistically significant differences were found in terms of BL Lesions were classified as proliferative or exudative depending when comparing diets, although HFD-BCG animals tended to on their quality. Exudative lesions were present in all sham have higher medians (Figure S3). The type of challenge is a groups and were predominant after SI. Most of the lesions determinant factor in BL, and was found to be higher in almost in BCG-vaccinated animals were proliferative, with exudative all MCI groups (Figure S4). lesions being present only in ND-MCI mice. Amongst the sham As expected, BCG decreased BL by ∼1 log in the groups fed groups, HFD exhibited a higher percentage of exudative lesions with ND. This effect was maintained over time (Figures 2A,B). than ND (Figure 4). Given the absence of exudative lesions in the This improvement was not observed in HFD groups for either the acute phase, this measurement was only performed at week 16. SI or MCI challenge (Figures 2C,D). It should be noted, however, that BCG vaccination was administered at the same time as HFD, before having differences in weight between groups. Cytokine Profile Analysis Pro-inflammatory cytokines (IFN-γ, IL-2, IL-17, IL-12, IL-6, and TNF-α), neutrophil recruitment chemokines (CXCL1 and Early Infiltration Induced by BCG Does Not CXCL5) and anti-inflammatory cytokines (IL-10 and TGF-β) Preclude a Long-Term Reduction in were analyzed at week 4 p.i. in lung homogenates. Pathology Under HFD Table 3 shows the p-value summary of the three-way ANOVA Analysis of the histometry revealed different trends in the acute results. Infection was the factor responsible for the largest (weeks 3 and 4) and chronic (week 16) phases of infection variation between cytokines and chemokines (significant in all (Figure 3). except IL-2 and TGF-β). Similarly, IL-2 and IL-12 levels were In the early stages of infection, and after SI, BCG-vaccinated significantly affected by BCG vaccination. Thus, the two dietary animals exhibited less infiltration than sham-vaccinated animals conditions alone had no significant effect on the cytokine profile. FIGURE 3 | Progression of lung disease at different end time points (w3, w4, and w16) is shown as the percentage of damaged area. Each panel compares sham and BCG vaccinated groups: ND and SI (A), ND and MCI (B), HFD and SI (C), HFD and MCI (D). Bar plots show median with range. *p < 0.05; Mann–Whitney test. Frontiers in Immunology | www.frontiersin.org 7 October 2019 | Volume 10 | Article 2464 Arias et al. Microbiota High Fat Diet and Tuberculosis FIGURE 4 | Quality of the lesions found at week 16 post-infection. The tissue sections show a proliferative (A) and an exudative (B) lesion stained with H/E. (C) Relative percentage of exudative lesions at week 16. Black bars represent ND groups while red bars represent HFD groups. Hatched bars in each type of diet represent vaccination Bar plots show median with range. Asterisk indicates differences between vaccination and infection in each diet; hashtag indicates differences between diets. *p < 0.05, # p < 0.05; Mann–Whitney test. TABLE 3 | P-value summary of three-way ANOVA from cytokines quantified in lungs (*p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001). Source of variation IFN-γ IL-2 IL-6 IL-10 IL-12 CXCL5 IL-17 CXCL1 TNF-α TGF-β Infection **** ns **** ** **** * **** **** **** ns Diet ns ns ns ns ns ns ns ns ns ns BCG ns * ns ns * ns ns ns ns ns Infection × Diet *** ns ns ns ns ** * * ** ns Infection × BCG ** ns ns ns ns ns ** * ns ns Diet × BCG * **** ns ns ns ns ns * ** ns Infection × Diet × BCG ns ** ns ns ns ns ns ns ns ns Some of the cytokines analyzed presented an interaction between A PCA was performed for all inflammatory mediators different factors. For more detailed information, each individual (Figure 5). The first two components explain 71.51% of the value is illustrated in Figure S5. Briefly, pro-inflammatory variation among samples, with the main contributions to cytokines were present in higher concentrations in MCI animals. PC1 coming from CXCL1, IFN-γ, TNF-α, IL-17, and IL-6. In the case of IFN-γ, TNF-α, and IL-17, this increase was higher PC2, in contrast, is mostly explained by CXCL5 and TGF- in mice fed with HFD. In fact, among SI, ND mice had higher β. The unvaccinated HFD-MCI animals presented higher PC1 levels of these immune mediators, with the opposite being found values, whereas HFD-SI animals had higher levels of PC2. As for MCI. The same pattern was observed for CXCL1 and CXCL5. concluded from the three-way ANOVA analysis, the type of The levels of anti-inflammatory IL-10 or TGF-β did not differ infection was the factor with the greatest influence on the inter- significantly between the experimental groups. sample variation. Frontiers in Immunology | www.frontiersin.org 8 October 2019 | Volume 10 | Article 2464 Arias et al. Microbiota High Fat Diet and Tuberculosis FIGURE 5 | Principal component analysis of inflammatory mediators. (A) Representation of animals according to experimental groups with 95% confidence ellipses. (B) Samples colored by type of infection. BCG Vaccination Does Not Increase Once the dissimilarity between diets had been confirmed, Survival Under HFD, Whereas NI Does Shannon, Simpson, and Inverse Simpson indexes were used Differences in survival rates were analyzed using the Log-rank to demonstrate the existence of real differences in microbiota test (Figure 6). Irrespective of their vaccination status, MCI mice diversity between the different nutritional statuses. These indexes had lower survival rates than their SI counterparts, and HFD showed that ND mice had a statistically significantly higher always presented a lower median survival time (MST). BCG was diversity than HFD groups (Figure 8D). able to significantly increase survival rates irrespective of the The relative abundance of Firmicutes and Bacteroidetes phyla type of challenge under ND (MST from 28 to 34 weeks for SI and the ratio between them were analyzed at both time points, and MST from 19 to 24 weeks for MCI) but not under HFD namely week 16 (Figures 9A,G) and week 28 (Figures 10A,G). feeding. Interestingly, NI induced in animals following antibiotic This analysis showed an inverse correlation between the relative treatment after SI, until no detectable CFU was achieved, was abundance of Bacteroidetes and the animals’ weight, with the protective in HFD but not in ND (Figure 7). F/B ratio being higher in HFD mice due to the reduction in Bacteroidetes. The detailed gut microbiota composition results are shown HFD-Induced Obesity and Chemotherapy in Figure 9 (week 16) and Figure 10 (week 28). Analysis Reduce Gut Microbiota Diversity of the most abundant families showed a significant increase The gut microbiota studies were performed at weeks 4 and in Ruminococcaceae and Rikenellaceae, and a reduction in 16 post-infection, corresponding to weeks 16 and 28 from the Porphyromonadaceae with HFD at week 28. Among the genera beginning of the diet. In our study, OTUs can be considered as from the Firmicutes phylum, mice fed with HFD showed a higher the species found in fecal samples. Given the same number of abundance of Acetatifactor, Anaerotruncus, Clostridium XIVa, reads sequenced, a smaller number of OTUs were found in HFD and Oscillibacter at week 28. In contrast, analysis of genera from samples at both time points (Figures 8A,B), thus suggesting a the Bacteroidetes phylum showed that HFD mice had a lower lower microbial diversity. abundance of Alistipes at week 16 than ND, whereas the opposite We then assessed the dissimilarity between both diets in was the case at week 28. The levels of the Barnesiella genus are terms of microbiota composition using an NMDS plot and markedly higher in ND-fed animals at all time points, whereas the Bray–Curtis dissimilarity index. Samples were grouped Parasuterella (Proteobacteria) is higher in HFD. It is interesting according to the type of diet, showing that microbiota to note the predominance of the Akkermansia genus at both time composition differs substantially between diet conditions points, and the appearance of Mucispirillum at the last time point, (Figure 8C). Other comparisons were performed to evaluate in HFD-fed mice. the influence of BCG vaccination and infection dose on The effect of chemotherapy used to induce NI was also microbiota composition. This analysis was performed assessed on gut microbiota (Figure 11). Results indicated that separately for the ND and HFD groups, showing no significant chemotherapy decreased the F/B ratio in ND due to both differences (Figure S6). a decreased abundance of Firmicutes and increased levels of Frontiers in Immunology | www.frontiersin.org 9 October 2019 | Volume 10 | Article 2464 Arias et al. Microbiota High Fat Diet and Tuberculosis FIGURE 6 | Evolution of survival after Mtb infection for the first experiment. Each panel compares sham and BCG vaccinated groups: ND and SI (A), ND and MCI (B), HFD and SI (C), HFD and MCI (D). Median survival times and p-values are indicated in each panel. Log-rank test. Bacteroidetes. Interestingly, chemotherapy showed the same spectrum of human lesions and is considered to best reflect trend for HFD, although the differences were not statistically the progression from infection to disease (26). This progression significant. In the Firmicutes phylum, Lachnospiraceae and differs from the general tolerance toward Mtb infection shown by Ruminococcaceae were reduced with chemotherapy in both inbred mice, which only develop proliferative lesions (27, 28). diets, whereas for Bacteroidetes, the Porphyromonadaceae family In view of recent data, our model is more complex than we increased with chemotherapy. Finally, the genus Akkermansia initially thought. Dalby et al. (29) have demonstrated that refined was substantially increased with chemotherapy in ND. diets reduce gut microbiota diversity and increase the F/B ratio irrespective of obesity or dietary fat. They showed that a refined DISCUSSION low fat diet is also able to induce dysbiosis, with no link to obesity or hyperglycemia. In any case, a high F/B ratio favors glucose To the best of our knowledge, this is the first time that diet- intolerance and is thus linked to an impaired immune response induced obesity has been studied in an experimental TB model. (14). As such, we have two independent factors that can explain We have chosen to test the impact of an HFD as obesity inducer as the increased progression toward TB. First, although metabolic this is the most robust and best characterized model for inducing syndrome and DM2-related obesity, which are clearly linked to metabolic syndrome and glucose intolerance, which are the main an immune dysfunction, have been observed in other models, factors leading to DM2 (14). Although our model was able to this is not the case in ours. Pre-feeding with HFD for 8 weeks in induce a marked increase in body weight, we did not observe CD-1 mice reduced the lethality of Trypanosoma cruzi infection a constant hyperglycemia and animals had no need for insulin by 35% (7). Secondly, dysbiosis and an increase in the F/B ratio. treatment. Indeed, this is the current situation for millions of This has also been linked to a decreased resistance to disease in citizens with this syndrome who will go on to develop DM2 (25). a study comparing wild mice with lab mice (30). On the other These variables are introduced in the context of the hand, Mtb infection has consistently been shown to be linked to a C3HeB/FeJ mouse, a unique strain that is able to reproducethe reduction in gut microbiota diversity (15, 17). Our data show that Frontiers in Immunology | www.frontiersin.org 10 October 2019 | Volume 10 | Article 2464 Arias et al. Microbiota High Fat Diet and Tuberculosis stronger reaction after vaccination with BCG than with sham, in contrast to what occurs in ND animals. Even when this is not maintained throughout the chronic phase, it seems to be indicative of a delayed immune response. Moreover, MCI mice (both BCG and sham vaccinated) exhibit a stronger infiltration at all the time points studied, although a higher percentage of damaged area at week 3 after BCG vaccination in ND reflects a better immune response due to the vaccination. This different reaction in HFD correlates with the immune impairment of the response against TB linked to obesity and DM2, as seen above, although this chronic increase in the pro-inflammatory response does not result in the expected clear histological phenotype in terms of increasing the number of exudative lesions. This is in accordance with the intensity of the risk factor shown by DM2 (around ×3), which is much lower than the risk factor for immunosuppression by anti-TNF or AIDS (around ×100), where a clear phenotype can be seen. As described by several authors, aerosol infection is able to induce two kinds of lesions in this model, namely proliferative lesions, which are characterized by a slow progression and spreading via infected foamy macrophages, and exudative lesions, in which neutrophils accumulate at the periphery, thereby fueling the extracellular bacillary growth that finally results in large-scale central necrosis (20, 26, 33, 34). Surprisingly, our analysis of the quality of the lesions revealed larger exudative regions in SI animals than those MCI infected. As we are currently unable to explain this phenomenon, further studies in this regard are required. The pulmonary cytokine profile showed that, for SI animals, there is a more pro-inflammatory profile in ND than in HFD fed mice, irrespective of vaccination status, thus confirming the delay in the onset of the immune response in HFD seen in the histometry. However, this pattern reverts when animals are challenged with MCI, thus receiving far higher bacillary counts. This strong pro-inflammatory response in HFD mice has no advantage in terms of BL reduction, a fact than may reflect a FIGURE 7 | Evolution of survival after Mtb infection for the second experiment. Each panel compares sham and NI groups: ND (A) and HFD (B). Median higher basal pro-inflammatory milieu. This could also explain the survival times and p-values are indicated in each panel. Log-rank test. pathology results. Data published by Rosshart et al. (30) regarding the ability of natural microbiota, in terms of diversity and F/B ratio, to benefit a balanced systemic and local inflammatory response, HFD aggravates this tendency, thus resulting in a lower survival which in this case would have more similarities to ND than in animals fed with HFD. HFD fed mice, merit more in-depth studies, including mice Although there were no differences in BL between the diets, fed with refined low fat diets (29). This is an important issue BCG only had a positive effect in ND mice, thus revealing given the recent paper by Maji et al. (19), which found a a diminished protective capacity under HFD conditions. To higher diversity of the intestinal microbiome in TB patients the best of our knowledge, and with the exception of some compared with controls, thus resulting in a higher presence experiments with a Beijing strain challenge (31), this is the of short-chain fatty acids. These authors claim that these first time that no protective effect has been seen upon BCG products, especially butyrate, have a strong anti-inflammatory vaccination in mice in terms of BL reduction. However, these effect after stimulating Tregs in the gut, thus promoting strains are related to the induction of a different pattern of IL-10 production and facilitating a chronic infection, which immune response after triggering an IFN type I response (32), contradicts the concept of Rosshart et al. (30). Indeed, as therefore further studies are required to determine whether HFD highlighted by the authors, there is an important confusing also induces such a response. factor in that study, namely the BMI, which is significantly Histometry studies shortly after the SI challenge showed an lower in TB patients. It is well-known that a low BMI unusual reaction for the HFD group as these mice exhibit a has a major impact in TB and other diseases by inducing Frontiers in Immunology | www.frontiersin.org 11 October 2019 | Volume 10 | Article 2464 Arias et al. Microbiota High Fat Diet and Tuberculosis FIGURE 8 | Analysis of the microbiota diversity based on 16S rRNA sequencing. In each panel the ND group is colored in black and the HFD group in red. Rarefaction curves at week 16 (A) and week 28 (B) after starting the different diets are represented. (C) NMDS ordination in samples following different diets. (D) Shannon, Simpson and inverse Simpson diversity indexes. **p < 0.01, ***p < 0.001; Mann–Whitney test. a global immune dysfunction as a consequence of protein- effect” caused by the abundance of Porphyromonadaceae energy undernutrition, which impairs T-lymphocyte-mediated family, and specifically Barnesiella, found under responses (35). ND conditions. In our model, HFD caused gut microbiota dysbiosis, thus In our system, analysis of the microbiota showed differences inducing a pro-inflammatory milieu that limits the bacteriome with previous data obtained in mice fed with HFD. These diversity. This caused an increase in the F/B ratio. These are differences are probably related to both Mtb infection and the the most important phyla in humans, with Bacteroidetes being time of feeding, as in our case the analysis was done after 16 and reduced due to obesity and subsequently being restored, along 28 weeks of feeding, which is far longer than the usual 8–12 weeks with a reduction in Firmicutes, after losing body weight on reported previously (14, 29). a diet (36). The question now arises as to how this dysbiosis With regard to the Firmicutes phyla, ND presented lower influences the fate of TB progression. The recent study by levels (at week 28) of Ruminococcaceae, which is related to the Rosshart et al. (30) links a high diversity and low F/B ratio to the production of short-chain fatty acids (37), thus triggering an anti- promotion of host fitness and limited inflammatory responses inflammatory response. It is interesting to note that the epitope against influenza A and colorectal tumorigenesis. This is expressed on the flagellin of the Lachnospiraceae, which becomes coherent with our findings, which also point to a sort of “neutral the most important Firmicutes family in both diets, is a potent Frontiers in Immunology | www.frontiersin.org 12 October 2019 | Volume 10 | Article 2464 Arias et al. Microbiota High Fat Diet and Tuberculosis FIGURE 9 | Taxonomic composition of the intestinal microbiota based on 16S rRNA sequencing at week 16 after starting both diets. Analysis of the relative abundance of the most abundant phylum (A), class (B), order (C), family (D), and genus (E) in ND vs. HFD. (F) Bar plot representation of the OTUs of each sample according the taxonomic classification. (G) Regression plot of Firmicutes and Bacteroidetes phyla against animal weight and analysis of the F/B ratio in ND and HFD. *p < 0.05, **p < 0.01, ***p < 0.001; Mann–Whitney test. inducer of Tregs and TGF-β in dendritic cells in mice (38, 39). data from Namasyvayam et al. (16) in Mtb-infected C57BL/6J- Lachnospiraceae increased significantly in mice under stress. CD45a(Ly5a) mice also found this family, but in a very low In this study, stress caused an increase in systemic corticoid, relative abundance (around 3%). As such, this may be a critical IFN-γ and IL-17 levels, as well as mucosal injury (40). This is point in the gut microbiome of C3HeB/FeJ as it may be coherent with the stress induced by TB infection regardless of determinant for triggering the strong inflammatory response that diet. At a genus level, and although they only represented a small causes exudative lesions. Lactobacillaceae has also been related percentage of the total microbiota, HFD presented an increase to a decrease in the inflammatory response, and thus immune in Clostridium XIVa, which promotes the expansion of Tregs balance, although this proposal remains controversial as it was (39), as well as Anaerotruncus and Acetatifactor, which have been reported in the context of inflammatory bowel disease (37). related to arthritis progression in mice, thus suggesting a pro- The most remarkable differences caused by diet were seen inflammatory response (41). Interestingly, we have also detected in the Bacteroidetes phylum. Thus, HFD-fed mice showed a total absence of the Lactobacillaceae family in C3HeB/FeJ significantly lower levels of the Barnesiella genus at both time mice. This is a relevant factor as, according to Winglee et al. points, along with an increase in the Alistipes genus with (15), this is the predominant family (around 50% of relative time. These changes may be crucial in the different evolution abundance) in Mtb-infected C57BL/6 mice. Curiously, recent experimented by both groups. Thus, the most abundant family Frontiers in Immunology | www.frontiersin.org 13 October 2019 | Volume 10 | Article 2464 Arias et al. Microbiota High Fat Diet and Tuberculosis FIGURE 10 | Taxonomic composition of the intestinal microbiota based on 16S rRNA sequencing at week 28 after starting both diets. Analysis of the relative abundance of the most abundant phylum (A), class (B), order (C), family (D), and genus (E) in ND vs. HFD. (F) Bar plot representation of the OTUs of each sample according the taxonomic classification. (G) Regression plot of Firmicutes and Bacteroidetes phyla against animal weight and analysis of the F/B ratio in ND and HFD. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001; Mann–Whitney test. in both diets is Porphyromonadaceae, which is decreased under With regard to Proteobacteria, HFD caused a sustained HFD conditions. Porphyromonadaceae is increased in mice fed increase in the genus Parasutterella, which is precisely the with a high-protein/low carbohydrate diet (42) and decreased in opposite to what was found in C57BL/6 (47). Additional mice fed with omega-6-polyunsaturated fatty acids (43). This has microbiome studies in humans and rats have associated been related with the dysbiosis shown in patients with ankylosing an increased abundance of Parasutterella with dysbiosis spondylitis, a heritable immune-mediated arthropathy (44). The in the gut, with this genus being increased in the Barnesiella genus has proved useful for reducing the colonization submucosa of individuals with Crohn’s disease (48) and of Enterococcus in a “neutral way” as it is not thought to be in rats with hypertriglyceridemia-related acute necrotizing associated with immune development or inflammatory diseases pancreatitis (49). in the intestine (45). In contrast, Alistipes is the most significant Finally, HFD mice had a higher relative abundance of over-represented taxon related to frailty in the elderly (46) and Mucispirillum and Akkermansia genera. Unlike the majority is found to be increased under HFD conditions (47). Overall, of commensal bacteria, Akkermansia and Mucispirillum evade it appears that colonization with the Porphyromonadaceae T-independent IgA and elicit a T-dependent response (50). In family, and in particular the Barnesiella genus, may have a contrast to our observations, the abundance of Akkermansia positive impact on reducing the pro-inflammatory milieu in muciniphila is decreased during obesity and diabetes, and the gut. administration thereof can counteract the development of Frontiers in Immunology | www.frontiersin.org 14 October 2019 | Volume 10 | Article 2464 Arias et al. Microbiota High Fat Diet and Tuberculosis FIGURE 11 | Taxonomic composition of the intestinal microbiota based on 16S rRNA sequencing between different diets and under chemotherapy. Analysis of the relative abundance of the most abundant phylum (A), class (B), order (C), family (D), and genus (E). (F) Bar plot representation of the OTUs of each sample according the taxonomic classification. (G) Analysis of the F/B ratio. *p < 0.05, **p < 0.01; Mann–Whitney test. Frontiers in Immunology | www.frontiersin.org 15 October 2019 | Volume 10 | Article 2464 Arias et al. Microbiota High Fat Diet and Tuberculosis HFD-induced obesity and gut barrier dysfunction. Its effect ETHICS STATEMENT is increased when given inactivated by pasteurization, a mechanism linked to the increase in mucus production The animal study was reviewed and approved by Hospital (51). Mucispirillum schaedleri has been reported to be Universitari Germans Trias i Pujol (registered as B9900005). elevated in abundance during intestinal inflammation in mouse models, and has been described as a mucus-associated AUTHOR CONTRIBUTIONS bacterium adapted to the oxidative burst that occurs during inflammation (52). P-JC led the project. CV, IC, and P-JC conceived and planned Another unknown was the effect caused by chemotherapy the experiments. LA, PC, JD, YR, GG, MT-P, GT, and EG after aerosol infection. A more significant impact was found carried out the experiments. LA, GG, PC, IC, CV, and P- for the gut microbiota of ND mice, which could give a clue JC contributed to interpretation of the results. LA, CV, and as to why natural infection was less effective in this group. P-JC took the lead in writing the manuscript. All authors Animals fed with ND presented a drastic reduction in the contributed to drafting the work and gave final approval for F/B ratio with chemotherapy. This is mainly explained by a the version to be published. All authors agree to be accountable reduction in Lachnospiraceae for Firmicutes and an increase for all aspects of the work in ensuring that questions related in Porphyromonadaceae for Bacteroidetes, thus increasing to the accuracy or integrity of any part of the work are the “neutral” effect of the microbiota. Chemotherapy also appropriately investigated and resolved. All authors provided markedly promoted the presence of the Akkermansia genus critical feedback and helped shape the research, analysis, in ND, thus increasing T-cell stimulation. In view of these and manuscript. results, we can hypothesize that induction of a protective effect by natural infection might require a certain anti- inflammatory effect. This is coherent with recent findings FUNDING suggesting that excess effector T-cell stimulation might be This work was supported by the European Commission Horizon deleterious to a protective immune response (53, 54). In 2020 research and innovation program under grant agreement HFD mice, in contrast, chemotherapy did not impact the TBVAC2020 No. 643381, European Research Council research abundance of Firmicutes, with the increase in several Clostridium grant 638553-TB-ACCELERATE and Ministerio de Economía y appearing to “compensate” the loss of Ruminococcaceae and Competitividad (Spanish government) research grant SAF2017- Lachnospiraceae. It also failed to impact the abundance of 92345- EXP (to IC). CV has a project Servet II contract Bacteroidetes and triggered the presence of Enterobacteriaceae, funded by the Instituto de Salud Carlos III (ISCIII, CPII which is difficult to analyze since it did not affect the levels 18/00031). LA received an Improvement and Mobility Program of Akkermansia. grant from the Centro de Investigación Biomédica en Red Overall, our findings show that HFD accelerates the de Enfermedades Respiratorias (CIBERES). GG contract was progression toward active TB. Although we have not been able funded by Ministerio de Economia y Competitividad (Spanish to discern the exact mechanism of action, our results point to Government) (BES-2014-071066). PC contract was funded by the a dysregulation in the ability to react to Mtb infection, which Instituto de Salud Carlos III (IFI14/00015). The Experimental appears to slow the appearance of immunity in the early stages Tuberculosis Unit is accredited by the Catalan Agency for post-challenge. HFD-fed animals show a reduced ability to detect Management of University and Research Grants with code 2017 low dose infection, while reacting excessively to a high dose, thus SGR500 and the IGTP is a member of the CERCA network leading to a reduced survival. This may be related to the HFD of institutes. The funders played no role in study design, data causing early DM2, as supported by a dysbiosis in the intestinal collection and analysis, decision to publish, or preparation of microbiota that promotes an intolerance to glucose. Interestingly, the manuscript. even when natural immunity induced by Mtb followed by chemotherapy is able to induce protection, BCG vaccination protection is reduced in terms of BL and survival. This finding ACKNOWLEDGMENTS highlights the need to further explore the mechanisms of this difference in order to design a better vaccine that can protect We thank the IGTP Flow Cytometry Core Facility and staff for against this rapidly growing comorbidity, which is starting to their contribution to this publication. hamper current efforts to control TB. SUPPLEMENTARY MATERIAL DATA AVAILABILITY STATEMENT The Supplementary Material for this article can be found The datasets generated for this study are available on request to online at: https://www.frontiersin.org/articles/10.3389/fimmu. the corresponding author. 2019.02464/full#supplementary-material Frontiers in Immunology | www.frontiersin.org 16 October 2019 | Volume 10 | Article 2464 Arias et al. Microbiota High Fat Diet and Tuberculosis REFERENCES 23. Cooper AM, Callahan JE, Keen M, Belisle JT, Orme IM. Expression of memory immunity in the lung following re-exposure to Mycobacterium tuberculosis. 1. WHO. Global Strategy and Targets for Tuberculosis Prevention, Care and Tuber Lung Dis. (1997) 78:67–73. doi: 10.1016/S0962-8479(97)90017-4 Control After 2015. (2014). p. 1–23. Available online at: http://www.who.int/ 24. Orme IM. Characteristics and specificity of acquired immunologic memory to tb/post2015%7B_%7DTBstrategy.pdf Mycobacterium tuberculosis infection. J Immunol. (1988) 140:3589–93. 2. Narasimhan P, Wood J, Macintyre CR, Mathai D. Risk factors for tuberculosis. 25. Algoblan A, Alalfi M, Khan M. Mechanism linking diabetes mellitus Pulm Med. (2013) 2013:828939. doi: 10.1155/2013/828939 and obesity. Diabetes Metab Syndr Obes Targets Ther. (2014) 3. Cegielski JP, McMurray DN. The relationship between malnutrition and 7:587. doi: 10.2147/DMSO.S67400 tuberculosis: evidence from studies in humans and experimental animals. Int 26. Marzo E, Vilaplana C, Tapia G, Diaz J, Garcia V, Cardona P-J. Damaging J Tuberc Lung Dis. (2004) 8:286–98. role of neutrophilic infiltration in a mouse model of progressive tuberculosis. 4. Gupta K, Gupta R, Atreja A, Verma M, Vishvkarma S. Tuberculosis and Tuberculosis. (2014) 94:55–64. doi: 10.1016/j.tube.2013.09.004 nutrition. Lung India. (2009) 26:9. doi: 10.4103/0970-2113.45198 27. Gil O, Díaz I, Vilaplana C, Tapia G, Díaz J, Fort M, et al. Granuloma 5. Leung CC. Lower risk of tuberculosis in obesity. Arch Intern Med. (2007) encapsulation is a key factor for containing tuberculosis infection in minipigs. 167:1297. doi: 10.1001/archinte.167.12.1297 PLoS ONE. (2010) 5:e10030. doi: 10.1371/journal.pone.0010030 6. Roth J, Sahota N. Obesity paradox, obesity orthodox, and the 28. Gil O, Guirado E, Gordillo S, Díaz J, Tapia G, Vilaplana C, et al. metabolic syndrome: an approach to unity. Mol Med. (2016) Intragranulomatous necrosis in lungs of mice infected by aerosol with 22:1. doi: 10.2119/molmed.2016.00211 Mycobacterium tuberculosis is related to bacterial load rather than 7. Brima W, Eden DJ, Mehdi SF, Bravo M, Wiese MM, Stein J, et al. The brighter to any one cytokine or T cell type. Microbes Infect. (2006) 8:628– (and evolutionarily older) face of the metabolic syndrome: evidence from 36. doi: 10.1016/j.micinf.2005.08.014 Trypanosoma cruzi infection in CD-1 mice. Diabetes Metab Res Rev. (2015) 29. Dalby MJ, Ross AW, Walker AW, Morgan PJ. Dietary uncoupling of gut 31:346–59. doi: 10.1002/dmrr.2636 microbiota and energy harvesting from obesity and glucose tolerance in mice. 8. Jeon CY, Murray MB. Diabetes mellitus increases the risk of active Cell Rep. (2017) 21:1521–33. doi: 10.1016/j.celrep.2017.10.056 tuberculosis: a systematic review of 13 observational studies. PLoS Med. (2008) 30. Rosshart SP, Vassallo BG, Angeletti D, Hutchinson DS, Morgan AP, Takeda K, 5:1091–101. doi: 10.1371/journal.pmed.0050181 et al. Wild mouse gut microbiota promotes host fitness and improves disease 9. Little M, Humphries S, Patel K, Dewey C. Decoding the type resistance. Cell. (2017) 171:1015–28 e13. doi: 10.1016/j.cell.2017.09.016 2 diabetes epidemic in rural India. Med Anthropol. (2017) 31. Ordway DJ, Shang S, Henao-Tamayo M, Obregon-Henao A, Nold L, Caraway 36:96–110. doi: 10.1080/01459740.2016.1231676 M, et al. Mycobacterium bovis BCG-mediated protection against W-Beijing 10. Manyema M, Veerman JL, Chola L, Tugendhaft A, Labadarios D, Hofman strains of Mycobacterium tuberculosis is diminished concomitant with the K. Decreasing the burden of type 2 diabetes in South Africa: the emergence of regulatory T cells. Clin Vaccine Immunol. (2011) 18:1527– impact of taxing sugar-sweetened beverages. PLoS ONE. (2015) 10:1– 35. doi: 10.1128/CVI.05127-11 17. doi: 10.1371/journal.pone.0143050 32. Manca C, Tsenova L, Freeman S, Barczak AK, Tovey M, Murray PJ, et al. 11. Hotamisligil GS. Inflammation and metabolic disorders. Nature. (2006) tuberculosis W/Beijing strains upregulate type I IFNs and increase expression 444:860–7. doi: 10.1038/nature05485 of negative regulators of the jak-stat pathway. J Interf Cytokine Res. (2005) 12. Hodgson K, Morris J, Bridson T, Govan B, Rush C, Ketheesan N. 25:694–701. doi: 10.1089/jir.2005.25.694 Immunological mechanisms contributing to the double burden of 33. Dutta NK, Illei PB, Jain SK, Karakousis PC. Characterization of a novel diabetes and intracellular bacterial infections. Immunology. (2015) necrotic granuloma model of latent tuberculosis infection and reactivation in 144:171–85. doi: 10.1111/imm.12394 mice. Am J Pathol. (2014) 184:2045–55. doi: 10.1016/j.ajpath.2014.03.008 13. Lin H, An Y, Hao F, Wang Y, Tang H. Correlations of fecal metabonomic and 34. Irwin SM, Driver E, Lyon E, Schrupp C, Ryan G, Gonzalez-Juarrero M, et al. microbiomic changes induced by high-fat diet in the pre-obesity state. Sci Rep. Presence of multiple lesion types with vastly different microenvironments in (2016) 6:1–14. doi: 10.1038/srep21618 C3HeB/FeJ mice following aerosol infection with Mycobacterium tuberculosis. 14. Rabot S, Membrez M, Blancher F, Berger B, Moine D, Krause L, et al. High fat Dis Model Mech. (2015) 8:591–602. doi: 10.1242/dmm.019570 diet drives obesity regardless the composition of gut microbiota in mice. Sci 35. Cegielski JP, Arab L, Cornoni-Huntley J. Nutritional risk factors for Rep. (2016) 6:1–11. doi: 10.1038/srep32484 tuberculosis among adults in the United States, 19711992. Am J Epidemiol. 15. Winglee K, Eloe-Fadrosh E, Gupta S, Guo H, Fraser C, Bishai W. Aerosol (2012) 176:409–22. doi: 10.1093/aje/kws007 mycobacterium tuberculosis infection causes rapid loss of diversity in gut 36. Ley R, Turnbaugh P, Klein S, Gordon J. Microbial ecology: microbiota. PLoS ONE. (2014) 9:1–9. doi: 10.1371/journal.pone.0097048 human gut microbes associated with obesity. Nature. (2006) 16. Namasivayam S, Maiga M, Yuan W, Thovarai V, Costa DL, Mittereder 444:1022–3. doi: 10.1038/4441022a LR, et al. Longitudinal profiling reveals a persistent intestinal dysbiosis 37. Zheng H, Chen M, Li Y, Wang Y, Wei L, Liao Z, et al. Modulation of gut triggered by conventional anti-tuberculosis therapy. Microbiome. (2017) microbiome composition and function in experimental colitis treated with 5:71. doi: 10.1186/s40168-017-0286-2 sulfasalazine. Front Microbiol. (2017) 8:1–15. doi: 10.3389/fmicb.2017.01703 17. Wipperman MF, Fitzgerald DW, Juste MAJ, Taur Y, Namasivayam S, Sher A, 38. Cong Y, Feng T, Fujihashi K, Schoeb TR, Elson CO. A dominant, coordinated et al. Antibiotic treatment for Tuberculosis induces a profound dysbiosis of T regulatory cell-IgA response to the intestinal microbiota. Proc Natl Acad Sci the microbiome that persists long after therapy is completed. Sci Rep. (2017) USA. (2009) 106:19256–61. doi: 10.1073/pnas.0812681106 7:1–11. doi: 10.1038/s41598-017-10346-6 39. Zhao Q, Elson CO. Adaptive immune education by gut microbiota antigens. 18. Kaakoush NO. Insights into the role of erysipelotrichaceae in the human host. Immunology. (2018) 154:28–37. doi: 10.1111/imm.12896 Front Cell Infect Microbiol. (2015) 5:1–4. doi: 10.3389/fcimb.2015.00084 40. Li S, Wang Z, Yang Y, Yang S, Yao C, Liu K, et al. Lachnospiraceae shift in 19. Maji A, Misra R, Dhakan DB, Gupta V, Mahato NK, Saxena R, et al. the microbial community of mice faecal sample effects on water immersion Gut microbiome contributes to impairment of immunity in pulmonary restraint stress. AMB Express. (2017) 7:82. doi: 10.1186/s13568-017-0383-4 tuberculosis patients by alteration of butyrate and propionate producers. 41. Liu X, Zeng B, Zhang J, Li W, Mou F, Wang H, et al. Role of the gut Environ Microbiol. (2018) 20:402–19. doi: 10.1111/1462-2920.14015 microbiome in modulating arthritis progression in mice. Sci Rep. (2016) 20. Cardona PJ. The key role of exudative lesions and their encapsulation: 6:1–11. doi: 10.1038/srep30594 lessons learned from the pathology of human pulmonary tuberculosis. Front 42. Kim E, Kim DB, Park JY. Changes of mouse gut microbiota diversity and Microbiol. (2015) 6:1–8. doi: 10.3389/fmicb.2015.00612 composition by modulating dietary protein and carbohydrate contents: a pilot 21. Cardona PJ, Vilaplana C. Multiple consecutive infections might study. Prev Nutr Food Sci. (2016) 21:57–61. doi: 10.3746/pnf.2016.21.1.57 explain the lack of protection by BCG. PLoS ONE. (2014) 43. Liu T, Hougen H, Vollmer AC, Hiebert SM. Gut bacteria 9:e94736. doi: 10.1371/journal.pone.0094736 profiles of Mus musculus at the phylum and family levels are 22. Kamath AB, Behar SM. Anamnestic responses of mice following. Society. influenced by saturation of dietary fatty acids. Anaerobe. (2012) (2005) 73:6110–8. doi: 10.1128/IAI.73.9.6110-6118.2005 18:331–7. doi: 10.1016/j.anaerobe.2012.02.004 Frontiers in Immunology | www.frontiersin.org 17 October 2019 | Volume 10 | Article 2464 Arias et al. Microbiota High Fat Diet and Tuberculosis 44. Costello ME, Ciccia F, Willner D, Warrington N, Robinson PC, Gardiner bacterium improves metabolism in obese and diabetic mice. Nat Med. (2017) B, et al. Brief report: intestinal dysbiosis in ankylosing spondylitis. Arthritis 23:107–13. doi: 10.1038/nm.4236 Rheumatol. (2015) 67:686–91. doi: 10.1002/art.38967 52. Loy A, Pfann C, Steinberger M, Hanson B, Herp S, Brugiroux S, et al. 45. Ubeda C, Bucci V, Caballero S, Djukovic A, Toussaint NC, Equinda M, Schaedleri, a core member of the murine gut microbiota. mSystems. (2017) et al. Intestinal microbiota containing Barnesiella species cures vancomycin- 2:e00171-16. doi: 10.1128/mSystems.00171-16 resistant Enterococcus faecium colonization. Infect Immun. (2013) 81:965– 53. Zelmer A, Stockdale L, Prabowo SA, Cia F, Spink N, Gibb M, et al. High 73. doi: 10.1128/IAI.01197-12 monocyte to lymphocyte ratio is associated with impaired protection 46. Langille MGI, Meehan CJ, Koenig JE, Dhanani AS, Rose RA, Howlett SE, after subcutaneous administration of BCG in a mouse model of et al. Microbial shifts in the aging mouse gut. Microbiome. (2014) 2:1– tuberculosis. F1000Research. (2018) 7:296. doi: 10.12688/f1000research.1 12. doi: 10.1186/s40168-014-0050-9 4239.1 47. Zhang C, Zhang M, Pang X, Zhao Y, Wang L, Zhao L. Structural resilience of 54. Tzelepis F, Blagih J, Khan N, Gillard J, Mendonca L, Roy DG, the gut microbiota in adult mice under high-fat dietary perturbations. ISME J. et al. Mitochondrial cyclophilin D regulates T cell metabolic (2012) 6:1848–57. doi: 10.1038/ismej.2012.27 responses and disease tolerance to tuberculosis. Sci Immunol. (2018) 48. Chiodini RJ, Dowd SE, Chamberlin WM, Galandiuk S, Davis B, Glassing A. 3:eaar4135. doi: 10.1126/sciimmunol.aar4135 Microbial population differentials between mucosal and submucosal intestinal tissues in advanced Crohn’s disease of the ileum. PLoS ONE. (2015) 10:1– Conflict of Interest: The authors declare that the research was conducted in the 19. doi: 10.1371/journal.pone.0134382 absence of any commercial or financial relationships that could be construed as a 49. Huang C, Chen J, Wang J, Zhou H, Lu Y, Lou L, et al. Dysbiosis of potential conflict of interest. intestinal microbiota and decreased antimicrobial peptide level in paneth cells during hypertriglyceridemia-related acute necrotizing pancreatitis in rats. Copyright © 2019 Arias, Goig, Cardona, Torres-Puente, Díaz, Rosales, Garcia, Front Microbiol. (2017) 8:1–13. doi: 10.3389/fmicb.2017.00776 Tapia, Comas, Vilaplana and Cardona. This is an open-access article distributed 50. Million M, Tomas J, Wagner C, Lelouard H, Raoult D, Gorvel JP. New under the terms of the Creative Commons Attribution License (CC BY). The use, insights in gut microbiota and mucosal immunity of the small intestine. Hum distribution or reproduction in other forums is permitted, provided the original Microbiome J. (2018) 7–8:23–32. doi: 10.1016/j.humic.2018.01.004 author(s) and the copyright owner(s) are credited and that the original publication 51. Plovier H, Everard A, Druart C, Depommier C, Van Hul M, Geurts L, et al. A in this journal is cited, in accordance with accepted academic practice. No use, purified membrane protein from Akkermansia muciniphila or the pasteurized distribution or reproduction is permitted which does not comply with these terms. Frontiers in Immunology | www.frontiersin.org 18 October 2019 | Volume 10 | Article 2464