Abstract.The invasive front of carcinomas may vary in complexity from smooth to highly complex when the front splits up into small cell clusters or even single cancer cells. The degree of complexity is usually estimated visually and semiquantitatively by a pathologist, although more objective methods based on computer-assisted image analysis are available. In this study, we compared the visual estimation of the irregularity of the tumour invasion front of colon carcinomas to different quantitative image analytical techniques and defined a complexity index for the invasive margin. Sections from 29 archived colon carcinomas were stained immunohistochemically for cytokeratin 8. Images of the tumour invasion front were read into a computer and thresholded so that the tumour tissue became black and the background white or so that the tumour front was outlined by a single pixel line. The invasive front was visually classified into four degrees of irregularity by a pathologist. The complexity of the front was then assessed using four different image analysis techniques, i.e. the estimation of fractal dimension, tumour front length, number of tumour cell clusters and lacunarity. Fractal dimension and tumour cell clusters together gave the best correlation to visual grading using a discriminant analysis. A cluster analysis and a tree diagram analysis were then performed and were found to be superior to visual estimation. The clusters represent different degrees of complexity and the result of the tree diagram analysis can be used to assign complexity indices to colon tumours. The fractal dimension separated tumours up to a certain level (1.5-1.6) of complexity. When the tumour front split up into small cell clusters, the counting of tumour cell clusters separated the cells over and above the fractal dimension. This new technique can be used to objectively and quantitatively describe the complexity of the invasive front of tumours.

Background and aim: The mechanisms behind increased f-Calprotectin (FC) in healthy

IntroductionTargeting tumor suppressor genes by epigenetic silencing especially in the promotor region is one of the probable mechanisms of carcinogenesis. 1 However, differential methylation may also occur outside of the promoter region and previous studies have shown that gene expression and tissue differentiation are powerfully affected by such methylation sites. 1,2 So far, there are limited data available on claudin (CLDN), epigenetics, and the relation to claudin protein expression as well as its role in the development of colorectal carcinoma (CRC). While DNA methylation of claudin 6 has been shown to be involved in the invasion of breast cancer cells 3 Abstract Altered claudin expression has been described in colon, prostatic, ovarian, and breast carcinoma. However, the role of epigenetic modifications in these genes and their role in colorectal cancer is unknown. We aimed our study to investigate whether claudin protein expression and methylation of CLDN can influence the tumorigenesis of colorectal cancer. A total of 31 patients diagnosed with colorectal carcinoma was used in this study. Immunohistochemical staining was used to study protein expression in both tumor and the adjacent nonneoplastic mucosa of claudin 1, 4, and 7. To detect the DNA methylation pattern of CLDN1, 4, and 7, genomic DNA was extracted from both the tumor and the adjacent nonneoplastic mucosa. Methylation analysis was carried out using bisulfite pyrosequencing. Cell membrane staining intensity of all claudins was found significantly lower in colorectal cancer tissues when compared to paired normal mucosa (p ≤ 0.001). For claudin 4, the percentage of cells staining positively was also significantly reduced (p = 0.04). In normal mucosa, cytoplasm showed no staining for claudins in any patient, whereas in paired colorectal cancer tissues, significant cytoplasmic staining appeared both for claudin 1 (p = 0.04) and claudin 4 (p = 0.01). Tumor samples were significantly hypomethylated in CLDN1 (p < 0.05). In conclusion, our results show that CLDN1 is significantly hypomethylated in tumor samples and that the membrane staining intensity for claudin 1, 4, and 7 is significantly lower in colorectal cancer tissues than in adjacent nonneoplastic tissue. Colorectal cancer cells showed dystopic cytoplasmic location of claudins.

X‐linked thrombocytopenia with thalassemia (XLTT) is caused by the mutation 216R > Q in exon 4 of the GATA1 gene. Male hemizygous patients display macrothrombocytopenia, splenomegaly, and a β‐thalassemia trait. We describe two XLTT families where three males were initially misdiagnosed as having primary myelofibrosis (PMF) and all five investigated males showed mild‐moderate bone marrow (BM) reticulin fibrosis. Comparative investigations were performed on blood samples and BM biopsies from males with XLTT, PMF patients and healthy controls. Like PMF, XLTT presented with high BM microvessel density, low GATA1 protein levels in megakaryocytes, and elevated blood CD34+ cell counts. But unlike PMF, the BM microvessel pericyte coverage was low in XLTT, and no collagen fibrosis was found. Further, as evaluated by immunohistochemistry, expressions of the growth factors VEGF, AGGF1, and CTGF were low in XLTT megakaryocytes and microvessels but high in PMF. Thus, although the reticulin fibrosis in XLTT might simulate PMF, opposing stromal and megakaryocyte features may facilitate differential diagnosis. Additional comparisons between these disorders may increase the understanding of mechanisms behind BM fibrosis in relation to pathological megakaryopoiesis. Am. J. Hematol. 90:E44–E48, 2015. © 2014 Wiley Periodicals, Inc.

Adhesion proteins are responsible for the structural integrity of epithelial tissue and in tumors this integrity is often lost, resulting in a disorganization of the tissue. In the present study the complexity of the invasive front of colon carcinomas was correlated with cell adhesion protein expression and with polymorphisms in their genes. A complexity index was constructed from 32 colon carcinomas using computer-assisted morphometry estimating fractal dimension and tumor cell clusters followed by tree analysis. Immunohistochemical staining of beta-catenin, E-cadherin, occludin and claudin 2 was used for assessment of protein expression. Genetic screening of tissue from the tumor invasion front with laser microdissection was performed using SSCP and DNA sequencing. Adhesion protein distribution was significantly disturbed in most carcinomas. A single mutation in the gene of beta-catenin was found but there was no correlation between protein expression and genetic polymorphism. Nor was there any correlation between the complexity of the invasive border and protein distribution or genetic alterations. The results indicate that the complexity of colon carcinoma invasion is not dependent on genetic derangements in the genes of adhesion proteins or the protein distribution. Rather, aberrations in the function of other proteins related to the adhesive proteins could be responsible.

Tumor volume increases during growth and due to tumor progression various mutations appear that may cause phenotypic changes. The invasive pattern may thus be affected resulting in a more disorganized growth. This phenomenon might be due to mutations in the genome of the adhesion proteins, which are responsible for the structural integrity of epithelial tissue. Tumor volume was assessed in whole mount sections of 33 colon carcinomas using Cavalieri's principle. Images from the entire invasive border were captured and used for calculating the irregularity of the border (Complexity Index). The expression of the adhesion proteins E‐cadherin, β‐catenin, Claudin 2 and Occludin was assessed after immunohistochemical staining of two randomly selected areas of the invasive front of the tumor. Statistical significance for differences in volume was obtained for tumor Complexity Index, tumor stage (pT) and lymph node status (pN). Expression of adhesion proteins was significantly perturbed in the tumors compared with normal mucosa but only infrequently correlated to tumor differentiation or invasive pattern. The results show that when tumor volume increases the invasive pattern becomes more irregular which is compatible with tumor progression. A direct contribution of adhesion protein derangement to this process appears to be insignificant.

Mannan A, Hahn-Stro¨mberg V. K-ras mutations are correlated to lymph node metastasis and tumor stage, but not to the growth pattern of colon carcinoma. APMIS 2012; 120: 459-68.In colorectal carcinoma, pathological assessment of tumors is essential for determining therapy and prognosis of the disease. Molecular associations of tumor complexity index and genetic alternations can be helpful to understand the tumor progression mechanism. Oncogenic K-ras is one of the major colorectal cancer associated genes, and is mutated in up to 50% of colorectal cancers. In this current study, we correlated tumor complexity index with mutations in K-ras codon 12, 13, and 61 in association with different clinicopathological parameters such as TNM stage, localization, sex, and age. Formalin-fixed paraffin embedded tissue blocks from colon cancer samples was selected from 88 patients diagnosed with adenocarcinoma. Mutations in the K-ras gene were detected using pyrosequencing technique. Tumor complexity index was calculated using immunohistochemically stained images of the tumor outline of the specimens and then analyzing these pictures using Photoshop CS, Fovea Pro, and Image J computer programs. Statistical analysis was performed with SPSS. K-ras mutations were detected in 17 (19.3%) colon cancer samples. Most of the samples were at a lower complexity index. No correlation was observed between K-ras mutations and complexity index. However, K-ras mutations were correlated with regional lymph node metastasis and tumor stages and complexity index with tumor wall penetration. In conclusion, complexity index and K-ras mutations are independent events; however, both correlate with tumor progression and are important in the biologic development of colon carcinoma.

Folate has a central role in the cell metabolism. This study aims to explore the DNA methylation pattern of the folate transporter genes FOLR1, PCFT, and RFC1 as well as the corresponding protein expressions in colorectal cancer (CRC) tissue and adjacent non-cancerous mucosa (ANCM). Our results showed statistically significant differences in the DNA-methylated fraction of all three genes at several gene regions; we identified three differentially methylated CpG sites in the FOLR1 gene, five CpG sites in the PCFT gene, and six CpG sites in the RFC1 gene. There was a pronounced expression of the FRα and RFC proteins in both the CRC and ANCM tissues, though the expression was attenuated in cancer compared to the paired ANCM tissues. The PCFT protein was undetectable or expressed at a very low level in both tissue types. Higher methylated fractions of the CpG sites 3-5 in the RFC1 gene were associated with a lower protein expression, suggestive of epigenetic regulation by DNA methylation of the RFC1 gene in the colorectal cancer. Our results did not show any association between the RFC and FRα protein expression and tumor stage, TNM classification, or tumor location. In conclusion, this is the first study to simultaneously evaluate both DNA methylation and protein expression of all three folate transporter genes, FOLR1, PCFT, and RFC1, in colorectal cancer. The results encourage further investigation into the possible prognostic implications of folate transporter expression and DNA methylation.