Search

Abstract Variations in color and texture of histopathology images are caused by differences in staining conditions and imaging devices between hospitals. These biases decrease the robustness of machine learning models exposed to out-of-domain data. To address this issue, we introduce a comprehensive histopathology image dataset named PathoLogy Images of Scanners and Mobile phones (PLISM). The dataset consisted of 46 human tissue types stained using 13 hematoxylin and eosin conditions and captured using 13 imaging devices. Precisely aligned image patches from different domains allowed for an accurate evaluation of color and texture properties in each domain. Variation in PLISM was assessed and found to be significantly diverse across various domains, particularly between whole-slide images and smartphones. Furthermore, we assessed the improvement in domain shift using a convolutional neural network pre-trained on PLISM. PLISM is a valuable resource that facilitates the precise evaluation of domain shifts in digital pathology and makes significant contributions towards the development of robust machine learning models that can effectively address challenges of domain shift in histological image analysis.

Abstract Researchers have attempted to identify the factors involved in lymph node recurrence in cT1‐2N0 tongue squamous cell carcinoma (SCC). However, studies combining histopathological and clinicopathological information in prediction models are limited. We aimed to develop a highly accurate lymph node recurrence prediction model for clinical stage T1‐2, N0 (cT1‐2N0) tongue SCC by integrating histopathological artificial intelligence (AI) with clinicopathological information. A dataset from 148 patients with cT1‐2N0 tongue SCC was divided into training and test sets. The prediction models were constructed using AI‐extracted information from whole slide images (WSIs), human‐assessed clinicopathological information, and both combined. Weakly supervised learning and machine learning algorithms were used for WSIs and clinicopathological information, respectively. The combination model utilised both algorithms. Highly predictive patches from the model were analysed for histopathological features. In the test set, the areas under the receiver operating characteristic (ROC) curve for the model using WSI, clinicopathological information, and both combined were 0.826, 0.835, and 0.991, respectively. The highest area under the ROC curve was achieved with the model combining WSI and clinicopathological factors. Histopathological feature analysis showed that highly predicted patches extracted from recurrence cases exhibited significantly more tumour cells, inflammatory cells, and muscle content compared with non‐recurrence cases. Moreover, patches with mixed inflammatory cells, tumour cells, and muscle were significantly more prevalent in recurrence versus non‐recurrence cases. The model integrating AI‐extracted histopathological and human‐assessed clinicopathological information demonstrated high accuracy in predicting lymph node recurrence in patients with cT1‐2N0 tongue SCC.

Abstract One drawback of existing artificial intelligence (AI)-based histopathological prediction models is the lack of interpretability. The objective of this study is to extract p16-positive oropharyngeal squamous cell carcinoma (OPSCC) features in a form that can be interpreted by pathologists using AI model. We constructed a model for predicting p16 expression using a dataset of whole-slide images from 114 OPSCC biopsy cases. We used the clustering-constrained attention-based multiple-instance learning (CLAM) model, a weakly supervised learning approach. To improve performance, we incorporated tumor annotation into the model (Annot-CLAM) and achieved the mean area under the receiver operating characteristic curve of 0.905. Utilizing the image patches on which the model focused, we examined the features of model interest via histopathologic morphological analysis and cycle-consistent adversarial network (CycleGAN) image translation. The histopathologic morphological analysis evaluated the histopathological characteristics of image patches, revealing significant differences in the numbers of nuclei, the perimeters of the nuclei, and the intercellular bridges between p16-negative and p16-positive image patches. By using the CycleGAN-converted images, we confirmed that the sizes and densities of nuclei are significantly converted. This novel approach improves interpretability in histopathological morphology-based AI models and contributes to the advancement of clinically valuable histopathological morphological features.

In pathology, various tissue and cell components play diverse biological roles. The morphology of each component can vary markedly with differentiation status or pathological conditions, making it critical for understanding diseases. Traditional computational pathology methods typically employ patch-based feature extraction, which aggregates visual features across entire images. However, this approach does not differentiate between tissue types, limiting component analysis. To address this limitation, we introduce a novel concept in pathology image analysis, namely segment representation learning, and present an algorithm, SegRep, for this purpose. SegRep uses a unique dual-masking strategy that combines input masking and feature map masking. This approach effectively removes external influences for the targeted segment, identified via a segmentation model or manual annotation, allowing for the extraction of segment-specific feature representations. In addition, SegRep utilizes a self-supervised learning algorithm to achieve optimized segment representation. We evaluated SegRep’s efficacy in clustering and classification tasks using a dataset of human gastric cancer samples. The results demonstrate SegRep’s superior capability in extracting feature vectors that are highly specific to different pathology image segments. Compared with traditional methods, SegRep shows significant improvements in accuracy and specificity in both clustering and classification tasks. Segment representations obtained via SegRep can offer a more detailed and insightful perspective on computational pathology, paving the way for advanced applications in the field.

Abstract Although recent molecular analyses revealed that sporadic meningiomas have various genetic, epigenetic, and transcriptomic profiles, meningioma in patients with neurofibromatosis type 2 (NF2) have not been fully elucidated. This study investigated meningiomas' clinical, histological, and molecular characteristics in NF2 patients. A long-term retrospective follow-up (13.5 ± 5.5 years) study involving total 159 meningiomas in 37 patients with NF2 was performed. Their characteristics were assessed using immunohistochemistry (IHC), bulk-RNA sequencing, and copy number analysis. All variables of meningiomas in patients with NF2 were compared with those in 189 sporadic NF2-altered meningiomas in 189 patients. Most meningiomas in NF2 patients were stable, and the mean annual growth rate was 1.0 ± 1.8 cm3/year. Twenty-eight meningiomas (17.6%) in 25 patients (43.1%) were resected during the follow-up period. WHO grade I meningiomas in patients with NF2 were more frequent than in sporadic NF2-altered meningiomas (92.9% vs. 80.9%). Transcriptomic analysis for patients with NF2/sporadic NF2-altered WHO grade I meningiomas (n = 14 vs. 15, respectively) showed that tumours in NF2 patients still had a higher immune response and immune cell infiltration than sporadic NF2-altered meningiomas. Furthermore, RNA-seq/IHC-derived immunophenotyping corroborated this enhanced immune response by identifying myeloid cell infiltration, particularly in macrophages. Clinical, histological, and transcriptomic analyses of meningiomas in patients with NF2 demonstrated that meningiomas in NF2 patients showed less aggressive behaviour than sporadic NF2-altered meningiomas and elicited a marked immune response by identifying myeloid cell infiltration, particularly of macrophages.

Summary: The stomach is an important digestive organ with various biological functions. However, because of the complexity of its cellular and glandular composition, its precise cellular biology has yet to be elucidated. In this study, we conducted single-cell RNA sequencing (scRNA-seq) and subcellular-level spatial transcriptomics analysis of the human stomach and constructed the largest dataset to date: a stomach encyclopedia. This dataset consists of approximately 380,000 cells from scRNA-seq and the spatial transcriptome, enabling integrated analyses of transcriptional and spatial information of gastric and metaplastic cells. This analysis identified LEFTY1 as an uncharacterized stem cell marker, which was confirmed through lineage tracing analysis. A wide variety of cell-cell interactions between epithelial and stromal cells, including PDGFRA+BMP4+WNT5A+ fibroblasts, was highlighted in the developmental switch of intestinal metaplasia. Our extensive dataset will function as a fundamental resource in investigations of the stomach, including studies of development, aging, and carcinogenesis.

Summary: Deep texture representations (DTRs) produced from a bilinear convolutional neural network allow objective quantification of tumor histopathology images effectively. They can be used for various analyses, including visualization of morphological correlation between histology images, content-based image retrieval (CBIR), and supervised learning. This protocol describes the simplified workflow to analyze DTRs from data preparation, visualization of the histological profile, and CBIR analysis, to supervised learning model development to predict the profile from histological images.For complete details on the use and execution of this protocol, please refer to Komura et al. (2022).1 : Publisher’s note: Undertaking any experimental protocol requires adherence to local institutional guidelines for laboratory safety and ethics.

Abstract NF2 alteration is the most commonly–found genetic abnormality in meningiomas and is known to initiate events for aggressive-type meningiomas. Whereas the prognosis of meningiomas differs depending on their epigenomic/transcriptomic profile, the effect of NF2 alteration on the prognosis of benign meningiomas is not fully elucidated. This study aimed to probe the importance of NF2 alteration in prognosis of WHO grade I meningiomas. A long-term retrospective follow-up (5.3 ± 4.5 years) study involving 281 consecutive WHO grade I meningioma patients was performed. We assessed tumour recurrence in correlation with extent of resection (EOR), histopathological findings, tumour location, and NF2 alteration. “NF2 meningioma” was defined as meningiomas with presence of NF2 mutation and/or 22q loss. Overall, NF2 meningioma per se was not a predictor of prognosis in the whole cohort; however, it was a predictor of recurrence in supratentorial meningiomas, together with EOR and Ki-67. In a striking contrast, NF2 meningioma showed a better prognosis than non-NF2 meningioma in infratentorial lesion. Supratentorial NF2 meningiomas had higher Ki-67 and forkhead box protein M1 expression than those of others, possibly explaining the worse prognosis in this subtype. The combination of NF2 alteration, high Ki-67 and supratentorial location defines subgroup with the worst prognosis among WHO grade I meningiomas. Clinical connotation of NF2 alteration in terms of prognosis of WHO grade I meningioma differs in an opposite way between supratentorial and infratentorial tumors. Integrated anatomical, histopathological, and genomic classifications will provide the best follow-up schedule and proactive measures.

Abstract Genetic analysis and culturing techniques for gastric non-Helicobacter pylori Helicobacter (NHPH) are progressing. NHPH is reported to accompany nodular gastritis, gastric MALT lymphoma, and mild gastritis. However, only a few gastric cancer cases infected by NHPH have been reported. PCR analysis specific for NHPH and H. pylori was performed for DNA from gastric mucosa of 282 Korean gastric cancer patients, who were treated with endoscopic submucosal dissection. For more precise strain detection of NHPH, NHPH-positive mucosa was stained by immunohistochemistry specific for Helicobacter suis. The Cancer Genome Atlas (TCGA) classification was analyzed for these 3 gastric cancer sub-groups by in situ hybridization and immunohistochemistry. Among 281 patients, 3 patients (1.1%) were positive for NHPH. One patient (Patient 1) was also positive for H. pylori by PCR, another patient (Patient 3) was positive for serum IgG for H. pylori, and the other patient (Patient 2) had no evidence for H. pylori infection. Gastric mucosa of Patients 2 and 3 were positive for H. suis staining. All three NHPH-positive gastric cancers were located in the antrum, and belonged to the Chromosomal Instability Type of TCGA classification. Gastric NHPH can be a cause of gastric cancer, although likely with lower pathogenesis than H. pylori.

Abstract Background Schwannomas are neoplasms that typically arise from the myelin sheath of peripheral nerves and rarely originate within the brain parenchyma. Some case reports present schwannomas arising from the brainstem, but regrowth of the tumor and the efficacy of postoperative irradiation have not been examined. In addition, the genetic background of schwannomas arising from the brainstem has not been investigated. Case presentation A 21-year-old male presented with diplopia, dysphagia, and left-sided hemiparesis, dysesthesia, and ataxia. Intracranial imaging showed a heterogeneous mass with a cystic lesion in the pontomedullary junction. Since the tumor caused obstructive hydrocephalus, the patient underwent subtotal tumor resection. A histopathologic evaluation aided a diagnosis of brainstem intraparenchymal schwannoma. Gradual postoperative mass regrowth was recognized. Three-dimensional conformal radiotherapy was performed on the residual mass and surgical cavity. No tumor regrowth was observed 4 years after surgery. To investigate the genetic background of the tumor, target sequences for 36 genes, including NF2, SMARCB1, and LZTR1, and microsatellite analysis for loss of 22q did not show any somatic variants or 22q loss. Conclusions We suggest that brainstem schwannomas might differ from conventional schwannomas in their genetic background.

Summary: The patient-derived xenograft (PDX) model is a versatile tool used to study the tumor microenvironment (TME). However, limited studies have described multi-tumor PDX screening strategies to detect hub regulators during cancer-stroma interaction. Transcriptomes of cancer (human) and stroma (mouse) components of 70 PDX samples comprising 9 distinctive tumor types were analyzed in this study. PDX models recapitulated the original tumors' features, including tumor composition and putative signaling. Particularly, kidney renal clear cell carcinoma (KIRC) stood out, with altered hypoxia-related pathways and a high proportion of endothelial cells in the TME. Furthermore, an integrated analysis conducted to predict paracrine effectors in the KIRC cancer-to-stroma communication detected well-established soluble factors responsible for the hypoxia-related reaction and the so-far unestablished soluble factor, apelin (APLN). Subsequent experiments also supported the potential role of APLN in KIRC tumor progression. Therefore, this paper hereby provides an analytical workflow to find hub regulators in cancer-stroma interactions.

Abstract Background The recent success of immunotherapy in treating tumors has attracted increasing interest in research related to the adaptive immune system in the tumor microenvironment. Recent advances in next-generation sequencing technology enabled the sequencing of whole T-cell receptors (TCRs) and B-cell receptors (BCRs)/immunoglobulins (Igs) in the tumor microenvironment. Since BCRs/Igs in tumor tissues have high affinities for tumor-specific antigens, the patterns of their amino acid sequences and other sequence-independent features such as the number of somatic hypermutations (SHMs) may differ between the normal and tumor microenvironments. However, given the high diversity of BCRs/Igs and the rarity of recurrent sequences among individuals, it is far more difficult to capture such differences in BCR/Ig sequences than in TCR sequences. The aim of this study was to explore the possibility of discriminating BCRs/Igs in tumor and in normal tissues, by capturing these differences using supervised machine learning methods applied to RNA sequences of BCRs/Igs. Results RNA sequences of BCRs/Igs were obtained from matched normal and tumor specimens from 90 gastric cancer patients. BCR/Ig-features obtained in Rep-Seq were used to classify individual BCR/Ig sequences into normal or tumor classes. Different machine learning models using various features were constructed as well as gradient boosting machine (GBM) classifier combining these models. The results demonstrated that BCR/Ig sequences between normal and tumor microenvironments exhibit their differences. Next, by using a GBM trained to classify individual BCR/Ig sequences, we tried to classify sets of BCR/Ig sequences into normal or tumor classes. As a result, an area under the curve (AUC) value of 0.826 was achieved, suggesting that BCR/Ig repertoires have distinct sequence-level features in normal and tumor tissues. Conclusions To the best of our knowledge, this is the first study to show that BCR/Ig sequences derived from tumor and normal tissues have globally distinct patterns, and that these tissues can be effectively differentiated using BCR/Ig repertoires.

Pancreatic cancer has an extremely poor prognosis, and identification of novel predictors of therapeutic efficacy and prognosis is urgently needed. Chemoresistance-related molecules are correlated with poor prognosis and may be effective targets for cancer treatment. Here, we aimed to identify novel molecules correlated with chemoresistance and poor prognosis in pancreatic cancer. We established 10 patient-derived xenograft (PDX) lines from patients with pancreatic cancer and performed next-generation sequencing (NGS) of tumor tissues from PDXs after treatment with standard drugs. We established a gene-transferred tumor cell line to express chemoresistance-related molecules and analyzed the chemoresistance of the established cell line against standard drugs. Finally, we performed immunohistochemical (IHC) analysis of chemoresistance-related molecules using 80 pancreatic cancer tissues. From NGS analysis, we identified olfactomedin-4 (OLFM4) as having high expression in the PDX group treated with anticancer drugs. In IHC analysis, OLFM4 expression was also high in PDXs administered anticancer drugs compared with that in untreated PDXs. Chemoresistance was observed by in vitro analysis of tumor cell lines with forced expression of OLFM4. In an assessment of tissue specimens from 80 patients with pancreatic cancer, Kaplan-Meier analysis showed that patients in the low OLFM4 expression group had a better survival rate than patients in the high OLFM4 expression group. Additionally, multivariate analysis showed that high expression of OLFM4 was an independent prognostic factor predicting poor outcomes. Overall, our study revealed that high expression of OLFM4 was involved in chemoresistance and was an independent prognostic factor in pancreatic cancer. OLFM4 may be a candidate therapeutic target in pancreatic cancer.

Abundant accumulation of digital histopathological images has led to the increased demand for their analysis, such as computer-aided diagnosis using machine learning techniques. However, digital pathological images and related tasks have some issues to be considered. In this mini-review, we introduce the application of digital pathological image analysis using machine learning algorithms, address some problems specific to such analysis, and propose possible solutions. Keywords: Histopathology, Deep learning, Machine learning, Whole slide images, Computer assisted diagnosis, Digital image analysis

Summary: Recent successes in tumor immunotherapies have highlighted the importance of tumor immunity. However, most of the work conducted to date has been on T cell immunity, while the role of B cell immunity in cancer remains more elusive. In this study, immunogenetic repertoire profiling for tumor-infiltrating B and T cells in gastric cancers was carried out to help reveal the architecture of B cell immunity in cancer. Humoral immunity in cancer was shown to involve oligoclonal expansions of tumor-specific and private B cell repertoires. We find that B cell repertoires in cancer are shaped by somatic hypermutation (SHM) either with or without positive selection biases, the latter of which tended to be auto-reactive. Importantly, we identified sulfated glycosaminoglycans (GAGs) as major functional B cell antigens among gastric tumors. Furthermore, natural anti-sulfated GAG antibodies discovered in gastric cancer tissues showed robust growth-suppressive functions against a wide variety of human malignancies of various organs. : Katoh et al. profile immunoglobulin repertoires of tumor-infiltrating B cells, identifying sulfated glycosaminoglycan as a major functional B cell antigen in tumors. Repertoire profiling for tumor-infiltrating B cells may pave the way to the development of therapeutic antibodies. Keywords: repertoire sequence, immunogenetics, tumor immunity, tumor-infiltrating B cell, sulfated glycosaminoglycan, therapeutic antibody, gastric cancer

Tumor suppression by the extracts of Azadirachta indica (neem) works via anti-proliferation, cell cycle arrest, and apoptosis, demonstrated previously using cancer cell lines and live animal models. However, very little is known about the molecular targets and pathways that neem extracts and their associated compounds act through. Here, we address this using a genome-wide functional pooled shRNA screen on head and neck squamous cell carcinoma cell lines treated with crude neem leaf extracts, known for their anti-tumorigenic activity. We analyzed differences in global clonal sizes of the shRNA-infected cells cultured under no treatment and treatment with neem leaf extract conditions, assayed using next-generation sequencing. We found 225 genes affected the cancer cell growth in the shRNA-infected cells treated with neem extract. Pathway enrichment analyses of whole-genome gene expression data from cells temporally treated with neem extract revealed important roles played by the TGF-β pathway and HSF-1-related gene network. Our results indicate that neem extract affects various important molecular signaling pathways in head and neck cancer cells, some of which may be therapeutic targets for this devastating tumor.

Epstein-Barr virus (EBV)-associated gastric carcinoma (EBVaGC), one of four major gastric cancer types, consists of clonal growth of EBV-infected epithelial cells. However, the significance of viral loads in each tumor cell has not been evaluated. EBV-DNA is stably maintained in episomal form in the nucleus of each cancer cell. To estimate EBV copy number per genome (EBV-CN), qPCR of viral EBNA1 and host GAPDH, standardized by Namalwa DNA (one copy/genome), was applied to the formalin-fixed paraffin embedded (FFPE) surgically resected EBVaGC specimens (n = 43) and EBVaGC cell lines (SNU-719 and NCC-24). In surgical specimens, the cancer cell ratio (CCR) was determined with image analysis, and EBV-CN was obtained by adjusting qPCR value with CCR. Fluorescent in situ hybridization (FISH) was also applied to the FFPE sections using the whole EBV-genome as a probe. In surgical specimens, EBV-CN obtained by qPCR/CCR was between 1.2 and 185 copies with a median of 9.9. EBV-CN of SNU-719 and NCC-24 was 42.0 and 1.1, respectively. A linear correlation was observed with qPCR/CCR data up to 20 copies/genome (40 signals/nucleus), the limit of FISH analysis. In addition, substantial variation in the number of EBV foci was observed. Based on qPCR/CCR, high EBV-CN (>10 copies) correlated with PD-L1 expression in cancer cells (P = 0.015), but not with other pathological indicators. Furthermore, EBVaGC with high EBV-CN showed worse disease-specific survival (P = 0.041). Our findings suggest that cancer cell viral loads may contribute to expression of the immune checkpoint molecule and promotion of cancer progression in EBVaGC.

Evaluating HER2 gene amplification is an essential component of therapeutic decision-making for advanced or metastatic gastric cancer. A simple method that is applicable to small, formalin-fixed, paraffin-embedded biopsy specimens is desirable as an adjunct to or as a substitute for currently used HER2 immunohistochemistry and in situ hybridization protocols. In this study, we developed a microfluidics-based digital PCR method for determining HER2 and chromosome 17 centromere (CEP17) copy numbers and estimating tumor content ratio (TCR). The HER2/CEP17 ratio is determined by three variables-TCR and absolute copy numbers of HER2 and CEP17-by examining tumor cells; only the ratio of the latter two can be obtained by digital PCR using the whole specimen without purifying tumor cells. TCR was determined by semi-automatic image analysis. We developed a Tumor Content chart, which is a plane of rectangular coordinates consisting of HER2/CEP17 digital PCR data and TCR that delineates amplified, non-amplified, and equivocal areas. By applying this method, 44 clinical gastric cancer biopsy samples were classified as amplified (n = 13), non-amplified (n = 25), or equivocal (n = 6). By comparison, 11 samples were positive, 11 were negative, and 22 were equivocally immunohistochemistry. Thus, our novel method reduced the number of equivocal samples from 22 to 6, thereby obviating the need for confirmation by fluorescence or dual-probe in situ hybridization to < 30% of cases. Tumor content chart-assisted digital PCR analysis is also applicable to multiple sites in surgically resected tissues. These results indicate that this analysis is a useful alternative to HER2 immunohistochemistry in gastric cancers that can serve as a basis for the automated evaluation of HER2 status.

Interstitial cystitis (IC) is a chronic bladder disease with urinary frequency, bladder discomfort or bladder pain of unknown etiology. Based on cystoscopic findings, patients with IC are classified as either Hunner-type/classic IC (HIC), presenting with a specific Hunner lesion, or non-Hunner-type IC (NHIC), presenting with no Hunner lesion, but post-hydrodistension mucosal bleeding. Inflammatory cell infiltration, composed predominantly of lymphocytes, plasma cells and epithelial denudation, has in the past been documented as a major pathological IC finding. However, the significance of the pathological evaluation of IC, especially with regard to the difference between HIC and NHIC, has been downplayed in recent years. In this study, we performed immunohistochemical quantification of infiltrating T-lymphocytes, B-lymphocytes and plasma cells, and measured the amount of residual epithelium in urinary bladder biopsy specimens taken from patients with HIC and NHIC, and those with no IC, using image analysis software. In addition, in situ hybridization of the light chains was performed to examine clonal B-cell expansion. Lymphoplasmacytic infiltration was significantly more severe in HIC specimens than in NHIC specimens (P

Analysis of acetylation and tri-methylation of the same residue of histone molecules might identify a subset of hepatocellular carcinoma (HCC) with aggressive behavior. In the present study, we examined acetylation and tri-methylation of lysine 27 on histone H3 (H3K27ac and H3K27me3, respectively) because these two modifications are known to exhibit opposite effects (enhancing and silencing) on gene expression. Neoplastic and non-neoplastic tissues from 198 HCC cases were immunostained with specific monoclonal antibodies against H3K27ac and H3K27me3. The stained tissues were evaluated by an image analyzing program to generate histological scores (H-scores, range 0-300), which were determined by multiplying the percentage of positive-stained cells with the classified immunohistochemical marker intensity (0-3). HCC tissues showed significantly higher H3K27ac (156.7±86.8) and H3K27me3 H-scores (151.8±78.1) compared with the background liver (40.3±33.0 and 64.7±45.6, respectively) (both P

Merkel cell polyomavirus (MCPyV) has recently been identified in Merkel cell carcinoma (MCC), an aggressive cancer that occurs in sun-exposed skin. Conventional technologies, such as polymerase chain reaction (PCR) and immunohistochemistry, have produced conflicting results for MCPyV infections in non-MCC tumors. Therefore, we performed quantitative analyses of the MCPyV copy number in various skin tumor tissues, including MCC (n = 9) and other sun exposure-related skin tumors (basal cell carcinoma [BCC, n = 45], actinic keratosis [AK, n = 52], Bowen's disease [n = 34], seborrheic keratosis [n = 5], primary cutaneous anaplastic large-cell lymphoma [n = 5], malignant melanoma [n = 5], and melanocytic nevus [n = 6]). In a conventional PCR analysis, MCPyV DNA was detected in MCC (9 cases; 100%), BCC (1 case; 2%), and AK (3 cases; 6%). We then used digital PCR technology to estimate the absolute viral copy number per haploid human genome in these tissues. The viral copy number per haploid genome was estimated to be around 1 in most MCC tissues, and there were marked differences between the MCC (0.119-42.8) and AK (0.02-0.07) groups. PCR-positive BCC tissue showed a similar viral load as MCC tissue (0.662). Immunohistochemistry with a monoclonal antibody against the MCPyV T antigen (CM2B4) demonstrated positive nuclear localization in most of the high-viral-load tumor groups (8 of 9 MCC and 1 BCC), but not in the low-viral-load or PCR-negative tumor groups. These results demonstrated that MCPyV infection is possibly involved in a minority of sun-exposed skin tumors, including BCC and AK, and that these tumors display different modes of infection.

Embryogenesis is tightly regulated by multiple levels of epigenetic regulation such as DNA methylation, histone modification, and chromatin remodeling. DNA methylation patterns are erased in primordial germ cells and in the interval immediately following fertilization. Subsequent developmental reprogramming occurs by de novo methylation and demethylation. Variance in DNA methylation patterns between different cell types is not well understood. Here, using methylated DNA immunoprecipitation and tiling array technology, we have comprehensively analyzed DNA methylation patterns at proximal promoter regions in mouse embryonic stem (ES) cells, ES cell-derived early germ layers (ectoderm, endoderm and mesoderm) and four adult tissues (brain, liver, skeletal muscle and sperm). Most of the methylated regions are methylated across all three germ layers and in the three adult somatic tissues. This commonly methylated gene set is enriched in germ cell-associated genes that are generally transcriptionally inactive in somatic cells. We also compared DNA methylation patterns by global mapping of histone H3 lysine 4/27 trimethylation, and found that gain of DNA methylation correlates with loss of histone H3 lysine 4 trimethylation. Our combined findings indicate that differentiation of ES cells into the three germ layers is accompanied by an increased number of commonly methylated DNA regions and that these tissue-specific alterations in methylation occur for only a small number of genes. DNA methylation at the proximal promoter regions of commonly methylated genes thus appears to be an irreversible mark which functions to fix somatic lineage by repressing the transcription of germ cell-specific genes.