[CANCER RESEARCH 49. 7132-7140. December 15. 1989]

Heterogeneous Response to Differentiation Induction in Different ClonalSubpopulations of a Rat Rhabdomyosarcoma Cell Line (BA-HAN-1)1

C. D. Gerharz,2 H. E. Gabbert, R. Engers, U. Ramp, H. Mayer, H. K. Biesalski, and C. Luley

Departments oj Pathology ¡C.l). G.. H. E. G., K. E., U. K.J. internal Medicine ¡H.M., C. L.J, and Biochemistry ¡H.K. B.¡,Johannes Gutenberg University of Main:D-650ÕIMain:, Federal Republic of Germany

ABSTRACT

Three clonal subpopulations (A, B, C) isolated from the same rhabdomyosarcoma of the rat were tested and compared for their susceptibilityto differentiation induction using retinole acid (RA), dimethylformamide(DMF), and /V-monomethylformamide (NMF). These subpopulationsdiffer in that a block to spontaneous differentiation is imposed at differentstages which are characteristic for each subpopulation. Whereas tumorcell proliferation was significantly inhibited (/' < 0.001) in all three

subpopulations, the effects of RA, DMF, and NMF on tumor celldifferentiation were strikingly heterogeneous. The response was mostmarked in subpopulation C, as evidenced by a significant increase in thenumber of terminally differentiated myotube-like giant cells (/' < 0.001)

and in biochemical differentiation, as indicated by the creatine kinaseactivity (P < 0.05). Between 5% (DMF and NMF) and 30% (RA) of themononuclear cells in subpopulation C exhibited thick and thin myofila-ments, which were never observed in the mononuclear cells of the control.In contrast, subpopulation A and B responded to RA, DMF, and NMFquite heterogeneously with an increase in biochemical differentiation,whereas terminally differentiated myotube-like giant cells were neverobserved. These results demonstrate that the therapeutic potential ofdifferentiation induction in malignant tumors may be impaired by tumorheterogeneity.

INTRODUCTION

In many cancers, at least some of the tumor cells spontaneously exhibit abortive attempts at normal differentiation (1).Furthermore, induction of differentiation has been successfullyachieved in several in vitro systems using a wide variety ofchemical substances including RA' (2-10), DMF (11-15), and

NMF (14-20). These observations indicate that the malignantphenotype is not necessarily an irreversible state and that theconversion of malignant cells to a more benign phenotype,through induced differentiation, may become an alternativetherapeutic approach (21-25). Less attention, however, hasbeen paid to the possible limitations of therapeutically induceddifferentiation. These limitations arise from the fact that tumorcells are heterogeneous. Thus, in recent years ample evidencehas been accumulated to indicate the simultaneous coexistenceof diverse clonal subpopulations within one tumor, differing inboth phenotype and functional characteristics (26) or in theirresponse to chemotherapeutic agents (27-29). Consequently aheterogeneous response to differentiation induction is also tobe expected. We recently described the effects of retinoic acidon the differentiation and proliferation of the clonal rat rhab-domyosarcoma cell line BA-HAN-1 C (30). We have now ex-

Received10/24/88:revised4/17/89.9/18/89:accepted9/21/89.Thecostsof publicationof this article«eredefrayedin part bythe payment

of pagecharges.Thisarticlemustthereforebe herebymarkedadvertisementinaccordancewith18U.S.C.Section1734solelyto indicatethisfact.

1This work supported by the Gesellschaft der Gönner und Förderer derGrundlagenforschung des Krebses.

2To whom requests for reprints should be addressed, at. Department ofPathology. Johannes Gutenberg University of Mainz. D-6500 Main/.. FederalRepublic of Germany.

'The abbreviations used are: RA. retinoic acid; DMF. A'.A'-dimethylformam-ide; NMF, A'-monomethylformamide; HPLC. high-performance liquid chroma-

tography; FCS. fetal calf serum; IC!0. concentration inhibiting population doublings by 50r'c.

tended our investigations to three phenotypically distinct clonalsubpopulations (A, B, C) derived from the same dimethyl-benzanthracene-induced rhabdomyosarcoma of the rat. As recently described (31), the common constituents of these threeclonal subpopulations are morphologically undifferentiatedmononuclear stem cells. In subpopulation C, some of thesemononuclear tumor cells spontaneously fuse to form multinu-clear myotube-like giant cells with ultrastructural features ofrhabdomyogenic differentiation. These myotube-like giant cellswere shown to have become irreversibly withdrawn from themitotic cycle representing terminally differentiated postmitotictumor cells (32). The cells of subpopulation B do not exceedthe stage of mononuclear desmin-positive rhabdomyoblasts invitro, but exhibit morphological features of rhabdomyogenicdifferentiation after retransplantation in vivo. The least differentiated subpopulation A is composed of partially desmin-positive tumor cells without the morphological features ofrhabdomyogenic differentiation, both in vitro and in vivo. Inthis report, the effects of RA, DMF, and NMF on thesephenotypically divergent subpopulations are compared.

MATERIALS AND METHODS

Cells and Culture

The clonal subpopulations A, B, and C used in this study werederived in our laboratory from the same dimethylbenzanthracene-in-duced rhabdomyosarcoma in rat (31). The clonal origin of these celllines had been confirmed by repeated recloning procedures and investigations were performed with cultures between passages 5 and 20. Thestandard growth medium was Dulbecco's modified Eagle medium

(DMEM, GIBCO Europe, FRG), supplemented with 10% heat-inactivated FCS, penicillin, and streptomycin. The same batch of FCS wasused for all experiments, to eliminate any possible changes in quality.Unless otherwise noted, cultures were refed after 4 days. The tumorcells were cultured in 25 cm2 and 80 cm2 Nunclon-flasks (GIBCOEurope, FRG) and incubated in an atmosphere with 5% CO2 at 37°C.

The cells were detached from the surface of the tissue culture flasks byexposure to 0.05% EDTA. Cell counts were performed with a Neubauerhemocytometer chamber.

Induction of Differentiation

Retinoic Acid Treatment. As described previously (30), a stock solution of 5 m\i retinoic acid (Serva, FRG) was prepared in 95% ethanol,sterilized by filtration and stored at —¿�20°C.For differentiation induc

tion the stock solution was diluted in standard growth medium to aconcentration of 1 MM.By means of transmission electron microscopyand the trypan blue exclusion test, this concentration was shown to benoncytotoxic to the tumor cells. Prior to the experiments, the purity ofthe commercial standard and the calculated concentration of retinoicacid were checked using HPLC (33, 34). The HPLC-analysis demonstrated a ratio of 13-m-:all-ira/ii-retinoic acid of 1:3. Further derivativeswere not detected. The total vitamin A concentration in the normalstandard medium supplemented with 10% FCS was 0.3 n\t. For lightprotection, the culture flasks were wrapped in aluminum foil. To ensurethat the effects we observed with retinoic acid were not due to ethanol,analogous control experiments were performed, where the tumor cellswere exposed to growth medium supplemented with 0.02% ethanol.

7132

Research. on August 14, 2021. © 1989 American Association for Cancercancerres.aacrjournals.org Downloaded from

HETEROGENEOUS RESPONSE TO DIFFERENTIATION INDUCTION

These control experiments failed to produce any detectable effect onthe proliferation and differentiation of the tumor cells.

Treatment with DMF and NMF. DMF (Serva, FRG) and NMF(Serva, FRG) were prepared as a 10% (v/v) stock solution in DMEMmedium and added to the cultures to yield a final concentration of 1%(v/v). The trypan blue exclusion test and transmission electron microscopy showed that this concentration was noncytotoxic to the tumorcells.

Assessment of Differentiation in Vitro

In Vitro Morphology. For transmission electron microscopy, thetumor cells were seeded on glass cover slips. After incubation for 10days, the tumor cells were fixed in situ by exposure to a 2.5% sodium-cacodylate buffered glutaraldehyde solution (0.1 M; pH 7.4) and post-fixed in a 1% sodium-cacodylate buffered osmium tetroxyde solution(0.1 M;pH 7.4) prior to Epon embedding. Thin sections were contrastedwith uranyl acetate and lead citrate. Electron photomicrographs weretaken with an EM-410 Philips transmission electron microscope. Phasecontrast photomicrographs were taken with a Leitz Labovert invertedmicroscope.

Creatine Kinase Activity. The total creatine kinase activity was usedas a biochemical marker of differentiation (6, 35, 36). Five replicateculture flasks were exposed to growth medium supplemented with IJIMretinoic acid, 1% (v/v) DMF, and 1% (v/v) NMF, respectively. Theculture medium was completely replaced every 3 days by fresh differentiation inducing media. After 10 days, the tumor cells were harvestedby exposure to 0.05% EDTA. The number of cells harvested wasdetermined with the Neubauer hemocytometer chamber. The cells weredisrupted by sonication and the lysate was centrifuged at 10,000 rpmfor 30 min at 4°C.The total creatine kinase activity was determined at30°Con an Olympus AU-5031 analyzer using the CK-test (NAC-

activated) from Merck (Darmstadt, FRG).The kinetics of the total creatine kinase activity were analyzed

separately for subpopulation C. Briefly, 25 replicate culture flasks wereexposed to each differentiation inducing medium. As control, 25 replicate culture flasks were exposed to standard growth medium. In eachexperiment, cells from five culture flasks were harvested separatelyevery day for 5 days. Cells were not refed during this period. The basiccreatine kinase activity (time point 0) was determined separately in fivesamples. The cells harvested were further processed as described above.The data was statistically analyzed by the Wilcoxon test for unpairedsamples.

Creatine Kinase Isoenzymes. It has been shown for normal embryonicrhabdomyogenesis that the BB (fetal) isoenzyme is gradually replacedby the MM (adult) isoenzyme going through an intermediate MB type(37, 38). The lack of biochemical differentiation in rhabdomyosarcomacells is indicated by the persistent synthesis of the BB isoenzyme (6,36). The profile of creatine kinase isoenzymes was analyzed afterexposure to retinoic acid, DMF, and NMF for 10 days. The culturemedium was completely replaced every 3 days by fresh differentiationinducing media. After 10 days, the cells were harvested and furtherprocessed as described above. The creatine kinase ¡soenzymeswereseparated according to the electrophoretic mobility on agarose gel usingthe TITAN GEL REP CK-30 isoenzyme procedure (Helena Laboratories, TX). Quantitation of isoenzymes was accomplished by denso-metric scanning with the Helena REP densitometer (Helena Laboratories).

Fusion Assay. 3 x IO5 tumor cells of subpopulation C each wereseeded into 25-cnr culture flasks. On the bottom of these culture flasksfour arbitrarily located fields had been marked. The area marked outby these four fields was '/J2 the growth area of the culture flask. After24 h the standard growth medium was completely substituted by thedifferentiation inducing media, i.e., growth medium supplemented withRA, DMF, or NMF, respectively. The number of myotube-like giantcells in the marked fields was counted by phase contrast microscopy atintervals of 24 h. Cells that contained three or more nuclei wereclassified as myotube-like giant cells. The effects of RA (0.1 and l UM),DMF (0.1 and 1%), and NMF (0.1 and 1%) were evaluated in fivereplicate culture flasks per concentration. As a control, the frequencyof myotube-like giant cells was determined in five culture flasks eachwith standard growth medium or standard growth medium with 0.02%ethanol, respectively. At the end of the observation period, the totalnumber of tumor cells was determined in each culture flask. The cellswere refed after 5 days. The relative frequency of myotube-like giant

cells, i.e., the ratio between the number of myotube-like giant cells (in'/)2 the growth area of the culture flask) and the total number of cells

per culture flask, was calculated. This ratio was then analyzed by ananalysis of variance with two independent factors.

Assessment of Growth Properties in Vitro. Fifteen replicate 25-cnrculture flasks were exposed to each differentiation-inducing medium.As a control, 15 replicate 25-cnr flasks were exposed to standardgrowth medium. Each culture flask was seeded with 5 x IO4cells. In

each experiment, cells from three culture flasks were harvested separately each day, for 5 days, and hemocytometer cell counts with theNeubauer chamber were performed. Cells were not refed during thisperiod. The data were analyzed by an analysis of variance with twoindependent factors. The mean doubling time was graphically deducedfrom the growth curves.

The concentrations of RA, DMF, and NMF inhibiting populationdoublings by 50% (i.e., inhibitory concentrations, IC50)were determinedseparately as described by Hölzelet al. (39). Briefly, triplicate cultureflasks were exposed to various concentrations of retinoic acid (0.01,0.1, 0.5, 1, and 5 JIM), DMF, and NMF (0.01, 0.1, 0.5, 1, and 5%).After an exposure for 7 days without refeeding, the cells were harvestedand cell counts with the Neubauer hemocytometer chamber were performed. Interpolation from semilogarithmic plots of the dose-responsecurves was used for the determination of IC50.

RESULTS

Assessment of Differentiation

In Vitro Morphology

Subpopulation A. After exposure to RA, DMF, and NMF,the predominantly spindle-shaped tumor cells of subpopulationA (Fig. la) exhibited corresponding phenotypic alterations byphase-contrast microscopy. Ten days after plating in thesedifferentiation-inducing media, the tumor cells were larger andmore flattened (Fig. \b) and showed by transmission electronmicroscopy a marked increase of rough endoplasmatic reticu-lum (Fig. \d) when compared to the control (Fig. le). Morphological features of rhabdomyogenic differentiation such as myotube-like giant cells or myofilaments could not be demonstrated. Nonspecific cytotoxic effects were excluded by thetrypan blue exclusion test and transmission electron microscopy.

Subpopulation B. Phase-contrast microscopy showed that thelarge and polygonal tumor cells of subpopulation B (Fig. le)did not change their morphological phenotype after exposureto the differentiation-inducing media for 10 days (Fig. I/).Correspondingly, the ultrastructural aspects of the tumor cellsdid not change. Thus, the tumor cells exhibited abundant andoften dilated profiles of rough endoplasmatic reticulum both instandard growth medium (Fig. \g) and in medium supplemented with the different differentiation inducers (Fig. ih).Morphological features of rhabdomyogenic differentiation suchas myotube-like giant cells or myofilaments could not be observed. Nonspecific cytotoxic effects were excluded by the try-pan blue exclusion test and transmission electron microscopy.

Subpopulation C. After exposure to RA, DMF, and NMF for10 days, the mononuclear tumor cells of subpopulation C weremore elongated and spindle-shaped (Fig. 1b) when comparedto the control (Fig. 2a). Transmission electron microscopyshowed that between 5% (DMF and NMF) and 30% (RA) ofthe mononuclear tumor cells exhibited irregular bundles of thin(6-8 nm in diameter) and thick (12-15 nm in diameter) myofilaments (Fig. 2a"). '•£•.morphological features of rhabdomy

ogenic differentiation that had never been observed in theirmononuclear counterparts under standard growth conditions(Fig. 2c). The ultrastructural characteristics of the myotube-like giant cells observed after exposure to RA, DMF, or NMF

7133

Research. on August 14, 2021. © 1989 American Association for Cancercancerres.aacrjournals.org Downloaded from

HETEROGENEOUS RESPONSE TO DIFFERENTIATION INDUCTION

f ^;riM^vCr

Fig. 1. a-</, morphology of subpopulation A before and after differentiation induction: small spindle-shaped mononuclear cells exhibiting a criss-crossed growthpattern in standard growth medium (a) as opposed to the more regular arrangement of larger and flattened cells after differentiation induction, here with Ie; NMF(A). Tumor cell with abundant profiles of rough endoplasmatic reticulum after exposure to lrr NMF (d) when compared to the control culture (<•).e-h. morphologyof subpopulation B before and after differentiation induction: large polygonal mononuclear tumor cells in standard growth medium (e) exhibiting abundant roughendoplasmatic reticulum dilated by moderately electron dense material (#). No noteworthy morphological changes after differentiation induction, here with 1'; DMF(/. h). a.h.e.f: bar. 100 ^m. c.d.x.h: bar. I ^m.

did not differ from those of their multinuclear counterpartsunder standard growth conditions (30, 31) and myofibrils witha sarcomeric organization were not seen in vitro. Nonspecificcytotoxic effects were excluded by the trypan blue exclusiontest and by transmission electron microscopy.

Creatine Kinase Activity

The total creatine kinase activity of the three subpopulationsgrown in standard growth medium for 10 days differed according to their degree of spontaneous differentiation. The total

7134

Research. on August 14, 2021. © 1989 American Association for Cancercancerres.aacrjournals.org Downloaded from

HETEROGENEOUS RESPONSE TO DIFFERENTIATION INDUCTION

>< - ' -

Fig. 2. Morphology of subpopulalion C before and after differentiation induction: small spindle-shaped mononuclear cells intermingled with scarce myotubc-likegiant cells (arrow) under standard growth conditions (a) as opposed to numerous myotube-like giant cells (arrows), here after exposure to l /IM RA (b). Mononucleartumor cell in standard growth medium (e) lacking morphological features of rhabdomyogenic differentiation. Mononuclear tumor cell afler exposure to I <IMRA (</).exhibiting numerous irregular bundles of thick and thin myofilaments (star and inset), a.h: bar. 100 /jm: c. bar. 1 ^m; d. bar. 5 firn; inset: bar, I >im.

creatine kinase activity was minimal in the least differentiatedsubpopulation A (0.7 ±0.2 milliunit/10" cells) and maximal in

the spontaneously most differentiated subpopulation C (64 ±8milliunits/10" cells). Subpopulation B exhibited a creatine kinase activity of 18 ±14 milliunits/10" cells after 10 days in

standard growth medium. The response of the three differentsubpopulations to treatment with RA, DMF, or NMF, respectively, proved to be strikingly heterogeneous. Subpopulation A(Fig. 3) responded with a statistically significant increase (P <0.05) of creatine kinase activity to treatment with RA (1.5 ±0.4 milliunits/lO" cells) and NMF (2 ±0.6 milliunits/106 cells),but not to DMF (0.8 ±0.4 milliunits/10" cells). Subpopulation

B (Fig. 3) only responded to treatment with DMF (54 ±29milliunits/10" cells; P < 0.05), whereas both NMF (35 ±5milliunits/10" cells) and RA (28 ±23 milliunits/10" cells) failed

to induce a statistically significant increase in the creatinekinase activity. Subpopulation C exhibited a statistically significant increase (P < 0.05) in the total creatine kinase activityafter exposure to RA (1420 ±212 milliunits/10" cells), DMF(602 ±109 milliunits/106 cells) and NMF (243 ±32 milliunits/10" cells) for 10 days (Fig. 3). The kinetics of creatine kinase

activity in subpopulation C (Fig. 4) revealed a statistically

significant increase in enzyme activity, which became evident 3days after exposure to RA. The kinetics of creatine kinaseactivity after exposure to RA closely corresponded to the kinetics of myotube formation (see below; Fig. 5a). In contrast,the increase in creatine kinase activity became evident already24 h after exposure to DMF and NMF. Under the conditionsof this experiment (no refeeding), this increase, however, provedto be reversible and enzyme activity did not significantly differfrom the control 5 days after exposure to DMF and NMF.Thus, the effects of NMF on creatine kinase activity precededthe increase in the number of myotube-like giant cells (seebelow. Fig. 56).

Creatine Kinase isoenzymes

The profiles of creatine kinase isoenzymes in subpopulationsA, B, and C are presented in Table 1. Each subpopulationexhibited BB (fetal) and MM (adult) isoenzymes. In standardgrowth medium, the proportion of the MM (adult) isoenzymewas maximal in subpopulation A (89%) and minimal in sub-population B (38%). In subpopulation A, exposure to RA,DMF, and NMF resulted in a complete shift of the isoenzymeprofile, the creatine kinase activity being entirely of the MM

7135

Research. on August 14, 2021. © 1989 American Association for Cancercancerres.aacrjournals.org Downloaded from

HETEROGENEOUS RESPONSE TO DIFFERENTIATION INDUCTION

(adult) type after exposure to differentiation inducing media.In subpopulation B, only minor shifts of the isoenzyme profilewere observed after exposure to RA, DMF, and NMF. Insubpopulation C, exposure to RA resulted in a marked decreasein the proportion of the BB (fetal) isoenzyme from 52% (standard growth medium) to 30% (RA) and a concomitant increasein the MB ieoenzyme from 4% (standard growth medium) to34% (RA). In contrast, exposure of subpopulation C to DMFand NMF resulted in an increase in the proportion of the BB(fetal) isoenzyme from 52% (standard growth medium) to 68%

1500-

I 1000-

500-

100-

1 = control2 = RA3 = DMF4 = NMF

subpopulotion subpopulation subpopulotionABC

Fig. 3. Creatine kinasc activity of subpopulations A. B, and C in percentageof the control after exposure to 1 MMRA. \% DMF. or 1% NMF for 10 days.Each value represents the mean ±the standard deviation of five replicate experiments. *, results statistically significant (P < 0.05: Wilcoxon test for unpairedsamples).

creatine kinase activity

1000

100-

1 2 3 4 5time in cultureldays)

Fig. 4. Kinetics of creatine kinase activity in subpopulation C after exposureto 1 (jM RA. lr; DMF. or lcr NMF for 5 days without refeeding. Each value isexpressed as percentage of the control and represents the mean ±standarddeviation of five replicate experiments. The increase was statistically significant(P < 0.05: \\ ilcoxon test for unpaired samples) on days I to 4 after exposure toDMF. on days I to 3 after exposure to NMF and on days 3 to 5 after exposureto RA.

i :1 :

101

control 0.1pM RA

n= 33! 2

j

RAn = 2300 •¿�297

012345 012345 01231.5time in cultureldays) time in cultureldoysl time in culture Idaysl

control 0.1 % NMF

12345' m cultureldays)

01 2345time in culture (days)

012345time in cultureldays)

Fig. 5. Absolute number of myotube-like giant cells of subpopulalion C afterincubation in medium supplemented with RA (a) or NMF (A). Each valuerepresents the mean ±the standard deviation of five replicate experiments. Afurther numerical evaluation of myotube-like giant cells was omitted after exposure to RA for 3 days.

(DMF) and 62% (NMF). This increase in the proportion of thefetal BB isoenzyme, however, does not indicate a loss of biochemical differentiation, because the marked increase in thetotal creatine kinase activity (see above; Fig. 3) also resulted ina marked increase in the MM (adult) isoenzyme activity.

Fusion Assay

The fusion assay could only be performed in subpopulationC, because myotube-like giant cells were not observed in sub-population A and B after exposure to differentiation-inducingmedia. In subpopulation C the exposure to RA for 72 h resultedin a marked dose- and time-dependent increase (Fig. 5a) in theabsolute number of myotube-like giant cells from 33 ±2 (control) to 234 ±80 (0.1 MMRA) and 2300 ±297 (1 ^M RA).Because the fusion rate is cell density dependent, the possibilityhad to be excluded that the increased number of myotube-likegiant cells was only caused by a higher cell density in thosecultures exposed to RA. The relative frequency therefore of

7136

Research. on August 14, 2021. © 1989 American Association for Cancercancerres.aacrjournals.org Downloaded from

HETEROGENEOUS RESPONSE TO DIFFERENTIATION INDUCTION

Table I I'rofi/es of the creatine kinase isoen:ymes KB, MB, and MM before and after exposure lo I nM R.4, 1% DMF, and 1% NMF for IO days

Each value represents the mean of duplicate samples and is expressed as percentage of the total creatine kinase activity.

Subpopulation

ABC

IsoenzymeBB

MBMMControlII89RA°100DMF100NMF°100Control62 38RA60 40DM

P6832NMF72

28Control524

44RA°303436DMF"6832NMF°62 38

' Statistically significant (I' < 0.05) increases in the total creatine kinase activity (see Fig. 3).

myotube-like giant cells, i.e., the ratio between the number ofmyotube-like giant cells (in Vu the growth area of the cultureflask) and the total number of tumor cells per culture flask wascalculated. For this ratio, a significant (P < 0.001) dose-dependent increase became evident from 4 x IO"6 ±O (control)to 30 x 10 6 ±13 x 10~6(0.1 MMRA) and 310 x IO'6 ±30 xIO'6 (1 JIMRA) after 72 h.

Exposure to NMF resulted in a time- and dose-dependentincrease (Fig. 5b) in the absolute number of myotube-like giantcells from 276 ±64 (control) to 384 ±130 (0.1% NMF) and959 ±364 (1% NMF). This increase, however, became evidentonly 120 h after exposure to NMF. The increase of the relativefrequency of myotube-like giant cells from 15 x 10~6 ±4 xIO"6 (control) to 22 x IO'6 ±7 x 10~6(0.1% NMF) and 83 x10~6 ±33 x 10~" (1% NMF) was statistically significant (P <

0.001).After exposure to DMF, a significant increase in the relative

frequency of myotube-like giant cells could not be demonstratedeven after 120 h. Only after 10 days, did we have the impressionthat the number of myotube-like giant cells markedly increasedin cultures exposed to 1% DMF. However, no correct numericalevaluation of the myotube-like giant cells was possible at thistime, because there was a marked piling up of the tumor cells,especially in the control cultures.

Assessment of Proliferation

Under the conditions of our experiments, exposure to RA,DMF, or NMF, respectively, resulted in a significant (P <0.001) inhibition of proliferation in all three subpopulations.The response, however, proved to be strikingly heterogeneous.Subpopulation A (Fig. 6) responded most markedly to bothDMF and NMF, the cell number being reduced to 4 ±1%(DMF) and 3 ±1% (NMF) of the control after 7 days. Incontrast, the cell number was reduced to only 48 ±2% of thecontrol after exposure to RA over the same period of time.Subpopulation B (Fig. 7) responded best to DMF with a reduction of the cell number to 10 ±3% of the control after 7 days,whereas the response to NMF (33 ±0.3% of the control) andRA (63 ±0.3% of the control) was less pronounced. Subpopulation C (Fig. 8) most markedly responded to DMF (8 ±1% ofthe control) and RA (14 ±2% of the control), but to a lesserdegree to NMF (53 ±2% of the control).

These differences in the degree of growth inhibition are alsoreflected by differences in the concentrations necessary to inhibit population doublings by 50% (Table 2).

DISCUSSION

In this report, we compared the susceptibility to differentiation induction in three different clonal subpopulations (A, B,C) isolated from the same rhabdomyosarcoma of the rat. Thesesubpopulations had previously been shown to be intrinsic constituents of the primary tumor and to exhibit a marked inter-clonal and intraclonal phenotypic heterogeneity (31). Thus, the

1234567time in culture {days)

Fig. 6. Growth curves of subpopulation A in standard growth medium (c =control) and in medium supplemented with 1 MMRA, Kr DMF, or 1'V NMF,respectively. Each value represents the mean ±the standard deviation of threereplicate experiments. The difference between the growth curves is statisticallysignificant (P < 0.001; analysis of variance with two independent factors).

107:

= 10'

10s:

subpopulationB

1 2 3 ¿ 5 6 7time in culture(days)

Fig. 7. Growth cunes of subpopulation B in standard growth medium (c =control) and in medium supplemented with 1 ^M RA, \% DMF, or 1% NMF,respectively. Each value represents the mean ±the standard deviation of threereplicate experiments. The difference between the growth curves is statisticallysignificant (P < 0.001; analysis of variance with two independent factors).

subpopulations A, B, and C abortively recapitulate stages ofnormal embryonic rhabdomyogenesis or tissue renewal, butdiffer in that a block to differentiation seems to be imposed atdifferent stages characteristic for each subpopulation (Fig. 9).Whereas the mononuclear tumor cells of subpopulation C fuseto form terminally differentiated and irreversibly postmitoticmyotube-like giant cells, the cells of subpopulation B are confined to the stage of morphologically differentiated rhabdomyo-

7137

Research. on August 14, 2021. © 1989 American Association for Cancercancerres.aacrjournals.org Downloaded from

HETEROGENEOUS RESPONSE TO DIFFERENTIATION INDICTION

1234567time in culture (days I

Fig. 8. Growth curves of subpopulation C in standard growth medium (c =control) and in medium supplemented with 1 ^M RA, 1% DMF, or 1% NMF,respectively. Each value represents the mean ±the standard deviation of threereplicate experiments. The difference between the growth cunes is statisticallysignificant (P < 0.001; analysis of variance with two independent factors).

Table 2 .Mean doubling time of subpopulations A, B, and C in standard growthmedium and the concentrations of RA, DMF, and NMF inhibiting population

doublings by 50% (1C*,)

Subpopulation

B

Mean doubling time in standardgrowth medium

1C»RADMFNMF

15 h 17h 14 h

1 flM 2 flM 0.2 nM

0.6% 0.5% 0.5%0.6% 0.8% 1.0%

blasts with irregular bundles of myofilaments. Cells of subpop-ulation A do not exceed the stage of biochemically differentiatedrhabdomyoblasts with the synthesis of creatine kinase anddesmin. The present study demonstrates that these three different clonal subpopulations respond to differentiated inductionwith RA, DMF, and NMF. The susceptibility to differentiation

induction, however, was strikingly heterogeneous and mostpronounced in subpopulation C, as evidenced by an increase inboth biochemical and morphological differentiation and amarked increase in the proportion of irreversibly postmitoticmyotube-like giant cells. In contrast, the effects of differentiation induction in subpopulations A and B were confined to anincrease in biochemical differentiation as indicated by the creatine kinase activity. Neither the exposure to RA nor to DMFand NMF induced the formation of irreversibly postmitoticmyotube-like giant cells in these subpopulations. These observations suggest that the therapeutic potential of differentiationinduction in malignant tumors is not unlimited and that it mayonly be possible to utilize a predetermined range of differentiation. This predetermined range of differentiation, however, isnot the same in every subpopulation of a tumor. Thus, in ourtumor model, the range of differentiation can include the statusof terminally differentiated and irreversibly postmitotic tumorcells, but can also end at the level of only biochemically differentiated tumor cells (Fig. 9).

Comparing the effectiveness of RA, DMF, and NMF asdifferentiation inducers, none of these agents proved to beuniversally effective in all three subpopulations. Thus, RA wasa very potent inducer of biochemical and morphological differentiation in subpopulation C, but completely failed to inducedifferentiation in subpopulation B. On the other hand, thespontaneously least-differentiated subpopulation A obviouslyresponded to treatment with RA shown by an increase inbiochemical differentiation. The same holds true for DMF andNMF, which are closely related chemically and supposedly actvia related mechanisms (12, 40). Nevertheless, DMF failed toinduce an increase in creatine kinase activity in subpopulationA, whereas NMF successfully induced biochemical differentiation in subpopulation A. In contrast, NMF, but not DMF,failed to induce differentiation in subpopulation B. This inter-clonal heterogeneity of response to the same differentiationinducer might indicate subtle differences in signal transductionand biochemical pathways between the different subpopula-tions, the elucidation of which should provide further cluesabout the mode of action of RA, DMF, and NMF.

The response to differentiation induction was further modu-

Fig. 9. Range of differentiation of the threerhabdomyosarcoma subpopulations (A, B, andC) compared to normal muscle histogenesis:bars, limits of maximum differentiation characteristic for each subpopulation which cannotbe overcome by differentiation inducers. des,desmin: ck, creatine kinase: myo, myofilaments: *. observed after retransplantation inviro (31): **. observed after exposure to R A,

DMF, or NMF.

co

oDCLOQ-

JDDM

O

O

jQO

B

undifferentiated

7138

ck+, des* "Xk+.des+,myo+

ck+, des* ck+.des+.myo+

biochemically and morphologicallydifferentiated

terminally differentiated,irreversibly postmitotic

Research. on August 14, 2021. © 1989 American Association for Cancercancerres.aacrjournals.org Downloaded from

Ill I I K(K,I SI 01 S KISI'OSSI H) 1)11ITKINÕÕAÕÕONINDICTION

lated by an intraclonal heterogeneity of susceptibility as evidenced by the behavior of subpopulation C. Thus, not all themononuclear tumor cells exhibited morphological features ofrhabdomyogenic differentiation after exposure to RA, DMF.or NMF, and the great majority of the tumor cells failed to fuseto irreversibly postmitotic myotube-like giant cells. In a clonalsubpopulation, the coexistence of genetically diverse cell populations is not very likely to account for this partial refractoriness to differentiation induction as was suggested for othertumor models (41, 42). As previously shown (30). subclonesisolated from primarily RA-refractory tumor cells of subpopulation C responded to reexposure to RA in the same way as theoriginal population. This observation strongly suggests thatthose tumor cells, which do not primarily respond to differentiation induction, are not permanently differentiation defective,but may be epigenetically blocked in their response to differentiation inducers.

It is important to emphasize that differentiation induction inall three subpopulations was always accompanied by a significant inhibition of proliferation. Furthermore, a significant inhibition of proliferation was observed, even when no concomitant differentiation induction could be demonstrated (e.g., theeffects of RA and NMF on subpopulation B). As shown by thetrypan blue exclusion test and transmission electron microscopy, this antiproliferative effect of differentiation inducerscannot be explained by nonspecific cytotoxic effects. Therefore,the effects of differentiation inducers might encompass a widespectrum of functional and structural alterations that rangefrom a decrease of proliferative capacity to biochemical andmorphological differentiation, culminating in irreversibly postmitotic and terminally differentiated tumor cells. The relationship between proliferation and differentiation, however, is notwell understood. Interestingly in this context, cells of subpop-ulation C can also be induced to differentiate only by exposureto FCS-depleted medium that does not support cell proliferation (43). This observation might suggest a genetic programinteracting between the proliferation and differentiation of tumor cells, which is similar to that proposed for normal cells(25).

In summary, our results demonstrate some of the limitationsimposed on a successful therapeutic application of differentiation induction by the interclonal and intraclonal heterogeneityof tumors. Nevertheless, it is important to note that eachdifferentiation inducer inhibited tumor cell proliferation in allthree subpopulations. The limitations of conventional tumortherapy should therefore encourage a continued search forcombinations of differentiation inducing agents that synergis-tically act on the different subpopulations within a heterogeneous tumor.

ACKNOWLEDGMENTS

We would like to express our appreciation to A. Niederauer, K.Molter, H. Breitbach, and C. Burkner. as well as to K. Weber and W.Meyer for their excellent technical assistance. We are grateful to Dr.K. H. Schicketanz for his statistical evaluations.

REFERENCES

10.

11.

12.

13.

14.

15.

16.

17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

27.

28.

29.

30.

1. Pierce, G. B. The benign cells of malignant tumors. In: P. J. King (ed.).Developmental Aspects of Canccrogcnesis and Immunity, pp. 3-22. New 31.York: Academic Press. 1974.

2. Strickland, S.. and Mandavi. V. The induction of differentiation in terato-carcinoma stem cells by retinoic acid. Cell. IS: 393-403. 1978.

3. Lotan. R. Effects of vitamin A and its analogs (rctinoids) on normal and 32.neoplastic cells. Biochim. Biophys. Acta. 605: 33-91. 1980.

7139

Sporn. M. B., and Roberts. A. B. Role of rctinoids in differentiation andcarcinogenesis. Cancer Res.. 4Õ:3034-3040. 1983.Sherman. M. I.. Gubler. M. I,.. Barkai. I'.. Harper. M. !.. Coppola. G.. andYuan, J. Role of retinoids in differentiation and growth of embryonalcarcinoma cells. In: Retinoids. Differentialion and Disease, pp. 42-60. Cibafoundation Symposium 113. London: Pitman. 1985.Garvin. A. J.. Stanley. \V. S.. Bennett. D. lì..Sullivan. J. J.. and Sens. D. A.The in ritro growth, hetcrotransplantation. and differentiation of a humanrhabdomyosarcoma cell line. Am. J. Pathol.. I2S: 208-217. 1986.Paukovits. J. B.. Paukovits. \V. R., and l.aerum, (). D. Identification of aregulator) peplide distinct from normal granulocyte-derived hemoregulatorypeptide produced by human promyelocytic III.-60 leukemia cells after differentiation induction with retinoic acid. Cancer Res.. 46: 4444-4448. 1986.Sidcll. N.. Sarafin. T.. Kelly. M.. Tsuchida, T.. and Haussler. M. Retinoleacid-induced differentiation of human neuroblastoma: a cell variant systemshowing two distinct responses. Exp. Cell Biol.. 54: 287-300. 1986.I.ippman. S. M.. Kessler, J. F.. and Meyskens. F. L.. Jr. Retinoids aspreventive and therapeutic anticancer agents (Part I). Cancer Treat. Rep..71: 391 -405. 1987.Lippman. S. M.. Kessler. J. F., and Meyskens, F. L.. Jr. Retinoids aspreventive and therapeutic anticancer agents (Part II). Cancer Treat. Rep..71: 493-515. 1987.Dexter. D. L. A'JV-Dimethylformamide-induced morphological differentia

tion and reduction of tumorigenicity in cultured mouse rhabdomyosarcomacells. Cancer Res.. 37: 3136-3140. 1977.Dexter, D. L.. and Hager. J. C'. Maturation-induction of tumor cells using a

human colon carcinoma model. Cancer (Phila.), 45: 1178-1184. 1980.Hager. J. C'., Gold, D. V.. Barbosa, J. A.. Fligicl. /.. Miller, F., and Dexter.D. L. /V.A'-Dimethylformamide-induced modulation of organ- and tumor-

associated markers in cultured human colon carcinoma cells. J. Nail. CancerInst., 64: 439-446. 1980.Dexter. D. L., Spremulli. E. N., Mattok. G. M.. Diamond, J.. and Calabresi.P. Inhibition of the growth of human colon cancer xenografts by polarsolvents. Cancer Res.. 42: 5018-5022. 1982.Christensen, T. G.. Burke. B.. Dexter. D. 1... and /amcheck. N. llltrastruc-tural evidence of dimelhylformumidc-induccd differentiation of cultured human colon carcinoma cells. Cancer (Phila.). 56: 1559-1565. 1985.Barclay. R. K., and Garfinkel. E. The influence of .V-methylformamide onformatc-C14 incorporation. II. In nucleic acids of tumor-bearing rats. CancerRes., / 7: 345-351. 1957.Dexter. D. L., Lee, E. S., Bliven. S. F., Glicksman. A. S.. and Leith, J. T.Enhancement by ,V-methylformamidc of the effects of ioni/ing radiation ona human colon tumor xenografted in nude mice. Cancer Res., 44: 4942-4946. 1984.Dibner. M. D.. Ireland. K. A., Koerner. L. A., and Dexter. D. !.. Polarsolvent-induced changes in membrane lipid internal diffusion in human coloncancer cells. Cancer Res., 45: 4998-5003. 1985.Langdon. S. P.. Hickman. J. A., Gescher, A., Stevens, M. F. G., Chubb, D.,and Vickers. L. M. A'-Methylformamide (NSC 3051): a potential candidate

for combination chemotherapy. Eur. J. Cancer Clin. Oncol.. 21: 745-752,1985.Iwakawa, M., Tofilon. P. J.. Hunter. N.. Stephens. L. C'., and Milas, L.Antitumor and antimelastatic activity of the differentiating agent A'-methyl-formamide in murine tumor systems. Clin. Expl. Metastasis. 5: 289-300.1987.Metcalf. D. How many cancers are reversible or suppressible? Pathology, 15:1-3, 1983.Spremulli. E. N.. and Dexter. D. L. Polar solvents: a novel class of antineo-plastic agents. J. Clin. Oncol., 2: 227-241. 1984.Frcshney. R. I. Induction of differentiation in neoplastic cells. AnticancerRes.. 5: 111-130. 1985.Sartorelli, A. C. Malignant cell differentiation as a potential therapeuticapproach. Br. J. Cancer. 52: 293-302. 1985.Sachs. L. Cell differentiation and bypassing of genetic defects in the suppression of malignancy. Cancer Res.. 47: 1981-1986, 1987.Owens. A. H., Jr., Coffey, D. S., and Baylin. S. B. (eds.). Tumor cellheterogeneity: origins and implications. Orlando. Ft.: Academic Press, 1982.Calabresi. P., and Dexter. D. I.. Clinical implications of cancer cell heterogeneity. In: A. H. Owens, Jr., D. S. Coffey. and S. B. Baylin (eds.). TumorCell Heterogeneity: Origins and Implications, pp. 181-201. Orlando. FL:Academic Press, 1982.Goldic. J. Drug resistance and chemotherapeutic strategy. In: A. H. Owens.Jr., D. S. Coffey. and S. B. Baylin (eds.). Tumor Cell Heterogeneity: Originsand Implications, pp. 115-125. Orlando. FL: Academic Press. 19X2.Trope, C. Different susceptibilities of tumor cell subpopulations to cytotoxicagents and therapeutic consequences. In: A. H. Owens. Jr.. D. S. Coffey. andS. B. Baylin (eds.). Tumor Cell Heterogeneity: Origins and Implications, pp.147-168. Orlando. FL: Academ ¡cPress. 1982.Gabbcrt, H.. Gerhar/. C. D.. Bicsalski. H. K., Engers. R., and Luley, C.Terminal differentiation and growth inhibition of a rat rhabdomyosarcomacell line (BA-HAN-IC) in ritro after exposure to retinoic acid. Cancer Res.,48: 5264-5269, 1988.Gcrharz. C. D.. Gabbert. H.. Moll. R., Mcllin, W., Engers, R., and Gabbiani.G. The intraclonal and interclonal phenotypic heterogeneity in a rhabdomyosarcoma cell line with abortive imitation of embryonic myogenesis.Virchows Arch. B (Cell Pathol.). 55: 193-206. 1988.Gabbert. H. E.. Gerharz. C. D., Engers, R.. Müller-Klieser.W., and Moll,R. Terminal!) differentiated postmitotic tumor cells in a ral rhabdo-

Research. on August 14, 2021. © 1989 American Association for Cancercancerres.aacrjournals.org Downloaded from

HETEKOGENEOrs RESPONSE TO DIFFERENTIATION INDUCTION

myosarcoma cell line. Virchows Arch. B (Cell Pathol.). 55: 255-261, 1988.33. Annesley, F., Giacherio. O.. and SVilkerson, K. Analysis of retinoids by high

performance liquid chromatography using programmed gradient separation.J. Chromatogr.. 305: 199-203. 1984.

34. Biesalski. H. K.. and Weiser. H. Sensitive analysis of retinyl esters by isocraticadsorption chromatography. J. Clin. C'hem. Clin. Biochem., 27:65-74. 1989.

35. Delaporte. C.. Dautreaux. B.. and Fardeau. M. Human myotube differentiation in vitro in different culture conditions. Biol. Cell.. 57; 17-22, 1986.

36. Dexter. D. L., Konieczny. S. F., Lawrence, J. B.. Shaffer, M.. Mitchell. P..and Coleman, J. R. Induction by butyrate of differentiated properties incloned murine rhabdomyosarcoma cells. Differentiation, 18: 115-122, 1981.

37. Foxal, C. D., and Emery, A. E. H. Changes in creatine kinase and itsisoenzymes in human fetal muscle during development. J. Neurol. Sci.. 24:483-492, 1975.

38. Turner. D. C., Gmür.R., Siegrist. M.. Burckhardt. E., and Eppenberger. H.M. Differentiation in cultures derived from enmbryonic chicken muscle. I.Muscle-specific enzyme changes before fusion in EGTA-synchronlzed cultures. Dev. Biol.. 48: 258-283. 1976.

39. Holzel. F.. Albrecht. M.. Simon, \V. E., Hansel. M.. Metz. R.. Schweizer. J.,and Dietel. M. Effectiveness of antineoplastic drugs on the proliferation ofhuman mammary and ovarian carcinoma cells in monolayer culture. J.Cancer Res. Clin. Oncol., 109: 217-226, 1985.

40. Cordeiro. R. F.. and Savarese. T. M. Role of glutathione depletion in themechanism of action of A'-niethylformamide and A'.A'-dimethylformamide ina cultured human colon carcinoma cell line. Cancer Res.. 46: 1297-1305,1986.

41. Sherman, M. I.. Eglitis. M. A., and Thomas, R. Reversible and irreversibleeffects of retinol upon the phenotypic properties of embryonal carcinomacells. J. Embryol. Exp. Morphol.. 93: 179-196. 1986.

42. Zile. M. H., Cullum. M. E.. Simpson. R. U.. Barua. A. B.. and Swartz. D.A. Induction of differentiation of human promyelocytic leukemia cell lineHL-60 by retinyl glucoronidc. a biochemically active metabolite of vitaminA. Proc. Nati. Acad. Sci. USA. 84: 2208-2212, 1987.

43. Gerharz. C. D.. Gabbert, H. E.. Biesalski. H. K.. Engers. R., and Luley. C.Fetal calf serum and retinole acid affect proliferation and terminal differentiation of a rat rhabdomyosarcoma cell line. Br. J. Cancer. 59: 61-67, 1989.

7140

Research. on August 14, 2021. © 1989 American Association for Cancercancerres.aacrjournals.org Downloaded from

1989;49:7132-7140. Cancer Res   C. D. Gerharz, H. E. Gabbert, R. Engers, et al.   Cell Line (BA-HAN-1)Different Clonal Subpopulations of a Rat Rhabdomyosarcoma Heterogeneous Response to Differentiation Induction in

  Updated version

  http://cancerres.aacrjournals.org/content/49/24_Part_1/7132

Access the most recent version of this article at:

   

   

   

  E-mail alerts related to this article or journal.Sign up to receive free email-alerts

  Subscriptions

Reprints and

  .pubs@aacr.orgDepartment at

To order reprints of this article or to subscribe to the journal, contact the AACR Publications

  Permissions

  Rightslink site. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC)

.http://cancerres.aacrjournals.org/content/49/24_Part_1/7132To request permission to re-use all or part of this article, use this link

Research. on August 14, 2021. © 1989 American Association for Cancercancerres.aacrjournals.org Downloaded from