Evaluation of a Deep Learning Model for Metastatic Squamous Cell Carcinoma Prediction From Whole Slide Images

CONTEXT.—

Squamous cell carcinoma (SCC) is a histologic type of cancer that exhibits various degrees of keratinization. Identifying lymph node metastasis in SCC is crucial for prognosis and treatment strategies. Although artificial intelligence (AI) has shown promise in cancer prediction, applications specifically targeting SCC are limited.

OBJECTIVE.—

To design and validate a deep learning model tailored to predict metastatic SCC in radical lymph node dissection specimens, using whole slide images (WSIs).

DESIGN.—

Using the EfficientNetB1 architecture, a model was trained on 6587 WSIs (2413 SCC and 4174 nonneoplastic) from several hospitals, encompassing esophagus, head and neck, lung, and skin specimens. The training exclusively relied on WSI-level labels without annotations. We evaluated the model on a test set consisting of 541 WSIs (41 SCC and 500 nonneoplastic) of radical lymph node dissection specimens.

RESULTS.—

The model exhibited high performance, with receiver operating characteristic curve areas under the curve between 0.880 and 0.987 in detecting SCC metastases in lymph nodes. Although true positives and negatives were accurately identified, certain limitations were observed. These included false positives due to germinal centers, dust cell aggregations, and specimen-handling artifacts, as well as false negatives due to poor differentiation.

CONCLUSIONS.—

The developed artificial intelligence model presents significant potential in enhancing SCC lymph node detection, offering workload reduction for pathologists and increasing diagnostic efficiency. Continuous refinement is needed to overcome existing challenges, making the model more robust and clinically relevant.

Establishing protein expression profiles involved in tooth development using a proteomic approach

Various growth and transcription factors are involved in tooth development and developmental abnormalities; however, the protein dynamics do not always match the mRNA expression level. Using a proteomic approach, this study comprehensively analyzed protein expression in epithelial and mesenchymal tissues of the tooth germ during development. First molar tooth germs from embryonic day 14 and 16 Crlj:CD1 (ICR) mouse embryos were collected and separated into epithelial and mesenchymal tissues by laser microdissection. Mass spectrometry of the resulting proteins was carried out, and three types of highly expressed proteins [ATP synthase subunit beta (ATP5B), receptor of activated protein C kinase 1 (RACK1), and calreticulin (CALR)] were selected for immunohistochemical analysis. The expression profiles of these proteins were subsequently evaluated during all stages of amelogenesis using the continuously growing incisors of 3-week-old male ICR mice. Interestingly, these three proteins were specifically expressed depending on the stage of amelogenesis. RACK1 was highly expressed in dental epithelial and mesenchymal tissues during the proliferation and differentiation stages of odontogenesis, except for the pigmentation stage, whereas ATP5B and CALR immunoreactivity was weak in the enamel organ during the early stages, but became intense during the maturation and pigmentation stages, although the timing of the increased protein expression was different between the two. Overall, RACK1 plays an important role in maintaining the cell proliferation and differentiation in the apical end of incisors. In contrast, ATP5B and CALR are involved in the transport of minerals and the removal of organic materials as well as matrix deposition for CALR.

Deep Learning Approach to Classify Cutaneous Melanoma in a Whole Slide Image

Although the histopathological diagnosis of cutaneous melanocytic lesions is fairly accurate and reliable among experienced surgical pathologists, it is not perfect in every case (especially melanoma). Microscopic examination-clinicopathological correlation is the gold standard for the definitive diagnosis of melanoma. Pathologists may encounter diagnostic controversies when melanoma closely mimics Spitz's nevus or blue nevus, exhibits amelanotic histopathology, or is in situ. It would be beneficial if diagnosing cutaneous melanocytic lesions can be automated by using deep learning, particularly when assisting surgical pathologists with their workloads. In this preliminary study, we investigated the application of deep learning for classifying cutaneous melanoma in whole-slide images (WSIs). We trained models via weakly supervised learning using a dataset of 66 WSIs (33 melanomas and 33 non-melanomas). We evaluated the models on a test set of 90 WSIs (40 melanomas and 50 non-melanomas), achieving ROC-AUC at 0.821 for the WSI level and 0.936 for the tile level by the best model.

Editorial on Special Issue "Artificial Intelligence in Pathological Image Analysis"

The artificial intelligence (AI), especially deep learning models, is highly compatible with medical images and natural language processing and is expected to be applied to pathological image analysis and other medical fields [...].

Inference of core needle biopsy whole slide images requiring definitive therapy for prostate cancer

BACKGROUND

Prostate cancer is often a slowly progressive indolent disease. Unnecessary treatments from overdiagnosis are a significant concern, particularly low-grade disease. Active surveillance has being considered as a risk management strategy to avoid potential side effects by unnecessary radical treatment. In 2016, American Society of Clinical Oncology (ASCO) endorsed the Cancer Care Ontario (CCO) Clinical Practice Guideline on active surveillance for the management of localized prostate cancer.

METHODS

Based on this guideline, we developed a deep learning model to classify prostate adenocarcinoma into indolent (applicable for active surveillance) and aggressive (necessary for definitive therapy) on core needle biopsy whole slide images (WSIs). In this study, we trained deep learning models using a combination of transfer, weakly supervised, and fully supervised learning approaches using a dataset of core needle biopsy WSIs (n=1300). In addition, we performed an inter-rater reliability evaluation on the WSI classification.

RESULTS

We evaluated the models on a test set (n=645), achieving ROC-AUCs of 0.846 for indolent and 0.980 for aggressive. The inter-rater reliability evaluation showed s-scores in the range of 0.10 to 0.95, with the lowest being on the WSIs with both indolent and aggressive classification by the model, and the highest on benign WSIs.

CONCLUSION

The results demonstrate the promising potential of deployment in a practical prostate adenocarcinoma histopathological diagnostic workflow system.

Novel Applications of Artificial Intelligence in Cancer Research

Cutting-edge developments in machine learning and deep learning are improving all aspects of cancer research and treatment. Nowadays, the applications of artificial intelligence, machine learning, and deep learning to clinical aspects of cancer research have received more attention from scholars, with particular emphasis on diagnosis, prognosis, detection, and treatment.

Weakly Supervised Learning for Poorly Differentiated Adenocarcinoma Classification in GastricEndoscopic Submucosal Dissection Whole Slide Images

Objective: Endoscopic submucosal dissection (ESD) is the preferred technique for treating early gastric cancers including poorly differentiated adenocarcinoma without ulcerative findings. The histopathological classification of poorly differentiated adenocarcinoma including signet ring cell carcinoma is of pivotal importance for determining further optimum cancer treatment(s) and clinical outcomes. Because conventional diagnosis by pathologists using microscopes is time-consuming and limited in terms of human resources, it is very important to develop computer-aided techniques that can rapidly and accurately inspect large number of histopathological specimen whole-slide images (WSIs). Computational pathology applications which can assist pathologists in detecting and classifying gastric poorly differentiated adenocarcinoma from ESD WSIs would be of great benefit for routine histopathological diagnostic workflow. Methods: In this study, we trained the deep learning model to classify poorly differentiated adenocarcinoma in ESD WSIs by transfer and weakly supervised learning approaches. Results: We evaluated the model on ESD, endoscopic biopsy, and surgical specimen WSI test sets, achieving and ROC-AUC up to 0.975 in gastric ESD test sets for poorly differentiated adenocarcinoma. Conclusion: The deep learning model developed in this study demonstrates the high promising potential of deployment in a routine practical gastric ESD histopathological diagnostic workflow as a computer-aided diagnosis system.

Weakly supervised learning for multi-organ adenocarcinoma classification in whole slide images

The primary screening by automated computational pathology algorithms of the presence or absence of adenocarcinoma in biopsy specimens (e.g., endoscopic biopsy, transbronchial lung biopsy, and needle biopsy) of possible primary organs (e.g., stomach, colon, lung, and breast) and radical lymph node dissection specimen is very useful and should be a powerful tool to assist surgical pathologists in routine histopathological diagnostic workflow. In this paper, we trained multi-organ deep learning models to classify adenocarcinoma in biopsy and radical lymph node dissection specimens whole slide images (WSIs). We evaluated the models on five independent test sets (stomach, colon, lung, breast, lymph nodes) to demonstrate the feasibility in multi-organ and lymph nodes specimens from different medical institutions, achieving receiver operating characteristic areas under the curves (ROC-AUCs) in the range of 0.91 -0.98.

Diagnostic Categories and Key Features for Pathological Diagnosis of Endoscopic Ultrasound-Guided Fine Needle Aspiration Biopsy Samples of Pancreatic Lesions: A Consensus Study

OBJECTIVES

This study aimed to establish a reliable and reproducible categorized diagnostic classification system with identification of key features to achieve accurate pathological diagnosis of endoscopic ultrasound-guided fine needle aspiration biopsy (EUS-FNAB) samples of pancreatic lesions.

METHODS

Twelve pathologists examined virtual whole-slide images of EUS-FNAB samples obtained from 80 patients according to proposed diagnostic categories and key features for diagnosis. Fleiss κ was used to assess the concordance.

RESULTS

A hierarchical diagnostic system consisting of the following 6 diagnostic categories was proposed: inadequate, nonneoplasm, indeterminate, ductal carcinoma, nonductal neoplasm, and unclassified neoplasm. Adopting these categories, the average κ value of participants was 0.677 (substantial agreement). Among these categories, ductal carcinoma and nonductal neoplasm showed high κ values of 0.866 and 0.837, respectively, which indicated the almost perfect agreement. Key features identified for diagnosing ductal carcinoma were necrosis in low-power appearance; structural atypia/abnormalities recognized by irregular glandular contours, including cribriform and nonuniform shapes; cellular atypia, including enlarged nuclei, irregular nuclear contours, and foamy gland changes; and haphazard glandular arrangement and stromal desmoplasia.

CONCLUSIONS

The proposed hierarchical diagnostic classification system was proved to be useful for achieving reliable and reproducible diagnosis of EUS-FNAB specimens of pancreatic lesions based on evaluated histological features.

Deep learning models in medical image analysis

BACKGROUND

Deep learning is a state-of-the-art technology that has rapidly become the method of choice for medical image analysis. Its fast and robust object detection, segmentation, tracking, and classification of pathophysiological anatomical structures can support medical practitioners during routine clinical workflow. Thus, deep learning-based applications for diseases diagnosis will empower physicians and allow fast decision-making in clinical practice.

HIGHLIGHT

Deep learning can be more robust with various features for differentiating classes, provided the training set is large and diverse for analysis. However, sufficient medical images for training sets are not always available from medical institutions, which is one of the major limitations of deep learning in medical image analysis. This review article presents some solutions for this issue and discusses efforts needed to develop robust deep learning-based computer-aided diagnosis applications for better clinical workflow in endoscopy, radiology, pathology, and dentistry.

CONCLUSION

The introduction of deep learning-based applications will enhance the traditional role of medical practitioners in ensuring accurate diagnoses and treatment in terms of precision, reproducibility, and scalability.

A Deep Learning Model for Cervical Cancer Screening on Liquid-Based Cytology Specimens in Whole Slide Images

Simple Summary

In this pilot study, we aimed to investigate the use of deep learning for the classification of whole-slide images of liquid-based cytology specimens into neoplastic and non-neoplastic. To do so, we used a large training and test sets. Overall, the model achieved good classification performance in classifying whole-slide images, demonstrating the promising potential use of such models for aiding the screening processes for cervical cancer.

Abstract

Liquid-based cytology (LBC) for cervical cancer screening is now more common than the conventional smears, which when digitised from glass slides into whole-slide images (WSIs), opens up the possibility of artificial intelligence (AI)-based automated image analysis. Since conventional screening processes by cytoscreeners and cytopathologists using microscopes is limited in terms of human resources, it is important to develop new computational techniques that can automatically and rapidly diagnose a large amount of specimens without delay, which would be of great benefit for clinical laboratories and hospitals. The goal of this study was to investigate the use of a deep learning model for the classification of WSIs of LBC specimens into neoplastic and non-neoplastic. To do so, we used a dataset of 1605 cervical WSIs. We evaluated the model on three test sets with a combined total of 1468 WSIs, achieving ROC AUCs for WSI diagnosis in the range of 0.89–0.96, demonstrating the promising potential use of such models for aiding screening processes.

Deep Learning Models for Poorly Differentiated Colorectal Adenocarcinoma Classification in Whole Slide Images Using Transfer Learning

Colorectal poorly differentiated adenocarcinoma (ADC) is known to have a poor prognosis as compared with well to moderately differentiated ADC. The frequency of poorly differentiated ADC is relatively low (usually less than 5% among colorectal carcinomas). Histopathological diagnosis based on endoscopic biopsy specimens is currently the most cost effective method to perform as part of colonoscopic screening in average risk patients, and it is an area that could benefit from AI-based tools to aid pathologists in their clinical workflows. In this study, we trained deep learning models to classify poorly differentiated colorectal ADC from Whole Slide Images (WSIs) using a simple transfer learning method. We evaluated the models on a combination of test sets obtained from five distinct sources, achieving receiver operating characteristic curve (ROC) area under the curves (AUCs) up to 0.95 on 1799 test cases.

Breast Invasive Ductal Carcinoma Classification on Whole Slide Images with Weakly-Supervised and Transfer Learning

Simple Summary

In this study, we have trained deep learning models using transfer learning and weakly-supervised learning for the classification of breast invasive ductal carcinoma (IDC) in whole slide images (WSIs). We evaluated the models on four test sets: one biopsy (n = 522) and three surgical (n = 1129) achieving AUCs in the range 0.95 to 0.99. We have also compared the trained models to existing pre-trained models on different organs for adenocarcinoma classification and they have achieved lower AUC performances in the range 0.66 to 0.89 despite adenocarcinoma exhibiting some structural similarity to IDC. Therefore, performing fine-tuning on the breast IDC training set was beneficial for improving performance. The results demonstrate the potential use of such models to aid pathologists in clinical practice.

Abstract

Invasive ductal carcinoma (IDC) is the most common form of breast cancer. For the non-operative diagnosis of breast carcinoma, core needle biopsy has been widely used in recent years for the evaluation of histopathological features, as it can provide a definitive diagnosis between IDC and benign lesion (e.g., fibroadenoma), and it is cost effective. Due to its widespread use, it could potentially benefit from the use of AI-based tools to aid pathologists in their pathological diagnosis workflows. In this paper, we trained invasive ductal carcinoma (IDC) whole slide image (WSI) classification models using transfer learning and weakly-supervised learning. We evaluated the models on a core needle biopsy (n = 522) test set as well as three surgical test sets (n = 1129) obtaining ROC AUCs in the range of 0.95–0.98. The promising results demonstrate the potential of applying such models as diagnostic aid tools for pathologists in clinical practice.

Deep Learning Models for Gastric Signet Ring Cell Carcinoma Classification in Whole Slide Images

Signet ring cell carcinoma (SRCC) of the stomach is a rare type of cancer with a slowly rising incidence. It tends to be more difficult to detect by pathologists, mainly due to its cellular morphology and diffuse invasion manner, and it has poor prognosis when detected at an advanced stage. Computational pathology tools that can assist pathologists in detecting SRCC would be of a massive benefit. In this paper, we trained deep learning models using transfer learning, fully-supervised learning, and weakly-supervised learning to predict SRCC in Whole Slide Images (WSIs) using a training set of 1,765 WSIs. We evaluated the models on two different test sets (n = 999, n = 455). The best model achieved a ROC-AUC of at least 0.99 on all two test sets, setting a top baseline performance for SRCC WSI classification.

Deep learning to predict subtypes of poorly differentiated lung cancer from biopsy whole slide images

8536

Background: Lung cancer is the leading cause of cancer-related death in many countries, and its prognosis remains unsatisfactory. Since treatment approaches differ substantially based on the subtype, such as adenocarcinoma (ADC), squamous cell carcinoma (SCC) and small cell lung cancer (SCLC), an accurate histopathological diagnosis is of great importance. However, if the specimen is solely composed of poorly differentiated cancer cells, distinguishing between histological subtypes can be difficult. The present study developed a deep learning model to classify lung cancer subtypes from whole slide images (WSIs) of transbronchial lung biopsy (TBLB) specimens, in particular with the aim of using this model to evaluate a challenging test set of indeterminate cases. Methods: Our deep learning model consisted of two separately trained components: a convolutional neural network tile classifier and a recurrent neural network tile aggregator for the WSI diagnosis. We used a training set consisting of 638 WSIs of TBLB specimens to train a deep learning model to classify lung cancer subtypes (ADC, SCC and SCLC) and non-neoplastic lesions. The training set consisted of 593 WSIs for which the diagnosis had been determined by pathologists based on the visual inspection of Hematoxylin-Eosin (HE) slides and of 45 WSIs of indeterminate cases (64 ADCs and 19 SCCs). We then evaluated the models using five independent test sets. For each test set, we computed the receiver operator curve (ROC) area under the curve (AUC). Results: We applied the model to an indeterminate test set of WSIs obtained from TBLB specimens that pathologists had not been able to conclusively diagnose by examining the HE-stained specimens alone. Overall, the model achieved ROC AUCs of 0.993 (confidence interval [CI] 0.971-1.0) and 0.996 (0.981-1.0) for ADC and SCC, respectively. We further evaluated the model using five independent test sets consisting of both TBLB and surgically resected lung specimens (combined total of 2490 WSIs) and obtained highly promising results with ROC AUCs ranging from 0.94 to 0.99. Conclusions: In this study, we demonstrated that a deep learning model could be trained to predict lung cancer subtypes in indeterminate TBLB specimens. The extremely promising results obtained show that if deployed in clinical practice, a deep learning model that is capable of aiding pathologists in diagnosing indeterminate cases would be extremely beneficial as it would allow a diagnosis to be obtained sooner and reduce costs that would result from further investigations.

A deep learning model to detect pancreatic ductal adenocarcinoma on endoscopic ultrasound-guided fine-needle biopsy

Histopathological diagnosis of pancreatic ductal adenocarcinoma (PDAC) on endoscopic ultrasonography-guided fine-needle biopsy (EUS-FNB) specimens has become the mainstay of preoperative pathological diagnosis. However, on EUS-FNB specimens, accurate histopathological evaluation is difficult due to low specimen volume with isolated cancer cells and high contamination of blood, inflammatory and digestive tract cells. In this study, we performed annotations for training sets by expert pancreatic pathologists and trained a deep learning model to assess PDAC on EUS-FNB of the pancreas in histopathological whole-slide images. We obtained a high receiver operator curve area under the curve of 0.984, accuracy of 0.9417, sensitivity of 0.9302 and specificity of 0.9706. Our model was able to accurately detect difficult cases of isolated and low volume cancer cells. If adopted as a supportive system in routine diagnosis of pancreatic EUS-FNB specimens, our model has the potential to aid pathologists diagnose difficult cases.

Weakly-supervised learning for lung carcinoma classification using deep learning

Lung cancer is one of the major causes of cancer-related deaths in many countries around the world, and its histopathological diagnosis is crucial for deciding on optimum treatment strategies. Recently, Artificial Intelligence (AI) deep learning models have been widely shown to be useful in various medical fields, particularly image and pathological diagnoses; however, AI models for the pathological diagnosis of pulmonary lesions that have been validated on large-scale test sets are yet to be seen. We trained a Convolution Neural Network (CNN) based on the EfficientNet-B3 architecture, using transfer learning and weakly-supervised learning, to predict carcinoma in Whole Slide Images (WSIs) using a training dataset of 3,554 WSIs. We obtained highly promising results for differentiating between lung carcinoma and non-neoplastic with high Receiver Operator Curve (ROC) area under the curves (AUCs) on four independent test sets (ROC AUCs of 0.975, 0.974, 0.988, and 0.981, respectively). Development and validation of algorithms such as ours are important initial steps in the development of software suites that could be adopted in routine pathological practices and potentially help reduce the burden on pathologists.

Reliable detection of the presence of pulmonary carcinoma on whole-slide images by a deep learning model

9025

Background: Lung cancer is one of the leading causes of cancer-related death worldwide, and its histopathological diagnosis is crucial for deciding on optimum treatment strategies. Recently, artificial intelligence (AI) models have been widely shown to be useful in various medical fields, particularly image and pathological diagnoses; however, AI models for the pathological diagnosis of pulmonary lesions that have been validated in large-scale test sets are yet to be seen. Methods: We trained a convolution neural network based on the Efficient Net B3 architecture to classify carcinoma from whole slide images (WSIs) using a training dataset of 3640 images. WSI diagnoses were available. We used a transfer learning approach, in which the starting weights were obtained from a pre-trained model on ImageNet. The model was then trained on our dataset using multiple instance learning, a semi-supervised learning approach. To classify a WSI, the model was applied in a sliding window fashion with an input tile size of 512x512 and a stride of 256. The maximum probability was then used as a WSI diagnosis. Results: We evaluated our model on a total of 2680 WSIs originating from five independent sources (two hospitals in Japan and three public datasets from around the world). The model achieved a Receiver Operator Curve Area Under the Curves (ROC AUCs) of 0.974, 0.974, 0.996, 0.988, and 0.981, respectively. Conclusions: We successfully established a reliable AI model for differentiating between lung carcinoma and non-neoplasm with a high ROC AUC on five independent test sets. If used in clinical practice, our model could help reduce the burden on pathologists and be useful for diagnosing pulmonary lesions in areas in which there are shortages of pathologists. Further prospective multicenter studies are warranted in order to validate the results obtained in the current study.

Rac1-dependent phagocytosis of apoptotic cells by oral squamous cell carcinoma cells: A possible driving force for tumor progression

Apoptotic cell death frequently occurs in human cancer tissues including oral squamous cell carcinoma (SCC), wherein apoptotic tumor cells are phagocytosed not only by macrophages but also by neighboring tumor cells. We previously reported that the engulfment of apoptotic SCC cells by neighboring SCC cells frequently occurs at the invading front. Therefore, we hypothesized that the phagocytosis of these apoptotic cells by tumor cells contributes to disease progression. Herein, using cultured oral SCC cells, we aimed to confirm whether tumor cells actually phagocytose apoptotic cells and to examine whether cellular activities are regulated by the phagocytosis of apoptotic cells. Co-culture experiments showed that living cells could ingest apoptotic cells into phagolysosomes. NSC23766, an inhibitor of Rac1, which is a key regulator of phagocytic cup formation in professional phagocytes, dramatically suppressed the phagocytosis of apoptotic cells by living cells. Additionally, cell migration and the secretion of DKK1, a tumor-promoting protein, were enhanced by co-culture with apoptotic cells, whereas NSC23766 inhibited these effects. These results show that tumor cells can actively phagocytose apoptotic neighbors in a Rac1-dependent manner and that such activity increases their migration. The regulation of apoptotic cell phagocytosis thus represents new directions for therapeutic intervention for oral cancer.

Deep Learning Models for Histopathological Classification of Gastric and Colonic Epithelial Tumours

Histopathological classification of gastric and colonic epithelial tumours is one of the routine pathological diagnosis tasks for pathologists. Computational pathology techniques based on Artificial intelligence (AI) would be of high benefit in easing the ever increasing workloads on pathologists, especially in regions that have shortages in access to pathological diagnosis services. In this study, we trained convolutional neural networks (CNNs) and recurrent neural networks (RNNs) on biopsy histopathology whole-slide images (WSIs) of stomach and colon. The models were trained to classify WSI into adenocarcinoma, adenoma, and non-neoplastic. We evaluated our models on three independent test sets each, achieving area under the curves (AUCs) up to 0.97 and 0.99 for gastric adenocarcinoma and adenoma, respectively, and 0.96 and 0.99 for colonic adenocarcinoma and adenoma respectively. The results demonstrate the generalisation ability of our models and the high promising potential of deployment in a practical histopathological diagnostic workflow system.

K1 gene transformation activities in AIDS-related and classic type Kaposi's sarcoma: Correlation with clinical presentation

Kaposi's sarcoma-associated herpesvirus (KSHV) causes both AIDS-related Kaposi's sarcoma (KS) and classic KS, but their clinical presentations are different, and respective mechanisms remain to be elucidated. The KSHV K1 gene is reportedly involved in tumorigenesis through the immunoreceptor tyrosine-based activation motif (ITAM). Since we found the sequence variations in the K1 gene of KSHV isolated from AIDS-related KS and classic KS, we hypothesized that the transformation activity of the K1 gene contributes to the different clinical presentations. To evaluate our hypothesis, we compared the transformation activities of the K1 gene between AIDS-related KS and classic KS. We also analyzed ITAM activities and the downstream AKT and NF-κB. We found that the transformation activity of AIDS-related K1 was greater than that of classic K1, and that AIDS-related K1 induced higher ITAM activity than classic K1, causing more potent Akt and NF-κB activities. K1 downregulation by siRNA in AIDS-related K1 expressing cells induced a loss of transformation properties and decreased both Akt and NF-κB activities, suggesting a correlation between the transformation activity of K1 and ITAM signaling. Our study indicates that the increased transformation activity of AIDS-related K1 is associated with its clinical aggressiveness, whereas the weak transformation activity of classic type K1 is associated with a mild clinical presentation and spontaneous regression. The mechanism of spontaneous regression of classic KS may provide new therapeutic strategy to cancer.

Possible role of p53/Mieap-regulated mitochondrial quality control as a tumor suppressor in human breast cancer

Mitochondria‐eating protein (Mieap), encoded by a p53‐target gene, plays an important role in mitochondrial quality control (MQC). Mieap has been reported to have a critical role in tumor suppression in colorectal cancer. Here, we investigated its role as a tumor suppressor in breast cancer. The enforced expression of exogenous Mieap in breast cancer cells induced caspase‐dependent apoptosis, with activation of both caspase‐3/7 and caspase‐9. Immunohistochemistry revealed endogenous Mieap in the cytoplasm in 24/75 (32%) invasive ductal carcinomas (IDC), 15/27 (55.6%) cases of ductal carcinoma in situ (DCIS) and 16/18 (88.9%) fibroadenomas (FA) (IDC vs DCIS; P = 0.0389, DCIS vs FA; P = 0.0234, IDC vs FA; P < 0.0001). In IDC, the Mieap promoter was methylated in 6/46 (13%) cases, whereas p53 was mutated in 6/46 (13%) cases. Therefore, the p53/Mieap‐regulated MQC pathway was inactivated in 12/46 IDC (26.1%). Interestingly, all tumors derived from the 12 patients with Mieap promoter methylation or p53 mutations pathologically exhibited more aggressive and malignant breast cancer phenotypes. Impairment of p53/Mieap‐regulated MQC pathway resulted in significantly shorter disease‐free survival (DFS) (P = 0.021), although p53 status is more prognostic in DFS than Mieap promoter methylation. These results indicate that p53/Mieap‐regulated MQC has a critical role in tumor suppression in breast cancer, possibly in part through mitochondrial apoptotic pathway.

Survivin: A novel marker and potential therapeutic target for human angiosarcoma

Human angiosarcoma is a rare malignant vascular tumor associated with extremely poor clinical outcome and generally arising in skin of the head and neck region. However, little is known about the molecular pathogeneses and useful immunohistochemical markers of angiosarcoma. To investigate the mechanisms of angiosarcoma progression, we collected 85 cases of human angiosarcoma specimens with clinical records and analyzed ISO-HAS-B patient-derived angiosarcoma cells. As control subjects, 54 cases of hemangioma and 34 of pyogenic granuloma were collected. Remarkably, consistent with our recent observations regarding the involvement of survivin expression following Hippo pathway inactivation in the neoplastic proliferation of murine hemangioendothelioma cells and human infantile hemangioma, nuclear survivin expression was observed in all cases of angiosarcoma but not in hemangiomas and pyogenic granulomas, and the Hippo pathway was inactivated in 90.3% of yes-associated protein (YAP) -positive angiosarcoma cases. However, survivin expression modes and YAP localization (Hippo pathway activation modes) were not correlated with survival. In addition, we confirmed that survivin small interference RNA (siRNA) transfection and YM155, an anti-survivin drug, elicited decreased nuclear survivin expression and cell proliferation in ISO-HAS-B cells which expressed survivin consistently. Conclusively, these findings support the importance of survivin as a good marker and critical regulator of cellular proliferation for human angiosarcoma and YM155 as a potential therapeutic agent.

CD44 Promotes Inflammation and Extracellular Matrix Production During Arteriovenous Fistula Maturation

OBJECTIVE

Arteriovenous fistulae (AVF) remain the optimal conduit for hemodialysis access but continue to demonstrate poor patency and poor rates of maturation. We hypothesized that CD44, a widely expressed cellular adhesion molecule that serves as a major receptor for extracellular matrix components, promotes wall thickening and extracellular matrix deposition during AVF maturation.

APPROACH AND RESULTS

AVF were created via needle puncture in wild-type C57BL/6J and CD44 knockout mice. CD44 mRNA and protein expression was increased in wild-type AVF. CD44 knockout mice showed no increase in AVF wall thickness (8.9 versus 26.8 μm; P=0.0114), collagen density, and hyaluronic acid density, but similar elastin density when compared with control AVF. CD44 knockout mice also showed no increase in vascular cell adhesion molecule-1 expression, intercellular adhesion molecule-1 expression, and monocyte chemoattractant protein-1 expression in the AVF compared with controls; there were also no increased M2 macrophage markers (transglutaminase-2: 81.5-fold, P=0.0015; interleukin-10: 7.6-fold, P=0.0450) in CD44 knockout mice. Delivery of monocyte chemoattractant protein-1 to CD44 knockout mice rescued the phenotype with thicker AVF walls (27.2 versus 14.7 μm; P=0.0306), increased collagen density (2.4-fold; P=0.0432), and increased number of M2 macrophages (2.1-fold; P=0.0335).

CONCLUSIONS

CD44 promotes accumulation of M2 macrophages, extracellular matrix deposition, and wall thickening during AVF maturation. These data show the association of M2 macrophages with wall thickening during AVF maturation and suggest that enhancing CD44 activity may be a strategy to increase AVF maturation.

Increased Oxidative Stress and Hypoxia Inducible Factor-1 Expression during Arteriovenous Fistula Maturation

BACKGROUND

The poor clinical results that are frequently reported for arteriovenous fistulae (AVF) for hemodialysis are typically due to failure of AVF maturation. We hypothesized that early AVF maturation is associated with generation of reactive oxygen species and activation of the hypoxia-inducible factor-1 (HIF-1) pathway, potentially promoting neointimal hyperplasia. We tested this hypothesis using a previously reported mouse AVF model that recapitulates human AVF maturation.

METHODS

Aortocaval fistulae were created in C57Bl/6 mice and compared with sham-operated mice. AVFs or inferior vena cavas were analyzed using a microarray, Amplex Red for extracellular H₂O₂, quantitative polymerase chain reaction, immunohistochemistry, and immunoblotting for HIF-1α and immunofluorescence for NOX-2, nitrotyrosine, heme oxygenase-1 (HO-1), and vascular endothelial growth factor (VEGF)-A.

RESULTS

Oxidative stress was higher in AVF than that in control veins, with more H₂O₂ (P = 0.007) and enhanced nitrotyrosine immunostaining (P = 0.005). Immunohistochemistry and immunoblot showed increased HIF-1α immunoreactivity in the AVF endothelium; HIF-1 targets NOX-2, HO-1 and VEGF-A were overexpressed in the AVF (P < 0.01). AVF expressed increased numbers of HIF-1α (P < 0.0001) and HO-1 (P < 0.0001) messenger RNA transcripts.

CONCLUSIONS

Oxidative stress increases in mouse AVF during early maturation, with increased expression of HIF-1α and its target genes NOX-2, HO-1, and VEGF-A. These results suggest that clinical strategies to improve AVF maturation could target the HIF-1 pathway.

CD44 Influences Fibroblast Behaviors Via Modulation of Cell-Cell and Cell-Matrix Interactions, Affecting Survivin and Hippo Pathways

CD44 has been studied in a wide variety of cell types, in a diverse array of cell behaviors and in a diverse range of signaling pathways. We now document a role for CD44 in mediating fibroblast behaviors via regulation of N-cadherin, extracellular matrix expression, Survivin and the Hippo pathway. Here, we report our findings on the roles of CD44 in modulating proliferation, apoptosis, migration and invasion of murine wild-type (WT-FB) and CD44 knockout dermal fibroblasts (CD44KO-FB). As we have documented in microvascular endothelial cells lacking CD44, we found persistent increased proliferation, reduced activation of cleaved caspase 3, increased initial attachment, but decreased strength of cell attachment in high cell density, post confluent CD44KO-FB cultures. Additionally, we found that siRNA knock-down of CD44 mimicked the behaviors of CD44KO-FB, restoring the decreases in N-cadherin, collagen type I, fibronectin, Survivin, nuclear fractions of YAP and phospho-YAP and decreased levels of cleaved caspase 3 to the levels observed in CD44KO-FB. Interestingly, plating CD44KO-FB on collagen type I or fibronectin resulted in significant decreases in secondary proliferation rates compared to plating cells on non-coated dishes, consistent with increased cell adhesion compared to their effects on WT-FB. Lastly, siRNA knockdown of CD44 in WT-FB resulted in increased fibroblast migration compared to WT-FB, albeit at reduced rates compared to CD44KO-FB. These results are consistent with CD44's pivotal role in modulating several diverse behaviors important for adhesion, proliferation, apoptosis, migration and invasion during development, growth, repair, maintenance and regression of a wide variety of mesenchymal tissues.

Membrane-mediated regulation of vascular identity

Vascular diseases span diverse pathology, but frequently arise from aberrant signaling attributed to specific membrane-associated molecules, particularly the Eph-ephrin family. Originally recognized as markers of embryonic vessel identity, Eph receptors and their membrane-associated ligands, ephrins, are now known to have a range of vital functions in vascular physiology. Interactions of Ephs with ephrins at cell-to-cell interfaces promote a variety of cellular responses such as repulsion, adhesion, attraction, and migration, and frequently occur during organ development, including vessel formation. Elaborate coordination of Eph- and ephrin-related signaling among different cell populations is required for proper formation of the embryonic vessel network. There is growing evidence supporting the idea that Eph and ephrin proteins also have postnatal interactions with a number of other membrane-associated signal transduction pathways, coordinating translation of environmental signals into cells. This article provides an overview of membrane-bound signaling mechanisms that define vascular identity in both the embryo and the adult, focusing on Eph- and ephrin-related signaling. We also discuss the role and clinical significance of this signaling system in normal organ development, neoplasms, and vascular pathologies.

Modulation of Sox10, HIF-1α, Survivin, and YAP by Minocycline in the Treatment of Neurodevelopmental Handicaps following Hypoxic Insult

Premature infants are at an increased risk of developing cognitive and motor handicaps due to chronic hypoxia. Although the current therapies have reduced the incidence of these handicaps, untoward side effects abound. Using a murine model of sublethal hypoxia, we demonstrated reduction in several transcription factors that modulate expression of genes known to be involved in several neural functions. We demonstrate the induction of these genes by minocycline, a tetracycline antibiotic with noncanonical functions, in both in vitro and in vivo studies. Specifically, there was induction of genes, including Sox10, Hif1a, Hif2a, Birc5, Yap1, Epo, Bdnf, Notch1 (cleaved), Pcna, Mag, Mobp, Plp1, synapsin, Adgra2, Pecam1, and reduction in activation of caspase 3, all known to affect proliferation, apoptosis, synaptic transmission, and nerve transmission. Minocycline treatment of mouse pups reared under sublethal hypoxic conditions resulted in improvement in open field testing parameters. These studies demonstrate beneficial effects of minocycline treatment following hypoxic insult, document up-regulation of several genes associated with improved cognitive function, and support the possibility of minocycline as a potential therapeutic target in the treatment of neurodevelopmental handicaps observed in the very premature newborn population. Additionally, these studies may aid in further interpretation of the effects of minocycline in the treatment trials and animal model studies of fragile X syndrome and multiple sclerosis.

Disturbed shear stress reduces Klf2 expression in arterial-venous fistulae in vivo

Laminar shear stress (SS) induces an antiproliferative and anti-inflammatory endothelial phenotype and increases Klf2 expression. We altered the diameter of an arteriovenous fistula (AVF) in the mouse model to determine whether increased fistula diameter produces disturbed SS in vivo and if acutely increased disturbed SS results in decreased Klf2 expression. The mouse aortocaval fistula model was performed with 22, 25, or 28 gauge needles to puncture the aorta and the inferior vena cava. Duplex ultrasound was used to examine the AVF and its arterial inflow and venous outflow, and SS was calculated. Arterial samples were examined with western blot, immunohistochemistry, and immunofluorescence analysis for proteins and qPCR for RNA. Mice with larger diameter fistulae had diminished survival but increased AVF patency. Increased SS magnitudes and range of frequencies were directly proportional to the needle diameter in the arterial limb proximal to the fistula but not in the venous limb distal to the fistula, with 22-gauge needles producing the most disturbed SS in vivo. Klf2 mRNA and protein expression was diminished in the artery proximal to the fistula in proportion to increasing SS. Increased fistula diameter produces increased SS magnitude and frequency, consistent with disturbed SS in vivo. Disturbed SS is associated with decreased mRNA and protein expression of Klf2. Disturbed SS and reduced Klf2 expression near the fistula are potential therapeutic targets to improve AVF maturation.

CD44 regulation of endothelial cell proliferation and apoptosis via modulation of CD31 and VE-cadherin expression

CD44 has been implicated in a diverse array of cell behaviors and in a diverse range of signaling pathway activations under physiological and pathophysiological conditions. We have documented a role for CD44 in mediating vascular barrier integrity via regulation of PECAM-1 (CD31) expression. We now report our findings on the roles of CD44 in modulating proliferation and apoptosis of microvascular endothelial cells via its modulation of CD31 and VE-cadherin expression and the Hippo pathway. In this report, we demonstrate persistent increased proliferation and reduced activations of both effector and initiator caspases in high cell density, postconfluent CD44 knock-out (CD44KO), and CD31KO cultures. We found that reconstitution with murine CD44 or CD31 restored the proliferative and caspase activation rates to WT levels. Moreover, we have confirmed that the CD31 ecto-domain plays a key role in specific caspase cascades as well as cell adhesion-mediated cell growth and found that CD31 deficiency results in a reduction in VE-cadherin expression. Last, we have shown that both CD44KO and CD31KO endothelial cells exhibit a reduced VE-cadherin expression correlating with increased survivin expression and YAP nuclear localization, consistent with inactivation of the Hippo pathway, resulting in increased proliferation and decreased apoptosis. These findings support the concept that CD44 mediates several of its effects on endothelia through modulation of adhesion protein expression, which, in addition to its known modulation of junctional integrity, matrix metalloproteinase levels and activation, interactions with cortical membrane proteins, and selected signaling pathways, plays a key role as a critical regulator of vascular function.

The mouse aortocaval fistula recapitulates human arteriovenous fistula maturation

Several models of arteriovenous fistula (AVF) have excellent patency and help in understanding the mechanisms of venous adaptation to the arterial environment. However, these models fail to exhibit either maturation failure or fail to develop stenoses, both of which are critical modes of AVF failure in human patients. We used high-resolution Doppler ultrasound to serially follow mice with AVFs created by direct 25-gauge needle puncture. By day 21, 75% of AVFs dilate, thicken, and increase flow, i.e., mature, and 25% fail due to immediate thrombosis or maturation failure. Mature AVF thicken due to increased amounts of smooth muscle cells. By day 42, 67% of mature AVFs remain patent, but 33% of AVFs fail due to perianastomotic thickening. These results show that the mouse aortocaval model has an easily detectable maturation phase in the first 21 days followed by a potential failure phase in the subsequent 21 days. This model is the first animal model of AVF to show a course that recapitulates aspects of human AVF maturation.

Extracellular heat shock protein A9 is a novel interaction partner of podoplanin in oral squamous cell carcinoma cells

In previous studies, we have shown several lines of evidence that podoplanin (PDPN) plays an important role in cell adhesion via its association with extracellular components in neoplastic conditions, though there has been no trial to search for PDPN-interaction molecules in the extracellular milieu. To screen for those molecules, we performed proteomics-based analysis using liquid chromatography-tandem mass spectrometry followed by co-immunoprecipitation for PDPN in ZK-1, an oral squamous cell carcinoma (SCC) cell system whose cell membrane molecules were cross-linked with each other in their extracellular compartments, and we identified heat shock protein (HSP) A9 as one of the extracellular PDPN bound molecules. Effects of transient PDPN knockdown by siRNA in ZK-1 were also comparatively examined for cellular behaviors in terms of HSPA9 expression and secretion. Finally, HSPA9 expression modes were immunohistochemically visualized in oral SCC tissue specimens. HSPA9 was secreted from ZK-1 cells, and the expression and secretion levels of HSPA9 gene and protein were well coordinated with those of PDPN. Immunohistochemically, HSPA9 and PDPN were co-localized in ZK-1 cells and oral SCC foci, especially in the peripheral zone. In conclusion, the results indicate that HSPA9 secreted by oral SCC cells interacts with PDPN on their cell surface in an autocrine manner and regulates their growth and invasiveness.

Inflammatory histopathogenesis of nasopalatine duct cyst: a clinicopathological study of 41 cases

OBJECTIVE

The aim of this study is to characterize immunohistochemical profiles of lining epithelia of nasopalatine duct cyst (NPC) as well as to correlate those findings with their clinicopathological features to understand the histopathogenesis of NPC.

MATERIALS AND METHODS

Forty-one surgical specimens from NPC were examined for clinical profiles and expression of keratin-7, 13, MUC-1, and P63 by immunohistochemistry, compared to radicular cyst (RC) and maxillary sinusitis.

RESULTS

Nasopalatine duct cyst was clinically characterized by male predominant occurrence: 44% of the cases involved tooth roots, and 70% with inflammatory backgrounds. Lining epithelia of NPCs without daughter cysts were immunohistochemically distinguished into three layers: a keratin 7-positive (+) ciliated cell layer in the surface, a keratin-13+ middle layer, and a MUC-1+/P63+ lower half, indicating that they were not respiratory epithelia, and the same layering pattern was observed in RC. However, those immunolocalization patterns of the main cyst lining with daughter cyst were exactly the same as those of daughter cyst linings as well as duct epithelia of mucous glands.

CONCLUSIONS

Two possible histopathogenesis of NPC were clarified: one was inflammatory cyst like RC and the other was salivary duct cyst-like mucocele.

Parenchymal-stromal switching for extracellular matrix production on invasion of oral squamous cell carcinoma

It is poorly understood which cell type, tumor cells, or stromal cells are responsible for the production of extracellular matrix molecules in the neoplastic stroma. We studied the expression of 4 extracellular matrix molecules at the protein and messenger RNA levels in monocellular and 2 kinds of coculture systems between human squamous cell carcinoma (ZK-1) and fibroblast (OF-1) cell lines, which may correspond to carcinoma in situ and squamous cell carcinoma, respectively. Squamous cell carcinoma and carcinoma in situ tissue sections were also investigated by immunohistochemistry and in situ hybridization for extracellular matrix. Immunohistochemically, perlecan and tenascin C were localized in carcinoma cells in carcinoma in situ, whereas they were in the stromal space in squamous cell carcinoma. In monocellular culture conditions, expression levels for perlecan, tenascin C, and laminin were more predominant in ZK-1 than in OF-1, although those for fibronectin were more enhanced in OF-1. However, these extracellular matrix expression levels of OF-1 were elevated, whereas those of ZK-1 dropped when they were in coculture conditions. The differences between ZK-1 and OF-1 were significantly more evident in direct contact (ZK-1/OF-1, 56%-22%) than in indirect contact (63%-39%). These results indicate that oral squamous cell carcinoma cells produce extracellular matrix in the absence of stromal fibroblasts (or in carcinoma in situ) and that they stop producing extracellular matrix in the presence of fibroblasts (or in squamous cell carcinoma). It is hence suggested that stromal fibroblasts after direct contact with invading squamous cell carcinoma cells are more responsible than squamous cell carcinoma cells for the formation of neoplastic stroma, whereas carcinoma in situ cells have to produce and deposit extracellular matrix by themselves to form intraepithelial microstromal spaces.

Podoplanin expression profiles characteristic of odontogenic tumor-specific tissue architectures

Podoplanin, a representative immunohistochemical marker for lymphatic endothelial cells, is also expressed in many other kinds of cancer cells, although its pathophysiological function is largely unknown. Our aim was to determine immunolocalization modes of podoplanin among odontogenic tumors to discuss possible roles of podoplanin in their characteristic tissue architecture formation. Immunohistochemical profiles of podoplanin were investigated in 40 surgical specimens from ameloblastoma (AM), adenomatoid odontogenic tumor (AOT), calcifying cystic odontogenic tumor (CCOT), and keratocystic odontogenic tumor (KCOT) in comparison with those of proliferating cell nuclear antigen (PCNA), integrin β1, fibronectin, and matrix metalloproteinase 9 (MMP-9). Podoplanin was localized in the basal cell layer or in the peripheral zone of AM foci. It was found in spindle-shaped tumor cells of AOT, in both the basal and polyhedral cells of CCOT, and in the basal and parabasal cells of KCOT linings. Podoplanin-positive (+) cells were located within areas of PCNA+ cells, and integrin β1 was localized in the cell membrane of podoplanin+ cells in the intercellular space where fibronectin and MMP-9 were deposited. In conclusion, podoplanin+ cells and areas in odontogenic tumors are in close associations with extracellular matrix signalings as well as cell proliferation.

Cleavage of the ER-targeting signal sequence of parathyroid hormone-related protein is cell-type-specific and regulated in Cis by its nuclear localization signal

Prepro-parathyroid hormone-related protein (ppPTHrP) has two targeting signals, an N-terminal signal sequence and a nuclear localization signal (NLS). In fact, the protein is not only secreted from the cell but also found in the nucleus and/or nucleolus. In order to understand the function of the PTHrP signal sequence for the dual localization, the signal sequence cleavage of a series of ppPTHrP deletion mutants fused to Escherichia coli leader peptidase was analysed in vitro and in several cell lines. Efficiency of the PTHrP signal sequence cleavage was intrinsically low in the in vitro reconstitution system. In cultured cells, cleavage efficiency of the PTHrP signal sequence varied significantly, being lowest in COS-1 cells, but rising in HeLa, HEK293 and CV-1 cells. However, virtually complete signal sequence cleavage was observed in CHO cells. In addition, the NLS of PTHrP had a negative effect on its own signal sequence cleavage, which could be enhanced by deletion of the spacer sequence between the signal sequence and the NLS. There was a roughly inverse relationship between the signal sequence cleavage and the nuclear localization of PTHrP. Thus, the final destination of PTHrP could be regulated at the ER membrane.