Single nucleotide polymorphism array analysis of microsatellite-stable, diploid/near-diploid colorectal carcinomas without the CpG island methylator phenotype.

Colorectal carcinomas are considered to progress by chromosomal instability (CIN), or microsatellite instability (MSI) and/or epigenetic gene silencing; however, in previous studies we observed a small fraction of tumours without this molecular phenotype. To further investigate these 'X-type' tumours, neoplastic glands from five tumours were isolated by laser-capture microdissection and used for single nucleotide polymorphism (SNP) array analyses. DNA from our own low-passage primary colorectal carcinoma cell lines (n=9) was used for comparison. Two of these 'X-type' tumours had very low numbers of aberrations (totals of four and five, respectively), consisting of trisomies and arm amplifications. Conversely, aberrations were markedly more frequent in the control cases and three of the 'X-type' tumours (range, 11-40). These aberrations included deletions of chromosomes and chromosome arms, uniparental disomies (UPD), trisomies and arm amplifications. Recurrent microdeletions (<1 MB) were observed at 3p14.2 (FHIT), 16p13.2 (A2BP1) and 20p12.1 (MACROD2). Microsatellite analyses with polymorphic markers at five 'canonical' colorectal carcinoma loci demonstrated a complete loss of one allele in all but one case. When compared to the SNP arrays, concordant results were observed in 93% of tests; however, this was only if DNA from cell lines or laser-capture microdissections was used. In conclusion, colorectal carcinomas may develop without the classic molecular features of CIN, MSI and/or CpG island methylator phenotype (CIMP), but this is a rare event. UPD is frequent but does not define a separate molecular phenotype. Furthermore, our study supports the notion that SNP arrays are reliable for genome-wide detection of deletions and UPD, but discourages the use of microsatellite analyses to detect loss of heterozygosity with DNA from whole tissues.

Ex-vivo clonally expanded B lymphocytes infiltrating colorectal carcinoma are of mature immunophenotype and produce functional IgG.

BACKGROUND: Tumor infiltrating B cells (TiBc) have not yet been investigated in detail. This may at least in part be due to technical difficulties. Here we describe a straightforward and reproducible method to isolate and culture TiBc from primary colorectal carcinomas (CRC). METHODS/RESULTS: TiBc cultures were generated by Epstein-Barr virus (EBV) immortalization. With this method, monoclonal TiBc cultures were obtained for 14/19 CRCs. As assessed by flow cytometry and ELISA, TiBc showed an activated immunophenotype (CD23(+), CD80(+)) and produced immunoglobulin (Ig; IgG secretion in 55% of the cultures). In functional in vitro analysis, most of the IgGs specifically bound to allogeneic CRC target cells. These data suggest that TiBc are antigen-experienced and thus may exhibit functionality in situ. Additionally, mini-cultures generated from 12 further CRCs revealed TiBc outgrowth exclusively in the presence of EBV. CONCLUSION: In summary, this simple method provides a cellular tool and our data set the stage for analysing the bivalent role of TiBc; being antigen-presenting cells on the one hand and tumor-specific antibody producers on the other. Additionally, the generation of long-term TiBc cultures and their monoclonal Ig may serve to identify novel tumor-specific antigens.

Establishment, characterization and chemosensitivity of three mismatch repair deficient cell lines from sporadic and inherited colorectal carcinomas.

BACKGROUND: Colorectal cancer (CRC) represents a morphologic and molecular heterogenic disease. This heterogeneity substantially impairs drug effectiveness and prognosis. The subtype of mismatch repair deficient (MMR-D) CRCs, accounting for about 15% of all cases, shows particular differential responses up to resistance towards currently approved cytostatic drugs. Pre-clinical in vitro models representing molecular features of MMR-D tumors are thus mandatory for identifying biomarkers that finally help to predict responses towards new cytostatic drugs. Here, we describe the successful establishment and characterization of three patient-derived MMR-D cell lines (HROC24, HROC87, and HROC113) along with their corresponding xenografts. METHODOLOGY: MMR-D cell lines (HROC24, HROC87, and HROC113) were established from a total of ten clinicopathological well-defined MMR-D cases (120 CRC cases in total). Cells were comprehensively characterized by phenotype, morphology, growth kinetics, invasiveness, and molecular profile. Additionally, response to clinically relevant chemotherapeutics was examined in vitro and in vivo. PRINCIPAL FINDINGS: Two MMR-D lines showing CIMP-H derived from sporadic CRC (HROC24: K-ras(wt), B-raf(mut), HROC87: K-ras(wt), B-raf(mut)), whereas the HROC113 cell line (K-ras(mut), B-raf(wt)) was HNPCC-associated. A diploid DNA-status could be verified by flow cytometry and SNP Array analysis. All cell lines were characterized as epithelial (EpCAM(+)) tumor cells, showing surface tumor marker expression (CEACAM(+)). MHC-class II was inducible by Interferon-γ stimulation. Growth kinetics as well as invasive potential was quite heterogeneous between individual lines. Besides, MMR-D cell lines exhibited distinct responsiveness towards chemotherapeutics, even when comparing in vitro and in vivo sensitivity. CONCLUSIONS: These newly established and well-characterized, low-passage MMR-D cell lines provide a useful tool for future investigations on the biological characteristics of MMR-D CRCs, both of sporadic and hereditary origin. Additionally, matched patient-derived immune cells allow for comparative genetic studies.

Cryopreservation of human colorectal carcinomas prior to xenografting.

BACKGROUND: Molecular heterogeneity of colorectal carcinoma (CRC) is well recognized, forming the rationale for molecular tests required before administration of some of the novel targeted therapies that now are rapidly entering the clinics. For clinical research at least, but possibly even for future individualized tumor treatment on a routine basis, propagation of patients' CRC tissue may be highly desirable for detailed molecular, biochemical or functional analyses. However, complex logistics requiring close liaison between surgery, pathology, laboratory researchers and animal care facilities are a major drawback in this. We here describe and evaluate a very simple cryopreservation procedure for colorectal carcinoma tissue prior to xenografting that will considerably reduce this logistic complexity. METHODS: Fourty-eight CRC collected ad hoc were xenografted subcutaneously into immunodeficient mice either fresh from surgery (N = 23) or after cryopreservation (N = 31; up to 643 days). RESULTS: Take rates after cryopreservation were satisfactory (71%) though somewhat lower than with tumor tissues fresh from surgery (74%), but this difference was not statistically significant. Re-transplantation of cryopreserved established xenografts (N = 11) was always successful. Of note, in this series, all of the major molecular types of CRC were xenografted successfully, even after cryopreservation. CONCLUSIONS: Our procedure facilitates collection, long-time storage and propagation of clinical CRC specimens (even from different centres) for (pre)clinical studies of novel therapies or for basic research.

Chromosomally and microsatellite stable colorectal carcinomas without the CpG island methylator phenotype in a molecular classification.

We hypothesized that in a comprehensive analysis of colorectal carcinomas (CRC) the three currently known major molecular mechanisms of carcinogenesis (i.e., chromosomal instability, microsatellite instability, and CpG island methylator phenotype, CIMP) would associate with the molecular features indicative of these pathways, allowing a molecular classification. A prospectively collected clinicopathologically well-characterized series of 130 CRCs was tested for chromosomal instability (DNA-flow cytometry and analysis of allelic imbalance with microsatellite markers 5q21, 8p21, 9q21, 17p13, and 18q21), microsatellite instability (Bethesda panel), CIMP (MethyLight), and mutations of K-ras, B-raf, APC, and p53. Morphology was reviewed, and nuclear beta-catenin translocation was assessed by immunohistochemistry. Based on the molecular features, sporadic high-degree microsatellite instable tumours, tumours of the hereditary non-polyposis coli carcinoma syndrome, and 'sporadic standard-type' CRC could be delineated (14, 4, and 55, respectively). However, overlap between classes was seen for 46 of the remaining tumours where widespread or occasional methylations (excluding MLH1) were observed, and the majority had chromosomal instability. Importantly, a group of 11 tumours was observed without either microsatellite or chromosomal instability, nor any methylation. Morphologically, these tumours were without any distinguishing features, all had tumour budding and 10 showed nuclear beta-catenin translocation. Overall, the data give an overview of the molecular classes in CRC that should be taken into account in studies on carcinogenesis and clinicopathological studies. Specifically, the absence of CIN, MSI, and CIMP in an 8.46% fraction of tumours delineates a group to be aware of.

High-degree tumor budding and podia-formation in sporadic colorectal carcinomas with K-ras gene mutations.

In vitro ras activation enhances the epithelial-mesenchymal transition of colorectal carcinoma cells. But ras effects are known to be highly dependent on cell types and the tissue context. Therefore, this study was made to test the hypothesis that in clinical colorectal carcinoma specimens, aggressive invasion phenotypes, specifically tumor budding and podia formation, would correlate with K-ras gene mutations. In a series of 95 clinically sporadic primary colorectal carcinomas collected ad hoc, tumor budding and podia formation were counted using pan-cytokeratin immunohistochemistry, and K-ras gene mutations in codons 12 and 13 were determined. Consistent with the hypothesis, tumor budding and podia formation were observed to be significantly higher in the 32 (34.7%) of the tumors with K-ras gene mutations (29 mutations in codon 12, 3 in codon 13), and this correlation was observed independent of the patterns of invasion (expansive versus infiltrative). Microsatellite status, numbers of losses of heterozygosity, adenomatous polyposis coli and p53 gene mutations, and degree of promoter methylations (CIMP status) were not associated with K-ras gene mutations. Besides their effects on the tumor cell cycles, oncogenic K-ras gene mutations in colorectal carcinomas could be important for aggressive tumor invasion. This may be important in metastasizing disease and could provide a rationale for developing drugs that interrupt ras-signaling cascades.

Do thymidylate synthase gene promoter polymorphism and the C/G single nucleotide polymorphism predict effectiveness of adjuvant 5-fluorouracil-based chemotherapy in stage III colonic adenocarcinoma?

Since 5-fluorouracil (5-FU)-based chemotherapy has become standard adjuvant treatment for patients with node-positive colonic adenocarcinoma, there has arisen the need for predictive factors. Thymidylate synthase (TS) is a major target of 5-FU's action, and high TS expression in carcinoma cells could reduce its cytostatic effect. Both, a 28-base pair repeat polymorphism and a cytosine vs. guanine single nucleotide polymorphism in the promoter region of the TS gene are known to modulate its expression. All patients with a single, non-metachronous node-positive colonic adenocarcinoma who underwent a potentially curative resection at this institution in the years 1994-2002, and who received adjuvant 5-FU (n=95) were included in this study. Ninety-four of the 95 patients were successfully genotyped: 70 patients were classified as TS gene low-expressors (2R-2R, 2R-3C and 3C-3C), and 24 patients were classified as high-expressors (2R-3G, 3C-3G and 3G-3G). Contrary to the hypothesis, Kaplan-Meier survival analysis did not reveal any differences between the groups (power of 0.8 to detect an absolute survival difference >30%). In a Cox model, venous angioinvasion and the infiltrative pattern of tumour invasion were strong adverse factors. These results argue against a practical role for the TS gene repeat polymorphism or the C/G single nucleotide polymorphism as a predictive factor. However, by careful histopathological examination a high-risk group of node-positive patients can be defined that could be candidates for studies of alternative (more aggressive) adjuvant treatment.

p16(INK4a) promoter methylation and 9p21 allelic loss in colorectal carcinomas: relation with immunohistochemical p16(INK4a) expression and with tumor budding.

In colorectal carcinomas, p16(INK4a) inactivation is known to occur by allelic loss and by promoter methylation, but mutations are rare. p16(INK4a) is up-regulated in tumor buds, and the consequent shutdown of proliferation may be a prerequisite for tumor budding. Fifty-seven colorectal carcinomas from a consecutive series were investigated. Using DNA from tissue homogenates, p16(INK4a) promoter methylation was seen in 17 of 57 tumors by methylation-specific polymerase chain reaction, and this could be confirmed using DNA from laser-capture microdissected material in 16 of these cases. A total loss of immunohistochemical p16(INK4a) expression was seen in 6 of 17 tumors with promoter methylation. Quantification of immunohistochemical p16(INK4a) expression for the remaining 11 cases revealed statistically lower frequencies of expression as compared with cases without p16(INK4a) promoter methylation. 9p21 allelic loss was observed in 9 cases, but p16(INK4a) expression in these carcinomas was not reduced. Attempted linear regression of p16(INK4a) expression in tumor buds on the degree of tumor budding, as counted on pan-cytokeratin immunostains, did not show a correlation. p16(INK4a) promoter methylation can completely abrogate p16(INK4a) expression in colorectal carcinomas. In many cases, however, it has an appreciable but only modulatory influence on p16(INK4a) expression. Possibly, methylations are heterozygous, and/or mosaic in colorectal carcinomas and/or methylations are not totally stable but can be lost between carcinoma cell replication cycles. Up-regulation of p16(INK4a) does not seem to be a strict requirement for tumor budding, hence, the absence of a correlation.

Expression profiling of colorectal carcinomas using tissue microarrays: cell cycle regulatory proteins p21, p27, and p53 as immunohistochemical prognostic markers in univariate and multivariate analysis.

With the rapidly growing understanding of tumor biology, molecular staging of cancer is expected to improve prognostication. This would be particularly important for cancers amenable to adjuvant treatment, such as colorectal carcinomas. To generate data for this, the tissue microarray technique may prove useful. Tissue microarrays were constructed with triplicate cores (0.6 mm diameter) from the invasive margins of a consecutive single-institution series of 184 colorectal carcinomas. Immunostaining for p53, p21, p27, Ecadherin, and beta-catenin was scored. Tumor cell proliferation was assessed by mitotic indices and Ki-67 labeling, apoptosis by quantification of apoptotic bodies. Reduced nuclear immunostaining for p21 (<10%) and p27 (< or =50%) and reduced membranous expression of Ecadherin were significantly associated with a poorer clinical course by univariate analysis. beta-catenin immunostaining had no prognostic impact. Mitotic and apoptotic indices as well as Ki-67 labeling below the median were indicators of poor prognosis. Complete absence of p53 nuclear staining was a significant adverse prognostic factor. By Cox regression, p53 = 0%, p53 = 0%, in combination with p27 < or = 50%, the mitotic index and the combined mitotic and apoptotic index added prognostic information to UICC stage. The authors found that growth pattern, lymphohistiocytic response, lymphatic permeation, and venous spread, too, each was a strong prognosticator in addition to UICC stage. The results support that tissue microarrays are a useful tool for screening immunohistochemical markers for prognostic use. An immunopanel of p21, p27, and p53 could be useful for prognostication in colorectal carcinoma in addition to UICC stage.

Prognostic role of CD8+ tumor-infiltrating lymphocytes in stage III colorectal cancer with and without microsatellite instability.

Previous studies have identified high numbers of intraepithelial T lymphocytes to be associated with good prognosis in various types of cancer. Few studies addressing this issue have been published for colorectal cancer. In a simulated prospective approach ("phase II prognostic factor study"), all nonmetachronous International Union Against Cancer (UICC) stage III colorectal cancers that were accessioned in the years 1994 to 1999 were included in the study (152 cases). Follow-up information as to vital status and occurrence of metachronous metastases could be obtained for all patients in the years 2001 and 2002. CD8+ intratumoral lymphocytes were quantified after immunostaining and referred to tumor cell area (CD8+ densities). Microsatellite status was determined by using the Bethesda panel of microsatellite markers. CD8+ densities ranged from 0 per square millimeter to 1436 per square millimeter of tumor area in a nonnormal distribution that was skewed toward low values. Univariate survival analyses revealed the 66th percentile as a stringent cutoff (CD8+high versus CD8+low), with CD8+high cases taking a significantly better clinical course. This prognostic impact appeared even more pronounced in the subset of patients with colon carcinomas who were receiving 5-fluouracil/leucovorin as adjuvant treatment (79 patients). Seventeen patients had carcinomas with high microsatellite instability (MSI-H). MSI-H-CD8+high cases (n = 11) showed an excellent prognosis, with tumor-free survival for 9 of the 11 patients. The prognostic effect of CD8+high was retained in Cox regression analyses when including UICC substages (IIIA to IIIC). Our results identify CD8+ tumor-infiltrating lymphocytes as a promising candidate for further evaluation in the ongoing search for prognostic and predictive factors of colorectal cancer, particularly if combined with microsatellite status.

Human papillomavirus 6/11, 16 and 18 in oral carcinomas and benign oral lesions.

We have examined 118 oral squamous cell carcinomas, 72 oral leukoplakias, 12 cases of cheilitis and 65 of oral lichen planus for the presence of human papillomavirus (HPV) 6/11, 16 and 18 DNA by PCR/Southern blot hybridization. HPV DNA were found in 51/118 carcinomas (43.2%), in 16/72 (22.2%) leukoplakias, 3/12 (25.0%) cheilitic lesions and 10/65 (15.4%) lichen planus cases. These differences were even stronger when analyzing separately for the high-risk types HPV 16 and 18 as compared to low-risk types 6/11. HPV 16 and 18 DNA were present in 41/118 (34.7%) oral carcinomas, 12/72 (16.7%) leukoplakias, 2/12 (16.7%) cheilitic lesions and 6/65 (9.2%) lichen planus. In contrast to this, oral carcinomas displayed the lowest HPV 6/11 detection rate (4.2%), compared with 11.1% for leukoplakias, 8.3% for cheilitic lesions and 7.7% in lichen planus. These results indicate a successive increase of the detection rate of HPV 16 and 18 from low level in non or questionably preneoplastic lesions (lichen planus) to preneoplastic lesions (leukoplakia and cheilitis) and to oral carcinoma. In conclusion, our results suggest an association of oral carcinogenesis and infection with the high-risk HPV types 16 and 18.