IN VITRO Vol. 11, No..5, 197.5

DEVELOPMENT OF AN IN VITRO COLONY FORMATION ASSAY F O R T H E E V A L U A T I O N O F I N VIVO C H E M O T H E R A P Y O F A R A T B R A I N T U M O R 1' ~ MARK L. ROSENBLUM/KATHY D. KNEBEL, KENNETH T. WHEELER, 4 MARVIN BARKER, AND CHARLES B. WILSON

:The Howard C. Naffziger Laboratories, Department of Neurological Surgery, University of California Medical Center, San Francisco, California 9~1~3 SUMMARY

An in vitro colony formation assay for the evaluation of in vivo brain tumor therapy has been developed. When plated, disaggregated cells derived from solid tumors proliferated to form relatively homogeneous colonies after a latency period of 2 to 6 days. Increasing concentrations of fetal calf serum enhanced colony-forming efficiency (CFE) with a plateau between 7 and 16%. Supplementation with either irradiated feeder cells (10 a to 105 cells per dish), or medium conditioned by 1 to 3 days of in vitro incubation with the same cell line, doubled the CFE. The density of tumor cells (untreated or previously treated with chemotherapeutic agents) did not affect the CFE when a minimum of 104 total cells (tumor plus feeder) were plated. Therefore, in this system the optimal experimental conditions for evaluating chemotherapy and radiotherapy require incubation of disaggregated tumor cells for 12 days in medium containing 10% of fetal calf serum and enough feeder cells to provide a minimum of 104 cells per dish. The CFE for untreated tumors was 18 • 10% (• demonstrating that there is significant biological variation. The assay appeared sensitive, with reproducible results, when applied to individual chemically treated tumors. An estimate of the percentage of clonogenic cells affected by in vivo chemotherapy may be obtained by comparing the CFE of cells from treated and untreated tumors. This assay can measure up to a 5 log10 cell kill, and it should prove to be valuable in developing more effective regimens for the treatment of solid tumors in animals and man.

Key words: colony; clonogenic; assay; therapy; brain tumor. cause tumor size is affected by alterations in growth fraction, cell volume, and rate of dead cell removal. Survival studies, although more reliable and applicable to clinical chemotherapy, are based on the assumption that results will vary minireally within groups of similarly treated specimens despite possible differences in tumor kinetics, location, and host response. Both of these assays 1 This work was supported by NIH Center Grant CA 13525, the National Phi Beta Psi Sorority, the require many animals to obtain statistically sigJoe Gheen Medical Foundation, and the Associa- nificant results, and are often insensitive to minor tion for Brain Tumor Research. perturbations. 2 Presented at the 25th Annual Meeting of the Recently, single cell suspensions from previTissue Culture Association, Miami Beach, Fla., ously treated solid tumors have been returned to June 3-6, 1974. 3 Recipient of NIH Individual Postdoctoral isogeneic hosts in an attempt to quantify cellular Research Fellowship Award 1 F22 CA02196-01. survival directly. The number of viable tumor 4 Also : Laboratory of Radiobiology, University of~California, San Francisco, Ca. 94143. Supported cells has been estimated from the length of animal survival after intracerebral or intraperitoneal by N I H Career Development Award. 264 The evaluation of solid tumor therapy will greatly benefit from the development of a reliable and sensitive assay of cell survival. Previously, investigators have relied upon indirect measures such as changes in tumor size (1-5), or prolongation of animal survival time (4-6). The results of the former method are difficult to evaluate be-

COLONY FORMATION ASSAY cell injection (4, 5, 7-9), and from the development of new tumors after subcutaneous transplantation (TDs0) (10-12). These in vivo techniques often entail long observation periods on many animals. Till and McCulloch (13) developed a spleen colony-forming unit (CFU) assay which has been used by many other investigators (1418) to measure normal and tumor cell survival in the hematological system of the mouse. Lung colony-forming assays have also been developed using similar procedures (19-22). However, these tumor cell bioassays do not allow for possible effects of the immunological system of the newly challenged host. We have modified an in vitro colony formation assay that has been used for the evaluation of in vivo chemotherapy (23-25) and radiotherapy (23, 24, 26-30) of subcutaneous animal tumors. Inasmuch as detailed analyses of such assays have not been published, we are reporting the evolution of a method for the in vitro analysis of cell survival after in vivo chemotherapy of a rat brain tumor. ~ATERIALS AND METHODS

An N-nitrosomethylurea-induced rat brain tumor was transplanted into the brains of 150 to 200 g adult male Fischer 344 rats, using a previously reported technique (6). The original tumor was described as maJignant astrocytoma (31). However, this tumor is now classified as a gliosarcoma or sarcoma because of changes occurring during serial animal and cell culture passage (6). General assay procedure. Rats were sacrificed by cervical dislocation and their brain tumors were removed aseptically and placed in Hanks' balanced salt solution at 4~ Specimens were weighed, minced to l-ram pieces with scalpel blades and placed in trypsinizing flasks. Twentyfive milliliters of 0.5% trypsin solution (reconstituted from lyophilized trypsin, GIBCO, Santa Clara, Calif.) were added and the suspension was decanted through a stainless steel mesh (80 #m) and centrifuged at 200 • g for 10 rain at 4~ The supernatant fluid was decanted, 10 ml of fresh medium added, and the cells resuspended. The cells were counted using a hemocytometer and viability was estimated using the trypan blue dye exclusion test. The number of single cells ranged from 1.5 • 105 to 3.0 • 105 per mg of solid tumor tissue, and cell viability usually exceeded 95%. Dilutions of the cell suspension were measured

265

into five or more 60-mm plastic Petri dishes with 2-mm grids (Microbiological Associates, In% Albany, Calif.) into which complete medium and sometimes feeder cells had been placed. The complete medium consisted of Eagle's basal medium supplemented with 10% of fetal calf serum, 1% of L-glutamine, 1% of nonessential amino acids, 1% of vitamins, and antibiotics (penicillin, 80.5 units per ml and streptomycin, 80.5 #g per ml) (Microbiological Associates, Inc., Albany, Calif.). Feeder cells were obtained by trypsinization of a previously irradiated (4000 fads) confluent monolayer of rat brain tumor cells, When used, feeder cells were added to provide 1 • 104 cells per dish in a total volume of 5 ml of medium. All cultures were incubated in 5% COs for 12 days at 37~ The cells were then fixed in situ with modified Lavdowsky fixative (32) for 45 min and stained with Giemsa for at least 1 hr. The number of colonies containing more than 25 cells was determined for each dish. The colonyforming efficiency (CFE) was calculated by dividing the number of colonies by the number of cells plated and expressed as a percentage. In vitro growth kinetics. To evaluate the in vitro growth characteristics of colonies obtained from primary explants of the rat brain tumor, the CFE and maximum colony diameter were determined for groups of 10 plates, each containing 200 tumor and 104 feeder cells. The proliferation kinetics of single cells were studied during incubation in conditioned medium for periods up to 21 days. In plates containing small numbers of cells, single cells were identified and their location marked with a wax pencil. Each area was observed daily and the number of cells was scored using an inverted phase contrast microscope. Effect of fetal calf serum concentration. To evaluate the effect of fetal calf serum concentration on the CFE, groups of five plates containing 200 untreated tumor cells and 104 feeder cells were incubated in media containing serum concentrations of 2, 7, 10, 13, 16, and 22%, and the CFE was determined for each. Effect of feeder cells. To evaluate the effect of feeder cell concentration on CFE, groups of five plates containing 200 untreated tumor cells were incubated alone and with various numbers of feeder cells (102, 5 • 102, 103, 1@, and 105) and the CFE was determined for each group. Because the CFE for each individual tumor is different., studies that require comparison or combination

266

ROSENBLUM ET AL.

of results from experiments performed at different times must utilize a common denominator. Therefore, the expression "relative CFE" has been used and is defined as the CFE for each group of plates in an experiment, divided by the CFE obtained for 200 tumor cells incubated with 104 feeder cells. The relative CFE is calculated separately for each experiment and is expressed as a percentage by multiplying by 100. In this manner, the effect of feeder cell density was determined from four separate experiments. To determine whether feeder cells could be plated before an experiment, 200 cells from each of four tumors were incubated with 1@ feeder cells prepared simultaneously, and 1 and 2 days previously. The CFE of each tumor was determined and the different preparations compared. .Effect of conditioned medium. To evaluate the effect of conditioned medium on CFE, medium collected from cultures of exponentially growing rat tumor cells after 1, 2, and 3 days of incubation was filtered through a cellulose ester Millipore filter (0.22 /~m) and stored at 4~ until used. Concurrently with a feeder cell density experiment, tumor cells were incubated with 4 ml of conditioned medium without the addition of feeder cells. The results obtained were adjusted to relative CFE as previously described. Effect of tumor cell density. To determine the effect of cell density on CFE, untreated tumor cells (200, 400, 800, 1600, 3200, and 6400) were plated and feeder cells were added to obtain a total of 1@ cells per plate. Similar cell dilutions were plated without the addition of feeder cells. The relative CFE was determined for each cell ~|

I

I

I

concentration and the results from three experiments were combined. Retransplantation study. To determine the origin of the colony-forming cell, colonies were grown from untreated tumors in the usual manner. Cells were harvested from these plates and 4 • 104 cells were implanted intracerebrally into 15 animals. A second group of 15 animals were transplanted with 4 • 104 cells obtained from long-term tissue culture in the routine fashion (6). The growth pattern and histological characteristics of tumors developing in both groups were determined. Assay sensitivity for trea~d tumors. Dilutions of tumor cells that had been treated in vivo with cytotoxic agents were plated, to evaluate the effects of the density of treated cells and diluting error on the CFE. Tumor-bearing animals were treated with 1-[2-ehloroethyl]-3-cyclohexyl-1nitrosourea (CCNU, NSC 79037), or 1,3-bis[2chloroethyl]-l-nitrosourea (BCNU, NSC 409962), 24 hr before tumor removal. The agents were administered intraperitoneally at approximate LDz0 doses. Seven therapeutic trials were conducted in six experiments and the CFEs were determined. ]:~ESULTS

In vitro growth kinetics. Colonies first became visible to the unaided eye by the 7th day; following this, the maximum colony diameter and number of colonies increased daily (Table t and Fig. 1). Little change in CFE was noted after the l l t h day of incubation when the maximal colony diameter was approaching 4 mm. Colonies were I

I

I

5 u.I

U

9 z 0._J 0 u

2

,/]

I

I

I

I

I

I

'7

8

9

10

11

12

13

DAYS

OF

14

INCUBATION

FIG. 1.~The effect, of duration of in vitro incubation on the CFE. Each point represents the mean of 10 plates :t: S.E.

COLONY FORMATION ASSAY TABLE 1 T H E K I N E T I C S OF COLONY G R O W T H IN V I T R O a Duration of Incubation

Mean No. Colonies/ Plate

Colony-forming Efficiency (% ~ S.E,)

days

7 8 9 10 11 12 14

Maximum Colony Diameter mm

3.8 5.8 7.6 9.2 10.0 10.6 11.2

1.9 2.9 3.8 4.6 5.0 5.3 5.6

4444444-

0.3 0.4 0.6 0.5 0.4 0.5 0.5

0.3 0.9 1.3 2.6 3.0 3.8 5.5

267

a confluent monolayer of irradiated, nonproliferating cells which often peeled off the dish after several days of incubation. Feeder cells could be irradiated and plated 1 or 2 days before use without a deleterious effect on C F E (Table 3). Effect of conditioned medium. The relative C F E obtained using conditioned media from 1, 2, and 3 days of incubation was 94, 99, and 109%, respectively. Comparing the CFE enhancement of 1000

Each plate contained 200 untreated tumor cells plus 104 feeder cells. Each value represents results from 10 plates. composed of densely packed, bipolar, sarcoma-like cells growing in a homogeneous monolayer. Clones assumed a circular contour after attaining sizes of greater than several hundred cells. Giant cells were observed in a few of the larger colonies. Clones composed of epithelioid cells were encountered rarely. When individual cells were observed, five of ten were seen to proliferate after a latency period of 2 to 6 days (Fig. 2). Four of these five groups continued to proliferate, forming large colonies with typical growth curves; whereas one group, which had contained five cells on day 7, was observed to diminish in size over the remainder of the experiment. Effect of fetal calf serum concentration. The C F E increased up to 7% serum concentration, remained essentially unchanged until 16%, and thereafter again increased (Fig. 3). A serum concentration of 10% was considered appropriate for subsequent studies, inasmuch as minor variations above or below this concentration did not significantly affect the CFE. Effect of feeder cells. The C F E of 200 untreated tumor cells is approximately doubled b y the addition of 103 to 105 irradiated feeder cells (Table 2 and Fig. 4). Optimal stimulation of C F E consistently occurred at a feeder cell density of 104 to 105 cells. Both the number of feeder cells and the total number of cells (feeder plus tumor) are shown on the abscissa of Fig. 4, because the plated untreated tumor cells appear tO act as feeder cells (by conditioning the medium), a phenomenon t h a t makes the total number of cells t h e more important factor influencing the CFE. The addition of more than 5 X 105 feeder cells resulted in

o

100

8

L. 10

4

2

6

DAYS

8

|0

12

14

OF INCUBATION

Fro. 2. The formation of colonies from single Cells as a function of the incubation period.

g (.J

z

1

I

I

4

I 8

I 12

16

E

m

12 o 8 o u_

z o_1 o

(,J

, 0

16

20

FETAL CALF SERUM C O N C E N T R A T I O N

(%}

FIG. 3. The effect of fetal calf serum concentration on the CFE. Each point represents the mean from five plates + S.E.

268

R O S E N B L U M E T AL. TABLE

2

T H E E F F E C T OF VARIOUS NUMBERS OF IRRADIATED F E E D E R C E L L S [4000 r] ON THE COLONY-FORMING EFFICIENCY OF 2 0 0 UNTREATED TUMOR C E L L S Colony-forming Efficiencya

Exp.

Relative Colonyforming Efficiencyb

(% 4- S.E.)

(% :~ S.E.)

2.2 3.1 6.0 5.4 5.8

:i: • • :i: •

0.2 0.1 0.3 0.1 0.3

41.1 54.4 112.0 100.0 108.3

0.4 0.5 0.5

54.4 4- 7.8 70.2 • 8.6 100.0 4- 9.4

III

8.3 • 0.7 9.2 4- 0.8 13.9 4- 1.2

59.7 :t: 4.8 66.2 4- 5.8 100.0 • 8.4

IV

11.5 4- 0.9 15.3 -4- 1.4 15.5 -4- 1.5

75.2 • 100.0 • 101.3 •

3.1 • 4.0 • 5.7 •

II

• • -4• 4-

3.0 1.9 5.9 1.8 5.1

larger n u m b e r s of cells were plated, t h e C F E decreased because t h e large n u m b e r of colonies b e g a n to overlap. Differences in t h e C F E w i t h a n d w i t h o u t feeder cells a p p e a r to d h n i n i s h a t 3200 cells a n d disapp e a r a t 6400 cells (Fig. 5). This finding indicates t h a t t u m o r cells b y themselves condition t h e m e d i u m w i t h a b o u t t h e same efficiency as irr a d i a t e d feeder cells, w i t h t h e o p t i m a l t o t a l cell n u m b e r b e i n g b e t w e e n 108 a n d 104 . Retransplantation study. I n t r a c e r e b r a l t u m o r s developed after t r a n s p l a n t a t i o n of cells o b t a i n e d from plates c o n t a i n i n g colonies of previously TABLE

5.9 9.1 9.7

Tumor No.

Daysa

t

1

lOO

Fresh Feeder Cellsb

Colony-forming efficiency c

Colony-forming efficiency ~

0.05 3.15 12.35 1.70

b R e l a t i v e colony-forming efficiency = [colonyforming efficiency of each d i l u t i o n / c o l o n y - f o r m i n g efficiency w i t h 104 feeder cells] X 100. Calculated for each experiment individually. I

Old Feeder Cells~

(% * S.E.)

a E a c h value represents the m e a n for five plates.

~/I

3

T H E EFFECT OF PREPARING F E E D E R CELLS B E F O R E A COLONY-FORMING E X P E R I M E N T

• • 4•

(% +_ S.E.)

0.20 3.65 12.17 1.90

0.05 0.38 0.68 0.53

4. 4• •

0.05 0.32 0.17 0.27

Feeder cells irradiated and 104 p l a t e d 1 or 2 days before t u m o r experiment. b Feeder cells irradiated and 10 ~ plated immediately before t u m o r experiment. c Each value represents a mean for five plates.

ou

,

[

J

'

I

+

w >

50

U

x,~

_u

0

o

L•l

0 1

TOTAL

I 3

I 4

I 5

NUMBER OF TUMOR CELLS (log)

FIG. 4. T h e effect of feeder cell density (inserted numbers) and t o t a l cell n u m b e r on the CFE. E a c h point r e p r e s e n t s the combined m e a n • S.E. from at least two e x p e r i m e n t s containing five plates each. c o n d i t i o n e d m e d i a to s i m u l t a n e o u s e x p e r i m e n t s u s i n g feeder cells, conditioned m e d i a h a d t h e same effect as 103 to 103 feeder cells. Effect of tumor cell density. W h e n t u m o r cells are p l a t e d w i t h feeder cells, t h e C F E is a p p a r e n t l y i n d e p e n d e n t of t h e n u m b e r of t u m o r cells p l a t e d over t h e r a n g e of 200 to 1600 cells (Fig. 5). W h e n

u~ 100 o z

2 z>-

75

o. o

9 +,i,+-+---r

U

.