Studies on drug resistance in a human melanoma xenograft system

Cancer

Treatment

Reviews

(1984)

11 (Supplement A), 85-97

Studies on drug resistance xenograft system Rainhardt

in a human

melanoma

Osieka*

Innere Universitiitsklinik ( Tumorforschung), Federal Republic of Germany

GHS Essen, Hufelandstrasse

55, D-4300

Essen I,

Introduction Alkylating agents and their functional analogues rank prominently among antineoplastic drugs classified as useful against malignant melanoma. Despite some similarities in their mechanisms of action, cross-resistance among DNA-damaging agents has been incomplete both in clinical trials and experimental tumour systems (2-4, 13, 14). The problem of multidrug resistance in human cancer cells has been addressed for anthracyclines, actinomycin D and vinca alkaloids but not for alkylating agents (16). Whereas in patients cross-resistance can only be evaluated after sequential drug exposure, preclinical tumour models based on human cancer cells allow for simultaneous drug testing while reference is maintained to specific entities of malignant disease. Human tumour xenografts have been described as a valid model of specific tumour entities with respect to drug treatment ( 17) and the need to condition the host prior to heterotransplantation is obviated by use of congenitally athymic (nu/nu) mice (9, 12). In conjunction with an open clinical phase II trial evaluating response to the combination of cisplatin and ifosfamide in patients mostly refractory to dacarbazine ( 1)) xenografts were established and exposed to a panel of DNA-damaging agents. Cisplatin, dacarbazine, dibromodulcitol, ifosfamide, methyl-CCNU, mitomycin C, and malonatodiaminocyclohexane-platinum(I1) all react covalently with cellular DNA or release moieties with such potential. In order to achieve a therapeutic ranking, all doses were kept close to or within the LD lo/30 range. This experimental design was to provide two answers: First, the validity of the xenograft model could be verified both by comparing individual responses of each xenograft line to the clinical treatment results of the donor patient and overall response rates in the panel of xenografts to published response rates to individual agents in malignant melanoma. Second, cross-resistance patterns could be analyzed independently from clinical results. * Supported

by SFB 102 of the Dcuochc

0305-7372/84/11AOO85+

Forschungsgemeinschaft. 0

14 $03.00/O 85

1984 Academic

Press Inc.

(London)

Limited

86

R. OSlEKA

In additional studies the degree of resistance was to be quantified by dose-response curves. Investigations on mechanisms of drug resistance are often carried out on tumour lines derived from originally sensitive tumours by suboptimal drug treatment. The question arises, how valid such models of drug resistance are with respect to patterns of resistance encountered in donor patients and xenografts. Finally, tumour pharmacology can be compared in resistant and sensitive xenograft lines after in uivo drug exposure in order to identify mechanisms of drug resistance. DNA damage after exposure to alkylating agents below the LD lo/30 level has been difficult to measure, but the recently introduced technique of alkaline elution allows for quantification of macromolecular DNA damage without radioactive labelling of DNA (5).

Materials

and methods

The phase II clinical trial evaluating the combination of cisplatin and ifosfamide against disseminated malignant melanoma has been described previously (1). Patients with disseminated malignant melanoma gave informed consent to surgical removal of subcutaneous metastatic tissue prior to establishing xenograft lines. Line ‘St?, however, was established from a surgically removed brain metastasis, after the donor patient had achieved a partial remission on the combination of cisplatin and ifosfamide. NIH-Swiss background nude (nu/nu) mice were used as recipients for xenografts. Details on transplantation and calculation of relative growth delay values from serial tumour volume measurements have been published previously (9, lo), and details of drug application are given in Table 1. Assays ofmacromolecular DNA damage were performed after in uivo treatment of human melanoma xenografts according to the methods outlined previously (5, 18), except that cells from xenografts treated in vivo were not irradiated.

Table

1. Details

Drug DBD

DDP DTIC IF MeCCNU

Mel

MMC PHM Mel

of drug

administration

Vehicle 20% 20% 60% Phys. Aqua Phys. 10% 10% 80% 0.8% 7.2”/0 92% Phys. Phys. Phys.

DMSO Emulphor phys. N&l N&l dest. N&l Ethanol Emulphor phys. NaCl Ethanol Propylenglycol phys. NaCl NaCl NaCl NaCl

Dose’ 80

9 100x4 300 18

6

3x4 33 x 2 8

Schedule

LD 10/30b

Single dose

Single dl, 8, Single Single

dose 15, 22 dose dose

Single dose

dl, 8, 15, 22 dlx2 Single dose

6.6 194.0 244.0 23.7

4.3

61.0

‘All drugs were given intraperitoneally. Doses in mg/kg body weight. bLD IO/30 values were determined from single-dose &posures. DBD, dibromodulcitol; DDP, cisplatin; IF, Ifosphamide; Mel, melphalan; MMC, mitomycin C; PHM, malonato-diaminocyclohexane-platinum.

DRUG

87

RESISTANCE

Results Sensitivity patterns From 50 melanoma tissue established and propagated a minimum of three drugs Figure 1 (a) illustrates

specimens of different donor origin, 25 xenograft lines were at least once. Seventeen of these xenograft lines were exposed to and form the subject of this report. tumour volume responses of individual nude mice bearing

I

0

I

I

I

1020304050 Days

.

0 1020304050 from iniliotion

I

of

I1

I

I

1

I

0 1020304050 treatment

I t i

0.1

t 2 .-: ‘0

IO

i I

0.0 I 0 lb)

b’igure 1. Relative turnour sin& dose of dacarbazinc. F, 200 mg/kg. (b) with

IO 20

30

40

50 Days

0 IO 20 30 from initiotion

40 of

50 0 IO 20 treatment

30

40

50

volumes of individual nude mice bearing xenograft line ‘Str’ after treatment (a) with a A, untrcatcd controls; B, 1.5 mg/kg; C, 3.12 mg/kg; D, 6.25 mg/kg; E, 12.5 mg/kg; single doses of methyl-CCNU. A, untreated controls; B, 1.12 mg/kg; C, 2.25 mg/kg; D, 4.5 mg/kg; E, 9.0 mg/kg; F, 18 mg/kg.

88

R.

OSIEKA

Dose

PA LDlo]

Figure 2. Dose-response curves for dacarbazine, melphalan and methyl-CCNU using human melanoma xenograft line ‘St?. Growth delay values are plotted as the dependent variable and drug doses are expressed as fractions of the respective LD lo/30 value. 1, dacarbazine; 0, melphalan; A, methyl-CCNU.

xenograft line ‘Str’ to dacarbazine. Increasing doses lead to more pronounced transient tumour regressions with complete and permanent regressions occurring at doses in excess of 12.5 mg/kg. This dose is far below the LD IO/30 value of 194 mg/kg dacarbazine. A similar behaviour of relative tumour volumes was recorded for methyl-CCNU (Fig. 1 (b)), but permanent tumour regressions were only achieved with doses approaching the LD lo/30 level of 23.4 mg/kg. The relative therapeutic ranking ofdacarbazine, methyl-CCNU and melphalan against human melanoma xenograft line ‘Str’ is best illustrated by dose-response curves, that display the tumour inhibitory effect by growth delay values and doses as fractions of respective LD lo/30 values. It can be seen in Figure 2 that more than a hundredfold differences in dose must be risked for equal growth delay depending on the drug in use. Thus, a high degree of intraindividual heterogeneity of response to drug treatment is evident, and the notion ofglobal sensitivity or resistance ofa tumour line is repudiated. The donor patient ofxenograft line ‘Str’ achieved only a partial remission with the combination

- (0) 0.01

0

1 IO

I 20

I 30 Days

Figure

3. Relative

turnour

volumes

I 40

I 50

from

of individual nude (b) 200 mg/kg

0

initiation

(b) I IO

I 20

I 30

I 40

I 50

of treatment

mice bearing dacarbazine.

xenograft

‘GrII’.

(a) untreated

controls,

DRUG Table Tumours

2.

Growth Cont

GR II

4. I

Ja 1

4.0

Ja II

2.8

KZI

8.9

Ki

2.5

KU

6.2

M0

3.4

Ni

9.5

Str

4.1

St

3.6

Zi

12.1

delay DDP -0.2 R 1.3 S I.0 R 0.3 R 5.1 S 1.5 R 4.1 R 2.4 S

Flo

12.4

R 0.2 R -0.4 R 0.7

Avo

9.1

0.8

Kr 1

8.1

Da I

19.8

Dc II

14.5

Ah

9.4

-0.4 R >2

values”. DBD -0.4 5.1

0.2 0.7 0.8 0.1 0.9

~ ~

89

RESISTANCE

Clinical

responses

DTIC -0.3 R -0.2

1.3 R I.0 R 0.6 R -0.1 R r2 R >2 S >2 S 0.1 R >0.5 R -0.2 R -0.1 R >2 R >2 S -0.1 R

IF

MeCCNIJ

-0.1 R 2.6 S 1.5 R 0.1 R 4.4 S 0.5 R 0.2 R 0.8

-0.2

>2

>2

MMC

PHM

-0.4

0.8

2.1

1.5

0.9 0.8

4.5

-0.1

2.5

1.7

0.5

3.9 -0.1

0.9

1.0

-0.1

-0.1

.~ -~ >2

0.1

S R 0.4 R -0.2 R >2

2.9

3.1

-0.2 0.5

-0.4 R >2

b-2

1.4

Cant, tumour volume doubling time for untrcatcd , not evaluated. “Values>2 were taken for growth delay, if tumour reached within 60 days.

animals; volume

R, resistant; doubling

S. srnsitivc; times

were

nor

of &platin and ifosfamide, but a complete remission when treatment was switched to dacarbazine. In contrast to this very sensitive line, melanoma xenograft line ‘GrII’, which was established from a donor patient considered refractory to cisplatin, dacarbazine and ifosfamide, never responded to any drug treatment. Thus, there was no difference in relative tumour volumes between untreated nude mice bearing xenograft line ‘GrII’ and animals treated with 200 mg/kg of dacarbazine (Fig. 3). S ince no agent was effective when applied at the LD lo/30 level, this xenograft line serves as an example of uniform drug resistance. In Table 2 the responses of all melanoma xenograft lines to treatment with the panel of DNA-damaging agents is indicated in a comprehensive manner by growth delay values. Clinical responses to cisplatin, dacarbazine and ifosfamide are noted below, although donor patients received cz’splatin and ifosfamide always in combination. With a growth delay value of 2 as a cutoff point, clinical responses correlated well with preclinical

90

R.

OSIEKA

drug evaluation. If the xenograft model yielded growth delay values < 2, progressive disease occurred in the donor patient in 26/27 instances. Conversely, a growth delay value > 2 corresponded to the absence of progressive disease in lo/13 instances. Actually, one complete remission, two partial remissions and one no-change status were observed in donor patients after treatment with dacarbazine. The combination of cisplatin and ifosfamide yielded no complete remissions, two partial remissions and one no-change status. In one patient (Jai) a partial remission obtained with the combination of Lisplatin and ifosfamide corresponded to a growth delay value < 2 for cisplatin and > 2 for ifosfamide, thus hinting at ifosfamide as the superior component. When this patient progressed on the combination of cisplatin and ifosfamide, another xenograft line (JaII) was established and reduced growth delay values were obtained for both drugs. Dacarbazine was falsely indicated as effective in 2 instances, with one of the donor patients receiving dacarbazine at 600 mg/m2 instead of the usual schedule of 250 mg/m2 daily x 5. Sensitivity to cisplatin was falsely indicated by a growth delay value of 4.1 in a patient, who died from progressive pleural and pulmonary disease manifestations within a week after receiving the combination of cisplatin and ifosfamide. In this case there was a growth delay < 2 for ifosfamide. Drugs not given to the donor patients yielded growth delay values > 2 in nine instances, indicating a therapeutic potential not fully exploited by conventional treatment approaches. Overall 12/ 17 human melanoma xenograft lines responded with a growth delay > 2 to at least one drug, which again underscores the potential of sensitivity testing not harnessed by this retrospective evaluation. The scatter of preclinical positive drug responses (delined by growth delay values > 2) within a given xenograft line or to a given drug among xenograft lines of different donor origin support the contention of tumour heterogeneity with respect to therapy. The number of xenograft lines investigated in this manner is still too small to estimate the true prevalences of unique vs. uniform chemosensitivity patterns.

123456 Fraction

number

Fipm.4. Elution patterns of DNA from cell suspemiom prepared from panel: Ice-cold tumour cells rcccivcd increasing doses of irradiation. Right panel: Elution patterns ofDNAfrom cells that were isolated 6 h increasing doses of dacarbazine. E, 50 mg/kg; F, 100 mg/kg, and experiments for A and D, n = 20; for F and G, n = 4, for E, n = 3; included for fraction 5.

human melanoma xenograft line ‘St?. Left A, 0 Gy; B, 1.5 Gy; C, 3 Gy; and D, 6 Gy. after in vim treatment ofxenograft ‘St? with G, 200 mg/kg. Number of independent C, n = 2; and B, n = 1. S.E.M. values are

DRUG

Molecular

91

RESISTANCE

pharmacology

The high degree of differential sensitivity to dacarbazine encountered among xenograft lines ‘Str’ and ‘GrII’ was analyzed at the level of molecular pharmacology. Decomposition in vivo liberates a monomethylating species from dacarbazine giving rise to ‘alkali labile sites’ in cellular DNA. Under alkaline conditions (pH = 12.1) DNA transits to singlestranded state and ‘alkali labile sites’ convert into single-strand breaks. These are detected by an increased penetration or elution rate from filters with the appropriate pore geometry. Ionizing irradiation results in lesions of the DNA macromolecule that affect DNA elution rates in an analogous manner. The left panel of Figure 4 shows typical elution patterns of 7 6E’..-.... 5-

. ..

4x-y

=./ .

3s

‘1

2-

‘\

“‘3 ‘\

I-

\ ‘&--.

!zI -----------

o-l -

-2-t

I

a)

I

I

I

6.,........

5-

.f.O

-Id -2-

( b) I

I

I

I

6543-E.

..

2-

“.E., ..

Io-

/

;;$&iGj

-I I:

/ ‘“;cj-al I I

I 6

I 12

I 24 Hours

fG~ure 5. Kinetics of DNA damage in human melanoma xenografts. (a) ‘St?, (h) ‘Grll’ and (c) murinc hone marrow ‘BM’. DNA damage was expressed in Gy-equivalents and plotted as function of time after in viva trcatmcnt with increasing doses of dacarhazinc. (- - - -) 200 mg/kg (-) 100 mg/kg; and (- ~ -) 50 “g/kg. Ncgatiw values result from calculation of irradiation dose cquivalcnts by linear regression. Mean values *S.E.M. are given for 1, 6, 12 and 24 h.

92

K.

OSIEKA

DNA from human melanoma xenograft ‘Str’, which were obtained after irradiating ice-cold cell suspensions from this xenograft. The right panel shows elution patterns of DNA from the same source, when cells were isolated 6 h after in vivo exposure to increasing doses of dacarbazine. Since similar elution patterns are obtained after irradiation regardless of the source of cells (data not shown here), macromolecular DNA damage was always expressed in radiation dose (Gy) equivalents. DNA damage was then assayed over 24 h following drug treatment of mice bearing sensitive xenograft ‘Str’ or resistant xenograft ‘GrII’. Murine bone marrow cells were included in the analysis as a readily available normal reference tissue. Figure 5 displays the expression and removal of macromolecular damage expressed in Gy-equivalents for all three types of cells. Initially, at least for the highest dose level, DNA damage is expressed in all cells to a comparable degree, with bone marrow cells sustaining most changes. At 24 h DNA damage persists in the sensitive line ‘St? but not in resistant line ‘GrII’ and less so in murine bone marrow cells. This parallels the potential of murine bone marrow to recover to normal cellularity within a week after drug exposure. If lower doses of dacarbazine are employed, the differential response at the level of macromolecular DNA damage becomes less evident, in contrast to differential tumour volume responses which could be observed within a dose range of more than two decades. Secondary drug resistance Starting from initially high levels ofdrug sensitivity resistant sublines were developed from human melanoma xenograft ‘St?. The principle of suboptimal drug treatment and use of

0.11

0

I 5 Ooyr

I IO from

initiation

I I5

I 20

of tr*atm*nt

Figure 6. Selection of drug resistant xenograft lines after suboptimal drug exposure. At each generation some animals remained untreated (- - -) but the majority received suboptimal doses of treatment (pm ) causing different increases of tumor volume doubling times (TD). The xenograft with the shortest TD (arrow) is used for propagation of the resistant subline. Generation 8 of subline ‘Str/DTIC’ is used as an example.

DRUG

93

RESISTANCE

the xenograft with the shortest tumor volume doubling time for subsequent propagation is illustrated in Figure 6. Reiteration of this procedure over 14 transplant generations virtually abolished response to 3.1 mg/kg dacarbazine as shown by Figure 7. The degree of resistance is best described by changes of the dose-response curves which were constructed for the parent line ‘Str’ and the resistant sublines ‘Str/DTIC 3.1’ at transplant generations 8 and 16 (Fig. 8). The slope of the dose-response curve differs by a factor of 25 when the parent line ‘Str’ is compared to the resistant line ‘Str/DTIC 3.1’ at generation 16. 100

GO

50 20IO -

0.2

01 _--_

-

I

I

I

50 20 -

I

'Z 0 :

I

I

I

I

I

I

,f--

I

I

04

I

I

I

I

I

50 20 IO 5-

I

I

I

1

I

I

I

I

I

I

I

,

I

I

I

I

I

I

G6

I

I

I

I

010

I

I

I

I

012

I 013

I

I

I

011

,/*

I

I

05

67

09

50 20 -

I

,A

66

0.50.2-

I

/'

_----

5020IO -

0.5 0.2 -

1

,--

NM

0.2 =t 2 :

I

03

IO5- 2-

Me I'

(12

1

I

I

I

r.

I

I

I

I 20

I 25

014

IO -

0.5 0.2 0.1 0

I 5

1 IO

I 15

1 20

1 25

0 Days

Figure 7. Development mean

relative

tumour

I 5 from

I IO

I I5

initiation

I 20

I 26

0

I 5

I 10

I 15

of trratmrnt

ofresistance to dacarbazine for subline ‘Str/DTIC’ volumes of untreated control animals; (-),

over 14 transplant generations. ( treatment with 3.1 mg/kg dacarbazine.

-),

94

R.

OSIEKA

Dose DTIC Figure 8. Dose-response Equations for the linear

[y/kg]

curves for generations 0 (a), 8 (m), and 16 (A) of resistant regression lines are for generation 0:y = - 1. I + 2.67%; generation generation 16:~ = -0.3+0.10x.

sublines ‘Str/DTIC’. 8:y = - 1.3 + 0.78x;

Resistant sublines were not only developed to the presumably monofunctional alkylator dacarbazine but also to the bifunctionally alkylating agents melphalan and methylCCNU. In order to illustrate changes ofdose-response curves in a single graph, doses were expressed as fractions of the respective LD IO/30 values, and growth delay values were plotted as the dependent variable. The development of resistance to each agent becomes apparent by a shift of the dose-response curves for the resistant sublines to the right (Fig. 9). Ifthe resistant sublines were challenged with an agent other than the resistance-inducing drug, partial cross-resistance to ifosfamide, melphalan and methyl-CCNU was found for the dacarbazine-resistant line ‘Str/DTIC’. Cross-resistance was more complete, when

0.1

I

IO

100

1

1000

Dose (% LD,o)

Figure 9. Development A (

) ‘Str’;

A’(O)

of resistance to dacarbazine (A; A’; A”), methyl-CCNU (B; B’), and melphalan: ‘Str/DTIC’ gen 8; A” (0) ‘Str/DTIC’ gen 16. B (0) ‘Str’; B’ (A) ‘Str/MeCCNU’ C ( n ) ‘Str’; C’ (*) ‘Str/Mel’ gen 21.

(C, C’). gen 13.

DRUG

RESISTANCE

95

50 A

B

IO B2 9

5

i 2 .-:! 5

I

B 0.5

0.1

Days

from

initiation

of

trcotmanl

Figure IO. Cross-resistance patterns among resistant sublines from human melanoma xenograft line ‘Str’. Mean rclativc tumor volumes of A: parental xcnograft lint ‘Str’; B, sublinc ‘Str/DTIC’ gen 16; and C, ‘Str/Mcl gcn 2 1. -, untreated controls; 0, 3.1 mg/kg dacarbazine; 0, 300 mg/kg ifosfamide; 0, 5.3 ma/kg melphalan; n , 4.5 mg I kg methyl-CCNU.

melphalan or methyl-CCNU some activity (Fig. 10).

were

the challenging

agents,

but ifosfamide

still retained

Discussion The data obtained with the human melanoma xenografts in conjunction with the open phase II trial extend the validation of this new tumour model beyond published data for gastrointestinal and pulmonary neoplasms ( 17). Although the overall response rates in the xenograft model are slightly higher than response to single agents in clinical trials (2-4) resistance correlated in 26/27 and sensitivity in lo/13 instances. This analysis, however, is somewhat confounded by the combined use of cisplatin and ifosfamide in contrast to use of single agents in the xenograft model. Since 5/ 16 (3 1%) d onor patients responded to either dacarbazine or the combination of cisplatin and ifosfamide, successful heterotransplantation may itself constitute a selection factor in favour of chemosensitivity. A similar argument has recently been raised by Meyskens in view of results obtained with soft agar cloning of human tumour cells (8). The data illustrate that response to anti-neoplastic agents is quite heterogeneous, even if the drugs tested share a common mechanism of action. This statement depends, of course, on the assumption that murine LD lo/30 levels bear sufficient relevance to clinical drug exposures (7). Most xenograft lines revealed unique chemosensitivity patterns, but some were uniformly resistant to the panel of DNA-damaging drugs. Considering the limited range of this drug panel and data obtained independently with in vitro tumour models (8, 19) unique chemosensitivity patterns probably prevail in many human melanoma tumour samples. Using the subrenal capsule assay in nude mice, similar results pointing

96

K. OSIEKA

at individualistic chemosensitivity patterns were obtained (6). The long time required to obtain enough xenograft-bearing mice after tumour excision virtually excludes subcutaneous xenografts from predictive testing. Subrenal capsule assays or human tumour colony forming assays must then help to uncover the hitherto unexploited therapeutic potential. Subcutaneous xenografts have a useful application in tumour pharmacology, which is restricted by ethical considerations in the cancer patients ( 18). The range of sensitivity to dacarbazine found in melanoma xenografts from different donors exceeds the range previously reported for human melanoma lines in tissue culture (11) by almost a decade. Macromolecular damage reflected this range of sensitivity less impressively. This may be due to cellular heterogeneity within human melanoma tissue samples, where vital, necrobiotic, and dead cells are present. Since nonspecifically degraded DNA from dead cells confounds the analysis of drug-induced damage, cell separation procedures have to be employed. Therefore, no clear distinction is possible between drug-induced or spontaneous necrosis 24 h after drug exposures. Furthermore, host clearing mechanisms cannot be accounted for after in vivo drug exposure. Despite these obvious limitations the xenograft model opens an avenue that extends pharmacologic studies from determinations of drug levels to molecular pharmacology within the neoplastic tissue (10, 18). The findings with DNA damage after in vivo exposure to dacarbazine are at variance with a report from Parsons el al. (1 I), who found no change in elution rates after induction of resistance in a human melanoma cell line in permanent tissue culture. This indicates a need to study mechanisms of resistance both after in vitro and in vivo drug exposure. Studies on cross-resistance have often relied on tumour lines made resistant by some selection procedure after suboptimal drug exposures (13, 14, 15). Induction of resistance in a sensitive line from a panel of human melanoma xenografts with known clinical responsiveness allowed for a comparison of primary cross-resistance patterns in the panel with those found for secondary resistance. While cross-resistance was shown to be incomplete in both instances, secondary resistance was not representative of the whole spectrum of cross-resistance patterns. Thus xenograft line ‘Str’ and the donor patient were more sensitive to dacarbazine than to the bifunctional alkylating agents cisplatin and ifosfamide. In contrast to this finding, cross-resistance was probably more complete if bifunctional alkylating agents rather than dacarbazine were the resistance-inducing drugs. Furthermore, a subline of murine EAT-tumours retained sensitivity to malonatodiaminocyclohexane platinum (PHM) after induction of resistance to cisplatin (15), whereas the derivative was in no instance superior to the parent compound in the battery of xenografts. The phenomenon of pleiotropic or multidrug resistance has been described in Chinese hamster cells as well as murine cell systems. Using anthracyclines, actinomycin, colchicine and vinblastine, the occurrence of .pleiotropic drug resistance was determined with the human tumour colony-forming assay (HTCFA). The highest degree of crossresistance was noted between agents with similar modes of action, but the patterns of drug resistance more often were heterogenous ( 16). Th ese investigations were restricted to DNAdamaging agents but the conclusions are similar. In conventional murine tumour systems cross-resistance patterns among alkylating agents have recently been described as incomplete (13, 14). The problem of pleiotropic or multidrug resistance is most urgent in medical oncology, since cells with such phenotype escape combination chemotherapy even if applied in predetermined alternating ‘non cross-resistant’ regimens. The xenograft model offers exciting potential for studies of resistance since, in contrast to

DRUG

HTCFA, manifold

RESISTANCE

97

permanent lines can be established for extended investigations to overcome the problems of resistance to chemotherapy encountered among human neoplasms. Summary

Alkylating agents and their functional analogues belong to the most useful antineoplastic drugs in the treatment of disseminated malignant melanoma. In conjunction with an open clinical phase II trial evaluating the combination of cisplatin and ifosfamide, 17 melanoma xenograft lines were established from patients often refractory to dacarbazine (DTIC). These xenograft lines were exposed to cisplatin, dacarbazine, dibromodulcitol, ifosfamide, methyl-CCNU, mitomycin C, and malonato-diaminocyclohexane-platinum II (PHM) at the respective LD lo/30 doses. Growth delay values < 2 corresponded in 26127 instances with progressive disease, whereas values > 2 corresponded in only lo/13 instances with achievement of a no-change status or a partial remission of the donor patient’s disease. Among the panel of DNA-damaging agents tested, cross-resistance was incomplete. Some xenograft lines revealed unique chemosensitivity patterns in contrast to a uniform pattern of drug resistance in others (pleiotropic or multidrug resistance). The data confirm independently of results obtained in the phase II study that the combination of cisplatin and ifosfamide is effective against malignant melanoma refractory to dacarbazine. Suboptimal drug exposure, repeated up to 2 1 transplant generations, was employed to induce secondary resistance to either dacarbazine, melphalan or methyl-CCNU in a melanoma xenograft line originally quite sensitive to drug treatment. When the resistant sublines were exposed to the other agents, only partial cross-resistance was observed. Tumour volume responses to treatment with dacarbazine correlated with persisting DNA damage assayed 24 h after in vivo drug exposure in a sensitive line and the absence of such lesions in a resistant line. References I.

Becher,

R.,

Seeber,

S. & Schmidt,

dichlorodiammineplatinum 2. Bellet, R. E., Mastrangelo, DTIC (NSC-45388) metastatic malignant

C. G.

(1980)

Combination

chemotherapy

with

ifosfamide

and

cir-

(II) in advanced malignant melanoma. 3. Cancer Res. Clin. Oncol. 97: 301-306. M. J., Laucius, J. F. & Bodurtha, A. J. (1976) Randomized prospective trial of

alone versus BCNU (NSC-409962) plus Vincristine melanoma. Cancer Treat. Rep. 66: 595-600.

3. BcIIct, R. E., Catalano, R. B., Mastrangclo, in patients with metastatic melanoma

(NSC-67574)

M. J. & Bcrd, D. (1978) Positive phase refractory to DTIC and a nitrosourea.

in the treatment

II trial ofdibromodulcitol Cancer Treat. Rep.

of

62:

2095-2099. 4. B&et, R. E., Mastrangelo, M. J., Berd, D. & Lustbader, E. (1979) Chemotherapy of metastatic malignant melanoma. In: Clark, W. H. Jr., Goldman, L. I. & Ma?trangelo, M. J. (eds) Human malignant melanoma. New York/San 5. Erickson, unlabeled 169~- 174.

Francisco/London: Grune & Stratton, pp. 325 351. L. C., Osieka, R., Sharkey, N. A. & Kahn, K. W. (1980) Measurement of DNA damage in mammalian cells analyzed by alkaline elution and a fluorometric DNA assay. Anal. Biochem. 106:

between response to 6. Giovanella, B. C., Stehlin, J. S., Shepard, R. C. & Williams, L. J. (1983) C orrelation chemotherapy of human tumors in patients and in nude mice. Cancer 52: 1146-l 152. 7. Guarino, A. M., Rozencweig, M., Kline, I., Penta, J. S., Venditti, J. M., Lloyd, H. H., Holzworth, D. A. & Muggia, F. M. (1979) Adequacies and inadequacies in assessing murine toxicity data with antineoplastic agents. Cancer Res. 39: 2204--22 10. 8. Meyskens, F. L. Jr. (1983) In vitro sensitivity clinical response: Results and limitations. International

Congress

of Chemotherapy,

of clonogenic human melanoma cells to therapeutic In: Spitzy, K. H. & Karrer, K. (eds) Proceedings Vienna,

28 August

to 2 September,

1983, Part

agents and of the 13th

224, pp. 5 8.

98

R.

9. Osieka, R., Houchens, xenografts in athymic 10. Osieka, R. & Schmidt, ofhuman melanomas .Nude Mice. Bozcman, Il.

Parsons, resistant

OSIEKA

D. P., Coldin, A. & Johnson, R. K. (1977) nude mice. Cancer 40: 264c-2650. and acquired resistance C. G. (1982) 1’ rlmary

Chemotherapy to alkylating

of human

colon

cancer

agents in hctcrotransplants

and colon carcinomas. In: Reed, N. D. (cd.) Proceedings ofthe 3rdInternational Workshot, MT, Sept. 6-9, 1979. Stuttgart, New York: G. Fischer Vcrlag, pp. 675 683.

P. G., Smellie, S. G., Morrison, L. E. & Hayward, to 5-(3’,3’-dimethyl-l-triazeno)imidazole-4-carboxamide

on

I. P. (1982) Properties of human melanoma ccl15 and other methylating agents. Cancer Res.

42: 1454-1461. 12. Povlsen, C. 0. &Jacobsen, G. K. (1975) Chemotherapy ofa human malignant melanoma transplanted in the nude mouse. Cancer Re.s. 35: 2790-2796. 13. Schabel, F. M. Jr, Skipper, H. E., Trader, M. W., Laster, W. R., Griswold, D. P. Jr. & Corbett, T. H. (1983) Establishment of cross-resistance profiles for new agents. Cancer Treat. Rep. 67: 9055922. 14. Schmid, F. A., Otter, G. M. & Stock, C. C. (1980) Resistance patterns of Walker Carcinosarcoma other rodent tumors to cyclophosphamide and L-phenylalanine mustard. Cancer Res. 40: 83Cb-833.

256 and

15. Sccbcr, S., Osicka, R., Schmidt, C. G., Achterrath, W. & Crooke, S. T. (1982) In viva resistance towards anthracyclines, etoposide, and cis-diamminedichloroplatinum (II). Cancer Res. 42: 4719-4725. 16. Shoemaker, R. H., Curt, G. A. & Carney, D. N. (1983) Evidenceformultidrug-resistant cellsin human tumor ccl1 populations. Cancer Treat. Rep. 67: 883~-888. 17. Steel, G. G., Courtenay, V. D. & Peckman, M. J. (1983) turnour xenografts. Br. J. Cancer 47: 1~ 13. 18. Thomas,

C. B., Osieka,

R. &

Kahn,

K. W.

chloroethyl)-3-(4-methylcyclohexyl)-l-nitrosourea grafts in nude mice. Cancer Res. 38: 2448-2454.

(1978)

The response DNA

of sensitive

to chemotherapy

cross-linking and resistant

ofa variety

by in uzuo treatment human

19. Von HOE, 1). D., Forscth, B., Mctclmann, H. R., Harris, G., Rowan, S. & Coltman, cloning of human malignant melanoma in soft agar culture. Cancm 50: 696 701.

colon

ofhuman with

carcinoma

C. A. (1982)

l-(2xenoDirect