Cross-scale multi-instance learning for pathological image diagnosis.

Analyzing high resolution whole slide images (WSIs) with regard to information across multiple scales poses a significant challenge in digital pathology. Multi-instance learning (MIL) is a common solution for working with high resolution images by classifying bags of objects (i.e. sets of smaller image patches). However, such processing is typically performed at a single scale (e.g., 20× magnification) of WSIs, disregarding the vital inter-scale information that is key to diagnoses by human pathologists. In this study, we propose a novel cross-scale MIL algorithm to explicitly aggregate inter-scale relationships into a single MIL network for pathological image diagnosis. The contribution of this paper is three-fold: (1) A novel cross-scale MIL (CS-MIL) algorithm that integrates the multi-scale information and the inter-scale relationships is proposed; (2) A toy dataset with scale-specific morphological features is created and released to examine and visualize differential cross-scale attention; (3) Superior performance on both in-house and public datasets is demonstrated by our simple cross-scale MIL strategy. The official implementation is publicly available at https://github.com/hrlblab/CS-MIL.

Gastric coinfection with thiopeptide-positive Cutibacterium acnes decreases FOXM1 and pro-inflammatory biomarker expression in a murine model of Helicobacter pylori-induced gastric cancer.

IMPORTANCE: H. pylori infects half of the world population and is the leading cause of gastric cancer. We previously demonstrated that gastric cancer risk is associated with gastric microbiota. Specifically, gastric urease-positive Staphylococcus epidermidis and Streptococcus salivarius had contrasting effects on H. pylori-associated gastric pathology and immune responses in germ-free INS-GAS mice. As gastritis progresses to gastric cancer, the oncogenic transcription factor Foxm1 becomes increasingly expressed. In this study, we evaluated the gastric commensal C. acnes, certain strains of which produce thiopeptides that directly inhibit FOXM1. Thiopeptide-positive C. acnes was isolated from Nicaraguan patient gastric biopsies and inoculated into germ-free INS-GAS mice with H. pylori. We, therefore, asked whether coinfection with C. acnes expressing thiopeptide and H. pylori would decrease gastric Foxm1 expression and pro-inflammatory cytokine mRNA and protein levels. Our study supports the growing literature that specific non-H. pylori gastric bacteria affect inflammatory and cancer biomarkers in H. pylori pathogenesis.

BRD4 Regulates Glycolysis-Dependent Nos2 Expression in Macrophages Upon H pylori Infection.

BACKGROUND & AIMS: Metabolic reprogramming is essential for the activation and functions of macrophages, including bacterial killing and cytokine production. Bromodomain-containing protein 4 (BRD4) has emerged as a critical regulator of innate immune response. However, the potential role of BRD4 in the metabolic reprogramming of macrophage activation upon Helicobacter pylori infection remains unclear. METHODS: Bone marrow-derived macrophages (BMDMs) from wild-type (WT) and Brd4-myeloid deletion conditional knockout (Brd4-CKO) mice were infected with H pylori. RNA sequencing was performed to evaluate the differential gene expression between WT and Brd4-deficient BMDMs upon infection. An in vivo model of H pylori infection using WT and Brd4-CKO mice was used to confirm the role of BRD4 in innate immune response to infection. RESULTS: Depletion of Brd4 in BMDMs showed impaired H pylori-induced glycolysis. In addition, H pylori-induced expression of glycolytic genes, including Slc2a1 and Hk2, was decreased in Brd4-deficient BMDMs. BRD4 was recruited to the promoters of Slc2a1 and Hk2 via hypoxia-inducible factor-1α, facilitating their expression. BRD4-mediated glycolysis stabilized H pylori-induced nitric oxide synthase (Nos2) messenger RNA to produce nitric oxide. The NO-mediated killing of H pylori decreased in Brd4-deficient BMDMs, which was rescued by pyruvate. Furthermore, Brd4-CKO mice infected with H pylori showed reduced gastric inflammation and increased H pylori colonization with reduced inducible NO synthase expression in gastric macrophages. CONCLUSIONS: Our study identified BRD4 as a key regulator of hypoxia-inducible factor-1α-dependent glycolysis and macrophage activation. Furthermore, we show a novel regulatory role of BRD4 in innate immunity through glycolysis to stabilize Nos2 messenger RNA for NO production to eliminate H pylori infection.

Myeloid deletion of talin-1 reduces mucosal macrophages and protects mice from colonic inflammation.

The intestinal immune response is crucial in maintaining a healthy gut, but the enhanced migration of macrophages in response to pathogens is a major contributor to disease pathogenesis. Integrins are ubiquitously expressed cellular receptors that are highly involved in immune cell adhesion to endothelial cells while in the circulation and help facilitate extravasation into tissues. Here we show that specific deletion of the Tln1 gene encoding the protein talin-1, an integrin-activating scaffold protein, from cells of the myeloid lineage using the Lyz2-cre driver mouse reduces epithelial damage, attenuates colitis, downregulates the expression of macrophage markers, decreases the number of differentiated colonic mucosal macrophages, and diminishes the presence of CD68-positive cells in the colonic mucosa of mice infected with the enteric pathogen Citrobacter rodentium. Bone marrow-derived macrophages lacking expression of Tln1 did not exhibit a cell-autonomous phenotype; there was no impaired proinflammatory gene expression, nitric oxide production, phagocytic ability, or surface expression of CD11b, CD86, or major histocompatibility complex II in response to C. rodentium. Thus, we demonstrate that talin-1 plays a role in the manifestation of infectious colitis by increasing mucosal macrophages, with an effect that is independent of macrophage activation.

Topological-Preserving Membrane Skeleton Segmentation in Multiplex Immunofluorescence Imaging.

Multiplex immunofluorescence (MxIF) is an emerging imaging technology whose downstream molecular analytics highly rely upon the effectiveness of cell segmentation. In practice, multiple membrane markers (e.g., NaKATPase, PanCK and ß-catenin) are employed to stain membranes for different cell types, so as to achieve a more comprehensive cell segmentation since no single marker fits all cell types. However, prevalent watershed-based image processing might yield inferior capability for modeling complicated relationships between markers. For example, some markers can be misleading due to questionable stain quality. In this paper, we propose a deep learning based membrane segmentation method to aggregate complementary information that is uniquely provided by large scale MxIF markers. We aim to segment tubular membrane structure in MxIF data using global (membrane markers z-stack projection image) and local (separate individual markers) information to maximize topology preservation with deep learning. Specifically, we investigate the feasibility of four SOTA 2D deep networks and four volumetric-based loss functions. We conducted a comprehensive ablation study to assess the sensitivity of the proposed method with various combinations of input channels. Beyond using adjusted rand index (ARI) as the evaluation metric, which was inspired by the clDice, we propose a novel volumetric metric that is specific for skeletal structure, denoted as clDiceSKEL. In total, 80 membrane MxIF images were manually traced for 5-fold cross-validation. Our model outperforms the baseline with a 20.2% and 41.3% increase in clDiceSKEL and ARI performance, which is significant (p<0.05) using the Wilcoxon signed rank test. Our work explores a promising direction for advancing MxIF imaging cell segmentation with deep learning membrane segmentation. Tools are available at https://github.com/MASILab/MxIF_Membrane_Segmentation.

A Specialized Epithelial Cell Type Regulating Mucosal Immunity and Driving Human Crohn's Disease.

Crohn's disease (CD) is a complex chronic inflammatory disorder that may affect any part of gastrointestinal tract with extra-intestinal manifestations and associated immune dysregulation. To characterize heterogeneity in CD, we profiled single-cell transcriptomics of 170 samples from 65 CD patients and 18 non-inflammatory bowel disease (IBD) controls in both the terminal ileum (TI) and ascending colon (AC). Analysis of 202,359 cells identified a novel epithelial cell type in both TI and AC, featuring high expression of LCN2, NOS2, and DUOX2, and thus is named LND. LND cells, confirmed by high-resolution in-situ RNA imaging, were rarely found in non-IBD controls, but expanded significantly in active CD. Compared to other epithelial cells, genes defining LND cells were enriched in antimicrobial response and immunoregulation. Moreover, multiplexed protein imaging demonstrated that LND cell abundance was associated with immune infiltration. Cross-talk between LND and immune cells was explored by ligand-receptor interactions and further evidenced by their spatial colocalization. LND cells showed significant enrichment of expression specificity of IBD/CD susceptibility genes, revealing its role in immunopathogenesis of CD. Investigating lineage relationships of epithelial cells detected two LND cell subpopulations with different origins and developmental potential, early and late LND. The ratio of the late to early LND cells was related to anti-TNF response. These findings emphasize the pathogenic role of the specialized LND cell type in both Crohn's ileitis and Crohn's colitis and identify novel biomarkers associated with disease activity and treatment response.

Turicibacterales protect mice from severe Citrobacter rodentium infection.

One of the major contributors to child mortality in the world is diarrheal diseases, with an estimated 800,000 deaths per year. Many pathogens are causative agents of these illnesses, including the enteropathogenic or enterohemorrhagic forms of Escherichia coli. These bacteria are characterized by their ability to cause attaching and effacing lesions in the gut mucosa. Although much has been learned about the pathogenicity of these organisms and the immune response against them, the role of the intestinal microbiota during these infections is not well characterized. Infection of mice with E. coli requires pre-treatment with antibiotics in most mouse models, which hinders the study of the microbiota in an undisturbed environment. Using Citrobacter rodentium as a murine model for attaching and effacing bacteria, we show that C57BL/6 mice deficient in granzyme B expression are highly susceptible to severe disease caused by C. rodentium infection. Although a previous publication from our group shows that granzyme B-deficient CD4+ T cells are partially responsible for this phenotype, in this report, we present data demonstrating that the microbiota, in particular members of the order Turicibacterales, have an important role in conferring resistance. Mice deficient in Turicibacter sanguinis have increased susceptibility to severe disease. However, when these mice are co-housed with resistant mice or colonized with T. sanguinis, susceptibility to severe infection is reduced. These results clearly suggest a critical role for this commensal in the protection against enteropathogens.

Pancreatic islet α cell function and proliferation requires the arginine transporter SLC7A2.

Interrupting glucagon signaling decreases gluconeogenesis and the fractional extraction of amino acids by liver from blood resulting in lower glycemia. The resulting hyperaminoacidemia stimulates α cell proliferation and glucagon secretion via a liver-α cell axis. We hypothesized that α cells detect and respond to circulating amino acids levels via a unique amino acid transporter repertoire. We found that Slc7a2ISLC7A2 is the most highly expressed cationic amino acid transporter in α cells with its expression being three-fold greater in α than ß cells in both mouse and human. Employing cell culture, zebrafish, and knockout mouse models, we found that the cationic amino acid arginine and SLC7A2 are required for α cell proliferation in response to interrupted glucagon signaling. Ex vivo and in vivo assessment of islet function in Slc7a2-/- mice showed decreased arginine-stimulated glucagon and insulin secretion. We found that arginine activation of mTOR signaling and induction of the glutamine transporter SLC38A5 was dependent on SLC7A2, showing that both's role in α cell proliferation is dependent on arginine transport and SLC7A2. Finally, we identified single nucleotide polymorphisms in SLC7A2 associated with HbA1c. Together, these data indicate a central role for SLC7A2 in amino acid-stimulated α cell proliferation and islet hormone secretion.

Predicting Crohn's disease severity in the colon using mixed cell nucleus density from pseudo labels.

Crohn's disease (CD) is a debilitating inflammatory bowel disease with no known cure. Computational analysis of hematoxylin and eosin (H&E) stained colon biopsy whole slide images (WSIs) from CD patients provides the opportunity to discover unknown and complex relationships between tissue cellular features and disease severity. While there have been works using cell nuclei-derived features for predicting slide-level traits, this has not been performed on CD H&E WSIs for classifying normal tissue from CD patients vs active CD and assessing slide label-predictive performance while using both separate and combined information from pseudo-segmentation labels of nuclei from neutrophils, eosinophils, epithelial cells, lymphocytes, plasma cells, and connective cells. We used 413 WSIs of CD patient biopsies and calculated normalized histograms of nucleus density for the six cell classes for each WSI. We used a support vector machine to classify the truncated singular value decomposition representations of the normalized histograms as normal or active CD with four-fold cross-validation in rounds where nucleus types were first compared individually, the best was selected, and further types were added each round. We found that neutrophils were the most predictive individual nucleus type, with an AUC of 0.92 ± 0.0003 on the withheld test set. Adding information improved cross-validation performance for the first two rounds and on the withheld test set for the first three rounds, though performance metrics did not increase substantially beyond when neutrophils were used alone.

Hypusination Maintains Intestinal Homeostasis and Prevents Colitis and Carcinogenesis by Enhancing Aldehyde Detoxification.

BACKGROUND & AIMS: The amino acid hypusine, synthesized from the polyamine spermidine by the enzyme deoxyhypusine synthase (DHPS), is essential for the activity of eukaryotic translation initiation factor 5A (EIF5A). The role of hypusinated EIF5A (EIF5AHyp) remains unknown in intestinal homeostasis. Our aim was to investigate EIF5AHyp in the gut epithelium in inflammation and carcinogenesis. METHODS: We used human colon tissue messenger RNA samples and publicly available transcriptomic datasets, tissue microarrays, and patient-derived colon organoids. Mice with intestinal epithelial-specific deletion of Dhps were investigated at baseline and in models of colitis and colon carcinogenesis. RESULTS: We found that patients with ulcerative colitis and Crohn's disease exhibit reduced colon levels of DHPS messenger RNA and DHPS protein and reduced levels of EIF5AHyp. Similarly, colonic organoids from colitis patients also show down-regulated DHPS expression. Mice with intestinal epithelial-specific deletion of Dhps develop spontaneous colon hyperplasia, epithelial proliferation, crypt distortion, and inflammation. Furthermore, these mice are highly susceptible to experimental colitis and show exacerbated colon tumorigenesis when treated with a carcinogen. Transcriptomic and proteomic analysis on colonic epithelial cells demonstrated that loss of hypusination induces multiple pathways related to cancer and immune response. Moreover, we found that hypusination enhances translation of numerous enzymes involved in aldehyde detoxification, including glutathione S-transferases and aldehyde dehydrogenases. Accordingly, hypusination-deficient mice exhibit increased levels of aldehyde adducts in the colon, and their treatment with a scavenger of electrophiles reduces colitis. CONCLUSIONS: Hypusination in intestinal epithelial cells has a key role in the prevention of colitis and colorectal cancer, and enhancement of this pathway via supplementation of spermidine could have a therapeutic impact.

Deep multimodal fusion of image and non-image data in disease diagnosis and prognosis: a review.

The rapid development of diagnostic technologies in healthcare is leading to higher requirements for physicians to handle and integrate the heterogeneous, yet complementary data that are produced during routine practice. For instance, the personalized diagnosis and treatment planning for a single cancer patient relies on various images (e.g. radiology, pathology and camera images) and non-image data (e.g. clinical data and genomic data). However, such decision-making procedures can be subjective, qualitative, and have large inter-subject variabilities. With the recent advances in multimodal deep learning technologies, an increasingly large number of efforts have been devoted to a key question: how do we extract and aggregate multimodal information to ultimately provide more objective, quantitative computer-aided clinical decision making? This paper reviews the recent studies on dealing with such a question. Briefly, this review will include the (a) overview of current multimodal learning workflows, (b) summarization of multimodal fusion methods, (c) discussion of the performance, (d) applications in disease diagnosis and prognosis, and (e) challenges and future directions.

Influence of Cell-type Ratio on Spatially Resolved Single-cell Transcriptomes using the Tangram Algorithm: Based on Implementation on Single-Cell and MxIF Data.

The Tangram algorithm is a benchmarking method of aligning single-cell (sc/snRNA-seq) data to various forms of spatial data collected from the same region. With this data alignment, the annotation of the single-cell data can be projected to spatial data. However, the cell composition (cell-type ratio) of the single-cell data and spatial data might be different because of heterogeneous cell distribution. Whether the Tangram algorithm can be adapted when the two data have different cell-type ratios has not been discussed in previous works. In our practical application that maps the cell-type classification results of single-cell data to the Multiplex immunofluorescence (MxIF) spatial data, cell-type ratios were different, though they were sampled from adjacent areas. In this work, both simulation and empirical validation were conducted to quantitatively explore the impact of the mismatched cell-type ratio on the Tangram mapping in different situations. Results show that the cell-type difference has a negative influence on classification accuracy.

Turicibacterales protect mice from severe Citrobacter rodentium infection.

One of the major contributors to child mortality in the world is diarrheal diseases, with an estimated 800,000 deaths per year. Many pathogens are causative agents of these illnesses, including the enteropathogenic (EPEC) or enterohemorrhagic (EHEC) forms of Escherichia coli. These bacteria are characterized by their ability to cause attaching and effacing lesions in the gut mucosa. Although much has been learned about the pathogenicity of these organisms and the immune response against them, the role of the intestinal microbiota during these infections is not well characterized. Infection of mice with E. coli requires pre-treatment with antibiotics in most mouse models, which hinders the study of the microbiota in an undisturbed environment. Using Citrobacter rodentium as a murine model for attaching and effacing bacteria, we show that C57BL/6 mice deficient in granzyme B expression are highly susceptible to severe disease caused by C. rodentium infection. Although a previous publication from our group shows that granzyme B-deficient CD4+ T cells are partially responsible for this phenotype, in this report we present data demonstrating that the microbiota, in particular members of the order Turicibacterales, have an important role in conferring resistance. Mice deficient in Turicibacter sanguinis have increased susceptibility to severe disease. However, when these mice are co-housed with resistant mice, or colonized with T. sanguinis, susceptibility to severe infection is reduced. These results clearly suggest a critical role for this commensal in the protection against entero-pathogens.

A Roadmap for the Human Gut Cell Atlas.

The number of studies investigating the human gastrointestinal tract using various single-cell profiling methods has increased substantially in the past few years. Although this increase provides a unique opportunity for the generation of the first comprehensive Human Gut Cell Atlas (HGCA), there remains a range of major challenges ahead. Above all, the ultimate success will largely depend on a structured and coordinated approach that aligns global efforts undertaken by a large number of research groups. In this Roadmap, we discuss a comprehensive forward-thinking direction for the generation of the HGCA on behalf of the Gut Biological Network of the Human Cell Atlas. Based on the consensus opinion of experts from across the globe, we outline the main requirements for the first complete HGCA by summarizing existing data sets and highlighting anatomical regions and/or tissues with limited coverage. We provide recommendations for future studies and discuss key methodologies and the importance of integrating the healthy gut atlas with related diseases and gut organoids. Importantly, we critically overview the computational tools available and provide recommendations to overcome key challenges.

Electrophilic reactive aldehydes as a therapeutic target in colorectal cancer prevention and treatment.

Colorectal cancer (CRC) is a major health problem worldwide. Dicarbonyl electrophiles, such as isolevuglandins (isoLGs), are generated from lipid peroxidation and form covalent adducts with amine-containing macromolecules. We have shown high levels of adducts of isoLGs in colonic epithelial cells of patients with CRC. We thus investigated the role of these reactive aldehydes in colorectal cancer development. We found that 2-hydroxybenzylamine (2-HOBA), a natural compound derived from buckwheat seeds that acts as a potent scavenger of electrophiles, is bioavailable in the colon of mice after supplementation in the drinking water and does not affect the colonic microbiome. 2-HOBA reduced the level of isoLG adducts to lysine as well as tumorigenesis in models of colitis-associated carcinogenesis and of sporadic CRC driven by specific deletion of the adenomatous polyposis coli gene in colonic epithelial cells. In parallel, we found that oncogenic NRF2 activation and signaling were decreased in the colon of 2-HOBA-treated mice. Additionally, the growth of xenografted human HCT116 CRC cells in nude mice was significantly attenuated by 2-HOBA supplementation. In conclusion, 2-HOBA represents a promising natural compound for the prevention and treatment of CRC.

Shotgun Metagenomics of Gastric Biopsies Reveals Compositional and Functional Microbiome Shifts in High- and Low-Gastric-Cancer-Risk Populations from Colombia, South America.

Along with Helicobacter pylori infection, the gastric microbiota is hypothesized to modulate stomach cancer risk in susceptible individuals. Whole metagenomic shotgun sequencing (WMS) is a sequencing approach to characterize the microbiome with advantages over traditional culture and 16S rRNA sequencing including identification of bacterial and non-bacterial taxa, species/strain resolution, and functional characterization of the microbiota. In this study, we used WMS to survey the microbiome in extracted DNA from antral gastric biopsy samples from Colombian patients residing in the high-risk gastric cancer town Túquerres (n = 10, H. pylori-positive = 7) and low-risk town of Tumaco (n = 10, H. pylori-positive = 6). Kraken2/Bracken was used for taxonomic classification and abundance. Functional gene profiles were inferred by InterProScan and KEGG analysis of assembled contigs and gene annotation. The most abundant taxa represented bacteria, non-human eukaryota, and viral genera found in skin, oral, food, and plant/soil environments including Staphylococus, Streptococcus, Bacillus, Aspergillus, and Siphoviridae. H. pylori was the predominant taxa present in H. pylori-positive samples. Beta diversity was significantly different based on H. pylori-status, risk group, and sex. WMS detected more bacterial taxa than 16S rRNA sequencing and aerobic, anaerobic, and microaerobic culture performed on the same gastric biopsy samples. WMS identified significant differences in functional profiles found between H. pylori-status, but not risk or sex groups. H. pylori-positive samples were significantly enriched for H. pylori-specific genes including virulence factors such as vacA, cagA, and urease, while carbohydrate and amino acid metabolism genes were enriched in H. pylori-negative samples. This study shows WMS has the potential to characterize the taxonomy and function of the gastric microbiome as risk factors for H. pylori-associated gastric disease. Future studies will be needed to compare and validate WMS versus traditional culture and 16S rRNA sequencing approaches for characterization of the gastric microbiome.

Epithelial talin-1 protects mice from citrobacter rodentium-induced colitis by restricting bacterial crypt intrusion and enhancing t cell immunity.

Pathogenic enteric Escherichia coli present a significant burden to global health. Food-borne enteropathogenic E. coli (EPEC) and Shiga toxin-producing E. coli (STEC) utilize attaching and effacing (A/E) lesions and actin-dense pedestal formation to colonize the gastrointestinal tract. Talin-1 is a large structural protein that links the actin cytoskeleton to the extracellular matrix though direct influence on integrins. Here we show that mice lacking talin-1 in intestinal epithelial cells (Tln1Δepi) have heightened susceptibility to colonic disease caused by the A/E murine pathogen Citrobacter rodentium. Tln1Δepi mice exhibit decreased survival, and increased colonization, colon weight, and histologic colitis compared to littermate Tln1fl/fl controls. These findings were associated with decreased actin polymerization and increased infiltration of innate myeloperoxidase-expressing immune cells, confirmed as neutrophils by flow cytometry, but more bacterial dissemination deep into colonic crypts. Further evaluation of the immune population recruited to the mucosa in response to C. rodentium revealed that loss of Tln1 in colonic epithelial cells (CECs) results in impaired recruitment and activation of T cells. C. rodentium infection-induced colonic mucosal hyperplasia was exacerbated in Tln1Δepi mice compared to littermate controls. We demonstrate that this is associated with decreased CEC apoptosis and crowding of proliferating cells in the base of the glands. Taken together, talin-1 expression by CECs is important in the regulation of both epithelial renewal and the inflammatory T cell response in the setting of colitis caused by C. rodentium, suggesting that this protein functions in CECs to limit, rather than contribute to the pathogenesis of this enteric infection.

Interleukin-23 receptor signaling impairs the stability and function of colonic regulatory T cells.

The cytokine interleukin-23 (IL-23) is involved in the pathogenesis of inflammatory and autoimmune conditions including inflammatory bowel disease (IBD). IL23R is enriched in intestinal Tregs, yet whether IL-23 modulates intestinal Tregs remains unknown. Here, investigating IL-23R signaling in Tregs specifically, we show that colonic Tregs highly express Il23r compared with Tregs from other compartments and their frequency is reduced upon IL-23 administration and impairs Treg suppressive function. Similarly, colonic Treg frequency is increased in mice lacking Il23r specifically in Tregs and exhibits a competitive advantage over IL-23R-sufficient Tregs during inflammation. Finally, IL-23 antagonizes liver X receptor pathway, cellular cholesterol transporter Abca1, and increases Treg apoptosis. Our results show that IL-23R signaling regulates intestinal Tregs by increasing cell turnover, antagonizing suppression, and decreasing cholesterol efflux. These results suggest that IL-23 negatively regulates Tregs in the intestine with potential implications for promoting chronic inflammation in patients with IBD.

The nutraceutical electrophile scavenger 2-hydroxybenzylamine (2-HOBA) attenuates gastric cancer development caused by Helicobacter pylori.

Stomach cancer is a leading cause of cancer death. Helicobacter pylori is a bacterial gastric pathogen that is the primary risk factor for carcinogenesis, associated with its induction of inflammation and DNA damage. Dicarbonyl electrophiles are generated from lipid peroxidation during the inflammatory response and form covalent adducts with amine-containing macromolecules. 2-hydroxybenzylamine (2-HOBA) is a natural compound derived from buckwheat seeds and acts as a potent scavenger of reactive aldehydes. Our goal was to investigate the effect of 2-HOBA on the pathogenesis of H. pylori infection. We used transgenic FVB/N insulin-gastrin (INS-GAS) mice as a model of gastric cancer. First, we found that 2-HOBA is bioavailable in the gastric tissues of these mice after supplementation in the drinking water. Moreover, 2-HOBA reduced the development of gastritis in H. pylori-infected INS-GAS mice without affecting the bacterial colonization level in the stomach. Further, we show that the development of gastric dysplasia and carcinoma was significantly reduced by 2-HOBA. Concomitantly, DNA damage were also inhibited by 2-HOBA treatment in H. pylori-infected mice. In parallel, DNA damage was inhibited by 2-HOBA in H. pylori-infected gastric epithelial cells in vitro. In conclusion, 2-HOBA, which has been shown to be safe in human clinical trials, represents a promising nutritional compound for the chemoprevention of the more severe effects of H. pylori infection.

Feasibility of Universal Anomaly Detection without Knowing the Abnormality in Medical Images.

Many anomaly detection approaches, especially deep learning methods, have been recently developed to identify abnormal image morphology by only employing normal images during training. Unfortunately, many prior anomaly detection methods were optimized for a specific "known" abnormality (e.g., brain tumor, bone fraction, cell types). Moreover, even though only the normal images were used in the training process, the abnormal images were often employed during the validation process (e.g., epoch selection, hyper-parameter tuning), which might leak the supposed "unknown" abnormality unintentionally. In this study, we investigated these two essential aspects regarding universal anomaly detection in medical images by (1) comparing various anomaly detection methods across four medical datasets, (2) investigating the inevitable but often neglected issues on how to unbiasedly select the optimal anomaly detection model during the validation phase using only normal images, and (3) proposing a simple decision-level ensemble method to leverage the advantage of different kinds of anomaly detection without knowing the abnormality. The results of our experiments indicate that none of the evaluated methods consistently achieved the best performance across all datasets. Our proposed method enhanced the robustness of performance in general (average AUC 0.956).