Motivation While deep-learning algorithms have demonstrated outstanding performance in semantic image segmentation tasks, large annotation datasets are needed to create accurate models. Annotation of histology images is challenging due to the effort and experience required to carefully delineate tissue structures, and difficulties related to sharing and markup of whole-slide images. Results We recruited 25 participants, ranging in experience from senior pathologists to medical students, to delineate tissue regions in 151 breast cancer slides using the Digital Slide Archive. Inter-participant discordance was systematically evaluated, revealing low discordance for tumor and stroma, and higher discordance for more subjectively defined or rare tissue classes. Feedback provided by senior participants enabled the generation and curation of 20 000+ annotated tissue regions. Fully convolutional networks trained using these annotations were highly accurate (mean AUC=0.945), and the scale of annotation data provided notable improvements in image classification accuracy. Availability and Implementation Dataset is freely available at: https://goo.gl/cNM4EL. Supplementary information Supplementary data are available at Bioinformatics online.

BackgroundThe Wnt pathway mediates differentiation of epithelial tissues; depending on the tissue types, Wnt can either drive or inhibit the differentiation process. We hypothesized that key genes in the Wnt pathway are suppressed in the human airway epithelium under the stress of cigarette smoking, a stress associated with dysregulation of the epithelial differentiated state.Methodology/Principal FindingsMicroarrays were used to assess the expression of Wnt-related genes in the small airway epithelium (SAE) obtained via bronchoscopy and brushing of healthy nonsmokers, healthy smokers, and smokers with COPD. Thirty-three of 56 known Wnt-related genes were expressed in the SAE. Wnt pathway downstream mediators β-catenin and the transcription factor 7-like 1 were down-regulated in healthy smokers and smokers with COPD, as were many Wnt target genes. Among the extracellular regulators that suppress the Wnt pathway, secreted frizzled-related protein 2 (SFRP2), was up-regulated 4.3-fold in healthy smokers and 4.9-fold in COPD smokers, an observation confirmed by TaqMan Real-time PCR, Western analysis and immunohistochemistry. Finally, cigarette smoke extract mediated up-regulation of SFRP2 and down-regulation of Wnt target genes in airway epithelial cells in vitro.Conclusions/SignificanceSmoking down-regulates the Wnt pathway in the human airway epithelium. In the context that Wnt pathway plays an important role in differentiation of epithelial tissues, the down-regulation of Wnt pathway may contribute to the dysregulation of airway epithelium differentiation observed in smoking-related airway disorders.

Even after quitting smoking, the risk of the development of chronic obstructive pulmonary disease (COPD) and lung cancer remains significantly higher compared to healthy nonsmokers. Based on the knowledge that COPD and most lung cancers start in the small airway epithelium (SAE), we hypothesized that smoking modulates miRNA expression in the SAE linked to the pathogenesis of smoking-induced airway disease, and that some of these changes persist after smoking cessation. SAE was collected from 10th to 12th order bronchi using fiberoptic bronchoscopy. Affymetrix miRNA 2.0 arrays were used to assess miRNA expression in the SAE from 9 healthy nonsmokers and 10 healthy smokers, before and after they quit smoking for 3 months. Smoking status was determined by urine nicotine and cotinine measurement. There were significant differences in the expression of 34 miRNAs between healthy smokers and healthy nonsmokers (p<0.01, fold-change >1.5), with functions associated with lung development, airway epithelium differentiation, inflammation and cancer. After quitting smoking for 3 months, 12 out of the 34 miRNAs did not return to normal levels, with Wnt/β-catenin signaling pathway being the top identified enriched pathway of the target genes of the persistent dysregulated miRNAs. In the context that many of these persistent smoking-dependent miRNAs are associated with differentiation, inflammatory diseases or lung cancer, it is likely that persistent smoking-related changes in SAE miRNAs play a role in the subsequent development of these disorders.

The toll-like receptors (TLRs) are important components of the respiratory epithelium host innate defense, enabling the airway surface to recognize and respond to a variety of insults in inhaled air. Based on the knowledge that smokers are more susceptible to pulmonary infection and that the airway epithelium of smokers with chronic obstructive pulmonary disease (COPD) is characterized by bacterial colonization and acute exacerbation of airway infections, we assessed whether smoking alters expression of TLRs in human small airway epithelium, the primary site of smoking-induced disease. Microarrays were used to survey the TLR family gene expression in small airway (10th–12th order) epithelium from healthy nonsmokers (n=60), healthy smokers (n=73) and smokers with COPD (n=36). Using the criteria of detection call of present in ≥50%, 6 of 10 TLRs (1, 2, 3, 4, 5 and 8) were expressed. Compared to nonsmokers, the most striking change was for TLR5, which was down-regulated in healthy smokers (1.4-fold, p<10−10) and smokers with COPD (1.6-fold, p<10−11). TaqMan RT-PCR confirmed these observations. Bronchial biopsy immunofluorescence studies showed that TLR5 was expressed mainly on the apical side of the epithelium and was decreased in healthy smokers and smokers with COPD. In vitro, the level of TLR5 downstream genes, IL-6 and IL-8, were highly induced by flagellin in TLR5 high-expressing cells compared to TLR5 low-expressing cells. In the context that TLR5 functions to recognize pathogens and activate innate immune responses, the smoking-induced down-regulation of TLR5 may contribute to smoking-related susceptibility to airway infection, at least for flagellated bacteria.

No abstract