The purpose of this study was to develop a computer-aided diagnosis (CAD) system for automatic classification of histopathological images of lung tissues. Two datasets (private and public datasets) were obtained and used for developing and validating CAD. The private dataset consists of 94 histopathological images that were obtained for the following five categories: normal, emphysema, atypical adenomatous hyperplasia, lepidic pattern of adenocarcinoma, and invasive adenocarcinoma. The public dataset consists of 15,000 histopathological images that were obtained for the following three categories: lung adenocarcinoma, lung squamous cell carcinoma, and benign lung tissue. These images were automatically classified using machine learning and two types of image feature extraction: conventional texture analysis (TA) and homology-based image processing (HI). Multiscale analysis was used in the image feature extraction, after which automatic classification was performed using the image features and eight machine learning algorithms. The multicategory accuracy of our CAD system was evaluated in the two datasets. In both the public and private datasets, the CAD system with HI was better than that with TA. It was possible to build an accurate CAD system for lung tissues. HI was more useful for the CAD systems than TA.

Tumor-associated macrophages (TAMs) contribute to the progression and mortality of various malignancies. We reported that high numbers of infiltrating TAMs were significantly associated with tumor progression and poor prognosis in esophageal squamous cell carcinoma (ESCC). In our previous investigation of TAMs’ actions in ESCC, we compared gene expression profiles between peripheral blood monocyte (PBMo)-derived macrophages and TAM-like macrophages stimulated with conditioned media of ESCC cell lines. Among the upregulated genes in the TAM-like macrophages, we focused on CC chemokine ligand 3 (CCL3), which was reported to contribute to tumor progression in several malignancies. Herein, we observed that not only TAMs but also ESCC cell lines expressed CCL3. A CCL3 receptor, CC chemokine receptor 5 (CCR5) was expressed in the ESCC cell lines. Treating the ESCC cell lines with recombinant human (rh)CCL3 induced the phosphorylations of Akt and ERK, which were suppressed by CCR5 knockdown. Migration and invasion of ESCC cells were promoted by treatment with rhCCL3 and co-culture with TAMs. TAMs/rhCCL3-promoted cell migration and invasion were suppressed by inhibition of the CCL3–CCR5 axis, PI3K/Akt, and MEK/ERK pathways. Treatment with rhCCL3 upregulated MMP2 and VEGFA expressions in ESCC cell lines. Our immunohistochemical analysis of 68 resected ESCC cases showed that high expression of CCL3 and/or CCR5 in ESCC tissues was associated with poor prognosis. High CCR5 expression was associated with deeper invasion, presence of vascular invasion, higher pathological stage, higher numbers of infiltrating CD204⁺ TAMs, and higher microvascular density. High expression of both CCL3 and CCR5 was an independent prognostic factor for disease-free survival. These results suggest that CCL3 derived from both TAMs and cancer cells contributes to the progression and poor prognosis of ESCC by promoting cell migration and invasion via the binding of CCR5 and the phosphorylations of Akt and ERK. The CCL3–CCR5 axis could become the target of new therapies against ESCC.

Cancer-associated fibroblasts (CAFs) contribute to the progression of various cancers. Previously, we reported the significance of CAFs in esophageal squamous cell carcinoma (ESCC); however, the functions of CAFs in the ESCC microenvironment remain unknown. To investigate CAFs’ function, we established an indirect coculture assay between human bone marrow-derived mesenchymal stem cells (MSCs) and ESCC cells. Cocultured MSCs expressed more fibroblast activation protein, one of the markers of CAFs, compared with monocultured MSCs. Therefore, we defined cocultured MSCs as CAF-like cells. To identify molecules associated with the ESCC progression in CAFs, we conducted a cDNA microarray analysis on monocultured MSCs and CAF-like cells to compare their gene expression profiles. We found that SERPINE1, which encodes plasminogen activator inhibitor-1 (PAI-1), was more abundant in CAF-like cells than in monocultured MSCs, and the PAI-1 derived from CAF-like cells induced the abilities of migration and invasion in both ESCC cells and macrophages by the Akt and Erk1/2 signaling pathways via the low-density lipoprotein receptor-related protein 1 (LRP1), which is a PAI-1 receptor. Based on immunohistochemistry assays of ESCC tissues, higher expression levels of PAI-1 and LRP1 were correlated with poor prognosis in ESCC patients. These results suggest that the PAI-1/LRP1 axis contributes to the progression of ESCC, making it a potential target for ESCC therapy.