Background: Bone marrow aspiration and biopsy remain the gold standard for the diagnosis of hematological diseases despite the development of flow cytometry (FCM) and molecular and gene analyses. However, the interpretation of the results is laborious and operator dependent. Furthermore, the obtained results exhibit inter- and intravariations among specialists. Therefore, it is important to develop a more objective and automated analysis system. Several deep learning models have been developed and applied in medical image analysis but not in the field of hematological histology, especially for bone marrow smear applications. Objective: The aim of this study was to develop a deep learning model (BMSNet) for assisting hematologists in the interpretation of bone marrow smears for faster diagnosis and disease monitoring. Methods: From January 1, 2016, to December 31, 2018, 122 bone marrow smears were photographed and divided into a development cohort (N=42), a validation cohort (N=70), and a competition cohort (N=10). The development cohort included 17,319 annotated cells from 291 high-resolution photos. In total, 20 photos were taken for each patient in the validation cohort and the competition cohort. This study included eight annotation categories: erythroid, blasts, myeloid, lymphoid, plasma cells, monocyte, megakaryocyte, and unable to identify. BMSNet is a convolutional neural network with the YOLO v3 architecture, which detects and classifies single cells in a single model. Six visiting staff members participated in a human-machine competition, and the results from the FCM were regarded as the ground truth. Results: In the development cohort, according to 6-fold cross-validation, the average precision of the bounding box prediction without consideration of the classification is 67.4%. After removing the bounding box prediction error, the precision and recall of BMSNet were similar to those of the hematologists in most categories. In detecting more than 5% of blasts in the validation cohort, the area under the curve (AUC) of BMSNet (0.948) was higher than the AUC of the hematologists (0.929) but lower than the AUC of the pathologists (0.985). In detecting more than 20% of blasts, the AUCs of the hematologists (0.981) and pathologists (0.980) were similar and were higher than the AUC of BMSNet (0.942). Further analysis showed that the performance difference could be attributed to the myelodysplastic syndrome cases. In the competition cohort, the mean value of the correlations between BMSNet and FCM was 0.960, and the mean values of the correlations between the visiting staff and FCM ranged between 0.952 and 0.990. Conclusions: Our deep learning model can assist hematologists in interpreting bone marrow smears by facilitating and accelerating the detection of hematopoietic cells. However, a detailed morphological interpretation still requires trained hematologists.

Cortactin and fascin-1 are important factors in tumor progression. We tested the hypothesis that cortactin and fascin-1 expression correlates with clinicopathological parameters of gastric adenocarcinoma. Immunohistochemical analysis of cortactin and fascin-1 was done using tissue microarrays of 100 surgical specimens, including 20 well-differentiated, 20 moderately differentiated, and 60 poorly differentiated gastric adenocarcinomas. Among the 20 well-differentiated gastric adenocarcinomas, 15 cases (75%) showed negative or weak staining (1+); 5 cases (25%) had moderate (2+) or strong (3+) cortactin expression. Among the 60 poorly differentiated gastric adenocarcinomas, more than three-quarters of the cases (76.7%) had moderate or strong cortactin expression; 14 cases (23.3%) had weak staining. Of 20 well-differentiated gastric adenocarcinoma cases, 14 (70%) showed negative or weak staining of fascin-1, whereas nearly one-third (30%) had moderate or strong expression. Among the 60 poorly differentiated gastric adenocarcinomas, 32 (53.3%) exhibited moderate or strong fascin-1 expression; fewer than half of the cases showed negative or weak staining. Higher intensity of cortactin and fascin-1 staining correlated directly with more-advanced cancer stages (TNM) and inversely with survival rates. Our findings suggest the possibility that pharmacological inhibitors of cortactin and fascin-1 activity may slow down tumor progression and prolong survival time in patients with gastric adenocarcinomas.

Glioblastoma (GBM) is a fatal cancer. Existing therapies do not have significant efficacy for GBM patients. Previous studies have shown that the collagen family is involved in the regulation of the extracellular environment of cancer cells, and these conditions could become an important factor for effective treatment. Therefore, we screened various collagen types and observed that the type V collagen α1 chain (COL5A1) gene plays a pivotal role in GBM. We further examined whether the overexpression of COL5A1 is common in mesenchymal subtypes and is related to the survival rate of GBM patients through several in silico cohorts. In addition, our cohort also showed a consistent trend in COL5A1 protein levels. Most importantly, we validated the cell mobility, metastatic ability and actin polymerization status caused by COL5A1 with two-way models. Based on these results, we established a transcriptomics dataset based on COL5A1. Moreover, PPRC1, GK and ESM1 were predicted by ingenuity pathway analysis (IPA) to be transcription factors or to participate downstream. We investigated the involvement of COL5A1 in extracellular remodeling and the regulation of actin filaments in the metastasis of GBM. Our results indicate that the COL5A1−PPRC1−ESM1 axis may represent a novel therapeutic target in GBM.

Bladder cancer (BC) has a high recurrence rate worldwide. The aim of this study was to evaluate the role of fatty acid binding protein 6 (FABP6) in proliferation and migration in human bladder cancer cells. Cell growth was confirmed by MTT and colony formation assay. Western blotting was used to explore protein expressions. Wound healing and Transwell assays were performed to evaluate the migration ability. A xenograft animal model with subcutaneous implantation of BC cells was generated to confirm the tumor progression. Knockdown of FABP6 reduced cell growth in low-grade TSGH-8301 and high-grade T24 cells. Cell cycle blockade was observed with the decrease of CDK2, CDK4, and Ki67 levels in FABP6-knockdown BC cells. Interestingly, knockdown of FBAP6 led to downregulation of autophagic markers and activation of AKT-mTOR signaling. The application of PI3K/AKT inhibitor decreased cell viability mediated by FABP6-knockdown additionally. Moreover, FABP6-knockdown reduced peroxisome proliferator-activated receptor γ and retinoid X receptor α levels but increased p-p65 expression. Knockdown of FABP6 also inhibited BC cell motility with focal adhesive complex reduction. Finally, shFABP6 combined with cisplatin suppressed tumor growth in vivo. These results provide evidence that FABP6 may be a potential target in BC cells progression.

Acute lung injury (ALI) is characterised by severe pulmonary inflammation, alveolar-capillary barrier disruption, and pulmonary oedema. Therefore, establishing effective therapeutic targets for ALI prevention is crucial. The present study reports a novel function of RNF128 in regulating LPS-induced ALI. Severe lung damage and increased immune cell infiltration were detected in RNF128-deficient mice. In vitro experiments revealed that RNF128 inhibits neutrophil activation by binding to myeloperoxidase (MPO) and reducing its levels and activity. Moreover, RNF128 regulates alveolar macrophage activation and neutrophil infiltration by interacting with TLR4, targeting it for degradation, and inhibiting NF-κB activation, hence decreasing pro-inflammatory cytokines. Our results demonstrate for the first time that RNF128 is a negative regulator of MPO and TLR4 in neutrophils and alveolar macrophages, respectively. However, AAV9-mediated RNF128 overexpression alleviated lung tissue damage and reduced inflammatory cell infiltration. Thus, RNF128 is a promising therapeutic candidate for pharmacological interventions in ALI.

Glioblastoma, also known as glioblastoma multiforme (GBM), is a fast‐growing tumor and the most aggressive brain malignancy. Proteasome subunit beta type‐8 (PSMB8) is one of the 17 essential subunits for the complete assembly of the 20S proteasome complex. The aim of this study was to evaluate the role of PSMB8 expression in GBM progression and angiogenesis. PSMB8 expression in glioblastoma LN229 and U87MG were knockdown by siRNA or inducible shRNA both in vitro and in vivo. After PSMB8 reduction, cell survival, migration, invasion, angiogenesis and the related signaling cascades were evaluated. An orthotopic mouse tumor model was also provided to examine the angiogenesis within tumors. A GEO profile analysis indicated that high expression of PSMB8 mRNA in GBM patients was correlated with a low survival rate. In immunohistochemistry analysis, PSMB8 expression was higher in high‐grade than in low‐grade brain tumors. The proliferation, migration and angiogenesis of human GBM cells were decreased by PSMB8 knockdown in vitro. Furthermore, phosphorylated focal adhesion kinase (p‐FAK), p‐paxillin, MMP2, MMP9, and Cathepsin B were significantly reduced in LN229 cells. Integrin β1 and β3 were reduced in HUVECs after incubation with LN229‐conditioned medium. In an orthotopic mouse tumor model, inducible knockdown of PSMB8 reduced the expression of VEGF, VEGFR and CD31 and the progression of human glioblastoma. In this article, we demonstrated the role of PSMB8 in glioblastoma progression, especially neovascularization in vitro and in vivo. These results may provide a clue target for the anti‐angiogenic effect of PSMB8 in glioblastoma therapy in the future.