Characteristics of three human gastric cancer cell lines, NU-GC-2, NU-GC-3 and NU-GC-4

Characteristics of Three Human Gastric Cancer Cell Lines, NU-GC-2, NU-GC-3 and NU-GC-4

Seiji AK1YAMA 1, Hiroyuki AMO 2, Tadashi WATANABE 3, Mutsushi MATSUYAMA 2, Junichi SA~MOTO 3, Munehisa IMAIZUMI 4, Hidehito ICmHASm 3,

Tatsuhei KONDO 3 and Hiroshi TAKAGI 3

ABSTRACT: Three human gastric cancer cell lines, NU-GC-2, NU-GC-3 and NU-GC-4 were established in vitro from the cancer tissues obtained from 3 patients during surgery. The pathological findings of the gastric tumors of these cases revealed poorly differentiated adenocarc inoma (and partial signet-ring cell carcinoma in the case of NU-GC-4). NU-GC-2 and NU-GC-4 were originally obtained from metastatic paragastric lymph nodes and NU-GC-3 was obtained from a metastatic tumor in the brachial muscle. The cells of NU-GC-2 and NU-GC-3 are polygonal in shape and grow as a monolayer sheet. NU-GC-4 cells, however, are mainly spherical in shape with a few free floating cells. Electron microscopy revealed epithelial characteristics in all 3 cell lines. The average doubling time o f NU-GC-2 was 36.1 hours, that of NU-GC-3 was 38.2 hours and that of NU-GC-4 was 29.9 hours. The modal chromosome number of NU-GC-2 was 62, that of NU-GC- 3 was 58 and those o f NU-GC-4 grown in in vitro and in vivo were 52-54 and 53, respectively. In vitro and in vivo lines of NU-GC-4 were established from the same tumor. These two cell lines are quite similar in morphology, but slightly different in karyotype. The in vitro sensitivity to anticancer agents was highest in NU-GC-4 and lowest in NU-GC-2. O f the anticancer agents, mitomycin C and adriamycin were most effective on the cells of all 3 cell lines.

KEY WORDS: gastric carcinoma, in vitro cell lines, in vivo cell lines, gastric cancer cell lines

INTRODUCTION

Department of Surgery, Nagoya National Hospital, Nagoya, Japan

aLaboratory of Ultrastructure Research, Aichi Cancer Center Research Institute, Nagoya, Japan

3Second Department of Surgery, School of Medicine, Nagoya University, Nagoya, Japan

4Department of Thoracic Surgery, School of Medi- cine, Nagoya University, Nagoya, Japan

Reprint requests to: Seiji Akiyama, MD, Department of Surgery, Nagoya National Hospital, 4-1-1, San- nomaru, Naka-ku, Nagoya 460, Japan

S i n c e gastric cancer is the most prevailing malignancy in Japan, many studies have been conducted on its diagnosis and treatment. The establishment of human gastric cancer cell lines may provide a powerful tool for the analysis of surface antigens of cells and the development of new anticancer agents for gastric cancer. To establish cell lines, however, is very difficult and only a few cell lines are available for study? We there-

JAPANESE JOURNAL OF SURGERY, VOL. 18, No. 4 pp. 438-446, 1988

Volume 18 Number 4 Gastric cancer cell lines 439

fore attempted to establish cancer cell l i ne s f rom fresh tumor specimens obtained at the time of operation. The results produced three in vitro cell lines and a transplantable tumor line in nude mice originating f rom three poorly differentiated adenocarcino- mas. In this paper, the characteristics o f these three cell lines are described.

MATERIALS AND METHODS

Source o f tumor cells NU-GC-2 was originally obtained from the

metastatic paragastric lymph nodes of a 56 year-old female patient in December, 1981. The tumor was diagnosed as poorly differen-

Fig. 1. Micrograph of a metastatic tumor in a paragastric lymph node which gave rise to NU-GC-2, showing an irregular glandular pattern. (HE, X100)

tiated adenocarc inoma (Fig. 1). NU-GC-3 was obtained f rom a metastatic tumor in the brachial muscle of a 72 year-old male patient in February, 1982 and this tumor was also diagnosed as poorly differentiated adenocarcinoma (Fig. 2). NU-GC-4 was obtained from the metastatic paragastric lymph nodes of a 35 year-old female patient in May, 1982. This tumor was diagnosed as poorly differentiated adenocarc inoma with areas of signet- ring cell carcinoma (Fig. 3).

Cell culture Fresh tumor tissues were scissored finely

and cell suspensions were obtained. Free cells o f these suspensions were seeded to culture bottles (Falcon Tissue Culture Flask, 25 cm 2, Becton Dickson, USA) and were cultured in a CO~ incubator (5 per cent CO~, 37~ 100 per cent moisture). The culture medium was RPMI 1640 supplemented with 20 per cent heat inactivated fetal calf serum (GIBCO Laboratories, USA) and penicillin- streptomycin solution (100 U / m l GIBCO Laboratories, USA). The culture medium was changed twice a week. Cells were observed through a phase contrast microscope.

Morphological studies Pr imary tumors a n d tumors grown in

BALB/c nude mice (Shizudokyo, J a p a n ) were fixed in 10 per cent formalin and embedded in paraffin. Sections were stained with hematoxylin and eosin. For electron

Fig. 2. Micrograph of a metastatic tumor in a brachial muscle which gave rise to NU-GC-3, showing a medullary pattern. (HE,.XIO0)

Fig. 3. Micrograph of a metastatic tumor in a paragastric lymph node which gave rise to NU-GC-4, showing a scirrhous pattern. (HE, X100)

440

microscopy, cell monolayers grown in plastic Falcon flasks were fixed in situ in 2.5 per cent glutaraldehide and 2 per cent paraformalde- hide in 0.05 M phosphate buffer, postfixed in 1 per cent OsO4, and dehydrated in graded ethanol. The monolayers were then removed from the bottom of the flasks after the inner surface of the plastic had been dissolved by propylene oxide and cell pellets were made by centr i fugat ion. T h e pellets were embedded in Epon 812. Thin sections were cut with an LKB ul t ra tome (LKB, Bromma, Sweden), stained with uranyl acetate and lead 2 and examined through a Hitachi HU- 12 electron microscope.

Chromosomal analysis Exponentially growing cells were treated

with colcemid solution (Grand Island Bio- logical, USA) for 3 hours at a final concentra- tion of 0.2 ktg/ml. Tumor cells dispersed by trypsin treatment were washed with phosphate-buffered saline (Ca free) and resus- pended for 20 minutes in a hypotonic solution (75 mM KC1). They were then fixed in acetic alcohol (acetic acid 1: methyl alcohol 3) with several changes , d r o p p e d on to chilled glass slides, and stained with Giemsa. Well-spread preparat ions o f metaphases were photographed and karyotyped (Inter- national System for H u m a n Cytogenetic Nomenclature, 1981). 3 G-banded prepara- tions were made with a slightly modified method of Sumner et al. 4

Cell growth 2.5 X 105 cells were plated in Falcon tissue

culture flasks and the medium was changed every second day. The cells were dispersed by 0.25 per cent trypsin and counted in hemocytometer chambers. The cell viability was assessed by trypan blue exclusion.

Hetero-transplantation 2 X 106 cells were inoculated in BALB/c

n u d e mice subcutaneously. For NU-GC-2, 150 /~1 o f anti-asialo GM 1 (Wako Junyaku, Japan) was administered intraperitoneally, 4 times before the tumor cell inoculation.

Detection of tumor markers Tumor markers in the used culture media

Jpn. J. Surg. Akiyama et al. July 1988

were measured by radioimmunoassay and the tumor markers in the unused medium were also measured as a control (Otsuka Assay Laboratories, Japan).

In vitro sensitivity to anticancer agents The sensitivities o f these cell lines to

anticancer agents were examined by the microplate labelling method? The anticancer agents used were mitomycin C (MMC, 10 /lg/ml), 5-fluorouracil (5-FU, 100 /tg/ml), adriamycin (ADM, 1 /~g/ml), nimustine hy- drochloride (ACNU, 1000/tg/ml), actinomy- cin D (ACTD, 0.1 /~g/ml), cytosine arabino- side (40497-S, 100 /~g/ml). They were dissolved in the culture medium and poured into the wells of the microplate. The sensitivity was shown by an inhibition index (I.I., positive >75 per cent) calculated by the following formula.

i . i . = ( l _ CPM(agent) ) CPM (control) X 100

RESULTS

Primary culture and establishment of cell lines NU-GC-2 and NU-GC-3 grew as a mono-

layer sheet and NU-GC-4 grew partly as free floating cells. Fibroblasts in NU-GC-2 and NU-GC-3 cultures were eliminated by using the difference in sensitivity to trypsin. The fibroblasts were more sensitive to trypsin than these cancer cells and dispersed in a shorter time. Most of the attached cells were cancer cells. NU-GC-4 was floating in an early stage of culture and it was very easy to eliminate the fibroblasts at that time. NU-GC- 2 has been maintained for 4 years and the passage number is 27. NU-GC-3 has been maintained for 3 years and 10 months and the passage number is 35. NU-GC-4 has been maintained for 3 years and 7 months and the passage number is 20.

Morphological studies The histological patterns of NU-GC-3 and

NU-GC-4 grown in nude mice were similar to those of the parent tumors. In vitro NU-GC-2 and NU-GC-3 were polygonal and NU-GC-4 was spherical in shape. The cytoplasm of

Volume 18 Number 4 Gastric cancer cell lines

NU-GC-2 cells was positive for PAS staining, however, the cells o f NU-GC-3 and NU-GC-4 were negative for PAS staining. All of the three cell lines were negative for Alcian blue staining. Submicroscopically, the in vitro cells of NU-GC-2 had poorly developed cytoplasmic organelle (Fig. 4), but those of NU-GC-3 and NU-GC-4 had well developed cytoplasmic organelle (Figs. 5, 6, 7). Numerous cytoplasmic fine filaments (Figs. 5, 6) and a few desmosome-like apparatuses (Fig. 6) were observed, but no mucous granules were found in these cell lines. Nu-GC-4 contained a few cells with an intracytoplasmic microcyst (Fig. 7).

Fig. 6. Electron micrograph of NU-GC-4 cells at passage 12, showing well developed endoplasmic reticulum and Golgi apparatuses, bundles of fine filaments and Desmosome-like apparatuses. (• loo)

441

Fig. 4. Electron micrograph of NU-GC-2 cells at passage 23, showing scattered bundles of cytoplasmic fine filaments. (•

Fig. 5. Electron micrograph of NU-GC-3 cells at passage 18, showing well developed Golgi apparatuses and many bundles of fine filaments. (X14,000)

Fig. 7. Electron micrograph of NU-GC-4 cells at passage 12, showing a cell with a large nucleolus, well developed cytoplasmic organelle and a large intracytoplasmic microcyst. (X3,400)

Chromosomal analysis The cells of all four lines had excess

chromosomes, but no marker chromosomes were found. The modal number of NU-GC-2 was 62 (Fig. 8A) and that o f NU-GC-3 was 58 (Fig. 8B). In vitro NU-GC-4 had a wide range of chromosomal numbers with a modal number range o f 52-54 (Figs. 8C, 9) and in vivo NU-GC-4 had a narrow range, with a modal number of 53 (Fig. 8D). Trisomy was con- stantly found in the chromosomes of groups

442 Akiyama et al. 1988

20,

0 54 56 58 60 62 64 66 86 90--93 120<-

1 2 3 4 5

lill tli1 l l t l i i l l I l l i I l i t till 6 7 8 9 10 11 12

B

60, ~o

5o. e.D

40'

30.

~ 2o. m lO-

0 54 56 58 60 62 64 66 86 90~93120<-

Chromosome Number

C

50 t 40 '~ 3O

~ ]0-

d ~ a a Jib i l l t ~ x l 13 14 15 16 17 1 f3

19 20 21 22 X X Fig. 9. Karyotype of in vitro NU-GC-4,

showing trisomies in Nos. 2, 4, 8, 19, and 20 chromosomes, and quadrisomy in No. 11 chromosome.

0 <~4950 52 54 56 58 60 92 99 102<~

D

!::1 ] I

0 <-49 50 52 54 56 58 60 92 99 102_< Chromosome Number

Fig. 8. Distribution of the cells of NU-GC- 2 (A), NU-GC-3 (B), NU-GC-4 (C), and in vivo NU-GC-4 (D), with various chromosome numbers.

A, C, E and F in all 4 cell lines and sometimes in those o f groups B, D and G. Monosomy was found in group D and G in NU-GC-2 and

107

10 ~ .D

J i I i 1 0 ~ 2 ~ 0 8 10

Culture Days

Fig. 10. Growth curves of NU-GC-2 (O), NU-GC-3 (&) and NU-GC-4 ( l ) .

in the in vivo NU-GC-4. Minute chromosomes and 4p(--) in NU-GC-2 and 5p(--) in NU-GC-3 were also observed.

Cell growth T h e rates of cell growth slightly differed

Volume 18 Number 4 Gastric. cancer cell lines 443

a m o n g the 3 celt lines (Fig. 10). T h e doubl ing time o f NU-GC-2 was 36.1 hours , that o f N U - GC-3 was 38.2 hours and that o f NU-GC-4 was 29.9 hours.

Hetero-transplantation NU-GC-3 and NU-GC-4 were transplant-

able to nude mice, deve lop ing tumors over 15 m m in diameter, 2 weeks after the tumor inoculat ion. NU-GC-2 was no t t ransplantable to nude mice even t h o u g h they had b e e n pretreated with anti-asialo GM 1 antibody.

Tumor markers Alpha-fe toprote in (AFP), carc inoembryo-

nic ant igen (CEA), ferritin, a n d beta-2-micro- globulin (BMG) were measu red (Table 1). AFP showed a h igh titer in the used culture med ium o f NU-GC-4 a nd CEA showed high titers in the used media o f all cell lines. Ferritin showed high titers in NU-GC-3 and NU-GC-4. BMG showed a h igh titer in NU- GC-2. T h e AFP titer was h igh in the serum o f the pat ient with NU-GC-2, but was low in the used culture medium.

In vitro sensitivities to anticancer agents The results o f the sensitivity test are shown

in Table 2. NU-GC-4 was the mos t sensitive cell line to the ant icancer agents and NU-GC-

2 was the least sensitive. All o f the three cell lines were sensitive to MMC and ADM. NU- GC-3 and NU-GC-4 were sensitive to ACTD and CA, and NU-GC-4 was sensitive to 5-FU. None o f the cell lines were sensitive to ACNU or 40497-S.

DISCUSSION

T h r e e h u m a n gastric cancer cell lines have b e e n establ ished f rom surgically re- sected tumor tissues and the in vitro growth properties, morpho logy , karyotype, hetero- t ransplantat ion, t u m o r markers and sensitivity to ant icancer agents have b e e n charac- terized. The es tabl i shment o f a h u m a n gastric cancer cell l ine is very difficult and only 18 such cell lines have b e e n established in the world (Table 3). T h e major problems o f es tabl ishment are con tamina t ion o f fibroblasts and infection. Since the tumor specimens f rom the digestive tract tend to be infected, it is bet ter to use cancer cells f rom the metastatic lymph nodes. Cancer cells o b t a i n e d f rom pleural effusion or ascites are also of ten con tamina t ed with mesothelial cells and fibroblasts. Since mesothel ial cells

Table 1. Titers of Tumor Markers in Media Cultured NU-GC-2, NU-GC-3 and NU-GC-4

Tumor Marker Titers in Cultured Media

NU-GC-2 NU-GC-3 NU-GC-4 None (control)

AFP (ng/ml) 5.0> 5.0> 8.0 5.0> CEA (ng/ml) 13.1 5.8 4.0 2.5 Ferritin (ng/ml) 1.5> 6.0 4.5 1.5> BMG (ng/ml) 0.9 0.5> 0.5> 0.5>

Table 2. In vitro Sensitivities ofNU-GC-2, NU-GC-3 and NU-GC-4 to Anticancer Agents

Sensitivities Anticancer Agents NU-GC-2 NU-GC-3 NU-GC-4

MMC ( 10/~g/ml) + + + 5-FU (100/~g/ml) -- -- + ADM ( 1/~g/lnl) + + + ACNU ( 1/~g/ml) -- - -- ACTD ( 0.1/~g/ml) -- + + CA (100/~g/ml) -- + + 40497-S (100/~g/ml) -- -- --

+, positive _-->75% 'I.I.); --, negative <75% (I.I.)

444 Akiyama et al. JPn)uJl'~ Surg. 1988

[...

�9

~. ~ . ~ ~. ~. ~. ~. ~. ~ ~ ~ . ~ - ~

~ N N N ~ N ~

~ . ~ ~ ~ . ~ ~ ~ 0 0 0 0 0 0 0 0 0

! d ~ ~

0

"r.

�9

. . . . ~ ~ ~ T ~ . . . . ~ . ~ T ~ ~ ~

o ~ ~ = ~ 0 � 9 1 6 9

~_. 04 Cc~ ~: ~

Volume 18 Number 4

Gastric cancer cell lines 445

tend to proliferate more easily than cancer cells, it is very difficult t o o b t a i n only cancer cells f rom the pr imary cultured cells. 8 of the 17 gastric cancer cell lines were established f rom metastatic lymph nodes, 4 f rom the primary tumor, 4 f rom ascites or pleural effusion and 1 f rom liver metastasis. Con- sidering these figures, the cancer tissue f rom metastatic lymph node is thought to be the best material for the establishment of gastric cancer cell lines. The elimination of fibroblasts is also very difficult and troublesome. C o n c e r n i n g NU-GC-2 and NU-GC-3, the fibroblasts were eliminated by the difference in sensitivity to trypsin between the cancer cells and fibroblasts. Since NU-GC-4 was composed of floating cells, we could easily obtain a cancer cell suspension. Laboisse et al. established a h u m a n gastric cancer cell line using soft agar 6 and they could easily eliminate fibroblasts using this technique.

Morphologically, most of the 18 cell lines are polygonal and attach to the bot tom of the flask. The only floating, round-shaped cell lines are Kato-III, Okajima and NU-GC-4. The cells o f Kato-III contained no adhesive apparatus, including desmosome, but the cells of NU-GC-4 and Okajima became attached by desmosome-like apparatuses. 7,8

Chromosomal abnormalit ies have been seen in all of the cell lines, including the three lines of the present study (Table 3), however, the details o f these abnormali t ies varied from cell line to cell line, and no clear common tendency was found. Two cell lines, MKN-7 and HGT-1, had marker chromosomes, but the others did not. The modal chromosomal n u m b e r of these cell lines exceeds 2n in all except 5 cell lines (MKN-1, MKN-45, MKN-74, AGS and KWS-1).

The cell growth property slightly differed among the three cell lines. NU-GC-4 had the fastest line in cell growth and its doubling time was 29.9 hours. (The doubling time of most gastric cancer cell lines established in the world range f rom 20 to 40 hours-Table 3).

Most cell l ines are capable of he tero-

transplantation, but a few are not (Table 3). Likewise, NU-GC-2 in the present study was not tumorigenic in nude mice and treatment with anti-asialo GM 1 ant ibody was unsuccess- ful in affecting the transplantability of this cell line into nude mice.

It was shown in the present study that the cells of these three cell lines produced either AFP, CEA, ferritin a n d / o r BMG. A high AFP titer was found in the serum of the patient with NU-GC-2 whereas none was found in the cultured medium. High CEA titers were also found in the cultured media of all 3 lines. In other gastric cancer cell lines, only HGC-Y1 cells have b e e n shown to secrete CEA in cultured medium. 9 In the other cell lines, no description of CEA was made, except in HGT-1, in which CEA was negative in the cultured medium, cultured cells, parent tumor and tumors grown in nude mice.

The sensitivity to ant icancer agents was quite different in all three cell lines and this finding may be analogous to the fact that chemotherapeut ic effects differ f rom patient to patient even among those with the same kind of cancer. A sensitivity test should therefore be done pr ior to the commence- ment of cancer chemotherapy.

ACKNOWLEDGMENTS

This work was supported in part by a Grant-in-aid for Cancer Research f rom the Ministry of Education, Science and Culture, and f rom the Ministry of Heal th and Wel- fare. We thank Mrs. H. Fukami and Mrs. C. Yamada for expert technical assistance.

(Received for publication on Sep. 9, 1987)

REFERENCES

1. Sekiguchi M, Suzuki T. List of human cancer cell lines established and maintained in Japan. Soshiki Baiyo 1980; 6: 527-547. (in Japanese)

2. Sato T. A modified method for lead staining of thin sections. J Electron Microscopy Jpn 1968; 17: 158-159.

446 Akiyama et al. jp% Surg 1988

3. Committee Human Cytogenetic Nomenclature. An in te rna t iona l system for h u m a n cytogenetic nomenclature. Cytogenet Cell Genet 1978; 21: 313-398.

4. Sumnar AT, Evans HJ, Buchland RA. New technique for distinguishing between human chromosomes. Nature 1971; 232: 31-32.

5. Akiyama S, Ichihashi H, Imaizumi M, Watanabe T, Kondo T. In vitro sensitivity test of anticancer drugs and their clinical application. Nippon Gan- chiryo Gakkai Zasshi (] Jpn Soc Cancer Ther) 1985; 20: 1285-1293. (English Abst.)

6. Laboisse CL, Augeron C, Coutrier-Turpin M, Gespach C, Cheret A, Potet F. Characterization of a newly established human gastric cancer cell line HGT-1 bearing histamine H2-receptors. Cancer Res 1982; 42: 1541-1548.

7. Sekiguchi M, Sakakibara K, Fujii G. Establishment of cultured cell lines derived from a human gastric carcinoma. J p n J Exp Med 1978; 48: 61-68.

8. Motoyama T, Hojo H, Watanabe H. Comparison of seven cell lines derived from human gastric car- cinomas. Acta PatholJpn 1986; 36: 65-83.

9. Yamaguchi N, Kawai K. Factors affecting the CEA secretions of human adenocarcinoma cell lines into the spent medium. Gastroenterologia Japo- nica 1983; 18: 428-435.

10. Barranco SC, Townsend CMJr, Casartelli C, Macik BG, Burger NL, Boerwinkle WR, Gourley WK.

Establishment and characterization of an in vitro model system for human adenocarcinoma of the stomach. Cancer Res 1983; 43: 1703-1709.

11. Dobrynin YaV. Establishment and characteristics of cell strains from some epithelial tumors of human origin.J Natl Cancer Inst 1963; 31: 1173-1195.

12. Hojo H. Establishment of cultured cell lines of human stomach cancer origin and their morphological characteristics. Niigata Igakkai Zasshi 197,7; 91: 737-763. (in Japanese)

13. Maunoury R. Etablissement et caracterization cellu- laires humaines derivees de tumours metastatiques intracerebrales. C R Acad Sci Paris 1980; 284: 991-994.

14. Nomura H, Tokumitsu S, Takeuchi T. Ultrastruc- rural, cytochemical and biochemical characterization of alpha-amylase produced by human gastric cancer cells in vitro. J Natl Cancer Inst 1980; 64: 1015-1024.

15. Tokumitsu S, Tokumitsu K, Kohnoe K, Takeuchi T. Characterization of liver-type alkaline phospha- tase from human gastric carcinoma cells (KMK-2) in vitro. Cancer Res 1979; 39: 4732-4738.

16. Oboshi S, Seido T, Shibata H. In vitro culture of human cancer cells for cancer chemotherapy. A long term cultured suspended cell line established from lymph nodes with metastasis from gastric cancer. Gann 1969; 60: 205-210.

Documents

Characteristics of three human gastric cancer cell lines, NU-GC-2, NU-GC-3 and NU-GC-4