215 Cancer Letters, (1991) 215 - 220

Elsevier Scientific Publishers Ireland Ltd.

y-Interferon and retinoic acid synergize in inhibiting the growth of human neuroblastoma cells in nude mice

Paolo Cornaglia-Ferraris, Gian Luigi Mariottini, and Mirco Ponzoni

Pediatric Oncology Research Laboratory. G. Gash Children’s Hospital, 16148 Genoa (Italy)

(Received 9 July 1991) (Revision received 18 September 1991) (Accepted 1 October 1991)

Summary

We have investigated the effects of retinoic acid @A), human recombinant gamma in- terferon f-y-IFN), and the association of both agents on the growth of human neuroblastoma (Nf3) cells in [CDl(nu/nu)] nude mice. Two human NB cell lines, namely LAN-5 and Gl- Ll-N, were previously adapted to grow in syngeneic animals for 7 consecutive passages. At the eighth passage, only animals which developed lo-mm diameter tumors within 40 days from xenograft were admitted to the study. RA and/or y-IFN were administered subcutaneously 3 - 5 days per week for 3 con- secutive weeks. The number of days necessary for each tumor mass to grow up to 20 mm di- ameter (in uiuo doubling time, ivDT) was then evaluated. Tumor growth was significantly in- hibited in y-1FN (P < 0.005) and RA (P < 0.05) treated mice grafted with Gl-Ll-N. The combination of the two agents did not further enhance iuDT. The tumor growth inhibition was not statistically significant in LAN-5 bear- ing mice treated with RA or y-1FN alone, while a synergistic effect between the two drugs was observed (P < 0.05). We conclude that par- enteral combined administration of RA and y-

Correspondence to: P. Cornaglia-Fenaris, Pediatric Oncology

Research Laboratory, G. Gaslini Children’s Hospital, 16148

Genoa, Italy.

1FN may proue to be useful in inhibiting the growth of tumors derived from human NB cells resistant to single inducers.

Keywords: y-interferon; retinoic acid; neuroblastoma cells; nude mice

Introduction

Neuroblastoma (NB) is the most common extra-cranial solid tumor of childhood account- ing for 100 - 120 new cases per year in Italy [8]. The disseminated form of the disease has an extremely poor prognosis in children over 1 year of age [5]. During the last 5 years, new therapeutic protocols including high dose chemotherapy, followed by autologous bone marrow transplantation [5] and/or interleukin- Z-LAK cells administration [7], have signifi- cantly modified the disease-free survival. How- ever, the clinical outcome remains unfavorable for at least 70-80% of the patients [5,7,8]. Thus, the development of research models ex- ploring new therapeutic strategies is presently necessary.

NB cells derive from neural crest elements and are capable of spontaneous or induced dif- ferentiation both in vitro and in vivo [3,6,9, 131. These observations present the possibility that pharmacologically-induced differentiation may be clinically useful in controlling the dis-

0304.3835/92/$05.00 0 1992 Elsevier Scientific Publishers Ireland Ltd

Printed and Published in Ireland

ease spread after surgically removing the pri- mary tumor mass. Based upon this strategy, few clinical studies have already been publish- ed, where only partial or very limited response to therapy has been reported, possibly due to the advanced stage of the disease affecting pa- tients admitted to the various clinical trials [lO,ll]. On this regard, another possibility to be considered is that there is not a simple rela- tionship between the degree of cellular dif- ferentiation and the clinical prognosis of neuro- blastoma [ 11.

In the last few years, several synthetic and recombinant molecules have been tested by our group in various established NB cell lines in vitro including all-trans retinoic acid (RA), cytosine-arabinoside (ARA-C) , gamma-inter- feron (T-IFN), nerve growth factor (NGF), and other cytokines [ 14 - 171. The differentiation- induction potential of the various agents have been extensively tested in vitro at various doses and times. Among all agents, RA and y-IFN seem to be th e most promising induc- ers. In vitro, cellular proliferation of LAN-5 and GI-LI-N NB cell lines, evaluated as [3H]- thymidine incorporation, is significantly modi- fied by RA, -r-IFN and even more by the association of both agents [17,X3], as in- dicated also by others [20], suggesting that in- duced cells have a reduced tumorigenic potential. Moreover, LAN-5 cells differentiated in vitro showing dramatic changes at both mor- phologic and biochemical level [ 171. These fin- dings encouraged us to evaluate in nude mice the impact of new different therapeutic schedules such as daily S.C. peritumoral ad- ministrations of various doses of RA and -r-IFN as single agents or in combination, in order to determine whether this experimental thera- peutic approach would influence the in vivo tumor growth.

Materials and Methods

Chemicals Human recombinant y-IFN was obtained

from Genzyme Corporation (Boston, MA, U.S.A.) and kept in aliquots (1000 units/PI) in phosphate buffered saline (PBS) (Flow Lab-

oratories, Milan, Italy) plus 0.1% bovine serum albumine (BSA) at -8OOC. All-trans- RA was purchased from Sigma (St. Louis, MO, U.S.A.) and kept in aliquots in ethanol (1 mg/ml) at -3OOC for 1 week. All other chemicals used were from Sigma.

Human NB cell lines The LAN-5 NB cell line was originally ob-

tained from the bone marrow metastases of a 4-month-old infant and kindly provided to us by R. Seeger [19]. GI-LI-N cell line was deriv- ed from peripheral blood metastatic NB cells and established in our laboratory as described [12]. Both cell lines were grown in a hum- idified incubator at 37OC in a 5% CO2 at- mosphere and maintained in RPMI-1640 me- dium (Flow) supplemented with 15% fetal calf serum (FCS) (PAA Labor, Technogenetics, Turin, Italy), penicillin (50 II-J/ml) and strep- tomycin (50 pg/ml), as described elsewhere in detail [ 161. Cells were prepared for inoculation by washing in sterile PBS (Flow), resuspending in fresh medium and counting with a hemo- cytometer.

Mice Seven-week-old CD1 nu/nu female athy-

mic nude mice were obtained from Charles River nude mice colony (Charles River, Calco, Italy) 1 week before each experiment. Mice were housed in a barrier facility in sterile isolated cages (Techniplast, Gazzada, Italy). All manipulation of cages and animals were conducted in laminar air flow cabinets. Animals received a standard sterile diet. Drink water was supplemented with a poly-vitamin complex and gentamycin every other week. All animals were randomly assigned to four dif- ferent treatment groups just before tumor in- oculum. Animals were inoculated by subcut- aneous dorsal injection of 3 x lo6 cells in 0.1 ml sterile saline. Cell lines were established in mice matched for sex and age for 7 subsequent passages in order to reduce variations in laten- cy time (LT). LT is here considered as the number of days necessary to develop a 10 mm diameter S.C. tumor mass, while survival time (ST) is the number of days necessary to

develop a 20 mm diameter S.C. tumor. From 3 to 5 mice per passage were used. A total of 62 animals were injected. Tumors from the last group of mice (passage no. 7) were surgically removed, mechanically minced in fresh culture medium, washed, pooled, counted and resus- pended as described above. Three to five mice per group were then injected with GI-LI-N or LAN-5 NB cells in three subsequent ex- periments. At the VIII passage 45 animals were injected, 37 of which were admitted to the protocols. Both protocols were utilized in GI-LI-N carrying mice while only protocol 2 in LAN-5 grafted animals.

ing mixing RA and y-IFN in the same syringe. Treatment was discontinued after 3 weeks.

(C) Protocol 2. The same amount of the previous weekly dose of RA was administered every other day. The weekly dose of y-IFN was augmented to 5.000 units/mouse per day. Animals were treated for 3 days a week for 3 consecutive weeks.

Therapeutic protocols (A) Admission/exclusion criteria. Only mice

which developed a lo-mm diameter S.C. tumor mass between 15 and 40 days from in- oculum were admitted to the study.

(0) iuDT evaluation. The in vivo doubling time was evaluated by recording the number of days necessary for developing a 20-mm diam- eter subcutaneous mass (survival time, ST) using calipers and subtracting from it the laten- cy time according to the simple formula: ivDT = ST - LT. ivDT was recorded for each treated animal. Control and statistical anal- ysis were performed using Student’s t-test for independent samples with non-homoge- neous variance.

(B) Protocol 1. Mice were injected S.C. daily per 5 days a week with placebo (solvent or culture medium), 30 ng RA/mouse per day, 1000 units y-IFN/mouse per day or the association of both inducers. Injections (vol- ume 100 ~1) were given peri-tumorally avoid-

Results

In uiuo passages Changes observed in LT and ST by stabiliz-

ing GI-LI-N and LAN-5 for 7 consequent in vivo passages are summarized in Table I. A

Table 1. Growth of human NB cell lines in nude mice !. A total of 62 CD1 nu/nu female 7-week-old mice were in-

217

oculated (3 - 5 animals per passage) for 7 consecutive passages.

Passage no. LT ST ivDT

Gl-Ll-N

II 111 IV

V VI VII

LAN-5 I II III

IV V VI VII

76.33 + 38.99 40.0 f 9.89 33.0 + 7.02 22.33 + 5.51

24.0 ztz 9.52 17.60 zt 5.59 26.54 f 8.54

65.30 ztz 46.70 28.64 zt 18.59 28.75 f 3.86

67.67 ztz 3.12 33.67 f 13.58 22.08 ztz 6.87 14.25 + 3.86

136.67 + 42.52

94.0 f 45.25

64.33 + 4.16

41.0 f 3.19

42.67 f 6.03

38.43 + 4.30

43.50 f 2.12

136.43 zt 48.19

57.50 f 12.02 45.03 f 13.63

97.33 f 21.08 56.03 +z 23.58 36.64 f 14.01 34.50 i 2.08

60.33 zt 17.47

54.0 f 35.35

29.67 + 5.03

21.50 zt 3.53

22.0 f 2.65

20.40 + 3.91

22.03 f 2.83

72.13 +z 28.39

29.50 + 12.02 17.03 + 4.63

29.67 ztz 5.77 22.33 zt 11.06

13.6 +z 7.86

23.25 ztz 2.36

218

progressive reduction of both LT and ST can be observed in both cell lines together with a significant reduction of inter-animal variations. A stabilization of both parameters was obtained after 5 - 6 passages for both LAN-5 and GI-LI- N engrafted mice.

Effects of protocol 1 and 11 in Gl-Ll-N bearing mice

At the end of the third week of treatment it was found that ivDT in controls was 18.7 f 0.9 days, while ST was 36.5 f 1.9 days. A significant rise in ST and ivDT was observed in mice treated with RA and y-IFN regardless of protocol assignment which allowed the results achieved to be pooled. Drug combination did not substantially modify the -y-IFN induced tumor growth inhibition observed with the single agents (Table II).

Effects of protocol 11 in LAN-5 bearing mice The changes observed in both RA and y-IFN

treated mice were not statistically significant (Table II), possibly due to the small number of animals tested. However, by combining the in- ducers, a significant change in ivDT was finally observed (P < 0.05).

Table II. Effect of RA, y-IFN and drug combination on growth of human NB cell lines in 37 nude mice.

Treatment group In vivo DT P

Gl-Ll-N Control 18.75 f 0.95 RA 25.16 f 4.4 < 0.05 y-IFN 31.5 f 5.4 < 0.005 yIFN + RA 27.0 zt 4.85 < 0.01

LAN-5 Control 17.1 f 3.8 RA 18.4 zt 3.6 ns. yIFN 15.5 f 3.5 n.s.

y-IFN + RA 25.0 zt 3.9 < 0.05

Discussion

The data presented in this report demon- strate that the proliferation of human NB cell lines in vivo can be significantly inhibited by the combined parenteral administration of y-IFN and RA, while the drugs alone do not seem to be always effective. The results are consistent with other studies showing inhibition by RA of in vivo NB cell growth [l]. In addition, this report compares the in vivo activity of RA and -y-IFN with th e combination of both agents. Results achieved with -r-IFN are particularly in- teresting since a lack of cross reactivity be- tween the recombinant human lymphokine and murine cellular receptor has been clearly documented [Z]. Thus, the inhibition of tumor growth which reached a statistical significance only in GI-LI-N grafted mice should be ascrib- ed to the drug-induced cellular differentiation.

It is possible that due to the clinical trials, tumor masses which already reached a con- siderable volume (10 mm diameter has been here considered the starting point for treat- ment) could not be penetrated by the drug in spite of S.C. peritumoral injection. The ratio be- tween cell number and bioavailability of the drug is certainly critical as it can be deduced from experiments reported by Abemayor et al. who tested the growth of LAN-5 cells in nude mice orally treated with RA [ 11. These authors observed a statistical significant inhibition of tumor growth that could be ascribed either to the fact that the treatment was scheduled to begin when subcutaneous tumor was limited to 5 mm diameter or because it was prolonged for more than 21 days. From the series of ex- periments reported by Abemayor et al., it was actually possible to conclude that retinoids in- hibit the growth of LAN-5 cells in vivo. How- ever, a statistically significant inhibition was reached only after 15 days of oral treatment and became prominent after 30 days. Thus, it appears that both length of treatment and dose/mass ratio are essential parameters to be considered in this animal model. It should be noticed on this regard that the appreciation and measurement of 5 mm diameter tumor in

nude mice is not always easy and the standard error in evaluating the real tumor mass may reach 20-25% (* 1.0 - 1.5 mm) of the en- tire neoplasm. By treating animals when larger tumor masses are present, this error can be avoided or reduced. However, as evident from our protocol, one should be prepared to face difficulties arising from drug-availability to the central part of the tumor.

Another important consideration should be made when comparing the in, vitro versus in, vivo sensitivity of NB cell lines to differentiation inducers. GI-LI-N cells actually appears to be much less sensitive than LAN-5 cells in vitro to both RA and y-IFN as far as [3H]thymidine in- corporation are concerned [18]. while a dif- ferent drug sensitivity for both cell lines has been here recorded in vivo. Thus, the much pronounced synergistic effect observed in vitro may be due to a culture artifact rather than resulting from a biologically induced process. Alternatively, a cascade of events is likely to occur when a growth/inhibition factor stimul- ates a cell. It appears that many of the cellular responses known to take place in vitro can be explained as secondary responses to activation of serum factors. However, it must also be considered that in vitro culture conditions allow drug/cells concentration ratios to op- timize, which was probably not achievable in our experiments with either protocol I or II.

Significantly, both in vitro and in vivo, the combined differentiation protocol was always effective on both cell lines.

Preliminary histopathologic findings observ- ed in tumors removed from untreated GI-LI-N- carrying mice were composed of a fairly homogeneous population of small cells with scanty cytoplasm having generally indistinct borders and round to oval nuclei (C. Cozzutto, unpublished data). Necrosis and picnosis were both present together with a high number of metaphases. Pseudorosettes were very rarely observed. In y-IFN-treated mice prominent changes were observed in nuclei which were slightly larger and more pleomorphic. More- over, cytoplasm was more abundant, ap- pearance of neurofibrillary stroma was clearly

219

noticed and pseudorosettes were frequently observed. Less relevant changes were observ- ed in LAN-5 treated mice.

From the present series of experiments we conclude that the combined administration of RA and y-IFN inhibits the growth of both highly malignant, N-myc amplified, human NB cell lines herein adapted to grow in vivo. How- ever, the lack of sensitivity of y-IFN murine receptor to recombinant human y-IFN make impossible to exclude undesired toxic effects which presently represent a major obstacle for systemic administration of the recombinant lymphokine.

Acknowledgements

We are grateful to C. Cozzutto and M. Brisigotti for providing unpublished data on histologic features of human NB in nude mice. We also thank A. Encalada for animal care and L. Malacrida for preparation of the manuscript. This work has been supported by AIRC, Italian Association for Cancer Research and G. Gaslini grant 901702C.

References

Abemayor. E. and Sidell.. N. (1989) Human neuro-

blastoma cell lines as models for the in vitro study of

neoplastic and neuronal cell differentiation. Environ.

Health Perspect., 80. 3- 15.

Aguet. M., Dembic, 2. and Merlin, G. (1988) Molecular

cloning and expression of the human interferon-g receptor.

Cell. 55, 273-280.

Cushing, H. and Wollbach, S.B. (1927) The transforma-

tion of a malignant paravertebral symphaticoblastoma into

a benign ganglioneuroma. Am J. Pathol. 3. 203-215.

D’Angio, C.J.. Evans, A.E. and Koop. C.E. (1971) Spe-

cial pattern of widespread neuroblastoma with a favorable

prognosis. Lancet i. 1046- 1049.

Dini. G.. Lanino, E., Garaventa. A., Paolucci, P., Amici

A., Locatelli. F., Coleselli. P., Boni, L., Dallorso, S..

Abla. 0. and De Bernardi, B. (1991) Unpurged auto-

logous bone marrow transplantation for neuroblastoma.

The A.I.E.O.P. BMT group experience. Proceedings of

the International Course on BMT in Children, vol. 7.

suppl. 3, pp. 109 - 111. Science and Medical Division

MacMillan Press Ltd., London.

Evans, A.E., Gerson. J. and Schnanfer, L. (1976). Spon-

taneous regression of neuroblastoma. Natl. Cancer Inst. Monogr. 44. 49 - 54.

220

7

8

9

10

11

12

13

14

15

Favrot, M.. Floret, D., Negrier. S.. Cochat. P.. Bouffet.

E., Zhou Dacheng, Franks, C.R.. Bijman. N J.T Brunat-

Mentigny. M.. Philip. I. and PhIlip. T (1989) Transplan- tation 4, 499-503.

Garaventa, A., Guerra. P.. Arrighini. A , Bertolazzi, L , Bestagno. M , De Bernardi. B , Lanino. E., Villavecchia.

G.P. and Claudiani, F. (1991) Treatment of advanced

neuroblastoma with l-131 Metaiodobenzylguanidine Can-

cer 67. 4. 922 -928.

Haas. D. (1988) Complete pathologic maturation and

regression of the stage IV-S neuroblastoma without treat-

ment. Cancer 62, 818 - 825.

Helson. L. Investigational chemotherapy of neuro-

blastoma. (1987) J. FI. Med. Assoc., 66, 284 - 287.

imashukii, S.. Sugano. T.. Fujiwara, K.. Todo. S., Ogita.

S. and Goto. Y. (1982) Intraaortic prostaglandin E, mfu-

sion in maturation of neuroblastoma. Experientia, 38.

932 - 933.

Longo. L , Christiansen, H.. Chrtstiansen. N.M.,

Cornaglia-Ferraris. P. and Lampert. F. (1988) N-myc

amplification at chromosome band 1~32 in neuroblastoma

cells as investigated by in situ hybridization. J. Cancer Res.

Clin. Oncol. 114. 636-640.

Lopez-lbor. B. and Schwartz. A D (1985) Neuro-

blastoma. Pediatr. Clin. N. Am., 32, 755- 788.

Parodi, M.T., Cornaglia-Ferraris. P. and Ponzoni.. M.

(1989) Effects of y-Interferon on the growth. morphology,

and membrane and cytoskeletal proteins expression of

LAN-l cells. Exp. Cell Res.. 185, 327 -341.

Ponzoni. M., Melodia. A., Cirillo. C., Casalaro. A and

16

17

18

19

20

Cornaglia-Ferraris, P. (1987) Effect of Cytosine Arab-

inoside on the growth and phenotypic expression of GI-

ME-N, a new human neuroblastoma cell line. In. Advances

in Neuroblastoma Research 2.. pp 437, Editors: A.E. Ev-

ans, G.J. D’Angio. A.G. Knudson and R.C. Seeger. Alan

R. Liss. New York.

Ponzoni, M.. Lanciotti, M.. Meclodia, A., Casalaro. A.

and Cornaglia-Ferraris, P (1989) Morphologic and

phenotypic changes of human neuroblastoma cells in

culture induced by Cytosine Arabinoside. Exp. Cell Res..

181, 226-237.

Ponzoni, M.. Lanciotti. M.. Montaldo. P.G. and

Cornaglia-Ferraris, P. (1991) Gamma-Interferon, Retinoic

Acid an Cytosin Arabinoside induce neuroblastoma dif-

ferentiatlon by different mechanisms. Cell. Mol.

Neurobiol , 11 (4), 397 -413.

Ponzoni, M.. Casalaro, A. and Cornaglia-Ferraris, P

(1991) Interferon gamma synergizes with retinoic acid and

tumor necrosis factor in inducing human neuroblastoma

cell differentiation. Proceedings of the Eighty-second

Meeting of the American Association for Cancer Research.

Vol. 32. p. 34. Waverly Press, Inc.. Baltimore.

Seeger. R.C.. Danon. Y.L., Rayner, S.A. and Hoover, F.

(1982) Definition of a thy-l determinant of human

neuroblastoma. glioma, sarcoma, and teratoma cells with

a monoclonal antibody. J. lmmunol., 128, 983- 989.

Wuarin, L., Verity, M.A. and Sidell. N. (1991) Effects of

Interferon-gamma and its interaction with Retinoic Acid on

human neuroblastoma differentiation. Int. J. Cancer, 48.

1. 136- 141.