Search

Fluorescence in situ hybridization (FISH) is the most widely used technique to detect HER-2/neu gene amplification; however, it is only available in some institutions. In contrast, chromogenic in situ hybridization (CISH) can be evaluated by routine light microscopy. In endometrial carcinoma there are few data concerning HER-2/neu status and prognosis. Therefore, we determined HER-2/neu gene status by CISH using a digoxigenin-labelled probe on 60 formalin-fixed paraffin-embedded endometrial carcinomas. The data were compared with the immunohistochemistry of HER-2/neu (A0485, TAB250), p53, Ki-67, clinicopathological factors, and survival. By conventional light microscopy, HER-2/neu amplification (>/=6 copies >50% cancer cells) was detected in 14% (8/59) tumours, HER-2/neu overexpression (>10% cells moderate/strong complete membrane staining) in 22% (13/60) for A0485, and 18% (11/60) for TAB250, p53 (>10% +cells) in 61% (36/59), and Ki-67 (>50% +cells) in 50% (30/60). Discordant cases for CISH and immunohistochemistry, as well as all (2+) were further analysed by FISH (Vysis). Among 10 cases (2+) and not amplified by CISH, two showed low-level amplification by FISH. Significant correlation was found between amplification and protein overexpression (P

K-RAS mutations are the most frequent molecular genetic alteration in serous ovarian tumors of borderline malignancy (SBOT). The pathogenesis of associated contralateral tumors and extraovarian implants and Müllerian inclusion cysts is obscure. We hypothesized that the comparison of K-RAS mutations in these lesions might help to distinguish multifocal from metastatic foci. Eight cases of SBOT with known K-RAS mutation (RAS+) and two cases without mutation (RAS-) were analyzed for comparison. DNA was extracted from multiple samples of 58 paraffin-embedded and laser-microdissected ovarian and extraovarian lesions (10 ovarian borderline tumors, 8 contralateral tumors, 25 implants, 15 inclusion cysts; total: 97 samples). K-RAS exon 1 was amplified by PCR and analyzed by denaturing gradient gel electrophoresis and cycle sequencing. In 12 of 14 SBOT and in 2 of 2 extraovarian implants, the K-RAS mutation could be found in different areas of the same lesion, indicating monoclonality. All RAS+ ovarian borderline tumors with contralateral tumors (six of six) harbored an identical mutation in both ovaries (in one case, a separate surface borderline tumor component contained a different mutation in addition). In 4 of 5 RAS+ ovarian tumors with extraovarian lesions, RAS mutations were also found in implants (15/21 implants [71%]) and more rarely in inclusion cysts (3 of 12 lesions [25%]). These extraovarian mutations were always identical to the one in the ovary (18 of 18 [100%]). Regarding the contralateral and extraovarian lesions of the two RAS- SBOT, only one extraovarian implant contained a RAS mutation. The demonstration of K-RAS mutations in Müllerian inclusion cysts and implants of SBOT suggests that K-RAS mutations represent a pivotal event during neoplastic transformation of ovarian and extraovarian serous epithelium. Considering our observations, the two putative pathogenetic mechanisms for the development of implants and endosalpingiosis-multifocal tumorigenesis and spread from the ovarian primary tumor-seem to coexist.

Endometrial carcinoma is the most common gynecologic malignancy in perimenopausal and postmenopausal women. A role of mismatch repair genes, like hMLH1 and hMSH2 in their pathogenesis, has been suggested. Loss of their function leads to the accumulation of replication errors (mutator phenotype), which are responsible for further mutations in genes with microsatellite sequences in their coding region, such as Bax. We analyzed the expression of hMLH1, hMSH2, and Bax genes in 89 formalin-fixed paraffin-embedded endometrial carcinomas. The immunostains were scored with regard to percentage of positive tumor cells (0%, <10%, 10 to 50%, >50%), and relative staining intensity (1+, 2+, 3+). The staining results were correlated with clinicopathologic features and survival. Loss of hMSH2 expression (0% positive cells) was observed in 1.1% (1/89) of the tumors; loss of hMLH1 was seen in 12.4% (11/89) of the cases, particularly in endometrioid tumors with mucinous differentation (5/11; 45%; P =.03). No significant association was found between the immunoscores and grade, stage criteria of the International Federation of Obstetrics and Gynecology (FIGO), or age of the patients. Among 11 tumors with loss of Bax expression (12.4%), 4 had also loss of hMLH1 (4/11; 36.4%; P =.017). In multivariate analysis (Cox model), significantly longer survival was found for patients with tumors in FIGO Stage I-II (P <.0001), endometrioid type (P =.001), low grade (P =.001), and absence of hMLH1 expression (P =.027). Our results suggest that loss of function of hMLH1 and Bax occur in a subgroup of endometrial carcinoma. In addition to the classical prognostic factors, absence of hMLH1 expression is associated with better outcome of patients.

The beta-catenin pathway plays a central role in transcriptional signaling and cell-cell interactions in colonic epithelium. Alterations of the expression of beta-catenin, and its binding partners E-cadherin and the adenomatous polyposis coli protein (APC), are frequent events in sporadic colorectal cancer. Ulcerative colitis (UC)-related cancers originate in a field of chronic inflammation and therefore may have different alterations in the beta-catenin pathway than sporadic cancers. To test this hypothesis, expression and subcellular localization of beta-catenin, E-cadherin, and APC were detected by immunohistochemistry in paraffin sections from 33 UC-related and 42 sporadic colorectal cancers. Although beta-catenin and E-cadherin expression were predominantly limited to the lateral cell membrane in normal colonic epithelium, both tumor groups showed an overall shift from membranous to cytoplasmic expression for these proteins. An increase in nuclear localization of beta-catenin and a decrease in cytoplasmic APC expression also were seen in both cancer groups compared with normal epithelium. Abnormal beta-catenin expression was more closely linked to E-cadherin alterations in UC-related cancers than in sporadic cancers. In contrast, abnormal beta-catenin expression was more closely linked to APC alterations in sporadic cancers than in UC-related cancers. These data suggest that alterations of the beta-catenin pathway are important in both UC-related and sporadic colorectal cancers. However, differences in the expression patterns of beta-catenin, E-cadherin, and APC between UC-related and sporadic colorectal cancers suggest that the specific alterations in this pathway may differ in these two cancer groups.

Genetic mechanisms leading to androgen-independent growth in advanced prostatic carcinomas (PC) are still poorly understood. Analysis of genes potentially involved in the regulation of tumor cell proliferation and apoptosis might confer better insight into this process and might lead to improved therapeutic strategies. Fluorescence in situ hybridization (FISH) analysis of dissociated nuclei with DNA probes for MYC (8q24)/#8, cyclin D1 gene (CCND1; 11q13)/#11, ERBB2 (17q13)/#17, the androgen receptor gene (AR; Xq12)/#X, and the retinoblastoma gene (RB; 13q14) was applied to formalin-fixed tissue from 63 patients with advanced PC after androgen deprivation therapy (ADT); matched tumor tissue before ADT was also available in 22 of these cases. The cut-points used were: "increased copy number," > or = 30% of all nuclei with increased FISH signals (centromere and/or gene); "amplification," > or = 15% of nuclei with "increased gene copy number." CCND1 and MYC gene "amplifications" were present before ADT in 25% and 33% of the cases, respectively; the frequency of these "amplifications" increased to 37% and 57% after ADT. Loss of the RB gene was nearly four times more frequent after ADT than before therapy (22% versus 6%). AR and ERBB2 gene "amplifications" occurred only after ADT in 36% and 30% of cases, respectively. With the exception of the AR gene, the copy number increase was low. After treatment, MYC and AR gene "amplifications" correlated with the proliferation rate (Ki-67/MIB1 index; p = 0.01 and p = 0.04), whereas ERBB2 "amplifications" were associated with increased apoptotic index (PCD/TUNEL; p = 0.016). However, no correlation between FISH results and clinical follow-up could be established. FISH analysis of genes putatively involved in PC progression revealed characteristic patterns of aberrations in advanced PC before and after ADT. Distinct changes in gene copy number before and after therapy suggests possible involvement of these genes in the escape from androgen control.

Few and conflicting cytogenetic data are available concerning the chromosomal constitution of (mainly gastric) extranodal marginal zone B-cell non-Hodgkin's lymphoma arising from mucosa-associated lymphoid tissue (MALT)-type lymphoma. The majority of salivary gland MALT lymphomas are thought to develop from longstanding Sjögren's syndrome/benign lymphoepithelial lesion (BLEL). We tried to achieve a better comprehension of related cytogenetic alterations by comparing DNA-ploidy and numerical chromosomal (#) aberrations, assessed by different techniques of DNA cytometry (image cytometry) and interphase cytogenetics using nonradiographic in situ hybridization (centromere specific probes for #3, 7, 12, 18) on 12 cases of BLEL, 13 low-grade MALT lymphomas (LG-MALT-L) and 4 high-grade MALT lymphomas (HG-MALT-L) of salivary gland. Both techniques were applied on tissue sections preferentially, enabling a reliable measurement of histomorphologically identified areas. No case of BLEL showed cytogenetic abnormalities. Three of 4 HG- and 2 of 13 LG-MALT-L exhibited complex chromosomal gains in nonisotopic in situ hybridization, which were reflected by DNA nondiploidy in image cytometry. In 6 of 13 LG- and lof 4 HG-MALT-L, one or two numerical chromosomal aberrations were demonstrated by nonisotopic in situ hybridization, which could not be resolved by image cytometry. In the 11 DNA-diploid LG-MALT-L, trisomies 18, 3, and 12 were found in 36, 12, and 9%, respectively. In conclusion, comparing BLEL, which showed no chromosomal aberrations, with LG- and HG-MALT-L, an increase in frequency and number of numerical aberrations and DNA nondiploidy was seen. Peritetraploid DNA nondiploidy might be characteristic for HG-MALT-L of salivary gland as it is a rare finding in MALT lymphomas of other sites. It is unclear whether the documented chromosomal aberrations in LG-MALT-L, especially increased rate of trisomy 18, indicate a pathogenic impact or merely reflect genetic instability.

Extending our previous efforts to characterize ovarian neoplasms by interphase cytogenetics, we analyzed a series of 32 mucinous tumors by nonisotopic in situ hybridization with seven different centromere-specific probes as well as by flow and image DNA cytometry; we then compared the data with results of p53 and Ki67 immunohistochemistry and MYC DNA-PCR analysis and of the clinical follow-ups. Of the tumors studied, 11 of 14 (78.6%) mucinous carcinomas, 7 of 7 (100%) mucinous tumors of low malignant potential (LMP), and 7 of 11 (63.6%) mucinous cystadenomas demonstrated chromosomal aberrations. The mean number of chromosomal aberrations (+/- SD) was slightly higher in DNA cytometrically nondiploid cases than in diploid cases (2.0 +/- 1.6 versus 1.6 +/- 1.2, not significant) but did not differ significantly among the study groups (carcinomas: 1.7 +/- 1.4; tumors of LMP; 1.9 +/- 0.7; adenomas: 1.4 +/- 1.4). Aberrations affected chromosomes 1 (14 of 27 cases) and 6 (12 of 31) most frequently, followed by chromosomes 17 (7 of 28), 7 (6 of 29), and X (6 of 28). Signal gain for centromere 1, which was the most prevalent finding (13 of 27), was observed in 3 of 10 mucinous cystadenomas, 2 of 4 mucinous tumors of LMP, and 8 of 13 mucinous carcinomas. All six moderately and poorly differentiated carcinomas demonstrated this aberration. Signal gain of centromere 6 (3 of 13) and centromere 7 (4 of 13) were found only in carcinomas (p < 0.05 and p < 0.025, respectively). The interphase cytogenetic results correlated neither with proliferative activity, immunohistochemical p53 accumulation, MYC DNA amplification, nor postoperative outcome. Compared with serous ovarian neoplasms (Lab Invest 1996, 75:473-485), mucinous tumors demonstrated signal gain for chromosome 1 (p < 0.0001) and signal loss for chromosomes 6 (p < 0.001) and X (p < 0.01) significantly more often. Loss of centromere 17 was more characteristic for serous than for mucinous carcinomas (p < 0.05). Our observations show that chromosomal aberrations in mucinous ovarian neoplasms are apparently not random. These results support the notion that the molecular genetic changes in mucinous neoplasms differ from those in serous tumors.

The cytogenetic and molecular genetic changes in serous tumors of low malignant potential (LMP) of the ovary have not been well characterized so far. Therefore, we analyzed 20 serous tumors of LMP, 10 invasive serous ovarian carcinomas, and 7 benign serous cystadenomas by nonisotopic in situ hybridization (seven different centromere-specific probes) as well as by flow and image DNA cytometry and compared the data with results of p53 and Ki67 immunohistochemistry, MYC DNA PCR analysis and with the clinical follow-up. All but two tumors of LMP were DNA cytometrically diploid; 9 of 10 invasive carcinomas proved to be DNA nondiploid (p < 0.0001). Nonisotopic in situ hybridization revealed a mean number of 1.5 chromosomal aberrations in tumors of LMP, which differed statistically significantly from cystadenomas (mean, 0.4) and from invasive carcinomas (mean, 3.4) (rho < 0.01). The main changes in tumors of LMP were +6 (7 of 18 cases) and +7 (6 of 19) followed by -3 (5 of 20), -1 (4 of 17) and +X (3 of 20). In the group of invasive carcinomas, the number of cases with signal gains for chromosomes 6 (5 of 8), 7 (7 of 10) and X (4 of 10) and signal loss for chromosome 1 (4 of 9) was even larger. In addition, statistically significantly more cases showed gain of 8 (5 of 10) and loss of 17 (5 of 10) (p < 0.05). Proliferative activity (Ki67 index) was positively correlated with the number of chromosomal aberrations (p < 0.05). There was no association between changes in the centromere signal number of chromosomes 8 and 17 and MYC DNA amplification and immunohistochemical p53 accumulation, respectively. Clinical follow-up showed prognostic differences between tumors of LMP and invasive carcinomas as expected (rho < 0.001) but did not reveal differences within the group of tumors of LMP with regard to the number or type of the chromosomal abnormalities detected. In conclusion, the patterns of chromosomal gains and losses in serous tumors of LMP and invasive serous carcinomas of the ovary do not seem random and suggest a close relation between these neoplasms compatible with sequential stages in a multistep model of ovarian carcinogenesis.

Normal adipose tissue as well as 13 benign and 17 malignant lipomatous tumors (lipomas, hibernomas, lipoblastomas, and liposarcomas) were immunohistochemically analyzed for their expression of the basement membrane components collagen IV, laminin, heparan sulfate proteoglycan, and fibronectin. Monovacuolar cells in normal white fat tissue and in lipomas generally exhibited a distinctive pericellular basement membrane composed of collagen IV and laminin, whereas heparan sulfate proteoglycan and fibronectin were almost completely missing. In brown fat tissue and hibernomas the characteristic multivacuolated cells differed from the monovacuolated white fat cells by the additional content of heparan sulfate proteoglycan and fibronectin and the more intensive staining for the other components tested. In contrast, multi-/monovacuolated cells in lipoblastomas exhibited no characteristic immunohistochemical feature because they reacted irregularly and only faintly for collagen IV, laminin, and heparan sulfate proteoglycan. Spindle cell areas in benign lipomatous tumors displayed more fibronectin than laminin and heparan sulfate proteoglycan indicating a "preadipose" fibroblast-like cellular differentiation. In liposarcomas, only well-differentiated lipoma-like neoplasms revealed a basement membrane pattern resembling that of white fat tissue. Otherwise, in nonlipoma-like liposarcomas, a marked decrease particularly of collagen IV staining was evident. Poorly differentiated liposarcomas mostly failed to express any of the basement membrane components, but showed a relative increase of fibronectin. Our results provide evidence that the staining pattern of basement membrane components parallels the histogenetic derivation of benign lipomatous tumors from either brown or white adipose tissue and, additionally, may reflect such a derivation in liposarcomas.