Selective Killing of Oncogenic Human Cells Cocultivated With Normal Human Fibroblasts 1,2 Robert C. Warrington
3,4
ABSTRACT -Hlstldlnol and theophylline reversibly arrested the growth of low-passage human foreskin fibroblasts (HFF) and protected these cells from an otherwise lethal combination of the S-phase drugs ,B-cytosine arabinoside and hydroxyurea. In contrast, HeLa cells continued cell cycle transit in the presence of hlstldinol and theophylline. Consequently, these reagents did not alter the vulnerability of He La cells to combined S-phase drugs. The differential S-phase drug sensitivity mediated by histldlnol and theophylline persisted under cocultivatlon conditions, thereby allowing the selective and total eradication of colonies of HeLa cells In the presence of subconfluent HFF cells. The HFF cells, spared S-phase drug toxicity by their unique response to histldlnol and theophylline, showed total survival and vlability.-J Natl Cancer Inst 61: 69-73, 1978.
MATERIALS AND METHODS HFF were obtained from D. L. Engelhardt (Columbia University) and used at less than 20 population doublings. HFF were propagated in minimum essential medium supplemented with 10% fetal bovine serum. Stocks were maintained by passage at low density, whereas cells for experiments were harvested from confluent, quiescent cultures. HeLa cells, obtained from P. I. Marcus (University of Connecticut), were propagated in Dulbecco's modified Eagle's medium supplemented with 5% calf serum. Colonies of HeLa cells were established by plating approximately 103 cells onto 100mm culture dishes. L-Histidinol (Calbiochem) was prepared, and cell numbers and volumes were determined as described earlier (8). Photographs were taken with a Nikon AFM 35-mm photomicrographic attachment on a Nikon inverted MS microscope with the use of phase-contrast optics supplied by P. Goetinck (University of Connecticut). Hydroxyurea (Calbiochem) and Ara-C (Aldrich Chemical Co.) were used at 2 mM and 10 ILg/ml, respectively, either alone or in combination, as described in the figure legends. Histidinol and theophylline (Calbiochem) were used individually or in combination at various concentrations (see figure legends).
A major limitation of currently used cancer chemotherapy is the failure of anticancer agents to differentiate between normal and malignant cells. In general, proliferating cells are more vulnerable to cancer drugs than are their nonproliferating counterparts (1). Consequently, crucial populations of cycling normal cells, such as those of the myelopoietic, gastrointestinal, and epidermal tissues, are at risk during chemotherapy. This suggests that appropriate manipulations of the proliferative status of normal and malignant cell populations could improve the therapeutic index of cell cycledependent cancer drugs. Since these drugs are preferentially toxic for cycling cells, their administration under conditions wherein only neoplastic cells are allowed to proliferate should result in selective cell killing. This prediction has been verified in vitro. Pardee and James (2) and O'Neill (3) in 1975 published procedures that allow selective killing of transformed cells, as have others more recently (4, 5, 6). Unfortunately, the reagents used to date to modify the proliferation of normal and malignant cells possess too great an inherent toxicity for in vivo applicability (4, 7). L-HistidinoI. a structural analog of the essential amino acid histidine, elicits differential growth responses from oncogenic and non oncogenic mouse cell lines. The analog maintains the nontumorigenic Balb/3T3 and 3T3 cells in the Go state [(8); Yen A, Warrington R, Pardee AB: Submitted for publication] but does not restrict the proliferation of a variety of tumorigenic mouse cell lines including L929, SVIOI, SVA31, and 3Tl2 (Warrington R, Hechtman R: Submitted for publication). Thus normal mouse cells are protected from S-phase drugs by the histidinol-mediated Go arrest, whereas transformed mouse cells, which continue cell cycle transit in the presence of the analog, retain their vulnerability to S-phase drugs [(9); Warrington R, Hechtman R: Submitted for publication]. I report here that histidinol and theophylline evoke similar responses from normal (low-passage HFF) and tumorigenic (HeLa) human cells. VOL. 61. NO.1. JULY 1978
RESULTS HFF, chosen as an In vitro prototype of nontransformed human cells, were harvested from confluent cultures and plated into fresh medium (text-fig. IA). The confluent state of cultures prior to harvest provided a population of cells most of which were in the quiescent or Go state. This was indicated by the prolonged delay, typical of Go-recovery, observed before the control cultures divided (text-fig. lA) and was verified by the observation that such populations of cel.ls doubled their cell volume and entered an S-phase pnor to the first division (Warrington R: Unpublished ABBREVIATIONS liSEn: Ara-C=,B-cytosine arabinoside; HFf=human foreskin fibroblast(s); HU=hydroxyurea. Received November l. 1977; accepted february l. 1978. Supported by Public Health Service research grant CAl4733 from the National Cancer Institute. This research benefitted significantly [rom the use of the University of Connecticut Cell Culture facility, also supported by this grant. 3 Microbiology Section. University of Connecticut. Storrs, Conn. I
2
06268. 4 Address reprint requests to Dr. Warrington at his present address:
Department of Cancer Research. University of Saskatchewan. Saskatoon. Saskatchewan S7N OWO. Canada.
69
J
NATL CANCER INST
70 Warrington data). Unlike nontransformed mouse cells quiescent HFF cells were not maintained in Go for extended periods by the addition of L-histidinol, although the initiation of division and the apparent doubling time were influenced in a dose-dependent manner. Theophylline had a similar effect, but neither histidinol nor theophylline alone gave significant protection from S-phase drugs. However, a persistent and reversible growth arrest was observed (text-fig. IB) when these reagents were combined. Furthermore, histidinol and theophylline offered near total protection from the Sphase drugs Ara-C and HU. No such protection was offered by these reagents alone. Detailed studies with Balb/3T3 cells, harvested from quiescent monolayers and subsequently treated with histidinol, established that the analog maintained these cells in Go (8). The similarities of the responses, recovery kinetics, and
1A
1B CONTROL CONTROL
prolonged protection from S-phase drugs shown by HFF treated with histidinol and theophylline suggest that these cells are also maintained in the Go state. HeLa cells, known to be oncogenic by the nude mouse assay (10), responded differently to histidinol and theophylline (text-fig. 2). Histidinol inhibited the growth of HeLa cells, and the appearance of dead cells in the medium of treated cultures indicated that it was toxic at all concentrations studied. Theophylline altered the morphology and reduced the growth rate of HeLa cells but was not obviously toxic. However, the combination of histidinol and theophylline was more toxic than histidinol alone. Furthermore, the combination altered the susceptibility of HeLa cells to S-phase drugs, since it reduced significantly the apparent killing rate of HU but enhanced slightly the killing rate of Ara-C and HU in combination. The inability of histidinol and theophylline to protect HeLa cells from Ara-C and HU indicated that these cells continued cell cycle transit in the presence of these reagents. The differential responses of HFF and HeLa cells to histidinol and theophylline suggested that HeLa cells could be eradicated selectively in the presence of HFF by appropriate pretreatments. The subline of HeLa cells
W III
:::IE ::I
Z
...J ...J W
;r-
T 105i""~~-(H+T)
105
()
H +T +(AraC+HU)
...J
i!
let
~ rr-
T
Z
c:
i:
m
:JI
II:
w
:::IE ::I
Z
CONTROL+(AraC+HU)
10'~~
= Histidinol = Theophylline =.Hydroxyurea = Cytosine arabinoside
__L - - L______~L--L~__J-~__L -______-L10'
1234
NATL CANCER INST
....w....
~
0
H+T
~
r5
10"
10 ~ rr-
z c
AraC
(.)
....
i:
j!
III
m jJ
~ HU
10'
10'
= =
Histidinol H T Theophylline HU = Hydroxyurea AraC Cytosine arabinoside
123456 DAYS
TEXT·FIGURE I.-Response of human foreskin fibroblasts to L-histidinol. theophylline. and S-phase drugs. Tissue culture dishes (60 mm) were seeded with approximately 105 HFF cells harvested from quiescent monolayers. Cell numbers were determined in duplicate at 24-hr intervals. Ara-C and HU were added at 10 #,g/ml and 2 mM. respectively. as indicated for the duration of the experiment. A: Cell numbers are plotted for untreated control cultures (0) and cultures receiving 5 (0) and 10 (e) mM histidinol. 2 mM theophylline (t.). or histidinol and theophylline (8) (5 and 2 mM. respectively) shortly after plating. Ara-C and HU were added on day 2 to cultures pretreated with 2 mM theophylline (+). B: Cell numbers are plotted for untreated control cultures (0). for control cultures that received Ara-C and HU on day 2 (e). and for cultures treated (after plating) with histidinol and theophylline (5 and 2 mM. respectively) with (II) or without (t.) the addition of Ara-C and HU on day 2. On day 3. many of the plates of the set treated with histidinol and theophylline were washed several times and incubated in fresh medium (A). ]
106
H
III
H T HU AraC
CONTROL
CONTROL
III
T+(AraC+HU)
~
28
2A
a
H(2x)
II:
107
107
H+T (Reverse)
=
1~ 1
2
3
4
5
2
6
3
4
5
103
DAYS
TEXT·FlGURE 2.-Response of HeLa cells tOL-histidinol. theophylline. and S-phase drugs. Cell numbers were determined and Ara-C and HU were used as noted in the legend of text-fig. I. A: Cell numbers are plotted for control cultures (0) and cultures receiving 3 mM histidinol (e).2 mM theophylline (t.). or histidinol and theophylline (8. 3 and 2 mM. respectively) shortly after plating. HU (+). Ara-C (0). or HU and Ara-C (~) were added on day 2 to cultures pretreated with histidinol and theophylline (~). B: Cell numbers are plotted for control cultures (0) and cultures receiving Ara-C (0). HU (+). or Ara-C and HU (.) on day 2. VOL. 61. NO.1. JULY 1978
Selective KIlling of Oncogenic Human CeUs In Vitro 71
used in this study could form discrete and uniform colonies when grown from single Eells (fig. la). This property allows clear distinctions between HFF cells (fig. lb) and HeLa cells in mixed cultures prepared by adding HFF cells to plates bearing established HeLa colonies. These cell types did not intergrow freely under these cocultivation conditions, and, consequently, the HeLa colonies remained visible as islands of morphologically distinct cells surrounded by fibroblasts (fig. le).
Mixed cultures were prepared and, after the HFF cells attached, were treated with histidinol and theophylline. One day later (fig. Id) the HFF cells assumed quiescent morphology, whereas mitotic activity and toxicity effects were evident in HeLa colonies. Ara-C and HU were added, and a progressive and total destruction of HeLa colonies was observed over the subsequent 48 hours (figs. Ie, f). Drugs and debris were then removed and, after several washings with fresh medium, the cultures were incubated to assess the status of the HFF survivors. These grew and established typical fibroblastic monolayers (figs. Ig-i). The monolayers were stained and no foci of HeLa cells (visible in mixed cultures not treated with S-phase drugs) were observed. The survival of oncogenic cells was significant if histidinol and theophylline were not added prior to Sphase drug administration. Further, large numbers of HFF cells were killed and those surviving were megaloblastic and incapable of proliferating when the pretreatment was omitted. Therefore, histidinol and theophylline are required not only for the survival of HFF cells but also for the successful eradication of the tumorigenic component of this cocultivation system. These experiments indicate that the differential growth responses of HFF and HeLa cells to histidinol and theophylline are maintained under conditions of cocultivation and demonstrate that appropriate manipulation of the proliferative status of normal and tumorigenic cells can enhance the selectivity of cell cycledependent anticancer drugs. DISCUSSION
Cultured eukaryotic cells possessing an untransformed phenotype enter a viable state of quiescence termed Go (11) when subjected to a multiplicity of conditions that are not optimal for proliferation. In contrast, transformed cells continue cell cycle transit and often show substantial cell death under similar restrictive conditions. This difference between normal and transformed cells in vitro, documented and discussed by many workers, is lucidly formalized in Pardee's "restriction point" control hypothesis (12). Thus the distinct responses of normal and transformed cells to essential amino acid deprivation provide a partial rationale for the observations reported here. LHistidinol, a competitive inhibitor of histidyl-tRNA synthetase (13), establishes a persistent functional deprivation for the essential amino acid L-histidine in a variety of cultured eukaryotic cells (13-16). This observation, considered with the ability of L-histidinol not VOL. 61, NO. I. JULY 1978
only to inhibit specific aspects of macromolecular synthesis in uninfected (17) and virus-infected (18, 19) mouse cells but also to evoke physiologic alterations (16) characteristic of the negative pleiotypic response (20), suggested that the analog would arrest normal mouse cells in the Go-state but would not arrest the growth of transformed cells. As indicated in the introduction, this prediction was verified for a variety of mouse cell lines. In view of the foregoing, we were surprised that no concentration of histidinol alone provided a persistent growth arrest for HFF cells. Similarly, HeLa cells showed an unanticipated resistance to S-phase drugs in the presence of histidinol. Thus HFF and HeLa cells do not respond in the same manner to this analog as do their mouse counterparts. Whether these differences reflect a peculiarity limited to these particular human cell types or indicate a general distinction between cultured mouse and human cells is uncertain. However, theophylline, an inhibitor of cAMP phosphodiesterase (21), sufficiently augmented the differential effects of histidinol on HFF and HeLa cells to allow selective killing. Although the reason for these apparently synergistic interactions between histidinol and theophylline is unknown, the effect appears to be general. As reported here, theophylline prolongs the duration of the histidinol-mediated Go arrest in normal mouse cell lines' but enhances the sensitivity of histidinol-treated transformed mouse cell lines to certain S-phase drugs (Warrington R, Hechtman R: Submitted for publication). These observations were derived serendipitously from experiments designed to determine whether a theophylline-mediated elevation in intracellular cAMP would counteract the effects of histidinol on transformed cells (Warrington R: Unpublished observations). Such an effect is an a priori deduction from the demonstration (22, 23) that increases in intracellular cAMP cause certain transformed cells to undergo a morphologic reversion to a habitus approaching that of normal cells (reverse transformation). Although theophylline does alter the morphology of transformed mouse and human cells, this reagent does not diminish the propensity of transformed cells to continue cell cycle transit in the presence of histidinol. On the contrary, theophylline increased the toxicity of histidinol in all transformed cells tested. Rationalizations for the various effects of the histidinol-theophylline combination cited herein await further investigation. Experiments similar to those reported here have been performed with an HFF-SVIOI (simian virus 40-transformed 3T3 cell) cocultivation system. In contrast to the HFF and He La cell combination used in this study, HFF and SVIOI cells intergrow freely under cocultivation conditions. Nevertheless, the administration of Sphase drugs, after appropriate pretreatment with histidinol and theophylline, allowed total eradication of SVIOI cells and total (apparent) survival and viability of the HFF cells (Warrington R: In preparation). Thus the selective killing of transformed cells reported in the present study appears to depend on specific alterations of the proliferative status of the cell types used rather
J
NATL CANCER INST
72 Warrington than a result of any peculiar interaction between them. The ability of histidinol and theophylline to restrict the cytoxicity of S-phase drugs to the oncogenic component of in vitro cocultivation systems suggests a potential applicability in cancer chemotherapy. The possibility that such an approach can be exploited to enhance the specificity of cell cycle-dependent anticancer drugs in vivo is under investigation.
REFERENCES (1) VALERIOTE F, VAN PUTIEN L: Proliferation-dependent cytotoxicity of anticancer agents: A review. Cancer Res 35:2619-2630, 1975 (2) PARDEE AB, JAMES LJ: Selective killing of transformed baby hamster kidney (BHK) cells. Proc Natl Acad Sci USA 72:49944998, 1975 (3) O'NEILL FJ: Selective destruction of cultured tumor cells with uncontrolled nuclear division by cytochalasin B and cytosine arabinoside. Cancer Res 35:3111-3115, 1975 (4) ROZENGURT E, Po CC: Selective cytotoxicity for transformed 3T3 cells. Nature 261:701-702, 1976 (5) SCHIAfFONATI L, BASERGA R: Different survival of normal and transformed cells exposed to nutritional conditions nonpermissive for growth. Cancer Res 37:541-545, 1977 (6) BRADLEY MO, KOHN KW, SHARKEY NA, et al: Differential cytotoxicity between transformed and normal human cells with combinations of aminonucleoside and hydroxyurea. Cancer Res 37:2126-2131, 1977 (7) BEN-ISHAY Z, FARBER E: Protective effects of an inhibitor of protein synthesis, cycloheximide, on bone marrow damage induced by cytosine arabinoside or nitrogen mustard. Lab Invest 33: 478-490, 1975 (8) WARRINGTON R, HECHTMAN R: The .histidine analogue L-histidinol arrests the growth of Balb/3T3 cells in Go. Cell BioI Int Rep 1:571-578, 1977 (9) - - - : Differential responses of normal and transformed mouse cell lines to L-histidinol. Am Soc Microbiol 1:120, 1977
J
NATL CANCER INST
(10) FREEMAN VH, SHIN S: Cellular tumorigenicity in nude mice: Cor-
relation with cell growth in semi-solid medium. Cell 3:355-359, 1974 (11) LAJTHA IG: On the concept of the cell cycle. J Cell Physiol 62:143-145, 1963 (12) PARDEE AB: A restriction point for control of normal animal cell proliferation. Proc Nat! Acad Sci USA 71:1286-1290, 1974 (13) HANSEN BS, VAUGHAN MH, WANG L: Reversible inhibition by histidinol of protein synthesis in human cells at the activation of histidine. J BioI Chern 247:3854-3857, 1972 (14) VAUGHAN M, HANSEN B: Control of initiation of protein synthesis in human cells. J BioI Chern 248:7087-7096, 1973 (15) HAMILTON TA, LITI M: Biosynthesis of mammalian transfer RNA. Evidence for regulation by deacylated transfer RNA. Biochim Biophys Acta 435:362-375, 1976 (16) GRUMMT F, GRUMMT I: Studies on the role of uncharged tRNA in pleiotypic response of animal cells. Eur J Biochem 64:307312. 1976 (17) WARRINGTON RC, WRATIEN N, HECHTMAN R: L-Histidinol inhibits specifically and reversibly protein and ribosomal RNA synthesis in mouse L·cells. J BioI Chern 252:5251-5257, 1977 (18) WARRINGTON RC, WRATIEN N: Regulation of macromolecular synthesis in reovirus-infected L·929 cells: Effect of L-histidinol. J Virol 16:1503-1511, 1975 (19) - - - : Differential action of L-histidinol in reovirus-infected and uninfected L-929 cells. Virology 81:408-418, 1977 (20) HERSHKO A, MAMOT P, SHIELDS R, et al: Pleiotypic response. Nature 232:206-211, 1971 (21) BUTUIER R W, SUTHERLAND EW: Adenosine 3',5'-phosphate in biological materials. I. Purification and properties of cyclic 3', 5'-nucleotide phosphodiesterase and use 'of this enzyme to characterize adenosine 3',5'-phosphate in human urine. J BioI Chern 237:1244-1250. 1962 (22) HsIE AW, PUCK IT: Morphological transformation of Chinese hamster cells by dibutyryl adenosine cyclic 3':5'-monophos· phate and testosterone. Proc Nat! Acad Sci USA 68:358-361, 1971 (23) JOHNSON GS, FRIEDMAN RM, PASTAN I: Restoration of several morphological characteristics of normal fibroblasts in sarcoma cells treated with adenosine-3' ,5' -cyclic monophosphate and its derivatives. Proc Nat! Acad Sci USA 68:425-429, 1971
VOL. 61, NO. I, JULY 1978
Selective KIlling of Oncogenic Human Cells In Vitro 73
FIGURE I.-Selective killing of colonies of HeLa cells cocultivated with human fibroblasts. a) Typical colony of HeLa cells grown from a single cell after 6 days' incubation. b) Low-passage HFF at mid-log phase. c) Cocultivation control. Colonies of HeLa cells were established as in la and subsequently flooded with HFF. HeLa colonies remained discrete (central area). HFF cells grew, establishing islands of HeLa cells surrounded by typical fibroblasts. d-f) Eradication of HeLa colonies. d) HeLa cells (1O~ / lOO-mm dish) were plated and allowed to grow to discrete colonies (6 days). HFF cells (\()6 / dish), harvested from quiescent monolayers, were added to dishes bearing HeLa colonies. When the HFF had attached, I.-histidinol (10 mM) and theophylline (6 mM) were added. Photograph d, taken 24 hr later, reveals the typical quiescent morphology of HFF and the toxic effects of the pretreatment reagents on HeLa cells. Ara-C (10 /lg / ml) and ' HV (2 mM) were added. Photographs e and f. taken 24 and 48 hr after S-phase drugs were added, reveal a progressive and total destruction of HeLa cells. HFF appear to be unaffected. g-i) Viability of HFF cells. Following 2 days' incubation in S-phase drugs, the cultures were washed repeatedly in fresh medium and incubated. Photographs g, h. and i were taken 2, 4, and 7 days later. HFF cells survived and proliferated to establish a monolayer essentially indistinguishable from the monolayers of control HFF. No foci derived from surviving HeLa cells were evident.
VOL. 61. NO. I. JULY 1978
J
NATL CANCER INST
3,4
ABSTRACT -Hlstldlnol and theophylline reversibly arrested the growth of low-passage human foreskin fibroblasts (HFF) and protected these cells from an otherwise lethal combination of the S-phase drugs ,B-cytosine arabinoside and hydroxyurea. In contrast, HeLa cells continued cell cycle transit in the presence of hlstldinol and theophylline. Consequently, these reagents did not alter the vulnerability of He La cells to combined S-phase drugs. The differential S-phase drug sensitivity mediated by histldlnol and theophylline persisted under cocultivatlon conditions, thereby allowing the selective and total eradication of colonies of HeLa cells In the presence of subconfluent HFF cells. The HFF cells, spared S-phase drug toxicity by their unique response to histldlnol and theophylline, showed total survival and vlability.-J Natl Cancer Inst 61: 69-73, 1978.
MATERIALS AND METHODS HFF were obtained from D. L. Engelhardt (Columbia University) and used at less than 20 population doublings. HFF were propagated in minimum essential medium supplemented with 10% fetal bovine serum. Stocks were maintained by passage at low density, whereas cells for experiments were harvested from confluent, quiescent cultures. HeLa cells, obtained from P. I. Marcus (University of Connecticut), were propagated in Dulbecco's modified Eagle's medium supplemented with 5% calf serum. Colonies of HeLa cells were established by plating approximately 103 cells onto 100mm culture dishes. L-Histidinol (Calbiochem) was prepared, and cell numbers and volumes were determined as described earlier (8). Photographs were taken with a Nikon AFM 35-mm photomicrographic attachment on a Nikon inverted MS microscope with the use of phase-contrast optics supplied by P. Goetinck (University of Connecticut). Hydroxyurea (Calbiochem) and Ara-C (Aldrich Chemical Co.) were used at 2 mM and 10 ILg/ml, respectively, either alone or in combination, as described in the figure legends. Histidinol and theophylline (Calbiochem) were used individually or in combination at various concentrations (see figure legends).
A major limitation of currently used cancer chemotherapy is the failure of anticancer agents to differentiate between normal and malignant cells. In general, proliferating cells are more vulnerable to cancer drugs than are their nonproliferating counterparts (1). Consequently, crucial populations of cycling normal cells, such as those of the myelopoietic, gastrointestinal, and epidermal tissues, are at risk during chemotherapy. This suggests that appropriate manipulations of the proliferative status of normal and malignant cell populations could improve the therapeutic index of cell cycledependent cancer drugs. Since these drugs are preferentially toxic for cycling cells, their administration under conditions wherein only neoplastic cells are allowed to proliferate should result in selective cell killing. This prediction has been verified in vitro. Pardee and James (2) and O'Neill (3) in 1975 published procedures that allow selective killing of transformed cells, as have others more recently (4, 5, 6). Unfortunately, the reagents used to date to modify the proliferation of normal and malignant cells possess too great an inherent toxicity for in vivo applicability (4, 7). L-HistidinoI. a structural analog of the essential amino acid histidine, elicits differential growth responses from oncogenic and non oncogenic mouse cell lines. The analog maintains the nontumorigenic Balb/3T3 and 3T3 cells in the Go state [(8); Yen A, Warrington R, Pardee AB: Submitted for publication] but does not restrict the proliferation of a variety of tumorigenic mouse cell lines including L929, SVIOI, SVA31, and 3Tl2 (Warrington R, Hechtman R: Submitted for publication). Thus normal mouse cells are protected from S-phase drugs by the histidinol-mediated Go arrest, whereas transformed mouse cells, which continue cell cycle transit in the presence of the analog, retain their vulnerability to S-phase drugs [(9); Warrington R, Hechtman R: Submitted for publication]. I report here that histidinol and theophylline evoke similar responses from normal (low-passage HFF) and tumorigenic (HeLa) human cells. VOL. 61. NO.1. JULY 1978
RESULTS HFF, chosen as an In vitro prototype of nontransformed human cells, were harvested from confluent cultures and plated into fresh medium (text-fig. IA). The confluent state of cultures prior to harvest provided a population of cells most of which were in the quiescent or Go state. This was indicated by the prolonged delay, typical of Go-recovery, observed before the control cultures divided (text-fig. lA) and was verified by the observation that such populations of cel.ls doubled their cell volume and entered an S-phase pnor to the first division (Warrington R: Unpublished ABBREVIATIONS liSEn: Ara-C=,B-cytosine arabinoside; HFf=human foreskin fibroblast(s); HU=hydroxyurea. Received November l. 1977; accepted february l. 1978. Supported by Public Health Service research grant CAl4733 from the National Cancer Institute. This research benefitted significantly [rom the use of the University of Connecticut Cell Culture facility, also supported by this grant. 3 Microbiology Section. University of Connecticut. Storrs, Conn. I
2
06268. 4 Address reprint requests to Dr. Warrington at his present address:
Department of Cancer Research. University of Saskatchewan. Saskatoon. Saskatchewan S7N OWO. Canada.
69
J
NATL CANCER INST
70 Warrington data). Unlike nontransformed mouse cells quiescent HFF cells were not maintained in Go for extended periods by the addition of L-histidinol, although the initiation of division and the apparent doubling time were influenced in a dose-dependent manner. Theophylline had a similar effect, but neither histidinol nor theophylline alone gave significant protection from S-phase drugs. However, a persistent and reversible growth arrest was observed (text-fig. IB) when these reagents were combined. Furthermore, histidinol and theophylline offered near total protection from the Sphase drugs Ara-C and HU. No such protection was offered by these reagents alone. Detailed studies with Balb/3T3 cells, harvested from quiescent monolayers and subsequently treated with histidinol, established that the analog maintained these cells in Go (8). The similarities of the responses, recovery kinetics, and
1A
1B CONTROL CONTROL
prolonged protection from S-phase drugs shown by HFF treated with histidinol and theophylline suggest that these cells are also maintained in the Go state. HeLa cells, known to be oncogenic by the nude mouse assay (10), responded differently to histidinol and theophylline (text-fig. 2). Histidinol inhibited the growth of HeLa cells, and the appearance of dead cells in the medium of treated cultures indicated that it was toxic at all concentrations studied. Theophylline altered the morphology and reduced the growth rate of HeLa cells but was not obviously toxic. However, the combination of histidinol and theophylline was more toxic than histidinol alone. Furthermore, the combination altered the susceptibility of HeLa cells to S-phase drugs, since it reduced significantly the apparent killing rate of HU but enhanced slightly the killing rate of Ara-C and HU in combination. The inability of histidinol and theophylline to protect HeLa cells from Ara-C and HU indicated that these cells continued cell cycle transit in the presence of these reagents. The differential responses of HFF and HeLa cells to histidinol and theophylline suggested that HeLa cells could be eradicated selectively in the presence of HFF by appropriate pretreatments. The subline of HeLa cells
W III
:::IE ::I
Z
...J ...J W
;r-
T 105i""~~-(H+T)
105
()
H +T +(AraC+HU)
...J
i!
let
~ rr-
T
Z
c:
i:
m
:JI
II:
w
:::IE ::I
Z
CONTROL+(AraC+HU)
10'~~
= Histidinol = Theophylline =.Hydroxyurea = Cytosine arabinoside
__L - - L______~L--L~__J-~__L -______-L10'
1234
NATL CANCER INST
....w....
~
0
H+T
~
r5
10"
10 ~ rr-
z c
AraC
(.)
....
i:
j!
III
m jJ
~ HU
10'
10'
= =
Histidinol H T Theophylline HU = Hydroxyurea AraC Cytosine arabinoside
123456 DAYS
TEXT·FIGURE I.-Response of human foreskin fibroblasts to L-histidinol. theophylline. and S-phase drugs. Tissue culture dishes (60 mm) were seeded with approximately 105 HFF cells harvested from quiescent monolayers. Cell numbers were determined in duplicate at 24-hr intervals. Ara-C and HU were added at 10 #,g/ml and 2 mM. respectively. as indicated for the duration of the experiment. A: Cell numbers are plotted for untreated control cultures (0) and cultures receiving 5 (0) and 10 (e) mM histidinol. 2 mM theophylline (t.). or histidinol and theophylline (8) (5 and 2 mM. respectively) shortly after plating. Ara-C and HU were added on day 2 to cultures pretreated with 2 mM theophylline (+). B: Cell numbers are plotted for untreated control cultures (0). for control cultures that received Ara-C and HU on day 2 (e). and for cultures treated (after plating) with histidinol and theophylline (5 and 2 mM. respectively) with (II) or without (t.) the addition of Ara-C and HU on day 2. On day 3. many of the plates of the set treated with histidinol and theophylline were washed several times and incubated in fresh medium (A). ]
106
H
III
H T HU AraC
CONTROL
CONTROL
III
T+(AraC+HU)
~
28
2A
a
H(2x)
II:
107
107
H+T (Reverse)
=
1~ 1
2
3
4
5
2
6
3
4
5
103
DAYS
TEXT·FlGURE 2.-Response of HeLa cells tOL-histidinol. theophylline. and S-phase drugs. Cell numbers were determined and Ara-C and HU were used as noted in the legend of text-fig. I. A: Cell numbers are plotted for control cultures (0) and cultures receiving 3 mM histidinol (e).2 mM theophylline (t.). or histidinol and theophylline (8. 3 and 2 mM. respectively) shortly after plating. HU (+). Ara-C (0). or HU and Ara-C (~) were added on day 2 to cultures pretreated with histidinol and theophylline (~). B: Cell numbers are plotted for control cultures (0) and cultures receiving Ara-C (0). HU (+). or Ara-C and HU (.) on day 2. VOL. 61. NO.1. JULY 1978
Selective KIlling of Oncogenic Human CeUs In Vitro 71
used in this study could form discrete and uniform colonies when grown from single Eells (fig. la). This property allows clear distinctions between HFF cells (fig. lb) and HeLa cells in mixed cultures prepared by adding HFF cells to plates bearing established HeLa colonies. These cell types did not intergrow freely under these cocultivation conditions, and, consequently, the HeLa colonies remained visible as islands of morphologically distinct cells surrounded by fibroblasts (fig. le).
Mixed cultures were prepared and, after the HFF cells attached, were treated with histidinol and theophylline. One day later (fig. Id) the HFF cells assumed quiescent morphology, whereas mitotic activity and toxicity effects were evident in HeLa colonies. Ara-C and HU were added, and a progressive and total destruction of HeLa colonies was observed over the subsequent 48 hours (figs. Ie, f). Drugs and debris were then removed and, after several washings with fresh medium, the cultures were incubated to assess the status of the HFF survivors. These grew and established typical fibroblastic monolayers (figs. Ig-i). The monolayers were stained and no foci of HeLa cells (visible in mixed cultures not treated with S-phase drugs) were observed. The survival of oncogenic cells was significant if histidinol and theophylline were not added prior to Sphase drug administration. Further, large numbers of HFF cells were killed and those surviving were megaloblastic and incapable of proliferating when the pretreatment was omitted. Therefore, histidinol and theophylline are required not only for the survival of HFF cells but also for the successful eradication of the tumorigenic component of this cocultivation system. These experiments indicate that the differential growth responses of HFF and HeLa cells to histidinol and theophylline are maintained under conditions of cocultivation and demonstrate that appropriate manipulation of the proliferative status of normal and tumorigenic cells can enhance the selectivity of cell cycledependent anticancer drugs. DISCUSSION
Cultured eukaryotic cells possessing an untransformed phenotype enter a viable state of quiescence termed Go (11) when subjected to a multiplicity of conditions that are not optimal for proliferation. In contrast, transformed cells continue cell cycle transit and often show substantial cell death under similar restrictive conditions. This difference between normal and transformed cells in vitro, documented and discussed by many workers, is lucidly formalized in Pardee's "restriction point" control hypothesis (12). Thus the distinct responses of normal and transformed cells to essential amino acid deprivation provide a partial rationale for the observations reported here. LHistidinol, a competitive inhibitor of histidyl-tRNA synthetase (13), establishes a persistent functional deprivation for the essential amino acid L-histidine in a variety of cultured eukaryotic cells (13-16). This observation, considered with the ability of L-histidinol not VOL. 61, NO. I. JULY 1978
only to inhibit specific aspects of macromolecular synthesis in uninfected (17) and virus-infected (18, 19) mouse cells but also to evoke physiologic alterations (16) characteristic of the negative pleiotypic response (20), suggested that the analog would arrest normal mouse cells in the Go-state but would not arrest the growth of transformed cells. As indicated in the introduction, this prediction was verified for a variety of mouse cell lines. In view of the foregoing, we were surprised that no concentration of histidinol alone provided a persistent growth arrest for HFF cells. Similarly, HeLa cells showed an unanticipated resistance to S-phase drugs in the presence of histidinol. Thus HFF and HeLa cells do not respond in the same manner to this analog as do their mouse counterparts. Whether these differences reflect a peculiarity limited to these particular human cell types or indicate a general distinction between cultured mouse and human cells is uncertain. However, theophylline, an inhibitor of cAMP phosphodiesterase (21), sufficiently augmented the differential effects of histidinol on HFF and HeLa cells to allow selective killing. Although the reason for these apparently synergistic interactions between histidinol and theophylline is unknown, the effect appears to be general. As reported here, theophylline prolongs the duration of the histidinol-mediated Go arrest in normal mouse cell lines' but enhances the sensitivity of histidinol-treated transformed mouse cell lines to certain S-phase drugs (Warrington R, Hechtman R: Submitted for publication). These observations were derived serendipitously from experiments designed to determine whether a theophylline-mediated elevation in intracellular cAMP would counteract the effects of histidinol on transformed cells (Warrington R: Unpublished observations). Such an effect is an a priori deduction from the demonstration (22, 23) that increases in intracellular cAMP cause certain transformed cells to undergo a morphologic reversion to a habitus approaching that of normal cells (reverse transformation). Although theophylline does alter the morphology of transformed mouse and human cells, this reagent does not diminish the propensity of transformed cells to continue cell cycle transit in the presence of histidinol. On the contrary, theophylline increased the toxicity of histidinol in all transformed cells tested. Rationalizations for the various effects of the histidinol-theophylline combination cited herein await further investigation. Experiments similar to those reported here have been performed with an HFF-SVIOI (simian virus 40-transformed 3T3 cell) cocultivation system. In contrast to the HFF and He La cell combination used in this study, HFF and SVIOI cells intergrow freely under cocultivation conditions. Nevertheless, the administration of Sphase drugs, after appropriate pretreatment with histidinol and theophylline, allowed total eradication of SVIOI cells and total (apparent) survival and viability of the HFF cells (Warrington R: In preparation). Thus the selective killing of transformed cells reported in the present study appears to depend on specific alterations of the proliferative status of the cell types used rather
J
NATL CANCER INST
72 Warrington than a result of any peculiar interaction between them. The ability of histidinol and theophylline to restrict the cytoxicity of S-phase drugs to the oncogenic component of in vitro cocultivation systems suggests a potential applicability in cancer chemotherapy. The possibility that such an approach can be exploited to enhance the specificity of cell cycle-dependent anticancer drugs in vivo is under investigation.
REFERENCES (1) VALERIOTE F, VAN PUTIEN L: Proliferation-dependent cytotoxicity of anticancer agents: A review. Cancer Res 35:2619-2630, 1975 (2) PARDEE AB, JAMES LJ: Selective killing of transformed baby hamster kidney (BHK) cells. Proc Natl Acad Sci USA 72:49944998, 1975 (3) O'NEILL FJ: Selective destruction of cultured tumor cells with uncontrolled nuclear division by cytochalasin B and cytosine arabinoside. Cancer Res 35:3111-3115, 1975 (4) ROZENGURT E, Po CC: Selective cytotoxicity for transformed 3T3 cells. Nature 261:701-702, 1976 (5) SCHIAfFONATI L, BASERGA R: Different survival of normal and transformed cells exposed to nutritional conditions nonpermissive for growth. Cancer Res 37:541-545, 1977 (6) BRADLEY MO, KOHN KW, SHARKEY NA, et al: Differential cytotoxicity between transformed and normal human cells with combinations of aminonucleoside and hydroxyurea. Cancer Res 37:2126-2131, 1977 (7) BEN-ISHAY Z, FARBER E: Protective effects of an inhibitor of protein synthesis, cycloheximide, on bone marrow damage induced by cytosine arabinoside or nitrogen mustard. Lab Invest 33: 478-490, 1975 (8) WARRINGTON R, HECHTMAN R: The .histidine analogue L-histidinol arrests the growth of Balb/3T3 cells in Go. Cell BioI Int Rep 1:571-578, 1977 (9) - - - : Differential responses of normal and transformed mouse cell lines to L-histidinol. Am Soc Microbiol 1:120, 1977
J
NATL CANCER INST
(10) FREEMAN VH, SHIN S: Cellular tumorigenicity in nude mice: Cor-
relation with cell growth in semi-solid medium. Cell 3:355-359, 1974 (11) LAJTHA IG: On the concept of the cell cycle. J Cell Physiol 62:143-145, 1963 (12) PARDEE AB: A restriction point for control of normal animal cell proliferation. Proc Nat! Acad Sci USA 71:1286-1290, 1974 (13) HANSEN BS, VAUGHAN MH, WANG L: Reversible inhibition by histidinol of protein synthesis in human cells at the activation of histidine. J BioI Chern 247:3854-3857, 1972 (14) VAUGHAN M, HANSEN B: Control of initiation of protein synthesis in human cells. J BioI Chern 248:7087-7096, 1973 (15) HAMILTON TA, LITI M: Biosynthesis of mammalian transfer RNA. Evidence for regulation by deacylated transfer RNA. Biochim Biophys Acta 435:362-375, 1976 (16) GRUMMT F, GRUMMT I: Studies on the role of uncharged tRNA in pleiotypic response of animal cells. Eur J Biochem 64:307312. 1976 (17) WARRINGTON RC, WRATIEN N, HECHTMAN R: L-Histidinol inhibits specifically and reversibly protein and ribosomal RNA synthesis in mouse L·cells. J BioI Chern 252:5251-5257, 1977 (18) WARRINGTON RC, WRATIEN N: Regulation of macromolecular synthesis in reovirus-infected L·929 cells: Effect of L-histidinol. J Virol 16:1503-1511, 1975 (19) - - - : Differential action of L-histidinol in reovirus-infected and uninfected L-929 cells. Virology 81:408-418, 1977 (20) HERSHKO A, MAMOT P, SHIELDS R, et al: Pleiotypic response. Nature 232:206-211, 1971 (21) BUTUIER R W, SUTHERLAND EW: Adenosine 3',5'-phosphate in biological materials. I. Purification and properties of cyclic 3', 5'-nucleotide phosphodiesterase and use 'of this enzyme to characterize adenosine 3',5'-phosphate in human urine. J BioI Chern 237:1244-1250. 1962 (22) HsIE AW, PUCK IT: Morphological transformation of Chinese hamster cells by dibutyryl adenosine cyclic 3':5'-monophos· phate and testosterone. Proc Nat! Acad Sci USA 68:358-361, 1971 (23) JOHNSON GS, FRIEDMAN RM, PASTAN I: Restoration of several morphological characteristics of normal fibroblasts in sarcoma cells treated with adenosine-3' ,5' -cyclic monophosphate and its derivatives. Proc Nat! Acad Sci USA 68:425-429, 1971
VOL. 61, NO. I, JULY 1978
Selective KIlling of Oncogenic Human Cells In Vitro 73
FIGURE I.-Selective killing of colonies of HeLa cells cocultivated with human fibroblasts. a) Typical colony of HeLa cells grown from a single cell after 6 days' incubation. b) Low-passage HFF at mid-log phase. c) Cocultivation control. Colonies of HeLa cells were established as in la and subsequently flooded with HFF. HeLa colonies remained discrete (central area). HFF cells grew, establishing islands of HeLa cells surrounded by typical fibroblasts. d-f) Eradication of HeLa colonies. d) HeLa cells (1O~ / lOO-mm dish) were plated and allowed to grow to discrete colonies (6 days). HFF cells (\()6 / dish), harvested from quiescent monolayers, were added to dishes bearing HeLa colonies. When the HFF had attached, I.-histidinol (10 mM) and theophylline (6 mM) were added. Photograph d, taken 24 hr later, reveals the typical quiescent morphology of HFF and the toxic effects of the pretreatment reagents on HeLa cells. Ara-C (10 /lg / ml) and ' HV (2 mM) were added. Photographs e and f. taken 24 and 48 hr after S-phase drugs were added, reveal a progressive and total destruction of HeLa cells. HFF appear to be unaffected. g-i) Viability of HFF cells. Following 2 days' incubation in S-phase drugs, the cultures were washed repeatedly in fresh medium and incubated. Photographs g, h. and i were taken 2, 4, and 7 days later. HFF cells survived and proliferated to establish a monolayer essentially indistinguishable from the monolayers of control HFF. No foci derived from surviving HeLa cells were evident.
VOL. 61. NO. I. JULY 1978
J
NATL CANCER INST
