CA1291432C

CA1291432C - Reconstituted basement membrane complex with biological activity

Info

Publication number: CA1291432C
Application number: CA000516643A
Authority: CA
Inventors: Hynda Karen Kleinman; George Reilly Martin
Original assignee: US Department of Commerce
Current assignee: US Department of Commerce
Priority date: 1985-08-30
Filing date: 1986-08-22
Publication date: 1991-10-29
Anticipated expiration: 2008-10-29
Also published as: EP0218065B1; JPS62129222A; EP0218065A2; US4829000A; EP0218065A3; JPH0740932B2; ATE66247T1; DE3680851D1

Abstract

Abstract The present invention discloses a biologically active basement membrane composition. When polymerized under physiological conditions, the composition forms gel-like structures whose ultrastructure resembles interconnected thin sheets of the lamina densa zone of basement membrane. The major components of the composition include laminin, type IV collagen, heparan sulfate proteoglycan, entactin and nidogen. These components polymerize in constant proportions when redissolved and allowed to reconstitute. Molecular sieve studies on the soluble extract demonstate that laminin, entactin and nidogen are associated in a large but dissociable complex. The reconstituted matrix is biologically active and stimulates the growth and differentiation of a variety of cells, including epithelial cells, nerve cells, hair follicles and the like. The reconstituted matrix can also be used for determining metastatic potential of tumor cells and for isolating metastatic tumor cells.

Description

3~:

Description Reconstituted Basement Membrane Complex With Biological Activity BACKGROUND OF THE INVENTION

Technical Field The present invention relates generally to basement membrane complex. More particularly, the present invention relates to reconstituted, basement-membrane-derived extracellular substratum (matrigel~
which polymerizes on heating and promotes cell growth and differentiation in vitro and-in vivo.

Backqround Art ' Basement membranes are thin, but continuous sheets that separate epithelium from stroma and surround nerves, muscle fibers, smooth muscle cells and fat cells. Ba~sement membranes comprise type IV collagen, the glycoproteins laminin, entactin, nidogen and heparan sulfate proteoglycans. In various studies, these materials show a codistribution within both the lamina densa and ~ts extensions across the lamina lucida. In the electron microscope, the components appear as a network of 5 nm wide cords and their codistribution suggests that the formation of basement membrane occurs through the interaction of various components. Type IV collagen molecules form ,: ,, ~9~

intermolecular disulfide bonds and associate in a continuous network which can be visualized in basement membranes digested with plasmin (Inoue et al., J. Cell Biol. 97, 1524~1537, 1983).

Various components oE the basement membrane are known to interact with each other. In vitro studies with purified components show that laminin binds through its short chains to native but not to denatured type IV collagen and through a domain in its long chain to the heparan sulfate proteoglycan. Each of these basement membrane components is soluble. However, when these macromolecules are mixed together in vitro, they form a floccular precipitate containing laminin to type IV collagen to heparan sulfate proteoglycan in a ~ 0:.1 molar ratio (Kleinman et al, Biochemistry 22, 4969-4974, 1983). However, this precipitate lacks the resiliency and consistency expected of basement membranous structures.

Purified components of basement membrane have been used previously as a coating for cultured cells (Terranova et al., Cell 22:719; 1980). ~owever, such material was soluble and did not form a three dimensional matrix as has been achieved by the composition of the present invention.

Disclosure of the Invention It is, thereEore, an object of the present invention to provide a reconstituted, basement~
membrane-derived extracellular composition (matrigel) capable of polymerizing on heating and forming a three dimensional matrix which promotes cell growth and diEferentiation in vitro and in vivo.

It is a further object of the present invention to provide a method of preparing the "matrigel" and promoting cell growth and differentiation thereln.

It is a still further object of the present invention to prepare matrigel from human placental extract.

An other object of the present invention is to provide a method of determining metastatic potential of tumor cells and of isolating metastatic tumor cells.

Other objects and advantages of the present invention will become apparent as the detailed description of the invention proceeds.

Brief Description of the Drawi~

These and other objects, features and many of the attendant advantages of the invention will be better understood upon a reading of the following detailed description when considered in connection with the accompanying drawings wherein:

Fig. 1 shows the effect of type IV collagen, heparan sulfate proteoglycan, and heparan on the gelation of basement membrane components from the basement membrane extract. Increasing amounts of each component were added to 100 ~1 of the extract and incubated for one hour at 35 in 0.15M NaCl, O.OSM
Tris-HCl, pH 7.4c The samples were then centrifuged and insoluble material was dissolved in sample buffer. Equal aliquots of the samples were electrophoresed in 5~ acrylamide. Densitometric scans of negatives of photographs of the gels were used to ~ ~, ~9~

quantikate the amount of protein pelleted. (A) The effect of type IV collagen on the amount of total protein in the gel. (B) The quantitative ef~ects of type IV collagen, heparan sulfate proteoglycan, and heparan on the amount of total protein present in the gel;

Fig. 2 shows the effect of time and added type IV
collagen on the gelation of the basement membrane extract. The conditions are similar to those described in the legend for FIGURE 1. This ~igure compares gelation in the presence ( ~) and absence ( ~) of type IV collagen (50 ~g );

FigO 3 shows the effect of temperature on the gelation of extracts of basement membrane. The experiment was carried out in the presence of type IV
collagen (50 ~g) as described in the legend for FIGURE 2. Gelation was stopped by centrifugation at the times indicated;

Fig. 4 shows the ability o~ the basement membrane extract to regel following dissolution of the gel. The first lane, designated "whole extract" demonstrates the components in the starting material. The "first gel1' designates the components in the gel formed in the presence of type IV collagen. The material present in the gel for~ned in the absence of type IV collagen (not shown) was solubilized ~or 20 minukes in 2.0M
yuanidine, dialyzed against 0.05t~ Tris-HCl, p~l 7.4, containing O.l5M NaCl and allowed to regel in the absence (not shown) and presence o~ type IV collagen ("designated 2nd gel"). The cycle was repeated two additional times ("3rd" and "4th gels"). Shown are equal aliquots of the gels electrophoresed in a 5%

acrylamide gel;

Fig. 5 shows SEPHAROSE*4B column chromatography of the 2.OM urea extract. Two ml o~ the whole extract equilibrated in either 2M urea, 0.15M NaCl, 0.05M Tris-HCl, pH 7.4 (dissociative), or 0.5M NaCl, 0.05M Tris-HCl, pH 7.4 (associative), were placed on a SEPHAROSE*
4B column (2X60 cm) equilibrated in the corresponding buffer (A). Aliquots of the designated fractions from the extract chromatographed in associative (B) or dissociative (C) conditions were analyzed by SDS
polyacrylamide gels. In addition, an aliquot of the material eluting from the column run under associative conditions was examined in-the electron microscope by rotary shadowing (A). The electron micrographs show the most common co~plex in the peak fractions involves a central heparan sulfate proteo~lycan and numerous peripheral laminin molecules. Entactin and nidogen are smaller molecules and are not readily visualized in these complexes, but are known to be present.

Fig. 6 shows electron micrographs of reconstituted gels and an authentic basement membrane. (A) Gel formed in the absence of added type IV collagen or heparan sulfate proteoglycan. The gel conslsts of dispersed segments with occasional interconnections.
(B) Gel formed in the presence of added type IV
collagen and heparan sulfate proteoglycan. The edge of the gel is at the top. The gel consists of an interconnected network; the network is made up o~
structures which are similar in width to the lamina densa part of native basement membranes. These lamina densa-like structures vary somewhat in thickness. (C) Kidney tubule basement membrane from a 100 gm rat. The basement membrane consists of the lamina lucida and * Trademark lamina densa. Extensions from the lamina densa attach it to the cell membrane (arrowheads). Bar=200 ~m x 47,500; and Fig. 7 shows the effect of the basement membrane gel on the morphology and differentiation of sl6c3 melanoma cells in culture. Sterile 2M urea extract of the EHS-tumors in 0.15M NaCl, 0.05M Tris-HCl, p~ 7.4, was allowed to gel on the surface of a petri dish for 30 minutes at 37. Then equal numbers of cells were plated onto the gel (left) or onto control tissue culture-plastic dishes (right)~ After one week in culture in DMEM containing 20 mM tyrosine, gentamicin, glutamine and 5~ fetal calf serum, the cells were photographed. ~A) Morphology and assessment of melanogenesis by the cells. (B) Direct view of the dishes. The gel at the edge has been deflected to show that the cells are attached to it.

Fig. 8 îs Coomassie blue stain of 2.0 M urea extract of placenta after heparin affinity chromatography showing the purification of laminin;

Fig. 9 shows the effect oE human placental matrigel on neuronal process formation. The results show that the human matrigel strongly promotes nevrite outgrowth;

Fig. 10 is a diagra~atic representation of Tumor Cell Invasiveness Assay usincl matrigel; and Fig. 11 is a diagramatic representation of Invasive (Metastatic) Cell Selection using matrigel.

~9~43~

Best Mode for Carrying Out the Invention The above objects and advantages of the present invention are achieved by a basement-membrane-derived composition comprising a biologically active polymerizable extract containing in parts by weight about 60-85~ laminin, 5-30% collag`en IV, 1-10~ nidogen, 1-10% heparan sulfate proteoglycan and 1-5% entactin.
The term "biologically active" as used herein means capable of supporting normal growth and differentiation of various cell types when cultured including epithelial cells.

It has been discovered during the studies described herein that under physiological conditions certain components including type IV collagen~ laminin, heparan sulfate proteoglycan, nidogen, and entactin interact in rather constant proportions to form a gel with lamellar structures rese~.bling in dimensions those in basement membranes. Under the conditions described herein, each of these components is required for the reconstitution of the matrix. Without being bound to any theory, it is proposed that the components of the gel form supramolecular complexes, which may be intermediates in the formation of the matrix. The gel of the present invention is designated "matrigel" as a descriptive terminology.

The reconstituted matrix (matrigel) promotes the growth and di~ferentation of a variety of ceIls. In particular, the reconstituted basement membrane gel of the present invention is an excellent substrate for epithelial cells in culture. The matrigel of the present invention has also been demonstrated to promote J ~; _ 143~

cell adhesion, growth and differentiation of a multiplicity of cells including neurons, hepatocytes, sertoli cells, hair follicles, thyroid cells and the like. In addition, sertoli cells cultured within the gel have been subsequently transplanted back into the animal with good survival and maturation of the spermatids. The composition of the present inven~ion has also been found to promote nerve regeneration (optic and sciatic) in vivo ana allows for organ reconstitution as well. Preparing matrigel using an extract from human placenta also reduces the possi~i-lity of immunological interactiorl or rejection when such matrigel is used in humans.

Although any similar or equivalent methods and materials as described herein can be used for testing or the practice o~ the present invention, the preferred methods and materials are now described.

Materials - Type IV collagen, laminin and heparan sulfate proteoglycan were prepared from the EHS
(Engelbreth Holm-Swarm) tumor (Timpl et al, J. Biol.
Chem. 254:9933-9937; 1979; Hassell et al, Proc. Natl.
Acad. Sci. USA 77:4494~4498; 1980; Kleinman, et al, Biochemistry 21:6188-6193; 1982). After washing the tumor tissue in 3.4~ NaCl, 0.05 M Tris-HCl, pH 7.4, containing protease inhibitors (Orkin et al, J. Exp.
Med. J~ ?0a~220; 1977; Timpl et al, suDra), the basement membrane matrix was extracted with 0.5~1 NaC1 in 0.n5~1 Tris-HCl, pH 7.4. Laminin was isolated from the 0.5M, NaCl extract as described by Timpl et al, supra~ The residue of tumor tissue from lathyritic animals was sxtracted with 2.0~ guanidine in 0.05M

g Tris-HCl, pH 7O4~ followed by an extraction with the same buffer containing 0.005M dithiothreitol to solubilize the type IV collagen (Kleinman et al, ~ ). Low density heparan sulfate proteoglycan was puri~ied from 6.0M urea extracts of the tumor by ion exchan~e chromatography followed by cesium chloride density centrifugation and molecular sieve column chromatography (Hassell et al, supra). Heparan was obtained from Sigma Chemical Company.

Unfractionated extracts of the basement membrane matrix were prepared by treating the tissue which had been washed with high salt with an equal volume (1 ml/gm) of 2M urea, 0.05 M Tris-HCl, pH 7.4, overnight at 4 and centrifuging at lO,OOOg for 30 minutes. The residue was washed once with the same volume of buffer. Then the extract and wash were combined, dialyzed against 0.15~ NaCl in 0.05M Tris-HCl, pH 7.4 (TBS), and centrifuged to remove a small amount of insoluble material. The supernatant fraction was stored at -20C in small aliquots and used in the reconstitution assays described below. Using established, quantitative ELISA assays, this extract was found to contain laminin ~3.5 mg/ml), type IV
collagen (0.1 mg/ml) and heparan sulfate proteoglycan (0.1 mg/ml). Entac~in, nidogen, and other minor components were also present. For the column chromatography, the extract was dialyzed into 0.5M
NaC1, 0.05M Tris-HCl~ p~l 7.4, and centrifuged to remove insoluble material.

Reconstitution Assavs - Gelation was carried out in a centrifuge tube to which 0.05-0.1 ml of the 2M
urea extract was added in physiological buffer.
Purified components dissolved in 0.15M NaC1, 0.05M

~29~4~

Tris-HCl, pH 7.4, were added to the extract or were incubated together at the concentrations indicated.
The final volume was made up to 0.5 or 1.0 ml with 0.15M NaCl, 0.05M Tris-HCl/ pH 7.4, and the samples were incubated for 1 hour at 35C. Insoluble material was isolated by centri~ugation and the pellets were dissolved in sample buffer and electrophoresed in either 5% or 7.5% acrylamide under reducing conditions (Laemmli, 1970, Nature - London, 227:680-682). Each experiment was repeated a minimum of three times. The total amount of protein in the precipitate was determined by the standard Lowry procedure. The amount of nidogen and entactin in the gel was related to the total amount of material present~in the 400K band of laminin by scanning negatives of photographs of the gels in a Helena densitometer (Quick Scan Model, Helena ~ab Corp., Beaumont, Texas). Entactin and nido~en were identified based on their migration in SDS gels and cross reactivity in Western blot analyses with suitable antibodies. Type IV collagen in the gel was quantitated using 14C-labeled type IV collagsn and heparan sulfate proteoglycan was quantitated using 35S-sulfate labeled material of known specific activities in separate but parallel experiments.

Rotary Shadowing - The 2. OM urea extract equilibrated in 0. 5M NaCl, 0. 05M Tris-HCl, pH 7.4, was placed on a SEPHAROSE 4B column. An aliquot (30 ~1) of the peak raction ~0.1 m9/ml) eluting from the column was diluted with 300 ~1 of 0.155 M ammonium acetate, pH 7.4, and 600 ~] of glycerol. For rotary shadowing, the mixture was sprayed onto mica, shadowed with platinum-palladium, carbon coatedl and examined in a ~EOL 100C electron microscope.

.. , .. ~ .

~x~

Ultrastructure of Reconstitued Components - The gel was prepared essentially as described above.
Briefly, 0.2 ml of the extract was incubated alone or in the presence of type IV collagen and heparan sulfate proteoglycan overnight at 35~C. The gel was isolated by centrifugation and then fixed in 2.5~
glutaraldehyde, treated with 1% OsO4, block stained with 2~ uranyl acetate, and dehydrated. The gel was then processed through EPON*(Ladd Research Industries, Inc., Burlington, VT; LX-112 resin) for electron microscopy. Thin sections were stained with uranyl acetate-and lead citrate, and examined in a JEOL 100C
electron microscope. Thin sections of rat kidney tubule basement membranes were obtained as described by Laurie et al, (Am. J. Anat. 169:463-481; 1984).

Cell Culture - B16C3 cells were cultured either directly on tissue culture plastic or on a 1 mm thick basement membrane gel in a mixture of F12 medium and DMEM (Dulbecco's modified Eagle's medium, lacking phenol red for visualization of the pigmentation of the cells) containing glutamine, antibiotics, 20mM tyrosine and 5~ fetal calf serum. After one week, the cells were photographed.

The assembly of basement membrane components was analyzed using purified basement membrane components as well as unfractionated extracts of basement membrane.
Purified type IV collagen, laminin and heparan sulfate proteoglycan formed a flocculent precipitate when incubated under physiological conditions for one hour at 35. In contrast, a gel formed when urea extracts of basement membrane are dialyzed against physiological saline and then warmed to 35~ for one hour. The components of the gel were isolated by centrifugation * Trademark 9~

and examined by SDS gel electrophoresis. As shown in Fig. lA, the amount of laminin, entactin, and nidogen present in the gel increased in proportion to the amount of type IV collagen added until 50-60% of the material in the extract was incorporated into the gel. Heparan sulfate proteoglycan also caused increasing amounts of basement membrane components to precipitate (Fig. 1~). Separation by gel electrophoresis and quantitation of the major components in the gel indicated that constant ratios of laminin, entactin, and nidogen are obtained in the presence of added type IV collagen (Figure lA) or of heparan sulfate proteoglycan (Figure lB). When both type IV collagen (150 ~g ) and heparan sulfate proteoglycan (10 ~g ) were added to the extract, up to 80% of the protein in the incubation was incorporated into the gel. The smaller chain of laminin co-elec~rophoresed with the chains of typs IV collagen and prevented its visualization in the SDS gel.

To estimate the amount of type IV collagen in the gel, 3H-labeled type IV collagen of known specific activity was used and the amount of 3H-label in the precipitate was used as a measure of type IV
collagen. Likewise, the heparan sulfate prot~oglycan cannot be visualized in the gels and 35S-labeled heparan sulfate proteoglycan was used. These studies showed that laminin accounted for almost 60~
(264~56 ~g) of the material in the gel in a typical experiment~ whereas type IV collagen was 30%
(125~7 ~g), heparan sulfate proteoglycan was less than 2% (8~0.7 ~g), nidogen was 5~, and entactin was less than 1%. In contrast, supplementation of the extract with either type I collagen, ~ibronectin or heparan (Figure lb) did not cause any increased precipitation .

~2~

indicating that specific interactions are involved.
Removal of the protein core of the proteoglycan by incubation overnight with 0.5M NaCH destroyed its ability to induce polymerization suggesting that the protein portion of the proteoglycan is involved in binding to other components.

Under physiological conditions, the gelation process is complete within 20 minutes (Figure 2). The formation of t~e gel is strongly dependent on temperature with maximum polymerization at 35~C ~Figure 3). The lack of interaction at 50C suggests that thermal denaturation inactivates a critical constituent.

The stability of the gel to dissolution was examined by using various solvents. The gel was not dissolved by cold aqueous salt but was partially dissolved by acidic solutions (Table 1) and completely dissolved in guanidine or in urea solutions. This suggests that the components are linked by relatively strong non-covalent bonds. When the guanidine-dissolved gel was dialyzed against physiological buffers and warmed in the presence of type IV collagen, gel-like structures were reconstituted. This process could be repeated several times with similar proportions of laminin, nidogen, and entactin being deposited at each step as determined by SDS
polyacrylamide gels (Figure 4t. In the presence of added type IV collagen, re~ormation of the gel occurred more rapidly and greater amounts oE the components were deposited.

... .

TABLE l SOL~BILIZATION OF ~CONSTTTUTED BASEMENT MEMBRANES

Solvent ~ Solubilized 0.15M NaCl 0 0.5M NaCl 0.5M E~Ac 43 1.0M ~rea ~ Dithiothreitol 40 2.0M urea 73 2.0M Guanidine 97 All solutions except the 0.5M HAc were buffered with 0.05M Tris-HCl at pH 7.4. The solubilization was carried out at 24 for 20 minutas. After 20 minutes with frequent vortexing, the solutions were centrifuged and the pellets were redissolved and electrophoresed in SDS gels with a reducing agent. The relative amount of material in the pellets was determined by scanning the SDS gels.

3~

Determination was also made whether soluble complexes of base~ent components existed. When the urea extract was dialyzed free of urea and passed over a SEPHAROSE 4B column in O.SM NaCl (associative conditions), laminin, nidogen, and entactin eluted in a major included peak (Figure 5A & B). When the material in the major included peak was pooled and rerun over the same molecular sieve column in 4M guanidine (dissociative conditions), these components separated in the manner expected from their molecular weights (Figure 5A & C). These results indicated that there are str~ng but non-covalent bonds joining laminin, nidogen, and entactin in the complex. Rotary shadowing electron microscopy o~ the major included peak material confirmed the presence of soluble complexes (Figure 5A). The complexes involved the large proteoglycan which appears as a large globule due to collapse of the heparan sulfate side chains in this kind of preparation surrounded by several laminin molecules. The nidogen and entactin molecules could not be distinguished but are known to be in the complexes from SDS
polyacrylamide gels (Figure SB).

The ultrastructure of the reconstituted basement membrane either with or without type IV collagen and heparan sulfate proteoglycan was also examined. In the absence of added type IV collagen and heparan sulfate proteoglycan, the gel consisted of numerous widely separated thin, filamentous aggregates ~Figure 6~).
The addition of type IV collagen or of heparan sulfate pro~eoglycan plus type IV collagen (Figure 6B) resulted in the formation of thin sheets which were interconnected (Figure 6B) or were confluent. The individual se~ments of the network had an average width similar to that of the lamina densa of kidney tubule .:
:, basement membrane (Figure 6C). However, unlike native basement membranes in which lamina densa-like layers are arranged in parallel, such as for example the PYS
tumor basement membranes (Martinez-Hernandez et al, Lab Invest. 47:247-257, 1982) or Reichert's metnbrane (Inoue et al, J. Cell. Biol. 97:1524-1537, 1983), the la~ina densa-like structures were interconnected and did not form parallel multilamellar structures. At very high power in the electron microscope, each segment could be resolved into 5nm cords as previously described in other basement membranes (Incue et al, supra, Laurie et al, J. ~ell. Biol. 99, 78a; 1984).

The ~atrigel (reconstituted basement membrane~ was used to coat the surfaces o~ bacteriological petri dishes and tested as a substrate for the growth and differentiation of a variety o~ cells at different laboratories. Melanoma cells (B16C3) showed considerable differences in morphology when grown on the basement membrane gel as compared to tissue culture surfaces (Figure 7). Further, there was a much earlier and more extensive pigmentation of the cells on this substrate. Studies of other cells showed that endothelial cells formed tube-like structures on the gel and that hepatocytes survived longer on basement membrane gel substrates than on tissue culture plates or on type I collagen. In vivo, the basement membrane gel was Eound to promote peripheral nerve regeneration (Madison et al, 1985. Exptl. Neurol., 88:767 772).

Such studies indicate that the reconstituted basement mernbrane is a biologically active substrate which induces diverse cellular responses. Since it can support cell adhesion, growth and diEferentiation beyond that known Eor the individual components, without being bound to any theory it is postulated that the reconstituted basement membrane gel contains these molecules in a unique and active conformation.

EXAMPL~ - 1 Preparation of EHS Tumor Extract for Gelation and Cell Culture The procedure is based on about 1009 of tumor and all steps are carried out at 4C unless indicated otherwise.
1. Hornogenize tumor in 200 ml of 3.4 NaCl buffer comprising 3.4M NaCl (397g) 0.05M Tris (12.lg) 0.004M EDTA (3.0g) 0.002M NEM (N-ethylmaleimide; 0.5g).
Add H20 to 2 Liters, adjust pH to 7.4.
2. Centrifuge at 10,000 RPI~ for 15 minutes, discard supernatant.
3. Homogenize tumor residue in 3.4M NaCl buffer.

4. Centrifuge at 10,000 RPM for 15 minutes, discard supernatant.

5. Homogenize tumor residue in 3.4M NaCl buffer.

6. Centrifuge at 10,000 RPM for 15 minutes, discard supernatant.

7. Homogenize tumor in 100 ml of 2~S urea buffer comprising 2M urea (240g) 0.05M Tris (12.1g) O.l5M NaCl (189) Add H20 to 2 liters and adjust pH to 7.4.

8. Stirlovernight at 4C.

9. Centrifuge at 14,000 RPM for 20 minutes. Save supernatant.

~9~

10. Add 50 ml of 2M urea buffer to the kumor residue and homogenize.

11. Centrifuge at 14,000 RPM for 20 minutes. Save supernatants.

12. Combine supernatants and dialyze vs Tris-saline comprising 0.05M Tris (12.1g) 0.15M NaCl (18.0g) Add H20 to 2 liters adjust pH to 7.4.
Use a graduated cylinder for 1 lite~. Add 900 ml of Tris-saline ~5ml of chloroform (this is a sterili~ation step).

13. Dialyze 2 hours - rotate bags at end.

14. Change dialyses to Tris-saline alone.

15. Dialyze one more change with Tris-saline.

16. Last dialysis steps should be against medi~ salts such as DMEM or Dulbecco-Vogt or the like.

17. Inside of the bag is sterile. Render the outside of the bag sterile with alcohol, rinse hemostat and scissors in alcohol and empty bags in sterile hood into sterile containers. Aliquot as needed. Collagen IV
can be optionally added at this stage to the liquid phase in an amount ranging from about 0.1 to 1 mg/ml depending on the desired consistency or strength of the polymerized gel matrix. The thicker gels have been found to be more durable.

18. Immediately cool. Cryopreserve at ~20C if storage is desired. For gelation: pour extract into desired containers and warm (about 24-35C) for 30-120 minutes for polymerization. For a 35mm petri dish, use less than 1 ml of the extract. Spread it thin.

19. To use the gel as a cell culture substratum, add about 3 ml of suitable growth medlum on top of the polymerized gel obtained from step 18 and inoculate the medium with the dispersions of the cells which are 43~:

desired to be grown. Of course, the growth medium to be used will depend on-the type of the cell which is desired to be gro-~ni specific standard growth medium and conditions (e.g. C02 concentration, temperature, pH
and the like) for different types of cells being well known in the art.

An alternate procedure for promoting the growth of some cell types is to inoculate or disperse the cells in the cold liquid extract just before polymerization in step 18 and then proceed with polymerization and subsequ~nt steps the same as describsd in step l9. For example, hair follicle, sertoli cells and the like are apt to be better cultured if first dispersed in the liquid phase prior to polymerization whereas epithelial cells, exocrine acinar cells, sciatic nerve cell, spinal cord neuron, thyroid organ culture, and the like are better cultured on top of the polymerized gel.
. .

Human Placental Preparation of Basement Membrane Matrigel Extracts comparable in composition and in biological activity can also be obtained from human placenta using a process similar to that used for the EHS mouse tumor described herein. However, since placenta is not composed o pure basement membrane like the EHS mouse tumor, an additional step is necessary as described hereunder:
(a) Placenta is freed of cord and amnion.
(b) Placenta is then washed and homogenized in about 3.4 M NaCl in 0.05 M Tris-HCl, pH 7.4 containing standard protease inhibitors such as phenyl~ethyl sulfonyl fluoride, N-ethyl~aleimide, EDTA, ' ~. X9~4~

pepstatin and the like.
(c) The tissue residue is extracted overniyht at about 4C with an equal volume (g/ml) of 0.5 M NaCl in 0.0~ M Tris-HCl, pH 7.4 (d) The tissue after buffer extraction is washed with an equal volume of the same buffer and combined with the extract.
(e) The tissue residue is extracted overnight at about 4C with an equal volume (g/ml) of 2.0 M urea in 0.05 t~ Tris-HCl, pH 7.4~
(f) Both the 0.5 M NaCl extract and the 2.0 M urea extract are dialyzed against 0.02 M sodium phosphate buffer, pH 7.~ overnight at 4 C and the dialyzed samples are separately chromatographed on a heparin SEPHAROSE column equilibrated in 0.02 M sodium phosphate buffer, pH 7.4, containing 0.15 M NaCl. The bound material is eluted with 1.0 M NaCl and dialyzed into Eagle's minimal essential medium.

Characterization and Bloloqical Activit~ of Placenta Extract The placental extract before and after heparin sepharose chromatography were compared for the purity of the extract. To do so, the 0.5 M NaCl and 2.0 M
urea extracts and the bound materials from the heparin column were dialyzed against water, lyophillized, and electrophoresed in SDS pol~acrylamide gels. The samples were then skained with Coomassie blue for a pro~ile on the protein content and immunoreacted with anti-laminin antibodies after tran~fer to nitrocellulose. These results demonstrated that laminin, the major component of basement membranes, as of the mouse basement membrane preparation described herein, is present in the placenta extracts in intact ~,.~.,.; , , . ..

;3X

form as clearly demonstrated after heparin affinity chromatography ~Figure 8). Placental laminin contains chains of Mr=~00,000 and Mr=200,000 componentsc The biological activity of this material on neurite outgrowth was tested using NG108-15 neuroblastoma plus glioma hybrid cells in culture.
These cells respond rapidly (within 2 hours) to the extracts as well as to the heparin bound material by sending out long neuritic processes tFigure 9). The material to be tested is added in Eagle's minimal essenti~l medium lacking serum or some other culture medium along with freshly dissociated cells. After two hours on tissue culture plastic, extended processes are observed in the cells exposed to the placental material. Thus, the placenta materials have comparable activity to the murine tumor material in stimulating neurite process development.

Use of_Basement Membrane Material From Murine Tumor to Assay For Tumor Cell Invasiveness and To Select For Invasive Tumor Cells In order for all tumor cells to metastasize, they must enter the blood stream and then exit from it to grow at a distant site. Tumor cells must therefore adhere to, de(Jrade, and mi~rate through endothelial basement membranes in order to metastasi2e. These steps are critical in tumor cell metastasis. A uni~ue in vitro assay to measure these critical steps in the invasion process has now been devised. The assay is fast, quantitative, reproducible, and distinguishes between nonmetastatic and metastatic cells. This assay ernploys the murine reconstituted basement membrane described herein.

4~%

Assay for Tumor Cell Invasiveness .

A porous filter (Nucleopore) is placed inside a blind well Boyden chamber. The lower compartment contains an attractant'such as fibroblast conditioned medium or laminin. Fifty microliters or murine basement membrane extract are''placed on top of the filter in the upper compartment and allowed to poly~erize at 37C. Then cells in Eagle's minimal essential medium or some other suitable culture medium are added to the upper well and the entire chamber is incubated at 37C in 95~ air, 5% C02 for 5 hours.
During this time r the invasive cells adhere to the matrix, degrade the matrix, and migrate through the matrix and the porous holes in the filter. 'This process is diagrammatically shown in FIGURE 10. The number of cells which have invaded the matrix can be quantitated on the lower side of the filter, ~or example by direct counting in a microscope after the ceils have been stained with DIFQUICK* (Harelco).
Alternatively, if the cells are radiolabeled, they can be measured directly in a scintillation counter.

It was observed that cells which are known to be non metastatic, in vivo, do not invade the matrix, i.e., less than 5 cells per field are observed on the lower surace of the filter. Whereas, cells which are known to be metastatic ln vivo invade the matri~, i.e , greater than 10 cells per high power Eield are observed on the lower surface of the filter. More tha'n 10 murine and human tu~or cell lines of known metastatic potential have been tested and it was found that there is a dir'ect relationship between the ability of the cells to adhere to, degrade, and migrate through the * Trademark .

reconstituted murine basement membrane and their metastatic potential (Table II).

Table II
Tumorgenicity and Invasiveness of Human and Murine Cell Lines Tumor Formation Invasiveness In VivoIn Vi-tro 10Tl/2 fibroblasts No No NIH 3T3 fibroblasts No No NIH 3T3 transfected with RAS Yes Yes NIH 3T3 transfected with MOS Yes Yes NIH 3T3 transfected with SSV Yes - ~es NIH 3T3 transfected with MMSV Yes Yes B16 Fl melanoma Yes Yes B16 F10 melanoma Yes Yes B16 BL6 melanoma Yes Yes B16 Br2 melanoma Yes Yes K-1735 melanoma Cl 10 No No K-1735 melanoma Cl 10 Yes Yes MC-180 epidermoid carcinoma Yes Yes A-204 Rhabdomyosarcoma Yes Yes PA-l Teratocarcinoma Yes Yes PC-3 prostate carcinoma Yes Yes MALME 3m Melamona Yes Yes SW 620 Colon Adenocarcinoma Yes Yes MCF-7 breast carcinoma No No MCF-7 breast carcinoma & estradiol Yes Yes MCF~7 breast carcinoma & ras Yes Yes Cells were assayed for 5 hours in the Boyden Chamber assay described above. Cell lines which had less than 5 cells/field on the lower surace of the filter were considered non invasive (No), whereas cell lines with 10 cells or more/field on the lower surface were considered invaslve (Yes). The ability of the cell lines to form tumors ln vivo is available in published literature.
Invas:Lve ~Metastatic) Cell Selection Highly invasive tumor cells can also be.selected Eor and obtained in pure Eorm based on their ability to adhere, degrade, and migrate through the reconstituted basement membrane. Here the murine basement membrane extract is placed on a tissue culture dish ~0.5 ~.X 9~4~32 -2~-ml/35 mm diameter dish~ and allowed to polymerize for 30 minutes at 37 C. The cells are plated in a sterile manner in complete culture medium as required for the growth of the specific cells. After two days, the invasive cells attach to, degrade, and migrate through the matrix to the surface of the plastic dish where they are concentrated. This is shown in FIGURE 11.
The invasive cells on the plastic surface can be recovered after removal of the reconstituted basement membrane gel.

It-was found that non-metasta~ic cells are unable to complete this process and very few, if any, cells are observed in the Matrigel or on the plastic dish after 2 days (TableIII). Metastatic cells are able to adhere to, degrade, and migrate through the matrix to the surface of the culture dish in quantity. When the "selected" cells, i.e., those which have migrated through the matrix, are retested in the Tumor Cell Invasion Assay, a more homogeneous and invasive population of cells is observed.

Table III
Number of Cell Aggregates Invading Matrigel Before and After Selection in Matrigel Cells# of aggregates Unselected Cells Cl 10 ~non metastatic) 2~1 Cl 3 (low metastatic)6-~1 M2 (high metastatic)16~1 Selected cells Cl 10 19~1 M2 19~3 _ All cells were obtained from M.D. Anderson Hospital, Houston, TX. The selected cells are the parent lines. The selected cells are those from the parent line which have chemoinvaded the matrigel, been isolated and grown. Upon retesting their ability to invade the matrigel, they are much more invasive. Data are expressed as number of cell aggregates which have invaded the matrigel ~ 1 S.D.
It is understood that the examples and embodiments described herein are for illustrative purposes only and that various modiEications or changes in light thereof will be-suggested to persons skilled in the art and are to be included within the spirit and purview of this Application and the scope of the appended Claims.

Claims

1. A cell culture composition comprising an extract containing in parts by weight about 60-85%
laminin, 5-30% collagen IV, 1-10% nidogen, 1-10%
heparan sulfate proteoglycan and 1-5% entactin;
said composition being capable of polymerizing on heating and promoting cell growth and differentiation.

2. The composition of Claim 1 wherein said extract is derived from Engelbreth-Holm-Swarm tumor.

3. The composition of Claim 1 wherein said extract is derived from human placenta.

4. The composition of Claim 3 having three dimensional matrix upon polymerization.

5. The composition of Claim 4 promoting cell growth in vitro or in vivo.

6. The composition of Claim 5 capable of supporting growth and differentiation of epithelial cells.

7. The composition of Claim 5 capable of supporting neurite growth and differentiation.

8. A biologically active extract capable of promoting cell growth and differentiation derived from Engelbreth-Holm-Swarm tumor by a process carried out at about 4°C comprising the sequential steps of:

(a) homogenizing the tumor multiple times in a first buffer and discarding the soluble fraction after each homogenization so as to remove undesirable components;

(b) extracting the residual tumor in a urea buffer and keeping the extract stirred for about 16-18 hours;

(c) separating the extract from step (b) and saving the extract while repeating step (b) with the residual tumor fraction and again saving the extract resulting therefrom;

(d) combining the extracts from step (c) and dialyzing against a sterilizing buffer with multiple changes of the dialyzing buffer, and (e) utilizing the dialysate as a substratum for cell culture either by polymerizing the dialysate before or after inoculating the dialysate with the cell to be cultured.

9. The extract of Claim 8 wherein the first buffer is a 3.4M NaCl buffer having a pH of about 7.4.

10. The extract of Claim 8 wherein the urea buffer is a 2M urea buffer having a pH of about 7.4.

11. The extract of Claim 8 wherein the sterilizing buffer is Eagel's minimal essential medium containing sterilizing amount of chloroform.

12. The extract of Claim 8, wherein the dialysate is polymerized at about 24°-35°C.

13. The extract of Claim 8 capable of supporting growth and differentiation of epithelial cells.

14. The extract of Claim 8 capable of supporting growth and differentiation of neuritic cells.

15. A biologically active extract capable of promoting cell growth and differentiation derived from human placental tissue by a process carried out at about 4°C comprising the sequential steps of:

(a) homogenizing placental tissue multiple times in a first buffer discarding the soluble fraction after each homogenization so as to remove undesirable components;

(b) extracting the residual tissue in a urea buffer and keeping the extract stirred for about 16-18 hours;

(c) separating the extract from step (b) and saving the extract while repeating step (b) with the residual tissue fraction an again saving the extract resulting therefrom;

(d) combining the extracts from step (c) and transferring to a heparin affinity column to allow adsorption and then eluting the bound material with a salt solution of greater than physiological concentration;

(e) dialyzing the eluate against a sterilizing buffer with multiple changes of the sterilizing buffer; and (f) utilizing the dialysat as a substratum for cell culture either by polymerizing the dialysate before or after inoculating the dialysate with the cell to be cultured.

16. A method of determining metastatic potential of tumor cells, comprising the steps of:

(a) layering about 50 µl to 100 µl of an extract as defined in Claim 1 on a suspended nucleopore filter and allowing the extract to polymerize;

(b) placing a single cell layer of suspended tumor cells to be tested in a buffered medium on said polymerized layer and allowing said cells to attach and migrate through said polymerized extract at about 37°C for about 5 hours; and (c) then determining the presence of migrated cells on the opposite surface of the nucleopore filter.

17. A method of isolating metastatic tumor cells, comprising the steps of:

(a) layering 1 to 2 ml of the extract of Claim 1 on a tissue culture dish and allowing the extract to polymerize;

(b) placing a single cell layer of suspended tumor cells in a buffered medium on said polymerized layer;

(c) incubating the culture dish for about 48 hours at about 37°C;

(d) removing polymerized matrix from said culture dish; and (e) then allowing cells adhering to the culture dish to grow in a growth medium.