WO1987006589A1

WO1987006589A1 - Method for preparing contact-inhibitory factor

Info

Publication number: WO1987006589A1
Application number: PCT/US1987/000928
Authority: WO
Inventors: Martin J. Rosenberg; George Lipkin
Original assignee: New York University
Priority date: 1986-04-23
Filing date: 1987-04-21
Publication date: 1987-11-05
Also published as: EP0266414A4; JPS63503113A; EP0266414A1

Abstract

A method for preparing a biologically active factor that is capable of restoring contact-inhibition of growth in malignant cell type. The method comprises loading a mixture comprising culture medium conditioned by growth of a contact-inhibited cell culture and an ionic buffer to a chromatography column equilibrated with the same buffer and comprising a non-ionic hydrophobic affinity material as the adsorbant. The column is washed once with the same buffer until a peak of material elutes, a second time with the same buffer with decreased ionic strength which lacks sodium chloride until a second peak elutes, and eluting the factor with unbuffered purified water. The eluate is desalted by membrane filtration and dried. The method is conducted in the absence of a volatile precipitating agent.

Description

METHOD FOR PREPARING CONTACT-INHIBITORY FACTOR FIELD OF THE INVENTION

This invention pertains to a novel method for the preparation of a contact inhibitory factor obtained from the culture medium of contact inhibited cells. Another aspect of this invention pertains to a concentrated form of said factor with increased activity. Yet another aspect of this invention pertains to methods for the administration and use of said factor.

BACKGROUND OF THE INVENTION

In vitro, normal cells exhibit a form of growth control termed "contact-inhibition of growth", or "density- dependent inhibition of growth". These terms refer to a cellu lar interaction that is dependent on population density and limits the growth of normal cells. Normal cells in culture multiply until confluent monolayers are formed; then their proliferation is halted. This phenomenon occurs despite the fact that the culture medium in which the normal cells are growing is still capable of supporting growth at lower cell densities.

In contrast, transformed or malignant cells do not exhibit this type of growth control. Upon reaching confluence under identical growth and culture conditions, these cells wil continue to multiply and form ultilayered, disorganized cell groups. Although it is likely that contact inhibition is main tained by multiple factors, a diffusible substance, "contact inhibitory factor" (CIF) found in the culture medium of a . contact-inhibited culture of hamster revertant amelanotic me anoma cells has been observed to restore contact inhibition a variety of malignant cell lines.

The prior art describes several techniques for th

- extraction of crude CIF from biologically active fractions obtained from a 48- to 72-hour old, serum-free conditioned culture medium (CM) of said contact-inhibited hamster amelan tic melanoma cell line. Lipkin, G. et al., Proc. Nat^/1. Aca Sci.. (U.S.A.) 21:849-853, 1974, disclose extraction of CIF

₁₀ using molecular sieving on a Sephadex G-200 column. Based o the existence of multiple protein bands on a polyaery1amide gel, it was recognized that the crude extract was relatively impure.

U.S. Patent 4,307,082 to Rosenberg discloses a me

₁₅ thod for the extraction of CIF by sequentially eluting a fra tionating column comprising a nonionic hydrophobic affinity with buffers of decreasing ionic strength. Although this te nique yields a more highly purified CIF-containing fraction, both this procedure and the prior molecular sieving procedu

₂₀ described above result in relatively small yields of CIF. I addition, the hydrophobic interaction chromatography techni disclosed in U.S. Patent 4,307,082 required approximately 9 days between the first and the last steps of separation.

U.S. Patent 4,530,784 of Rosenberg discloses a me

₂₅ thod for extracting large quantities of biologically active from serum-free conditioned medium without the use of colum chromatography. The procedure involved mixing media condi¬ tioned by growth of a contact-inhibited cell culture therei together with a volatile nondenaturing precipitating agent,

30 such as methanol or ethanol. The precipitate formed by thi reaction was separated from the mixture and extracted with biologically acceptable ionic buffering agent. CIF activity found in the supernatant fraction. However, the CIF obtain via this non-columnar technique did not have the same high

35 degree of purity as the chromatographic isolate, and furthe purification would be necessary before administration to ma mals or humans. In order to test the efficacy of CIF in the treat¬ ment of malignancies in mammals and humans, a less time-cons ing yet efficient preparation method was necessary. In addi tion, a method of drug delivery was needed that would maximi the effects of the CIF administered to mammals or humans. T present inventors have unexpectedly found that a modified preparatory technique using hydrophobic column chromatograph on phenyl sepharose leads to a substantial purification of C while requiring only one day from start to finish. Moreover, they discovered that CIF activity is diminished after treatm with ethanol. Since all the prior art techniques involved a least one step where the CIF-containing material was exposed ethanol, substantial amounts of activity were lost during th procedures. The present inventors have also found that CIF ca be entrapped in artificial lipid vesicles, or liposomes, and still retain its biological activity. Such liposomes provid convenient drug delivery system for the depot, in situ, ad¬ ministration of CIF. The term "depot administration" means that the lo cally injected (i.e. in or near the tumor), liposome-entrapp CIF is released slowly, over a period of several days. Thus under these circumstances, the site of injection acts as a storage depot for CIF.

OBJECTS OF THE INVENTION

The present invention has several objects includi but not limited to, the following:

- to increase the understanding of the nature and role of CIF:

- to devise an improved method for preparing CIF without substantially inactivating it?

- to devise a method for stably incorporating CIF into liposomes;

- to devise a method for administration of CIF to mammals and humans so as to maximize the effects of CI in said individuals. - to decrease the tumor burden of a mammal.

These and other objects of the present invention will be apparent to those skilled in the art in light of th present description, accompanying claims, and appended drawings.

BRIEF DESCRIPTION OF THE DRAWING

Figure 1 is a plot of the rate at which K₂Cr0₄ or

CIF flows out of liposomes as a function of time. Figure 2 is a bar graph of the effectiveness of

NK-mediated cytolysis on CIF-treated and untreated tumor cells.

Figure 3 shows the morphology of the hamster amelanotic melanoma cell line RPMI 1846 (also available as ATCC No. CCL49) grown in RPMI 1640 medium containing 1% serum (a) , 1% serum + CM (b) or 10% serum and CM (c) .

Figure 4 is a graph of the effect of CM upon the serum dependent growth of RPMI 1846 cells in culture.

Triangles, RPMI 1846 cells in RPMI 1640 medium; circles, - RPMI 1846 cells in RPMI 1640 medium conditioned by non- contact inhibited hampster melanoma cells; squares, RPMI

1846 cells in RPMI 1640 medium conditioned by contact ^• inhibited FF cells, and containing CIF. Open symbols: growth in 10% calf serum. Closed symbols: growth in 1% calf serum.

Figure 5 is a linear plot of melanoma tumor volume, versus days of treatment with aqueous CIF. Square control; +, experimental.

Figure 6 is a semilog plot of melanoma tumor volume versus days of treatment with liposome-entrapped CI

Squares, control; +, experimental. _χ SUMMARY OF THE INVENTION

The present inventors have found that, in additio to restoring density-dependent growth to transformed cells,

CIF treatment can also restore anchorage- and serum-depen-

_ dence to melanoma cells. CIF was also able to induce expression of another phenotypic characteristic of benign human pigmen cells, e.g., the vitiligo-related surface antigens on hamster and mouse amelanotic melanoma cells. Furthermore, ₁₀ susceptibility to NK cell lysis was markedly decreased in tumor cells upon restoration of contact-inhibited-growth induced by CIF treatment.

An improved method of preparation of CIF is dis¬ closed herein. This method comprises column chromatography ₁₅ using phenyl sepharose and requires only one day from start to finish.

Another aspect of the present invention relates t a method for entrapping CIF into anionic ultilamellar lipi vesicles (liposomes) .

20 Another aspect of the present invention relates t a method of in situ administration of said liposome-entrap¬ ped CIF for the treatment of malignant melanoma, colon car¬ cinoma, mammary carcinoma and other types of malignancies. This method comprises the local (i.e. at or near the site o 5 the tumor or in the tumor directly) administration of lipsome-entrapped CIF. The material is released in a slow, sustained fashion, the liposomes acting as a "depot" for th delivery of CIF.

Yet another aspect of the present invention relates to a method of reducing the tumor burden in a mamma

30 comprising treatment with either aqueous or liposome-entrap ped CIF.

DETAILED DESCRIPTION OF THE INVENTION Source of CIF

A contact-inhibited cell line has been used to

35 produce the crude extract containing CIF that is used as th raw material for the present purification process. A pre- ferred cell line that can be used to obtain CIF-containg preparations was obtained during pigment transformation of a highly malignant, hamster amelanotic melanoma cell line (Roswell Park Memorial Institute, No. 1846; ATCC-CRL-1479) by nucleic acids derived from 7,12-dimethylbenzene-

(a)anthracene-induced (DMBA-induced) , benign, highly pig- mented blue nevi of Syrian hamsters. In practice, highly pigmented benign blue nevus tumors are induced in Syrian hamsters by topical application of a single dose of DMBA. These benign tumors correspond histologically to the cellu¬ lar blue nevi of humans. To obtain nucleic acids for incu¬ bation, 110 12-week old female Syrian hamsters were painted with 1% DMBA and observed for 8 to 12 months until maximum blue nevus formation occurred in each animal. At such times, all palpable nevi were carefully excised, trimmed of overlying skin and underlying subcutaneous fat, and frozen at -70^βC until ready for use. When sufficient material had been collected, the frozen, pooled nevi were used as a source of nucleic acids, the latter being extracted by the well-known phenol method.

Prior to use, the pooled nucleic acids were stored for up to six months at -20^βC. The purity and concentration of RNA and DNA were determined by employing UV spectroscopy and the Orcinol and Indol assays. Using conventional tissue culture and extraction procedures, the pooled nucleic acids from the benign melanotic (pigmented) nevus cells were added to tissue cultures of amelanotic (RPMI No. 1846) malignant melanoma cells. As a result of this operation, a new cell line (maintained as ATCC No. CRL 1479) was derived from the hamster amelanotic malignant melanoma cell line (RPMI 1846) . Details of the transformation are described in an article by Lipkin, G. in the Journal of Investigative Dermatology. 57: 49-65 (1971).

The new cell line contained amelanotic cells with stable, hereditary properties that differ from those of the parent malignant (RPMI 1846) cell line. The new cell line is available from the American Type Culture Collection as ATCC-CRL-1479. Although aneuploid, the amelanotic (nonpig- mented) transformant had the property of density-dependent inhibition of cell division and showed features of increased contact inhibition such as growth in monolayers of parallel

₅ oriented cells, decreased maximum plate density and increased adherence to the culture plate.

To obtain the starting materials for practice of the present invention, cultures of cell line ATCC-CRL-1479 were initiated and maintained in Falcon T60 plastic flasks

-₀ (No. 3024, Falcon Plastics, Los Angeles, CA) on RPMI 1640 medium (GIBCO, Grand Island, NY) containing a 10% fetal calf serum and antibiotics (penicillin, 100 units/ml; strepto¬ mycin, 100 mg/ml; gentamycin, 10 micrograms/ml; Fungizone [amphotericin], 2.5 micrograms/ml). Stock cultures were

-_ sub-cultured twice weekly.

Tissue cultures of the contact-inhibited cell line (ATCC CRL-1479). were grown in RPMI 1640 medium containing 10% fetal calf serum (Irvine Scientific, Santa Ana, CA.) and antibiotics until they achieved confluence. Cells were then

_₀ washed twice with phosphate-buffered saline (PBS, GIBCO, Grand Island, N.Y.) and re-fed with serum-free medium. After 48 hours, the serum-free conditioned medium (CM) con¬ taining CIF, was collected and centrifuged gently to remove cells and debris. Some of this CM was stored at 4'C for up

25 to 14 days prior to use in certain studies, while some was immediately lyophilized and stored as a dry powder until it was further processed by hydrophobic affinity chromatog¬ raphy.

30 Preparation of CIF

A phenyl sepharose (Pharmacia, Piscataway, NJ) column, 1.5 x 30cm, was equilibrated with 500ml of a buffer containing 4M NaCl (or 4M NH4HCO3, see below), lOmM Tris, lOmM EDTA at pH 8.0. The lyophilized CM was dissolved in

35 20ml of the same buffer and applied to the column. There¬ after, 500ml of the same buffer was run through the column until a peak (monitored at 280 nm) appeared (10ml fractions . were collected) . After return of the peak to baseline, a new elution buffer (500ml containing lOmM Tris and lOmM EDTA) was started and a second peak eluted. In a similar manner, a third peak was obtained using 500ml of unbuffered _c distilled water as the eluant.

In a preferred embodiment of the present inven¬ tion, ammonium bicarbonate (NH4HCO3 between about IM and 4M) can be used instead of NaCl, Tris HC1 and EDTA. The advan¬ tage here lies in the fact that both the ammonium and bicar-

₁₀ bonate ions can be easily removed from the solution by vo¬ latilization during lyophylization whereas the sodium ions must be removed by A icon filtration. In this case, the low ionic strength washing buffer contained lOmM N H4HCO3 pH 8.0. The loading, washing and eluting steps are carried out

₁₅ in the same manner as when NaCl is used as above.

Each fraction was desalted by Amicon filtration with a 10,000 Dalton membrane, then lyophilized and stored at 4^βC.

For bio-assay, each sample was dissolved in PBS.

₂₀ An aliquot was removed for a protein assay, culture medium added, and the mixture applied directly to the test cells. The contact-inhibitory activity was found in the third, i.e. relatively most hydrophobic peak. The activity of the CIF purified by the present procedure was between about 50 and

₂₅ about 100 micrograms/ml but the total yield was between 30 and 50mg. Preparation of Liposomes^'

Negatively charged multila ellar lipid vesicles (liposomes) were prepared as described below. Lipid ix-

_₀ tures containing phosphatidylcholine (GIBCO, Grand Island, NY), dicetyl phosphate (K & K Laboratories, Inc., Plainview, NY) , and cholesterol (Fisher Scientific Company, Spring¬ field, NJ) (molar ratios, 7:2:1, respectively) in chloroform were rotoevaporated in vacuo to form a thin lipid film in a

„_ round bottom flask. Phosphate buffered saline (Gibco, Grand Island, NY) with or without CIF (peak III from the phenyl sepharose column) (100-200 micrograms protein per ml) was . then added to the flask to a final lipid concentration of 15 micro oles/ml. The flask was then agitated for 10 minutes on a laboratory vortex to produce a suspension of multi- lamellar vesicles which were then allowed to swell for two

₅ hours at room temperature.

When liposomes are prepared in the presence of a solute, a portion of the solute is entrapped within the lipsomes but some remains in free solution; it is sometimes necessary to remove the free solute. Two methods were used

-₀ to achieve this: gel filtration and centrifugation, as noted below.

To demonstrate that the method of preparation was generating stable liposomes, two flasks with lipid film on their walls were prepared. An aliquot of a dilute solution

₁₅ of K2Crθ4 (0.29M) was added to the first flask, and an iden¬ tical aliquot of K2Cr0₄ containing added CIF at a concentra¬ tion of 100-200 microgra s protein/ml was added to the second flask. Liposomes were then generated as described above. In this case, liposomes were separated from entrap-

₂₀ ped solute by gel filtration on Sepharose 2B (Pharmacia, Piscataway, NJ) . In each case, the most concentrated liposome fraction, which eluted in 2 ml was divided into equal aliquots. Triton-X-100 (Sigma Chemical Company, St. Louis, MO) detergent was added to one aliquot to disrupt the

₂₅ liposomes in order to permit determination of the total

K₂Cr0 concentration spectrophotometrically by absorbance at 370nm. K₂Cr0 concentration was determined by the method of Schieren et al. (Biochem. Bioohys. Acta. 542: 137-153, 1978; the entire disclosure of this article is incorporated

₃₀ by reference here) as follows: 0.8-0.9 ml aliquots of the liposomes was diluted to 1 ml with phosphate buffered saline (PBS, pH 7.3-7.4). The samples (1 ml) were dialyzed in 0.25 inch tubing against 5 ml of PBS at 37"C in a water bath. The tubing was progressively moved to fresh 5 ml solutions at 30 minute intervals for an elapsed time of 90 minutes. Chromate anion release was measured by its absorbance at 370 nm directly from the dialysate at each interval. g. s ows e oss o c rom um ons rom an liposomes (closed circles) or CIF-containing liposomes (open circles) . After an initial accelerated decline, the curve assumes a relatively flat slope on a se ilogarithmic scale. Half of the K2Crθ4 was released from CIF-containing liposomes in 77 hours; an equivalent amount of K₂Crθ4 was released from control liposomes in 40 hours. Thus, the association of CIF with liposomes reduced the efflux rate of

Liposomes with __2Crθ4 were prepared through the swelling stage as outlined above. However, passage of these liposomes through Sepharose 2B leads to a significant reduc¬ tion in the final concentration of liposomes when compared to that in the initial swell solution. Thus, in order to

__ insure a maximally high concentration of liposomes for addi- 15 tion to cultures, an alternative method for the final prep¬ aration of liposomes was selected. The liposome-suspension was centrifuged (2,000 g, 20 in, 4'C) to separate the unas- sociated solute. Pellets were gently re-suspended in PBS •and added to tissue culture medium at appropriate dilutions.

20

Method of Reducing the Tumor Burden in a Mammal by Treatment with Aqueous or Liposome-Encapsulated CIF

When cancer cells metastasize, they migrate from the primary site of the tumor, enter the blood stream and

25 invade other organs in the body. This can occur even though the patient is undergoing treatment for the primary cancer. Therefore, what is needed is an adjunct to traditional chemotherapy and radiation treatments which can forestall the spread of cancer to other regions of the body.

30

This can be accomplished, for example, by the intravenous (i.v.) administration of either aqueous or, preferably liposome-entrapped CIF. As is demonstrated below in Example 5, both aqueous and liposome-entrapped CIF were effective in arresting the early growth of melanoma in

35 hamsters. Liposome-entrapped CIF had a substantially longer-lasting effect. Such CIF treatments will broadly comprise i.v. administration of 100 mg to 1000 mg of CIF per kg body weight. The number of treatments will vary with the severity of disease and the type of CIF employed: e.g., liposome-entrapped CIF is released more slowly than aqueous CIF and the effects are longer lasting (see Example 5 below) .

The present invention is further described below by reference to specific examples, which are intended to illustrate the present invention without limiting its scope. Example 1

Tumor Cells Treated with CIF Are Much Less Susceptible to NK Cell Lysis

Natural killer (NK) cells recognize and lyse most neoplastic, virus-infected and fetal cells, but do not attack their normal or fully differentiated counterparts. NK-mediated cell lysis is predicated on cell-to-cell con¬ tact, and it is likely that some surface configuration or substance on the target cells is responsible for recogni¬ tion. To date, no cell surface property has been identifie that is shared by all neoplastic cells or that is invariabl associated with cells subject to NK lysis. Like NK cell recognition, contact inhibition of growth is also presump¬ tively dependent on cell surface signals that are operative when the cells have made contact, i.e., reached confluence. The availability of CIF afforded the inventors the oppor¬ tunity to determine whether CIF-induced surface changes which made the cells capable of contact inhibition would also render them resistant to attack by NK cells. Two huma colon carcinoma cell lines, HT-29 (ATCC HTB-38) and SK CO-l (ATCC HTB-39) , the human erythroleukemia cell line K562 (ATCC CCL-243) , and three human melanoma cell lines, SK23 (obtained from Dr. J. Fogh, Sloan-Kettering Cancer Insti¬ tute) , Rob and HN-54 (obtained from Dr. D. Zucker-Franklin, N.Y.U., Medical Center) (or ATCC HTB-70 could have been use instead) were used in this study; all had been found to be sensitive targets for NK cells. Cultures were maintained (except line K562) in 150 cm² polystyrene culture flasks (Corning, Corning NY) at 37^βC in a humidified atmosphere containing 5% CO₂. Medium consisted of RPMI 1640 with 10% heat-inactivated fetal calf serum and antibiotics. Stock cultures were refed and subcultured weekly. Each cell line to be tested (except K562) was seeded in duplicate 25 cm² polystyrene flasks at an initial density of 100,000 cells per flask in complete medium including 10% heat-inactivated (56^βC, lhr) fetal calf serum. Following cell attachment and spreading after 2 to 3 hours, the medium was replaced, with one of each pair of flasks receiving the usual complete medium with 10% fetal calf serum and antibiotics, and the other receiving 48 hour con¬ ditioned medium (CM) with fresh 10% fetal calf serum and antibiotics added. In the case of K562 cells, which grow in suspension, cells were placed directly in either regular or CIF-containing medium at the same initial density of 100,000 cells per flask.

In order to assay for NK cell cytotoxicity, the cytolysis of attached target cells by NK cells was assayed by the method of Bean et al.. Natl. Cancer Inst. Mongr.

37-41. 1973, incorporated by reference, in which targets are pre-labelled with ³H-proline. Target cells (2,000 per well) plated in 96-well Costar plates (Data Packaging Corp., Cambridge, MA) were incubated with NK cells in ratios of 1:100 and 1:150. After 24 hours, the lysed cells were removed by gentle washing of the wells with Hanks Balanced Salt solution (GIBCO, Grand Island, NY) (HBSS) . Cytolysis was determined by counting the radioactivity of the remain¬ ing attached cells. Cytolysis of K562 cells was evaluated by the well-known ⁵¹Cr release assay. Following labelling of 5 x 10⁶ K562 with 100 micro Ci of Na₂ ⁵¹Cr0₄ (Amersham Corp., Arlington Hgts., Illinois) in one ml RPMI 1640 for one hour at 37^βC, the cells were washed twice in the same medium. Subsequently, they were exposed to NK cells in ratios of 1:25 and 1:50 for 4 hours at 37"C in a 5% C0₂ atmosphere. After incubation, 100 microliters of the supernatants were removed and radioactivity was measured in a gamma counter. Each experiment was performed at least 3 times and the average of 4 determinations was recorded.

NK cells were isolated as follows: Mononuclear cells were isolated by Ficoll-Hypaque (Pharmacia, Piscataway, NJ) gradient centrifugation of 100 ml heparin- ized blood obtained from healthy donors. Monocytes and other adherent cells were eliminated by Lymphocyte Separat¬ ing Reagent (LSR) (Technicon Products, New York, NY) . The isolated non-adherent cells were washed twice in HBSS and resuspended in RPMI 1640. They were centrifuged in a dis¬ continuous Percoll gradient (Pharmacia, Piscataway, NJ) ranging from 37.5% (V/V) to 52.5% with 2.5% diminution steps. The cells at the 47.5% to 50% interface were identi¬ fied as large granular lymphocytes enriched for NK cells, following which they were washed and resuspended in RPMI 1640 with 10% fetal calf serum.

The.results of the above-described cytotoxicity assay are presented in Fig. 2. As is shown in the bar-graph in Fig. 2, NK cell-enriched fractions were able to kill 70-85% of three different human melanoma cell lines which had been used as targets (Fig.2, A, C, and E) . However, after these melanoma targets had been cultured for 5 days in the presence of CIF, they had developed a marked resistance to cytolysis by NK cells (Fig.2, B, D,and F) . Inspection of cultures which had been grown in the presence of CIF revealed the morphologic alterations which are consistently associated in these melanoma lines with a restoration of contact inhibition of growth. This was evidenced by conflu¬ ent, well-oriented monolayers of the uniformly bipolar cells rather than the multilayered overgrowth of pleomorphic cells seen in cultures grown in the absence of CIF. Human colon carcinoma cell lines, HT-29 and SK CO-l also proved to be good targets for NK cells before exposure to CIF (Fig. 2, G and I) . Following culture with CIF, their susceptibility to lysis by NK cells was reduced by 70-90%(Fig. 2, H, and J) . To rule out the possibility that CIF was toxic to NK cells, NK cell fractions were incubated in media containing CIF for one hour. Following washing, the NK cells were incubated with targets as routine. In addition, NK cells were incu¬ bated with untreated control target cells in media contain ing CIF for 24 hours. This did not reduce cytolysis. Thu CIF treatment of NK cells did not in any way alter their ability to kill the targets used in this study.

Treatment of K562 cells with CIF also rendered this target cell line resistant to NK cell lysis (Fig. 2, and L) . The above effects of CIF were reversible. When

CIFtreated target cells were washed and cultured in media lacking CIF for 24 hours, they resumed growth which was no contact- inhibited and they became as vulnerable to NK cel lysis as before their exposure to CIF. Thus, CIF produced transient, reversible alteration of tumor cells which renders them more "normal" to NK cells.

Example 2: Induction of Pigment Cell Differentiation Anti gens in Melanoma Cells by CIF

Vitiligo is a disease characterized by the absen of melanocytes, which causes hypopigmented areas in the patient's skin. Most vitiligo sera contain antibodies to surface antigens on normal human melanocytes but not to human, hamster or mouse amelanotic melanoma cells. Thus, the antigens reacting with vitiligo serum appear to be mar ers of pigment cell terminal differentiation. Therefore, the inventors investigated whether an additional phenotypi characteristic of benign differentiated pigment cells migh also be induced by CIF in malignant melanoma cells.

Hamster (RPMI 1846) and mouse (F106LR) melanoma cell lines were each seeded in 8-chamber Lab-Tek slide-wel (Miles Laboratories, Elkhart, IN) at initial densities of 3,500 cells per well (for RPMI 1846 cells) or 5,000 per we (for F106LR cells) in complete medium including 10% calf serum. Following cell attachment and spreading after 2 hours, the media were replaced with half of the wells receiving the usual complete medium with 10% calf serum an antibiotics, and half receiving 48-hour conditioned medium (containing CIF activity) with fresh 10% calf serum and antibiotics. In some cases, medium conditioned for 48 hours by the highly malignant, non-contact inhibited RPMI 1846 cells with fresh 10% calf serum and antibiotics added, was used as the replacement medium in wells containing hamster melanoma cells. Media were changed on day 3 if cultures had neither reached confluence nor undergone morphologic rever¬ sion. When either contact-inhibited monolayers or hyper- confluent cultures had formed (at 3 days for F106LR and at 7 days for RPMI 1846) , the slides were processed for immunolo¬ gical studies as described below.

Antigens were assayed by Enzyme Linked Immunosorb¬ ent Assay (ELISA) performed on whole cells as a quantitative measure of antibody reaction to reverted and non-reverted hamster cells. Normal and CIF-treated cells were counted and equal volumes were added to wells of flexible, dispos¬ able, polyvinyl chloride microtiter plates (Dynatech Labora¬ tories, Alexandria, VA) . Fifty microliters of vitiligo sera (obtained from Dr. Bystryn, the Department of Dermatology, New York University School of Medicine, New York, NY) dilu¬ ted 1:5, 1:10, 1:20, 1:40 with a 1% solution of bovine serum albumin (BSA) in phosphate buffered saline (PBS), pH 7.4, were added in duplicate to wells of microtiter plates con¬ taining a fixed number of cells, and allowed to incubate for 30 minutes at room temperature. Three vitiligo sera, previ¬ ously shown to have high titers of anti-melanocyte anti¬ bodies, and three normal human sera were used. Wells were washed three times with phosphate bufferred saline (PBS) containing 0.5% Tween 20 (Sigma Chemical Company, St. Louis, MO) prior to addition of peroxidase-conjugated goat anti-human IgG (Cappel Laboratories, Cochranville, PA) dilu¬ ted 1:10 with a 1% solution of BSA in PBS. After a 30- minute room temperature incubation, excess labeled antibody was removed by 5 washes with 0.05%.Tween 20 in BSA and 75 microliters of ortho-phenylenediamine were added to each row. Following a 30-minute incubation at 37*C, 25 microli¬ ters of 8N sulfuric acid was added to each well, stopping the enzyme activity. Optical density was determined for each well at 492 nm using a Titertek Multiskan ELISA Reader (Flow Laboratories, McLean, VA) .

The use of ELISA for cell-associated antigens on intact cells has been well described in the art and found to be a highly sensitive method of detection for specific antibody-antigen reactions. Absorbance (O.D.) values ob¬ tained after reaction of vitiligo and normal human sera with reverted and non-reverted cells are listed in Table I below.

TABLE I

ELISA using vitiligo (V) and normal (N) sera on CIF-treated and untreated hamster melanoma cells

Cells Serum Sera Peroxidas Dilution Control

VI V2 V3 Nl N2 N3

RlrMI 1846 1:5 0.69 0.64 0.41 0.19 0.30 0.24 0.08 cells 1:10 0.53 0.55 0.44 0.21 0.23 0.17 0.06 treated with 1:20 0.45 0.48 0.38 0.14 0.07 0.05 0.06

CIF 1:40 0.30 0.25 0.16 0.08 0.06 0.03 0.02

RIMI 1846 1:5 0.31 0.19 0.10 0.30 0.29 0.27 0.09 cells 1:10 0.22 0.17 0.11 0.27 0.29 0.22 0.04 iBitreated 1:20 0.06 0.07 0.09 0.19 0.10 0.11 0.01 control 1:40 0.02 0.05 0.05 0.07 0.09 0.08 0.01

A linear relationship was seen between concentra¬ tions of test sera utilized and O.D. values, indicating that this method could serve to assay cell antigen-antibody reac¬ tions in a quantitative manner. Highest values were noted with CIF-treated RPMI 1846 cells. The latter and vitiligo sera showed a sequential decrease in O.D. readings ranging from 0.69 to 0.16 units (mean 0.58 at 1:5 dilution down to 0.24 at 1:40 dilution). By contrast, control hamster RPMI 1846 cells incubated with vitiligo sera exhibited minimal antigen-antibody interaction, with O.D. values ranging between 0.31 to 0.02 units (0.20 at 1:5 dilution down to 0.04 at 1:40 dilution). A similar low reactivity was seen with both CIF-treated and untreated RPMI 1846 cells incubat¬ ed with normal human sera (mean 0.24 down to 0.06 for CIF-treated cells, 0.29 down to 0.08 for untreated cells). No significant background staining due to non-specific bind¬ ing of peroxidase to the conjugated IgG occurred, as illus¬ trated by O.D. values of 0.08 to 0.01 resulting from incuba¬ tion of cells with conjugate in the absence of sera. These

_ results substantiate and further demonstrate that CIF in- duced the expression of cell surface antigens not normally seen on hamster malignant melanoma cells.

The induction by CIF of pigment cell antigens characteristic of differentiated human melanocytic cells, on ._ both hamster and mouse amelanotic melanoma cells, demon¬ strates that the effects of CIF transcend species barriers. It also provides further evidence of the potential reversi¬ bility of the malignant phenotype. Example 3: CIF Restores Anchorage and Serum Dependence to 1 ,5_ Melanoma Cell*—s.

Various characteristics of the transformed pheno¬ type can be correlated reliably with the capacity for neo¬ plastic growth in vivo. The properties analyzed have in¬ cluded morthology, saturation density, fibrinolytic

_2Q activity, nutrient uptake, fibronectin production, lectin agglutinability, surface structure, cytoskeleton, cell shape, serum requirement for growth and anchorage depen¬ dence. Loss of the requirement for attachment to a sub¬ strate in order to grow (anchorage dependence), has, despite

₂_ exceptions, been shown repeatedly to be significantly cor¬ related with malignant behavior in vivo. Another property of most transformed cells, the ability to grow progressively in 1% serum, is also significantly correlated with in vivo growth of transformed hamster cell lines.

Therefore, the ability of CIF to restore both anchorage and serum dependence to hamster melanoma cells was examined.

In order to test for anchorage dependence, agarose cultures were prepared using a micro-well technique. The underlayer contained 0.5% agarose, the overlayer 0.33%. For experimental wells, phenyl sepharose-fractionated material containing CIF (prepared as above) was incorporated into the overlay mixture at concentrations ranging from 0 to 1,000 micrograms protein per ml (Bio Rad Laboratories, Rockville Centre, NY) . The warmed overlay mixture was added to a pellet of RMPI 1846 hamster cells to give a final concentra- tion of 50,000 cells per ml; 0.2 ml of this warm mixture was then layered over 0.2 ml of semi-solidified underlay in each well to give a final concentration of 10,000 cells per well, and allowed to harden. All cells were then observed daily and scored at 6 days for clones of 8 or more cells. At the 0 same time that wells were being prepared for anchorage dependence studies, similar aliquots of the phenyl sepharose-fractionated material were also tested over the entire concentration range, on freshly-plated RPMI 1846 hamster cells in Falcon Microtest II wells for effects on 5 cell viability. The latter was judged by the well-known dye exclusion method using trypan blue both immediately after exposure of cells to CIF and 24 hours later.

The results are presented below in Table II ₀ TABLE II

Protein(micrograms/ml) Cloning efficiency

0 51.5

67.5 38.0

125.0 25.0 ₅ 250.0 13.5

500.0 9.7

1000.0 4.9

In wells containing CIF, clonal growth was inhib- _Q ited in a concentration-dependent manner. At the highest concentration (1,000 micrograms per ml), about 95% of cells failed to proliferate, only 4.9% forming aggregates of 8 or more cells. Control wells lacking CIF exhibited excellent clonal growth of RPMI 1846 cells, with 51.5% of cells for - ₅ ing aggregates of 8 cells or more. Simultaneous testing of the phenyl sepharose fractionated material used in preparing the agarose showed all concentrations to be non-toxic to RPMI 1846 cells as judged by dye exclusion (90% viability at the highest concentration immediately after plating, 88% after 24 hours) .

In order to test for serum dependence, RPMI 1846 cells were plated out in Falcon Microtest II wells at ini¬ tial concentrations of 5,000 cells per 0.2ml complete medium per well. Following attachment of cells after two hours, cells were washed with PBS, and media were replaced to define six experimental groups: cultures containing RPMI 1640, antibiotics and 1% or 10% calf serum; those cultures with CIF-containing CM, antibiotics and 1% or 10% calf serum; and those cultures with mock CIF-containing CM (pre¬ pared from non-contact-inhibited RPMI 1846 cells) , antibio¬ tics and 1% or 10% calf serum. All groups were refed after three days. Growth and morphology were observed for up to six days.

Results are shown in Figs. 3 and 4. RPMI 1846 cells grew well in 10% serum in the presence of control medium (Fig. 3A, Fig. 4) or in 10% serum in the presence of ςiF-containing medium (Fig. 3C,Fig. 4). By.contrast, RPMI 1846 cells grown in 1% serum-containing CIF conditioned medium failed to proliferate at all (Fig. 3B, Fig. 4) . RPMI 1846 cells grew to much higher densities in 10% calf serum then in 1% calf serum with both the control medium and CIF-containing medium, but did exhibit the usual relative reduction in saturation density when grown to confluence in 10% calf serum and CIF-containing CM.

As a further control, conditioned medium obtained from non-contact inhibited RPMI 1846 cells (not containing CIF) was tested. Although growth in 10% serum was inhibited over control values, this inhibition was not as great as that found using CM from CRL-1479 cells. Moreover, growth in 1% serum was the same as control cells when CM from RPMI 1846 cells was tested. CIF, previously shown to mediate density dependent growth of melanoma cells of human, mouse and hamster ori¬ gins, has now been found, in hamster melanoma, to restore ., anchorage and serum dependence as well. Loss of one or more of these properties is a prominent accompaniment of the transformed phenotype; the ability of CIF to restore all three suggests that it is acting as a central or critical point to modulate growth control. That such a mechanism for regulation of growth may be a general one is suggested by the lack of species or tissue restrictions upon the biologic effects of hamster CIF. Example 4: Transfer of Capacity for Density-Dependent

₁₀ Growth to Melanoma Cells By Liposome-Entrapped CIF

In order to test the efficacy of liposome- entrapped CIF for the treatment of malignant melanoma, liposomes containing CIF were prepared as described above. RPMI 1846 melanoma cells were first plated in

₁₅ complete medium from stock culture into wells of a Falcon microtest II plate at subconfluent densities of 10,000 per well. Following cell attachment after 2-3 hours, some of the wells were refed with complete medium containing lipo¬ somes (14.6 micrograms of lipid per ml) with entrapped CIF

₂₀ (14 to 112 micrograms of protein per ml at final dilution in medium) . A second group of wells received complete medium with blank liposomes containing identical concentrations of lipids. A third group was refed with complete medium con¬ taining added CIF (25, 50, 100 or 200 micrograms protein per

₂₅ ml) without liposomes. Finally, a fourth group received only fresh medium without CIF or liposomes. All cultures were refed once after 48 hours with fresh aliquots of their respective media. Cultures were observed daily for one week for morphologic and growth patterns. The CIF used in this

_ experiment was prepared using phenyl sepharose as described

I above.

All cultures (except those receiving plain CIF at 200 mg per ml, which showed toxicity) , grew to confluence. Control cultures receiving complete growth medium only, as __ well as those receiving blank liposomes, displayed essen- tially similar morphology, e.g., multilayered, high-density overgrowth of pleomorphic, disoriented melanoma cells. By contrast, those cultures receiving either aqueous CIF

(25-100 micrograms protein per ml) or liposome-entrapped CIF (at 1:8 and 1:4 dilutions, equivalent to protein concentra¬ tions of 56-112 micrograms per ml) , underwent the character- istic morphologic and growth pattern changes previously seen with aqueous CIF treatment, that is, the formation of cohesive monolayers of well oriented, flattened, uniform bi¬ polar cells which failed to pile up but instead stopped proliferating at confluence with saturation densities 45% ₀ lower than in untreated controls at the 1:8 dilution. Fur¬ thermore, the earliest observed effect, change of cell shape, was noted sooner (after 24 hours) in the case of aqueous CIF, while requiring 48 hours in cultures receiving liposome-entrapped CIF. The contact-inhibitory effects ₅ persisted for five days after the refeeding on day 2 in the liposome CIF cultures, but only for three days in the cul¬ tures refed on day 2 with aqueous CIF. Thus, entrapment of CIF resulted in some delay in the appearance of the characteristic changes, but also produced a more prolonged 0 ^effect-

The ability of CIF-bearing liposomes to effec¬ tively restore density-dependent growth in vitro coupled with the observed longer latency and more sustained duration of CIF effects, make this an ideal system in which to study

₂₅ the in vivo effects of this potent growth inhibitor on mela¬ noma and other neoplasms.

Example 5: Treatment of Melanoma in Hamsters By Depot Administration of Aqueous or CIF-Encapsulated CIF

All of the data presented above involved the in

3_Q vitro effects of CIF on transformed cells. In order to examine the in vivo efficacy of CIF, an in situ model of hamster melanoma was employed. The highly malignant hamster amelanotic melanoma line RPMI 1846 was inoculated into the dorsal skin of each hamster at an initial tumor burden of

-,_ 40,000 cells in 0.2 ml of phosphate-buffered saline (PBS). Four groups of animals were used in each experiment (6 animals/group) . Beginning at day 0 each of the four groups received 1 ml of either (a) PBS (control) , (b) PBS 4- CIF (1000 mg protein/ml), (c) liposomes (control), or (d) lipo¬ somes + CIF (1000 mg protein/ml) . The injections were given n situ around the tumor cells, either 3 times weekly [for the aqueous groups (a) and (b)], or 2 times weekly [for the liposome groups (c) and (d)]. Injections were continued for only 30 days in the experiment illustrated in Figure 6, but for up to 9 weeks in the aqueous group of experiments illus¬ trated in Figure 5. The growth of tumors was measured with ια calipers 1-3 times weekly, with volume calculated according to the formula, V = (lxw )/2, where V = volume, 1 = length of tumor, w = width of tumor. Both the aqueous-CIF (Figure 5) and liposome-CIF (Figure 6) groups showed marked retarda¬ tion of tumor growth compared to controls receiving carrier

15 vehicles alone. Despite cessation of treatment after 30 days, the liposome-CIF group of 6 animals all experienced subsequent resolution and disappearance of their tumor nodules, with 100% survival to a normal life span. The aqueous-CIF group.subsequently experienced regrowth of their 0 suppressed tumors, but only after a long interval of 1 month after discontinuation of treatment.

5

0

5

Claims

WHAT IS CLAIMED IS:

1. A method for preparing a biologically active factor that is capable of restoring contact-inhibition of growth in malignant cell types, said method comprising the steps of:

loading a mixture comprising culture media condi¬ tioned by growth of a contact-inhibited cell cul¬ ture, and an ionic buffer to a chromatography column equilibrated with the same buffer and com¬ prising a nonionic hydrophobic affinity material as the adsorbant;

washing the column once with the same buffer until a peak of material elutes; a second time with the same buffer with decreased ionic strength, said second buffer being free of sodium chloride until the second peak elutes; and eluting said factor with unbuffered purified water;

desalting said eluate by membrane filtration; and

drying said filtrate containing said factor;

said method being conducted in the absence of a volatile precipitating agent.

2. The method of claim 1, wherein said hydro¬ phobic affinity material is phenyl sepharose.

3. The method of claim 1, wherein said loading buffer comprises 4M NaCl, 10 mM Tris-HCl, and 10 mM disodium ethylenedia ine tetraacetic acid at pH 8.0.

4. The method of claim 1, wherein said loading buffer comprises between about IM and about 4M NH4HCO3 and lOmM disodium ethylenediamine tetracetic acid.

5. The method of claim 1, wherein said buffer of decreased ionic strength comprises 10 mM Tris-HCl and 10 mM disodium ethylenediamine tetraacetic acid at pH 8.0.

6. The method of claim 1, wherein said drying step comprises lyophilization.

7. The method of claim 1 wherein said conditioned medium is serum-free.

8. The method of claim 1, wherein said condi¬ tioned medium is obtained from a 48- to 72-hour old culture of contact-inhibited cells maintained in the absence of serum.

9. The method of claim 8, wherein said cells are ATCC CRL-1479 cells.

10. Biologically active factor capable of restor¬ ing contact-inhibition of growth in malignant cells prepared by the method of claim 1.

11. A method of restoring contact-inhibition of growth to malignant melanoma tumor cells vitro comprising the steps of:

providing contact-inhibitory factor produced by the method of claim 1 entrapped in liposomes;

exposing a malignant melanoma tumor to said lipo¬ somes by placing said liposomes in the immediate vicinity of said tumor, said liposomes being capable of releasing sufficient CIF to restore contact inhibition in said tumor cells, said lipo¬ somes and said tumor being immersed in a culture medium; and waiting for said factor to be released from said liposomes and inhibit growth in said cells in vitro.

12. A method for restoring contact-inhibition of growth in malignant melanoma tumor cells in a mammal com¬ prising:

providing contact inhibitory factor produced in accordance with claim 1 and entrapped in lipo¬ somes; and

depositing an amount of said liposomes in the tissue of said mammal in the immediate vicinity of said tumor, said amount being sufficient to re¬ lease contact-inhibitory factor in an amount suf¬ ficient to restore contact inhibition in the cells of said tum or.

13. A method for reducing the tumor burden in a mammal in need of such treatment comprising parenterally administering to said mammal a composition comprising a tumor-burden-reducing- effective amount of contact-inhibi¬ tory factor and a pharmaceutically acceptable carrier.

14. The method of claim 13 wherein said composi¬ tion comprises contact-inhibitory factor in an aqueous form¬ ulation.

15. The method of claim 13 wherein said composi¬ tion comprises contact-inhibitory factor entrapped in lipo¬ somes.

16. A method for managing tumor growth in a mam¬ mal in need of such treatment comprising the steps of: restoring contact inhibition of growth in the cells of said tumor by administering to said mam¬ mal a composition comprising a contact-inhibitory- effective amount of contact-inhibitory factor; and

administering to said mammal a tumor-regressive effective amount of an anticancer agent selected from the group consisting of chemotherapeutic agents and radiation.

17. The method of claim 16, wherein said factor has been prepared in accordance with the method of claim 1.

18. The method of claim 16 wherein said composi¬ tion comprises contact-inhibitory factor in an aqueous form¬ ulation.

19. The method of claim 16 wherein said composi¬ tion comprises contact-inhibitory factor entrapped in lipo- so e.

20. A composition for reducing the tumor burden in a mammal comprising a tumor-burden-reducing effective amount of contact-inhibitory factor and a pharmaceutically acceptable carrier or diluent.

21. The composition of claim 20 wherein said composition comprises contact-inhibitory factor in an aqueous formulation.

22. The composition of claim 20 wherein said composition comprises contact-inhibitory factor entrapped in liposomes.

23. The method of claim 13 wherein said adminis¬ tration is via intravenous injection.

24. The method of claim 13 wherein said amount of contact-inhibitory factor is between about 100 mg and 1000 mg per kg body weight.