Specimens of bone marrow trephine biopsy (BMT) are transported and fixed in acetic acid–zinc–formalin fixative, decalcified in 10% formic acid–5% formaldehyde and processed with other specimens to paraffin-wax embedding. Sections, 1-μm-thick, are cut by experienced histotechnologists and used for haematoxylin and eosin, Giemsa, reticulin silver and other histological stains. Further, all immunohistochemical procedures used in the laboratory, including double immunostaining, can be used on these sections with no or minimal modifications. About 10 000 BMT specimens have been analysed using this procedure since 1997 and diseases involving the bone marrow have been classified successfully. More recently, standardised polymerase chain reaction-based analysis and mRNA in situ hybridisation studies have been conducted. Excellent morphology with good antigen, DNA and RNA preservation is offered by the Hammersmith Protocol.

A unique feature of SW480 and SW620 colon carcinoma cell lines is that they are derived from primary and secondary tumours resected from a single patient. As such, they may represent a valuable resource for examining genetic changes late in colon cancer progression. In order to verify this, both cell lines have been characterized to determine whether phenotypic differences have been retained despite long-term cell culture in vitro. The primary tumour-derived SW480 cells have an epithelioid morphology in vitro, while metastasis-derived SW620 cells have a fibroblast-like appearance. Xenografts of SW480 cells form gland-like structures in vivo, while SW620 xenografts form solid sheets of tumour cells. SW620 cells have a higher BrdU labelling index than SW480 cells, and are more highly tumourigenic and metastatic. Furthermore, SW620 cells show less susceptibility to apoptosis induction by TNFalpha and anti-Fas monoclonal. Findings from these investigations therefore indicate that SW480 and SW620 cell lines do show appropriate phenotypic differences and represent an interesting model for studying the genetic events in the late stages of colon cancer progression.

PURPOSE:This trial was designed to test the safety and efficacy of a tumor-specific genetic prodrug activation therapy targeted by use of the human erbB-2 gene promoter. The erbB-2 oncogene is overexpressed in approximately 20% of cases of breast cancer and is associated with poor prognosis.PATIENTS AND METHODS:Twelve breast cancer patients received transcriptionally targeted gene therapy in a phase I clinical trial using direct intratumoral injection of plasmid construct combined with systemic administration of prodrug. The genetic prodrug activation therapy is specifically targeted to erbB-2-overexpressing breast cancer cells by use of a therapeutic cassette that contains the Escherichia coli cytosine deaminase gene driven by the tumor-specific erbB-2 promoter, thus allowing activation of fluorocytosine to the active cytotoxic fluorouracil only within tumor cells that express the oncogene.RESULTS:The approach was shown to be safe and to result in targeted gene expression in up to 90% of cases. Using a number of different assays, we demonstrated that significant levels of expression of the suicide gene were specifically restricted to erbB-2-positive tumor cells, confirming the selectivity of the approach.CONCLUSION:The results of this study, the first targeted gene therapy for breast cancer and the first to use the cytosine deaminase system in human subjects, are encouraging for the development of genetic prodrug activation therapies that exploit the transcriptional profile of cancer cells.

We have defined the host leukocyte infiltrate in epithelial ovarian tumors and related this to the expression of C-C chemokines. Immunohistochemical analysis of 20 paraffin-embedded biopsies showed that the infiltrate was primarily composed of CD68+ macrophages and CD8+/CD45RO+ T cells (median values, 3700 cells/mm3 and 2200 cells/mm3, respectively). Natural killer cells, B cells, and mast cells occurred in lower numbers (median values, 0 to 200 cells/mm3). Eosinophils were rarely seen and neutrophils were mainly confined to blood vessels. More infiltrating cells were found in stromal than in tumor areas. Only macrophages occurred in significant numbers in areas of necrosis (P < 0.0005). Using in situ hybridization to mRNA, we examined expression of the chemokines MCP-1, MIP-1 alpha, MIP-1 beta, and RANTES. MCP-1 and MIP-1 alpha were expressed by significantly more cells than MIP-1 beta and RANTES (P < 0.005). In tumor epithelial areas, the predominant chemokine was MCP-1. MCP-1 and MIP-1 alpha were the predominant stromal chemokines. A significant correlation was found between the total number of CD8+ T cells and the number of cells expressing MCP-1 (rs = 0.63 and P < 0.003, respectively) and between the CD8+ population and RANTES-expressing cells (rs = 0.6 and P < 0.003). A correlation was also found between CD68+ macrophages and the number of cells expressing MCP-1 (rs = 0.50 and P = 0.026). We suggest that MCP-1 may be responsible for the leukocyte infiltrate in ovarian carcinomas, but the expression of other chemokines may determine its exact nature.

The expression of PCNA was examined by immunohistochemistry on 95 formalin-fixed breast surgical specimens. Seventy-four were benign breast disease, 3 atypical hyperplasia; 5 in situ carcinoma; 13 invasive carcinoma. Forty-two percent of benign breast disease, 75% of atypical hyperplasia/in situ carcinoma and 92% of invasive breast carcinoma expressed PCNA nuclear staining. Forty percent of cysts, 15.78% of sclerosing adenosis, 22.22% of adenosis, and 17.39% of epitheliosis expressed low cyclin protein levels (PCNA score I, < 18%); 20% of cysts, 31.57% of sclerosing adenosis, 5.55% of adenosis and 43.47% of epitheliosis expressed moderate levels (PCNA score II, 18 to 50%); whereas 50% of atypical hyperplasia/in situ carcinoma and 54% of invasive breast carcinoma contained high levels (PCNA score III, > 50%). This study on PCNA/cyclin localization suggests that PCNA is an early indicator of ongoing cellular proliferation.

Chemokines may control the macrophage infiltrate found in many solid tumors. In human ovarian cancer, in situ hybridization detected mRNA for the macrophage chemokine monocyte chemoattractant protein-1 (MCP-1) in 16/17 serous carcinomas, 4/4 mucinous carcinomas, 2/2 endometrioid carcinomas, and 1/3 borderline tumors. In serous tumors, mRNA expression mainly localized to the epithelial areas, as did immunoreactive MCP-1 protein. In the other tumors, both stromal and epithelial expression were seen. All tumors contained variable numbers of cells positive for the macrophage marker CD68. MCP-1 mRNA was also detected in the stroma of 5/5 normal ovaries. RT-PCR demonstrated mRNA for MCP-1 in 7/7 serous carcinomas and 6/6 ovarian cancer cell lines. MCP-1 protein was detected by ELISA in ascites from patients with ovarian cancer (mean 4.28 ng/ml) and was produced primarily by the cancer cells. Human MCP-1 protein was also detected in culture supernatants from cell lines and in ascites from human ovarian tumor xenografts which induce a peritoneal monocytosis in nude mice. We conclude that the macrophage chemoattractant MCP-1 is produced by epithelial ovarian cancer and that the tumor cells themselves are probably a major source. MCP-1 may contribute to the accumulation of tumor-associated macrophages, which may subsequently influence tumor behavior.

We have used an avidin-biotin immunoperoxidase technique to localise epithelial cadherin (E-cadherin), a calcium-dependent cell-cell adhesion molecule, in 107 paraffin-embedded sections from 93 patients consisting of 24 with colorectal adenoma, 55 with rectal carcinoma and 14 with liver metastases. The corresponding primary colorectal tumours were also studied in these cases. E-cadherin was expressed by normal colorectal epithelial cells with typical membranous staining at the intercellular junctions. Loss of normal membranous E-cadherin expression and presence of cytoplasmic staining were found frequently in adenomas larger than 1 cm (P < 0.01), with high grade dysplasia and villous histology (P < 0.01). In primary rectal cancers, loss of membranous expression correlated with high tumour grade. No correlation was seen with Dukes and Jass stage, local extramural spread and 5-year recurrence rate. Complete loss of membranous E-cadherin immunoreactivity was seen in 7/14 (50%) liver metastases in which 6/7 (86%) showed intense membranous E-cadherin immunoreactivity in the corresponding primary tumour. Our data indicate that changes in E-cadherin immunoreactivity and cellular localisation correlate with size, severe dysplasia in adenomas and tumour grade in carcinomas. However, there seems to be no correlation between loss of membranous E-cadherin immunoreactivity and the invasive and metastatic potential of the carcinomas.

substitution determine that the strength of linear pull exerted by a particular muscle is indeterminable, and if this pull is delivered at an unknown distance from the fulcrum of rotation of the joint in question, so obscuring its effective moment, how can the clinician measure individual muscle strength? Can he possibly tell? The truth is that most often we do not know, and that unsubstantiated claims to be able to tell have long been an impediment to the re-integration of musculoskeletal medicine into orthodoxy. While hypotheses are perfectly acceptable in the absence of evidence to the contrary, it is NOT acceptable to cling to beliefs which have been shown to be untenable. What is clear is that, 'I can tell' really means, 'I will not learn'. Nonetheless, shorn of ideological fantasies, vertebral manipulation has been shown to be a simple, safe therapeutic option, unpredictably successful in outcome, legitimate to orthodoxy3. It is surely time to sound the death-knell of 'I can tell'. True wisdom lies in understanding how little we know; and acknowledging this.

A novel cDNA clone was isolated using a polyclonal serum directed against partially purified ovarian carcinoma antigen CA125. The deduced peptide sequence lacked membrane protein characteristics expected for CA125 but encompassed a B-box/coiled coil motif present in many genes with transformation potential. The gene was mapped by fluorescence in situ hybridization within the minimum region known to contain the familial breast/ovarian carcinoma gene, BRCA1. YAC and cosmid clones were isolated and used to refine the location of this gene adjacent and proximal to the RNU2 locus. The exon structure of the gene was determined. Extensive SSCP and sequence analysis of over 100 tumour and normal DNAs from familial and sporadic breast cancers and sporadic ovarian cancers failed to detect mutations in the coding region of this gene.

In order to understand the ability of human ovarian cancers to degrade the basement membrane, we have studied the localization and activity of matrix metalloproteases (MMPs) 2 and 9, using in site hybridization and quantificative zymography on sequential sections of tumor biopsies. We have related these data to expression of some of the controlling elements of the enzymes, namely tissue inhibitors of metastasis (TIMPs) and tumor necrosis factor (TNF). mRNA for MMP-2 was found in the majority of cases and localized to stromal areas with maximal expression adjacent to neoplastic areas. MMP-9 expression was associated with cells in epithelial and stromal areas, consistent with distribution of macrophages. Zymography revealed higher levels of MMP-9 activity in the ovarian cancer biopsy samples than in other cancers studied, but in contrast to our previous observations in breast and bladder cancer, there was no correlation between MMP levels and tumor grade. Nor was there any association between amount of TNF mRNA and levels of MMP enzymes. TIMP-1 expression was localized to stromal areas adjacent to tumor epithelial cells as well as, in some cases, to epithelial cells. The pattern of TIMP-2 expression was similar to that of MMP-2. We conclude that the stromal elements of ovarian tumors express MMP-2 and 9 and their specific inhibitors, but these do not seem to be controlled by endogenous TNF in the tumor microenvironment. (C) 1994 Wiley-Liss, Inc.

Recent data suggest an inverse correlation between human papillomavirus (HPV) infection and p53 tumor-suppressor gene mutation. In an attempt to elucidate the role of p53 mutations in cervical neoplasias, 65 cervical lesions, ranging from normal to malignant, were examined for overexpression of p53 protein by immunohistochemistry in paraffin-embedded tissue, and correlated with proliferating cell nuclear antigen (PCNA). An overexpression was seen in 35% of well-differentiated and 32.5% of poorly-differentiated squamous carcinomas, in 43.33% of microinvasive and 21.66% of CIS. More than 50% of neoplastic cells were immunoreactive for p53 protein in 10% of well-differentiated squamous carcinomas. Other positive specimens showed reactivity in < 24% of tumor cells. No staining was found in adenocarcinoma, dysplastic tissue, condylomas and normal tissue (83.07%). In contrast PCNA was detected in all cases of invasive squamous carcinomas, adenocarcinoma, CIS, CIN III, in 32.5% of CIN II, in 29.54% CIN I, and in 53.52% of wart. More than 50% of tumor cells showed nuclear staining for PCNA protein in 61.17% of invasive squamous carcinomas, in 21.66% of CIS, in 39% CIN III, in 32.5% CIN II and in 7.64% of wart. In the remain cases the positivity of nuclear staining was < 24%. No staining was present in 20% of cases including in normal cervix. Our data suggest that the viral-host protein interactions result in loss of the negative growth control normally exerted by p53. The consequence of HPV infection is a loss of functional wild-type p53 protein within the cells.(ABSTRACT TRUNCATED AT 250 WORDS)

Several studies have linked abnormal actin cytoskeletal organization and mutant actin genes with neoplastic transformation. Using in situ hybridization techniques we looked at expression of beta actin mRNA at a cellular level in normal, benign and neoplastic human colorectal tissues and correlated the level of expression with the extent of actin cytoskeletal organization in sequential sections. Normal tissues showed light labelling with the actin riboprobe, but non-neoplastic crypts, with evidence of regeneration and repair, showed greater levels. Higher levels of actin mRNA expression were found in adenomas and metaplastic polyps, but the highest levels were found in carcinomas, particularly those that were poorly differentiated. Cytoskeletal actin organization was, however, reduced in colorectal neoplasia. Normal mucosa showed the highest level of cytoskeletal actin organization, as assessed by phalloidin binding, and adenomas and metaplastic polyps also stained strongly. In contrast, phalloidin binding to poorly differentiated carcinomas was absent or very weak. The inverse relationship between actin mRNA expression and actin filament organization was confined to the carcinomas studied and may indicate defects in actin binding proteins or in post-transcriptional or translational events.

We undertook an in situ hybridization study to localize the MRNAs for the 72 kda type IV collagenase (MMP-2) and its specific inhibitor (TIMP-2) in 12 colorectal carcinomas, 3 adenomas, and 4 uninvolved resection margins to see how their distributions correlated with that of the reported distribution of MMP-2 protein. Labeling for MMP-2 and TIMP-2 mRNAs was detectable in 10 of 12 carcinomas and in 2 of 3 adenomas. Unexpectedly, we found much stronger signals for MMP-2 and TIMP-2 mRNAs within the mesenchymal cells in the desmoplastic stroma, of endothelial and/or (myo)fibroblastic nature, rather than in tumor epithelial cells in which localization of MMP-2 was anticipated Our data indicate that stromal cells may have the ability to synthesize a metalloproteinase that degrades basement membrane, and may together with the neoplastic epithelial cells participate actively in the tissue remodeling and disruption of the basement membrane integrity which is characteristic of invasive tumors.

The Very Late Antigens (VLAs) are alpha beta heterodimeric transmembrane proteins mediating cell-substratum as well as cell-cell interactions. Changes in their expression and/or function seem to occur in a number of invasive carcinomas and may at least in part explain their abnormal patterns of growth and differentiation. Using monoclonal antibodies to the beta 1 (DH12, A1A-5), alpha 2 (B1.515) and alpha 3 (E1.56) chains, VLA-2 (alpha 2 beta 1) and VLA-3 (alpha 3 beta 1) were studied on cryostat sections of three fibroadenomas and 43 invasive breast carcinomas (29 ductal, 14 lobular) by the avidin-biotin complex immunoperoxidase technique. In non-neoplastic breast tissue and in fibroadenomas VLA-2 and VLA-3 were expressed by myoepithelial cells and on the basolateral surface of the luminal cells. There was weak or absent expression of alpha 2, alpha 3 and the common beta 1 chain in the majority of invasive carcinomas compared to the adjacent normal breast epithelium and preinvasive (in-situ) carcinomas. In addition, the expression of the alpha 2 chain of VLA-2 was reduced significantly (P less than 0.005) in the poorly differentiated ductal breast carcinomas (Grade III) compared to the well (Grade I) and moderately (Grade II) differentiated ductal tumours. These data give further evidence that loss or down-regulation of VLA-2 and VLA-3 occur relatively frequently in invasive cancers, and, at least in the invasive ductal breast carcinomas. Loss of an extracellular matrix receptor controlling growth and differentiation seems to be one of the abnormalities underlying the progression towards an undifferentiated morphology.

The integrins are alpha beta heterodimeric transmembrane proteins mediating cell-substratum as well as cell-cell interactions. Changes in their expression and/or function seem to occur in a number of malignant epithelial neoplasms and may in part explain their abnormal patterns of growth and differentiation. Using monoclonal antibodies to the beta 1 (DH12), alpha 1 (TS2/7), alpha 2 (B1.515), and alpha 3 (E1.56) integrin chains, the alpha 1 beta 1 (VLA-1), alpha 2 beta 1 (VLA-2), and alpha 3 beta 1 (VLA-3) integrin receptors were studied on cryostat sections of 22 basal cell carcinomas (BCCs) and adjacent normal tissues by a standard peroxidase-antiperoxidase technique. In non-neoplastic skin, VLA-2 and VLA-3 were found in the basal layer, eccrine glands, and cells of the outer root sheath in which VLA-1 was detected. In BCCs, there was a considerably higher expression of VLA-2 and VLA-3 compared with epidermal basal cells but similar to that seen in hair bulb and outer root sheath. In two cases of nodular BCC showing evidence of regression, both VLA-2 and VLA-3 were completely negative, in contrast to non-regressing foci which were strongly positive. The high level of expression of two adhesion molecules (VLA-2 and VLA-3) involved in cell-substratum as well as cell-cell interactions may account for the more indolent pattern of growth characteristic of BCC and perhaps reflect its high degree of differentiation towards the hair follicle.

Immunohistochemistry is now an established technique in larger diagnostic histopathology units, but its full potential is not always achieved due to irrational application and incorrect interpretation. This review attempts to place the use of monoclonal antibodies in relation to the wider aspects of morphological diagnosis. The necessity of obtaining familiarity with the techniques and reagents in order to recognise artefacts is emphasized, and the practical benefit of this technique in difficult diagnostic problems is discussed.

A case of an unusual hypertrophic gastropathy confined to the gastric antrum which presented with chronic anaemia is described. The clinical and pathological features are contrasted with Menetrier's disease and hypersecretory hypertrophic gastropathy, and a possible relation to solitary hyperplastic (regenerative) polyps is discussed.