CLIN. EXPL. METASTASIS, 1989, VOL. 7, NO. 3, 353--359

Effects of l a m i n i n on the a t tachment of g l ioma ce l l s to type IV co l lagen

G I L L I A N H U N T t and G. V. S H E R B E T ~

Cancer Research Unit, The Medical School, University of Newcastle upon Tyne, Newcastle upon Tyne NE2 4HH, U.K.

(Received 20 August 1987; accepted 12 July 1988)

The interaction of tumour cells with basement membrane components is thought to be important in influencing their invasive and metastatic properties. This paper describes the effect of laminin on the attachment of radiolabelled glioma and B16 murine melanoma cells to tissue culture plastic and type IV collagen. With the exception of the non-metastatic B16 F1 variant, laminin (and fibronectin) stimulated cell attachment to tissue culture plastic. Although laminin stimulated the attachment of the B 16 BL6 metastatic variant to type IV collagen, it consistently inhibited the attachment of the glioma cells under the same conditions. Laminin appeared to exert its effect by adsorption to the collagen and was not cytotoxic to the glioma cells. In contrast, fibronectin had very little effect on cell attachment to type IV collagen. One of the most unusual features of glioma is the rarity of metastasis to extraneural sites. However, the effect of laminin observed here may not be the only factor involved in the metastatic inefficiency of this tumour type.

Introduct ion L a m i n i n regula tes a var ie ty of b io logica l p h e n o m e n a , i nc lud ing a t t a chmen t ,

g rowth , m o r p h o l o g y and m i g r a t i o n of cells. T h e s e p r o p e r t i e s can be re la ted to its large size and m u l t i - d o m a i n s t ruc tu re [13].

T h e ab i l i ty of t u m o u r cells to in te rac t wi th b a s e m e n t m e m b r a n e s is c ons ide r e d i m p o r t a n t in in f luenc ing the i r invas ive and me tas t a t i c p r o p e r t i e s [6] and m u c h a t t en t ion has been focused on the role of l amin in in th is process . L a m i n i n increases b o t h the rate and n u m b e r of me tas t a t i c m u r i n e t u m o u r cells a t t ach ing to t ype IV col lagen and cells a t t ached in the p resence of l amin in give rise to m o r e p u l m o n a r y metas tases than cells which have no t been ' s e l ec ted ' in th is way. P r e - i n c u b a t i o n of cells wi th an t ibod ie s to l amin in is k n o w n to reduce the n u m b e r of s u b s e q u e n t lung metas tases [1 2]. In cont ras t , l amin in inh ib i t s the a t t a c h m e n t of m a c r o p h a g e s [3] and s q u a m o u s ca r c inoma cells [14] to type IV col lagen.

One of the mos t unusua l fea tures of gl iorna is the i n f r e q u e n c y of ' s p o n t a n e o u s ' metas tas i s to ex t r aneu ra l sites, a l t hough the b ra in i tself is a fa i r ly c o m m o n site for the me tas t a t i c sp read of sys temic t u m o u r s . T h e m a j o r i t y of d o c u m e n t e d cases of g l ioma metas tas i s fol low surgical in te rven t ion , and d i s s e m i n a t i o n in the absence of

t Present address: Department of Pathology, Royal Victoria Infirmary, Newcastle upon Tyne NE1 4LP, U.K.

To whom correspondence should be addressed.

0262--0898/89 $3"00 �9 1989 Taylor & Francis Ltd.

354 G. Hunt and G. V. Sherbet

craniotomy, although not unknown, is a rare occurrence [9]. In this s tudy we investigated the effects of laminin on glioma cell a t tachment to tissue culture plastic and type IV collagen.

Materials and m e t h o d s

Preparation of labelled cells H u m a n malignant astrocytic glioma cell line U373 M G was obtained f rom

the American T y p e Culture Collection (Rockville, M D , U.S.A.) and rat glioma line C6 was obtained f rom Flow Laboratories. H u m a n glioma cell lines G - I J K t and G - U V W , originating f rom anaplastic astrocytomas, were provided by Dr R. I. Freshney (Depar tment of Medical Oncology, Universi ty of Glasgow, U.K.) . The metastatic variants BL6 and F1 of murine melanoma B16 were obtained f rom E. G. and G. Mason Research Institute, (Worcester, MA, U.S.A.). BL6 was shown to metastasize extensively to the lung, whereas F1 had no such ability [ | 1]. I t was not intended to use the latter cell lines in direct comparison with the gliomas: the effect of laminin on the a t tachment of the BL6 variant had been described previously [12] and therefore this cell line was useful as a 'positive control ' in this investigation.

Cells at 50 per cent confluence in 75 cm / flasks (Nunc) were cultured for a further population doubling time in 25 ml of med ium containing 10 #1 of 1 mCi /ml [methyl- 3H]thymidine (Amersham). After m i n i m u m trypsin harvesting, the cells were resuspended in serum-free medium ( D M E M supplemented with 200pg/ml of bovine serum albumin), henceforth known as assay medium. T h e cells were washed three times in this med ium and their concentration was adjusted to 1 x 10 6 viable cells/ml in assay medium.

Attachment assay Plastic tissue culture wells (35 mm) (Nunc) were coated with 10 #g of human type

IV (basement membrane) collagen (Sigma) [5], then rinsed three.t imes with assay med ium and incubated for 1 h at 37~ with 1 ml of assay med ium containing either 3 #g of laminin isolated f rom the E H S sarcoma (BRL, Cambridge, U.K. ) or 8 #g of human plasma fibronectin (Sigma). Following the addition of the cell suspension, these concentrations were equivalent to 2#g/ml of laminin and 5"3 #g/ml of fibronectin. L, aminin (at 1 pg/ml initial concentration) had previously been shown to stimulate the a t tachment of B16 BL6 cells to type IV collagen [12], whereas in this study an initial concentration of 3 #g/ml of laminin had a m a x i m u m effect on the a t tachment of G - U V W cells (data not shown); previous similar studies [3, 12] used initial fibronectin concentrations of 5 and 10 pg/ml, respectively. Control collagen- coated wells containing only 1 ml of assay med ium were also incubated. In a parallel experiment, uncoated wells were also incubated in the presence or absence (control) of the attachment factors. Viable, labelled cells (5 x 105) in 0"5 ml of assay medium were added and the wells incubated at 37~ for a further 1"5 h (previous studies have utilized at tachment times of between 30min [14] and 3 h [12]). At the end of the incubation period, the supernatant med ium containing the non-adherent cells was removed and each well was gently rinsed twice with 1 ml of assay medium. The rinse solution was added to the supernatant med ium and centrifuged to obtain a cell pellet. This pellet of non-adherent cells and also the cells adhering to the wells were lysed by the addition of I ml of 0"1 M N a O H followed by incubation at 37~ for I h. The

Laminin as an attachment factor 355

solutions were neutralized by the addition of 0"1 ml of 1 M HCI and counted for radioactivity. The percentage attachment (A) was calculated from the equation

A = 100 x cpm in adherent cells cpm in (adherent + non-adherent) cells

Effects of different laminin treatments on the attachment of human glioma cells to type I V collagen

In addition to the experiments described, attachment of glioma cells to type IV collagen was investigated under three other conditions [3]. The basic technique was as described previously.

Laminin 'washed'. The coated wells were pre-incubated with laminin as previously. Before the cells were added, the medium was removed and the wells rinsed three times with assay medium. A volume of 1"5 ml of assay medium (without laminin) containing 5 x 105 cells was added and the wells were incubated at 37~ for a further 1"5 h.

Laminin 'late'. The assay was performed as for the collagen control but 3 #g laminin were added to each well 45 min after the cells and the wells incubated for a further 45 min.

Laminin-treatedcells. Cells (1 • 106) were suspended in assay medium containing laminin at 2 #g/ml and incubated at 37~ for 1 h with gentle agitation. The cells were washed three times and resuspended in assay medium without laminin at 1 • 1 0 6

viable cells/ml; 5 x 105 cells were added to collagen-coated wells which had been pre- incubated for 1 h in assay medium and the wells were then incubated at 37~ for a further 1"5 h.

R e s u l t s

Effects of laminin on cell attachment to tissue culture plastic and type I V collagen The results are summarized in table 1; figures are means + standard deviations

of at least five wells from two independent experiments for the type IV collagen and at least three wells for the plastic. Significant differences are seen with respect to the type IV collagen or plastic control, as appropriate, analysed by the two-sample t-test. The standard deviation of counts recovered per well did not exceed 10 per cent of the means within experiments and release of label into the medium was negligible (data not shown).

There was a wide variation in the attachment abilities of the glioma cells in the absence of added attachment factors, although it was not intended to use the C6 rat cells in direct comparison with the human cells. G- I JKt , G - U V W and B16 BL6 cells were significantly more adherent to type IV collagen than to tissue culture plastic in the absence of exogenous attachment factors (p<0"01, p<0"001 and p<0"01, respectively), whereas C6 and B16 F1 attached equally to both surfaces under these conditions. The non-metastatic F1 cells were significantly more adherent to plastic than the metastatic BL6 cells (p<0-001). In contrast to these findings, the B16 variants did not differ in their attachment to type IV collagen. With the exception of the F1 cells, laminin and fibronectin significantly increased the attachment of all the cell lines to tissue culture plastic. Laminin reduced the attachment of the F1 variant to tissue culture plastic relative to control and fibronectin had no effect. The only significant effect of fibronectin on attachment to type IV collagen was to inhibit the

356

Table 1.

G. Hunt and G. V. Sherbet

Effects o f l a m i n i n a n d f ibronect in on c e l l a t t a c h m e n t to type IV c o l l a g e n a n d t i ssue cu l ture plast ic .

Mean attachment (A)__ S.D. (per cent)

Type IV collagen Tissue culture plastic

Cell line Control +Lamin in +Fibronect in Control +Lamin in +Fibronectin

Gliomas: C6 58-5___4'4 34"2_3'6 c 67"5+8"3 55-9___7-4 74"2+ 5"9 b 77"5__ 10"0 b G-IJKt 66"8_+5"4 44 '5_ 5'2 b 55'2+5"6" 23"4_+2-5 34"9_ 5"4 a 35"2+3"6 a G-UVW 84"7+7-0 41"5+8"0 b 78"7+7-0 32"1+1"6 37'6+2"0 b 83"3+2"5 r U373 MG 59"9+0'6 43"9+4"4 ~ ND d ND ND ND

Melanomas: B16 BL6 71"0+5"2 81'1 i- 3"9 b 72"4+6"2 23"3+8"0 70"9+5"5 c 76"1___4"8 r B16 F1 73"5+6-0 68"8+10'0 85"3+9"0 84"5+3"8 34"3+6"9 c 79'1+12-8

a,b,c Significantly different to control by two-sample t-test: ap<0"05; bp<0"01; Cp<0"001. d ND: not done.

a t t achment of cell l ine G - I J K t . Its s t imula tory effect on F1 a t t achment to type IV collagen did not reach statistical significance. In contrast , l a mi n i n signif icantly increased the a t t achmen t of the BL6 cells to type IV collagen, did no t affect F1 a t t achment bu t inh ib i ted the a t t achmen t of all the g l ioma cell l ines relative to the

type IV collagen control . F r o m viabi l i ty studies on the n o n - a d h e r e n t cells, l a mi n i n was not found to be cytotoxic, the viabil i t ies of these cells not differing signif icantly f rom those of freshly harvested cells (data not shown).

Effects of different laminin treatments on the attachment of human glioma cells to type I V collagen

T h e results of this s tudy to de te rmine the mechan i sms of the inh ib i to ry effect of l amin in on gl ioma cell a t t achment to type IV collagen are shown in table 2. T h e

figures given are means - t - s t andard deviat ions of at least five wells f rom two i n d e p e n d e n t exper iments for the control and ' + l a m i n i n ' as previous ly and of at

Table 2. Effects o f d i f ferent l a m i n i n *~reatments on the a t t a c h m e n t o f h u m a n g l i o m a ce l l s to type IV co l lagen . E x p e r i m e n t a l t r e a t m e n t s are d e t a i l e d in the text.

Mean attachment (A)+ S.D. (per cent)

G- IJKt U373 MG Treatment cell line cell line

Type IV collagen control 66"8 + 5-4 52"9_ 0"6 + Laminin 44-5 + 5-2 b 43"9 _ 4"4" + Laminin 'washed' 49"5 + 3-5 a 33"2 + 5"3 b + Laminin 'late' 62-5 + 0"7 62"4 + 7"4 + Laminin-treated cells 68"5 + 6"5 68-9 + 8"3"

a,b Significant differences with respect to the type IV collagen control by the two-sample t-test: ap<0"05; bp<0"01.

Laminin as an at tachment fac tor 357

least three wells for the other experimental conditions. Removing the laminin- containing assay med ium before the addition of the cells did not prevent the inhibitory effect observed in the original assay and these inhibitory effects were not significantly different f rom one another by the two-sample t-test. Adding laminin after the cells abolished its inhibitory effect completely. Laminin pre- t rea tment of G - I J K t cells did not alter their ability to attach to type IV collagen. In contrast, pre- treating U373 M G cells enhanced their ability to attach (p < 0"05 by the two-sample t-test).

D i s c u s s i o n This study has demonstrated some of the effects of laminin on the a t tachment of

glioma cells using B16 BL6 cells as a 'posit ive ' control and fibronectin as a control a t tachment factor. The general increase in a t tachment to tissue culture plastic in the presence of these factors reflects the findings of a previous study [14], while the increased at tachment of the B16 BL6 cells to type IV collagen in the presence of laminin was as expected [12].

The inhibitory effect of laminin on gl ioma cell a t tachment was specific to this cell type in this study: the assays were per formed at the same t ime as those on the melanoma variants using the same batches of a t tachment factors, med ium and coated plates.

The laminin molecule has separate binding domains for type IV collagen and a cell surface receptor. In the assay described here, pre- incubat ion of the collagen- coated wells with laminin allowed the at tachment factor to adsorb to the collagen before the addition of the cells. T h e K,1 of this reaction is approximately 5 • 10-7 M [10]. The effects of the different laminin t reatments suggest that the inhibitory effect is mediated by laminin adsorption to the collagen rather than by a direct effect on the cells. However, U373 M G cells pre- treated with laminin showed enhanced at tachment to type IV collagen.

I t has been proposed [6] that the cell surface laminin receptor can stimulate haematogenous metastasis by at least two mechanisms: if the receptor is 'unoc- cupied' , it can bind directly to laminin in the basement membrane or, if the receptor is 'occupied' with laminin, this laminin can be used as an a t tachment bridge to type IV collagen. The latter mechanism may be one explanation for the enhanced at tachment of U373 M G cells pre- t reated with laminin. I t is possible that the pre- t reatment allowed laminin to bind to the cell surface and subsequently be utilized to attach to the type IV collagen. T h e fact that the a t tachment of G - I J K t cells was unchanged by this t reatment suggests that laminin was unable to interact with these cells.

The presence of laminin on the surface of the glioma cells may provide an explanation for some of the observations described here. C6 and human glioma cells express laminin at low passage num ber [8] and immunofluorescence studies have indicated laminin product ion by G - I J K t and G - U V W cells (Dr R. I. Freshney, personal communication). While residual cell surface laminin may participate in a t tachment in the absence of exogenous at tachment factor (and may account for the observed variability in control at tachment), a t tachment to type IV collagen in the presence of exogenous laminin may be restricted owing to compet i t ion between the exogenous and endogenous molecules. Squamous carcinoma cells, the adherence of which was similarly inhibited by laminin [14], had been shown to synthesize

358 G. H u n t and G. V. Sherbet

laminin which was found at the cell surface and in the extracellular matrix [1]. However, Fligiel et al. [2] have demonstrated that B16 variants BL6 and F1 secrete similar amounts of laminin into the culture medium and, from immunofluorescence studies, appear to have similar amounts of the molecule at the cell surface. Laminin also has an almost identical effect on the motility of the two cell variants. This study has demonstrated that, whereas exogenous laminin stimulated the attachment of BL6, it had no effect on the F1 variant. Hence it would appear possible that exogenous laminin may have different effects on cells which express cell surface laminin.

The effects of laminin and fibronectin on glioma cell attachment depend on the underlying substratum, laminin and fibronectin stimulating attachment to tissue culture plastic, fibronectin having little effect on attachment to type IV collagen and laminin inhibiting attachment to type IV collagen. These substratum-related effects of laminin have also been noted for the F10 variant of the B16 melanoma, laminin increasing attachment to tissue culture plastic but having no effect on attachment to type IV collagen [3], and also some squamous carcinoma cell lines, laminin increasing attachment to tissue culture plastic but inhibiting attachment to type IV collagen [14], suggesting at least two mechanisms for laminin-mediated cell attachment.

This study has confirmed that the B16 metastatic BL6 variant uses laminin as an attachment factor to type IV collagen. Terranova et al. [12] demonstrated that non- metastatic C3H cells used fibronectin rather than laminin as an attachment factor to type IV collagen. In this investigation, the non-metastatic B16 F1 variant did not utilize laminin as an attachment factor: attachment was stimulated by fibronectin, although this did not reach statistical significance. The inhibitory effect of laminin on the attachment of certain cell types to type IV collagen does not appear to be a definitive characteristic of poorly metastatic cells, the attachment of the highly metastatic M5706 cell line being similarly inhibited by laminin [3].

The true metastatic potential of human gliomas is extremely difficult to demonstrate. Although tumorigenic in athymic 'nude ' mice, there has been no report of metastasis from these subcutaneous xenografts. However, this may not be a good model for this tumor type. It has been demonstrated [4] that the barrier properties of vessels growing into brain tissue transplanted systemically are characteristic of brain capillaries. If it is argued that the blood-brain barrier is still effective within primary brain tumours and is one of the causes of the metastatic inefficiency of this tumour type, systemic transplantation of tumour material may not overcome this problem. Intravenous injection of human and C6 glioma cells resulted in disseminated tumours in nude mice, mainly in the lung but for C6 also in the liver [7]. This experimental model of metastasis circumvents the early events of the process, i.e. invasion of host blood vessels and release of tumour cells into the circulation. Failure to invade cerebral blood vessels may be one of the constraints to 'spontaneous' glioma metastasis. Also, the tumour cells are moving from the lumen of the vessel across the endothelium and underlying basement membrane whereas in 'natural ' metastasis the direction of migration would be from the abluminal side. Another complication of the nude mouse model is that these animals still have natural killer cells which may be involved in surveillance against tumour cells.

In conclusion, laminin stimulated the attachment of glioma cells to tissue culture plastic but inhibited (although for one cell line under special conditions, stimulated) the attachment of these cells to type IV collagen. Which is the more important

Laminin as an at tachment fac tor 359

funct ion of this a t t achment factor in v ivo and whether the observat ions are linked to the metastat ic inefficiency of this t u m o u r type require fur ther investigation.

Acknowledgment T h e authors thank D r R. I. F reshney for provid ing cells and the N o r t h of

England Cancer Research Campa ign for financial support .

References [1] BROWN, K. W., and PARKINSON, E. K., 1985, Alteration of the extracellular matrix of

cultured human keratinocytes by transformation and during differentiation. Intern- ational Journal of Cancer, 35, 799-807.

[2] FLIGIEL, S. E. G., LAYBOURN, K. A., PETERS, B. P., RUDDON, R. W., HISERODT, J. C., and VARANI, J., 1986, Laminin production by murine melanoma cells: possible involvement in cell motility. Clinical and Experimental Metastasis, 4, 259-272.

[3] GIAVAZZI, R., LIOTTA, L., and HART, I., 1982, Laminin inhibits the adhesion ofa murine tumour of macrophage origin. Experimental Cell Research, 140, 315-322.

[4] HIRANO, A., GHATAK, N. R., BECKER, N. H., and ZIMMERMAN, H. M., 1974, A comparison of the fine structure of small blood vessels in intracranial and retroperitoneal malignant lymphomas. Acta Neuropathology, 27, 93-104.

[5] KLEBE, R. J., 1974, Isolation of a collagen-dependent cell attachment factor. Nature (London), 250, 248-251.

[6] LIOTTA, L. A., 1986, Tumour invasion and metastasis--role of the basement membrane. Cancer Research, 46, 1 7.

[7] MANDYBUR, Y. I., LIWNICZ, B. H., and WECHSLER, W., 1984, Disseminated (metastatic) tumours in nude mice produced by intravenous injection of cells of human and non- human neurogenic tumour lines. Acta Neuropathology, 63, 203-209.

[8] McKEEVER, P. E., FLIGIEL, S. E. G., VARANI, J., HUDSON, J. L., SMITH, D., CASTLE, R. L., and McCoY, J. P., 1986, Products of cells cultured from gliomas. IV. Extracellular matrix proteins of gliomas. International Journal of Cancer, 37, 867-874.

[8] PASQUIER, B., PASQUIER, D., GOLET, A. N., PANH, M. H., and COUDERE, P., 1980, Extraneural metastases of astrocytomas and glioblastomas. Cancer, 45, 112-125.

[10] RAO, C. N., MARGULIES, I. M. K., and LIOTTA, L. A., 1984, Molecular regions mediating binding of type IV collagen to laminin. Federation Proceedings, 43, 992 (Abstract).

[11] SHERBET, G. V., 1987, The Metastatic Spread of Cancer: Reflections on its Mechanisms Based on the Malignant Glioma Model (London: Macmillan), p. 80.

[12] TERRANOVA, V. P., LIOTTA, L. A., Russo, R. G., and MARTIN, G. R., 1982, Role of laminin in the attachment and metastasis of murine tumour cells. Cancer Research, 42, 2265-2269.

[13] TIMPL, R., ENGEL, J., and MARTIN, G. R., 1983, Laminin--a multifunctional protein of basement membranes. Trends in Biochemical Sciences, 8, 207 209.

[14] VARANI, J., CAREY, T. E., FLIGIEL, S. E. G., McKEEVER, P. E., and DIXIT, V., 1987, Tumour type-specific differences in cell-substrate adhesion among human tumour cell lines. International Journal of Cancer, 39, 397403.