Field of the Invention

This invention relates to growth factor antagonists, compositions comprising said antagonists, and a method of making compositions suitable for treatment of certain diseases.

Background of the Invention

A large number of compounds are known which may be described as sulphated polysaccharides. Sulphated polysaccharides exhibit extreme diversity in their characteristics, depending on their molecular weight and degree of sulphation (D.S). They include natural substances such as heparin and artificial compounds such as dextran sulphate. Also classified as sulphated polysaccharides are natural compounds obtainable from seaweeds, such as the carrageenans and laminar ^' , ns.

The carrageenans are a family of sulphated polysaccharides which may be isolated from red algae. The backbone structure of most carrageenans consists substantially of repeating galactose and 3, 6 anhydrogalactose residues linked beta-1,4 and alpha-1,3 respectively.

Several sulphated polysaccharides have been shown to

possess anti-angiogenic activity. This is significant because angiogenesis (the formation of new blood capillaries, or "neovascularisation" ) appears to play a key role in the growth and development of solid tumours. Thus substances which inhibit angiogenesis might be able to inhibit the growth of certain solid tumours. Thus, for example, Tanaka et al. have shown that a bacterial substance, which they have termed sulphated polysaccharide peptidoglycan complex (SP-PG), inhibits angiogenesis and has antitumour effects (Cancer Research 4_9_, 6727-6730 and Int. J. Radiat. Biol 6 , 79-83.

The same researchers, in collaboration with other groups, described SP-PG as having effects on angiogenesis and lesions caused by Kaposi's Sarcoma. However, other sulphated compounds (suramin, pentosan polysulphate) were found to be far less active.

urata et al. (Cancer Research jiJL, 22-26) disclosed that some chitin derivatives were able to inhibit tumour- induced angiogenesis. Confusingly however, the same authors found that heparin had no such activity.

It is believed that one mechanism by which the sulphated polysaccharides may exert anti-angiogenic activity is by virtue of their antagonistic effects on a number of growth factors. There is one report of the ability of sulphated polysaccharides to inhibit a growth factor. Thus ell.εtein et al . (1991, Journal of the National Cancer Institute 8_3, 716-720) described the inhibition of Kaposi's Sarcoma-derived FGF by pentosan polysulphate.

WO 88/05301 discloses anti-metastatic properties for a number of sulphated polysaccharides, including a type of

carrageenans called carrageenan lambda. Thus it was suggested that carrageenan lambda inhibited the spread (but not the growth) of tumours. The mechanism shown to be responsible for this property was inhibition of endoglycosidases (particularly heparanase), enzymes which facilitate the spread of tumour cells. This is far removed from the present invention which, inter alia, concerns the discovery of carrageenans which are extremely effective antagonists of growth factors, including the group of growth factors termed Fibroblast Growth Factor (FGF) and Transforming Growth Factor-beta (TGF-beta) and thereby exert anti-tumour effects in a very different manner to the prior art.

Summary of the Invention

In one aspect the invention provides a compound being a carrageenan or derivative thereof for use as a growth factor suppressor or inhibitor.

The finding that carrageenans possess such activity was published (by one of the present inventors) after the priority date of the present application (Hoffman, 1993 Biochem. J. 289, 331-334).

Preferably ne growth factor suppressed or inhibited is one which binds heparin. However, the present inventors have found that carrageenans also inhibit growth factors by means other than or in addition to direct competition for binding, thus the present invention extends beyond use of carrageenans or their derivatives to inhibit growth factors to which they bind.

For example, carrageenans have been found by the present

inventors not to inhibit binding of TGF-alpha to target cells but nonetheless to inhibit TGF-alpha-mediated DNA synthesis.

Generally the growth factor suppressed or inhibited by the carrageenan or its derivative will be basic Fibroblast Growth Factor (bFGF), Platelet-derived Growth Factor (PDGF) or Transforming Growth Factor-Beta (TGF-beta). However it is believed that a number of other growth factors may be inhibited by the compounds of the invention, especially those which bind to heparin.

Generally, the carrageenans of the present invention will have very large molecular weights, typically over 100,000 Daltons. However it is probable that lower molecular weight derivatives are obtainable which also possess desirable anti-tumour activities. For example, one might expect certain derivatives of the carrageenans (such as oligosaccharides) to have similar properties. This has been shown for low molecular weight derivatives of heparin (Ishihara et al . , 1992, Analytical Biochemistry 202, 310- 315).

It will be readily apparent to those skilled in the art that derivatives of carrageenans may be prepared in a number of ways. For example, lower molecular weight derivatives may be prepared by chemical or enzymatic hydrolysis using mild acid or carrageenase (respectively) to produce a range of compounds. Such methods will also produce substances with a range of degree of sulphation. By a simple process of trial and error, various product fractions can be screened for activity. Once the most potent fractions have been identified, these derivatives will probably be preferred for therapeutic use (rather

than the intact carrageenan) as they will be more effective per unit weight of therapeutic substance. Another method of producing lower molecular weights derivatives is by freeze-dryin , described below.

Conveniently the carrageenans of the present invention are those obtainable commercially, such as the types known as iota, lambda and kappa.

It will be readily apparent to those skilled in the art that compounds having the properties of the present invention may well find therapeutic applications.

Growth factors are implicated in the growth of tumours. A carrageenan having a growth factor suppressing or inhibiting activity might possess anti-tumour activity. The present inventors have found this to be the case. Indeed, the carrageenans of the present invention are thought to exert anti-tumour activity by at least two distinct mechanisms (described below). It is a highly preferred feature that the carrageenan or derivative thereof exhibits anti-tumour activity, preferably anti- tumour growth activity.

Diseases involving aberrant or undesirable production of growth factors, which might be susceptible to treatment with carrageenans or derivatives, include the following: psoriasis (which, as demonstrated by Elder et al . , [1989, Science 243, 811-814], is associated with the over¬ production of TGF-alpha), pyogenic granulo a, alopecia (Epidermal Growth Factor [EGF] and TGF-alpha have both been shown to inhibit hair follicle elongation [Philpott et al . , 1992, British Journal of Dermatology 127, 600- 607], and inhibition of inappropriate keratinocyte

proliferation (which is stimulated by TGF-alpha).

It will be apparent that such dermatological conditions are highly preferred for treatment by carrageenans, as there are no complications arising from the anti-coagulant activity of some of the carrageenans.

Thus in a second aspect the invention provides a composition for use in the treatment of tumours and/or of diseases associated with aberrant production of growth factors and/or diseases involving inappropriate neovascularisation, the composition comprising a carrageenan or derivative thereof.

It is to be expected that such compositions will be particularly effective in the treatment of solid tumours such as brain tumours (gliomas), and prostate gland cancers. Other angiogenesis-dependent diseases which may be susceptible to treatment with compositions according to the invention are in the areas of: ophthalmology (diabetic retinopathy, corneal graft neovascularisation, neovascular glaucoma, trachoma and retrolental fibroplasia) ; and cardiology (atherosclerotic plaques), paediatrics (hemangioma, angiofibroma, haemophilic joints). Other conditions which might be usefully treated include: Osier-Weber Syndrome, Nonunion fractures, arteriovenous malformations, arthritis, scleroderma, vascular adhesions, hypertrophic scars and delayed wound healing.

Thus in a further aspect the invention provides a method of treating a human or animal body comprising administering an effective amount of a composition in accordance with the invention.

- 1 -

The formulation of the composition, and the route of delivery, will depend on the disease or condition to be treated. For example, in treating solid, internal tumours, one could inject or infuse a sterile, isotonic, aqueous solution of a carrageenan salt or derivative in a neutral pH buffer (such as phosphate-buffered saline).

As an alternative, the composition could be given orally (e.g. for the treatment of cancers of the gastrointestinal tract). The ideal formulation would then be a viscous, aqueous solution or gel containing a carrageenan salt (or derivative). For the treatment of diabetic retinopathy, the composition could be administered as eyedrops (in a sterile, isotonic, aqueous solution). As yet a further example, the composition could take the form of a solution, lotion or hydrophilic cream for topical application (in the treatment of psoriasis or skin cancers). Appropriate doses without significant toxic effects would appear to be in the range of l-50mg carrageenan per kilogram body mass, more especially in the range 2-20 mg per Kg body mass.

The invention will be better understood by reference to the following illustrative examples and drawings in which :

Figure 1 is a graph of bFGF binding against concentration of carrageenan;

Figure 2 is a graph of receptor bound TGF beta 1 against concentration of carrageenan;

Figure 3 is a graph of specific binding against

concentration of lambda carrageenan or iota carrageenan;

Figure 3b is a graph of cpm against concentration of iota carrageenan;

Figure 4 is a graph of PDGF bound against concentration of carrageenan or heparin;

Figure 5 is a graph of radiolabelled bFGF bound against concentration of unlabelled bFGF;

Figures 6a-6d are graphs of bFGF bound against concentration of carrageenan or heparin;

Figures 7a and 7b are graphs of bFGF bound against concentration of heparin;

Figure 8 is a graph of bFGF bound against time (in minutes) ;

Figures 9a-9c are graphs of cell numbers against time (in days) ;

Figures lOa-lOf are graphs of cell proliferation against concentration of carrageenan;

Figure 11 is a graph of cell proliferation (as a percentage of controls) against concentration of carrageenan iota;

Figure 12 is a graph of optical density against time (in days ) ;

Figures 13a-13d are graphs of DNA synthesis against

concentration of carrageenan;

Figure 14a is a bar chart showing DNA synthesis in LNCaP cells under various conditions;

Figure 14b is a graph of relative thymidine incorporation (as a percentage of the control) against concentration (in ug/ ) of iota carrageenan;

Figure 15 is a picture showing the results of electrophoresis of DNA on an agarose gel;

Figure 16a and 16b are graphs of thymidine incorporation against concentration of carrageenan for MCF-7 and T47D cells respectively;

Figures 17a and 17b are graphs of O.D. against concentration of carrageenan for MCF-7 and T47D cells respectively;

Figure 18 is a graph of relative thymidine incorporation (%) against concentration of iota carrageenan;

Figure 19 is a bar chart showing thymidine incorporation under different conditions;

Figures 20a-d are graphs of relative thymidine incorporation or relative OD against concentration of iota carrageenan (ug/ml) for MCF-7 cells (a,b), T47D (c) or SKBR3 cells (d) ;

Figure 21 is a graph of clotting time against concentration of polysulphated polysaccharide; and

Figures 22 and 23 are bar charts showing bFGF binding and DNA synthesis respectively occurring in FBHE cells in the presence of different iota carrageenan derivatives.

Examples

The ability of the carrageenans to inhibit binding of various growth factors to target cells was investigated.

All carrageenans were purchased from Sigma Chemical Co., Poole, Dorset. Carrageenans were freshly made up prior to each experiment at 2 mg/ml in an appropriate buffer (25mM Hepes, pH7.4) heated to 60°C, cooled and sterilised by filtration (0.2 u filter). The essential features of the carrageenans used in this study are shown in Table 1.

Tabl e 1

Carrageenan Kappa Iota Lambda (CAR/K) (CAR/I) (CAR/L)

Source Eucheuma Eucheuma Gigartina cottonii spinosa aciculaire and G. pistillata

Sigma C 1263 C 4041 C 3889 Catalogue No.

Ester suplhate 25-30% 28-35% 32-39%

3,6 Anhydro-D- 28-35% 0% 30% galactose

Solubility in Na ⁺ salt Na ^"1" salt All salt cold water soluble soluble soluble

Gelation Gels most Gels most Non-gelling strongly strongly with K ⁺ with Ca ²⁺

Example 1

In this example, the ability of the carrageenans to reduce binding of radiolabelled bFGF ([ 125I] bFGF, purchased from

Amersham International, UK), to baby hamster kidney (BHK) cells was examined. The BHK cells employed, (clone 13), were purchased from the European Collection of Animal Cell

Cultures (ECACC, Porton Down, Salisbury, UK), and were

maintained in MEM supplemented with 10% foetal calf serum (FCS) and 5% tryptose phosphate broth (TPB). All cell lines used in this and the following examples were maintained in the presence of antibiotics (100 units/ml penicillin and 100 ug/ml streptomycin).

All of the carrageenans tested were potent antagonists of bFGF binding to BHK cells, as shown by Figure 1. Figure 1 is a graph showing binding of O.lnM radiolabelled bFGF (as a percentage of positive controls) against concentration of carrageenan (in ug/ml). Determinations were made after a 3 hour incubation at 4°C. The order of potency of antagonistic effect was iota (most, filled circles on the graph) lambda (intermediate, triangles), kappa (least, boxes), with IC50 values of 0.4, 2.4 and 5.7 ug/ml respectively. These and other data are summarised in Table 2.

Table 2. Inhibition of growth factor binding by carrageenans

Growth factor IC50 ( μg/ml) ¹

i-carrageenan κ-carrageenan λ-carrageenan

¹ Values represent mean concentrations±S.D. (n =3) required to reduce specific binding (for PDGF, TGFβl, IGFl, TGFα) or total binding (for bFGF) to 50% of control values.

Examp le 2

Another growth factor investigated for inhibition by carrageenans was TGF beta 1. Binding of radiolabelled TGF beta 1 ([ ¹²⁵ I] TGF beta 1, purchased from NEN [Du Pont UK Limited, Stevenage, Hertfordshire]) to Swiss 3T3 cells was determined by the method of Wakefield (1987, Methods Enzymol . 146, 167-173). The Swiss 3T3 target cells employed were purchased from ECACC and routinely maintained in DMEM supplemented with 10% new-born calf serum (NBCS) and penicillin/streptomycin. Briefly, cells were plated out at 5 x 10 cells/well in 24 well plates and grown for 2 days until about 70% confluent. Cells were washed with binding buffer and incubated with 20pM [ 125I]TGF beta 1 and test compound or an excess of unlabelled ligand for 2 hours at room temperature. Cells were then washed with ice-cold binding buffer and solubilised with 1% Triton X-100, 10% glycerol, 25mM

Hepes, pH7.4 (400 ul/well) and the solubilisate counted in a gamma counter .

The results are shown in Figure 2 (Key as for Figure 1), which is a graph of receptor-bound TGF beta 1 (as a percentage of positive controls) against concentration of carrageenan (in ug/ml). As can be seen from the Figure, the carrageenans did inhibit binding of TGF beta 1 to Swiss 3T3 cells, but this inhibition was weaker than the inhibition of bFGF binding to BHK cells.

Example 3

A third growth factor studied was Insulin-like Growth factor 1 (IGFl). The target cells for IGFl used in this assay were L23/P cells (a large cell adenocarcinoma line),

provided by P. Twentyman, MRC Unit of Clinical Onoclogy and Radiotherapeutics, Cambridge, UK. These cells were routinely maintained in RPMI supplemented with 10% FCS and penicillin/streptomycin.

For the assay, L23/P cells were plated out at 1 x 10 cells/well in 24 well plates and grown to 90% confluency (24 hours). Cells were washed once with binding buffer

(bicarbonate-free DMEM/ 0.1% BSA/ 25mM Hepes, pH 7.4) and incubated with binding buffer containing 0.1 nM [ 125I]IGFl

(purchased from Amersham International, UK) and either test substance or excess ligand for 90 in at 4°C. Cells were washed four times with unlabelled binding buffer, solubilised with IM NaOH and the solubilisate counted in a gamma counter.

The results are shown in Figure 3, which is a graph (triangles) of specific binding of IGFl (as a percentage of the positive control) against concentration of lambda carrageenan (in ug/ml). Suramin was used as a positive control in this experiment and caused 100% inhibition of IGFl Dinding at a final concentration of 200 ug/ml. Figure 3 shows that carrageenan lambda did not inhibit binding of IGFl to L23/P cells at concentrations up to 250 ug/ml. Similarly, carrageenans kappa and iota failed to inhibit binding of IGFl at concentrations of 200 ug/ml (data not shown). Thus the carrageenans do not behave in the same way as suramin with regard to all growth factors (suramin inhibits TGF-beta 1).

Figure 3 also shows data (filled circles) showing that iioottaa ccaarrrraaggeeeennaann ddooeess nnoot inhibit binding of lnM[ ¹²⁴ I] TGF-alpha to A431 cells.

Further experiments, using standard thymidine incorporation assays, were performed. The results are shown in Figure 3b, which is a graph of thymidine incorporation (cpm) against concentration (ug/ml) of iota carrageenan. The experiments demonstrated that iota carrageenan was able to inhibit DNA synthesis stimulated by bFGF or TGF-alpha, but not IGFl.

Example 4

Effects of carrageenans on the binding of platelet-derived growth factor (PDGF) to target cells (Swiss 3T3) were also studied. The assay was essentially as that described for TGF beta 1 except that [ ¹²⁵ I] PDGF (purchased from NEN) was used at 40pM. The results are shown in Figure 4, which is a graph of PDGF bound (as a percentage of the positive control) against concentration of carrageenan. The key is as for Figure 1. This shows that carrageenans kappa and lambda are particularly effective antagonists of PDGF binding to 3T3 cells (IC50s of approximately 0.4 and 0.9 ug/ml respectively).

Example 5

As demonstrated in Example 1, carrageenans iota, lambda and Kappa are potent antagonists of bFGF binding to BHK cells. Further experiments were performed to investigate the nature of this mechanism.

bFGF binds to both low and high affinity receptors on BHK cells, with the low affinity site probably representing heparin or heparin-like substances (Moscatelli, 1987). Binding of bFGF to low and high affinity sites on BHK cells can be distinguished by first washing the cells with

2M NaCl at pH 7.5 and then by extracting the cells with 0.5% Triton X-100 in 0.1M sodium phosphate, pH 8.1 (Moscatelli, 1987).

Competition experiments were carried out between 0.1 nM radiolabelled bFGF and various concentrations of unlabelled bFGF. The results are shown in Figure 5, which demonstrate that similar amounts of bFGF were detected in NaCl washes (filled circles) and Triton X-100 extracts (open circles) from the BHK cells. Increasing the concentration of unlabelled bFGF caused a significant increase in bound radiolabelled bFGF which could be released by an NaCl wash. By contrast, increasing concentrations of unlabelled bFGF reduced binding of radiolabelled bFGF to Triton X-100 extractable sites. Scatchard analysis of the competition curve for Triton extractable sites gave a Kd value of InM.

Figures 6a-6d compare the effects of carrageenans and heparin on NaCl-extractable and triton-extractable bFGF respectively (Key as for Figure 5). Carrageenan iota (Fig. 6a) and heparin (6d) were about equally potent at antagonising binding to the NaCl-extractable (low affinity) binding sites with IC50 values of about 0.3 ug/ml (Figure 6a). Heparin was a weak antagonist of bFGF binding to high affinity sites whereas carrageenan iota inhibited binding of bFGF to the high affinity sites with an IC50 value comparable with that for the low affinity site (Figure 6b). The data for Kappa carrageenan (6b) and lambda carrageenan (6c) are qualitatively similar to those for iota carrageenan.

The interaction of carrageenan iota (CAR/I) with bFGF was further investigated by examining competition between

heparin and CAR/I for inhibition of bFGF binding to low and high affinity sites. The results are shown in Figures 7a and 7b, which illustrate the effects of heparin on the inhibition of bFGF binding to NaCl-extractable and triton- extractable sites respectively. The symbols are as follows :

heparin alone (filled circles); heparin plus CAR/I at 0.1 ug/ml (filled boxes); at lug/ml (triangles); at lOug/ml (empty circles); at lOOug/ l (empty boxes). The inhibition of bFGF binding to low affinity sites by low concentrations of CAR/I was enhanced in a dose dependent manner by increasing concentrations of heparin (Figure 7a). Heparin had little effect on the ability of carrageenan iota to inhibit bFGF binding to low affinity sites. By contrast, heparin had less of an effect on inhibition of bFGF binding to high affinity sites by low (up to 0.1 ug/ml) concentrations of CAR/I (Figure 7b). However, increasing concentrations of heparin reduced the ability of CAR/I to inhibit bFGF binding to high affinity sites. These data indicate that CAR/I competes with heparin for binding to bFGF and that inhibition of bFGF binding to the high affinity site by CAR/I involves binding of CAR/I to the heparin binding site. Similar results were obtained when heparin was pre-incubated with bFGF for lh prior to adding CAR/I (data not shown).

Basic FGF was pre-bound to BHK cells in order to examine the abilities of CAR/I and heparin to displace bFGF from high and low affinity receptors. The results are shown in Figure 8. Figure 8 shows the displacement of bFGF by CAR/l (circles) or by heparin (boxes) from triton- extractable (high affinity, closed symbols) and NaCl (low affinity, open symbols) sites respectively. Carrageenan

iota and heparin exhibited similar activities causing maximum displacement of bFGF from low affinity sites within the first 15 minutes and slower displacement from high affinity sites.

Example 6

In addition to studying the growth factor antagonist properties of carrageenans, experiments were performed to study the in vitro inhibition of mitogenesis, using a foetal bovine heart endothelial (FBHE) cell model.

FBHE cells were purchased from ECACC and were maintained in DMEM supplemented with 10% FCS, antibiotics (100 units /ml penicillin and 100 ug/ml streptomycin) and supplied with bFGF (10 ng/ml every alternate day).

To perform the cell proliferation assay, cells were plated

3 out at 10 /well. The following day medium was replaced with fresh medium containing test substance. Cell proliferation was assessed by an 3- (4,5-diτnethylthiazol-

2-yl)-2 ,5-diphenyltetrazolium bromide (MTT) assay, as described by Alley et al . , (1988, Cancer Research 4_8_, 589-

601) .

Growth of the FBHE cells in vitro was investigated. The results are shown in Figures 9a - 9C, which are graphs of optical density (O.D., a measure of cell numbers) against time for starting concentrations of 5 x 10 cells/ml, 10 /ml and 2 x 10 /ml respectively. FBHE cells were plated out at these different concentrations and growth in the presence and absence of bFGF was compared. Growth in the presence of bFGF was linear up to 8 days for cells plated out at 5 x 10 /ml and 10 /ml (Figures 9a, 9b).

Growth was reduced in the absence of bFGF.

An anti-bFGF antibody reduced the growth of the FBHE cells grown in the absence of exogenously added bFGF by approximately 30% (data not shown).

Once the growth of FBHE cells in vitro was characterised in this way, the effects of carrageenans could be studied. The results of these experiments are shown in Figures 10a, 10b and 10c. These are graphs of proliferation (as measured by O.D.) against concentration of carrageenan lambda (10a), iota (10b) or Kappa (10c), in ug/ml, for cells grown in medium alone (empty circles) or medium supplemented with lOng bFGF (filled circles).

Carrageenan iota (CAR/I) was found to be a potent inhibitor of FBHE cell proliferation (plated at 4 x

3 10 /ml) as determined after 3 days following a single drug administration. CAR/I (Figure 10b) was a considerably more potent inhibitor of FBHE proliferation than CAR/L

(Figure 10a). Carrageenan Kappa (CAR/K) was a significantly weaker antagonist (Figure 10c) than either

CAR/I or CAR/L. Basic FGF-independent growth was also inhibited by the carrageenans. The carrageenans reduced this growth by up to 63%. This is greater than growth reduced by an anti-bFGF antibody, indicating that the carrageenans may inhibit another growth factor which stimulates FBHE cell proliferation.

Essentially the same experiment was performed but with repeated daily administrations of the respective carrageenans. The results are shown in Figures lOd-f (symbols as for Figures lOa-c), which demonstrate that the growth inhibition is greatly enhanced by repeated exposure

to the carrageenans.

The time course of inhibition of proliferation by CAR/I was examined by plating the cells out at 10 /well and determining cell proliferation at two-daily intervals. The results are shown in Figure 11, which is a graph of cell proliferation (as a percentage of positive controls) against concentration of CAR/I (in ug/ml) for cells treated with iota carrageenan for 2 hours (filled circles), 8 hours (empty circles), 24 hours (filled boxes) or 48 hours (empty boxes). Cell counts were made 6 days after the initial treatment.

Figure 12, which is a graph of O.D. against time (in days) shows FBHE cell proliferation in the presence (boxes) or absence (circles) of CAR/I (at 40 ug/ml) in the presence (closed symbols) or absence (open symbols) of lOng/ml bFGF. Error bars represent the standard deviation of triplicate determinations and are smaller than the symbol not shown. This was confirmed by staining the cells with trypan blue (data not shown).

Further experiments were performed to investigate the inhibition of FBHE proliferation by carrageenans, by studying the effects of the compounds on DNA synthesis.

To perform this assay, FBHE cells (or NRK cells) were seeded in 96-well plates at 2 x 10 ,/well. After plating down overnight, cells were washed once with MDMEM (DMEM/0.1% BSA/25 mM Hepes, pH 7.4) and incubated with MDMEM for 1 hour at 37°C to dissociate any bound bFGF. Cells were quiesced in MDMEM/0.5% FCS for 2 days. The medium was then replaced with fresh MDMEM/0.5% FCS containing test substance in the presence or absence of

bFGF (10 ng/ml ) . [ ³ H]thymidine (0.5 uCi/well) was added after 6 hours and uptake into DNA was determined after a further 18 hour incubation (Rozengurt and Heppel, 1975, PNAS 21, 4492 ).

The results are shown in Figures 13a-d, which are graphs of DNA synthesis (as a percentage of controls) against concentration of CAR/I (13a), CAR/K (13b) and CAR/L (13c) respectively. CAR/I and CAR/L were potent antagonists of bFGF-stimulated DNA synthesis and CAR/K was a weaker antagonist. Basic-FGF-independent DNA synthesis was also inhibited by the carrageenans.

Figure 13d is a graph showing inhibition of DNA synthesis in NRK cells by iota carrageenan.

FBHE cells were either left unstimulated (open circles) or were stimulated with bFGF (closed circles) and treated simultaneously with the various carrageenans. DNA synthesis was determined 24 hours later with a two hour pulse of thymidine.

NRK cells were either left unstimulated (open circles) or stimulated with 5% serum (closed circles). DNA synthesis was determined 8 hours later with a 16 hour exposure to thymidine. Error bars represent the SD of triplicate determinations and if not shown are smaller than the symbol .

In summary, the foregoing examples demonstrate that the carrageenans inhibit the binding to target cells of heparin-binding growth factors: bFGF (Example 1); TGF beta 1 (Example 2); and PDGF (Example 4), but do not inhibit the binding of IGFl (Example 3) or TGF alpha (Hoffman,

1993, cited previously). Moreover, for iota carrageenan at least, bFGF is inhibited from binding to both low and high affinity receptors (Example 5) and this inhibition is probably the result of iota carrageenan binding to the heparin-binding domain on bFGF rather than binding to the high affinity receptor (Example 5, and Hoffma-n & Sykes, 1993, Biochemical Pharmacology 4_5, 2348-2351).

Additional experiments were performed to investigate the possible mechanism of action of the carrageenans. In particular, in view of the lesser anti-coagulant activity of iota-car ageenan (compared to kappa and lambda carrageenans described in Example 10 below), this compound was selected for further analysis.

Example 7

The effects of iota carrageenan on the cell proliferation and DNA synthesis of a number of cell lines were investigated .

Details of the cell lines were as follows: U87MG, PC3, LNCaP, A375, Swiss 3T3, NRK (all purchased from ECACC) and BEAS-2B (provided by Dr C C Harris, National Cancer Institute, Bethesda, Maryland). These represent a variety of transformed cell lines. L23/P is a non-small cell lung carcinoma (NSCLC) , L23/R is a multi- drug resistant variant of L23/P. BEN and LUDLUl are 2 human squa ous NSCLC cell lines. BEAS-2B is a non- tumourigenic bronchial epithelial line. Swiss 3T3 is the well-known mouse fibroblast line. MCF-7 and T47D cells are both breast cancer cell lines. U87MG and G36 are both glioma cell lines. LNCaP (lymph node carcinoma of the prostate) is an androgen-responsive cell line, whilst PC3

is an androgen-independent prostate cell line. A375 is a melanoma cell line.

All the lung cell lines (except BEAS-2B) , and T47D and LNCaP cells were maintained in RPMI + 10% FCS. BEAS-2B, U87MG and MCF-7 cells were maintained in DMEM + 10% FCS.

Swiss 3T3 cells were maintained in DMEM/10% NBCS. PC-3 cells were maintained in Ham's F12, 1% AA, 1% NEAA. NRK cells were maintained in DMEM, 5% FCS, 1% NEAA. All media were supplemented with antibiotics (100 units/ml penicillin, 100 ug/ml streptomycin).

Cell proliferation was assessed by performing an MTT assay (as described previously). Briefly cells were plated in 96-well icrotitre plates (2000 cells/well, except L23/P plated at 1000 cells/well, and BEN and LUDLUl , plated at 4000 cells/well) overnight. Fresh medium containing carrageenan was then added and the numbers of viable cells were determined by an MTT assay 5d later according to the published method of Twentyman and Luscombe (1987, Brit. J. Cancer 5J>_, 279-285) .

DNA synthesis was determined by means of a thymidine incorporation assay. Essentially cells were plated in 96- well microtitre plates (4000 cells/well, except L23/P plated at 2000 cells/well, and BEN and LUDLUl plated at 8000 cells/well) for l-2d depending on the cell line. Fresh medium containing carrageenan was then added and after the desired time incorporation of [ H] thymidine (0.5 uCi/well) over a 2-6h period into TCA insoluble material was determined according to Dealtry and Balkwill (1987 "Lymphokines and Interferons" [Eds. Clemens, Morris & Gearing] ppl95-220, IRL Press, Oxford).

The results are summarised in Table 3, which is self- explanatory.

TABLE 3 Table 3. Effect of i-carrapeenan on the proliferation and DNA synthesis of various cell lines cell line DNA synthesis cell proliferation

IC50 (μg/ml) % inhibition by 100 IC50 (μg/ml) μg/ml i-carrageenan

Lung

L23/P 6 60 10

L23/R 6 53 1

MOR >100 42 >100

BEN >100 14 >1() ⁽⁾

LUDLUl >100 10 >100

BEAS-2B >100 16 >100

Fibroblast

Swiss 3T3 88 52 >100 Breast

MCF7 32±4 (n=3) 60 90 T47D 14±3 (n=7) 65 45 Glioma

U87MG >100 38 > 100 G36 Prostate

LNCaP 4.5±1.7 (n=4) 94

PC- >100 0 >!()( ⁾ Melanoma

A375 >100 >!()()

βUBSTITUTE SHEET

Example 8

Table 3 illustrates that DNA synthesis in LNCaP cells is particularly susceptible to inhibition by iota carrageenan. As discussed previously, the growth factor bFGF is greatly inhibited by iota carrageenan, and bFGF is one of a number of growth factors implicated in the growth regulation of prostate cancer lines (reviewed by McKeehan, 1991, Cancer Surveys Jl_l_, 165-175). In order to determine if bFGF is the target for the anti-proliferative activity of iota-carrageenan against LNCaP cells, the effect of excess bFGF on the inhibitory activity of iota-carrageenan was determined.

The results are shown in Figure 14a, which is a bar chart showing DNA synthesis (as measured by incorporation of tritiated thymidine) in LNCaP cells incubated in the absence (left hand side) or presence (right hand side) of lOug/ml iota carrageenan. The columns represent (from left to right): no exogenous growth factor; lug/ml EGF; lug/ml bFGF; and lug/ml IGFl.

The chart clearly shows that excess bFGF does not reverse the iotacarrageenan-mediated inhibition of DNA synthesis, thus indicating that inhibition of bFGF binding is not primarily responsible for anti-proliferative activity.

It was shown in another experiment (Figure 14b) that DNA synthesis in LNCaP cells is highly sensitive to inhibition by iota carrageenan (filled circles), whilst PC3 cells (open circles) were not significantly affected.

Similar results were obtained with DNA synthesis in FBHE

cells (data not shown), the proliferation of which was strongly inhibited by iota carrageenan. Indeed, it was found that incubating FBHE cells in the presence of lOug/ml iota carrageenan resulted in apoptosis, as shown by Figure 15. Figure 15 is a picture of an agarose gel. The contents of the lanes are as follows: 1-phi x 174 RF DNA/Hae III Digest (0.25ug DNA, reference); 2-DNA from untreated FBHE monolayer cells; 3-DNA from untreated "floater" (detached) cells; 4-DNA from FBHE monolayer cells incubated in serum-free DMEM without bFGF; 5-as lane 4, but DNA from "floater" cells; 6-DNA from FBHE monolayer cells incubated in serum-free DMEM without bFGF but with lOug/ml iota carrageenan; 7-as lane 6, but DNA from "floater" cells; 8-as lane 1, but reference DNA subjected to the extraction method used to obtain DNA from the FBHE cells, 9-as lane 1.

The appearance of a "smear" or "ladder" of DNA in lanes 6 and 7 shows that the DNA is being degraded, which is characteristic of apoptosis.

Example 9

A number of polypeptide growth factors are mitogenic for breast carcinoma cells including transforming growth factor alpha (TGFalpha), insulin-like growth factor-I (igFI), platelet-derived growth factor (PDGF) and basic fibroblast growth factor (bFGF) (Lippman et al . , 1988; J. Steroid Bioche . 3_0, 53-61; Peyrat et al . , 1992, J. Steroid Biochem. Molec. Biol. 4_3_, 87-94). Transforming growth factor beta (TGFbeta) was shown to inhibit the growth of the oestrogen-receptor (ER) positive breast cancer cell line MCF-7 (Zugmaier et al . , 1989, J. Cell Physiol. 141, 353-361). Mitogenic growth factors derived

from stromal cells may stimulate breast cancer cells in a paracrine manner.

As shown in Table 3, iota carrageenan was found to inhibit the proliferation of the breast cancer cell line MCF-7. This cell line was therefore selected for further experiments with iota carrageenan.

MCF-7 cells and T47D cells were purchased from ECACC. SKBR3 cells were from the American Type Culture Collection (Rockville, Maryland, USA). MCF-7 and SKBR3 cells were maintained in DMEM/10% FCS; T47D cells were maintained in RPMI/10% FCS. All cells received antibiotics (100 units/ml penicillin and 100 ug/ml streptomycin).

Cells were plated out at 10,000 cells/ml in medium containing 10% foetal calf serum (FCS). The following day, fresh medium/10% FCS containing carrageenan was added. Cell proliferation was assessed 5d later by determining the number of viable cells using an MTT assay (Twentyman and Luscombe, 1987 cited previously). Triplicate determinations were made for each experiment.

DNA synthesis was assessed by measuring the incorporation of a 3-6h pulse of [ H] thymidine into DNA (Dealtry and Balkwill, 1987 cited above). For proliferating cells, 10,000 cells/ml were plated out in medium containing 10% FCS. After a 24h incubation, fresh medium/10% FCS containing carrageenan was added and DNA synthesis was determined 24h later. Determination of the growth factor stimulated DNA synthesis of MCF-7 cells was based on the method of Karey and Sirbasku (1988, Cancer Res. 4_8_, 4083- 4092). 10,000 cells/ml were plated down for 2d in DMEM/10% FCS. The cells were washed in serum-free DMEM

and incubated in serum-free DMEM containing transferrin (10 ug/ml) and BSA (0.2 mg/ml) (defined medium) for l-2d iotacarrageenan was then added in the defined medium, and 3h later growth factors were added. DNA synthesis was determined after a further 24h. E2 stimulation of MCF-7 cells was determined as follows. 10,000 cells/ml were plated down for 2d in DMEM/10% FCS. Cells were washed with phenol red free DMEM and incubated in phenol red free

DMEM containing 10% charcoal-stripped serum for 2d. E2 and carrageenan were added and DNA synthesis was determined 24h later. Triplicate determinations were made for each experiment.

The results of the various experiments are shown in Figures 16-17.

Figure 16 is a graph of thymidine incorporation against concentration of carrageenan (ug/ml) and shows the results of an initial experiment comparing the effects of a single administration of iota, kappa, or lambda carrageenan on MCF-7 (Figure 16a) and T47D (Figure 16b) cells grown in 10% serum.

Iota carrageenan is represented by empty circles, kappa carrageenan by filled circles and lambda carrageenan by empty squares. Iota carrageenan partially inhibited the DNA synthesis of MCF-7 and T47D cells when determined 24h after a single administration; kappa and lambda carrageenan were less active. A 48h exposure of MCF-7 cells to iota carrageenan gave a similar inhibition curve (data not shown) .

A similar experiment involved determining inhibition of cell proliferation by means of an MTT assay performed five

days after administration of the carrageenan. The results are shown in Figure 17 (17a, for MCF-7 cells; 17b for T47D cells). The symbols are as for Figure 16.

In a further experiment, MCF-7 cells were growth arrested in serum-free conditions and their responsiveness to exogenous growth factors was examined (Fig. 18). Figure 18 is a graph of relative thymidine incorporation (%) against concentration of iota carrageenan. Cells maintained in serum-free conditions were treated with iotacarrageenan in the absence of growth factors (empty circles), or with 10 ng/ml IGF-I (filled squares), 10 ng/ml bFGF (filled circles) or lOng/ l TGFalpha (filled triangles) and DNA synthesis was determined 24h later.

Figure 18 shows that DNA synthesis was stimulated by bFGF, IGFl and TGF alpha. The cells did not respond to PDGF (B- chain homodimer). bFGF-stimulated DNA synthesis was inhibited by iota carrageenan to non-stimulated values. There was also substantial inhibition of TGFalpha- stimulated DNA synthesis. In some experiments TGF alpha- stimulated DNA synthesis was completely inhibited. By contrast, IGF-I-stimulated DNA synthesis was either not inhibited or there was only slight inhibition. Iota carrageenan did not significantly inhibit DNA synthesis of the cells not stimulated with exogenous growth factors.

In order to investigate if inhibition of growth factor stimulated DNA synthesis is due to antagonism of growth factor binding, the effects of excess growth factors were determined. It was found that excess bFGF caused only a slight reversal of the inhibition of bFGF-stimulated DNA synthesis caused by 10 ug/ml iota carrageenan in MCF-7 cells (Figure 19; compare Figure 14). Excess TGF alpha

did not reverse the inhibition of TGF alpha-stimulated DNA synthesis. The results are shown in Figure 19, which is a bar chart showing thymidine incorporation.

Essentially, cells were maintained in serum-free conditions and stimulated with 10 ng/ml or 1000 ng/ml growth factors in the absence (solid bars) or presence of 10 ug/ml iota carrageenan (hatched bars). DNA synthesis was determined 24 hours later.

Preliminary experiments had shown that iota carrageenan only partially inhibits DNA synthesis of MCF-7 cells grown in serum. Serum contains many mitogenic factors including hormones and growth factors which may be insensitive to iota carrageenan. Tamoxifen has been shown to decrease the responsiveness of MCF-7 cells to mitogenic growth factors and to increase the production of the inhibitory growth factor TGF beta. The present inventors decided to investigate the effects of tamoxifen in combination with iota carrageenan.

The results are illustrated in Figure 20 (a-d). Briefly, cells in medium supplemented with 10% serum were treated with various concentrations of iota carrageenan (shown in ug/ml) in the absence (empty circles) or presence (filled circles) of luM tamoxifen. Figure 20a shows the results for relative MCF-7 DNA synthesis (as measured by thymidine incorporation), Figure 20b shows the results for relative MCF-7 proliferation (as measured by O.D.), and Figures 20c and 20d show relative DNA synthesis for T47D and SKBR3 cells respectively.

A 2d exposure of proliferating MCF-7 cells to 1 uM tamoxifen alone caused a modest reduction in DNA

synthesis, but a combination of luM tamoxifen and iota carrageenan resulted in a significantly enhanced inhibitory effect compared to the effect of either agent alone (Fig. 20a). A combination of tamoxifen and iota carrageenan also significantly enhanced the anti- proliferative activity compared to the activity of either agent alone (Fig. 20b). By comparison, tamoxifen inhibited T47D cells more than MCF-7 cells, but a combination of tamoxifen and iota carrageenan only caused slightly greater inhibition than tamoxifen alone (Fig. 20c). The ER negative cell line SKBR3 was not inhibited by tamoxifen alone, but iota carrageenan caused more inhibition than in the two ER positive lines, and this inhibition was potentiated by tamoxifen (Fig. 20d) .

The increase in tamoxifen-induced inhibition of DNA synthesis by iota carrageenan is interesting as it occurs at physiologicaly relevant concentrations of the anti- oestrogen. Two ways by which tamoxifen could contribute to the inhibitory activity of iota carrageenan are by reducing the responsiveness of MCF-7 cells to serum- derived IGF-like activity (Wosikowski et al . , 1993 Int. J. Cancer j3_3_, 290-297) or by inducing TGF beta (Knabbe et al., 1987 Cell 4_8_, 417-428). Although there is no direct evidence for the involvement of IGF-I, the enhanced inhibition by the anti-IGF-I antibody (data omitted for brevity) was reminiscent of the effect of tamoxifen. In our experiments the MCF-7 cells were exposed to tamoxifen for 2d, which is the time required for maximal production of TGF beta (Knabbe et al . , 1987). Although an antibody to TGF beta did not reduce the enhanced inhibition by tamoxifen (data omitted for brevity), it is possible that there was insufficient antibody to neutralize the TGF beta produced by the MCF-7 cells in response to tamoxifen. The

concentration of antibody used in our experiments neutralizes approximately 0.1-0.5 ng/ml of total TGF beta in a thy ocyte growth inhibition assay (Dasch et al . , 1989, J. Immunol. L4_2. 1536-1541). 10 ng/ml TGF beta was required to inhibit DNA synthesis of MCF-7 cells in our experiments and this relatively high concentration is consistent with other workers (Zugmaier et al . , 1989 cited previously) . Although 1 ng/ml TGF beta alone did not inhibit DNA synthesis, this concentration did enhance the activity of iota carrageenan. This observation suggests that serum-derived growth factors, sensitive to inhibition by iota carrageenan, may attentuate the anti-proliferative activity of 1 ng/ml TGF beta.

There is good evidence from the present experiments that the inhibition of growth factor-stimulated DNA synthesis by iota carrageenan is not simply due to growth factor antagonism: excess bFGF did not reverse inhibition of bFGF-stimulated DNA synthesis by iota carrageenan, and iota carrageenan did inhibit TGF alpha-stimulated DNA synthesis, although this growth factor is not antagonised by iota carrageenan (Hoffman, 1993 cited previously). The ability of iota carrageenan to inhibit DNA synthesis stimulated by growth factors which are not directly antagonised suggests that this agent may have anti- proliferative activity against a number of cell lines. The present inventors have found this indeed to be the case.

It would appear from the work disclosed herein that carrageenans are able to inhibit tumour growth by at least 2 mechanisms.

Firstly, by inhibiting bFGF binding and inhibiting

proliferation of endothelial cells, carrageenans can inhibit the neovascularisation which is essential for tumours to progress beyond a certain limited size, thereby inhibiting tumour growth indirectly.

Secondly, carrageenans can inhibit tumour growth directly by interfering with growth factor binding and subsequent growth "signals". The mechanism(s) have not been fully elucidated, but it seems possible that carrageenans block the autocrine and paracrine growth factor stimulation of several different tumour cells. This inhibition can be potentiated by other growth antagonists, such as the anti- oestrogen tamoxifen, which block other growth signal pathways insensitive to the action of carrageenans.

Example 10

A simple experiment was performed which demonstrated that iota carrageenan has less anti-coagulant activity than the other carrageenans (kappa and lambda). Thus, if systemic administration of carrageenans is envisaged for treating tumours, the preferred compound would be iota carrageenan or derivatives thereof.

Briefly, lOul of a solution of the relevant carrageenan (diluted with 0.9% NaCl) was mixed with 90ul of pooled plasma, and the coagulation time measured with a KC-10 coagulometer (Amelung, Lemgo, Germany). Activated partial thromboplastin time (APTT) was determined using Pathrombin™ (Behring, Marburg, Germany).

The results are shown in Figure 21, which is a graph of clotting time (in seconds) against concentration of polysulphated polysaccharide (in ug/ml). Data points are

as follows: heparin (filled circles), lambda carrageenan (boxes), kappa carrageenan (crosses) and iota carrageenan (5-pointed stars).

Example 11

Mention has been made of carrageenan derivatives. A range of lower molecular weight derivatives of iota carrageenan were prepared using the novel method devised by the instant inventors, described below.

Carrageenans (e.g. iota carrageenan, with a molecular weight in the range 200 - 500, kD) were dissolved in water to form a 1% solution. This solution was dialysed against distilled water to remove low molecular weight impurities (with a molecular weight less than lOkD). When the solution reaches a pH below 3.5 it is subjected to freeze drying. The product is a white powder and is the low molecular weight form of free acidic carrageenans (mw 0- 125kD) . Carrageenans freeze dried from a solution with a pH below 4.0 are unstable and become degraded, giving a range of lower molecular weight products.

To produce stable sodium salts of the acidic carrageenans, the pH of the carrageenan solution is adjusted to about pH7.0 with sodium hydroxide. Freeze drying as above gives the sodium salt.

This method is preferable to, for example, chemical or enzymatic hydrolysis as it is particularly "gentle", leaving the anhydrogalactose residues and sulphate groups unaltered. Also, chemical hydrolysis is inefficient due to the high viscosity of carrageenan solutions.

The various products can be fractionated according to their charge and/or molecular weight. Separation according to charge can be achieved on either a weak or a strong anion-exchanger using a sodium chloride gradient. Separation accoding to molecular weight can be achieved by gel permeation chromatography. The fractions can be characterised by various conventional techniques. For example, molecular weight and molecular weight distribution can be determined by gel permeation chromatography; the galactose and anhydrogalactose content can be assayed by gas chromatography; and the sulphate content can be measured by a conductometric or photometric method.

Fractions made by the methods described were analysed for any inhibition of bFGF binding and inhibition of FHBE cell proliferation. Sample resuls are shown in Figures 22 (a,b) and 23, which are bar charts showing bFGF binding (cpm) and FHBE DNA synthesis (cpm) respectively for each fraction. The respective assays were performed as described for Examples 1, 5 and 6. Figure 22a shows the amount radiolabelled bFGF bound to low affinity receptor sites in the presence of 100 ug/ml (solid bar) or lOug/ml (hatched bar) of each fraction. Figure 22b shows the results for bFGF binding to high affinity receptors (same symbols as for Figure 22a).

Figure 23 shows the amount of DNA synthesis as measured by cpm of tritiated thymidine incorporation in FHBE cells incubated in the presence of lOOug/ml (solid bars) or lOug/ml (hatched bars) of each fraction.

As can be seen from the figures, several fractions showed inhibitory activity when present at lOOug/ml, especially

fraction 67, although no fraction was as active as iota carrageenan itself.

Fraction 67 was prepared by ion exchange chromatography on a strong anionic column, using a 0-4M NaCl gradient. The fraction has a molecular weight range of 3,200-105,000

Daltons, with an average molecular weight of about 18,000.