NO872267L - PREPARATIONS CONTAINING LIPID MOLECULES WITH IMPROVED ANGIOGEN ACTIVITY AND PREPARATION THEREOF. - Google Patents

Description

The present invention relates to lipid-containing preparations with improved angiogenic properties.

The invention also relates to a method for producing it.

Angiogenesis is the process by which new blood vessels are formed with accompanying increased blood circulation. The area of angiogenesis has been a research and investigation favorite for more than 100 years, see e.g. Virchau, R., "Die Krankhaftern Geshwulste"r Hirshwald, Berlin (1863); Thiersch, C, "Die Haut Mit Atlas", Leipzig (1865); (significance of the interaction between host vasculature and survival and growth of solid malignant tumors observed). Interest has been fueled by the observation that angiogenic factors have been found, in trace amounts, in normal tissue, see e.g. D'Amore et al., "PNAS" 78: 3068-3072 (1981); Kissun, et al., "Br. J. Ophthalmol". 6.6:165-169 (1982); (retinal tissue); DeCarvellho, et al., "Angiology" 34:231-243 (1983); (activated lymphocytes and macrophages); Frederick, et al., "Science" 224:289-390 (1984); (human follicular fluid); Burgos, "Eur. J. Clin. Invest" 13:289-296 (1983); (amniochondrion and placenta); Catellot, et al., "PNAS" 79.: 5597-5601 (1982); (cooling medium of 3T3 cells).

The trace amounts of angiogenic factors observed in these tissues, however, show no anglogenic activity other than in normal growth and development of tissues and organs. In a similar way, angiogenic factors have been observed in tissue of pathological origin, see e.g. Weiss, et al., "Br. J. Cancer" 40:493-496 (1979); Fencelau, et al., "J.Biol.Chem." 256:9605-9611 (1981), McAslau, et al., "Exp. Cell Res." 119:181-191 (1979); (tumor cells); Kumar,, et al., The Lancet 2:364-367 (1983); Brown, et al., "Lancet" 1:682-685 (1980) (synovial fluid from arthritis); Hill, et al., "Experientia" 39:583-585 (1983)

(vitreous material of diabetics); Banda, et al., "PNAS" 79:7773-7777 (1982) (wound fluid).

Goldsmith et al, "JAMA" 252:2034-2036 is the first report of an Anglophone factor showing activity beyond normal weight and development, and large amounts of the factor. The factor was found in chlor of worm/methanol fraction J onates of feline omenta (CMFr). Reference should be made here to the parallel US-SN 672 624 of 20 August 1984.

It has now been found that the crude lipid extract of Goldsmith et al. can be purified into different fractions that have angiogenic properties far beyond what is observed in SMFr.

In addition, commercially available gangliosides such as gangliosides derived from brain tissue or other lipid-containing compounds have also been found to have angiogenic properties. Furthermore, it has been found that new mixtures of known lipid-containing compounds can be created and which also have angiogenic properties.

The discovery of lipid-containing compounds that have angiogenic properties is new in this art. Previous attention has been focused on proteinaceous angiogenic factors, see e.g. Kumar et al., "Lancet" IJ_:364-367 (1983) (proteinaceous factors from 300 to 10<5>dalton); Kissun, et al., supra (proteinaceous factors up to 70 kd); Banda, et al., supra.

(proteins of about 2-14 kd); Burgos et al. supra. (protein complexes with from 100 to 200 kd). It has now unexpectedly been shown that the mixtures containing lipid-containing molecules such as gangliosides, glucolipids, ceramides, cerebrosides, phospholipids, sphingosides, etc., show increased anglogenic action.

Lipid-containing compositions derived from mammalian sources and having angiogenic properties are described below. -In addition, new mixtures containing mixtures of old lipid-containing compounds, which also have angiogenic properties, will be described. Furthermore, it is described that known lipid-containing mixtures have unexpected angiogenic properties. Figure 1 which is also described in US-SN 642 624 indicates a method for obtaining CMFr. Figures 2 and 3 show methods for further fractionation of the CMFr fraction obtained by the method according to Fig. 1. Figure 3 further shows the further fractions obtained by the methods described below. Figure 4 shows capillary formation in the rabbit cornea after treatment with CMFr.

Figure 5 shows multiple capillary formation in the stroma.

Figures 6 to 17 are graphic illustrations of linear categorization studies based on CÅM analyzes according to the invention. Figures 8 to 24 are photographs and graphical representations of in vivo revascularization studies conducted using CMFr.

Embodiments of the invention.

I. Obtaining the chloroform methanol llpidic (CMFr)

faction.

Adult female cats with a body weight of 2.4 to 3.2 kg were anesthetized by an intramuscular injection of ketamine at a preferred dose of 7.0 mg/kg. After anesthetization, a laparotomy was performed through a midline incision according to conventional surgical procedures. Incidentally, this moment is an abdominal cavity body tissue. The omenta were surgically removed and placed in sterile plastic bags kept at 4°C for immediate treatment. At the same time, subcutaneous fat was removed and processed in a manner identical to the omental tissue for use in procedures as a non-omental lipid control. Using appropriate aseptic techniques, the omentum was weighed, spread out on a plastic surface, and cut into individual pieces approximately 4 cm<2>in size using surgical equipment. These individual omental pieces ranging in weight from 7 to 66 g were placed in a sterile Waring blender containing 300 ml of phosphate-buffered saline, hereafter called PBS, which had been pre-cooled to 4°C. The omental pieces were mixed for 5 min. at 2500 rpm. thereby giving an omental homogenate which was then placed in sterile 250 ml plastic bottles and centrifuged at 1600 x g in a cooled centrifuge at 4°C for 20 min. After centrifugation, three distinct and separable fractions were observable in the bottles: a pellet of mixed composition; a turbid homogenate containing essentially all proteinaceous material, and a liquid, cream-colored lipid cake. Each of these fractions was isolated individually.

The pellet - of mixed composition was discarded in its entirety. The turbid homogenate fraction was completely saturated, i.e. 100$, with aqueous ammonium sulphate which caused precipitation of the total protein content in this fraction. Testing of the turbid homogenate fraction and the total protein precipitate (resuspended in PBS) via the cornea analysis showed that none of these preparations had good anglobus activity.

The lipid fraction that was isolated as a liquid lipid cake consisted of two distinct layers: an upper foam-like mixture and a denser, more compact layer that was also darker in color than the upper one. Each layer was evaluated and found to contain an active Anglophone factor in significant quantities. For this reason, each of these lipid layers individually and in combination contains the active 1 lipid fraction per se according to the invention. The weight of the lipid cake comprising both layers was found to be approx. 93& of the total weight of the omentum from which it is derived and it is this lipid cake that is used to prepare the concentrated organic extract comprising the active angiogenic factor 1 anding. Active lipid fractions were extracted using the amounts and proportions of the lipid cake indicated in Table I below:

The indicated amounts of lipid cake were combined with approx. 21 ml of an organic solvent comprising chloroform:methanol in a volume ratio of 2:1 in an Eberbach 8575 microblender and homogenized for two minutes. The lipid/organic solvent homogenate was then centrifuged at 200 x g in a clinical centrifuge at room temperature for 10 min. thereby obtaining a clear, golden supernatant and a particulate precipitate. The supernatant was isolated using conventional procedures and subjected to rotary evaporation at 37°C under vacuum to completely remove the chloroform/methanol solvent. Other methods of solvent removal are known in the art and may be used instead of rotary evaporation. A viscous liquid was obtained which was then preferably suspended in approx. 4 ml PBS for use in the cornea and CAM analyses. II. Obtaining purified fractions. The CMFr obtained above was dissolved in a mixture of hexane (approx. 60 ml hexane/10 g extract ) and 0.66 volumes of 95% ethanol were then added. The phases were mixed thoroughly and allowed to separate. The upper phase, hexane, was re-extracted with 95% ethanol and the resulting lower phase of the re-extraction was combined with the first ethanol fraction. The combined ethanol fractions were then re-extracted with hexane and the resulting hexane layer combined with the first hexane fraction. Both phases were dried to obtain "hexane upper phase mat aerial" and "methanol lower phase material", hereafter called hexane-UP and ethanol-LP.

The ethanol-LP fraction was then subjected to Folch distribution according to Folch, et al., "J. Biol.Chem." 226:497-509 (1957)

(that is, the fraction was dissolved in 20 volumes of chloroform:methanol in a 2:1 volume:weight ratio and 0.2 volumes of water was added. The phases were thoroughly mixed and allowed to separate). The upper phase of the Folch distribution was removed and the lower was washed with 0.4 volumes of methanol:water 1:1. This gave an upper methanol phase which was combined with the Folch upper phase and then dried to obtain a proportion hereafter called "Folch UP". The lower part was also dried and is henceforth called "Folch LP".

The Folch LP portion was dissolved in chloroform and then subjected to chromatography on a silicic acid, unisil column, as described by Vance, et al., "J. Lipid Res". 8:621-630 (1967). The column was eluted successively with 20 column volumes of chloroform, acetone:methanol 9:1, and methanol. In the successive elutions, neutral lipids (chloroform), glucolipids (acetone:methanol) and phospholipids (methanol) separate.

The Folch UP share was dissolved in approx. 3 ml/mg methanol:water 1:1 and was then loaded onto a C18 reverse phase cartridge as described by Williams, et al., "J. Neurochem", 35 (1): 266-269 (1980) and the cartridge was then washed with 4 volumes of methanol:water 1:1 to thereby obtain a "non-lipid UP material" after drying, and 4 volumes of chloroform:methanol 2:1 to obtain "lipid UP material" after drying.

The lipid UP material was then dissolved in methanol:chloroform:water 60:30:8 and applied to a DEAE-Sephadex acetate column according to Christie, "Lipid analysis", Pergamon Press, 2nd ed. pp. 109-110 (1982). This column was then eluted with 10 volumes of the methanol:chloroform:water mixture originally used to obtain what is called a "neutral lipid upper phase fraction" or "neutral lipid UP". Extraction with methanol:chloroform:0.8 M sodium acetate 60:30:8 gave ganglioside fractions. Both fractions were evaporated to dryness and the glycoside fraction was desalted using a C18 reverse phase cartridge.

The chloroform:methanol fraction was extracted with 1.0% acetic acid

in a volume ratio of 1:10 weight:volume by stirring with a magnetic stirrer for 10-12 min. The extract was centrifuged at 2,000 rpm. for 5 min. in 200 ml flask. The top layer, i.e. the acetic acid insoluble fraction, was then removed. The acetic acid-soluble fraction was combined with an equal volume of chloroform and centrifuged as above to achieve a clear separation of the two phases. Each phase was backwashed twice with the opposite solvent and all chloroform phases were pooled. Evaporation of the chloroform gave the "acetic acid soluble" fraction.

Affinity chromatography (heparin and gelatin binding) of the Folch UP "PBS homogenate" fractions was carried out as follows: Heparin-sepharose CL-6B beads or gelatin sepharose beads (approx.

3 g each) were washed with 450 ml of water in a sintered glass filter. The sepharose beads were suspended in water and packed into 2.5 x 9 cm chromatography columns. Excess water was drained off and the dry sample, i.e. Folch UP, was suspended in 0.01 phosphate buffer pH 7.0 and applied to the column. Elution was carried out with the same buffer (100 ml in total at a flow rate of 2 ml/min.) This was followed by a wash with 100 ml of water to remove salts from the column. Final elution of heparin- or gelatin-binding material was carried out with 50 ml of 0.5% acetic acid. After evaporation of the acetic acid, the material was resuspended in phosphate buffer for testing.

Chloroform:methanol 2:1 volume:volume 1 The lipid fraction was further characterized with respect to its component parts or subfractions using silica gel or iatrosphere liquid chromatography. For these chromatographic fractionations, the procedures described by A. Kuksin, "Chromatography Part B" (E. Heftmann, editor), Elsevier, New York, 1983, were used. In the present method, 5.0 ml of chloroform:methanol the lipid extract placed in a chromatography column containing silica gel (100 - 200 mesh) which had previously been equilibrated with chloroform. Using these gel columns, elutions were carried out in sequence using 100 ml amounts of the following solvents: chloroform; ethyl acetate; ethyl acetate:methanol 3:1; methanol:water 4:1 followed by 200 ml of a solvent mixture comprising chloroform:methanol:acetic acid:water 25:15:4:2. Five individual elution fractions were obtained, here numbered 1 to 5.

Gel permeation chromatography was carried out on a Sephadex LH-20 column. 100 mg of the chloroform:methanol extract or of the ethanol LP was fed to the column and the elution was carried out with a chloroform:methanol 1:1 solvent. Fractions were pooled and the solvent evaporated for testing.

III. Analysis of the fractions.

The subphase glycolipids were examined by HPTLC with chloroform:methanol:water 60:35:8 as developing agent and visualized with the orcinol spray reagent described by Svennerholm, "J.Neurochem", 1:42 (1956). It was also analyzed by HPLC as their perbenzoyl derivatives as previously reported by Ullman, et al., "J. Lipid Res." 19:910-913

(1978). The upper phase complex neutral glycolipid fraction was examined by HPTLC and by immunoblotting with Forsmann and SSEA-I antibodies. The major component of the upper phase complex neutral glycolipid fraction was further purified by preparative TLC or by chromatography on an Iatroball column (1 x 50 cm, 60 p), eluted with hexane:isopropanol:water mixtures as indicated by Kannagi, et al., "J. Biol .Chem" 237(24) 14865-14874 (1982); and Hakamori, et al., "J. Biol. Chem" 259(7) 4672-4680 (19849.

The ganglioside fraction was treated with 0.25 N sodium hydroxide in methanol for 2 hours at 37°C, neutralized with glacial acetic acid and desalted with a C18 reverse phase cartridge. The alkali-treated ganglioside fraction was then subjected to chromatography on a DEAE-Sephadex column and eluted with 0.02M, 0 .08M and 0.05M ammonium acetate in methanol to obtain mono-, di- and polysialoganglioside fractions respectively, see for this Ledeen et al., "Methods in Enzymol", v.83, part D, pp 139-191 (1982). The ganglioside fractions were separated into individual components by chromatography on a 0.4 x 50 cm 10 pm particle Iatrosphere column eluted with chloroform:methanol:water 65:35:8 Fractions of 1.2 ml were pooled, equal portions examined by HPTLC and fractions containing individual components combined appropriately. The non-lipid material was extracted with methanol, centrifuged and the supernatant removed. The insoluble residue was dissolved in water and the water- and methanol-soluble fractions examined by HPTLC in several resolution m iddel systems.

Purified gangliosides were dried under nitrogen and 100 µl of 0.05 M sodium acetate buffer pH 5.5 containing 0.02556 CaCl2 was added. V. cholerae neuraminidase (100 µl, 0.1 units) was added and the sample incubated for 3 hours at 37°C. The reaction was stopped by the addition of 2 ml of chloroform:methanol 2:1 and the mixtures passed over a reverse phase cartridge and non-lipid -ponents eluted with water. Remaining gangliosides and the lipid reaction product were eluted with methanol and chloroform:-methanol and examined by HPTLC. Released sialic acids were also examined by TLC as the trimethylsilyl derivatives according to Ledeen, supra.

For sugar and fat analysis, the glycolipids were subjected to methanolysis in anhydrous 0.75 NHCl in methanol according to Leeden, supra and Kozulec, et al., "Analytical Biochem" 94:36-39 (1979). The fatty acid methyl esters were analyzed by TLC. The methylglycosides were analyzed as the trimethylsilyl derivatives on the same OV-1 column as described by Kozulec, supra. For HPTLC analysis of the lower phase neutral glycolipids, the fraction was perbenzoylated with benzoyl chloride in pyridine and benzoylated "glycosphingolipids" were separated and quantified by HPLC on an uncoated Zipax column with gradient elution and 230 nm detection as previously indicated by Ullman, supra. For direct probe mass spectrometry, glycolipid or ganglioside samples of 5 to 50 pg were trimethylsilylated in 25 µl of pyridine/hexamethyldisilane/trimethylchlorosilane/N.O-bistrimethylsilyltrifluoroacetamide. Anything from 1 to 5 pg of derivative was placed in a sample cup and the sample was heated from 100 to 350°C at a rate of 30°/min. The mass spectrum was obtained with a Finnigan model 4500 quadrupole mass spectrometer equipped with a Teknivent, model 56K computer system. It was operated with an ionization current of 0.5 mA and an ionization voltage of 70 eV. The ionization temperature was 150°C .Repetitive sweeps of the mass range from 100 m/e to 950 m/e were conducted at 5 sec intervals.

The glycolipids were chromatographed on aluminum fired HPTLC plates with chloroform:methanol:water 60:35:8, dried, and soaked in 10.05% polyisobutyl methacrylate in hexane as described by Brockhause et al, "J.Biol.Chem" 256:13223-13225 (1981). The plates were then soaked in PBS containing 1% bovine serum albumin for 2 hours before corresponding exposure to antibody for 2 hours at 40°C. The plate of upper phase neutral glycolipid was treated with Forssmann monoclonal antibody IgM, obtained from ATCC (TIB 121). The TLC plates of the disialoganglioside fraction were treated with GD3 monoclonal antibody IgM prepared by the inventors. After washing in PBS, the plates were exposed to goat anti-mouse IgM, conjugated to horseradish peroxidase for 2 hours at 20"C. After washing in PBS, the plates were developed with 33 mN 4-chloro-naphthol in 0.02 M Tris-HCl buffer containing 20$ methanol and 0.025 56 HgOg.

IV. Characterization of the fractions.

Feline omentum was homogenised, centrifuged and the liquid lipid cake was extracted with chloroform and methanol and further fractionated as shown in Fig. 2. The hexane phase contained approx. 9856 of the material in CMFr and was shown to consist primarily of triglycerides as determined by TLC. Alkaline methanolysis and GC/MS analysis of the resulting fatty acid methyl esters showed that 14:0, 16:0, 16:1, 17:0, 18:0, 18:1 and 18:2 were the main triglyceride fatty acids (the that is, the first number indicates the carbon chain length for the fatty acid and the second the number of unsaturated bonds).

The ethanol phase material was subjected to Folch solvent partitioning and the lower phase lipids which constituted 8056 of the ethanol phase lipids were fractionated on a Unisil column. The neutral lipid fraction recovered from the Unisil column also consisted mainly of triglycerides and small amounts of cholesterol and free fatty acids were detected by TLC analysis. The acetone glycolipid fraction was examined by TLC and the components migrating as hexosylceramide, lactosylceramide, globotriaosylceramide and globoside were present. Quantitative analysis of these glycolipids by HPTLC is described infra. The methanol phospholipid fraction was examined by TLc and the components that migrated such as phosphatidylserine, phosphatidylcholine and sphingomyelin were present.

About. 20 wt-56 of the ethanol phase material was recovered at Folch UP. This Folch UP material was applied to a reverse phase cartridge and the non-lipid fraction eluted with methanol water and the lipids eluted with chloroformmethanol. The lipid UP material was applied to a DEAE column and the neutral lipid fraction not retained by the column was collected and found to constitute 4056 of the lipid UP material. When examined by HPTLC, this fraction was found to contain primarily a glycolipid that migrated below globoside and a small amount of more complex glycolipids.

The ganglioside fraction was eluted from the DEAE column with ammonium acetate in methanol and desalted using a reverse-phase cartridge. Examination by HPTLC shows the presence of recorsinol components which migrated as GM3, GM1, GD3 and several minor polysialoganglioside components. The further purification and identification of these gangliosides is described below.

The non-lipid upper phase fraction, not lipid UP, was brought to a dry state and extracted with methanol. The bulk of the material was not methanol soluble and the suspension was centrifuged and the supernatant removed. The insoluble material was completely soluble in water. These fractions were examined by TLC and the water-soluble fraction showed only a ninhydrin positive signal. The mass of this water-soluble material turned out to be salt. The methanol-soluble material contained at least 6 orcinol- and ninhydrin-positive components and a GC/MS analysis indicated after trimethylsilylation that this material was a complex mixture of the sugar, amino acids, peptides and glycopeptides. The weight distribution of the fractions from this omentum crude lipid extract is shown in Table II.

Aliquots of the glycolipid fraction were benzoylated with benzoyl chloride in pyridine and the perbenzoylated derivatives analyzed by HPLC at 230 nm detection. The results are shown in Table III. These data show that the percentage distribution of glycolipids in this fraction is, as: GlcCer (Nfa), 2656; GalCer (Nfa), 9.6*; GlcCer (Hfa) + GalCer (Hfa) + GaOse2Cer (Nfa), 12; LacCer, 11*; Gb0se3Cer, 10*; Gbose4Cer, 26*.

The upper phase neutral lipldf fraction was examined by HPTLC and was found to consist of approx. 90* of an orcinol positive material which migrated somewhat more slowly than the globoside standard, as well as minor amounts of 3 to 4 more polar orcinol positive components. Immunoblotting with Forsmann and SSA-1 antibodies indicated that the major component was Forsmann positive and that no SSEA-1 positive components were present. The main component was further purified by chromatography on an Iatrokule column and subjected to methanolysis and component analysis by GC/MS. The hexose ratios were found to be Gl c / Gal / NAcGal 1:2:2. The fatty acids present were primarily . The intact glycolipid was also siled and examined by direct probe mass spectrometry. The spectra, shown in Fig. 3, indicate the presence of terminal hexosamine, internal hexose residues, the presence of C-18 sphingosine and fatty acids. Taken together, these data indicate that the glycolipid is Forsmann pentaglycosylceramide. Although the position and configuration of bonds have not been directly determined, the activity and glycolide analytical data strongly support this structure.

The ganglioside fraction was treated with mild alkali to destroy any ester linkage that might be present and separated into mono-, di- and polysialoganglioside fractions by DEAE-Sephadex chromatography. The monosialoganglioside fraction was shown by HPTLC to consist primarily of components that migrated as a triplet of signals corresponding to the mobility of the GM3 standard and a small amount of material that migrated as GM1. The monosialoganglioside fraction was further purified by chromatography on an iatrosphere column and the fractions containing only components migrating as GM3 were pooled. This material was treated with neuraminidase and the 1 lipid product was characterized as lactosylceramide by HPTLC and direct probe MS. The released sialic acid was shown by GC analysis to consist only of N-acetylneuraminic acid. The intact ganglioside was subjected to methanolysis and the sugars and fatty acids examined by GC analysis. The Gle/Gal ratio was found to be 1:1 and the fatty acids consisted primarily of 16:0, 18:0, 18:1, 20:0, 22:0, 23:0, 24:0 and 24:1. The preparations were also examined by direct probe-mass spectrometry as the trimethylsilyl ether derivative. A mass spectrum was obtained corresponding to that emitted by the ganglioside GM3 standard (sialyl [2-3]galactosyl[1-4]glucosyl[1-1] ceramide.

The disialoganglioside fraction was shown by HPTLC to consist primarily of a component that migrated as GD3. This material was further purified by chromatography on an Iatroball column and the fractions containing only a single component that migrated as GD3 were pooled. The preparations were subjected to methanolysis and the methyl glycosides and fatty acid methyl esters analyzed by GC/MS. The Gle/Gal ratio was found to be 1:1 and the major fatty acid components were 16:0, 18:0, 18:1, 24:0, 24:1. The material was treated with neuraminidase and the lipid product identified as lactosylceramide by HPTLC and direct probe MS analysis. Released sialic acid was shown to consist only of N-acetylneuraminic acid by GC analysis. Direct probe MC of the PMS derivative gave spectra consistent with GD3. The material was also shown by immunoblotting to react with a monoclonal antibody produced in the inventors' laboratory with proven reactivity with GD3.

The polysyaloganglioside fraction was shown to contain components that migrated on HPTLC such as ganglioside GT1a, GT1b, but insufficient amounts were obtained for further analysis.

V. Angiogenesis.

The angiogenic effect of the lipid preparations as described in

Fig. 1 and by silica gel chromatography fractions 1 to 5 were tested by the rabbit cornea sample as follows: A series of New Zealand white rabbits were anesthetized with intravenous pentobarbitol in an amount of 30 mg/kg. From each preparation shown in Table I, a single 50 µl injection of the aqueous lipid suspension was carried out via a 25 needle placed intrastromally in the cornea in each eye. The animals' corneas were examined grossly and with a working microscope on the second, fourth, sixth, eighth and tenth days after the ocular injection. Vascular growth and the presence of possible corneal edema and/or inflammation were noted. On the tenth day after the visual examination, the rabbits were killed individually and histological sections, stained with hematoxylin and eosin in the usual way, were obtained from 6 µm thick sections, cut from the formalin-fixed enucleated eyes. Photographs of positive rabbit eyes were carried out.

The angiogenic response was graded as follows: 0, identified no angiogenesis and clear cornea; 1+ identified scleral vessel dilatation with red staining noted in the limbus; 2+ identified several individual blood vessels that migrated from the limbus two-thirds of the way to the injection site; 3+ identified multiple blood vessels extending from the limbus to the injection site and involving 10 - 20* of the cornea; 4+ identifies dense vascularization that extends from the limbus to the injection site and involves at least 30 - 40* of the cornea.

For comparative purposes, an aqueous suspension of the omental lipid cake and an aqueous preparation of the subcutaneous non-omental fat were also prepared and tested. The non-omental fat preparation was obtained by combining a 3 g portion of adipose subcutaneous tissue with 4 ml of PBS and homogenizing this mixture using an Eberbach microblender for 2 min. at 4°C. In a similar way, an aqueous suspension of the omental lipid cake was prepared by homogenizing 4 g of proteins of the lipid cake with 4 ml of PBS in the microblender for 2 min. at 4°C. The homogenate of the whole omentum before centrifugation into proteinaceous fractions and lipid fractions was also evaluated. The result is shown in Table II below.

These data indicate that excellent anglogenic activity was observed after a single 50 µl central corneal injection of the chloroform-methanol lipid extract. In comparison, only minimal Anglogenic activity was noted with the PBS homogenate of the total omentum and with the PBS homogenate of the lipid cake before extraction. Note, however, that a heparin-binding component was concentrated by affinity chromatography from the PBS homogenate and which showed a good Anglogenic effect in the CAM analysis (see below). No angiogenesis occurred at all in the cases following injection of PBS alone or subcutaneous non-omental fat PBS homogenate. A complication that was noted in the data according to Table II, however, is that the injected subcutaneous fat taken from the cat's abdominal wall caused severe inflammation of the cornea within two days of the corneal injection.

The course of the angiogenic response in the cornea to the injected aqueous suspended chloroform:methanol lipid preparation followed a consistent pattern of rapid development and intense activity. After injection of the extracted lipid fraction, a mild corneal inflammatory reaction was observed within 24 hours which resolved within 48 hours. This first inflammation is characterized by a slight opacity in the cornea with minimal erythema in the scleral area, often accompanied by a light drainage of fluid from the eye. A pannus, the appearance of a blanket of blood vessels around the corneal edge, with intestinal blood vessel formation, was clearly visible 3-4 days after the injection. On the 7-10. By today the blood vessels had formed a dense and richly structured network within the cornea. This is illustrated by the photograph in Fig. 4. Histological examination of the enucleated eye, taken on the tenth day, showed several capillaries in the corneal stroma; a photograph of the histological section showing such multiple capillaries in the stroma is shown in Fig. 5.

It should be particularly pointed out that solvent-extracted lipid fractions in aqueous medium initiate and support angiogenesis after only a 50 µl dose injection. Although the mechanism of this angiogenic process and the response are currently unknown, it is clear that the injection of the extracted lipid fraction initiates and induces new blood vessel formation which is organized into a dense, well-structured, vascular network within 7 to 10 days.

As shown in Fig. 3, further fractionation of CMFr is carried out by means of silica gel chromatography. Subsequent testing of each of the five lipid subfractions with which the cornea analysis showed angiogenic activity only in subfraction V with an unnoticeable angiogenic effect from subfractions I to IV. However, the total activity of subfraction V was measurable to a lesser extent than the originally obtained chloroform:methanol lipid extraction preparations. It was subsequently found that the subfactors I-IV, although having no angiogenic activity by themselves, together with subfraction V caused an increase and improvement in the activity and potency of the angiogenic mixture as a whole.

The experiments indicated above demonstrate that CMFr exhibits angiogenic activity. Further experiments were then carried out using further fractions prepared by following the setup according to Fig. I. The experiments consisted of carrying out CAM analyzes as described below. This leads to the derivatization of the "Angiogenic Index" which is a measure of the effect the fractions had in the CAM analyses. A further value, the "discrimination unit", is also derived. Both of these are explained below. The experiments described were not limited to the omental fractions obtained in the experiments described above. After the general molecular composition of the more effective fractions was determined to contain lipid-containing molecules, especially gangliosides, further lipid-containing molecules of a known type were used. It was unexpectedly found that many of these substances had a strong, unexpected angiogenic effect. In addition, trials were carried out using commercially available gangliosides in new combinations. Again, unexpected angiogenic properties were found. Of even greater interest is the fact that the mixtures with mixtures of different gangliosides had more than additive angiogenic effects.

WE. CAM analyses

Angiogenic properties of extracts, fractions and mixtures were determined by subjecting them to "Chick Embryo Chorioallan-toin membrane analysis", hereinafter called "CAM analyses".

In CAM analyses, fertile chicken eggs are used and involve the following steps: Production of eggs: Using a drill, 2 cm<2>shell removed from the fertilized egg on the 4th day after incubation. The opening is henceforth called "window". Cellophane tape seals the window against the outside environment. The eggs are then placed in a 37°C incubator for a further 4 days.

Preparation of plates: On the 8th day after incubation, 0.4 g of agarose and 10 ml of PBS were mixed and heated to 100°C in a small glass ampoule and then mixed at 50°C with a 2* BSA solution in PBS. The mixture (2* agarose plus 1* BSA o PBS) is kept warm in a water bath. Using a pipette, 20 - 40 ml of sample solution (e.g. extract, fractionate or mixture) was mixed with a drop of agar oseb land in nig under constant stirring. After the large disc has hardened by gelatinization, it is divided into four smaller discs.

Placement of the discs on membranes: On the 8th day after incubation, the discs were placed inside the eggs on CAM; areas on CAM with different degrees of blood vessel development were chosen. The selected area is approx. 1 cm away from the chick embryo but not so far away that the disc would lie outside the CAM or stick to the interior of the shell wall. The eggs are then incubated for a further 4 days. All instruments used are pre-washed with 98* ethanol.

Plastic discs: Plastic discs were prepared using a hole punch. After placing 2.5 ml of the sample solution on each disc, this was allowed to dry over a hot plate. Additional 2.5 ml amounts of the sample solution can be added to the disc and dried between applications. After the disc has been prepared, it is placed on the CAM as described above.

Evaluation of the effect: On the 12th day after incubation, the discs were located in the eggs and the windows were enlarged by breaking off pieces of the shell with tweezers. The eggs are then examined under a light microscope. The vascularization in the rest of the egg is compared to that surrounding the disc. Degrees of neovascularization in the direction of the disc are determined and compared with the effect of the discs in other eggs. The effects of each disc are rated on a 1:5 scale as follows: 1 = one or two small areas of increased branching around the disc; essentially negative; 2 = three or more small areas of increased branching around the disc; weak response; 3 = formation of a self-explanatory "wheel effect"; increased branching around the disc; moderate response; 4 = similar to "wheel action" with increased branching around the disk to a greater degree than No. 3, a strong response; and 5=equal to "pinwheel action" with extensive branching around the disc; a very strong response.

Plus and minus indications are also used with each numerical value so that a CAM analysis can have a value that goes from 1- (No response at all) to 5+ (exceptionally strong response with extensive branching.

Based on the above-mentioned assessment system, the so-called angiogenic index ("A" Index or "A" in the following tables) is determined. The A index allows comparative analyzes of samples, expressed by angiogenic effect.

The angiogenic index is defined as:

100 x the total assessment on individual analyzes maximum mul^g assessment

For example, if in a sample containing 12 CAM assays 7 are "weak", 1 is "moderate", none are "strong" and 4 are "negative", the A index will be calculated as follows:

Table III indicates data obtained by analyzing extracts and solvent distribution components obtained using procedures as indicated in Figs. 1, 2 and 3. The terminology used is the same as used in the figures.

The samples were obtained from feline, bovine, porcine and rabbit omenta as indicated.

Table IV below contains data corresponding to those in Tables I-III. However, the samples are all on ganglioside materials. The first group is ganglioside obtained from cat omental extracts. The gangliosides were separated into mono-, di-, and tri-sialylated components and were also mixed in 1:1 or 1:1:1 ratios. Corresponding analyzes were carried out with porcine omenta-derived glycosides.

The "Supelco" group presented analyzes for known, commercially available gangliosides (items 1-4 in this group). However, points 5-8 represented new ganglioside mixtures.

This table also presents a value for these substances, namely the "DU" or "discriminator value".

To determine the DU value, the A index values were noted as well as the percentage negative mean values sj and Sj for the compounds in class I, and sj and the Sjj values for class II. These numbers, sj, Sj, sjj and Sjj determine centroids of distribution of each class of compounds. Using this, the values were W^ and W2 and X^-p and X2T; "weight coefficient" determined via the equations

The smaller the DU value, the greater the angiogenic effect of the sample. An assessment of the DU values per connection from the best to the worst, is shown in Table V.. These results show that, while CMFr has angiogenic activity vis a vis the CAM analysis, the additional fractions obtained by following the process as indicated in Fig. 2 have better angiogenic properties . With reference to Table III, e.g. Cat CMFr, the first point, an A value of 28.98 while 46.34* of the samples were negative. In contrast, the purer monosialogangliosides obtained on the DEAE column showed an A value of 34.1 where only 39* were negative. In contrast, non-lipid fractions, also from DEAE columns, 23.6 and 67* negative, showed a drop despite purification. Finally, for this comparison, a mixture of mono-, di- and tricyalogangliosides from cat omentum showed values of 38.9* with only 13* negative.

Further comparisons can be drawn from the data given in Table III. The DU value, shown in Tables IV and V, is a useful indication of showing real efficiency, as it takes into account not only the A value but also the percent negative response. The lower the DU value, the more efficient the material to be tested. To thereby refer to Table VII, it can be seen that the new mixture of known gangliosides (Supelco mono-, di- and tri-sialogangliosides), and the fraction containing feline mono-, di- and tri-sialogangliosides, are the most effective mixtures.

These results can also be displayed graphically as shown in figures 6-16. These figures are linear categorization diagrams for different substances. In the case of linear categorization as used here, the angiogenic index is plotted against the percentage of negative response. A "centroid" or "median" point is obtained for each group of substances plotted in this way and the T value is obtained from a comparison of each two groups of compounds. This T value is then taken as an index according to which certain mixtures are more effective than others. Fig. 6 sets up the guidelines for the T values when using all examined samples. Then, in figures 7-16, different groups are brought up against the T values. Anything listed to the left of Tg suggests possibilities such as angiogenic mixing. Fig. 16 shows the best mixtures.

Fig. 17 is incorporated to show a graph of the angiogenic index plotted against an inverted negative percent standard using novel mixtures of known di- and tri-sialogangliosides. The diagram shows that the best mixture is di- and tri-sialogangliosides in a 1:2 ratio. This diagram is interesting because the curves obtained strictly correspond to those obtained when the antigen-antibody complex formation is plotted. This suggests that a complex reaction, not unlike a precipitation and agglutination that occurs with antigen-antibody systems, occurs.

The following experiments show that CMFr, described above, has in vivo efficacy against angiogenesis. The experiments are set forth in US-SN 672,624 of August 20, 1984. As can be seen and with reference to Tables III-VII, CMFr has a lower angiogenic index and a higher discrimination unit value than the additional fractions and mixtures similarly tested (ie the CAM analysis). The person skilled in the art will therefore see that it should be expected that these experiments can be repeated with the additional fractions and with expected superior results.

Other commercially available 1 lipid compounds acquired primarily from Supelio and Sigma Chemical or supplied by individual researchers were tested for their ability to induce angiogenesis by CAM. These trials are shown in Table VI.

The person skilled in the art will see that additional tissues, characterized by the presence of lipid-containing molecules, can be analyzed in this way to obtain potentially active fractionated lipid-containing mammalian tissues such as liver, brain, epithelial tissue and so on, as well as plant tissues, especially seeds. Plants are known as good sources of lecithins and angiogenically active lecithins can be found. Synthetically produced lipids can also be used.

A set of experiments was conducted to demonstrate the neovascularization effects of the non-aqueous lipid preparation at a site where the normal vascularization of the tissue had been intentionally destroyed. Adult female cats were anesthetized with an intramuscular injection of Ketamine using a dose of 7 ml/kg. Each cat was placed in a supine position and an incision made between the knee and the inguinal fold in both hind legs. The femoral arteries were isolated, ligated and removed between the groin and the first deep femoral branch (Hunter's canal). The incision was closed and each cat immediately submitted to an intravenous injection of a tin II chloride/Technesium-99 which binds to and radioactively labels the red blood cells in the tissue. The location and amount of this radionuclide can be identified using gamma camera scanning. For this purpose, blood vessels and capillaries carrying radioactivity labeled red blood cells are specifically visualized.

The tin II chloride/phosphate preparation contained 10 mg sodium pyrophosphate, 30 mg sodium trimetaphosphate and 0.95 mg tin II chloride. The preparation was reconstituted by adding 2.0 ml of PBS and 1.0 ml of this solution was injected intravenously into the bronchial vein of the cat. 20 min. later, an intravenous dose of lOm Curie Technesium-99m was injected to radiolabel red blood cells in the tissue area. Nuclear imaging scanning and digital integration of the blood flow were observed and pursued.

After the cats were surgically prepared, a "baseline" scan for background radioactivity of the surgical sites was performed, followed by injection of between 6 and 7 ml of the chloroform: methanol-extracted and evaporated viscous liquid lipid, suspended in PBS, intramuscularly in equal amounts to two preselected and marked sites on the right leg in the area where the femoral artery had been removed. A placebo injection containing only PBS was carried out in two corresponding identified and marked sites on the left leg. Under normal circumstances, the body's recognized response to this type of surgery would be to try to establish collateral blood circulation to the damaged tissues by forming new capillaries and blood vessels in the areas where the femoral artery was damaged. By following and comparing the degree and speed of new blood circulation in each leg after the operation, a direct and verifiable assessment of the angiogenic properties and potency of the chloroform-methanol extracted lipid preparation could be carried out accurately.

After intravenous injection of tin II chloride/technesium-99m preparation carried out in pre-selected sites on each leg and each leg was subjected to nuclear scanning on the third, sixth and ninth days after surgery. The results of these nuclear scans are shown in Figures 18-20 which exemplify the effect of the lipid fraction on neovascularization in a representative cat. The data presented show that the increase in blood vessel formation in the right leg of the cat (injected with the omental lipid preparation) and substantially higher integrated radioactivity counts than in the left (control) leg. At 72 hours after surgery, a 29.6* difference in radioactivity was observed; 6 days after surgery, a 38.2* increase in radioactivity was observed in the right leg compared to the left; and after 9 days the rate of neovascularization in the right leg showed a 65.8* increase compared to the left leg. The photographs in figures 18-20 give a visual indication of the significant differences in the formation of new blood vessels in the use of chloroform: methanol extracted lipid fraction. A graph shows a linear increase in radioactivity (as counts) when comparing lipid-injected bone vascularization versus saline-injected bone vascularization as indicated in Fig. 21. These data give a rate of 0.25* per hour increase in neovascularization in the right leg, compared to the left. This clearly shows the angiogenic effect of the lipid fraction as demonstrated by the significant increase in the formation of new blood vessels and vascular organization, and the structure in the right leg. However, these data overlook the possibility of a common systemic effect using lipid extract preparations that were shown to be effective in the following control trial.

In this additional control, another cat was surgically operated to remove the femoral arteries as described above. In this case, however, no injection was given. Gamma camera scan, carried out 3 and 6 days after the operative intervention, is shown in figures 22-24. The sweep of the right and left leg is shown in Fig. 22 where no discernible difference is visible in the collateral circulation of new blood cells after a 3-day duration. Fig. 23 shows a front view of a leg on the 6th day after surgery and Fig. 24 shows a back view on the same day. The scan indicated no difference in count between the two legs at any time after the procedure and a much lower degree of neovascularization compared to the previous trials. Thus, the neovascularization was noticeably less in this further experiment than in the left (control) leg in the previous works. In light of this and seen in light of the fact that in the previous experiments the cat's left leg (injected control) showed a relatively higher degree of neovascularization (although to a significantly lesser extent than in the right leg), there is reason to assume that the part of the lipid preparation in the right leg was probably systemically transferred to the left in the previous experiments.

The in vivo experiments using CMFr can be repeated with the different materials obtained by following hexane:ethanol extraction. As a comparison between these fractions and CMFr can be carried out from the data set out in Tables III - VII above, the person skilled in the art will conclude that these purified extracts will result in an even faster and better angiogenesis.

Preparations having angiogenically active lipid-containing molecules have been obtained from mammalian tissue. The compounds in therapeutically active amounts have also been shown to affect angiogenic activity in a way that was not previously expected. Tissues equivalent to the omentum such as lipid-containing mammalian tissue and plant tissue can be expected to have angiogenically active molecules as well. Synthetic lipids based on the structure of the molecules shown to be angiogenically active are also intended.

In practice, the compositions may be administered by any known standard means including intravenous, intramuscular, oral and topical routes of administration. The amount or dose will of course vary from patient to patient. Expressions and formulations used here are meant to be descriptive and not limiting, the person skilled in the art can therefore of course carry out modifications and changes, also terminological, without going outside the spirit and framework of the invention.

Claims (44)

1. Preparation with angiogenic effect, characterized in that it comprises at least one type of angiogenically active lipid-containing molecule in an angiogenically effective amount. 2. Preparation according to claim 1, characterized in that the mixture comprises gangliosides. 3. Preparation according to claim 1, characterized in that the mixture comprises monosialogangliosides. 4. Preparation according to claim 1, characterized in that the mixture comprises disialogangliosides. 5. Preparation according to claim 1, characterized in that the mixture comprises trisialogangliosides. 6. Preparation according to claim 1, characterized in that it comprises a mixture of mono-, di- and trisialogangliosides. 7. Mixture according to claim 1, characterized in that it comprises mono- and di-sialogangliosides. 8. Mixture according to claim 1, characterized in that it comprises mono- and tri-sialogangliosides. 9. Mixture according to claim 1, characterized in that it comprises di- and tri-sialogangliosides. 10. Mixture according to claim 9, characterized in that the di- and tri-sialogangliosides are present in a ratio of approx. 0 parts disialogangliosides: 3 parts trisialogangliosides to approx. 3 parts disialogangliosides:0 parts trisialoganglioside. 11. Mixture according to claim 9, characterized in that said di- and trisialogangliosides are present in a ratio of approx. 1 part disialoganglioside to approx. 2 parts trisialoganglioside. 12. Mixture according to claim 1, characterized in that it comprises phospholipids. 13. Mixture according to claim 1, characterized in that it comprises ceramides. 14. Mixture according to claim 1, characterized in that it comprises cerebrosides. 15. Mixture according to claim 1, characterized in that it comprises neutral lipids. 16. Mixture according to claim 1, characterized in that it comprises lecithin. 17. Mixture according to claim 1, characterized in that it comprises sphingosides. 18. Mixture according to claim 1, characterized in that it comprises lipid-containing molecules derived from animal tissue. 19. Mixture according to claim 18, characterized in that the lipid-containing molecule is derived from omental tissue. 20. Mixture according to claim 19, characterized in that the omental tissue is an omental tissue selected from the group comprising feline, bovine, porcine or rabbit omental tissue. 21. Mixture according to claim 20, characterized in that the omental tissue is feline omental tissue. 22. Mixture according to claim 1, characterized in that the mixture comprises lipid-containing material derived from plant tissue. 23. Mixture according to claim 22, characterized in that the plant tissue is a lecithin-containing tissue. 24. Mixture According to claim 1, characterized in that the molecules are produced synthetically. 25. Method for obtaining preparations with increased angiogenic activity comprising contacting a sample of a mammalian tissue with a first solvent under conditions that favor the extraction of angiogenically active lipid containing molecules, and further contacting this angiogenically active lipid containing first solvent with a second solvent under conditions favoring the extraction of the angiogenically active lipid-containing molecule into a second solvent, and non-angiogenic inhibitory substances, the first and second solvents being insoluble in each other. 26. Method according to claim 25, characterized in that the first solvent is a chloroform:methanol mixture. 27. Method according to claim 25, characterized in that the second solvent is a hexane:ethanol mixture. 28. Method according to claim 25, characterized in that the first solvent is a chloroform-methanol mixture and that the second solvent is a hexane:ethanol mixture. 29. Product according to claim 28. , 30. Method according to claim 25, characterized in that the angiogenically active lipid-containing second solvent is separated into the components hexane and ethanol phases. 31. Method according to claim 30, characterized in that the ethanol phase is divided into an upper and a lower phase by combination with a mixture of chloroform, methanol and water. 32. Product according to claim 31. 33. Method according to claim 31, characterized in that the upper phase is separated into two phases by combining the upper phase in sequence with a mixture of methanol and water, and a mixture of chloroform and methanol. 34. Product according to claim 33. 35. Method according to claim 33, characterized in that the chloroform-methanol part is mixed with a mixture of methanol, chloroform and water, and then a mixture of methanol, chloroform and sodium acetate, to form two separate parts. 36. Product according to claim 35. 37. Method according to claim 31, characterized in that the lower phase is sequentially mixed with chloroform, acetone and methanol to form three separate parts. 38. Product according to claim 37. 39. Sodium acetate containing the product according to claim-35. 40. Process for obtaining mixtures of angiogenically active lipid-containing molecules, characterized in that it comprises contacting a mammalian tissue sample with a mixture of hexane and ethanol solvent under conditions that favor the extraction of angiogenically active lipid-containing molecules in the solvent, removing the solvent, contacting the extract with a second solvent of hexane and ethanol under conditions favoring the extraction of the angiogenically active lipid-containing molecule into said second solvent, removing the hexane, contacting the ethanol extract with a solvent of chloroform, methanol and water, separating the resulting mixture into the upper and lower phases, contacting the upper phase with a mixture of methanol and water and a mixture of chloroform and methanol, separating the thus treated upper phase into a methanol-chloroform portion and a methanol-water portion , and contacting the methanol-chloroform portion with a mixt ing of chloroform, methanol and water and a mixture of chloroform, methanol and an acetate-containing compound and to separate the thus treated methanol-chloroform portion into a chloroform:methanol:water portion and an acetate-containing portion, and to purify these portions. 41. Method according to claim 40, characterized in that the lower phase is brought into contact with separate proportions of chloroform, acetone and methanol to give separate fractions. 42. Method for improving angiogenesis in a patient, characterized in that it comprises subjecting the patient to a therapeutically effective amount of the preparation according to claim 1. 43. Method for improving angiogenesis in a patient, characterized in that it comprises subjecting the patient to a therapeutically effective amount of the preparation according to claim 2. 44. Method according to claim 42, characterized in that the preparation is administered orally, intravenously or topically.