Search

Purpose: The human CC3/TIP30 gene is a putative metastasis suppressor gene, based on the results of experimental studies using lung, colon and melanoma cell lines. However, there is very little evidence from studies on clinical material, in support of such a role for the gene. In this study, we evaluated the expression of CC3/TIP30 in human breast cancer tissue and investigated the possible associations with the clinicopathological parameters. Methods: Total RNA and proteins were extracted from the frozen breast tumor and matched normal tissues. Evaluation of CC3/TIP30 expression was assessed by reverse transcription (RT)-PCR and reverse phase protein array. Immunohistochemistry of CC3/TIP30 on breast tissue microarrays was also analyzed. Results: We have found that CC3/TIP30 expression is significantly associated with positive HER-2/neu status at both mRNA (P=0.023) and protein (P=0.016) levels. Immunohistochemical analysis on tissue microarrays also shows a positive correlation between CC3 expression and HER-2/neu status (P=0.0028). Conclusion: Our findings suggest a potential link between the expression of CC3/TIP30 gene and the HER-2/neu oncogene-mediated signal pathway.These findings could not have been predicted from previous experimental studies, and suggest that CC3/TIP30 may play a complex role in breast cancer.

Genetic changes occurring in the late stages of colonic tumour progression have received much less attention than those occurring in the early stages. As described in the accompanying paper, SW480 and SW620 cell lines provide a useful model for studying the advanced stages of progression for colon cancer. Comparison of the two cell lines by differential display reveals that SW620 cells express lower levels of the CC3 tumour suppressor gene and also lower levels of the tissue inhibitor of metalloproteinases-3 (TIMP-3) gene. Northern blot analysis for TIMP-3 confirms this finding and shows a similar difference in the expression of TIMP-2, which seems logical since TIMPs inhibit enzymes that play a role in tumour invasion. For this reason, it was surprising to find that TIMP-1 messenger RNA expression is markedly increased in SW620 cells. Consistent with this finding, western blot analysis shows a ten-fold increase in TIMP-1 protein secretion by SW620 cells. It is noteworthy that high TIMP-1 expression is associated with poor prognosis in colorectal cancer. This association between TIMP-1 expression and tumour progression may be related to additional growth factor-like effects described for TIMP-1 in some systems.

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.

Matrix metalloproteinases (MMPs), or matrixins, are a family of zinc endopeptidases that play a key role in both physiological and pathological tissue degradation. Normally, there is a careful balance between cell division, matrix synthesis and matrix degradation, which is under the control of cytokines, growth factors and cell matrix interactions. The MMPs are involved in remodelling during tissue morphogenesis and wound healing. Under pathological conditions, this balance is altered: in arthritis, there is uncontrolled destruction of cartilage; in cancer, increased matrix turnover is thought to promote tumour cell invasion. The demonstration of a functional role of MMPs in arthritis and tumour metastasis raises the possibility of therapeutic intervention using synthetic MMP inhibitors with appropriate selectivity and pharmacokinetics. As the process of drug discovery focuses on structure-based design, efforts to resolve the 3-dimensional structures of the MMP family have intensified. Several novel MMP inhibitors have been identified and are currently being investigated in clinical trials. The structural information that is rapidly accumulating will be useful in refining the available inhibitors to selectively target specific MMP family members. In this review, we focus on the role of MMPs and their inhibitors in tumour invasion, metastasis and angiogenesis, and examine how MMPs may be targeted to prevent cancer progression.

To invade and metastasize, carcinomas must penetrate or lose their epithelial basement membrane (EBM), and then penetrate basement membranes (BMs) surrounding blood vessels, lymphatics, nerves and muscle cells. Knowledge of the composition of different BMs is necessary, so that appropriate antibodies and DNA probes are used to analyse these events. Laminin and type IV collagen are the principal BM components. However, recent studies show these two proteins exist in various isoforms, each of which is a heterotrimer of different subunit polypeptides. In this study, we analysed the distribution of laminin subunits, alpha 1 (lam), alpha 2 (lam), beta 1(lam), beta 2(lam) and gamma 1 (lam), and collagen IV subunits, alpha 1(IV), alpha 3(IV), alpha 4(IV) and alpha 5 (IV), in normal and neoplastic tissues of colorectum and breast. Subunits alpha 1(IV), alpha 1(lam), beta 1(lam) and gamma 1(lam) were detected in all BMs, while the distribution of alpha 3(IV), alpha 4(IV), alpha 5(IV) and alpha 2(lam) was much more restricted. In carcinomas, EBM staining for all subunits was invariably discontinuous or absent, consistent with the presence of complete EBM breaks. Use of antibody to alpha 1(lam) selectively stained the EBMs of carcinomas. Strong vascular staining for alpha 1(lam), beta 1(lam), gamma 1(lam) and alpha 1(IV) suggests an abundance of BM proteins in vessel walls, which may aid tumour cell attachment before vascular invasion. Within carcinomas, vascular BM staining for beta 2(lam) was clearly weaker than in normal tissues, which may reflect incomplete maturation of these vessels. Images Figure 1 Figure 2 Figure 3 Figure 4 Figure 5

Tumor cell invasion is now viewed as dysregulated physiologic invasion. Investigators have started to define the molecular events that are involved in this process. We find that there are many functional similarities with molecular events involved in physiologic process such as angiogenesis and wound healing. Matrix metalloproteinase activity is a common denominator in these pathologic conditions and in normal responses. Studies using endogenous metalloproteinase inhibitors suggest that targeting matrix metalloproteinase activity may prevent tumor cell dissemination. The development and pre-clinical testing of novel, low molecular weight matrix metalloproteinase inhibitors support this concept and suggest that an exciting new era of cancer therapy is on the horizon.

There have been reports that squamous cell carcinomas (SCC) and basal cell carcinomas (BCC) are surrounded by continuous epithelial basement membranes (EBMs). This argues against the hypothesis that EBM breaks are required for tumour invasion. We have used morphometric techniques to re-examine the evidence for SCCs and BCCs as objectively as possible. We assessed sections stained for type-IV collagen from 12 SCCs, 14 keratoacanthomas (KAs), 9 morphoeic BCCs, 10 nodular BCCs and 7 superficial multifocal BCCs. In the centre of these tumours, the EBM was generally more continuous than at the periphery, and this difference was statistically significant for SCCs, KAs and morphoeic BCCs (p < 0.01 in all). By considering central and peripheral tumour regions separately, a significant difference was seen between SCCs and the difficult to distinguish benign tumour KA. In the centre of KAs, EBM was significantly more continuous than in SCCs (p = 0.0029), which may suggest new ways of distinguishing these lesions. All of the SCCs and morphoeic BCCs examined showed clear evidence of EBM breaks, but some nodular BCCs did not. As nodular BCCs show an expansile growth pattern without typical histological features of tumour invasion, we suggest that these tumours may be at a pre-malignant stage. In general, our findings are consistent with the hypothesis that EBM breaks are required for tumour invasion. © 1996 Wiley-Liss, Inc.

ゼラチナーゼAは、マトリックスメタロプロテイナーゼ (MMP) ファミリーの一員である。このMMPファミリーは、普通不活性なプロエンザイムとして分泌されて、N末端ドメインの切断により活性化され、またTIMPsにより阻害される。それらは、細胞外マトリックスの再構築や、正常な成長と分化の過程、および腫瘍の浸潤や転移のような病理学的過程に関わっている。ゼラチナーゼAは、多くの点で他のMMPと異なった特徴を持っているのでこれらの過程において特別な役割を担っていると推測されている。多くのMMPsと異なり、ゼラチナーゼAは多くの細胞で構成的に発現されており、あらゆる組織に分布している。そのほかに特別な点としては、このプロテアーゼは最近発見された細胞膜上に存在する膜型MMP (MT-MMP) により活性化されるということである。この酵素の発現は、多くの腫瘍の浸潤性に関係しており、それは癌治療に対する的確な標的となると同時に、有用な予後の指標となると推測されている。

Many previous qualitative studies have shown that tumours are less vascular in the centre, and that host tissues become more vascular in close proximity to tumours. However, quantitative findings presented here for human colorectal cancer reveal some significant differences. Sections from 20 colorectal carcinomas (ten moderately and ten poorly differentiated) were immunostained with the QB/end/10 monoclonal to demonstrate blood vessels. These were measured by interactive morphometry and vascular volume density, surface density (Sv) and length density were recorded. In poorly differentiated carcinomas, the tumour centre was significantly less vascular than the periphery for all three parameters (P = 0.008 for Sv). However, no significant difference was seen for moderately differentiated tumours, which constitute the majority of colorectal cancers. Surrounding host tissues did not show a general increase in vascular density close to tumours. Furthermore, when total viable tissue was considered, the vascular density of carcinomas was not markedly different from normal mucosa. In the centre of moderately differentiated carcinomas for example, the mean value for Sv was only 1.4 times higher than the mean value for normal mucosa. These findings suggest that colorectal cancers may elicit a relatively weak angiogenic response, consistent with their exceptionally slow growth rate. Images Figure 1 Figure 2 Figure 3

Some invasive tumours characteristically have an abundant stroma rich in collagen, the production of which is termed the desmoplastic response. It has been suggested that this response may have a protective effect, and act to limit the process of tumour invasion. To investigate this possibility, we have examined various colorectal tumours for inter- and intra-tumoural variations in the desmoplastic response. As markers of this response, the distributions of collagen-I protein and myofibroblasts have been demonstrated by immunocytochemistry, while collagen-I messenger RNA has been demonstrated by in situ hybridization (ISH). Evidence of a desmoplastic response was obvious in carcinomas, but not in non-invasive adenomas. In carcinomas, we found that the response was marked in the tumour centre, where morphological features of active invasion have been reported to be absent. By contrast, we found little evidence of a desmoplastic response at the invasive edge of these carcinomas, where features suggestive of active invasion are prominent: in this location, collagen-I immunostaining was limited and myofibroblasts were sparsely distributed or absent. While our ISH results suggested active collagen-I synthesis in the tumour centre, there was little evidence of collagen-I synthesis in host tissues ahead of the invasion front. On the basis of these and other reported findings, we suggest that, while the desmoplastic response may reduce the invasive activity of neoplastic cells in the tumour centre, it fails to prevent the spread of colorectal cancer because of its deficiency at the invasive edge.

Biobanking has been identified as a key area for development in order to accelerate the discovery and development of new drugs. This review describes the recent advances in the field of biobanking and biospecimen research, with special reference to tumour banks which are the biobanks of primary interest in oncology.There is a dramatic deficiency of high-quality, well annotated cancer biospecimens. Biospecimen research is a fast developing field that will improve biobanking methodology and biobanking is becoming more professionally organized with increased attention to quality management. Biobank networks are developing rapidly in order to combine and share resources.Biobanking services must improve rapidly to serve the needs of personalized medicine and biospecimen research should be encouraged and supported at all levels from project funding to publication of results. Biobanks need to be run to high professional standards and the importance of adequate funding, training and certification must be emphasized. The growing presence of national and international biobank networks will allow biobanks to synergize. The development of a biobanking community will facilitate teamwork to overcome common challenges and enhance communication with multiple stakeholder groups.

Increased collagenase activity in colorectal carcinomas has recently been shown to be associated with increased malignant potential. To determine the tissue distribution of collagenase and its specific inhibitor, tissue inhibitor of metalloproteinases (TIMP), we carried out an immunohistochemical study on colorectal carcinomas (n = 20), adenomas (n = 7) and normal mucosa (n = 6). We found increased staining for collagenase in the connective tissue stroma of carcinomas, as compared with adenomas and normal mucosa. Little evidence of epithelial cell staining for collagenase was seen in any tissue. In carcinomas, both stromal fibroblasts and collagen fibres stained strongly and stromal staining was strongest close to neoplastic glands. Vascular staining was more prominent in neoplastic than normal tissues, perhaps reflecting the increased proteolytic activity during tumour angiogenesis. The pattern of TIMP immunostaining was similar to that of collagenase, although basement membrane staining for TIMP was generally more intense. Another difference was that, unlike TIMP, staining for collagenase was often increased at the invasive edge of carcinomas, perhaps reflecting increased collagenase activity at this location.

In this study, we report on the distribution of tissue inhibitor of matrix metalloproteinase-3 (TIMP-3) mRNA expression in human normal colorectal mucosa, adenomas and adenocarcinomas. Northern blot analysis showed five TIMP-3 mRNA transcripts to be present in normal mucosal epithelium and in moderately and poorly differentiated carcinoma. Adenomas and well-differentiated carcinomas were not examined in this part of the investigation. In situ hybridization studies showed no detectable TIMP-3 mRNA in normal and adenomatous tissue. In contrast, TIMP-3 mRNA is localized to stromal fibroblast-like cells in colorectal carcinomas, with an increased incidence in moderately and poorly differentiated groups compared with well-differentiated carcinomas. Expression in both the moderately and the poorly differentiated tumour groups was strongest at the tumour invasive edge; none of the poorly differentiated carcinomas showed mRNA expression in regions ahead of the invasive edge, compared with 3 of 12 of the moderate group. To our knowledge, this is the first detailed report on the regional localization of TIMP-3 mRNA in colorectal tumours. We suggest that the lack of TIMP-3 mRNA expression in host stromal tissues ahead of poorly differentiated carcinomas may contribute to their increased invasiveness. Images Figure 2 Figure 3

Remodeling of the extracellular matrix (ECM), which occurs during many physiological and pathological processes, is one of the requisite events of cellular invasion. The matrix metalloproteinases (MMPs) are a family of zinc-dependent proteases that are responsible for proteolytic degradation of specific ECM components. Regulating the activity of the MMPs at both mRNA and/or protein levels modulates the degradation of the ECM components which in turn alter cellular invasion. Although most MMPs are regulated via similar mechanisms at the mRNA and protein levels, the modulation of gelatinase A is unique. Understanding the mechanisms that regulate gelatinase A is important since expression and activation of this particular MMP is consistently correlated with a majority of malignant phenotypes. In this report, we will contrast the mechanisms that regulate the expression, activation and inhibition of gelatinase A with the mechanisms that modulate the rest the MMP family.

Extracellular matrix (ECM) turnover is an event that is tightly regulated. Much of the coordinate (physiological) or discoordinate (pathological) degradation of the ECM is catalyzed by a class of proteases known as the matrix metalloproteinases (MMPs) or matrixins. Matrixins are a family of homologous Zn atom dependent endopeptidases that are usually secreted from cells as inactive zymogens. Net degradative activity in the extracellular environment is regulated by specific activators and inhibitors. One member of the matrixin family, gelatinase A, is regulated differently from other MMPs, suggesting that it may play a unique role in cell–matrix interactions, including cell invasion. The conversion from the 72 kDa progelatinase A to the active 62 kDa species may be a key event in the acquisition of invasive potential. This discussion reviews some recent findings on the cellular mechanisms involved in progelatinase A activation and, in particular, the role of tissue inhibitor of matrix metalloproteinases-2 (TIMP-2) and transmembrane containing metalloproteinases (MT-MMP) in this process.Key words: tissue inhibitors of metalloproteinases, metalloproteinase, gelatinases, extracellular matrix, activation.

Qualitative histological studies in the distribution of urokinase-type plasminogen activator (uPA) in human colorectal carcinomas have been well documented. However, to our knowledge the histological distribution of this enzyme has not been quantified in any tumour. For the present image analysis study, uPA was demonstrated in sections of human colorectal cancer using immunoperoxidase technique. A total of 9 colorectal carcinoma cases were used, in which 132 regions were analysed. Within each region, staining intensity measurements were made at evenly spaced intervals. Samples of normal mucosa from 6 cases were also studied. Enzyme levels were assessed with staining intensity measurements. For each section, a negative control section was included, in which the primary antibody was omitted. Staining for uPA was quantified for each region in the test section, and the measurement for the corresponding region of the negative control was then subtracted. The enzyme uPA was localised more at the tumour edge than in the tumour centre or distant host tissue. These differences were highly significant (p < 0.0001). There was also a highly significant difference in staining intensity when tumour regions adjacent to pushing edge were compared with those adjacent to infiltrating edge (p < 0.0001). Infiltrating tumours showed stronger staining for uPA than tumours with pushing edges. Since invasive activity is thought to be maximal at the edge of the tumour, localisation of uPA at this site is consistent with the role of this enzyme in the process of tumour invasion. © 1995 Wiley-Liss, Inc.

For colorectal carcinomas, there is evidence that marked discontinuity of the epithelial basement membrane (EBM) is associated with higher malignant potential. Since the meta-static process appears to be selective, more discontinuous EBMs might be expected in secondary rather than in primary tumours. To test this prediction, we examined a series of 60 cases of colorectal carcinoma for which samples of lymph-node or liver metastases were available. Sections were immunocyto-chemically stained for laminin, and the continuity of tumour EBM was then assessed by observational rating as well as by detailed morphometric analysis for a sample of cases. Contrary to the above prediction, we find that EBMs tend to be more continuous in secondary tumours than in corresponding primary tumours. These results could be explained by the possibility that local tissue environmental factors have a major influence on EBM continuity. Supporting evidence comes from our previous observation that EBM is very discontinuous at the advancing edge of primary colorectal carcinomas, where the tumour is adjacent to collagen-I-deficient stroma. Further evidence from the present study is that the EBM is extremely discontinuous at the interface between metastases and specialised parenchymal tissue, but more continuous at the interface between metastases and stromal connective tissue. Since basement membranes affect the differentiation and behaviour of adjacent cells, these findings suggest that host tissues may influence invasive activity through their effects on EBM continuity.

Many studies suggest that increased proteolysis accounts for the epithelial basement membrane (EBM) breaks commonly seen in carcinomas. As failure to produce or maintain EBM may also be important, we chose to investigate synthesis of basement membrane collagen-IV in human colorectal carcinomas. First, to determine the cellular origin of EBM collagen-IV, species-specific antibodies were used to analyse caecal xenografts of 4 different human colorectal-carcinoma-derived cell lines. The results of this study suggest an exclusively stromal cell origin for EBM collagen-IV. Next, the distribution of periglandular myofibroblasts in carcinomas was examined, since in normal mucosa their location and ultrastructural features suggest that they play a role in EBM maintenance. They were generally abundant in normal mucosa and adenomas, but sparsely distributed in carcinomas, particularly at the invasive periphery where EBM collagen-IV immunostaining is most deficient. Finally, the in situ hybridization technique was used to define cell populations synthesizing collagen-IV. In normal mucosa, no collagen-IV mRNA was detected in any component, while in carcinomas, the mRNA was clearly detectable in vascular endothelial cells but not in any other cell type. Increased vascular collagen-IV production in carcinomas may be at least partly due to tumour-induced angiogenesis, since new blood-vessel formation requires the synthesis of new vascular basement membranes.

Previous studies on colorectal carcinomas indicate that consistent differences in epithelial basement membrane (EBM) integrity are present between the tumour centre and periphery. We report that within the tumour centre, EBM staining between back-to-back (BTB) neoplastic glands (i.e., adjacent glands in direct contact with no intervening connective tissue) generally follows a pattern different from that of EBM staining at the tumour:stromal interface (TSI). Such distinctions are important, since the factors responsible for EBM deficiencies may vary with intra-tumoural location, as may the prognostic significance of these deficiencies. Analysis of paraffin sections from 130 colorectal carcinoma cases showed that EBM staining between BTB glands is generally weaker and more discontinuous than at the TSI, sometimes appearing as a linear array of immunostained granules on high-resolution light microscopy. By double-labelling immunofluorescence analysis of cryostat sections from 30 cases, a decrease in type-IV collagen:laminin staining intensity ratio was found between BTB glands. Hence, the composition of EBM between BTB glands appears to be abnormal. As much recent evidence indicates that epithelial:mesenchymal interactions play an essential role in EBM formation, the demonstration of immunostained EBM fragments between BTB glands requires an explanation: We suggest that the synthesis of EBM between BTB glands involved previously intervening stromal (mesenchymal) cells, and that EBM fusion and dissolution occur between BTB glands following the displacement of these cells.

In colorectal cancer an association has been found between lack of epithelial basement membrane (EMB) immunostaining in the tumour centre and more extensive malignant spread. Interestingly, ultrastructural investigations suggest that EBM loss at the tumour periphery may be part of an invasive mechanism. To further assess the significance of EBM deficiencies in different tumour areas, we carried out a detailed study of the basement membrane laminin immunostaining patterns in 130 cases of colorectal carcinoma. We find that discontinuous EBM staining in the tumour centre is associated with poor tumour differentiation (p less than 0.005), presence of lymph-node metastases (p less than 0.02), and more advanced Dukes stage (p less than 0.02). The latter association is strengthened by excluding cases in which numerous polymorphonuclear leukocytes (PMNs) are present adjacent to EBM breaks, suggesting that these inflammatory cells are a confounding factor. Discontinuous EBM staining is more frequently observed in tumour deep to muscularis propria than in submucosal tumour (p less than 0.02), indicating intra-tumoral variation. At the tumour periphery, extensive EBM discontinuity shows no association with lymph-node involvement, but is linked with deeper local invasion (p less than 0.05). While EBM staining patterns around central and peripheral tumour glands are related (p less than 0.001), staining around peripheral glands is almost invariably more discontinuous. However, EBM lack at the tumour periphery is not as absolute as previously suggested, since in 18% of tumours fewer than 25% of peripheral tumour glands show EBM breaks. This appears consistent with the hypothesis that invasive changes at the tumour periphery are temporary and reversible.

plays an essential r61e in computing the amount of this overshoot. While teaching a course in the theory of functions of a real variable, E. HEWITT found the value 1.71... listed for the integral (1) in HARDY & ROGOSINSKI [271, page 36. This anomaly, as well as others encountered in the literature, led us to a study of the GIBBS phenomenon and its history. In the course of this study we uncovered a maze of forgotten results, interesting and difficult generalizations, faulty constants, and some details about the GIBBS phenomenon that have escaped the attention of many writers on the subject. Despite the familiarity of our theme, we therefore entertain a hope that readers of the Archive will find some interest in a discussion of this corner of FOURIER analysis. The paper is divided into three Parts. In Part I, we examine GIBBS's phenomenon in some detail. In Part II, we take up its curious history and describe briefly some of its congeners. In Part III, we offer some conclusions. The computations given in this paper were carried out on two computers: a Hewlett-Packard 9810 and a Univac 1110. The graphs (barring the simplest) were drawn by a Hewlett-Packard 9862A plotter. All finite decimal expansions are truncated decimal expansions. It is a pleasure to record our indebtedness to GERALD B. FOLLAND, THOMAS L. HANKINS, EINAR HILLE, and STEPHEN P. KEELER, who have made valuable suggestions to us.