Int. 1.Cancer: 49,711-716 (1991) 0 1991 Wiley-Liss, Inc.
Publication of the InternationalUnion Against Cancer Publication de I'Union InternationaleContre le Cancer
PENTOXIFYLLINE ENHANCES LUNG COLONIZATION AND ALTERS CELL ADHESION AND GLYCOSAMINOGLYCAN SYNTHESIS BY METASTATIC B16 MELANOMA CELLS Michael EDWARD' and Rona M. MACKIE University of Glasgow, Department of Dermatology, Anderson College Building Glasgow, GI2 8QQ, UK. The effect of pentoxifylline on B16 melanoma cell lung colonization, synthesis and properties of glycosaminoglycans (GAGS), and adhesion to and degradation of subendothelial extracellular matrix was examined. Pentoxifylline inhibited cell growth, cell numbers being reduced by 50% following incubation for 4 days in the presence of 250 yg/ml pentoxifylline, while the treated cells appeared more flattened, possessed numerous but short dendritic processes, and exhibited greatly enhanced tyrosinase activity and melanin synthesis. Pentoxifylline treatment increased the cells' ability to colonize the lungs of syngeneic C57BL mice following tail-vein injection of I O5 cells. The number of lung tumours increased from 16.7 f 6.1 to 52.2 k 17.8. In addition, pentoxifylline-treatedcell GAG synthesis was reduced by 36%, and the charge density of chondroitin sulphate reduced, while tumour-cell aggregation and adhesion to subendothelial extracellular matrix was increased, as was the tumour-cell-mediated releaseof "SO, from radiolabelled subendothelial matrix. The observed changes in GAG synthesis may contribute toward the increased cell adhesiveness which, in addition to increased degradation of certain components of the subendothelial extracellular matrix, may account, at least in part, for the enhancement of lung colonization.
cell types (Raz et al., 1986; Hochman et al., 1984; Netland and Zetter, 1985). Pentoxifylline, an analogue of theobromine, originally used to increase peripheral blood flow, reduces fibroblast GAG, fibronectin and collagen synthesis, and increases collagenase production (Berman and Duncan, 1989). In view of the possible involvement of GAGS in adhesive interactions during the metastatic process, we have examined the effect of pentoxifylline on metastatic melanoma cell lung colonization, GAG synthesis and adhesion to various substrates. MATERIAL AND METHODS
Material Eagle's and Dulbecco's minimal essential medium (EMEM and DMEM), penicillin/streptomycin, trypsin, phosphatebuffered saline (PBS; Dulbecco's A), FCS and Nunclon tissue culture plastics were obtained from Life Technologies, Paisley, UK. Dextran (40 kDa), pentoxifylline, chondroitinase ABC lyase from Proteus vulgaris, and pronase E from Streptomyces grisius were supplied by Sigma (Poole, UK) and DE-52 DEAE-cellulose by Whatman (Maidstone, UK), while D-[114C]glucosamine (50 mCilmmol), D-[6-3H] glucosamine (32 Ci/mmol), NaFSO, (1154 Ci/mmol, [3,5-'H] tyrosine (52 Ciimmol), [2,3,4,5-'H] proline (85 Ci/mmol), and Na, 51Cr0, (403 mCi/mg Cr) were from ICN, High Wycombe, UK. All other reagents were of analytical grade.
Dissemination of tumour cells from a primary tumour to distant body sites with the establishment of a secondary tumour is a complex multi-step process. Substantial evidcnce indicates that a number of cell-surface molecules are involved in adhesive mechanisms which play crucial roles in metastasis (Juliano, 1987). The release of tumour cells into the vasculature as cell aggregates will facilitate non-specific entrapment in Cell culture the capillaries of the target organ (Updyke and Nicolson, B16FlM2 cells were derived from the weakly metastatic 1986); however, there is increasing evidence that the site of B16F1 melanoma cell line (obtained from Dr. I.R. Hart, metastasis is down-stream of non-specific lodgement, suggest- ICRF, London, UK) by injecting 5 times lo4 cells into the tail ing a specific interaction between tumour cells and endothe- veins of syngeneic C57BL mice, and isolating lung colonies 18 lium (Nicolson, 1988). For tumour-cell invasion of the target days later. The lung colonies were then cultured and passaged organ, the tumour cells must first adhere to the subendothelial through the lungs once more to produce a highly metastatic basement membrane, prior to degradation of certain basement cell line. The cells' lung-colonizing ability was stable up to 10 membrane components and migration across the membrane passages, while all experiments were performed between into the underlying matrix (Liotta, 1986). The importance of passages 2 and 6. The B16FlM2 cell line was comparable to these adhesive interactions has been substantiated by the the B16F10 line in that it colonized the lungs with no evidence blocking of tumour-cell lung colonization by pre-incubation of of localization in other organs, but had a lower plating tumour cells with cell binding domains of fibronectin efficiency (55% compared to 74% for B16F10 cells). The cells (Humphries et a/., 1986), and laminin (McCarthy et a/., 1988), were maintained in EMEM supplemented with 10% FCS, prior to intravenous injection into recipient mice. A number of penicillin (100 unitslml), and streptomycin (100 pglml) at extracellular matrix receptors on tumour cells have been 37°C in a humidified atmosphere of 5% CO, in air. Cells were identified including a 67-kDa laminin receptor (Liotta, 1986), incubated in the presence or absence of pentoxifylline (250 whose expression has been correlated with metastatic poten- yg/ml) for 4 days with a medium change on day 3. Calf tial, and several members of the integrin family of receptors pulmonary aortic endothelial cells were cultured in DMEM (Ramos et al., 1990; Ruoslahti and Giancotti, 1989). containing the above supplements. Cell viability, as assessed by Proteoglycans, and in particular heparan sulphate proteogly- Trypan-bluc exclusion, was in excess of 92% for both control cans, are integral components of the plasma membrane (Ra- and pentoxifylline-treated cultures. praeger and Bernfield, 1983), and are therefore ideally located to play an important role in cell-cell and cell-extracellular Assay of Lung-colonizingability matrix interactions, as they are known to interact with several B16FlM2 melanoma cells grown in the presence or absence extracellular matrix components including laminin, fibronectin, of 250 pg/ml of pentoxifylline for 4 days were detached from collagens, and other heparan sulphate proteoglycans. In addi- culture dishes with trypsin/EDTA and harvested in EMEM tion, compositional or structural changes in the GAG sidechains of proteoglycans from tumour cells have been observed 'To whom correspondence and reprint requests should be sent. with increasing metastatic potential (Steck et al., 1987; Timar et al., 1987), while correlations between adhesive properties Received: March 12,1991 and in revised form June 27,1991 and metastatic potential have been suggested in a number of
712
EDWARD AND MACKIE
containing 10% FCS. The cells were washed 3 times with 0.9% NaCI, and 105cells in 0.9% NaCl were injected into the tail veins of syngeneic C57BL mice. All mice were killed 18 days later, the lungs were removed and rinsed in 0.9% NaCI, then the number of lung tumours was determined.
Preparation of subendothelial extracellular matrix Calf pulmonary aortic endothelial cells were cultured in 35-mm Petri dishes as described above, except that 5% dextran was included in the medium, ascorbate (50pg/ml) was added daily, and the cultures were incubated for 7 days after reaching confluence. For radiolabelled subendothelial extracellular matrices, medium was changed 3 days after the cells had reached confluence, and Na, "SO, (20 pCi/ml) was added, and another addition of Na,SO, (20 pCi/ml) was made 48 hr later; or 3H-proline (5 pCi/ml) was added for the final 48 hr of incubation. The cultures were washed with PBS and exposed to 0.5% Triton XlOO in PBS for 15 rnin at 20°C with gentle shaking. Remaining nuclei and cytoskeletons were removed by a 3 min exposure to 0.025 M ammonium hydroxide followed by 4 washes in PBS. Adhesion assay B16F1M2 melanoma cells grown in the presence or absence of pentoxifylline (250 pg/ml) for 4 days, and radiolabelled with Na;'CrO (8 pCi/ml) for the last 16 hr of culture, were detached with trypsin/EDTA and harvested in EMEM containing 10% FCS, washed 3 times with serum-free EMEM, and finally resuspended in the same medium. To each 35-mm Petri dish of either endothelial cell monolayers or subendothelial extracellular matrix was added 1.5 ml of the cell suspension containing 2 times 105 cells, followed by incubation in a stationary mode at 37°C in a humidified atmosphere of 5% CO, in air for various times as indicated in Figure 3. The unattached cells were removed, and the dishes carefully washed with 2 times 2 ml of PBS before the addition of 1.5 ml of 0.1 M NaOH, after which duplicate samples (0.5 ml) were taken from each dish for gamma counting. Cell-cell aggregation was assessed by examining control and pentoxifyllinetreated cells seeded on to 0.5% agar at various time intervals.
rnin before centrifuging (3,000g 5 rnin), after which the pellets were extracted twice cach with 0.5 M perchloric acid and a 3:l v/v mixture of ethanokether, and finally once with ether. After drying of pellets and addition of 1 m10.85 M KOH, tubes were heated to 100°C for 10 min, after which the absorbance was measured at 400 nm.
Isolation and characterization of radiolabelled G A G S Qualitative analysis. GAGS were radiolabelled and isolated essentially as described by Edward and MacKie (1989). Briefly, control B16FlM2 cultures were radiolabelled for 48 hr with 3H-glucosamine (5 pCi/ml), and pentoxifylline-treated cultures with 14C-glucosamine(2 pCi/ml). Media from control cultures were pooled with media from pentoxifylline-treated cultures. Similarly, trypsin-releasable and cell-associated fractions were pooled, boiled and pronase-digested, after which GAGS were isolated by DEAE-cellulose ion-exchange chromatography. Peaks corresponding to sulphated GAGS were pooled and split into 2 equal portions, one half being treated with chondroitinase ABC, and the other half with nitrous acid. All samples were then dialysed and rechromatographed on the ion-exchange columns. Biosynthetic studies. Quantitative analysis of the GAGS was again determined as described by Edward and MacKie (1989). Briefly, control and pentoxifylline-treated cultures were radiolabelled with 'H-glucosamine and Na, 35S04,the cells harvested as described above, and the fractions digested with pronase, after which the sulphated GAGS were isolated by DEAE-cellulose chromatography. Individual GAGS were estimated by gel filtration chromatography on Sephadex G-50 columns following sequential degradation of the void volume material with chondroitinase ABC and nitrous acid. RESULTS
Cultures of B16F1M2 melanoma cells grown in the presence of 250 pg/ml pentoxifylline appeared more flattened than control cells, and possessed numerous but short dendritic processes (Fig. l), while growth was inhibited, cell numbers being reduced by 50% following 4 days' incubation. One of the most striking features of pentoxifylline-treated cultures was Extracellular matrix degradation assay the 22-fold increase in melanin synthesis, while tyrosinase B16F1M2 melanoma cells (3 times 10' cells/35-mm dish) activity was increased 4-fold following treatment with 250 were seeded onto radiolabelled endothelial cell extracellular pg/ml pentoxifylline, compared to control cultures (Fig. 2). matrix and allowed to attach for 2 hr. Unattached cells were The ability of the tumour cells to form pulmonary tumours was removed and monolayers washed twice with serum-free me- examined by injecting control and pentoxifylline-pre-treated dium, then 1ml of serum-free medium containing a range of cells into the tail veins of syngeneic mice and counting lung pentoxifylline concentrations was added. Following incubation tumours 18 days later while, for comparison, cells pre-treated for 48 hr, the medium was removed and centrifuged at 12,000g with lo-% retinoic acid were also injected, as we have for 5 min, then samples were taken for scintillation counting. previously observed retinoic acid-induced inhibition of lung colonization by these cells (Edward and MacKie, 1989). Tyrosinase assay The method is essentially that described by Halaban et al. (1983). Briefly, cell suspensions were centrifuged at 400g for 3 rnin and the cell pellets resuspended in 0.2 ml of PBS containing L-dopa (2 times ~O-,M),L-tyrosine (4 times 1 0 - 4 ~ ) and 2.0 FCi of 'H-tyrosinc. The mixture was incubated for 1 hr at 37"C, after which the reaction was terminated by the addition of 0.5 ml of 10% TCA. The mixture was then centrifuged at 2,OOOg for 5 min, and the supernatant transferred to tubes containing activated charcoal, mixed, and left for 30 min. The mixture was again centrifuged at 2,OOOg for 5 min, and 0.3 ml of the supernatant was taken for scintillation counting.
Melanin assay Melanin content was measured by the colorimetric method of Whittaker (1963). Cell suspensions werc centrifuged (400g, 3 min) and cell pellets resuspended in 0.5 ml distilled water, then subjected to 2 cycles of freeze/thawing. Perchloric acid (0.5 ml, 1 M) was then added and the mixture kept on ice for 10
FIGURE 1-Morphology of B16FlM2 melanoma cells grown in the absence ( a ) or presence (b) of 250 Fg/ml pentoxifylline following 3 days' incubation. Bar = 100 pm.
713
PENTOXIFYLLINE AND MELANOMA LUNG COLONIZATION
50
100
Pentoxifylline
250
500
cone.
1000
lyg/mll 0
FIGURE2 - Dose-response relationship of pentoxifyllineinduced changes in growth inhibition (H), tyrosinase activity (0), and melanin synthesis ( 0 ) .Each value represents the mean of 3 separate cultures ? SD. TABLE I - EFFECT O F PRE-TREATMENT O F B16FIM2 MELANOMA CELLS WITH PENTOXIFYLLINE (250 pgiml) AND RETINOIC ACID (lo-' M) ON LUNG COLONIZATION Cell culture
Lung colonies
Range (n)'
Control Retinoic acid - treated M\ Pehoxifylhne - treated (250 uelml)
16.68 & 6.06' 1.71 2.022
7 - 32 (25) 0 - 6 (17)
*
52.26
k
17.812
25 - 103 (23)
'Mean t SD.-'Significantly different from control; y < 0.001.-'(n) number of mice from 3 separate experiments.
Total
Pentoxifylline pre-treatment of the B16FlM2 cells greatly enhanced their ability to form pulmonary tumours, while retinoic acid pre-treatment, as expected, inhibited tumour formation (Table I). The size of the lung tumours formed varied greatly in mice injected with both control and pentoxifylline-treated cells. It is unlikely that the pentoxifylline pre-treatment of the tumour cells would significantly affect the growth rate of the lung tumours as its effect on the cells in culture is reversible. Tumour-cell attachment to the vascular endothelium and subsequently the subendothelial basement membrane and extracellular matrix is a prerequisite for successful arrest in, and invasion of, the target organ. To examine the possibility that the observed changes in lung colonization following pentoxifylline pre-treatment of the melanoma cells may be due to modulation of adhesive interactions, we studied the effect of pentoxifylline pre-treatment on the adhesion of B16FlM2 cells to endothelial cell monolayers and subendothelial extracellular matrix. Initial adhesion to endothelial cell extracellular matrix was similar in both control and pentoxifyllinetreated cultures, but, following 30 min incubation, the pretreated cells exhibited significantly enhanced adhesion ( p < 0.01), while both cultures adhered less well to monolayers of calf pulmonary aortic endothelial cells, with pentoxifylline-pretreated cells showing no significant difference in adhesion from that of control cultures (Fig. 3). The release of tumour cells as aggregates into the vasculature may enhance their lung-colonizing potential, and therefore tumour cell-cell adhesion may be an important factor in increasing the entrapment of turpour cells in the target organ. Following seeding of B16FlM2 cells onto 0.5% agar, the pentoxifylline-pre-treated cells formed aggregates more readily than did the control cells, although both cell cultures formed large aggregates and, eventually, spheroids, following prolonged incubation (24 hr), while retinoic-acid-pre-treated cells failed to form spheroids (not shown).
20
40
60
80
Incubation time (mi")
FIGURE 3 - Time-dependence of attachment of control (-), and pentoxifylline-pre-treated (-----)B16F1M2 melanoma cells to endothelial cell monolayers (H), and subendothelial extracellular matrix ( 0 ) .Each value represents the mean of duplicate samples taken from 3 separate cultures & SD. TABLE I1 -EFFECT OF PENTOXIFYLLINE ON B16FlM2 MELANOMA CELL MEDIATED RELEASE OF ''SO, AND 'H FROM "SO, AND 'H-PROLINE -
RADIOLABELLED SUBENDOTHELIAL EXTRACELLULAR MATRIX
Pentoxifylline conc. (pglml)
0 25 50 100 250 500 1,000
Radioactivity released (dpm) 330,
2,281 & 275 3,023 f 89 3,274 f 220 3,821 & 121 3.801 & 199 3$38 & 90 4,932 2 308
'H
1,679 -t 253 1,735 k 296 1,670 & 256 1,319 2 262 1.305 t 227 764 2 258 0
T h e ability of tumour cells to invade and metastasize is facilitated by their degradation of certain extracellular matrix components, and a correlation between metastatic potential and type-IV collagenase and heparanase activity has been shown (Turpeenniemi-Hujanen et al., 1985; Nakajima et al., 1983). Here we have examined the ability of B16FlM2 melanoma cells to release ''SO, and 3H from Na, 35S0,-and 3H-proline-radiolabelled endothelial cell extracellular matrix; the majority of the 3 5 S 0 4presumably being incorporated into sulphated GAGS, and in particular, heparan sulphate, and the 'H-proline into type-IV collagen (Kramer and Vogel, 1984). Pentoxifylline pre-treatment of the melanoma cells stimulated an increased release of '5S0, into the medium, but had no significant effect on 'H release, except at high concentrations of pentoxifylline, which inhibited 'H release (Table 11). T h e observed changes in lung colonization and adhesive properties may be due, in part, to changes in the composition or properties of cell-surface GAGS. The effect of pentoxifylline on total G A G synthesis and on the composition of the GAGS expressed by B16FlM2 cells is shown in Figures 4 and 5. Pentoxifylline reduces the incorporation of 3H-glucosamine into GAGS, overall incorporation being reduced by 36%, while 3 s S 0 4 incorporation was reduced by 40%. Analysis of the individual GAGS revealed that the incorporation of 3Hglucosamine was reduced in medium heparans by 46% and chondroitins by 23%, and in trypsin-releasable heparans by 40% and chondroitins by 33%, while cell-associated heparans showed an increased incorporation of 17% with chondroitins reduced by 48% following pentoxifylline treatment. In addition to these compositional changes, we have examined any differences in GAG charge densities using DEAE-cellulose
714
EDWARD AND MACKIE
T
Con Pent
w Medium
Can Pent
v
con Pent
w
Tryprin-releasable Cell-associated
FIGURE4 - Distribution of total GAGS synthesized by control (Con) and pentoxifylline-treated (Pent) B16F1M2 melanoma cells. Each value represents the mean of 3 separate cultures 2 SD. HS
I
cs
-
:I :S
4
E
:
HS
cs
cs
_L_
I
cs
2
(I
L
Medium
Tryprin-releasable
Pent
1 Cell-asrwiated
FIGURE 5 - Distribution of individual sulphated GAGS synthesized by control (Con), and pentoxifylline-treated(Pent) B16F1M2 melanoma cells. Each value represents the mean of 3 separate cultures k SD.
anion-exchange chromatography. Elution profiles (Fig. 6) demonstrate a reduced charge density of pentoxifyllinetreated B16FlM2 chondroitins, while heparans showed a slight but consistent increase in charge density. DISCUSSION
Our study shows that the lung-colonizing potential of metastatic B16FlM2 melanoma cells is significantly enhanced following pre-treatment with pentoxifylline (250 Fg/ml), prior to injection into the tail vein of syngeneic mice, while melanin synthesis is greatly increased. Melanin synthesis has been used as a readily measurable indicator of melanoma cell differentiation, but we have found that differentiation-inducing agents such as retinoic acid have variable effects on a number of melanoma cell lines (Edward et al., 1988). The mechanisms involved in pentoxifylline-induced changes in B16 melanoma cell melanin synthesis, growth, morphology, and lung colonization are poorly understood, but may be partly mediated by increased levels of intracellular CAMP, as methyl-xanthines such as pentoxifylline inhibit CAMP-phosphodiesterase and increase melanin synthesis (Kreider et al., 1975). Moreover, Bennett et al. (1986) reported that PH changes in the presence of melanocyte-stimulating hormone, which also increases c-AMP levels, induces melanin synthesis, reduces growth rate, and enhances lung colonization by B16F10 melanoma cclls.
Pentoxifylline may affect several of the complex steps involved in the establishment of pulmonary tumours. Adhesive mechanisms are undoubtedly involved in several stages of the metastatic process, and here we have observed pentoxifyllineinduced changes in cell-cell aggregation, and adhesion to subendothelial extracellular matrix. Pentoxifylline-pre-treated melanoma cells aggregated more rapidly than control cells when seeded onto an agar substrate, suggesting that they may readily form cell clumps within the vasculature. While cell-cell aggregates may exhibit enhanced non-specific arrest in the target organ, the formation of these clumps may be more important in enhancing the survival of the tumour cells within the vasculature, as large numbers of tumour cells localize in the target organ soon after i.v. injection, but do not necessarily form tumours. We found no increase in adhesion to endothelial cell monolayers following pentoxifylline pre-treatment, but adhesion to aortic endothelial cells may not reflect the tumour cells' ability to adhere to the vascular endothelium of the target organ. Auerbach et al. (1987) have demonstrated that adhesive interactions of malignant cells with organ-derived microvessel, but not large-vessel, endothelial cells correlates with their metastatic properties and organ preference of metastasis. Adhesion to the underlying subendothelial extracellular matrix may also be an important rate-limiting step in the metastatic process. We have shown that retinoic-acid-induced inhibition of B16 melanoma cell lung colonization is accompanied by a decrease in adhesion to basement membrane components (Edward et al., 1989), and now report that the pentoxifylline-induced increase in lung colonization is accompanied by an increase in adhesion to endothelial cell extracellular matrix. Such adhesive mechanisms may be mediated by a number of specific cell-surface receptors and by proteoglycans which may be present as integral components of the plasma membrane. Retinoic-acid-induced changes in lung colonization are accompanied by changes in both the synthesis and properties of GAGS (Edward and MacKie, 1989), and here we have shown an overall reduction in GAG synthesis, and both compositional and charge-density changes in GAGS synthesized by pentoxifylline-treated cells. Just how such changes in GAG composition and properties may affect lung colonization is unclear, but it is unlikely that they mediate adhesion directly, rather augmenting cell-surface receptor-mediated adhesion. However, McCarthy et al. (1988) showed that pre-treatment of both fibrosarcoma and melanoma cells with the heparinbinding domain of fibronectin inhibits pulmonary metastasis, suggesting that cell-surface proteoglycans are important in mediating adhesion of metastatic cells. It is clear that other cell-surface receptors such as integrins play significant roles in tumour invasion and metastasis, as lung colonization by B16F10 melanoma cells may be blocked by co-injecting the tumour cells with RGD-containing peptides (Humphries et aL, 1986). It is possible that pentoxifylline may modulate the expression of integrins, either quantitatively or structurally, thereby affecting receptor-ligand affinity. While adhesion to the capillary endothelium and subendothelial basement membrane of the target organ is an important step in the metastatic process, subsequent tumour-cell invasion of the underlying extracellular matrix is required for successful establishment of a secondary tumour in the tissue stroma. This invasive step is mediated by a variety of degradative enzymes of both tumour-cell and host-cell origin including plasminogen activators, collagenases and heparanases (Tryggvason et al., 1987). Heparan sulphate proteoglycan is an important component of basement membranes, both as a regulator of ion flux due to its highly anionic charge, and in the maintenance of membrane integrity, facilitated by its multiple interactions with other basement membrane components. The role of heparanase in tumour-cell invasion has been dcmonstrated by the inhibition of experimental metastasis by murine
715
PENTOXIFYLLINE AND MELANOMA LUNG COLONIZATION C H O N D R O I T I N SULPHATES
10.5
7.0
-
I
0 I
X
3.5
g
?
-
V
.f
0
20
40
60
0
20
40
60
HEPARAN SULPHATES tryprln
medium
-
24
4.8
16
3.2
1
I I
-1
,
0
N 0 I c
Y
X
En
1.6
?
h
P
*u
m I
c
0
20
40
60
0
20
40
60
Fraction number
FIGURE 6 - Anion-exchange profiles of chondroitin sulphates and heparan sulphates synthesized by control (-), treated (-----)B16FlM2 melanoma cells.
B16 melanoma cells using structural analogues of heparin that block heparanase activity (Irimura et al., 1986). Here w e report increased release of 35S0,from radiolabelled endothelial cell extracellular matrix by pentoxifylline-treated melanoma cells, and this may therefore contribute to enhanced lung colonization. In addition, heparanase may modify the activity of basic fibroblast growth factor (b-FGF), as this growth factor is found in t h e extracellular matrix complexed with heparan sulphate
and pentoxifylline-
(Vlodavsky et al., 1988), and tumours such as melanomas are known to respond t o b-FGF, and t o utilize it as an autocrine growth factor (Halaban et af., 1988). ACKNOWLEDGEMENT
The authors thank the Cancer Research Campaign for financial support.
REFERENCES
AUERBACH, R., Lu, W.C., PARDON, E., GUMKOWSKI, F., KAMINSKA, G. and KAMINSKI, M., Specificity of adhesion between murine tumour cells and capillary endothelium: an in vitro correlate of preferential metastasis in vlvo. CuncerRes., 47,1492-1496 (1987). BENNETT,D.C., DEXTER,T.J., ORMEROD, E.J. and HART, I.R., Increased experimental metastatic capacity of a murine melanoma following induction of differentiation. Cancer Res., 46,3239-3244 (1986). BERMAN, B. and DUNCAN, M.R., Pentoxifylline inhibits normal human dermal fibroblast in v i m proliferation, collagen, glycosarninoglycan, and fibronectin production, and increases collagenase activity.]. invest. Dennatol., 92,605-610 (1989). EDWARD, M., GOLD,J.A. and MACKIE,R.M., Different susceptibilities of melanoma cells to retinoic acid-induced changes in melanotic expression. Biochem. biophys. Res. Comm., 155,773-778 (1988). EDWARD,M., GOLD,J.A. and MACKIE,R.M., Modulation of melanoma cell adhesion to basement membrane components by retinoic acid.J. Cell Sci., 93,155-161 (1989). EDWARD, M. and MACKIE,R.M., Retinoic acid-induced inhibition of lung colonization and changes in the synthesis and properties of glycosaminoglycansof metastatic B16 melanoma cells. J. Cell Sci., 94, 537-543 (1989). HALABAN, R., KWON,B.S., GHOSH,S . , DELLIBOW,P. and BAIRD,A.,
b-FGF as an autocrine growth factor for human melanomas. Oncogene Rex, 3,177-186 (1988). HALABAN, R., POMERANTZ, S.H., MARSHALL, S., LAMBERT, D.T. and LERNER,A.B., Regulation of tyrosinase in human melanocytes grown in culture. J. Cell Biol., 97,480-488 (1983). HOCHMAN, J., LEVY,E., MADOR,N., GOTTESMAN, M.M., SHEARER, G.M. and OKON,E., Cell adhesiveness is related to tumourigenicity in malignant lymphoid cells. J. Cell Biol., 99, 1282-1288 (1984). HUMPHRIES, M.J., OLDEN,K. and YAMADA, K.M., A synthetic peptide from fibronectin inhibits experimental metastasis of murine melanoma cells. Science, 233,467-470 (1986). IRIMURA, T., NAKAJIMA, M. and NICOLSON, G.L., Chemically modified heparins as inhibitors of heparan sulphate specific endo-e-glucoronidase (heparanase) of metastatic melanoma cells. Biochemistty, 25, 5322-5328 (1986). JULIANO, R.L., Membrane receptors for extracellular matrix macromolecules: relationship to cell adhesion and tumour metastasis. Biochim. biophys.Acta, 907,261-278 (1987). UAMER,R.H. and VOGEL,K.G., Selective degradation of basement membrane macromolecules by metastatic melanoma cells. J. nut Cancer Inst., 72,889-899 (1984). KREIDER,J.W., WADE, D.R., ROSENTHAL,M. and DENSLEY,T.,
716
EDWARD AND MACKIE
Maturation and differentiation of B16 melanoma cells induced by theophylline treatment.]. nut. Cancer Inst., 54, 1457-1467 (1975). LIOTTA,L.A., Tumor invasion and metastasis-role of the extracellular matrix. Cancer Rex, 46,l-7 (1986). MCCARTHY,J.B., SKUBITZ,P.N., PALM, S.L. and FURCHT,L.T., Metastasis inhibition of different tumour types by purified laminin fragments and a heparin-binding fragment of fibronectin. J. nut. CuncerInst., 80,108-116 (1988). NAKAJIMA, M., IRIMURA, T., DI FERRANTE, D., DI FERRANTE, N. and NICOLSON, G.L., Heparan sulphate degradation: relation to tumour invasive and metastatic properties of mouse B16 melanoma sublines. Science, 220,611-613 (1983). NETLAND, P.A. and ZETTER, B.R., Metastatic potential of B16 melanoma cells after in vitro selection for organ-specific adherence. J. Cell Bid., 101,720-724 (1985). NICOLSON,G.L., Organ specificity of tumour metastasis: role of preferential adhesion, invasion and growth of malignant cells at specific secondary sites. Cancer Metast. Rev., 7,143-188 (1988). RAMOS,D.M., BERSTON, E.D. and KRAMER, R.H., Analysis of integrin receptors for laminin and type IV collagen on metastatic B16 melanoma cells. CuncerRes., 50,728-734 (1990). RAPRAEGER, A.C. and BERNFIELD, M., Heparan sulphate proteoglycans from mouse mammary epithelial cells. J. hiol. Chern., 258, 3632-3636 (1983). RAZ, A., ZOLLER,M. and BEN-ZE’EV,A,, Cell configuration and adhesive properties of metastasizing and non-metastasizing Bsp73 rat adenocarcinoma cells. Exp. Cell Res., 162,127-141 (1986).
RUOSLAHTI, E. and GIANCOTFI, F.G., Integrins and tumor cell dissemination. Cancer Cells, 1,119-126 (1989). STECK,P.A., CHEONG, P.H., NAKAJIMA, M., YUNG,W.K.A., MOSER, R.P. and NICOLSON, G.L., Altered expression of glycosaminoglycansin metastatic 13762NFrat mammary adenocarcinoma cells. Biochemistry, 26,1020-1028 (1987). TIMAR,J., MOCZAR,E., PAL,K., KOPPER,L., LAPIS,K. and JENEY,A., Comparative study on Lewis lung tumour lines with “low” and “high” metastatic capacity. 11. Cytochemical and biochemical evidence for differences in glycosaminoglycans.Clin. exp. Metust., 5479-87 (1987). TRYGGVASON, K., HOYHTYA, M. and SALO,T., Proteolytic degradation of extracellular matrix in tumour invasion. Biochim. biophys. Actu, 907, 191-217 (1987). TURPEENNIEMI-HUJANEN, T., THORGEIRSSON, U.P., HART,I.R., GRANT, S.S. and LIOTTA, L.A., Expression of collagenase IV (basement membrane collagenase) activity in murine tumour hybrids that differ in metastatic potential. J. nut. Cancer Inst., 75,99-103 (1985). UPDYKE,T.V. and NICOLSON, G.L., Malignant melanoma cell lines selected in vitro for increased homotypic adhesion properties have increased experimental metastatic potential. Clin. exp. Metust., 4, 273-284 (1986). VLODAVSKY, I., MICHAELI, R.I., BER-NER,M., FRIDMAN, R., HOROWITZ, A.T., FUKS,Z . and BIRAN,S., Involvement of heparanase in tumour metastasis and angiogenesis. Israel J. med. Sci., 24, 464-470 (1988). WHITTAKER, J.R., Changes in melanogenesis during the dedifferentiation of chick retinal pigment cells in cell cultures. Devel. B i d , 8, 265-274 (1963).
Publication of the InternationalUnion Against Cancer Publication de I'Union InternationaleContre le Cancer
PENTOXIFYLLINE ENHANCES LUNG COLONIZATION AND ALTERS CELL ADHESION AND GLYCOSAMINOGLYCAN SYNTHESIS BY METASTATIC B16 MELANOMA CELLS Michael EDWARD' and Rona M. MACKIE University of Glasgow, Department of Dermatology, Anderson College Building Glasgow, GI2 8QQ, UK. The effect of pentoxifylline on B16 melanoma cell lung colonization, synthesis and properties of glycosaminoglycans (GAGS), and adhesion to and degradation of subendothelial extracellular matrix was examined. Pentoxifylline inhibited cell growth, cell numbers being reduced by 50% following incubation for 4 days in the presence of 250 yg/ml pentoxifylline, while the treated cells appeared more flattened, possessed numerous but short dendritic processes, and exhibited greatly enhanced tyrosinase activity and melanin synthesis. Pentoxifylline treatment increased the cells' ability to colonize the lungs of syngeneic C57BL mice following tail-vein injection of I O5 cells. The number of lung tumours increased from 16.7 f 6.1 to 52.2 k 17.8. In addition, pentoxifylline-treatedcell GAG synthesis was reduced by 36%, and the charge density of chondroitin sulphate reduced, while tumour-cell aggregation and adhesion to subendothelial extracellular matrix was increased, as was the tumour-cell-mediated releaseof "SO, from radiolabelled subendothelial matrix. The observed changes in GAG synthesis may contribute toward the increased cell adhesiveness which, in addition to increased degradation of certain components of the subendothelial extracellular matrix, may account, at least in part, for the enhancement of lung colonization.
cell types (Raz et al., 1986; Hochman et al., 1984; Netland and Zetter, 1985). Pentoxifylline, an analogue of theobromine, originally used to increase peripheral blood flow, reduces fibroblast GAG, fibronectin and collagen synthesis, and increases collagenase production (Berman and Duncan, 1989). In view of the possible involvement of GAGS in adhesive interactions during the metastatic process, we have examined the effect of pentoxifylline on metastatic melanoma cell lung colonization, GAG synthesis and adhesion to various substrates. MATERIAL AND METHODS
Material Eagle's and Dulbecco's minimal essential medium (EMEM and DMEM), penicillin/streptomycin, trypsin, phosphatebuffered saline (PBS; Dulbecco's A), FCS and Nunclon tissue culture plastics were obtained from Life Technologies, Paisley, UK. Dextran (40 kDa), pentoxifylline, chondroitinase ABC lyase from Proteus vulgaris, and pronase E from Streptomyces grisius were supplied by Sigma (Poole, UK) and DE-52 DEAE-cellulose by Whatman (Maidstone, UK), while D-[114C]glucosamine (50 mCilmmol), D-[6-3H] glucosamine (32 Ci/mmol), NaFSO, (1154 Ci/mmol, [3,5-'H] tyrosine (52 Ciimmol), [2,3,4,5-'H] proline (85 Ci/mmol), and Na, 51Cr0, (403 mCi/mg Cr) were from ICN, High Wycombe, UK. All other reagents were of analytical grade.
Dissemination of tumour cells from a primary tumour to distant body sites with the establishment of a secondary tumour is a complex multi-step process. Substantial evidcnce indicates that a number of cell-surface molecules are involved in adhesive mechanisms which play crucial roles in metastasis (Juliano, 1987). The release of tumour cells into the vasculature as cell aggregates will facilitate non-specific entrapment in Cell culture the capillaries of the target organ (Updyke and Nicolson, B16FlM2 cells were derived from the weakly metastatic 1986); however, there is increasing evidence that the site of B16F1 melanoma cell line (obtained from Dr. I.R. Hart, metastasis is down-stream of non-specific lodgement, suggest- ICRF, London, UK) by injecting 5 times lo4 cells into the tail ing a specific interaction between tumour cells and endothe- veins of syngeneic C57BL mice, and isolating lung colonies 18 lium (Nicolson, 1988). For tumour-cell invasion of the target days later. The lung colonies were then cultured and passaged organ, the tumour cells must first adhere to the subendothelial through the lungs once more to produce a highly metastatic basement membrane, prior to degradation of certain basement cell line. The cells' lung-colonizing ability was stable up to 10 membrane components and migration across the membrane passages, while all experiments were performed between into the underlying matrix (Liotta, 1986). The importance of passages 2 and 6. The B16FlM2 cell line was comparable to these adhesive interactions has been substantiated by the the B16F10 line in that it colonized the lungs with no evidence blocking of tumour-cell lung colonization by pre-incubation of of localization in other organs, but had a lower plating tumour cells with cell binding domains of fibronectin efficiency (55% compared to 74% for B16F10 cells). The cells (Humphries et a/., 1986), and laminin (McCarthy et a/., 1988), were maintained in EMEM supplemented with 10% FCS, prior to intravenous injection into recipient mice. A number of penicillin (100 unitslml), and streptomycin (100 pglml) at extracellular matrix receptors on tumour cells have been 37°C in a humidified atmosphere of 5% CO, in air. Cells were identified including a 67-kDa laminin receptor (Liotta, 1986), incubated in the presence or absence of pentoxifylline (250 whose expression has been correlated with metastatic poten- yg/ml) for 4 days with a medium change on day 3. Calf tial, and several members of the integrin family of receptors pulmonary aortic endothelial cells were cultured in DMEM (Ramos et al., 1990; Ruoslahti and Giancotti, 1989). containing the above supplements. Cell viability, as assessed by Proteoglycans, and in particular heparan sulphate proteogly- Trypan-bluc exclusion, was in excess of 92% for both control cans, are integral components of the plasma membrane (Ra- and pentoxifylline-treated cultures. praeger and Bernfield, 1983), and are therefore ideally located to play an important role in cell-cell and cell-extracellular Assay of Lung-colonizingability matrix interactions, as they are known to interact with several B16FlM2 melanoma cells grown in the presence or absence extracellular matrix components including laminin, fibronectin, of 250 pg/ml of pentoxifylline for 4 days were detached from collagens, and other heparan sulphate proteoglycans. In addi- culture dishes with trypsin/EDTA and harvested in EMEM tion, compositional or structural changes in the GAG sidechains of proteoglycans from tumour cells have been observed 'To whom correspondence and reprint requests should be sent. with increasing metastatic potential (Steck et al., 1987; Timar et al., 1987), while correlations between adhesive properties Received: March 12,1991 and in revised form June 27,1991 and metastatic potential have been suggested in a number of
712
EDWARD AND MACKIE
containing 10% FCS. The cells were washed 3 times with 0.9% NaCI, and 105cells in 0.9% NaCl were injected into the tail veins of syngeneic C57BL mice. All mice were killed 18 days later, the lungs were removed and rinsed in 0.9% NaCI, then the number of lung tumours was determined.
Preparation of subendothelial extracellular matrix Calf pulmonary aortic endothelial cells were cultured in 35-mm Petri dishes as described above, except that 5% dextran was included in the medium, ascorbate (50pg/ml) was added daily, and the cultures were incubated for 7 days after reaching confluence. For radiolabelled subendothelial extracellular matrices, medium was changed 3 days after the cells had reached confluence, and Na, "SO, (20 pCi/ml) was added, and another addition of Na,SO, (20 pCi/ml) was made 48 hr later; or 3H-proline (5 pCi/ml) was added for the final 48 hr of incubation. The cultures were washed with PBS and exposed to 0.5% Triton XlOO in PBS for 15 rnin at 20°C with gentle shaking. Remaining nuclei and cytoskeletons were removed by a 3 min exposure to 0.025 M ammonium hydroxide followed by 4 washes in PBS. Adhesion assay B16F1M2 melanoma cells grown in the presence or absence of pentoxifylline (250 pg/ml) for 4 days, and radiolabelled with Na;'CrO (8 pCi/ml) for the last 16 hr of culture, were detached with trypsin/EDTA and harvested in EMEM containing 10% FCS, washed 3 times with serum-free EMEM, and finally resuspended in the same medium. To each 35-mm Petri dish of either endothelial cell monolayers or subendothelial extracellular matrix was added 1.5 ml of the cell suspension containing 2 times 105 cells, followed by incubation in a stationary mode at 37°C in a humidified atmosphere of 5% CO, in air for various times as indicated in Figure 3. The unattached cells were removed, and the dishes carefully washed with 2 times 2 ml of PBS before the addition of 1.5 ml of 0.1 M NaOH, after which duplicate samples (0.5 ml) were taken from each dish for gamma counting. Cell-cell aggregation was assessed by examining control and pentoxifyllinetreated cells seeded on to 0.5% agar at various time intervals.
rnin before centrifuging (3,000g 5 rnin), after which the pellets were extracted twice cach with 0.5 M perchloric acid and a 3:l v/v mixture of ethanokether, and finally once with ether. After drying of pellets and addition of 1 m10.85 M KOH, tubes were heated to 100°C for 10 min, after which the absorbance was measured at 400 nm.
Isolation and characterization of radiolabelled G A G S Qualitative analysis. GAGS were radiolabelled and isolated essentially as described by Edward and MacKie (1989). Briefly, control B16FlM2 cultures were radiolabelled for 48 hr with 3H-glucosamine (5 pCi/ml), and pentoxifylline-treated cultures with 14C-glucosamine(2 pCi/ml). Media from control cultures were pooled with media from pentoxifylline-treated cultures. Similarly, trypsin-releasable and cell-associated fractions were pooled, boiled and pronase-digested, after which GAGS were isolated by DEAE-cellulose ion-exchange chromatography. Peaks corresponding to sulphated GAGS were pooled and split into 2 equal portions, one half being treated with chondroitinase ABC, and the other half with nitrous acid. All samples were then dialysed and rechromatographed on the ion-exchange columns. Biosynthetic studies. Quantitative analysis of the GAGS was again determined as described by Edward and MacKie (1989). Briefly, control and pentoxifylline-treated cultures were radiolabelled with 'H-glucosamine and Na, 35S04,the cells harvested as described above, and the fractions digested with pronase, after which the sulphated GAGS were isolated by DEAE-cellulose chromatography. Individual GAGS were estimated by gel filtration chromatography on Sephadex G-50 columns following sequential degradation of the void volume material with chondroitinase ABC and nitrous acid. RESULTS
Cultures of B16F1M2 melanoma cells grown in the presence of 250 pg/ml pentoxifylline appeared more flattened than control cells, and possessed numerous but short dendritic processes (Fig. l), while growth was inhibited, cell numbers being reduced by 50% following 4 days' incubation. One of the most striking features of pentoxifylline-treated cultures was Extracellular matrix degradation assay the 22-fold increase in melanin synthesis, while tyrosinase B16F1M2 melanoma cells (3 times 10' cells/35-mm dish) activity was increased 4-fold following treatment with 250 were seeded onto radiolabelled endothelial cell extracellular pg/ml pentoxifylline, compared to control cultures (Fig. 2). matrix and allowed to attach for 2 hr. Unattached cells were The ability of the tumour cells to form pulmonary tumours was removed and monolayers washed twice with serum-free me- examined by injecting control and pentoxifylline-pre-treated dium, then 1ml of serum-free medium containing a range of cells into the tail veins of syngeneic mice and counting lung pentoxifylline concentrations was added. Following incubation tumours 18 days later while, for comparison, cells pre-treated for 48 hr, the medium was removed and centrifuged at 12,000g with lo-% retinoic acid were also injected, as we have for 5 min, then samples were taken for scintillation counting. previously observed retinoic acid-induced inhibition of lung colonization by these cells (Edward and MacKie, 1989). Tyrosinase assay The method is essentially that described by Halaban et al. (1983). Briefly, cell suspensions were centrifuged at 400g for 3 rnin and the cell pellets resuspended in 0.2 ml of PBS containing L-dopa (2 times ~O-,M),L-tyrosine (4 times 1 0 - 4 ~ ) and 2.0 FCi of 'H-tyrosinc. The mixture was incubated for 1 hr at 37"C, after which the reaction was terminated by the addition of 0.5 ml of 10% TCA. The mixture was then centrifuged at 2,OOOg for 5 min, and the supernatant transferred to tubes containing activated charcoal, mixed, and left for 30 min. The mixture was again centrifuged at 2,OOOg for 5 min, and 0.3 ml of the supernatant was taken for scintillation counting.
Melanin assay Melanin content was measured by the colorimetric method of Whittaker (1963). Cell suspensions werc centrifuged (400g, 3 min) and cell pellets resuspended in 0.5 ml distilled water, then subjected to 2 cycles of freeze/thawing. Perchloric acid (0.5 ml, 1 M) was then added and the mixture kept on ice for 10
FIGURE 1-Morphology of B16FlM2 melanoma cells grown in the absence ( a ) or presence (b) of 250 Fg/ml pentoxifylline following 3 days' incubation. Bar = 100 pm.
713
PENTOXIFYLLINE AND MELANOMA LUNG COLONIZATION
50
100
Pentoxifylline
250
500
cone.
1000
lyg/mll 0
FIGURE2 - Dose-response relationship of pentoxifyllineinduced changes in growth inhibition (H), tyrosinase activity (0), and melanin synthesis ( 0 ) .Each value represents the mean of 3 separate cultures ? SD. TABLE I - EFFECT O F PRE-TREATMENT O F B16FIM2 MELANOMA CELLS WITH PENTOXIFYLLINE (250 pgiml) AND RETINOIC ACID (lo-' M) ON LUNG COLONIZATION Cell culture
Lung colonies
Range (n)'
Control Retinoic acid - treated M\ Pehoxifylhne - treated (250 uelml)
16.68 & 6.06' 1.71 2.022
7 - 32 (25) 0 - 6 (17)
*
52.26
k
17.812
25 - 103 (23)
'Mean t SD.-'Significantly different from control; y < 0.001.-'(n) number of mice from 3 separate experiments.
Total
Pentoxifylline pre-treatment of the B16FlM2 cells greatly enhanced their ability to form pulmonary tumours, while retinoic acid pre-treatment, as expected, inhibited tumour formation (Table I). The size of the lung tumours formed varied greatly in mice injected with both control and pentoxifylline-treated cells. It is unlikely that the pentoxifylline pre-treatment of the tumour cells would significantly affect the growth rate of the lung tumours as its effect on the cells in culture is reversible. Tumour-cell attachment to the vascular endothelium and subsequently the subendothelial basement membrane and extracellular matrix is a prerequisite for successful arrest in, and invasion of, the target organ. To examine the possibility that the observed changes in lung colonization following pentoxifylline pre-treatment of the melanoma cells may be due to modulation of adhesive interactions, we studied the effect of pentoxifylline pre-treatment on the adhesion of B16FlM2 cells to endothelial cell monolayers and subendothelial extracellular matrix. Initial adhesion to endothelial cell extracellular matrix was similar in both control and pentoxifyllinetreated cultures, but, following 30 min incubation, the pretreated cells exhibited significantly enhanced adhesion ( p < 0.01), while both cultures adhered less well to monolayers of calf pulmonary aortic endothelial cells, with pentoxifylline-pretreated cells showing no significant difference in adhesion from that of control cultures (Fig. 3). The release of tumour cells as aggregates into the vasculature may enhance their lung-colonizing potential, and therefore tumour cell-cell adhesion may be an important factor in increasing the entrapment of turpour cells in the target organ. Following seeding of B16FlM2 cells onto 0.5% agar, the pentoxifylline-pre-treated cells formed aggregates more readily than did the control cells, although both cell cultures formed large aggregates and, eventually, spheroids, following prolonged incubation (24 hr), while retinoic-acid-pre-treated cells failed to form spheroids (not shown).
20
40
60
80
Incubation time (mi")
FIGURE 3 - Time-dependence of attachment of control (-), and pentoxifylline-pre-treated (-----)B16F1M2 melanoma cells to endothelial cell monolayers (H), and subendothelial extracellular matrix ( 0 ) .Each value represents the mean of duplicate samples taken from 3 separate cultures & SD. TABLE I1 -EFFECT OF PENTOXIFYLLINE ON B16FlM2 MELANOMA CELL MEDIATED RELEASE OF ''SO, AND 'H FROM "SO, AND 'H-PROLINE -
RADIOLABELLED SUBENDOTHELIAL EXTRACELLULAR MATRIX
Pentoxifylline conc. (pglml)
0 25 50 100 250 500 1,000
Radioactivity released (dpm) 330,
2,281 & 275 3,023 f 89 3,274 f 220 3,821 & 121 3.801 & 199 3$38 & 90 4,932 2 308
'H
1,679 -t 253 1,735 k 296 1,670 & 256 1,319 2 262 1.305 t 227 764 2 258 0
T h e ability of tumour cells to invade and metastasize is facilitated by their degradation of certain extracellular matrix components, and a correlation between metastatic potential and type-IV collagenase and heparanase activity has been shown (Turpeenniemi-Hujanen et al., 1985; Nakajima et al., 1983). Here we have examined the ability of B16FlM2 melanoma cells to release ''SO, and 3H from Na, 35S0,-and 3H-proline-radiolabelled endothelial cell extracellular matrix; the majority of the 3 5 S 0 4presumably being incorporated into sulphated GAGS, and in particular, heparan sulphate, and the 'H-proline into type-IV collagen (Kramer and Vogel, 1984). Pentoxifylline pre-treatment of the melanoma cells stimulated an increased release of '5S0, into the medium, but had no significant effect on 'H release, except at high concentrations of pentoxifylline, which inhibited 'H release (Table 11). T h e observed changes in lung colonization and adhesive properties may be due, in part, to changes in the composition or properties of cell-surface GAGS. The effect of pentoxifylline on total G A G synthesis and on the composition of the GAGS expressed by B16FlM2 cells is shown in Figures 4 and 5. Pentoxifylline reduces the incorporation of 3H-glucosamine into GAGS, overall incorporation being reduced by 36%, while 3 s S 0 4 incorporation was reduced by 40%. Analysis of the individual GAGS revealed that the incorporation of 3Hglucosamine was reduced in medium heparans by 46% and chondroitins by 23%, and in trypsin-releasable heparans by 40% and chondroitins by 33%, while cell-associated heparans showed an increased incorporation of 17% with chondroitins reduced by 48% following pentoxifylline treatment. In addition to these compositional changes, we have examined any differences in GAG charge densities using DEAE-cellulose
714
EDWARD AND MACKIE
T
Con Pent
w Medium
Can Pent
v
con Pent
w
Tryprin-releasable Cell-associated
FIGURE4 - Distribution of total GAGS synthesized by control (Con) and pentoxifylline-treated (Pent) B16F1M2 melanoma cells. Each value represents the mean of 3 separate cultures 2 SD. HS
I
cs
-
:I :S
4
E
:
HS
cs
cs
_L_
I
cs
2
(I
L
Medium
Tryprin-releasable
Pent
1 Cell-asrwiated
FIGURE 5 - Distribution of individual sulphated GAGS synthesized by control (Con), and pentoxifylline-treated(Pent) B16F1M2 melanoma cells. Each value represents the mean of 3 separate cultures k SD.
anion-exchange chromatography. Elution profiles (Fig. 6) demonstrate a reduced charge density of pentoxifyllinetreated B16FlM2 chondroitins, while heparans showed a slight but consistent increase in charge density. DISCUSSION
Our study shows that the lung-colonizing potential of metastatic B16FlM2 melanoma cells is significantly enhanced following pre-treatment with pentoxifylline (250 Fg/ml), prior to injection into the tail vein of syngeneic mice, while melanin synthesis is greatly increased. Melanin synthesis has been used as a readily measurable indicator of melanoma cell differentiation, but we have found that differentiation-inducing agents such as retinoic acid have variable effects on a number of melanoma cell lines (Edward et al., 1988). The mechanisms involved in pentoxifylline-induced changes in B16 melanoma cell melanin synthesis, growth, morphology, and lung colonization are poorly understood, but may be partly mediated by increased levels of intracellular CAMP, as methyl-xanthines such as pentoxifylline inhibit CAMP-phosphodiesterase and increase melanin synthesis (Kreider et al., 1975). Moreover, Bennett et al. (1986) reported that PH changes in the presence of melanocyte-stimulating hormone, which also increases c-AMP levels, induces melanin synthesis, reduces growth rate, and enhances lung colonization by B16F10 melanoma cclls.
Pentoxifylline may affect several of the complex steps involved in the establishment of pulmonary tumours. Adhesive mechanisms are undoubtedly involved in several stages of the metastatic process, and here we have observed pentoxifyllineinduced changes in cell-cell aggregation, and adhesion to subendothelial extracellular matrix. Pentoxifylline-pre-treated melanoma cells aggregated more rapidly than control cells when seeded onto an agar substrate, suggesting that they may readily form cell clumps within the vasculature. While cell-cell aggregates may exhibit enhanced non-specific arrest in the target organ, the formation of these clumps may be more important in enhancing the survival of the tumour cells within the vasculature, as large numbers of tumour cells localize in the target organ soon after i.v. injection, but do not necessarily form tumours. We found no increase in adhesion to endothelial cell monolayers following pentoxifylline pre-treatment, but adhesion to aortic endothelial cells may not reflect the tumour cells' ability to adhere to the vascular endothelium of the target organ. Auerbach et al. (1987) have demonstrated that adhesive interactions of malignant cells with organ-derived microvessel, but not large-vessel, endothelial cells correlates with their metastatic properties and organ preference of metastasis. Adhesion to the underlying subendothelial extracellular matrix may also be an important rate-limiting step in the metastatic process. We have shown that retinoic-acid-induced inhibition of B16 melanoma cell lung colonization is accompanied by a decrease in adhesion to basement membrane components (Edward et al., 1989), and now report that the pentoxifylline-induced increase in lung colonization is accompanied by an increase in adhesion to endothelial cell extracellular matrix. Such adhesive mechanisms may be mediated by a number of specific cell-surface receptors and by proteoglycans which may be present as integral components of the plasma membrane. Retinoic-acid-induced changes in lung colonization are accompanied by changes in both the synthesis and properties of GAGS (Edward and MacKie, 1989), and here we have shown an overall reduction in GAG synthesis, and both compositional and charge-density changes in GAGS synthesized by pentoxifylline-treated cells. Just how such changes in GAG composition and properties may affect lung colonization is unclear, but it is unlikely that they mediate adhesion directly, rather augmenting cell-surface receptor-mediated adhesion. However, McCarthy et al. (1988) showed that pre-treatment of both fibrosarcoma and melanoma cells with the heparinbinding domain of fibronectin inhibits pulmonary metastasis, suggesting that cell-surface proteoglycans are important in mediating adhesion of metastatic cells. It is clear that other cell-surface receptors such as integrins play significant roles in tumour invasion and metastasis, as lung colonization by B16F10 melanoma cells may be blocked by co-injecting the tumour cells with RGD-containing peptides (Humphries et aL, 1986). It is possible that pentoxifylline may modulate the expression of integrins, either quantitatively or structurally, thereby affecting receptor-ligand affinity. While adhesion to the capillary endothelium and subendothelial basement membrane of the target organ is an important step in the metastatic process, subsequent tumour-cell invasion of the underlying extracellular matrix is required for successful establishment of a secondary tumour in the tissue stroma. This invasive step is mediated by a variety of degradative enzymes of both tumour-cell and host-cell origin including plasminogen activators, collagenases and heparanases (Tryggvason et al., 1987). Heparan sulphate proteoglycan is an important component of basement membranes, both as a regulator of ion flux due to its highly anionic charge, and in the maintenance of membrane integrity, facilitated by its multiple interactions with other basement membrane components. The role of heparanase in tumour-cell invasion has been dcmonstrated by the inhibition of experimental metastasis by murine
715
PENTOXIFYLLINE AND MELANOMA LUNG COLONIZATION C H O N D R O I T I N SULPHATES
10.5
7.0
-
I
0 I
X
3.5
g
?
-
V
.f
0
20
40
60
0
20
40
60
HEPARAN SULPHATES tryprln
medium
-
24
4.8
16
3.2
1
I I
-1
,
0
N 0 I c
Y
X
En
1.6
?
h
P
*u
m I
c
0
20
40
60
0
20
40
60
Fraction number
FIGURE 6 - Anion-exchange profiles of chondroitin sulphates and heparan sulphates synthesized by control (-), treated (-----)B16FlM2 melanoma cells.
B16 melanoma cells using structural analogues of heparin that block heparanase activity (Irimura et al., 1986). Here w e report increased release of 35S0,from radiolabelled endothelial cell extracellular matrix by pentoxifylline-treated melanoma cells, and this may therefore contribute to enhanced lung colonization. In addition, heparanase may modify the activity of basic fibroblast growth factor (b-FGF), as this growth factor is found in t h e extracellular matrix complexed with heparan sulphate
and pentoxifylline-
(Vlodavsky et al., 1988), and tumours such as melanomas are known to respond t o b-FGF, and t o utilize it as an autocrine growth factor (Halaban et af., 1988). ACKNOWLEDGEMENT
The authors thank the Cancer Research Campaign for financial support.
REFERENCES
AUERBACH, R., Lu, W.C., PARDON, E., GUMKOWSKI, F., KAMINSKA, G. and KAMINSKI, M., Specificity of adhesion between murine tumour cells and capillary endothelium: an in vitro correlate of preferential metastasis in vlvo. CuncerRes., 47,1492-1496 (1987). BENNETT,D.C., DEXTER,T.J., ORMEROD, E.J. and HART, I.R., Increased experimental metastatic capacity of a murine melanoma following induction of differentiation. Cancer Res., 46,3239-3244 (1986). BERMAN, B. and DUNCAN, M.R., Pentoxifylline inhibits normal human dermal fibroblast in v i m proliferation, collagen, glycosarninoglycan, and fibronectin production, and increases collagenase activity.]. invest. Dennatol., 92,605-610 (1989). EDWARD, M., GOLD,J.A. and MACKIE,R.M., Different susceptibilities of melanoma cells to retinoic acid-induced changes in melanotic expression. Biochem. biophys. Res. Comm., 155,773-778 (1988). EDWARD,M., GOLD,J.A. and MACKIE,R.M., Modulation of melanoma cell adhesion to basement membrane components by retinoic acid.J. Cell Sci., 93,155-161 (1989). EDWARD, M. and MACKIE,R.M., Retinoic acid-induced inhibition of lung colonization and changes in the synthesis and properties of glycosaminoglycansof metastatic B16 melanoma cells. J. Cell Sci., 94, 537-543 (1989). HALABAN, R., KWON,B.S., GHOSH,S . , DELLIBOW,P. and BAIRD,A.,
b-FGF as an autocrine growth factor for human melanomas. Oncogene Rex, 3,177-186 (1988). HALABAN, R., POMERANTZ, S.H., MARSHALL, S., LAMBERT, D.T. and LERNER,A.B., Regulation of tyrosinase in human melanocytes grown in culture. J. Cell Biol., 97,480-488 (1983). HOCHMAN, J., LEVY,E., MADOR,N., GOTTESMAN, M.M., SHEARER, G.M. and OKON,E., Cell adhesiveness is related to tumourigenicity in malignant lymphoid cells. J. Cell Biol., 99, 1282-1288 (1984). HUMPHRIES, M.J., OLDEN,K. and YAMADA, K.M., A synthetic peptide from fibronectin inhibits experimental metastasis of murine melanoma cells. Science, 233,467-470 (1986). IRIMURA, T., NAKAJIMA, M. and NICOLSON, G.L., Chemically modified heparins as inhibitors of heparan sulphate specific endo-e-glucoronidase (heparanase) of metastatic melanoma cells. Biochemistty, 25, 5322-5328 (1986). JULIANO, R.L., Membrane receptors for extracellular matrix macromolecules: relationship to cell adhesion and tumour metastasis. Biochim. biophys.Acta, 907,261-278 (1987). UAMER,R.H. and VOGEL,K.G., Selective degradation of basement membrane macromolecules by metastatic melanoma cells. J. nut Cancer Inst., 72,889-899 (1984). KREIDER,J.W., WADE, D.R., ROSENTHAL,M. and DENSLEY,T.,
716
EDWARD AND MACKIE
Maturation and differentiation of B16 melanoma cells induced by theophylline treatment.]. nut. Cancer Inst., 54, 1457-1467 (1975). LIOTTA,L.A., Tumor invasion and metastasis-role of the extracellular matrix. Cancer Rex, 46,l-7 (1986). MCCARTHY,J.B., SKUBITZ,P.N., PALM, S.L. and FURCHT,L.T., Metastasis inhibition of different tumour types by purified laminin fragments and a heparin-binding fragment of fibronectin. J. nut. CuncerInst., 80,108-116 (1988). NAKAJIMA, M., IRIMURA, T., DI FERRANTE, D., DI FERRANTE, N. and NICOLSON, G.L., Heparan sulphate degradation: relation to tumour invasive and metastatic properties of mouse B16 melanoma sublines. Science, 220,611-613 (1983). NETLAND, P.A. and ZETTER, B.R., Metastatic potential of B16 melanoma cells after in vitro selection for organ-specific adherence. J. Cell Bid., 101,720-724 (1985). NICOLSON,G.L., Organ specificity of tumour metastasis: role of preferential adhesion, invasion and growth of malignant cells at specific secondary sites. Cancer Metast. Rev., 7,143-188 (1988). RAMOS,D.M., BERSTON, E.D. and KRAMER, R.H., Analysis of integrin receptors for laminin and type IV collagen on metastatic B16 melanoma cells. CuncerRes., 50,728-734 (1990). RAPRAEGER, A.C. and BERNFIELD, M., Heparan sulphate proteoglycans from mouse mammary epithelial cells. J. hiol. Chern., 258, 3632-3636 (1983). RAZ, A., ZOLLER,M. and BEN-ZE’EV,A,, Cell configuration and adhesive properties of metastasizing and non-metastasizing Bsp73 rat adenocarcinoma cells. Exp. Cell Res., 162,127-141 (1986).
RUOSLAHTI, E. and GIANCOTFI, F.G., Integrins and tumor cell dissemination. Cancer Cells, 1,119-126 (1989). STECK,P.A., CHEONG, P.H., NAKAJIMA, M., YUNG,W.K.A., MOSER, R.P. and NICOLSON, G.L., Altered expression of glycosaminoglycansin metastatic 13762NFrat mammary adenocarcinoma cells. Biochemistry, 26,1020-1028 (1987). TIMAR,J., MOCZAR,E., PAL,K., KOPPER,L., LAPIS,K. and JENEY,A., Comparative study on Lewis lung tumour lines with “low” and “high” metastatic capacity. 11. Cytochemical and biochemical evidence for differences in glycosaminoglycans.Clin. exp. Metust., 5479-87 (1987). TRYGGVASON, K., HOYHTYA, M. and SALO,T., Proteolytic degradation of extracellular matrix in tumour invasion. Biochim. biophys. Actu, 907, 191-217 (1987). TURPEENNIEMI-HUJANEN, T., THORGEIRSSON, U.P., HART,I.R., GRANT, S.S. and LIOTTA, L.A., Expression of collagenase IV (basement membrane collagenase) activity in murine tumour hybrids that differ in metastatic potential. J. nut. Cancer Inst., 75,99-103 (1985). UPDYKE,T.V. and NICOLSON, G.L., Malignant melanoma cell lines selected in vitro for increased homotypic adhesion properties have increased experimental metastatic potential. Clin. exp. Metust., 4, 273-284 (1986). VLODAVSKY, I., MICHAELI, R.I., BER-NER,M., FRIDMAN, R., HOROWITZ, A.T., FUKS,Z . and BIRAN,S., Involvement of heparanase in tumour metastasis and angiogenesis. Israel J. med. Sci., 24, 464-470 (1988). WHITTAKER, J.R., Changes in melanogenesis during the dedifferentiation of chick retinal pigment cells in cell cultures. Devel. B i d , 8, 265-274 (1963).
