direct cloning of human ovarian carcinoma cells in agar1 · tumor stem cells in soft agar (8, 9)....

[CANCER RESEARCH 38, 3438-3444, October 1978]0008-5472/78/0038-OOOOS02.00

Direct Cloning of Human Ovarian Carcinoma Cells in Agar1

Anne W. Hamburger, Sydney E. Salmon,2 Mary B. Kim, Jeff M. Trent, Barbara J. Soehnlen, David S.Alberts, and H. Jack Schmidt

Section ol Hematology and Oncology, Department of Internal Medicine Â¡A.W. H., S. E. S., M. B. K., J. M. T., B. J. S., D. S. A.Â¡,Department ol Pathology Â¡J.M. T.], Department ol Obstetrics and Gynecology [H. J. S./, The Cancer Center [S. Â£.S.Â¡, University of Arizona, College of Medicine, Arizona HealthSciences Center, Tucson, Arizona 85724

ABSTRACT

We have recently developed an in vitro assay for humantumor stem cells that permits cloning of human ovarianadenocarcinoma cells in soft agar. Tumor colonies grewfrom both effusions and biopsies from 85% of 31 ovariancancer patients. The cloning efficiency did not vary withthe histology of the tumor. Growth was induced withmedium conditioned by the adherent spleen cells ofmineral oil-primed BALB/c mice. Up to 2000 colonies

appeared after 10 to 14 days in culture, yielding a platingefficiency of 0.001 to 1%. Cells from nonmalignant effusions did not form colonies under these conditions. Thenumber of tumor colonies was proportional to the numberof cells plated between concentrations of 10" to 106 cells/

dish.Morphological and histochemical criteria showed that

the colonies consisted of cells with the same characteristics as those of the original tumor. Results of cytogenicstudies were also consistent with a malignant origin forthe tumor colonies with marked hyperdiploidy in coloniesfrom four patients and hypodiploidy in a fifth patient.

[3H]Thymidine and hydroxyurea suicide indices pro

vided evidence that in most cases a high proportion ofovarian tumor colony-forming cells were actively in transitthrough the cell cycle. Removal of phagocytic macrophages with carbonyl iron markedly reduced the platingefficiency, and 2-mercaptoethanol could only partially

substitute for macrophages.The assay appears useful for screening differential

cytotoxic effects of specific anticancer drugs (such as cis-

platinum) against the tumor stem cells from various ovarian cancer patients.

INTRODUCTION

We have recently demonstrated the ability to clone humantumor stem cells in soft agar (8, 9). In vitro colony-forming

assays for stem cells from transplantable murine myelomashave been predictive of therapeutic response in vivo (19,20). In applied studies of the bioassay, a range of in vitrosensitivities to various chemotherapeutic drugs has beenobserved. We are currently testing the predictive value ofour assay.

The ability to clone ovarian cancer cells in soft agar couldhave major importance. Epithelial tumors of the ovary arenow the most common fatal gynecological cancer in the

United States. Their incidence has tripled in the last 40years, yet the 5-year survival has not changed appreciably

during that period (29). Despite this the growth of ovariantumors in vitro has not been extensively pursued. Only afew groups have developed the ovarian cell lines that areessential to study the kinetic, biochemical, and immuno-chemical properties of these tumors (7, 10, 12, 30).

We now report on the application of our bioassay methodto the study of ovarian carcinoma cells. This paper describes the growth of the cells in vitro, their morphology,karyology, and the factors controlling their growth.

MATERIALS AND METHODS

Patient Studies. Patients with well-documented ovarian

carcinomas were selected for this study. The internationalsystem for clinical staging of ovarian cancer was utilized(26). Patients with ovarian cancer all had epithelial typecancers.

Collection of Cells. Malignant effusions were collected inheparinized vacuum bottles (100 units/ml). After centrifu-gation at 150 x g for 10 min, the cells were collected andwashed twice in HBSS3 (Grand Island Biological Co., GrandIsland, N. Y.) with 10% heat-inactivated PCS (Flow Labora

tories, Rockville, Md.). Tumor nodules obtained immediately after surgery were mechanically dissociated underaseptic conditions in a laminar flow hood. Tumors wereminced with a scalpel and then teased apart with needles.Cells were filtered through sterile gauze to remove cellclumps passed through 25-gauge needles, and then washed

by centrifugation as described previously. The viable nucleated cell counts determined in a hemocytometer withtrypan blue were routinely more than 90% if samples wereobtained within 2 hr of surgery. Only samples with morethan 90% viability were studied. Differential counts wereperformed on slides prepared with a cytocentrifuge andstained by the Papanicoulau (7) and/or Wright-Giemsa

methods (28).Culture Assay for Ovarian Colony-forming Cells. Cells

were cultured as described by Hamburger and Salmon (9).One ml under layers containing 0.25 ml of Millipore-filtered

medium conditioned by the adherent spleen cells of mineraloil-primed BALB/c mice in 0.5% agar were prepared in 35-

mm plastic Petri dishes. Cells to be tested were suspendedin 0.3% agar in enriched CMRL 1066 medium (Grand IslandBiological Co.) with 15% horse serum (Flow Laboratories).In this standard assay system, 2-ME was added to a final

1Supported by Grants CA 21839 and CA 17094 from the National CancerInstitute.

2To whom requests for reprints should be addressed.Received May 5, 1978; accepted July 19, 1978.

3The abbreviations used are: HBSS, Hank's balanced salt solution; PCS,fetal calf serum; 2-ME, 2-mercaptoethanol; PAS, periodic acid-Schiff,[3H]dThd, [3H]thymidine.

3438 CANCER RESEARCH VOL. 38

Research. on December 10, 2020. © 1978 American Association for Cancercancerres.aacrjournals.org Downloaded from

http://cancerres.aacrjournals.org/

Cloning of Ovarian Cells

concentration of 50 /U.M.However, the requirement for 2-ME

for ovarian carcinoma was specifically studied as an experimental variable and was found not to be necessary. Eachculture received 2 x 105 cells in 1 ml of agarmediummixture. Cultures were incubated at 37Â°in a 7.5% CO.,-

humidified atmosphere.Scoring of Cultures. Cultures were examined with a

Nikon MS inverted-phase microscope at x100 and x200.

Final colony counts were made 10 to 14 days after plating.Aggregates of 30 or more cells were considered colonies.Individual colonies were removed from the dishes with afine capillary pipet and were suspended in a drop of heat-inactivated PCS and deposited on slides that were air-driedfor 3 to 4 hr. Cells were stained routinely with Wright-

Giemsa, Papanicoulau, and methyl green pyronin stains (9)for morphology and for peroxidase, nonspecific esterase,oil red O, and PAS reactivity (16). Cells picked from colonieswere compared to the tumor cells in the original cellsuspension with the help of a clinical pathologist.

Determination of Percentage of Cells in DMA Synthesisby [3H]dThd or Hydroxyurea Suicide. The [3H]dThd suicide

method was used to measure the proportion of ovariancolony-forming cells in the DNA synthetic (S) phase of thecell cycle (13). Samples of 2 x 106 cells suspended in HBSSand 10% heat-inactivated PCS were added to 1.0 ml ofHBSS containing 40 /Â¿Ciof [3H]dThd (23 Ci/mmol, Amer-

sham/Searle, Arlington Heights, III.). Control samples wereadded to HBSS. Cell suspensions were incubated for 30min at 37Â°and washed twice with 20 ml of cold HBSS

containing 100 /xg/ml of unlabeled dThd and 10% PCS.Each suicide and control suspension was cultured in 4replicate plates at a concentration of 5 x 105 cells/plate.Alternatively, 2 x 106 cells in 1 ml of HBSS and 10% heat-

inactivated PCS were incubated with 76 /Â¿gof hydroxyureafor 1 hr at 37Â°.Cells were then washed twice in HBSS and

plated as described.Drug Sensitivity Testing. Single cell suspensions pre

pared as described previously for initial plating were adjusted to a final concentration of 1 x 106 cells/ml in HBSSand 10% PCS in the desired concentration of c/s-platinum

diamine dihydrochloride (National Cancer Institute). Cellswere incubated with drugs for 1 hr at 37Â°.Cells were then

centrifuged at 150 x g for 10 min, washed twice in HBSSand prepared for culture. For quantitative comparison of invitro curves, a "sensitivity index" was defined as the area

under linear survival-concentration curves to an upperboundary concentration. For c/s-platinum, the boundary

concentration was 0.06 /Â¿g/ml.Removal of Phagocytic Cells. One x 107 cells obtained

from effusions were incubated in 15 ml of McCoy's Medium

5A with 10% heat-inactivated PCS and 80 mg of dry-heat-

sterilized carbonyl iron (Tridom Fluka, Hauppage, N. Y.) in250-ml Falcon flasks. After incubation at 37Â°for 45 min in a

shaking water bath, the flasks were placed flat down on amagnet, and supernatant cells were carefully decanted andkept for subsequent study. This step was repeated as oftenas necessary (generally 4 to 5 times) to remove all the ironpowder and iron-laden phagocytic cells. The cells in the

supernatant were considered nonphagocytic. Cells werewashed twice in HBSS and 10% heat-inactivated FCS. Thisprocedure removed between 30 and 60% of cells.

Cytogenetic Analysis. Agar cultures were incubated at37Â°in the presence of 2.5 ml of Ham's Medium F-12 (Grand

Island Biological Co.) containing 0.1 /Â¿Mcolchicine. Four to15 hr later, cultures were removed from the incubator, andthe overlying medium was vigorously agitated to dislodgecolonies from the agar. The medium was removed andcentrifuged for 5 min at 150 x g. The supernatant wascarefully removed, and the pelleted colonies were resus-pended in 0.075 M KCI for 10 min at 37Â°.They were then

recentrifuged for 5 min, and the supernatant was discarded.Five ml of fresh, cold fixative (3:1 absolute methanol toglacial acetic acid) was added, and the suspension wasagitated with a vortex. They were then washed twice morewith fixative and stored at -9Â°.

Slides were prepared by placing 2 to 3 drops of thecolony suspension onto precleaned slides that had beendipped in 50% acetic acid. A sharp burst of air was thendelivered to each slide to further disperse the cells. Slideswere then air-dried for a minimum of 48 hr. Metaphasespreads were stained with 3% Giemsa (Gurr's R-66; Reboz,

Washington, D. C.) for 5 to 7 min. G banding was performedwith the technique of Sun ef al. (25).

RESULTS

Development and Identification of Colonies. Cell doublings were usually observed within 24 hr of plating andclusters of 8 to 20 cells appeared within 3 to 8 days.Individual cells appeared to increase in size within 24 hr ofplating. Colonies (collections of 30 or more cells) appeared10 to 14 days after plating. Cell lysis generally occurred 21days after plating. Cultures were not refed, and the averagelife of a culture was 3 weeks. No attempt was made tosubculture the colonies. During the first 5 days of incubation, there was a progressive increase in the number ofcells that commenced proliferation. Colonies consisted of30 to several hundred large (30 Â¿Â¿M)round cells. Many cellswere vacuolated with a characteristic signet ring appearance. Cells in large colonies were tightly packed withoutthe free cell layer at the periphery that is characteristic ofmacrophage colonies. The morphology of the colonies didnot vary with histological type (Fig. 1).

The number of colonies ranged up to 2000/plate, yieldinga maximum plating efficiency of up to 1%. A linear relationship was obtained between the number of nucleated cellsplated and the number of colonies found after 14 days(Chart 1). This line back extrapolated through zero, indicating origin of the colonies from a single clonogenic cell.Origin of colonies from single cells could also be observeddirectly by serial inverted-phase microscopy.

When cells from individual colonies were plucked fromagar and stained by the Wright-Giemsa and Papanicoulau

methods, they appeared to have the same morphologicalcharacteristics as did tumor cells in the original suspension.

In general, cells from colonies grown from patients withserous adenocarcinoma were oval with pale-staining cyto

plasm. The nucleus was pleomorphic with one or morenucleoli. Multinucleated cells were occasionally present.Cells usually contained a moderate amount of PAS-positive

material and occasional oil red O+ granules (Fig. 20).Mucinous adenocarcinoma cells were larger, polyhedral,

OCTOBER 1978 3439



A. W. Hamburger et al.

700

600

sao

200

100

Relationship Between Number ofOvarian Colonies and Cells Plated

1 5 10 20 50 XÂ»Number of Nucleated Cells Plated Per Dish (xio4)

Chart 1. Linear relationship between colony formation and the number ofnucleated effusion cells plated. Points, mean of 4 dishes Â±S.E. These arethe results of a typical experiment. Three other trials on different patientshave yielded similar results.

and frequently contained secretory vacuoles. Their nucleiwere often small, irregularly shaped, and hyperchromaticwhen pushed to the periphery by the secretory vacuoles.Special staining revealed PAS-positive material within the

vacuoles (Fig. 2a).Endometrioid cancer cells often showed PAS-positive

material and were large with large nuclei and prominentnucleoli.

Poorly differentiated cells, although pleomorphic withatypical nuclei, still showed faintly staining, PAS-positive

material. However, many cells contained vacuoles that didnot stain for either oil red O or PAS. Multinucleated cellswere often seen.

However, morphological features of cells from differentpatients with the same histological type varied. In addition,the appearance of cells from the same patient varied fromcolony to colony to a lesser degree. Moreover, the morphology and secretory activities of the tumor cells changedunder culture conditions. Quite often, cells from patientswith serous adenocarcinoma were more positive for PAS oroil red O stains after culture. Individual cells appearedlarger and more vacuolated than did cells in the originalsuspension.

Factors Affecting Cell Growth. Substitution of type 0human RBC for BALB/c spleen cell-conditioned medium

permitted cell growth in all but one case (Table 1). Substitution of medium conditioned by adherent spleen cells ofCD-1, DBA/2, or BALB/c mice that were not oil primed

decreased the number of ovarian colonies. Conditionedmedia from WI-38 cells, MA-184 cells (bone marrow fibro-

blasts), or primary expiants of human skin fibroblasts didnot support cell growth (Table 1).

Cell-free autologous effusions at a dilution of 1:4 did notsupport cell growth as well as did BALB/c spleen cell-

conditioned media (Table 1).Treatment of the BALB/c-conditioned medium with pro-

nase (1 mg/ml for 120 min at 37Â°)or heat (60Â°for 60 min)

decreased the activity of the preparation (Table 1). Ovariancells [in contrast to myeloma cells grown in this system (9)]did not require the presence of 2-ME for cell growth (Table

2).Effect of Histological Type on Cell Growth. Ovarian

colony growth has been achieved from 75% of patients withserous adenocarcinoma and 100% of patients with mu-

cinous, endometrioid, or undifferentiated types (Table 3).All patients but one were in relapse from treatment with

multiple drugs or radiotherapy.Cells from nonmalignant effusions (consisting of macro

phages, lymphocytes, mature granulocytes, and mesothe-lial cells) obtained from patients in cardiac failure failed to

Table 1Effect of substitution of alternative feeder layers on ovarian colony

growthVarious growth factors were added to underlayers al a dilution

of 1:4. Ovarian cells were then plated over feeder layers.No. of colonies/5 x 10*

cells

Control, (oil-primed condÃ¬- Experi-

Substituted material Trial tioned media)" mental

RBC1234306

Â±35''

2549Â±39420 Â±30380 Â±19320

Â±40833 Â±15300

Â±50350 Â±30

EffusionHeatPronase

CD-1 conditioned media"DBA/2 conditioned media"BALB/c conditioned media

MA-184 conditioned mediaWI-38 conditioned media

" See "Materials and Methods."" Mean Â±S.E.

Table 2Effect of 2-ME on ovarian colony growth

Ovarian tumor cells were cloned in the presence or absence of50 MM2-ME.

55561

667799

Â±999Â±999Â±9580

Â±27580Â±27580Â±27420

Â±50420Â±5029

Â±838Â±250Â±4320

Â±40240Â±36109Â±900

No. of colonies/200,000 cells

Trial1

23

4+

2-ME280

Â±14"183 Â±1691 Â± 1100 Â±11-

2-ME306

Â±48195 Â±2199 Â± 9146 Â±35

" Mean Â±S.E.

Table 3Effect of tumor type on ovarian cniony growth

Cells from patients with ovarian cancers of different histologicaltypes were plated in soft agar, and results were compared.

No. of successfulcases/total no. of

cases

Tumor typeSolid % sue- Colony

Effusion tumor cessful range

Serous adenocarcinomaMucinous

adenocarcinomaEndometrial

adenocarcinomaUndifferentiated

carcinoma10/15

5/53/3

1/12/25/575100100100127-50012-2549144-500020-220





form colonies at concentrations of 5 x 103 to 5 x 10s cells/

plate.Proliferate State of Ovarian Colony-forming Cells. In

cubation of cells with high specific activity [3H]dThd re

duced colony formation to as little as 35% of control. Nosuicide effect was seen in 2 patients with undifferentiatedcancer. Hydroxyurea, also known to kill cells in the S phaseof the cell cycle, reduced colony formation to about thesame levels as did [3H]dThd (Table 4).

Drug Sensitivity Assay. Survival plots of ovarian colony-

forming cells from 11 patients whose cells were tested withc/s-platinum diamine dihydrochloride are shown in Chart 2.

The wide difference observed in the plots indicates aspectrum of sensitivity to the drug. The dose of c/s-platinum

required to reduce colony formation to 50% of controlranged from 0.02 to 0.06 ng/m\ for 4 patients. A subsequentplateau effect, wherein increasing lethality with increasingdosage did not occur, was frequently observed and suggested the presence of a subpopulation of cells resistant toc/s-platinum. Unit areas under the survival concentrationcurves for c/s-platinum charts (to a defined upper boundary

of 0.06 /Â¿g/ml) ranged from 2.16 in the most sensitivepatient studied to 7.13 in the cells from the patient with thehighest degree of in vitro resistance. Resistant colonieswithin the population manifested continued growth onserial observation and did not differ in morphology fromsensitive colonies.

Effect of Depletion of Phagocytic Cells. The number ofovarian colonies seen after depletion of phagocytic macrophages was decreased in every case. The addition of 50 ^M2-ME only partially restored ovarian cell growth (Table 5).

Cytogenetics. Results presented are derived from thecytogenetic analysis of 11 cultures from 5 patients. Approximately 300 metaphases were observed, and over 100spreads were analyzed for chromosome number and recognizable chromosome aberrations.

The number of chromosomes per cell ranged from 38 to200. Three patients had a modal count near 58 to 60, onehad a modal count at 68 to 69, and one had a modal countof 42.

When cultures were harvested with our procedure, mostcolonies remained intact. Thus, the mitotic figures observed could be scored as occurring within or outside a

colony. Seventy % of the mitotic figures occurred within acolony, 30% near or outside a colony. In cultures of 4 of the5 patients, diploid metaphases were found primarily outsidethe colony, while hyper- or hypodiploid metaphases werefound within the colony. In 3 patients, all of the mitoseswithin a colony were hyperdiploid. In one patient (modalcount, 42), hypodiploid figures were found most oftenwithin colonies. In the final patient, all mitoses were outsidethe colonies, thus not allowing comparison. Double-minute

bodies, or dot chromosomes, were found in cultures from 2patients. One case contained multiple double minute chromosomes in approximately 20% of the cells, the second inapproximately 5%.

DISCUSSION

These studies have demonstrated that tumor cells frompatients with epithelial-type ovarian cancer can form colonies in soft agar. The growth pattern did not vary significantly with the histological type of the tumor.

The use of a battery of morphological and histochemicalstains indicated that cells picked from colonies had thesame features as did the original tumor cells. However,routine staining techniques are often inadequate for therecognition of malignant cells in pleural and peritonealeffusions (23). Even the use of histochemical stains forovarian tumor cells is not completely specific (12, 15). Thedevelopment and use of an ovarian tumor-specific anti-serum, such asthat raised by loachimef a/. (12) and Bhatta-

charya and Barlow (3), would be of use in confirming themalignant nature of the colony cells.

However, the good coincidence observed between properties of cultured cells and cells in the original suspensionand our cytogenetic observations suggested that the colonycells were derived from clonogenic neoplastic cells fromthe original tumor.

Cytogenetic analysis revealed a wide range of both chromosome number and modal count in cultured cells. This isconsistent with the few published cytogenetic analyses ofcultured or primary ovarian ascitic and solid tumors (1, 2,14, 27). The finding of double-minute bodies (2, 6) incultures from 2 patients is also of interest. Double-minute

bodies were originally observed in cultured cells from

Table 4Effect of [3H]dThd and hydroxyurea on human ovarian colony formation

Ovarian cancer cells were incubated with high specific activity [3H]dThd or hydroxyureafor 1 hr prior to plating.

Trial1

2345678Stage

at diag

nosis4

3333333Estimated

tumorbulk>1kg

>1 kg>2kg>1 kg<1 kg>1 kg>1 kg>1 kgSource

oftissueTumor

noduleTumor noduleAscitesTumor noduleAscitesAscitesAscitesAscitesControl

(No. ofcolonies/

500,000cells)54:

78:500

1832334720

120t

6Â°t 8

191640125.7

23%

survival of colony-formingcellsHydrox-

[3H]dThd yurea98

5865364683 113

120 15639 44

" Mean Â±S.E.

OCTOBER 1978 3441



A. W. Hamburger et al.

1004 06

ug/ml (1 hr)

Chart 2. Survival of ovarian tumor colonies after 1 hr exposure to c/s-platinum. Dose-response curves for 11 patients are depicted. Points, meanof 3 dishes. S.E. for each point averaged Â±10%. , 50% survival.

patients with neuroblastoma. Since this initial finding, theyhave been reported in nonneural tumors. Significantly,double minute bodies have only been found in malignanttissue. The fact that cells from nonmalignant effusions donot grow in our system at concentrations of 2 x 105 cells/

plate provides further supportive evidence that our culturesystem has some optimization for the growth of tumorcolonies.

Plating efficiency, although low, was in the same rangeas that observed for granulocyte or erythroid colony-forming cells or stem cells from transplantable mouse myeloma.Therefore, our assay should prove useful in determiningfactors governing the proliferation of ovarian tumor cells inthe same manner that soft agar assays have proven usefulfor elucidation of granulocyte and erythroid progenitorbehavior (4, 24).

The results reported here also indicate that macrophageswere needed for proliferation of ovarian tumor cells derivedfrom effusions. A number of authors have noted the stimulatory effect of macrophages on malignant cell growth (5,11, 18). Nathan and Terry (18) have demonstrated that DMAsynthesis of many (although not all) mouse lymphomas wasstimulated by normal mouse peritoneal macrophages. Similarly, Namba and Hanoka (17) reported that phagocyticcells or a glycoprotein released by them was required forgrowth of MOPC-104E mouse myeloma cells in spinner culture. CalderÃ³n and Unanue (5) also have found a cell proliferation factor for L1210 leukemic cells in the dialyzedsupernatants from mouse peritoneal macrophages.

2-ME only partially restored ovarian cell growth in theabsence of macrophages. Nathan and Terry (18) found thatstimulation of lymphoma cell DNA synthesis by normal macrophages was only evident when lymphoma lines could beequally stimulated by 2-ME.

Whether the stimulatory effect seen here was due todirect contact between macrophages and tumor cells or toa macrophage-derived factor is not known. Interestingly,autologous effusions, which presumably contain macrophage-derived factors, could not support ovarian tumor cellgrowth in the absence of medium conditioned by macro-

Table 5Effect of depletion of macrophages on ovarian tumor cell growth

Cell populations in malignant effusions from patients with ovarian cancer were depleted of phagocytic macrophages by ingestionof carbonyl iron prior to plating. The potential growth restorativeeffect of 2-ME (50 Â¿Â¿M)was also tested.

No. of colonies/500,000 cells

Trial Control- macrophages

macrophages + 2-ME

1234157 Â±30"146Â±3599

Â±9147Â±2031

Â±617Â±85Â±

358Â±1044

Â±825Â±1440Â±887Â± 7

" Mean Â±S.E.

phages from BALB/c mouse spleens. However, malignanteffusions probably contain a complex mixture of stimulatory and inhibitory substances derived from many celltypes, including activated lymphoid cells as well as macrophages and tumor cells. We speculate that the reason thatovarian cancer often fails to spread outside peritoneal orpleural cavities may be related to the dependence of tumorcells on macrophages in such exÃºdalesfor cell growth (21).

In addition to its use in studying the biological propertiesof ovarian tumor cells, this assay should prove to be ofclinical importance. For example, in vitro tumor colony-forming assays of transplantable mouse myeloma haveindicated that each subline had a different pattern of drugsensitivity that predicted that which could be observed invivo (19). The drug sensitivity study reported here indicatesa wide spectrum of sensitivities to the experimental agentc/s-platinum.

Additionally, the presence of a viable, resistant subpopulation of cells could be inferred from serial observation ofsurviving colonies. Clonal proliferation of such inherentlyresistant subpopulations could be responsible for the development of cellular resistance and overt relapse thateventually occurs in most patients with advanced-stageovarian cancer. Comparative study of the properties ofsensitive and resistant colonies might help to clarify resistance mechanisms that are of clinical importance and provide opportunity to devise techniques to overcome specifictypes of resistance.

Our initial studies (22) correlating in vitro and in vivosensitivity of myeloma and ovarian tumor colonies to various standard anticancer drugs have suggested that suchassay results frequently correspond with clinical sensitivityor resistance. Careful prospective clinical testing of thisdrug assay technique appears warranted to determinewhether it will have potential utility in screening useful newagents and selecting the most effective drugs for individualpatients with ovarian cancer.

REFERENCES

1. Atkin, N. B., and Pickthall, V.J. Chromosomes 1 in 14 Ovarian Cancers.Human Genet.,38. 25-33, 1977.

2. Balaban-Malenbaum, G., and Gilbert, F. Double Minute Chromosomesand the Homogenously Staining Regions in Chromosomes of a HumanNeuroblastoma Cell Line. Science, 798. 739-741, 1977.

3. Bhattacharya, M., and Barlow, J. Immunologie Studies of Human SerousCystadenocarcinoma of the Ovary. Demonstration of Tumor Associated





Antigens. Cancer, 37. 588-595, 1973.4. Bradley, T., and Metcalf, D. The Growth of Mouse Bone Marrow In Vitro

Aust. J. Exptl. Biol. Med. Sci., 4: 287-299, 1966.5. CalderÃ³n, J., and Unanue, E. Two Biological Activities Regulating Cell

Proliferation Found in Cultures of Peritoneal ExÃºdateCells. Nature. 253.359-361, 1975.

6. Cox, D., Yuncken, C., and Spriggs. A. I. Minute Chromatin Bodies inMalignant Tumors of Childhood Lancet, 2. 55-58,1965.

7. DiSaia, P., Morrow, M., Kanabus, J., Piechal, W., and Townsend, D. E.Two New Culture Lines from Ovarian Cancer. Gynecol. Oncol., 3: 215-219, 1975.

8. Hamburger, A. W., and Salmon, S. E. Primary Bioassay of Human TumorStem Cells. Science, 197: 461-463. 1977.

9. Hamburger, A. W. and Salmon, S. E. Primary Bioassay of HumanMyeloma Stem Cells. J. Clin. Invest., 60. 846-854, 1977.

10. Hecker, D., Saul, G., and Wolf, G.Untersuchungen Zum Einsatz vonZellund Organkultur Bei Zytostatika Sensibilitatistestungen Air Menschlichen Ovarialkarzinomen. Arch. Geschwulstforch., 46: 34-43. 1976.

11. Hewlett, G.. Opitz, H. G., Schlumberger, H. D., and Lemke, H. GrowthRegulation of a Murine Lymphoma Cell Line by a 2-Mercaptoethanol orMacrophage-Activated Serum Factor. European J. Immunol., 7. 781-785, 1977.

12. loachim, H., Sabbath, M., Anderson, B., and Barber, H. K. TissueCultures of Ovarian Carcinomas. Lab. Invest., 31: 381-390, 1974.

13. Iscove, N . Till, J., and McCulloch, E. The Proliferative State of MouseGranulopoietic Progenitor Cells. Proc. Soc. Exptl. Biol. Med., 134: 33-37,1970.

14. Kakati, S., Hayota, I., Oshimura, M., and Sandberg, A. A. Chromosomesand Causation of Human Cancer and Leukemia X. Banding Patterns inCancerous Effusions. Cancer, 36: 1729-1738, 1975.

15. Long, M., and Hosannah, Y. Comparative Histochemical Studies of u-leucyl-/3-naphthylamidase in Ovarian Epithelial Tumors. Cancer, )9:909-925, 1965.

16. Luna, L. Manual of Histologie Staining Methods, pp. 70-158. New York:McGraw-Hill. 1968.

17. Namba, Y., and Hanoka, M. Immunocytology of Cultured IgM Forming

Cells of Mice. J. Immunol., 109: 1193-1200, 1972.18. Nathan, C. F., and Terry. W. D. Differential Stimulation of Murine

Lymphoma Growth In Vitro By Normal and BCG-activated Macrophages.J. Exptl. Med., 142: 887-902, 1978.

19. Ogawa, M., Bergsagel. D., and McCulloch, E Chemotherapy of MouseMyeloma: Quantitative Cell Culture Predictive of Response In Vivo.Blood, 4: 7-15, 1973.

20. Park. C. H., Bergsagel, D., and McCulloch. E. Mouse Myeloma TumorStem Cells: A Primary Cell Culture Assay. J. Nati. Cancer Inst., 46: 411-416,1971.

21. Salmon, S. E.. and Hamburger, A. W. Immunoproliferation and Cancer:A Common Macrophage Derived Promotor Substance. Lancet. ) (8077):1289-1290,1978

22. Salmon, S. E., Hamburger, A. W., Soehnlen, B. J., Durie, B. G. M.,Alberts, D. S., and Moon, T. C. QuantitÃ¤ten of Differential Sensitivitiesof Human Tumor Stem Cells to Anticancer Drugs. New Engl. J. Med..298:1321-1327,1978.

23. Soendergaard, K. On the Interpretation of Atypical Cells in Pleural andPeritoneal Effusions. Acta Cytol., 21: 413-416, 1977.

24. Stephenson, J., Axelrod. A., McLeod, D.. and Shreeve, M. Induction ofColonies of Hemoglobin Synthesizing Cells by Erythropoietin in Vitro.Proc. Nati. Acad. Sci., 68: 1542-1546, 1972.

25. Sun, N. C.. Chu, E. H., and Chang. C. C. Staining Method For theBanding Patterns of Human Mitotic Chromosomes. Mammalian Chromosomes Newsletter, 14: 26-27,1973.

26. Technical Bulletin 47, pp. 1-3. Chicago: American College of Obstetricsand Gynecology, 1977.

27. Tiepolo, L., and Zuffardi. O. Z. Identification of Normal and AbnormalChromosomes in Tumor Cells. Cytogenet. Cell Genet., 12: 8-16, 1973

28. Todd, J., and Sandford, A. The Blood. In: Davidson and Wells, (eds.).Clinical Diagnosis by Laboratory Methods, pp. 131-132. Philadelphia:W. B.SaundersCo., 1962.

29 Young, R. C. Chemotherapy of Ovarian Cancer: Past and Present. Sem.Oncol.,2. 267-276, 1975.

30. Zaharia, M., and Samorlescu, M. In Vitro Growth Peculiarities of SomeHuman Ovarian Tumors. Morphology Embryology. 27: 283-287. 1975.

Fig. 1. Phase-contrast view of a typical ovarian cancer colony from an 11-day-old culture grown from a patient with serous adenocarcinoma in relapse. Apseudoacinar pattern was apparent, x 435.

Fig. 2. a, three cells plucked from an ovarian tumor colony from a patient with a mucinous adenocarcinoma of the ovary. The cytoplasm of the cells werestrongly PAS positive, x 2680. b, four cells plucked from an ovarian tumor colony from a patient with a serous adenocarcinoma of the ovary. The cytoplasmof these cells stained lightly with PAS. x 2680.

OCTOBER 1978 3443



1978;38:3438-3444. Cancer Res Anne W. Hamburger, Sydney E. Salmon, Mary B. Kim, et al. Direct Cloning of Human Ovarian Carcinoma Cells in Agar

Updated version

http://cancerres.aacrjournals.org/content/38/10/3438

Access the most recent version of this article at:

E-mail alerts related to this article or journal.Sign up to receive free email-alerts

Subscriptions

Reprints and

[email protected] at

To order reprints of this article or to subscribe to the journal, contact the AACR Publications

Permissions

Rightslink site. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC)

.http://cancerres.aacrjournals.org/content/38/10/3438To request permission to re-use all or part of this article, use this link



http://cancerres.aacrjournals.org/cgi/alerts

mailto:[email protected]



direct cloning of human ovarian carcinoma cells in agar1 · tumor stem cells in soft agar (8, 9)....

Documents