Journal of Neuro-Oncology 6: 361-370 (1988) © Kluwer Academic Publishers - Printed in the Netherlands

361

New sensitivity test using flow cytometry Keiji Kawamoto, Katsuhiro Kawakami, Yasuo Kawamura, Hiroshi Matsumura and Akio Ohyama*

Department of Neurosurgery and *Department of Microbiology, Kansai Medical University, Moriguchi, Japan

Key words: sensitivity test, brain tumor, flow cytometry, antineoplastic agents Summary Flow cytometry (FCM) has been used to evaluate not only the malignancy of tumor cells but also the effects of chemotherapy. Here a new application of FCM for selecting the best antineoplastic agent in the chemotherapy for brain tumors is reported. Through our preliminary study using established brain tumor cell lines, the system for this sensitivity test was developed. Antineoplastic agents were placed in contact with monolayer-cultured cells; then cell viability and changes in the DNA histogram were analyzed by FCM. Cell viabilities were measured with the fluorescein diacetate (FDA) staining method, and the DNA histogram was analyzed by the propidium iodide (PI) staining method. The best antineoplastic agent was determined based on changes in cell viability and cell cycle. In other words, when markedly decreased viability as compared with that of the control, is measured by FCM, then the agents can be considered to have had a cytocydal effect on the tumor cells, and thus the sensitivity of the agents is able to be evaluated. If the viability of the tumor cell is observed to be similar to that of the control, the cytostatic effects of the agents are able to be evaluated only if a marked change is observed in the DNA histogram. After the preliminary study, this system was applied clinically to malignant brain tumor cases, resulting in success in selecting the best antineoplastic agent for each individual case. Our sensitivity test using this FCM established in vitro system has much potential value for clinical use.

Introduction Malignant brain tumor cells of 280 cases, collected from surgery samples since 1980, were monolayercultured and placed in contact with various antineoplastic agents, and their cell kinetics were analyzed using flow cytometry (FCM; FACS III, Becton-Dickinson) [1-4]. Most malignant brain tumor cells obtained from surgical specimens could be cultivated and still keep their characteristic biological activities for a short term. Monolayer culture of brain tumor cells is convenient as a system for an experimental model for the sensitivity test of antineoplastic agents in individual cases; especially for glioma cells with slender cell processes which have

a tendency to grow easily in such a monolayerculture system. Flow cytometric evaluation is reliable and reproducible enough for the analysis of the effects of the antineoplastic agents so that FCM could become useful for clinical materials if a precise preparation is made from the pure sample of a single tumor cell suspension. This new sensitivity test of antineoplastic agents using FCM was established in monolayer-cultured cell lines of brain tumor cells by quantitatively evaluating the preliminary study and sensitivity of each agent. The present sensitivity test was also tried clinically to determine the selection of the best antineoplastic agent; the advantages of this system are discussed.

362 Materials and methods

Sensitivity test of antineoplastic agents using FCM Changes in the cell kinetics of brain tumor cells in a monolayer culture caused by antineoplastic agents were analyzed based on the results of FCM study. Sensitivity of antineoplastic agents to the individual cells was correlated with the viable rate; therefore, a definitive decrease in viability, compared with the control, represented cytocydal effects. Thus, cytocydal effects of the antineoplastic agents were judged based on the decreased rate of viable cells; cytostatic effects were evaluated according to the disturbance of the cell cycle as shown on the DNA histogram. Flow cytometric judgment about the effects of the antineoplastic agents was made by taking the viability caused by the disturbance of the cell cycle into account. In the flow cytometric analysis, a cell suspension of a brain tumor cell was obtained after enzymatical preparation with trypsin-EDTA and the suspension was separated into two portions. One portion was stained with 20/zg/ml of fluorescein diacetate (FDA) to measure the viability which was shown in percentage as the cytocydal effects of the agents [5]. We have already reported that this method produced results similar to those obtained by the conventional trypan blue method [6] and that we believed that this FDA method was a concise and significant method. The other portion was fixed with 99.6% methanol and then conventionally stained with a mixture of 50/~g/ml of propidium iodide (PI) and 1.139 mg/ml

of sodium citrate to evaluate the degree of cell cycle disturbance due to the cytostatic effects of the agents shown in the DNA histogram [7]. In judging the effects of antineoplastic agents in the present test, the results of the growth curves of the tumor cells and changes of the viability and the DNA histogram determined by FCM were totally evaluated. On the other hand, a high viability rate suggested the resistance of the tumor cells to the agents. Concerning the DNA histogram and cell viability, an accumulation of G2+M phase ceils with a high viability rate could be interpreted as a disturbance of cell cycle due to the cytostatic effects of the agents (Table 1).

Sensitivity of the established brain tumor cells Established T9 rat glioma cells [8] were used in this study. A monolayer-culture was performed in a medium of 4 ml containing the minimal essential medium (MEM) and a 10% phosphated buffer saline (PBS) solution. After inoculating 1.0×105 cells/ml, ACNU, 5-FU, Mitomycin C (MMC) and Bleomycin (BLM) respectively were placed in contact with the ceils at a logarithmic growth phase for 24 hours. Basic concentration of each agent was 1/1000 of the conventional administration dose used clinically: 1, 5, 10 and 20 tzg/ml for ACNU, BLM and 5-FU; and 0.01, 0.05, 0.1, and 0.2/~g/ml for MMC. For days after exposure to the agents, tumor cells were analyzed by FCM concomitant with the evaluation of the growth curves. The effects of antineoplastic agents were judged as follows: cytocydal

Table 1. Evaluation of the sensitivity test by flow cytometry.

Assay method grade

m

Growth inhibition (growth curve)

(-)

Flow cytometry Cell viability (%)

DNA histogram

No change

No change

~ Cytocydal effect

Cytostatic effect l Cytocydal effect

1

+

l (+)

363 effects were evaluated from the viabilities of the brain tumor cells in various concentrations of the agents, and the degree of the disturbance of the cell cycle due to cytostatic effects was measured by the percentage of G2 + M phase cells in the DNA histograms. ACNU and 5-FU, showing sensitivities with decreases in cell viabilities, were placed in contact with the cells for 24 hours in each basic concentration; their morphological change was chronologically observed for 6 days and then compared with the changes of the cell kinetics shown by FCM.

10

contrl

g5 "-

~¢/

oe

/

~

_

/

BLM lO#9/rnl MMC O.t/~g/rnl

~

11 contact I

i

[

1

2

3

I

4

I

5 aay

Fig. 1. Basic concentration ofACNU, 5-FU, MMC and BLM was administered, respectively.

Selection of the best antineoplastic agent in clinical cases Sterile surgical specimens were rinsed in PBS 2 - 3 times (to wash out corpuscles), minced, pipetted, treated with enzymes (trypsin and EDTA) and then monolayer-cultured in a 37 °C 5% CO 2 incubator (primary culture). In the present study, 15 specimens of malignant brain tumor cells were placed in contact with several kinds of antineoplastic agents to select sensitive agents at the second passage. Clinical evaluation of the cells was divided into three grades: 1. Cells in which viability decreased less than 50°70 compared with that of the control were evaluated to be sensitivity (+), 2. cells showing a slight decrease in viability (less than 50- 70°70 or a marked disturbance of cell cycle on the DNA histogram) were evaluated to be sensitivity ( - ) - (+), and 3. cells which showed a high viability and slight changes on the histogram and, therefore, had no cytocydal and cytostatic effects were evaluated to be sensitivity ( - ) .

Results

Results of established brain tumor cells Figure 1 shows growth curves of T9 rat glioma cells given basic concentrations of ACNU, 5-FU, MMC and BLM. In the control, the cells increased from 1.0x 105 cells/ml to 1.0× 105 cells/ml 5 days after in-

oculation. Each agent showed significant growth inhibitions from the basic concentrations, showing no clear differences among the 4 agents. In the comparative study of ACNU and 5-FU at various concentrations, both agents had slight growth inhibitions at a low concentration of 1 txg/ml and cell count reductions at a high concentration of 20 t~g/ml without showing a clear difference between the two agents (Fig. 2). Figure 3 shows the results of the sensitivity test using FCM. Figure 3a shows cell viabilities for each agent determined by FCM. With ACNU and 5-FU, the higher the concentrations, the viabilities, showing 24°70 and 38% at 20/zg/ml, respectively. MMC, however, shows high percentages 60- 66% of viability at 0.05-0.2 ~zg/ml. BLM shows higher percentages suggesting a poor reduction of viability. Figure 3b shows the percentage of G2 + M phase cells in the whole cells determined from the DNA histogram 4 days after contact with various concentrations of the agents. Compared with the value of 9°70 in the control, MMC shows a rather high rate, 21°70 at 0.1/xg/ml; on the other hand, ACNU shows the same value as the control at 10/xg/ml but a rather high value at 20/xg/ml. BLM and 5-FU show low values, 6°70 and 3°70, at 10/zg/ml, respectively. From the above results, ACNU and 5-FU were determined sensitivity (+) as they decreased cell viability due to their cytocydal effects. MMC was determined (_+) as a slight reduction of cell viability and an accumulation of G2 + M phase cell suggested that it possessed cytostatic effects rather than cytocydal

364 10¸

~

i

100

OntrOI

BLM

~J

C3 X

MMC

~5 CNU

c

;

50

I #g/mu

5-FU

I 0 #g/~l "~

contact

24hrs

10

~

-

-

-

e

~

4 ACNU

20 # g/,rn I

~oontrol

i

0.01

i

0.05

i

0.1

I

0.2

i

'MMC

0

1

5

10

20

~g/ml

concentration of drugs

100 %

+

~5 5-FU l~g/rn I

i

"1-

1

?

contact

24hrs

-

i ;lO#g/ml - -----'20#g/ml

T

Fig. 2. Growth curves of T9 rat glioma cells. Various concentrations of ACNU (a) and 5-FU (b) were delivered.

50

fo g u

~ i

0 1 MMC : 0.01

MMC ACNU BLM

5-FU

i 5 0 05

concentration

1O 0.1

20 #g/ml 0,2

of drugs

Fig. 3. Results of our sensitivity test by FCM for T9 rat glioma

effects for a T9 glioma cell. BLM was determined ( - ) because of high viabilities and no change in the DNA histogram suggesting the resistance of the cells to the antineoplastic agents. Microscopic findings are shown in Fig. 4 of T9 rat glioma cells on the 4th day after contact with 5-FU (c) and ACNU (b) at 10 #g/ml, and that of the control (a). The control cell represented a bipolar configuration of a tumor cell and a round nucleus. But abnormal nuclei were induced markedly by ACNU and marked vacuolization by 5-FU. Table 2 shows nucleic abnormality, disappearance of cell processes and the degree of vacuolization in 200 cells during

cells. Cell viability (a) and fractions of G 2 + M phase cell (b) by different drug concentrations are shown.

chronological observation. At 24 and 48 hours after contact with the agent, the cells had no clear morphological changes, but from 98 hours on morphological changes of abnormal nuclei became marked. A chronological study on the viability and DNA histogram for ACNU and 5-FU by FCM revealed a clearly decreased viability during 6 days and a rapid accumulation of G2 + M phase cells 2 4 - 48 hours after contact (Figs. 5, 6). The time when accumulation

365

Fig. 4. Control of T9 rat glioma cells (a), 10 ~g/ml of ACNU on the 4th day (b), and 10 ~g/rnl of 5-FU (c). Giemsa stain x200.

Results o f clinical cases

of G 2 + M phase cell was observed on the DNA histogram did not necessarily coincide with the appearance of morphological changes, i.e., G2+M phase cells markedly accumulated at the stage earlier than the appearance of morphological changes and at the appearance of morphological changes. Though stronger morphological changes occurred at the later stage, the accumulation of G2 + M phase cells had already disappeared (Fig. 5). In a comparative study of cell viability and morphological change, the morphological change became apparent with the decrease in cell viability following the changes of the DNA histogram, suggesting that such changes might result from the cytocydal effects of the antineoplastic agents (Fig. 6).

A representative case is shown in Fig. 7a, to which the present system was applied. A one-year old boy, who was admitted because of a consciousness disturbance and underwent a suboccipital craniotomy as a CT scan revealed medulloblastoma in the cerebellar vermis. A sensitivity test using ACNU, 5-FU, MMC and Cisplatin was performed for the surgical material from the above patient. Growth curves shown in Fig. 7b demonstrate growth inhibitions but no differences are seen among the agents. In flow cytometric studies, however, though there were no clear changes in the DNA histogram, and cell viabilities decreased with ACNU and Cisplatin, suggesting that ACNU and Cisplatin might be the best antineoplastic agents for this patient (Fig. 7d). Concomitant with the postoperative administration of ACNU, this patient's clinical symptoms and CT findings improved. Table 3 is the summary of this sensitivity test for 12 clinical cases: metastatic brain tumor, 6; malignant glioma, 5 and other. ACNU was selected as the best antineoplastic agent for 4 cases of glioblastoma, MMC and BLM for 4 cases of metastatic brain tumor. In 12 clinical cases of malignant brain tumors for which cultured cells antineoplastic agents were evaluated to be effective in vitro by FCM, effective antineoplastic agents were tried clinically. Judging from the performance status of the patients after 3 - 6 months and CT findings after contact with the agents, half of them showed a good prognosis, but

366 Table 2. Morphological chantes in the T9 rat glioma cell (according to antineoplastic agents). ACNU 10 #g/ml

5-FU 10 ,ug/ml

Abnormal shape of nuclei

Disappearance of cell processes

Vacuolization

Abnormal shape of nuclei

Disppearance of cell processes

Vacuolization

hrs

24

.

.

.

.

.

48 72

. ++

.

.

.

.

96 120 144

+++ +++ +++

-

4-

+ ++

+ + +

- : - 2 4 % , +: 2 5 - 4 9 % , + + : 5 0 - 7 4 % , + + + :

Change of the DNA histogram (T9 glioma cell )

ACNU10#g/rn(

5 FU10~:g/ml

24 hrs

DNA

,

~

,

content

48

.

.

.

++ ++ +++

i

+ + ++

100

i

~

control

i

=

96

+

+ ++ +++

Nitrosaurea compounds such as BCNU, ACNU, MCNU, and other antineoplastic agents have recently been used for the chemotherapy of brain tumors. ACNU, which was developed and is widely used in Japan, did not necessarily have good chemotherapeutic effects judging from our results which shows clinical improvements in 12 out of 20 glioblastoma patients (60%) and 8 out of 11 astrocytoma patients (~3%) by the combined therapy of ACNU and irradiation with Cobalt [9]. As an agent effective for one brain tumor case is not always effective for other tumor patients, and since hazardous side effects can be induced following the administration of an an-

.

i

-

Discussion

%

i

. +

75%-.

the other half showed a poor prognosis. In 3 glioblastoma cases out of 4 and in 1 ependymoblastoma case, the prognosis was correlated with the performance status and CT findings: there was no such correlation, however, in metastatic brain tumor cases.

control

.

50 5-FU (10#g/mt)

E

o 144 2c

4c

I

I

T

i

2c

4c

2c

4c

contact

24

hrs

- ACNU ( 1 0 # g / m l )

14

7'2 post

treatment

(contact

1 4,. :24hrs)

Fig. 5. Chronological changes in the DNA histogram following

Fig. 6. Chronological changes of cell viability following the ad-

the administration of ACNU and 5-FU.

ministration of ACNU and 5-FU.

367

Clinical

£

case

No. 255 (Medul Ioblastoma,

ly.o)

C 100-

%

20xi04.

10×104

B

~

Control 5-FU ACNU Cisplatin MMC

~

Cont rol 5-FU MMC

:0:tn

I ACNU MMC Cisplatin

w

i

i

2

3

4

5

6 day

D 1

2

3

4

5

6 day

Fig. 7. a. CT of 1-year old medulloblastoma patient, b, c, d. Results of our sensitivity test by FCM. Growth curves (B), fraction of G2+ M phase cell in the DNA histogram (C) and cell viability (D).

368 Table 3. Selection of antineoplastic agents (according to FCM and clinical application).

Case No.

Pathology

Effective agents by FCM

Clinical application

Prognosis ( 3 - 6 months later)

199 226 288 NM 259 282 222 243 274 235 395 329

Glioblastoma Glioblastoma Glioblastoma Metastatic tumor Hemangioendothelioma Ependymoblastoma Metastatic tumor Metastatic tumor Metastatic tumor Metastatic tumor Glioblastoma Metastatic tumor

ACNU ACNU, MMC ACNU 5-FU, MMC, BLM 5-FU, BLM, Cisplatin ACNU, 5-FU MMC, 5-FU, BLM MMC, 5-FU, BLM BLM ACNU, 5-FU ACNU 5-FU

ACNU ACNU ACNU MMC 5-FU MMC 5-FU 5-FU BLM 5-FU ACNU 5-FU

Good Good Good Good Good Good Poor Poor Poor Poor Poor Poor

tineoplastic agent, effective chemotherapy requires a decision on what is the best agent for a patient and prompt administration of it. Since it is impossible to remove all the cells of a malignant brain tumor by surgery, analytical study of the cell kinetics of the postoperative residual tumor becomes important for the prevention of its recurrence and for effective postoperative treatment. With the aid of current CT scanning, by which the location and size of a tumor can be precisely identified, postoperative follow-up and evaluation of chemotherapeutic effects are easier for brain tumors than for malignant tumors in other areas. Accordingly, if the best antineoplastic agent for each individual malignant brain tumor case can be selected based on the changes of cell kinetics according to the agents and their effects, and if this is followed up and reinforced by the findings of a CT scan, then this system, the sensitivity test, and its clinical application to malignant brain tumors will be of much value. There have been many reports of in vivo and in vitro studies on the sensitivity test for malignant tumors. In in vitro studies such as the INK method (which uses dehydrogenase as one of the enzymes related to respiration [10]), the Isotope method (in which synthesis of enzyme is used as the index of sensitivity [11]), and methods evaluating survival rates of the tumor cells or which utilize cell number counting or another method; however, these studies so far have not been clinically applied because of the

faults involved in each method. H u m a n tumor stem cell assay developed by Salmon [12] and Hamburger et al. [13] has recently been noted and von H o f f et al. [14] demonstrated the usefulness of it as a sensitivity test for antineoplastic agents. Inoue et al. [15], who introduced this method to Japan as the in vitro colony assay, reported its high correlation with clinical results. This assay, however, still needs further study on the identification of the clonogenic cell, or on the technique of colony formation, etc. before more effective and wider clinical applications can be made. Although in vivo studies using hatched chicken eggs [16] and nude mice have been noted because of their environmental resemblance to clinical tumors, their application is limited as there still remain various problems such as the low survival rates of tumor cells after inoculation or the means to judge the effects of antineoplastic agents. Studies on how to select the best antineoplastic agent for each individual malignant brain tumor case seems to be insufficient compared with that for carcinomas in other fields. There are, however, in vitro studies on sensitivity tests of brain tumor cells: a microtiter assay for BCNU by Kornblith [17] and a microtest plate for ACNU by Shibuya et al. [18]. Rosenblum et al. [19] reported the clonogenic cell positive to GFAP and the sensitivity to BCNU applying stem cell assay for brain tumor ceils. According to the report on an in vivo study of brain tumors by Shapiro et aL [20], in nude mice as an experimental

369 brain tumor model, the growth rate of transplanted tumors following explant generally paralleled those in the patients. Thus, methods for selecting the best antineoplastic agents are gradually being studied in the field of malignant brain tumors. The present study on a new sensitivity test of antineoplastic agents for monolayer-cultured brain tumor cells using FCM is based on the advantages of FCM - rapidity and simplicity. Thus, by using FCM, fluorescein-positive cells in a cell suspension can be rapidly analyzed quantitatively from the fluorescein strength and the obtained data are reproducible. Generally, FCM is widely used clinically for cancer cases; especially for the analysis of cell kinetics from the DNA histogram of cancer cells. In chemotherapy, some studies on the difference of sensitivities to antineoplastic agents at each phase in the cell cycle and on the inhibitory effects of antineoplastic agents on the progression of cell cycle through analyzing the changes of the DNA histogram using a cell suspension can be rapidly analyzed quantitatively from the fluorescein strength; the obtained data are reproducible. FCM is now widely used clinically for cancer cases, especially for the analysis of cell kinetics from the DNA histogram of the cancer cell. In chemotherapy, some studies have been done on the difference of sensitivities to antineoplastic agents at each phase in the cell cycle and on the inhibitory effects of antineoplastic agents on the progression of the cell cycle by analyzing the changes of the DNA histogram using FCM [21]. In Japan, combined therapy of ACNU and irradiation is clinically applied based on the changes of cell ',:inetics by the synchronization effects of ACNU [22]. Nomura et al. [23] reported BCNU- and ACNUinduced kinetic changes of brain tumor cells evaluated by the variation of S + G2+ M phase cells in the DNA histogram, referring to the usefulness of FCM for the sensitivity test. However, since the effects of the inhibition on cell cycle by antineoplastic agents might be considered as the cytostatic effects based on the degree of the DNA histogram if only the histogram is evaluated in chemotherapy, then the correlation between the effects on the DNA histogram and cytocydal effects of the agents is likely to be missed. It is reported that flow cytometric analysis of the antineoplastic agents requires a chronological follow-up of the cell cycle of the tumor [23]. It

is considered that, in addition to the evaluation of the effects on the DNA histogram based on the cell kinetic changes, indices for the cytocydal effects of the agents are necessary. When evaluating the present method using FCM, both the cytocydal and cytostatic effects of the antineoplastic agents on the tumor cells were used as the evaluation criteria. As the index to evaluate cytocydal effects, viability of tumor cells was measured by FDA (being changed into a fluorescent substance by hydrolase enzyme in the cytoplasm), instead of by conventional trypan-blue staining [24]. Viability of tumor cells, from which the degree of cytocydal effects of the anti-neoplastic agents become apparent, can be easily measured by FCM [6]. Such determination of viability by FDA could be widely used in the future. Thus, the present sensitivity test is characterized as follows: a) cytocydal effects can be evaluated by the FDA method, b) the assessment of cytostatic effects is possible even when a comparatively high viability is observed (i.e., when changes in the DNA histogram with a high viability by FDA suggest some cell cycle disturbance due to cytostatic effects of the antineoplastic agents), and c) no changes in the DNA histogram with a high viability suggest resistance of the tumor cells to antineoplastic agents. In comparative studies of various antineoplastic agents, slight differences in the sensitivity of each agent can be clearly demonstrated. The present test was first applied to established cell lines to evaluate its usefulness, and then followed by clinical application to 15 malignant brain tumor patients. As a result, the sensitivity or resistance to antineoplastic agents could be shown by our sensitivity test in individual cases. However, some problems still remained for the cultured cells of surgical materials. In particular, some tumor cells do not grow rapidly after inoculation and our present experiment requires a certain number of cells for some antineoplastic agents at the stage of the secondary culture. It currently takes about 2 or 3 weeks before the assessment of the sensitivity test because clinical specimens grow very slowly compared with the established cell lines. Antineoplastic agents were clinically applied to 12

370

patients, based on the present sensitivity test using FCM; the following results were obtained. The results of the sensitivity test and clinical application did not necessarily correlate with each other in metastatic brain tumors because of problems in the primary lesion or in the performance status of the patients; correlation was, however, observed in malignant brain tumors. The present test method could be useful for the selection of the best antineoplastic agents in individual cases. This study was presented at the 43th and 46th Meetings of the Japan Neurosurgical Society in 1984 and 1987 and at the 8th International Congress of Neurological Surgeons in 1985.

References 1. Kawamoto K, Herz F, Wolley RC, Hirano A, Kajikawa H, Koss LG: Flow cytometric analysis of the DNA distribution in human brain tumors. Acta Neuropath 46:39-44, 1979 2. Kawamoto K, Nishiyama T, Ikeda Y, Yamanouchi Y, Kawamura Y, Matsumura H, Hirano A, Herz F, Wolley RC: Flow cytometric studies of human brain tumors - Part I: Human malignant brain tumors. Neurol Surg 8:723-728, 1980 3. Kawamoto K, Herz F, Wolley RC, Hirano A, Koss LG: Flow cytometric analysis of the DNA content in cultured human brain tumor cells. Virchows Arch B Cell Path 35:11-17, 1980 4. Kawakami K, Kawamoto K, Oka N, Kawamura Y, Matsumura H, Ito T, Ohyama A: Flow cytometric studies on brain tumors - Part 5: New sensitivity test of antineoplastic agents for brain tumors and its clinical application. Neurol Surg 14:627-634, 1986 5. Pullen GR, Chalmers PJ, Nind APP, Nairn RC: Criteria of cell killing in vitro. J Immunol Methods 43:87-93, 1981 6. Oka N, Kawamoto K, Kawakami K, Fujiwara H, Shoda Y, Matsumura H: Experimental study on measurement of viability using fluorescein diacetate. Kansai Flow Cytometry 2:58-64, 1985 (in Japanese) 7. Vindelov LL: Flow microfluorometric analysis of nuclear DNA in ceils from solid tumors and cell suspensions. Virchows Arch B Cell Path 24:227-242, 1977 8. Yoshida J, Cravioto H, Ransohoff J: In vitro transformation of fetal brain cells from CDF rats exposed in utero to N-EthylN-nitrosourea: Morphological and immunologic studies. JNCI 64:1231-1239, 1980 9. Kawakami K, Kawamura Y, Kawamoto K, Oka N, Numa Y, Matsumura H, Akagi K, Ohyama A: Multidisciplinary therapy for gliomas. Experimental study using brain tumorbearing rats and analytical study of clinical cases. Neurol Med Chir (Tokyo) 26:133-139, 1986

10. Taguchi T: History and significance of sensitivity test of anticancer drugs. Cancer Chemother 9:570-574, 1982 11. Bickis I J, Henderson IWD, Quastel JH: Biochemical studies of human tumors. II. In vitro estimation of individual tumor sensitivity to anticancer agents. Cancer 19:103-113, 1966 12. Salmon SE, Hamburger AW, Soehnlen B, Durie BGM, A1berts DS, Moon TE: Quantitation of differential sensitivity of human-tumor stem ceils to anticander drugs. New Engl J Med 298:1321-1327, 1978 13. Hamburger AW, Salmon SE: Primary bioassay of human tumor stem cells. Science 197:461-463, 1977 14. von Hoff DD, Casper J, Bradley E, Sandbach J, Jones D, Makuch R: Association between human tumor colonyforming assay results and response of an individual patient's tumor to chemotherapy. Am J Med 70:1027-1032, 1981 15. Inoue M, Arakawa M, Ogawa K: Human tumor stem cell assay. Cancer Chemother 9:599-605, 1982 16. Nishiyama T, Kawamura Y, Kawamoto K, Matsumura H, Ito T, Ohyama A: Cultivation of human brain tumor cells on the chorioallantoic membrane of fertile chicken eggs. Neuropathology 3:13-28, 1982 17. Kornblith PL, Szypko PE: Variations in response of human brain tumors to BCNU in vitro. J Neurosurg 48:580-586, 1978 18. Shibuya N, Yoshida J, Kobayashi T, Kageyama N: Antitumor activity of ACNU against malignant glioma - A clinical application of in vitro sensitivity test of chemotherapeutic agent. Neurol Surg 11:361-367, 1983 19. Rosenblum ML, Dougherty DV, Deen DF, Wilson CB: Potentials and limitations of a clonogenic cell assay for human brain tumors. Cancer Treat Rep 65 (Suppl 2):61-66, 1981 20. Shapiro WR, Basler GA, Chernik NL, Posner JB: Human brain tumor transplantation into nude mice. J Natl Cancer Inst 62:447-453, 1979 21. Takamoto S, Ota K: Mechanism of action of antitumor agents by flow microfluorometry. Cancer Chemother 5:727-736, 1978 22. Shitara N, Kohno T, Nagamune A, Takakura K, Sano K: Pulsecytophotometric studies on experimental brain tumor under the effect of chemotherapeutic agents, microwave irradiation and hyperthermia. Neurol Med Chit (Tokyo) 18:199-207, 1978 23. Nomura K, Hoshino T, Knebel K, Deen DF, Barker M: BCNU-induced perturbations in the cell cycle of 9L rat brain tumor. Cancer Treat Rep 62:747-754, 1978 24. Ford WL: The preparation and labelling of lymphocytes. In DM Weir (ed) Handbook of Experimental Immunology, 3rd ed, Blackwell Scientific, Oxford, 1978, pp Chap 23, 1-22

Address for offprints." K. Kawamoto, Department of Neurosurgery, Kansal Medical University 1, Fumizonocho, Moriguchi, Osaka 570, Japan