Two recent studies have reanalyzed previously published data and found that when data sets were analyzed independently, there was limited support for the widely accepted hypothesis that changes in the microbiome are associated with obesity. This hypothesis was reconsidered by increasing the number of data sets and pooling the results across the individual data sets. The preferred reporting items for systematic reviews and meta-analyses guidelines were used to identify 10 studies for an updated and more synthetic analysis. Alpha diversity metrics and the relative risk of obesity based on those metrics were used to identify a limited number of significant associations with obesity; however, when the results of the studies were pooled by using a random-effect model, significant associations were observed among Shannon diversity, the number of observed operational taxonomic units, Shannon evenness, and obesity status. They were not observed for the ratio of Bacteroidetes and Firmicutes or their individual relative abundances. Although these tests yielded small P values, the difference between the Shannon diversity indices of nonobese and obese individuals was 2.07%. A power analysis demonstrated that only one of the studies had sufficient power to detect a 5% difference in diversity. When random forest machine learning models were trained on one data set and then tested by using the other nine data sets, the median accuracy varied between 33.01 and 64.77% (median, 56.68%). Although there was support for a relationship between the microbial communities found in human feces and obesity status, this association was relatively weak and its detection is confounded by large interpersonal variation and insufficient sample sizes. IMPORTANCE As interest in the human microbiome grows, there is an increasing number of studies that can be used to test numerous hypotheses across human populations. The hypothesis that variation in the gut microbiota can explain or be used to predict obesity status has received considerable attention and is frequently mentioned as an example of the role of the microbiome in human health. Here we assessed this hypothesis by using 10 independent studies and found that although there is an association, it is smaller than can be detected by most microbiome studies. Furthermore, we directly tested the ability to predict obesity status on the basis of the composition of an individual’s microbiome and found that the median classification accuracy is between 33.01 and 64.77%. This type of analysis can be used to design future studies and expanded to explore other hypotheses.

Rationale: Based on surface brushings and bronchoalveolar lavage fluid, Hilty and coworkers demonstrated microbiomes in the human lung characteristic of asthma and chronic obstructive pulmonary disease (COPD), which have now been confirmed by others. Objectives: To extend these findings to human lung tissue samples. Methods: DNA from lung tissue samples was obtained from nonsmokers (n = 8); smokers without COPD (n = 8); patients with very severe COPD (Global Initiative for COPD [GOLD] 4) (n = 8); and patients with cystic fibrosis (CF) (n = 8). The latter served as a positive control, with sterile water as a negative control. All bacterial community analyses were based on polymerase chain reaction amplifying 16S rRNA gene fragments. Total bacterial populations were measured by quantitative polymerase chain reaction and bacterial community composition was assessed by terminal restriction fragment length polymorphism analysis and pyrotag sequencing. Measurement and Main Results: Total bacterial populations within lung tissue were small (20–1,252 bacterial cells per 1,000 human cells) but greater in all four sample groups versus the negative control group (P < 0.001). Terminal restriction fragment length polymorphism analysis and sequencing distinguished three distinct bacterial community compositions: one common to the nonsmoker and smoker groups, a second to the GOLD 4 group, and the third to the CF-positive control group. Pyrotag sequencing identified greater than 1,400 unique bacterial sequences and showed an increase in the Firmicutes phylum in GOLD 4 patients versus all other groups (P < 0.003) attributable to an increase in the Lactobacillus genus (P < 0.0007). Conclusions: There is a detectable bacterial community within human lung tissue that changes in patients with very severe COPD.

Cancer-associated fibroblasts (CAF) represent a functionally heterogeneous population of activated fibroblasts that constitutes a major component of tumor stroma. Although CAFs have been shown to promote tumor growth and mediate resistance to chemotherapy, the mechanisms by which they may contribute to immune suppression within the tumor microenvironment (TME) in lung squamous cell carcinoma (LSCC) remain largely unexplored. Here, we identified a positive correlation between CAF and monocytic myeloid cell abundances in 501 primary LSCCs by mining The Cancer Genome Atlas data sets. We further validated this finding in an independent cohort using imaging mass cytometry and found a significant spatial interaction between CAFs and monocytic myeloid cells in the TME. To delineate the interplay between CAFs and monocytic myeloid cells, we used chemotaxis assays to show that LSCC patient–derived CAFs promoted recruitment of CCR2+ monocytes via CCL2, which could be reversed by CCR2 inhibition. Using a three-dimensional culture system, we found that CAFs polarized monocytes to adopt a myeloid-derived suppressor cell (MDSC) phenotype, characterized by robust suppression of autologous CD8+ T-cell proliferation and IFNγ production. We further demonstrated that inhibiting IDO1 and NADPH oxidases, NOX2 and NOX4, restored CD8+ T-cell proliferation by reducing reactive oxygen species (ROS) generation in CAF-induced MDSCs. Taken together, our study highlights a pivotal role of CAFs in regulating monocyte recruitment and differentiation and demonstrated that CCR2 inhibition and ROS scavenging abrogate the CAF–MDSC axis, illuminating a potential therapeutic path to reversing the CAF-mediated immunosuppressive microenvironment.

Rationale: The relatively sparse but diverse microbiome in human lungs may become less diverse in chronic obstructive pulmonary disease (COPD). This article examines the relationship of this microbiome to emphysematous tissue destruction, number of terminal bronchioles, infiltrating inflammatory cells, and host gene expression. Methods: Culture-independent pyrosequencing microbiome analysis was used to examine the V3–V5 regions of bacterial 16S ribosomal DNA in 40 samples of lung from 5 patients with COPD (Global Initiative for Chronic Obstructive Lung Disease [GOLD] stage 4) and 28 samples from 4 donors (controls). A second protocol based on the V1–V3 regions was used to verify the bacterial microbiome results. Within lung tissue samples the microbiome was compared with results of micro–computed tomography, infiltrating inflammatory cells measured by quantitative histology, and host gene expression. Measurements and Main Results: Ten operational taxonomic units (OTUs) was found sufficient to discriminate between control and GOLD stage 4 lung tissue, which included known pathogens such as Haemophilus influenzae. We also observed a decline in microbial diversity that was associated with emphysematous destruction, remodeling of the bronchiolar and alveolar tissue, and the infiltration of the tissue by CD4⁺ T cells. Specific OTUs were also associated with neutrophils, eosinophils, and B-cell infiltration (P < 0.05). The expression profiles of 859 genes and 235 genes were associated with either enrichment or reductions of Firmicutes and Proteobacteria, respectively, at a false discovery rate cutoff of less than 0.1. Conclusions: These results support the hypothesis that there is a host immune response to microorganisms within the lung microbiome that appears to contribute to the pathogenesis of COPD.

Previous reports have shown that the gastrointestinal (GI) bacterial microbiota can have profound effects on the lungs, which has been described as the “gut-lung axis”. However, whether a “lung-gut” axis exists wherein acute lung inflammation perturbs the gut and blood microbiota is unknown. Adult C57/Bl6 mice were exposed to one dose of LPS or PBS instillation (n = 3 for each group) directly into lungs. Bacterial microbiota of the bronchoalveolar lavage fluid, blood, and cecum were determined using 454 pyrotag sequencing and quantitative polymerase chain reaction (qPCR) at 4 through 168 hours post-instillation. We then investigated the effects of oral neomycin and streptomycin (n = 8) on the microbiota at 4 and 24 hours post LPS instillation versus control treatment (n = 5 at baseline and 4 hours, n = 7 at 24 hours). At 24 hours post LPS instillation, the total bacterial count was significantly increased in the cecum (P<0.05); whereas the total bacterial count in blood was increased at 4, 48, and 72 hours (P<0.05). Antibiotic treatment reduced the total bacteria in blood but not in the cecum. The increase in total bacteria in the blood correlated with Phyllobacteriaceae OTU 40 and was significantly reduced in the blood for both antibiotic groups (P<0.05). LPS instillation in lungs leads to acute changes in the bacterial microbiota in the blood and cecum, which can be modulated with antibiotics.

A steep decline in sequencing costs drove an explosion in studies characterizing microbial communities from diverse environments. Although a significant amount of effort has gone into understanding the error profiles of DNA sequencers, little has been done to understand the downstream effects of the PCR amplification protocol. We quantified the effects of the choice of polymerase and number of PCR cycles on the quality of downstream data. We found that these choices can have a profound impact on the way that a microbial community is represented in the sequence data. The effects are relatively small compared to the variation in human stool samples; however, care should be taken to use polymerases with the highest possible fidelity and to minimize the number of rounds of PCR. These results also underscore that it is not possible to directly compare sequence data generated under different PCR conditions.