Comparison of hyaluronidase expression, invasiveness and tubule formation promotion in ER (-) and ER (+) breast cancer cell lines in vitro.

BACKGROUND: Hyaluronidase (Hyase) is an enzyme which hydrolyses hyaluronan (HA), a large nonsulfated glycosaminoglycan. Several genes have been identified to code for hyaluronidases in humans. Its role has only recently been underlined in the invasion of prostate cancer, colonic cancer, and breast cancer. Moreover, the findings were in agreement with some experimental results which showed that HA-derived oligosaccharides had angiogenesis-promoting activity. All these findings prompted us to investigate factors that had been characterized as putative invasive factors in different human breast cancer-derived cell lines. METHODS: We selected two series of human breast cancer-derived cell lines whose expression of estrogen receptors (ER) was previously published. Hyaluronidase secretion in culture medium and expression of matrix metallo-proteinase (MMP)-9, cathepsin-D (cath-D) and vascular endothelial growth factor (VEGF) by cells were determined. We also investigated cell invasiveness in the Matrigel invasion assay, and studied the capability of cancer cells to promote in vitro formation of tubules by endothelial cells. RESULTS: ER(-) cells secreted significantly more hyaluronidase (P < 0.001) and expressed significantly more VEGF (P < 0.01), MMP-9 (P < 0.05) and cath-D (P < 0.0001) than ER(+) cells. Invasion through Matrigel by ER(-) Hyase(+) cells was significantly higher than that by ER(+) Hyase(-) cells (P < 0.05). In both cases, invasion was decreased by heparin (P < 0.05). When ECV-304 endothelial cells were co-cultivated in millicell chambers with cancer cells, ECV-304 cells were induced to form tubules. Tubule formation was demonstrated to be more prominent with ER(-) Hyase(+) cells than with ER(+) Hyase(-) cells (P < 0.05). CONCLUSION: Invasive features of ER(-) breast cancer cells can be characterized in vitro by an invasive Matrigel assay, as the induction of tubule formation by ECV-304 endothelial cells, higher secretion of hyaluronidase, and higher expression of proteinases MMP-9, cath-D, and the angiogenesis promoting factor VEGF.

Hyaluronectin modulation of lung metastasis in nude mice.

Hyaluronectin (HN) is a glycoprotein with a high affinity to hyaluronic acid (HA) and known to be a component of the extracellular matrix of tumours. Clinical studies have shown that a low level of HN correlates to tumours with poor prognosis, whereas a high level of HN correlates to tumours with good prognosis. We previously demonstrated in vitro that hyaluronidase activity, which promotes tumour progression and metastatic spread by degradation of HA into angiogenic oligosaccharides, was inhibited or promoted by HN, according to the level of HN-expression. This raises the question of the role played by HN in cancer, and particularly if high and low levels of HN-expression could trigger opposite effects on tumour growth and/or metastatic spread. To address this issue, we used a model of spontaneous lung fluorescent metastases that we characterised previously. We stably transfected the human HN cDNA into fluorescent H460MGFP cells and selected two clones characterised by different levels of HN-expression: HN110 and HN704, with a high and a low level of HN-expression, respectively. In vitro, we demonstrated that HN704 cell migration was significantly increased. Inoculation of clones to nude mice had no significant effect on tumour growth, but clearly revealed opposite effects on metastatic spread: HN110 significantly decreased the number of fluorescent metastases whereas HN704 significantly increased it. We also analysed HN, HA and hyaluronidase contents in sera and tumours. These results demonstrate that HN can play a role as either a suppressor or promoter of metastatic spread.

Hyaluronate and its receptors in bone marrow.

Cell-cell and cell-matrix interactions, which are mediated by cell adhesion molecules, play a fundamental role during many cellular processes including growth, differentiation, cell migration and cancer metastasis. One molecule playing a major role in these processes is the CD44 surface receptor, which is expressed in a wide range of cells including many cells of the hemopoietic system, where it mediates the interaction with its major ligand, hyaluronate. However, little is known about CD44 and hyaluronate in bone marrow and this was investigated immunohistochemically in trephine biopsies and in cultivated human bone marrow stromal cells. In biopsy specimens, patches of hyaluronate deposition were detected in the extracellular matrix (ECM). However, most of the areas of the ECM were devoid of hyaluronate. Single mast cells and lymphocytes scattered throughout the marrow were CD44 immunopositive. Marrow-derived stromal cells (MDSC) expanded in cell culture were immunopositive for CD44, hyaluronate synthase, and hyaluronate. Hence, a marked difference between CD44 immunolocalisation and hyaluronate deposition can be observed between in situ and under cell culture conditions. Since in normal marrow in situ the number of CD44 immunopositive cells was low, interactions of CD44 and hyaluronate would appear to not to play a major role in cell adhesion in the normal bone marrow.

Reticulated hyaluronan hydrogels: a model for examining cancer cell invasion in 3D.

The extracellular polysaccharide hyaluronan (HA) controls cell migration, differentiation and proliferation, and contributes to the invasiveness of human cancers. The roles of HA cell surface receptors and hyaluronidases (HAses) in this process are still controversial. In order to investigate their involvement in cancer pathogenesis, we developed a reticulated HA hydrogel, a three-dimensional matrix in which cells can invade and grow. We have studied thirteen cell lines, from primary tumors or metastases, that migrated into the HA hydrogel and proliferated giving rise to clusters and colonies. The number of colonies, which reflects tumor cell invasiveness, ranged from 7 to 193 after 5 days of culture. Invasion was dependent on the production of HAse as well as other factors. Optimal colonization occurred when cells released HAse, lacked HA-binding sites and did not secrete HA. Moreover, we describe for the first time a HAse activity at physiological pH that may be responding to the confinement of the enzyme in a three-dimensional structure. We show here that this reticulated matrix provides a three-dimensional model for investigating mechanisms involved in malignant invasion.

Biodistribution of injected tritiated hyaluronic acid in mice: a comparison between macromolecules and hyaluronic acid-derived oligosaccharides.

BACKGROUND: Hyaluronan (HA) has been reported to bind specifically and with high affinity to various cell types and to directly modify cell behaviour. In a previous report we demonstrated that both high molecular weight molecules (HA(H)) and HA-derived oligosaccharides were efficient at triggering terminal differentiation of acute myeloid leukemia (AML) blasts, in vitro, through CD44 ligation. MATERIALS AND METHODS: To explore the possibility of using HA for a differentiation therapy in AML, we investigated whether intravenous injection of tritiated HA(H) and/or HA-derived oligosaccharides (HA10-20) into mice accumulated in bone marrow, the main site of AML cell proliferation. RESULTS: The present work showed that the level of HA in bone marrow: 1) was maximum 5 hours after injection of either HA(H) or HA10-20; 2) was about 40 times higher after HA(H) than after HA10-20 injection. The amount of HA in bone marrow (5.8% ID/g) was two-fold higher than in serum, indicating that it was not due to circulating blood. Finally, using chromatographic analysis, we showed that about 34% of tritiated HA present in bone marrow 5 hours after HA(H) injection displayed a size higher or equal to HA10. CONCLUSION: After a single injection of macromolecular hyaluronan in mouse bone marrow we obtained a concentration of oligosaccharides close to the one shown to trigger AML cell differentiation in vitro. A part of the oligosaccharides had a size higher than or equal to the minimal one required to interact with HA receptors.

Inhibition of arterial cells proliferation in vivo in injured arteries by hyaluronan fragments.

It has been demonstrated previously that administration of high levels of high molecular mass hyaluronan (hyaluronic acid, HA) to rats was able to reduce in a significant way neointima formation in the injured arteries. In the present study, our aim was to verify whether small forms of HA (4-16 saccharides) are still able to reduce the proliferative response of ASMC to aortic injury. Treated rats received a total of 8 injections of a fixed dose of HA fragments (27 mg/kg rat contained in a volume of 550 microl). Two injections were given on the day of balloon catheter injury (BCI): one, intravenous, 10 min before BCI and one, subcutaneous, immediately after the BCI. The others injections (subcutaneous) were at 2, 4, 6, 8, 10 and 12 days after BCI. Control rats received an equivalent volume of the dissolving buffer containing only hyaluronidase, which has been destroyed before injection to rats. Neointima formation was analysed 14 days after the BCI. Intima-media wet weight and DNA content were significantly reduced in rats receiving HA fragments in comparison to controls (2P=0.01 for wet weight and 0.03 for DNA). This finding was confirmed by the histomorphometric study which showed that both neointima area and the ratio neointima/neointima+media were significantly decreased in treated rats (2P=0.03 for intima area and 0.049 for the ratio). Our data showed thus and for the first time that administration of HA fragments with a very low molecular mass (4-16 saccharides) reduces the proliferative reaction of aorta to injury in vivo. In conclusion, HA fragments, which are components with an excellent safety profile, may offer hope for the prevention of restenosis after angioplasty.

Hyaluronidase in sera of tumour-bearing nude mice.

Cancer cell lines often secrete hyaluronidase, suggesting that this enzyme could be used as a marker of growing tumours. We have measured hyaluronidase in the sera of non-grafted mice and mice grafted with human tumour-derived hyaluronidase-secreting H460M and SA87 cells or non-secreting CB 193 cells. Mouse serum hyaluronidase was measured at pH 3.8 using the enzyme-linked sorbent assay (ELSA) technique by reference to human serum whose activity at pH 3.8 was determined by the Reissig technique. The serum hyaluronidase in non-grafted mice ranged from 310-520 mU l(-1) (mean+/-SD 432+/-70 mU l(-1), median 440 mU l(-1)). Hyaluronidase increased in the sera of tumour-bearing mice grafted with H460M cells or with SA87 cells, but not in the sera of mice grafted with CB 193 cells. Serum hyaluronidase activity in H460M or SA87 tumour-bearing mice correlated with the tumour mass, increased with time, and decreased after tumour removal. Zymography detected two different hyaluronidase forms in the sera of non-grafted mice: type 1 had only one hyaluronidase band and type 2 had five different bands. In both types, enzyme augmentation in tumour-bearing mice correlated with the presence of an additional enzyme band that was not seen in normal sera and that migrated as the cancer cell enzyme did; there was no augmentation of the normal isoform(s). These results show that serum hyaluronidase can be used to follow the development of tumours in mice grafted with hyaluronidase-secreting cells.

Expression of hyaluronate and hyaluronate synthase in human primary tumours and their metastases in scid mice.

Hyaluronate and hyaluronate synthase expression were examined in primary tumours and if present in metastatic deposits of human breast, colon, ovarian and small-cell lung cancer cell lines transplanted into scid mice using biotinylated hyaluronectin and immunohistochemical staining of hyaluronate synthase. Very intensive hyaluronate and hyaluronate synthase expression could be observed in peripheral areas of tumours derived from highly metastatic cell lines (HT29, MCF-7). Even smaller lung metastases of up to 15 cells showed typically a focal binding of hyaluronectin predominantly at the host-tumour interface of the metastases, indicating that increased expression is closely correlated with the degree of invasiveness and metastatic potential of malignant tumours.

Human monocytes synthesize hyaluronidase.

The involvement of hyaluronic acid (HA) oligosaccharides and blood-derived mononuclear cells in inflammatory processes prompted us to determine whether peripheral blood mononuclear cells (PBMC) possess hyaluronidase activity. PBMC were incubated with macromolecular-tritiated HA at pH 3.8 and supernatants were analysed by size exclusion chromatography to reveal digestion of HA. This digestion was due to the CD14-positive (CD14+), adherent, non-specific esterase-positive, subpopulation of PBMC. Hyaluronidase activity (72 kDa) was found in aqueous and non-ionic detergent PBMC extracts but not in the medium in which the cells had been cultured. These results indicate that hyaluronidase is, at least in part, linked to the membrane rather than excreted. Hence, monocytes have one or more hyaluronidases that can generate a pool of active HA fragments within tissues. Hyaluronidase activity was also found in 3/3 myelomonocytic lineage leukaemias but not in 3/3 lymphoblastic leukaemias.

Hyaluronidase is more elevated in human brain metastases than in primary brain tumours.

BACKGROUND: Hyaluronidase is hypothesised to play a role in cancer invasion and metastasis formation. MATERIALS AND METHODS: Hyaluronidase activity was investigated at pH 3.8 in extracts of 30 human brain tumours (17 glioblastomas and 13 brain metastases of carcinomas) and in cancer cell cultures with the ELSA method and zymography. RESULTS: In brain metastases, hyaluronidase activities were significantly higher than in glioma extracts (9.16 +/- 4.48 mU/g vs 4.25 +/- 5.74) which was not explained by serum hyaluronidase contamination. Serum hyaluronidase of tumour patients' sera was within the normal values determined in 28 matched blood donors'sera (33.8 +/- 11 U/l). The maximum hyaluronidase/albumin (U/g) ratio was 0.9, below which the hyaluronidase content of tumours was below the maximum value calculated from the albumin content of the tumour extract and could not be considered as local production by tumour cells. The hyaluronidase content and hyaluronidase/albumin ratio of metastasis extracts was significantly higher than in glioma extracts and patients' sera, whereas no significant difference was found between the ratios of glioma extracts and sera. The production of hyaluronidase was studied in cell extracts and in culture media of 3 human glioma-derived cell lines and of the brain metastasis-derived cell line SA87. Hyaluronidase activity of the metastasis-derived cell line SA87 was 100 to 1000-fold that of glioma cell lines. CONCLUSION: These results suggest that hyaluronidase is associated with the more aggressive cancer cells and is directly or indirectly involved in brain metastasis phenotype.

Importance of hyaluronan length in a hyaladherin-based assay for hyaluronan.

Specific hyaladherin-based assays have been set up to measure the concentration of hyaluronan in biological fluids. Hyaluronectin (HN; a hyaladherin extracted from ovine brain) binds to hyaluronan (HA) that must be 10 units (HA10) or more long. It was therefore of interest to determine whether HN would continue to bind to HA10 in full-length HA since conformational changes might mask potential binding sites. We used the enzyme-linked sorbent assay (ELSA) to assay HA and hyaluronan-derived oligosaccharides, with different standard HAs, and the results were compared to results obtained with the carbazole technique. Oligosaccharide length was calculated from the ratio glucuronic acid/reducing N-acetylglucosamine in fractions of hyaluronidase-digested macromolecular hyaluronan prepared by chromatography; the size of the HA12 oligosaccharide was confirmed by matrix-assisted laser desorption ionization mass spectrometry. During the digestion of macromolecular HA with hyaluronidase, the binding of HN to HA first increased and then decreased as shown using the ELSA. The concentration of HA fragments of HA60 and below was overestimated when intact macromolecular HA was used as the reference for the ELSA, while the concentration of HA100 and above was underestimated when HA10 was used as the reference. The binding of HN to HA20, HA40, and HA60 saccharides was consistent with binding to multiples of HA10 sites. In conclusion, the level of HN binding is determined by the conformation of HA, which may mask binding sites. Hence, calibration HA used in the ELSA must be adapted to the size of HA to assay.

Inhibition of tumor growth and metastatic spreading by overexpression of inter-alpha-trypsin inhibitor family chains.

The inter-alpha trypsin inhibitor (ITI) family is a group of proteins built up from different combinations of I light chain (ITI-L) and 3 highly homologous heavy chains (ITI-HI, -H2 and -H3). To investigate a potential role of the ITI family chains in cancer and metastasis spreading, we engineered human H460M cell lines expressing both the green fluorescent protein (GFP) and one of these chains. These clones were subcutaneously injected in athymic nude mice, and lung metastasis number and primary tumor weight were determined after 28 days. Expression of the ITI-L chain considerably decreased tumor weight and fluorescent lung metastasis number. ITI-HI and ITI-H3 chain expression induced a significant decrease of metastasis number, whereas no decrease of tumor weight could be detected. In vitro, ITI-L expression significantly decreased chemotaxis and ITI-HI and ITI-H3 expression increased cell attachment. These results argue for the antitumoral or antimetastatic properties of ITI-L, -HI and -H3 chains.