Detection of mammaglobin in the sera of patients with breast cancer.

Current procedures for the diagnosis of breast cancer are cumbersome and invasive, making detection of this disease difficult. A rapid screening test for early detection of breast cancer would allow for better management of this deadly disease. In this report, we show that, with the exception of the skin, mammaglobin mRNA is specifically expressed in mammary tissue and commonly overexpressed in breast cancer. Mammaglobin is not expressed in other types of cancer including colon, lung, ovarian, and prostate cancer. Breast-specific expression of mammaglobin protein was shown using immunohistochemical methods. Mammaglobin is secreted from both established breast cancer cell lines and primary breast carcinoma cells cultured in vitro. Using a monoclonal antibody-based assay for monitoring the presence of mammaglobin in serum, elevated levels of mammaglobin were detected in sera of patients with breast cancer, but not in healthy women. Thus, mammaglobin, which is overexpressed and secreted from breast carcinoma cells, is detectable in sera of patients with breast cancer and may provide a rapid screening test for the diagnosis and management of breast cancer.

Wide spectrum screening keratin as a marker of metaplastic spindle cell carcinoma of the breast: an immunohistochemical study of 24 patients.

AIMS: Metaplastic spindle cell carcinomas may be difficult to distinguish histologically from other spindle cell lesions in the breast. Variable staining with cytokeratin immunomarkers has been reported for metaplastic carcinomas. We evaluated the diagnostic utility of anti-cytokeratin polyclonal antibody, wide spectrum screening keratin, to assess spindle cell breast lesions. METHODS AND RESULTS: Twenty-four patients with spindle cell breast carcinoma and 31 patients with benign or malignant spindle cell tumours were studied using a panel of antibodies directed against multiple cytokeratins (AE1/AE3, CAM5.2, wide spectrum screening keratin), epithelial membrane antigen (EMA), and vimentin. Sites of origin for the 31 controls included breast, bone, and soft tissue. All but one (95.8%) metaplastic carcinomas stained positively with wide spectrum screening keratin. Only rare or focal immunoreactivity was observed with AE1/AE3 in four cases; however, sensitivity of AE1/AE3 was improved in 13 cases using steam EDTA as an antigen retrieval technique. Three cases were immunoreactive with CAM5.2 and eight cases were immunoreactive with EMA. All control cases lacked immunoreactivity with the cytokeratin panel and EMA. The spindle cells in the metaplastic breast tumours (88%) and in the controls (97%) stained with vimentin. CONCLUSIONS: Wide spectrum screening keratin may be the most useful and convenient antibody in differentiating metaplastic spindle cell carcinoma from other spindle cell lesions in the breast.

Ontogeny of Phex/PHEX protein expression in mouse embryo and subcellular localization in osteoblasts.

PHEX, a phosphate-regulating gene with homologies to endopeptidases on the X chromosome, is mutated in X-linked hypophosphatemia (XLH) in humans and mice (Hyp). Although recent observations indicate that Phex protein is expressed primarily in bone and may play an important role in osteoblast function and bone mineralization, the pattern of the Phex protein expression in the developing skeleton and its subcellular localization in osteoblasts remain unknown. We examined the ontogeny of the Phex protein in the developing mouse embryo and its subcellular localization in osteoblasts using a specific antibody to the protein. Immunohistochemical staining of mouse embryos revealed expression of Phex in osteogenic precursors in developing vertebral bodies and developing long bones on day 16 postcoitum (pc) and thereafter. Calvaria from day 18 pc mice showed Phex epitopes in osteoblasts. No Phex immunoreactivity was detected in lung, heart, hepatocytes, kidney, intestine, skeletal muscle, or adipose tissue of mouse embryos. Interestingly, embryonic mouse skin showed moderate amounts of Phex immunostaining. In postnatal mice, Phex expression was observed in osteoblasts and osteocytes. Moderate expression of Phex was seen in odontoblasts and slight immunoreactivity was observed in ameloblasts. Confocal microscopy revealed the presence of immunoreactive PHEX protein in the Golgi apparatus and endoplasmic reticulum of osteoblasts from normal mice and in osteoblasts from Hyp mice transduced with a human PHEX viral expression vector. PHEX protein was not detected in untransduced Hyp osteoblasts. These data indicate that Phex protein is expressed in osteoblasts and osteocytes during the embryonic and postnatal periods and that within bone, Phex may be a unique marker for cells of the osteoblast/osteocyte lineage.

PTEN induces chemosensitivity in PTEN-mutated prostate cancer cells by suppression of Bcl-2 expression.

The tumor suppressor gene PTEN (MMAC1/TEP1) is lost frequently in advanced prostate cancer (PCa). However, the function of PTEN in tumorigenesis is not understood fully. In this study, we demonstrate that expression of Bcl-2 in prostate tumors correlates with loss of the PTEN protein. This finding was verified by studies in the PCa cell lines DU145, PC-3, LNCaP, and an androgen-refractory subline of LNCaP. Transient transfection of PTEN into the PTEN-null cells resulted in decreased levels of Bcl-2 mRNA and protein. These effects appear to be mediated at the level of gene transcription, since a Bcl-2 promoter-reporter construct was down-regulated by ectopic expression of PTEN in LNCaP cells. The inhibition of Bcl-2 required the lipid-phosphatase activity of PTEN and was blocked by overexpression of a constitutively active form of Akt. Moreover, the transcription-regulatory protein cAMP-response element-binding protein (CREB) may be involved, since decreased phosphorylation of CREB at Ser(133) was detected following PTEN expression, and ectopic expression of CREB repressed completely the PTEN-induced inhibition of Bcl-2 promoter activity. Furthermore, cotransfection of Bcl-2 and PTEN expression vectors rescued PTEN-induced cell death but not G(1) cell cycle arrest. Finally, forced expression of PTEN sensitized LNCaP cells to cell death induced by staurosporine, doxorubicin, and vincristine, and this chemosensitivity was attenuated by exogenous expression of Bcl-2. Taken together, these data demonstrate that loss of PTEN leads to up-regulation of the bcl-2 gene, thus contributing to survival and chemoresistance of PCa cells. These findings suggest that the PTEN gene and its regulated pathway are potential therapeutic targets in prostate cancer.

hMLH1 and hMSH2 expression in human hepatocellular carcinoma.

The role of microsatellite instability (MSI) in the pathogenesis of hepatocellular carcinoma (HCC) is incompletely defined. Although high-frequency MSI (MSI-H) is infrequently seen in HCC, some studies have suggested a role for MSI in HCC development. While MSI has been clearly defined for a subset of tumors, in particular colorectal, gastric and endometrial cancers, generally accepted criteria have not been developed for other tumors. Colorectal cancers (CRC) are classified as MSI-H if >30-40% of >5 microsatellite loci analyzed show instability. The MSI-H phenotype is associated with defective DNA mismatch repair (MMR) and is observed in the majority of tumors from patients with hereditary non-polyposis colon cancer (HNPCC) and also in 15% of sporadic CRCs. Inactivating mutations of the hMLH1 or hMSH2 genes lead to defects in MMR in HNPCC. In sporadic CRCs, MMR is usually due to hypermethylation of the hMLH-1 promoter. The role of defective MMR in hepatocellular carcinogenesis is controversial. Immunohistochemistry for hMLH1 and hMSH2 reliably indicates hMLH1 or hMSH2 loss in MSI-H CRC tumors. To investigate the role of defective MMR in HCC carcinogenesis, we performed immunohistochemistry for hMLH1 and hMSH2 on 36 HCCs. BAT26, a microsatellite marker that reliably predicts MSI-H was also examined. All 36 of the tumors stained positively for both hMLH1 and hMSH2, strongly suggesting an absence of either inactivating mutations of hMLH1 and hMSH2 or promoter hypermethylation of hMLH1. None of the tumors showed MSI at the BAT26 locus. These findings suggest that defective MMR does not contribute significantly to hepatocellular carcinogenesis.

Expression of an immediate early gene, IEX-1, in human tissues.

IEX-1 is an immediate early gene that is induced by ionizing radiation, ultraviolet radiation, and a variety of growth factors. It plays an important role in the regulation of cellular growth. Earlier, we performed studies on the distribution of IEX-1 messenger RNA in different tissues and on the subcellular localization of IEX-1 protein. No reports, however, have appeared concerning the distribution of IEX-1 protein in a variety of human tissues. We raised a polyclonal antibody against a synthetic IEX-1 peptide (amino acids 51-75) and used the antibody to study the distribution of the protein in human tissues. We demonstrate that IEX-1 is strongly expressed in epithelia of the skin, trachea, gastrointestinal, and genitourinary systems, as well as in the pancreas and breast. Endothelial cells within the vasculature of most tissue/organs also strongly express IEX-1. Liver, lung, lymph nodes, and placenta stain weakly. No IEX-1 epitopes were detected in the thymus, testes, ovary, myocardium, skeletal muscle, or spleen. We conclude that IEX-1 is widely expressed in epithelial and endocrine tissues, as well as in vascular endothelium.

Transcriptional complementarity in breast cancer: application to detection of circulating tumor cells.

BACKGROUND: We used a combination of genetic subtraction, silicon DNA microarray analysis, and quantitative PCR to identify tissue- and tumor-specific genes as diagnostic targets for breast cancer. METHODS AND RESULTS: From a large number of candidate antigens, several specific subsets of genes were identified that showed concordant and complementary expression profiles. Whereas transcriptional profiling of mammaglobin resulted in the detection of 70% of tumors in a panel of 46 primary and metastatic breast cancers, the inclusion of three additional markers resulted in detection of all 46 specimens. Immunomagnetic epithelial cell enrichment of circulating tumor cells from the peripheral blood of patients with metastatic breast cancer, coupled with RT-PCR-based amplification of breast tumor-specific transcripts, resulted in the detection of anchorage-independent tumor cells in the majority of patients with breast cancer with known metastatic disease. CONCLUSION: Complementation of mammaglobin with three additional genes in RT-PCR increases the detection of breast cancers in tissue and circulating tumor cells.

Structural analysis of the 17q22-23 amplicon identifies several independent targets of amplification in breast cancer cell lines and tumors.

A novel region of amplification in breast tumors has recently been identified on chromosome 17q22-23. In an effort to identify the oncogenes in the region that are targeted by the amplification process, we determined the structure of the amplicon in breast cancer cell lines and tumors. Physical and transcription maps of the approximately 3.5-Mb region were established and used as the basis for copy number analysis within the region by Southern blot and fluorescence in situ hybridization. Seven specific and independent amplification maxima were identified in breast cancer cell lines and breast tumors. We present correlative amplification and overexpression studies for the FLJ21316 and Hs.6649 genes suggesting a role for these candidates as amplification-dependent oncogenes.

Analysis of Hurthle cell neoplasms of the thyroid by interphase fluorescence in situ hybridization.

Recent studies have indicated that numerical chromosomal abnormalities including changes in p53 and cyclin D1 may be involved in Hurthle cell tumorigenesis. We analyzed a series of Hurthle cell neoplasms of the thyroid to evaluate the diagnostic and prognostic utility of numerical anomalies by DNA fluorescent probes for cyclin D1 and p53 gene loci and chromosomes 5, 7, 11, 12, 17, and 22. Interphase fluorescence in situ hybridization (FISH) analysis was performed on paraffin-embedded tissue sections from 10 Hurthle cell adenomas, 19 Hurthle cell carcinomas, and 7 normal thyroid tissues used as controls. Directly labeled fluorescent DNA probes for the centromere region of chromosomes 7, 11, 12, and 17 and locus-specific probes for chromosomes 5 and 22, cyclin D1, and p53 were utilized for dual-probe hybridizations. Sixty percent (6 of 10) Hurthle cell adenomas and 63% (12 of 19) Hurthle cell carcinomas showed chromosome gains. Twenty percent (2 of 10) Hurthle cell adenomas and 26% (5 of 19) Hurthle cell carcinomas showed chromosome losses. Normal thyroid tissues used as controls showed no chromosomal abnormalities. Among Hurthle cell tumors with chromosomal abnormalities, adenomas averaged 2.7 gains and 0.3 losses per case, and carcinomas averaged 3.3 gains and 0.6 losses per case. The two adenomas with chromosome losses each showed loss of one chromosome, whereas the five carcinomas with losses averaged 1.8 losses per case. Chromosome 22 was the most common loss identified, occurring in three of the 11 patients who died of disease. These results indicate that chromosomal imbalances as gains are common in both benign and malignant Hurthle cell neoplasms, but Hurthle cell carcinomas tend to have more chromosome losses than adenomas. Among Hurthle cell carcinomas in this study, chromosome losses were identified only from patients who died of disease. The loss of chromosome 22 may have prognostic value in Hurthle cell carcinoma of the thyroid.

Discrepancies in clinical laboratory testing of eligibility for trastuzumab therapy: apparent immunohistochemical false-positives do not get the message.

BACKGROUND: Several studies have reported what seem to be false-positive results using the Food and Drug Administration (FDA)-approved HercepTest (Dako Corp, Carpinteria, CA) to profile Her-2/neu amplification and overproduction in breast carcinoma. False-positive status has been based on comparisons with gene copy enumeration by fluorescence in situ hybridization (FISH) and with comparisons to immunohistochemistry (IMH) results using a monoclonal antibody. However, simple overexpression by tumor cells that have normal gene copy has not been evaluated by profiling mRNA expression, ie, such cases could simply represent true-positive, transcriptionally upregulated overexpression. MATERIALS AND METHODS: Four hundred infiltrating ductal carcinomas of breast were evaluated by IMH using monoclonal (CB11; Ventana Medical Systems, Inc, Tucson, AZ) and polyclonal (HercepTest; Dako) antibodies after antigen retrieval (AR). A polyclonal antibody sans AR (PCA/SAR) was also used. All IMH stains were evaluated and scored according to the guidelines for the FDA-approved HercepTest. A total of 145 of 400 carcinomas were subsequently evaluated by direct and digoxigenin-labeled (Dig) FISH, and 144 of 400 were evaluated by detection of mRNA overexpression via autoradiographic RNA:RNA in situ hybridization. RESULTS: Overall HercepTest/CB11 IMH discordance was 12%. Expression of mRNA was highly concordant with FISH and DigFISH amplification and with CB11 and PCA/SAR immunohistology. IMH false-positive cases (no Her-2/neu gene amplification) occurred with both HercepTest (23%) and CB11 (17%), and the majority of false-positive results (34 of 44) were scored as 2+. All 2+ false-positive cases were mRNA-negative. Combined results of HercepTest and CB11 showed that 79% (38 of 48) of 3+ cases were Her-2/neu gene amplified, but only 17% (seven of 41) of 2+ cases had increased gene copy. CONCLUSION: Discordant HercepTest/FISH results, and to a lesser extent discordance with CB11 IMH, are most commonly false-positive results with a score of 2+. The 2+ score as defined in the guidelines for the FDA-approved HercepTest should not be used as a criterion for trastuzumab therapy unless confirmed by FISH. Determination of Her-2 gene copy number by FISH may be a more accurate and reliable method for selecting patients eligible for trastuzumab therapy.

Immunohistochemical Analysis for hMLH1 and hMSH2 Expression in Colorectal Cancer.

Defective DNA mismatch repair (MMR) occurs in the majority of tumors from patients with hereditary non-polyposis colorectal cancer (HNPCC) and approx 15% of sporadic colorectal cancer (CRC) (1,2). In HNPCC-associated tumors, defective MMR is most often due to inactivating mutations of the DNA MMR genes hMLH1 and hMSH2 (3,4). Defective MMR in sporadic CRC, on the other hand, is generally due to hypermethylation of the hMLH1 promoter (5-7). As might be expected, inactivating mutations of hMSH2 and hMLH1 lead to a loss of hMSH2 or hMLH1 expression respectively and hypermethylation of the hMLH1 promoter to a loss of hMLH1 expression (5-8). One of the hallmarks of defective DNA MMR is a type of genetic instability known as microsatellite instability (MSI). Tumors with defective DNA MMR generally exhibit MSI at the majority of the loci examined (MSI-H phenotype) (8,9).

Loss of alternately spliced messenger RNA of the luteinizing hormone receptor and stability of the follicle-stimulating hormone receptor messenger RNA in granulosa cell tumors of the human ovary.

OBJECTIVES: The expression of messenger RNA (mRNA) for the LH (LHR) and FSH receptors (FSHR) was examined in normal human corpora lutea and granulosa cell tumors. METHODS: Expression was examined by RT/PCR and DNA sequencing techniques. RESULTS: The full-length (FL) coding region and seven additional isoforms were identified for normal LHR mRNA. Isoform 1 had portions of exons II and III deleted, and isoform 2 had exon IX omitted. Isoform 3 also had portions of exons II and III deleted and all of exon IX deleted. Exons III through VI were missing in isoforms 4-7. Isoform 5 also had exon IX omitted, and isoform 6 also had part of exon XI missing. Isoform 7 also had exon IX and part of exon XI deleted. An aberrant migration pattern of the LHR mRNA isoforms was observed for granulosa cell tumors with FIGO Stage I-IV. Five tumor samples of Stage III-IV had many isoforms absent. Seven Stage I samples had aberrant migration patterns that depended on the size of the tumor. As the size of the tumor increased the aberrant migration pattern of the LHR mRNA isoforms was more pronounced and some isoforms were not detected. The FL and at least one additional isoform were identified for FSHR mRNA. Isoform 1 had regions of exons IV and V deleted. The FSHR mRNA isoforms had a similar migration pattern for the normal ovary and the granulosa cell tumors. CONCLUSIONS: Alternately spliced forms of mRNA for the LHR and FSHR exist for normal human ovary and granulosa cell tumors. The aberrant migration and missing LHR mRNA isoforms in granulosa cell tumors do not appear to result from general genomic instability associated with tumor progression. These findings are important to understand the role of alternate splicing in the regulation of LHR and FSHR expression in different pathological states.

17q23 amplifications in breast cancer involve the PAT1, RAD51C, PS6K, and SIGma1B genes.

Amplification of the 17q23 region occurs frequently in breast tumors. To characterize the structure of 17q23 amplicons and to identify oncogene targets associated with this alteration, we performed a copy number analysis of 87 17q23 localized expressed sequence tags in seven breast cancer cell lines. Three major regions of amplification were detected in the MCF7 and BT474 cell lines. Amplification of at least one of four known genes (PAT1, PS6K, RAD51C, and SIGMA1B) was detected in the cell lines and in 28% of 94 breast tumors. In most cases, these four genes were overexpressed when amplified, but there was a particularly good association between amplification of the SIGMA1B gene and elevated expression in tumors, which suggested a possible role for this gene in tumor progression. Our data show that this region contains at least four independent targets of amplification, which suggests that there is considerable variability in the structure of the 17q23 amplicon.

Prognostic value of bone marrow angiogenesis in multiple myeloma.

We studied the prognostic value of angiogenesis grading and microvessel density estimation in newly diagnosed multiple myeloma. Seventy-five patients with newly diagnosed myeloma, treated on Eastern Cooperative Oncology Protocol E9486 and Intergroup study 0141 (S9321) at the Mayo Clinic, were studied. Bone marrow microvessels were examined using immunohistochemical staining for von Willebrand factor. Determination of microvessel density and angiogenesis grading was done in a blinded manner. There was a strong correlation between microvessel density and the plasma cell labeling index, rho 0.42, P < 0.001. Angiogenesis grade was also significantly associated with the plasma cell labeling index. Fifteen % of patients with low-grade angiogenesis had a high labeling index (>1%). In contrast, 47% of patients with intermediate or high-grade angiogenesis had high labeling indices (P = 0.02). Overall survival was significantly different among those with high-, intermediate-, and low-grade angiogenesis, with median times of 2, 4, and 4.4 years, respectively (P = 0.02). Similarly, patients with microvessel density >50/x400 field had poorer survival compared with those with 50 or fewer microvessels/field, median survival 2.6 versus 5.1 years, respectively (P = 0.004). There was a strong association between angiogenesis grade and microvessel density (P < 0.001). We conclude that bone marrow angiogenesis is a predictor of poor survival in newly diagnosed myeloma. Angiogenesis is correlated with the plasma cell labeling index but not the bone marrow plasma cell percentage. A simple visual grading of angiogenesis is an efficient alternative to microvessel density estimation.

Pretreatment assessment of prognostic indicators in endometrial cancer.

OBJECTIVE: The object of this study was to assess the association of histologic, cytokinetic, and molecular variables in preoperative endometrial samples with extrauterine disease, recurrence, and survival among patients with endometrial cancer. STUDY DESIGN: In a case-cohort study of 125 women, ploidy, S-phase fraction, proliferative index, deoxyribonucleic acid index, proliferating cell nuclear antigen, MIB-1 proliferation marker, p53 tumor suppressor gene, and cytoplasmic HER-2/neu oncogene and bcl-2 expressions were quantitated. RESULTS: A model with only one independent term predicted progression-free survival; that variable was p53 (P <. 0001; relative risk, 5.60). A model with two independent terms predicted disease-related survival; these variables were p53 (P =. 0002; relative risk, 7.39) and MIB-1 (P =.03; relative risk, 3.27). Among patients with tumors with both p53 and MIB-1 expression exceeding 33%, a total of 32% had died of disease by 2 years. A model for predicting extrauterine disease selected two independent variables: p53 (odds ratio, 3.20; P =.01) and ploidy (odds ratio, 2. 16; P =.04). An advanced surgical stage was encountered in 26% to 35% of cases in which either the p53 expression exceeded 33% or the deoxyribonucleic acid content was nondiploid and in 53% of cases in which both variables were unfavorable. CONCLUSIONS: Preoperative evaluation of quantifiable cytokinetic and molecular variables can assist in identifying tumor types that are predisposed toward a more aggressive clinical course.

Assessment of inhibin and p53 in granulosa cell tumors of the ovary.

OBJECTIVE: The goal of this work was to determine the cellular content of inhibin and p53 in granulosa cell tumors (GCTs). METHODS: Clinical records of 47 patients (mean age, 54 years; range, 20-85 years) presenting with GCT surgically managed at our institution were abstracted. International Federation of Gynecology stage I was assigned in 39 patients, stage II in 2, and stage III in 6. Concomitant endometrial carcinoma was identified in 6 patients. Mean follow-up was 13.6 years (range, 1 day to 37.6 years). Sections from paraffin-embedded tissue blocks were analyzed immunohistochemically for expression of tissue inhibin and p53 levels. Inhibin expression was graded by intensity and reactivity, and p53, by its presence or absence. RESULTS: The tumors of 27 patients (57%) stained strongly for inhibin intensity and showed >60% reactivity. Decreased intensity and reactivity of inhibin expression were associated with advanced-stage disease (P = 0.05 and P < 0.01, respectively, by Fisher exact test). Expression of p53 was detected in tumors from 27 patients (57%), and immunoreactivity was associated with compromised progression-free survival (P = 0.016, log-rank test). However, the association between p53 immunoreactivity and disease stage was not significant. Absence of p53 expression was significantly associated with concurrent endometrial carcinoma (P = 0.022), suggesting more molecularly intact tumors that retain functional activity. CONCLUSIONS: Although the majority of GCTs show strong expression of inhibin with regard to intensity and reactivity, weak expression is associated with advanced disease but not with decreased progression-free survival. By contrast, expression of p53 is not significantly associated with stage, but increased expression is associated with decreased disease-free survival. Absence of p53 expression appears to be associated with concurrent endometrial carcinoma.

Adenoma-specific alterations of protein kinase C isozyme expression in Apc(MIN) mice.

Members of the protein kinase C (PKC) family appear to play important roles in colorectal carcinogenesis. To investigate the potential involvement of PKC isozymes in adenomatous transformation induced by inactivation of the adenomatous polyposis coli (APC) gene product, we examined protein levels and localizations of ten PKC isozymes by immunohistochemistry in normal and adenomatous ileal epithelium of ApcMIN mice. Compared with surrounding normal epithelium, adenomas showed dramatically reduced staining for PKCs a, beta1, and zeta, as well as dysplasia-specific punctate nuclear staining of PKC mu. We conclude that reduced protein expression of PKC alpha, beta1, and zeta, and nuclear localization of PKC mu are markers of, and are perhaps involved in, adenomatous transformation induced by APC inactivation in ApcMIN mice.

HMSH6 alterations in patients with microsatellite instability-low colorectal cancer.

Two microsatellite instability (MSI) phenotypes have been described in colorectal cancer (CRC): MSI-H (instability at >30% of the loci examined) and MSI-L (MSI at 1-30% of the loci examined). The MSI-H phenotype, observed in both hereditary nonpolyposis colon cancer-associated CRC and approximately 15% of sporadic CRC, generally results from mutational or epigenetic inactivation of the DNA mismatch repair (MMR) genes hMSH2 or hMLH1. The genetic basis for the MSI-L phenotype, however, is unknown. Several other proteins, including hMSH3 and hMSH6, also participate in DNA MMR. Inactivating mutations of MSH6 in yeast and human tumor cell lines are associated with an impaired ability to repair single-base mispairs and small insertion-deletion loops but not large insertion-deletion loops. This suggests that hMSH6 mutations are more likely to be associated with a MSI-L phenotype than a MSI-H phenotype in CRC. To explore this possibility, we screened tumors from 41 patients with MSI-L CRC for hMSH6 mutations with conformation-sensitive gel electrophoresis (CSGE) and for hMSH6 protein expression by immunohistochemistry. Alterations found with CSGE were confirmed by DNA sequencing of normal and tumor tissue. One somatic (Asp389Asn) and 15 germ-line changes were found. Of the 15 germ-line changes, 9 were found in an intron (none involving splice junctions), and 6 were found in an exon (Gly39Glu, Leu395Val, and 4 silent alterations). Immunohistochemical staining for hMSH6 performed on 34 of the 41 tumors revealed strong nuclear hMSH6 expression in all of the cases. Overall, our results suggest that hMSH6 mutations do not play a major role in the development of sporadic CRC with a MSI-L phenotype.

Immunodiscrimination of colorectal neoplasia using MUC1 antibodies: discrepant findings in tissue versus stool.

Colorectal tumor-associated antigens are attractive targets for novel stool-screening assays. MUC1, a glycoprotein antigen, is aberrantly expressed in transformed colorectal mucosa and represents a candidate fecal biomarker. In this study, tissue staining and stool testing were performed to further clarify the discriminant potential of MUC1 in markedly different biologic media. One anti-MUC1 monoclonal antibody (MA5) was used for immunohistochemistry and two commercially available MUC1 assay kits (ELSA-CA 15-3 and Truquant BR) were used for stool detection. On tissue staining, MUC1 expression was strong in 40/40 (100%) adenocarcinomas, moderate in 42/55 (76%) adenomas, faint in 8/28 (29%) juxtatumoral mucosa specimens, and absent in 15/15 (0%) nonadjacent mucosa specimens. Conversely MUC1 levels in stool testing did not differ between colorectal cancer cases (N = 14) and controls (N = 14). Based on these results, MUC1 appears to be a functional tumor biomarker in colorectal tissue but not in stool. Bacterial metabolism within stool may unmask the core antigen of MUC1 and account for this discordance in immunoreactivity.