- Email: [email protected]
A new approach for realizing the “antioncogram”
Life Sciences, Vol. Printed in the U S A
50, pp.
1711-1718
Pergamon
Press
A NEW APPROACH FOR REALIZING THE "ANTIONCOGRAM"
Petra K6pf-Maier Institut ftir Anatomic, Freie Universitfit Berlin, K6nigin-Luise-Stral3e 15, D-W-1000 Berlin 33, Germany (Received
in final
form March
24,
1992)
Summary
On the basis of an organoid culture method which allows organoid reorganization and histotypical growth of human carcinomas under in vitro conditions, we propose in the present study an organoid culture assay (OCA) as "antioncogram", i.e., as in vitro model for testing the drug sensitivity and resistance of individual patients' carcinomas previous to clinical chemotherapy. At the beginning of the assay, specimens of human carcinomas are disaggregated to dense single cell suspensions and dropped on a filter sheet at the air-medium interface. Organoid culture nodules develop within several days to weeks. They are exposed to cytostatics by adding the drugs for 2-3 days to the growth medium below the filter sheet. At the end of the exposure period, the cytotoxic effects are estimated by determining the fraction of viable cells, measured by the uptake of neutral red in relation to the total cell mass. In the present study we could show that this assay actually reflects the different levels of experimentally induced resistance of three human carcinoma strains of an epidermoid hypopharynx carcinoma to the cytostatic drug cisplatin and is obviously suited to predict the response of individual human carcinomas to chemotherapy in a rapid and feasible manner. Clinical chemotherapy of human carcinomas is often unsatisfactory because of some major, insuperable problems. Although there is clinical experience and statistical evidence that certain types of carcinomas respond to certain cytostatics or combinations of cytostatic drugs, this comes true only for a varying proportion of the patients suffering from a certain cancer type. Unfortunately, there is currently no way to predict whether an individual patient will respond to chemotherapy or not, as it is possible and routinely done in the case of bacterial diseases where the sensitivity and resistance to antibiotics can be determined in vitro before the beginning of individual patients' therapy. In dependence upon the tumor type, this situation in cancer chemotherapy necessarily results in a varying number of ineffective treatment trials which encumber the non-responding patients needlessly with severe and troublesome side effects without the benefit of any positive therapeutic effect. Moreover, these ineffective therapy courses often induce drug resistances in the treated tumors which usually impede second-line therapy strategies. The differing response of human carcinomas has been attributed to cellular heterogeneity detectable even in carcinomas that are classified as identical according to their histopathological characteristics (1). This cellular heterogeneity was shown to influence the drug response of human carcinomas (2). In consequence, it is critical that all
Copyright
0024-3205/92 $5.00 + .00 © 1992 Pergamon Press Ltd All rights reserved.
1712
New Approach
to an A n t i o n c o g r a m
Vol.
50, No. 22, 1992
in vivo and in vitro systems that are used for testing the drug sensitivity of individual patients' carcinomas preserve any cellular heterogeneity that has been present in the original tumor and maintain the histotypical arrangement and differentiation of the diverse cell types. However, the realization of this goal in vitro is quite difficult. When single cell suspensions of human carcinomas are plated on plastic or glass supports in order to grow them as monolayer cultures, the carcinoma cells do not in many instances adapt to these rather modest culture conditions, whereas fibroblasts usually overgrow the carcinoma cells within a few days. Another approach to grow human carcinoma cells in vitro is the clonogenic colony culture, the so-called human tumor stem cell assay (HTSCA), which highly selects cells and only allows transformed cells with a high proliferative activity to grow in vitro and to form colonies (3,4). In an opposite approach, several authors tried to establish organ cultures of human carcinomas, to keep small fragments of carcinoma tissues in vitro and to use them for drug testing (5-7). However, under these culture conditions, the carcinoma tissues are very unstable, rapidly lose viability and are, thus, not suited for drug-testing experiments. In the present study, we made a compromise and established an in vitro assay (OCA) for drug testing on the basis of an organoid culture method (8) which allows the cells of individual human carcinomas to reorganize in vitro to histotypically growing tissue nodules, to maintain cellular heterogeneity, to set up cellular interactions and to express morphological differentiation features similar to those seen in vivo. We developed an experimental assay for exposing these cultures to cytostatic agents and to measure the drug-induced cytotoxic effects. It was shown that the patterns of drug sensitivity and drug resistance were preserved under the described culture conditions and that, especially, the neutral red (NR) uptake into the culture nodules was suited to determine the drug-induced cytotoxic damages in organoid carcinoma cultures by a rapid and feasible experimental approach. Or~anoid culture of human carcinomas
When human carcinomas were disaggregated by mechanical and enzymatic methods to single cell suspensions and seeded at high cell density on a membrane filter consisting of cellulose nitrate (pore diameter, 0.2 t~m) at the air-medium interface, they reorganized within several days to multicellular aggregates, which always consisted of heterogeneous cells, i.e., fibroblasts, myoepithelial cells and epithelial carcinoma cells, in organoid, histotypical architecture. Regarding adenocarcinomas of the lung and the gastrointestinal tract, glandular structures with central lumina, adjacent microvilli and junctional complexes were formed. Sometimes, typical brush borders and signs of mucous secretion were detectable. Indubitable symptoms of cellular differentiation such as desmosomes, tight junctions, and interdigitations between adjacent cell membranes appeared. They stabilized the three-dimensional structure of the multicellular, organoid aggregates. In the case of epidermoid carcinomas of the lung or the head and neck, the culture nodules again consisted of heterogeneous cell types. Most cells growing were epithelial carcinoma cells which imitated the features of the basal layers of stratified epithelia and consisted of multilayers of cuboid, polyhedral and flattened cells which exhibited widened intercellular spaces, spinous cellular processes provided with typical desmosomes, and bundles of tonofilaments traversing the carcinoma cell cytoplasm and radiating into the desmosomes (Fig. 1). The minority of the cells was represented by fibroblasts with a well-developed rough endoplasmic reticulum in the cytoplasm. A discontinuous fragmentary basal lamina was often detectable between the carcinoma and the connective tissue cells. These features confirm organoid, histotypical reorganization of disaggregated
Vol.
50, NO. 22,
1992
New Approach
to an A n t i o n c o g r a m
1713
human carcinoma tissues under the conditions of the described organoid culture. Recent and so far unpublished investigations have shown that, moreover, collagen fibrils are synthesized by fibroblasts present in vitro under the given culture conditions, fragments of a mono- or multilayered basal lamina assemble at various sites within the carcinoma nodules, the typical keratin pattern is detectable in carcinoma cells, and specific tumor markers and oncogens are expressed by analogy with the in vivo situation. This means that the described culture conditions obviously allow heterogeneous cell types derived from an
FIG. 1 Human squamous cell carcinoma of the hypopharynx in organoid culture on day 6. a, Semithick section, stained with toluidine blue; b,c, electron micrographs of ultrathin sections. The multicellular aggregates formed in vitro within a few days (a,b) consist of heterogeneous cell types (e.g., fibroblasts, F, and tumor cells, T) in three-dimensional, histotypical architecture. The cellular arrangement resembles that of the spinous layer of stratified squamous epithelia (a,b). Numerous signs of cellular differentiation such as spinous cellular processes, typical desmosomes connecting neighboring cells and bundles of tonofilaments within the tumor cell cytoplasm are detectable (a,b). x 350 (a); x 9.500 (b); x 90.000 (c)
1714
New Approach
to an A n t i o n c o g r a m
Vol.
50, No. 22,
1992
individual human carcinoma to coexist in vitro and to differentiate to histotypically organized, three-dimensional tissue nodules that preserve the cellular microenvironment as factor of cellular ease and as stimulus for cellular differentiation. In vitro assay for drug testing,
These organoid culture nodules growing on the top of a sheet of cellulose nitrate (pore diameter, 0.2/~m) at the air-medium interface can be easily exposed to cytostatic drugs by adding these agents to the growth medium beneath the membrane filter (Fig. 2). This situation mimics well the in vivo situation where cytostatic drugs approach carcinoma tissues via irradiating blood vessels and have to diffuse through the vascular wall, before they reach the tumor tissue just by diffusion along these guide rails. Ham's F12 medium supplemented by fetal calf serum, L-glutamine, penicillin and streptomycin, amphotericin B, ascorbic acid, non-essential amino acids and glucose was used as growth medium. Experimental investigations that are not shown here in detail have revealed that the optimum duration of drug exposure in organoid cultures lasts 2 to 3 days
~'[
NR tSRB
NOR~
ISRB
Removal of human tumors 4, Dispersion to a single cell suspension (5 x 10' cells/ul) 4, Dropping volumes of 20~1 of this cell suspension on a membrane filter at the gas-medium interface 4, Incubation at 37~C/5% CO 2 for several days 4, Organoid, histotypic growth of the cultures 4, Addition of cytostatic drugs to the growth medium 4, Detachment of the nodules, bisecting them, and determining (i) the viable cell fraction (e.g., by the neutral red method), (ii) the total cell mass (e.g., by the sulforhcdamin B assay)
FIG. 2 Synopsis of the main experimental steps of the organoid culture assay (OCA), proposed as method for realizing the pretherapeutic antioncogram, i.e., as assay which allows to predict the outcome of clinical chemotherapy in individual cancer patients.
Vol.
50, No. 22, 1992
New Approach
to an A n t i o n c o g r a m
1715
and that the most feasible method for establishing the drug-induced cytotoxic effects is to determine first the viable cells by measuring the uptake of the vital dye neutral red (NR) into treated culture nodules, submerged for 24 h in a NR solution, at the end of drug exposure (NR is known to be taken up selectively by live, viable cells and to be incorporated into their lysosomes (9-11)) and, then, to assess the total cell mass by measuring the total cell protein using the sulforhodamin B (SRB) assay (12-15). Both parameters are estimated by extracting the dyes and by determining the concentration of the dyes in defined volumes by measuring the absorbances spectrophotometrically at 540 (NR) and 564 nm (SRB), respectively. The quantification of the dyes was performed according to methods described before (9,10,12). The quotient NR:SRB absorbances, related as percentage to the control value, indicates the fraction of viable cells and gives a measure of the cytotoxic injury which was effected by the cytotoxic drug applied in vitro. In this connection it is worth mentioning that, by this procedure, all cells growing in the organoid culture nodules, e.g., poorly, moderately and highly differentiated carcinoma cells, fibroblasts and myoepithelial cells, are counted by dye uptake and extraction. Since cellular heterogeneity also characterizes human tumors growing in patients, the organoid culture assay (OCA) seems to be more reliable to the in vivo situation than other methods such as the colony assay or methods using monolayer cultures, which do not consider quiescent carcinoma or non-malignant cells. Detection of differences in drug sensitivity in vitro In order to determine whether the described organoid culture assay (OCA, Fig. 2) is suited for detecting differences in the drug sensitivity and resistance of human carcinomas in vitro, we applied this method to three strains of a human, moderately differentiated squamous carcinoma of the hypopharynx which distinguished in vivo by different levels of experimentally induced resistance to the inorganic cytostatic drug cisplatirl (cis-diamminedichloroplatinum(II)). The original strain was highly sensitive to cisplatin and regressed completely and irreversibly when it was treated with optimum doses of cisplatin (9 or 12 mg/kg) in athymic mice. Two resistant strains of this tumor had been developed in vivo by applying increasing, but always subeffective doses of cisplatin to nude mice bearing this tumor during several months. Before culturing these carcinoma strains in vitro, the partially resistant strain had been accustomed to a weekly dose of 6.5 mg/kg, whereas the highly resistant strain had been treated at last twice per week with 6.0 mg cisplatin/kg. These long-lasting, subeffective cisplatin applications had induced different levels of resistance which were reflected in graduated responses to in vivo treatment with cisplatin (15). When these differently resistant carcinoma strains were grown in vitro and exposed there to cisplatin at increasing concentration levels, the differences in drug resistance were clearly mirrored by differing dose-response relations (Fig. 3). The ECs0 values calculated from the trypan blue exclusion assay amounted to 0.9x10 ~ mol/1 (original sensitive strain), 5.5x10 ~ tool/1 (partially resistant strain), and 0.9x10 5 mol/l (highly resistant strain), respectively. This means that the different levels of resistance which had been induced experimentally in vivo can be detected in vitro by differences of the response levels varying by factors from 5 to 10. The statistical analysis using the Student's t-test confirmed these differences to be significant (p < 0.01). These results show that the organoid culture assay (OCA) is actually suited to reveal differences in the sensitivity and resistance of human carcinomas to cytostatic drugs in a rapid and feasible manner. Nevertheless, it is presently too early to decide whether the OCA can be developed as new approach to realize the antioncogram, i.e., as in vitro test for examining the pattern of drug sensitivity and resistance of individual patients' tumors before the beginning of clinical chemotherapy. For this decision, additional in vitro data
1716
New Approach
to an A n t i o n c o g r a m
Vol.
50, No. 22, 1992
% of c o n t r o l
80
j, "7"7,~,~ -' "°':~.~
40 20 0
Original sensitive strain
C
•
5x_10-7
-'.r_. . . . . . . . 5 x 1 0 -6
5x10 "5
~_ 5x10-'
/11 I•.
% %
80 60
~ ~-
2o o
Partiallyresistant strain
........ C,
, -7 5 x_lO
,r,A, I
80 60
1~..0
5 x 1, 0 -6
't' 5x10 "5
"',
5 xiO"
"~,ez . Highly resistant ~'~kl strain
2O
0
~
5xtlO -'
•- - - - * C o n c e n t r a t i o n
5x10 -6
5x10 "s
of cisplatin
5x10-'
(mol/I)
FIG. 3 Alterations of the cell viability in organoid cultures of three strains of a human epidermoid hypopharynx carcinoma at the end of a 72 h-exposure to increasing doses of the cytostatic drug cisplatin (abscissa).. The stratus distinguished by different levels of resistance to cisplatin. Cell viabdity was measured by trypan blue (TB) exclusion (22) of single cell suspensions in relation to the total cell number counted in a haematocounter (o---o), and by neutral red (NR) uptake into whole, non-dissociated culture nodules, either related to the total cell number ( o - 4 ) or to the total cell protein determined by the sulforhodamin B (SRB) assay (e-...o). These values were related as percentage to the control values (ordinate). The ECs0 values (ECs0, concentration effecting the loss of viability of 50 % of the cells) are indicated by downward arrows and amount to 0.9x10 ~ (original sensitive strain), 5.5x10 ~ (partially resistant strain), and 0.9x10 5 mol cisplatin/1 (highly resistant strain), respectively.
Vol.
50, No. 22,
1992
New Approach
to an A n t i o n c o g r a m
1717
based on numerous other carcinomas and cytostatic drugs are required. Comparison with other organoid in vitro models Other organoid culture methods which had been developed during the past years and were also used to determine the drug sensitivity of human carcinomas in vitro are the spheroid model, where separated human carcinoma cells aggregate to spheroids when they are submerged in growth medium under certain culture conditions (16-18), and the gelsupported culture system where small tumor tissue pieces are explanted on the top of hydrated collagen gels, allowing the cells at the periphery to grow out and to invade the collagen support (1,19,21). In both models, autoradiographic and histological investigations confirmed vivacious proliferation, but only limited differentiation. Correspondingly, the assay procedure elaborated for the gel-supported cultures by Hoffman et al. (1,20,21) only regards the proliferative activity in explanted carcinoma tissue pieces by determining the incorporation of ['H]thymidine and ['H]deoxyuridine and does not consider the influence of cytostatic drugs upon non-proliferating, i.e., quiescent and just non-cycling carcinoma cells. These cells, however, are often responsible for the later regrowth of treated carcinomas and the failure of chemotherapy. Moreover, the handling with radioactivity is disagreeable, not feasible in every laboratory and should be avoided as far as possible. Another disadvantage of the spheroid and gel-supported models is the experimental situation that, in both systems, the cultures grow submerged in medium, unable to set up their cell-specific microenvironment. They are bathed in drug-containing fluid during exposure to cytostatic drugs in a manner that is quite unphysiological and dissimilar to the in vivo situation where the drugs just reach the carcinoma nodules via and along the blood vessel rails, but do not flow homogeneously around the tumor tissue. This situation must lead to an overestimation of drug-induced cytotoxic effects both in the spheroid and the gel-supported assays. Conclusions and Perspectives These results confirm that the OCA obviously fulfils the criteria which are required for an in vitro model to determine the patterns of drug sensitivity and drug resistance of human carcinomas in close relation to the in vivo situation. It allows the growth of histotypically differentiated nodules of carcinoma tissues in vitro, determines the cytotoxic damage by the NR uptake indicating cellular integrity and the preservation of fundamental cellular functions, and apparently mirrors differences in drug sensitivity and resistance of human carcinomas. It offers this information within several days and is a rather feasible method. Extensive experimental and clinical trials must be done in the future in order to pass judgement on the actual value of this assay as "antioncogram", i.e., as test for predicting the outcome of clinical chemotherapy for individual cancer patients in a manner analogous to techniques used for bacterial antibiotic sensitivity testing. In vitro experiments that are currently performed with human breast carcinomas obviously confirm a strong correlation between the results of the OCA and the in vivo data of chemotherapy. Acknowledgement The author thanks Mrs. Birgit Kolon for her excellent and dedicated technical assistance.
1718
N e w A p p r o a c h to an A n t i o n c o g r a m
Vol.
50, No. 22, 1992
References
1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14.
15. 16. 17. 18. 19. 20. 21. 22.
A.E. FREEMAN and R.M. HOFFMAN, Proc. Natl. Acad. Sci. USA, 83 2694-2698 (1986). B.E. MILLER, F.R. MILLER and G.H. HEPPNER, Cancer Res., 45 4200-4205 (1985). A.W. HAMBURGER and S.E. SALMON, Science, 197 461-463 (1977). S.E. SALMON, Cancer Treat. Rep., 68 117-125 (1984). Et. WOLFF and Em. WOLFF, Human Tumor Cells in vitro, J. Fogh (ed), 207-240, Plenum Press, New York, London (1975). A.H. LAWSON, A.C. RICHES and J.P.A. WEAVER, J. Urol., 135 1061-1065 (1986). M.G. DAIDONE, R. SILVESTRINI, N. ZAFFARONI, E. GRIGNOLIO and F. LANDONI, Invest. New Drugs, 5_ 245-250 (1987). P. K(~PF-MAIER and B. ZIMMERMANN, Cell Tiss. Res., 264 563-576 (1991). E. BORENFREUND and J.A. PUERNER, J. Tiss. Cult. Methods, 9 7-9 (1984). E. BORENFREUND and J.A. PUERNER, Toxicol. Lett., 24 119-124 (1985). M. BORG, D. KIRK, H. BAUMGARTEN and R. RUCHEL, J. Med. Vet. Mycology, 22 357-367 (1984). P. SKEHAN, R. STORENG, D. SCUDIERO, A. MONKS, J. McMAHON, D. VISTICA, J.T. WARREN, H. BOKESCH, S. KENNEY and M.R. BOYD, J. Natl. Cancer Inst., 82 1107-1112 (1990). L.V. RUBINSTEIN, R.H. SHOEMAKER, K.D. PAULL, R.M. SIMON, S. TOS1NI, P. SKEHAN, D.A. SCUDIERO, A. MONKS and M.R. BOYD, J. Natl. Cancer Inst., 82 1113-1118 (1990). A. MONKS, D. SCUDIERO, P. SKEHAN, R. SHOEMAKER, K. PAULL, D. VISTICA, C. HOSE, J. LANGLEY, P. CRONISE, A. VAIGRO-WOLFF, M. GRAY-GOODRICH, H. CAMPBELL, J. MAYO and M. BOYD, J. Natl. Cancer Inst., 83 757-766 (1991):. M. J,~CKEL and P. KOPF-MAIER, Cancer Chemother. Pharmacol., 27 464-471 (1991). E.K. ROFSTAD, A. WAHL and T. BRUSTAD, Rad. Res., 102 324-335 (1985). R.M. SUTHERLAND, Science, 240 177-184 (1988). R.E. DURAND, Brit. J. Cancer, 62 947-953 (1990). T. HOLMSTROM, Human Tumor Cells in vitro, J. Fogh (ed), 161-206, Plenum Press, New York, London (1975). R.A. VESCIO, C.H. REDFERN, T.J. NELSON, S. UGORETZ, P.H. STERN, and R.M. HOFFMAN, Proc. Natl. Acad. Sci. USA, 84 5029-5033 (1987). R.A. VESCIO, K.M. CONNORS, T. YOUNGKIN, G.M. BORDIN, J.A. ROBB, J.N UMBREIT, and R.M. HOFFMAN, Proc. Natl. Acad. Sci. USA, 87 691-695 (1990). W.J. DURKIN, V.K. GHANTA, C.M. BALCH, D.W. DAVIS and R.N. HIRAMOTO, Cancer Res., 39 402-407 (1979).
50, pp.
1711-1718
Pergamon
Press
A NEW APPROACH FOR REALIZING THE "ANTIONCOGRAM"
Petra K6pf-Maier Institut ftir Anatomic, Freie Universitfit Berlin, K6nigin-Luise-Stral3e 15, D-W-1000 Berlin 33, Germany (Received
in final
form March
24,
1992)
Summary
On the basis of an organoid culture method which allows organoid reorganization and histotypical growth of human carcinomas under in vitro conditions, we propose in the present study an organoid culture assay (OCA) as "antioncogram", i.e., as in vitro model for testing the drug sensitivity and resistance of individual patients' carcinomas previous to clinical chemotherapy. At the beginning of the assay, specimens of human carcinomas are disaggregated to dense single cell suspensions and dropped on a filter sheet at the air-medium interface. Organoid culture nodules develop within several days to weeks. They are exposed to cytostatics by adding the drugs for 2-3 days to the growth medium below the filter sheet. At the end of the exposure period, the cytotoxic effects are estimated by determining the fraction of viable cells, measured by the uptake of neutral red in relation to the total cell mass. In the present study we could show that this assay actually reflects the different levels of experimentally induced resistance of three human carcinoma strains of an epidermoid hypopharynx carcinoma to the cytostatic drug cisplatin and is obviously suited to predict the response of individual human carcinomas to chemotherapy in a rapid and feasible manner. Clinical chemotherapy of human carcinomas is often unsatisfactory because of some major, insuperable problems. Although there is clinical experience and statistical evidence that certain types of carcinomas respond to certain cytostatics or combinations of cytostatic drugs, this comes true only for a varying proportion of the patients suffering from a certain cancer type. Unfortunately, there is currently no way to predict whether an individual patient will respond to chemotherapy or not, as it is possible and routinely done in the case of bacterial diseases where the sensitivity and resistance to antibiotics can be determined in vitro before the beginning of individual patients' therapy. In dependence upon the tumor type, this situation in cancer chemotherapy necessarily results in a varying number of ineffective treatment trials which encumber the non-responding patients needlessly with severe and troublesome side effects without the benefit of any positive therapeutic effect. Moreover, these ineffective therapy courses often induce drug resistances in the treated tumors which usually impede second-line therapy strategies. The differing response of human carcinomas has been attributed to cellular heterogeneity detectable even in carcinomas that are classified as identical according to their histopathological characteristics (1). This cellular heterogeneity was shown to influence the drug response of human carcinomas (2). In consequence, it is critical that all
Copyright
0024-3205/92 $5.00 + .00 © 1992 Pergamon Press Ltd All rights reserved.
1712
New Approach
to an A n t i o n c o g r a m
Vol.
50, No. 22, 1992
in vivo and in vitro systems that are used for testing the drug sensitivity of individual patients' carcinomas preserve any cellular heterogeneity that has been present in the original tumor and maintain the histotypical arrangement and differentiation of the diverse cell types. However, the realization of this goal in vitro is quite difficult. When single cell suspensions of human carcinomas are plated on plastic or glass supports in order to grow them as monolayer cultures, the carcinoma cells do not in many instances adapt to these rather modest culture conditions, whereas fibroblasts usually overgrow the carcinoma cells within a few days. Another approach to grow human carcinoma cells in vitro is the clonogenic colony culture, the so-called human tumor stem cell assay (HTSCA), which highly selects cells and only allows transformed cells with a high proliferative activity to grow in vitro and to form colonies (3,4). In an opposite approach, several authors tried to establish organ cultures of human carcinomas, to keep small fragments of carcinoma tissues in vitro and to use them for drug testing (5-7). However, under these culture conditions, the carcinoma tissues are very unstable, rapidly lose viability and are, thus, not suited for drug-testing experiments. In the present study, we made a compromise and established an in vitro assay (OCA) for drug testing on the basis of an organoid culture method (8) which allows the cells of individual human carcinomas to reorganize in vitro to histotypically growing tissue nodules, to maintain cellular heterogeneity, to set up cellular interactions and to express morphological differentiation features similar to those seen in vivo. We developed an experimental assay for exposing these cultures to cytostatic agents and to measure the drug-induced cytotoxic effects. It was shown that the patterns of drug sensitivity and drug resistance were preserved under the described culture conditions and that, especially, the neutral red (NR) uptake into the culture nodules was suited to determine the drug-induced cytotoxic damages in organoid carcinoma cultures by a rapid and feasible experimental approach. Or~anoid culture of human carcinomas
When human carcinomas were disaggregated by mechanical and enzymatic methods to single cell suspensions and seeded at high cell density on a membrane filter consisting of cellulose nitrate (pore diameter, 0.2 t~m) at the air-medium interface, they reorganized within several days to multicellular aggregates, which always consisted of heterogeneous cells, i.e., fibroblasts, myoepithelial cells and epithelial carcinoma cells, in organoid, histotypical architecture. Regarding adenocarcinomas of the lung and the gastrointestinal tract, glandular structures with central lumina, adjacent microvilli and junctional complexes were formed. Sometimes, typical brush borders and signs of mucous secretion were detectable. Indubitable symptoms of cellular differentiation such as desmosomes, tight junctions, and interdigitations between adjacent cell membranes appeared. They stabilized the three-dimensional structure of the multicellular, organoid aggregates. In the case of epidermoid carcinomas of the lung or the head and neck, the culture nodules again consisted of heterogeneous cell types. Most cells growing were epithelial carcinoma cells which imitated the features of the basal layers of stratified epithelia and consisted of multilayers of cuboid, polyhedral and flattened cells which exhibited widened intercellular spaces, spinous cellular processes provided with typical desmosomes, and bundles of tonofilaments traversing the carcinoma cell cytoplasm and radiating into the desmosomes (Fig. 1). The minority of the cells was represented by fibroblasts with a well-developed rough endoplasmic reticulum in the cytoplasm. A discontinuous fragmentary basal lamina was often detectable between the carcinoma and the connective tissue cells. These features confirm organoid, histotypical reorganization of disaggregated
Vol.
50, NO. 22,
1992
New Approach
to an A n t i o n c o g r a m
1713
human carcinoma tissues under the conditions of the described organoid culture. Recent and so far unpublished investigations have shown that, moreover, collagen fibrils are synthesized by fibroblasts present in vitro under the given culture conditions, fragments of a mono- or multilayered basal lamina assemble at various sites within the carcinoma nodules, the typical keratin pattern is detectable in carcinoma cells, and specific tumor markers and oncogens are expressed by analogy with the in vivo situation. This means that the described culture conditions obviously allow heterogeneous cell types derived from an
FIG. 1 Human squamous cell carcinoma of the hypopharynx in organoid culture on day 6. a, Semithick section, stained with toluidine blue; b,c, electron micrographs of ultrathin sections. The multicellular aggregates formed in vitro within a few days (a,b) consist of heterogeneous cell types (e.g., fibroblasts, F, and tumor cells, T) in three-dimensional, histotypical architecture. The cellular arrangement resembles that of the spinous layer of stratified squamous epithelia (a,b). Numerous signs of cellular differentiation such as spinous cellular processes, typical desmosomes connecting neighboring cells and bundles of tonofilaments within the tumor cell cytoplasm are detectable (a,b). x 350 (a); x 9.500 (b); x 90.000 (c)
1714
New Approach
to an A n t i o n c o g r a m
Vol.
50, No. 22,
1992
individual human carcinoma to coexist in vitro and to differentiate to histotypically organized, three-dimensional tissue nodules that preserve the cellular microenvironment as factor of cellular ease and as stimulus for cellular differentiation. In vitro assay for drug testing,
These organoid culture nodules growing on the top of a sheet of cellulose nitrate (pore diameter, 0.2/~m) at the air-medium interface can be easily exposed to cytostatic drugs by adding these agents to the growth medium beneath the membrane filter (Fig. 2). This situation mimics well the in vivo situation where cytostatic drugs approach carcinoma tissues via irradiating blood vessels and have to diffuse through the vascular wall, before they reach the tumor tissue just by diffusion along these guide rails. Ham's F12 medium supplemented by fetal calf serum, L-glutamine, penicillin and streptomycin, amphotericin B, ascorbic acid, non-essential amino acids and glucose was used as growth medium. Experimental investigations that are not shown here in detail have revealed that the optimum duration of drug exposure in organoid cultures lasts 2 to 3 days
~'[
NR tSRB
NOR~
ISRB
Removal of human tumors 4, Dispersion to a single cell suspension (5 x 10' cells/ul) 4, Dropping volumes of 20~1 of this cell suspension on a membrane filter at the gas-medium interface 4, Incubation at 37~C/5% CO 2 for several days 4, Organoid, histotypic growth of the cultures 4, Addition of cytostatic drugs to the growth medium 4, Detachment of the nodules, bisecting them, and determining (i) the viable cell fraction (e.g., by the neutral red method), (ii) the total cell mass (e.g., by the sulforhcdamin B assay)
FIG. 2 Synopsis of the main experimental steps of the organoid culture assay (OCA), proposed as method for realizing the pretherapeutic antioncogram, i.e., as assay which allows to predict the outcome of clinical chemotherapy in individual cancer patients.
Vol.
50, No. 22, 1992
New Approach
to an A n t i o n c o g r a m
1715
and that the most feasible method for establishing the drug-induced cytotoxic effects is to determine first the viable cells by measuring the uptake of the vital dye neutral red (NR) into treated culture nodules, submerged for 24 h in a NR solution, at the end of drug exposure (NR is known to be taken up selectively by live, viable cells and to be incorporated into their lysosomes (9-11)) and, then, to assess the total cell mass by measuring the total cell protein using the sulforhodamin B (SRB) assay (12-15). Both parameters are estimated by extracting the dyes and by determining the concentration of the dyes in defined volumes by measuring the absorbances spectrophotometrically at 540 (NR) and 564 nm (SRB), respectively. The quantification of the dyes was performed according to methods described before (9,10,12). The quotient NR:SRB absorbances, related as percentage to the control value, indicates the fraction of viable cells and gives a measure of the cytotoxic injury which was effected by the cytotoxic drug applied in vitro. In this connection it is worth mentioning that, by this procedure, all cells growing in the organoid culture nodules, e.g., poorly, moderately and highly differentiated carcinoma cells, fibroblasts and myoepithelial cells, are counted by dye uptake and extraction. Since cellular heterogeneity also characterizes human tumors growing in patients, the organoid culture assay (OCA) seems to be more reliable to the in vivo situation than other methods such as the colony assay or methods using monolayer cultures, which do not consider quiescent carcinoma or non-malignant cells. Detection of differences in drug sensitivity in vitro In order to determine whether the described organoid culture assay (OCA, Fig. 2) is suited for detecting differences in the drug sensitivity and resistance of human carcinomas in vitro, we applied this method to three strains of a human, moderately differentiated squamous carcinoma of the hypopharynx which distinguished in vivo by different levels of experimentally induced resistance to the inorganic cytostatic drug cisplatirl (cis-diamminedichloroplatinum(II)). The original strain was highly sensitive to cisplatin and regressed completely and irreversibly when it was treated with optimum doses of cisplatin (9 or 12 mg/kg) in athymic mice. Two resistant strains of this tumor had been developed in vivo by applying increasing, but always subeffective doses of cisplatin to nude mice bearing this tumor during several months. Before culturing these carcinoma strains in vitro, the partially resistant strain had been accustomed to a weekly dose of 6.5 mg/kg, whereas the highly resistant strain had been treated at last twice per week with 6.0 mg cisplatin/kg. These long-lasting, subeffective cisplatin applications had induced different levels of resistance which were reflected in graduated responses to in vivo treatment with cisplatin (15). When these differently resistant carcinoma strains were grown in vitro and exposed there to cisplatin at increasing concentration levels, the differences in drug resistance were clearly mirrored by differing dose-response relations (Fig. 3). The ECs0 values calculated from the trypan blue exclusion assay amounted to 0.9x10 ~ mol/1 (original sensitive strain), 5.5x10 ~ tool/1 (partially resistant strain), and 0.9x10 5 mol/l (highly resistant strain), respectively. This means that the different levels of resistance which had been induced experimentally in vivo can be detected in vitro by differences of the response levels varying by factors from 5 to 10. The statistical analysis using the Student's t-test confirmed these differences to be significant (p < 0.01). These results show that the organoid culture assay (OCA) is actually suited to reveal differences in the sensitivity and resistance of human carcinomas to cytostatic drugs in a rapid and feasible manner. Nevertheless, it is presently too early to decide whether the OCA can be developed as new approach to realize the antioncogram, i.e., as in vitro test for examining the pattern of drug sensitivity and resistance of individual patients' tumors before the beginning of clinical chemotherapy. For this decision, additional in vitro data
1716
New Approach
to an A n t i o n c o g r a m
Vol.
50, No. 22, 1992
% of c o n t r o l
80
j, "7"7,~,~ -' "°':~.~
40 20 0
Original sensitive strain
C
•
5x_10-7
-'.r_. . . . . . . . 5 x 1 0 -6
5x10 "5
~_ 5x10-'
/11 I•.
% %
80 60
~ ~-
2o o
Partiallyresistant strain
........ C,
, -7 5 x_lO
,r,A, I
80 60
1~..0
5 x 1, 0 -6
't' 5x10 "5
"',
5 xiO"
"~,ez . Highly resistant ~'~kl strain
2O
0
~
5xtlO -'
•- - - - * C o n c e n t r a t i o n
5x10 -6
5x10 "s
of cisplatin
5x10-'
(mol/I)
FIG. 3 Alterations of the cell viability in organoid cultures of three strains of a human epidermoid hypopharynx carcinoma at the end of a 72 h-exposure to increasing doses of the cytostatic drug cisplatin (abscissa).. The stratus distinguished by different levels of resistance to cisplatin. Cell viabdity was measured by trypan blue (TB) exclusion (22) of single cell suspensions in relation to the total cell number counted in a haematocounter (o---o), and by neutral red (NR) uptake into whole, non-dissociated culture nodules, either related to the total cell number ( o - 4 ) or to the total cell protein determined by the sulforhodamin B (SRB) assay (e-...o). These values were related as percentage to the control values (ordinate). The ECs0 values (ECs0, concentration effecting the loss of viability of 50 % of the cells) are indicated by downward arrows and amount to 0.9x10 ~ (original sensitive strain), 5.5x10 ~ (partially resistant strain), and 0.9x10 5 mol cisplatin/1 (highly resistant strain), respectively.
Vol.
50, No. 22,
1992
New Approach
to an A n t i o n c o g r a m
1717
based on numerous other carcinomas and cytostatic drugs are required. Comparison with other organoid in vitro models Other organoid culture methods which had been developed during the past years and were also used to determine the drug sensitivity of human carcinomas in vitro are the spheroid model, where separated human carcinoma cells aggregate to spheroids when they are submerged in growth medium under certain culture conditions (16-18), and the gelsupported culture system where small tumor tissue pieces are explanted on the top of hydrated collagen gels, allowing the cells at the periphery to grow out and to invade the collagen support (1,19,21). In both models, autoradiographic and histological investigations confirmed vivacious proliferation, but only limited differentiation. Correspondingly, the assay procedure elaborated for the gel-supported cultures by Hoffman et al. (1,20,21) only regards the proliferative activity in explanted carcinoma tissue pieces by determining the incorporation of ['H]thymidine and ['H]deoxyuridine and does not consider the influence of cytostatic drugs upon non-proliferating, i.e., quiescent and just non-cycling carcinoma cells. These cells, however, are often responsible for the later regrowth of treated carcinomas and the failure of chemotherapy. Moreover, the handling with radioactivity is disagreeable, not feasible in every laboratory and should be avoided as far as possible. Another disadvantage of the spheroid and gel-supported models is the experimental situation that, in both systems, the cultures grow submerged in medium, unable to set up their cell-specific microenvironment. They are bathed in drug-containing fluid during exposure to cytostatic drugs in a manner that is quite unphysiological and dissimilar to the in vivo situation where the drugs just reach the carcinoma nodules via and along the blood vessel rails, but do not flow homogeneously around the tumor tissue. This situation must lead to an overestimation of drug-induced cytotoxic effects both in the spheroid and the gel-supported assays. Conclusions and Perspectives These results confirm that the OCA obviously fulfils the criteria which are required for an in vitro model to determine the patterns of drug sensitivity and drug resistance of human carcinomas in close relation to the in vivo situation. It allows the growth of histotypically differentiated nodules of carcinoma tissues in vitro, determines the cytotoxic damage by the NR uptake indicating cellular integrity and the preservation of fundamental cellular functions, and apparently mirrors differences in drug sensitivity and resistance of human carcinomas. It offers this information within several days and is a rather feasible method. Extensive experimental and clinical trials must be done in the future in order to pass judgement on the actual value of this assay as "antioncogram", i.e., as test for predicting the outcome of clinical chemotherapy for individual cancer patients in a manner analogous to techniques used for bacterial antibiotic sensitivity testing. In vitro experiments that are currently performed with human breast carcinomas obviously confirm a strong correlation between the results of the OCA and the in vivo data of chemotherapy. Acknowledgement The author thanks Mrs. Birgit Kolon for her excellent and dedicated technical assistance.
1718
N e w A p p r o a c h to an A n t i o n c o g r a m
Vol.
50, No. 22, 1992
References
1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14.
15. 16. 17. 18. 19. 20. 21. 22.
A.E. FREEMAN and R.M. HOFFMAN, Proc. Natl. Acad. Sci. USA, 83 2694-2698 (1986). B.E. MILLER, F.R. MILLER and G.H. HEPPNER, Cancer Res., 45 4200-4205 (1985). A.W. HAMBURGER and S.E. SALMON, Science, 197 461-463 (1977). S.E. SALMON, Cancer Treat. Rep., 68 117-125 (1984). Et. WOLFF and Em. WOLFF, Human Tumor Cells in vitro, J. Fogh (ed), 207-240, Plenum Press, New York, London (1975). A.H. LAWSON, A.C. RICHES and J.P.A. WEAVER, J. Urol., 135 1061-1065 (1986). M.G. DAIDONE, R. SILVESTRINI, N. ZAFFARONI, E. GRIGNOLIO and F. LANDONI, Invest. New Drugs, 5_ 245-250 (1987). P. K(~PF-MAIER and B. ZIMMERMANN, Cell Tiss. Res., 264 563-576 (1991). E. BORENFREUND and J.A. PUERNER, J. Tiss. Cult. Methods, 9 7-9 (1984). E. BORENFREUND and J.A. PUERNER, Toxicol. Lett., 24 119-124 (1985). M. BORG, D. KIRK, H. BAUMGARTEN and R. RUCHEL, J. Med. Vet. Mycology, 22 357-367 (1984). P. SKEHAN, R. STORENG, D. SCUDIERO, A. MONKS, J. McMAHON, D. VISTICA, J.T. WARREN, H. BOKESCH, S. KENNEY and M.R. BOYD, J. Natl. Cancer Inst., 82 1107-1112 (1990). L.V. RUBINSTEIN, R.H. SHOEMAKER, K.D. PAULL, R.M. SIMON, S. TOS1NI, P. SKEHAN, D.A. SCUDIERO, A. MONKS and M.R. BOYD, J. Natl. Cancer Inst., 82 1113-1118 (1990). A. MONKS, D. SCUDIERO, P. SKEHAN, R. SHOEMAKER, K. PAULL, D. VISTICA, C. HOSE, J. LANGLEY, P. CRONISE, A. VAIGRO-WOLFF, M. GRAY-GOODRICH, H. CAMPBELL, J. MAYO and M. BOYD, J. Natl. Cancer Inst., 83 757-766 (1991):. M. J,~CKEL and P. KOPF-MAIER, Cancer Chemother. Pharmacol., 27 464-471 (1991). E.K. ROFSTAD, A. WAHL and T. BRUSTAD, Rad. Res., 102 324-335 (1985). R.M. SUTHERLAND, Science, 240 177-184 (1988). R.E. DURAND, Brit. J. Cancer, 62 947-953 (1990). T. HOLMSTROM, Human Tumor Cells in vitro, J. Fogh (ed), 161-206, Plenum Press, New York, London (1975). R.A. VESCIO, C.H. REDFERN, T.J. NELSON, S. UGORETZ, P.H. STERN, and R.M. HOFFMAN, Proc. Natl. Acad. Sci. USA, 84 5029-5033 (1987). R.A. VESCIO, K.M. CONNORS, T. YOUNGKIN, G.M. BORDIN, J.A. ROBB, J.N UMBREIT, and R.M. HOFFMAN, Proc. Natl. Acad. Sci. USA, 87 691-695 (1990). W.J. DURKIN, V.K. GHANTA, C.M. BALCH, D.W. DAVIS and R.N. HIRAMOTO, Cancer Res., 39 402-407 (1979).