Int. J. Cancer: 52,797-801 (1992) Cs 1992 Wiley-Liss, Inc.
(m '
Publication of the lniernational Union Against Cancer Publication de I'Union lniernationaleConire le Cancer
*e5
EARLY SUPEROXIDE DISMUTASE ALTERATIONS DURING SV40-TRANSFORMATION OF HUMAN FIBROBLASTS A. BRAVARD',F. HOFFSCHIR',L. SABATIER', M. RICOUL',A. PINTON',R. CASSINGENA~, S. ESTRADE~, C. LUCCIONI' and
B. D U T R I L L A U X ' . ~ ~ ~ 'CEAIDSVIDPTEILCG, B P 6 92265 Fonteriay aux Roses Cedex; W P R 42, IRSC, BP 8, 94802 Villejuf Cedex and 3URA 620, CNRS, Institiit Curie, Section Biologie, 26, nce d'Ulm, 75231 Paris Cedex 05, France. The expression of superoxide dismutases (SOD) I and 2 was studied in 4 clones of human fibroblasts after their infection by simian virus 40 (SV40). in parallel with the alterations of chromosomes 21 and chromosome 6q arms, carrying the genes that encode for SOD1 and SOD2 respectively. For all clones, a similar scheme with 2 main phases was observed for both chromosome and SOD variations. The first phase, defined as the pre-crisis phase, was characterizedby chromosomal instability, but maintenance of normal numbers of chromosome 6q arms and chromosomes 2 I. The level of SOD2 mRNA was high, while SODZ activity and immunoreactive protein were low. SOD I protein and activity were decreased. In the second phase, defined as the post-crisis phase, the accumulation of clonal chromosomal rearrangements led to the loss of 6q arms, while the number of chromosomes 21 remained normal. SOD2 mRNA level was decreased and SOD2 immunoreactive protein and activity remained low. SOD I protein and activity increased with passages, reaching values similar to those of control cells at late passages. As in established SV40-transformed human fibroblast cell lines, good correlation was found between SOD2 activity and the relative number of 6q arms. These results allow us to reconstruct the sequence of events leadingto the decrease of SODZ, a possible tumor-suppressor gene, during the process of SV40-transformation of human fibroblasts.
o 1992 Wiley-Liss, Inc.
The frequent deletions occurring in human cancers suggest the existence of tumor-suppressor genes. The deletion of the long arm of chromosome 6 (6q) is frequently observed in a number of malignancies such as lymphomas, melanomas, glioblastomas, and breast and ovarian carcinomas, among others (Sandberg, 1990), as well as in SV40 (simian virus 40)-transformed human cells (Hoffschir et al., 1988). This chromosome segment carries the gene encoding for the enzyme superoxide dismutase 2 (SOD2; E.C.1.15.1.1) located on band 6q21 (HGM 10,1989), which catalyzes the dismutation of supcroxide radicals. It has been repeatedly proposed that free radicals of oxygen are implicated in the initiation/promotion phases of carcinogenesis as well as in cellular differentiation and proliferation process (for review, see Oberley and OberIcy, 1988; Sun, 1990; Allen, 1991; Cerutti and Trump, 1991). A decrease of SOD2 expression has been observed in a great number of tumors (Galeotti et al., 1989; Sun, 1990; Kwee et al., 1991) and in SV4O-transformed human fibroblasts (Yamanaka and Deamer, 1974; Marklund et al., 1982; Marlhens et al., 1985; Oberley et af., 1989). In our model, we reported the decrease of SOD2 activity as well correlated with the loss of chromosome 6q (Bravard et af., 1992). These data led us to suggest that SOD2 might be a new type of tumor-suppressor gene. In the prcsent study, we tried to reconstruct the sequence of events resulting in chromosome 6q arm losses and decrease of SOD2 activity by studying these parameters during the early steps of human fibroblast transformation by SV40. SOD2 activity and loss of the 6q arm were also compared with SODl activity and chromosome 21 on which this enzyme is mapped (HGM10, 1989). MATERIAL AND METHODS
Cells Four clones of human fibroblasts were analyzed at different passages (ranging from 9 to 112) after infection by SV40. Two
clones, CHSV3 and CHSV4, were derived from cornea and 2, DHSV2 and DHSV4, from dermis. The corresponding normal non-infected cells in culture were used as controls. They were primary fibroblast cultures which became senescent around passage 20. Cells were grown in DMEM-F12 medium with 10% FCS.
Cytogenetic analysis Metaphases were spread and R-banded according to our usual methods (Dutrillaux and Couturier, 1981). For each passage, at least 10 karyotypes were established. After tentative identification of the rearranged chromosomes, we tried to determine the location of the breakpoints and counted the number of copies of each chromosome or chromosome segment. This allowed us to estimate the average number of copies of each chromosome per passage and in particular to count the relative number of 6q arms: the average number of chromosome 6q was divided by the average number of chromosomes and multiplied by 23 (normal haploid number), the balanced status being 1. Identical analysis was done for chromosome 21. SOD activities and proteins Cells were collected by trypsinization in the exponential growing phase, 72 hr after feeding, washed twice in PBS by centrifugation and stored in liquid nitrogen until extraction. Extraction was performed by successive freezing and thawing in 10 mM Tris HCI, p H 7.5, 0.1% Triton X-100 and 200 mM sucrose. Enzymatic and Western-blotting assays were performed on 20,000 g supernatant. S O D assay was adapted from the method of Beauchamp and Fridovich (1971). It was carried out at 30°C in 1 ml of 20 mM sodium carbonate buffer, p H 10.0, containing 0.1 mM DTPA, 0.2 mM xanthine, 12 FM nitro blue tetrazolium (NBT) and 1.95 munits xanthine oxidase. SOD2 activity was determined as the remaining SOD activity after addition of 5 mM KCN. The reaction was followed at 560 nm with a spectrophotometer. One unit of SOD activity was defined as the amount of enzyme inhibiting the rate of NBT reduction by 50%. Protein content was determined using a Biorad kit with albumin as standard. Antibodies against E. coli MnSOD (SOD2) and human CuZnSOD (SOD1) (Sigma, St Louis, MO) were raised in rabbit by serial injections of 1 mg purified protein. Total sera were used for immunoblotting. After pre-treatment with SDS and P-mercaptoethanol at 100°C for 4 min, electrophoresis of 10 to 20 pg of homogenates were performed in a 4% stacking, 15% separating polyacrylamide gel in 25 mM Tris, 192 mM glycine, 0.1% SDS buffer, p H 8.3. After electrotransfer of proteins, nitrocellulose sheets were saturated with dried milk and incubated for 1 hr with diluted anti-SOD antibodies. The specific immunoreactive S O D l or SOD2 band was revealed using an Amersham blotting-detection kit for rabbit bound ?To whom correspondence and reprint requests should be addressed. at the Institut Curie.
Received: April 30,1992 and in revised form June 27,1992.
798
B R A V A R D ET AL.
antibodies (RPN.23). For quantification, the density of each band was measured with a laser densitometer 2202 Ultrascan LKB (Bromma, Sweden).
Northern blot The technique and references have been described in Bravard et al. (1992). Briefly, total cellular RNA was prepared either from cell pellets after trypsinization or directly from cells in culture flasks. The purified R N A was quantified spectrophotometrically; 20 p,g per lane were size-fractionated by formaldehyde agarose (1.5%) gel electrophoresis and transferred to nylon. The filters were stained with methylene blue to confirm R N A integrity. For detection of R N A sequences, a random-primed cDNA SOD2 probe was utilized. Random priming was performed using a BRL kit (Paris, France). The levels of 4-kb and 1-kb SOD2 mRNA were estimated by comparison with methylene blue staining. RESULTS
Cytogenetic analysis The cytogenetic evolution of the 4 clones described by Hoffschir ef al. (1992), may be summarized as follows. Clone CHSV3: progressive decrease of chromosome number to about 35 (passage 36), followed by endoreduplication (between passages 36 and 48) and further chromosome losses. No deletion of 6q arm nor loss of chromosomes 21 were observed, all variations being related to the changes of ploidy of the metaphases. Clone CHSV4: progressive decrease of chromosome number to about 37 (passage 44), without polyploidization. One chromosome 6 was missing after passage 21. Chromosomes 21 remained unchanged. Clone DHSV2: progressive decrease of chromosome number to about 39 (passage 18), followed by endoreduplication and further chromosome losses. A deletion of 6q arm occured between passages 18 and 32. Chromosomes 21, not lost, increased in relative number. Clone DHSV4: progressive decrease of chromosome number to about 42 (passage 44). The occurrence of various rearrangements of chromosome 6 resulted in its frequent loss at late passages. Chromosomes 21 remained stable. The relative numbers of chromosome 6q arms and of chromosomes 21 at the passages studied for enzyme activities are given in Table I. For all clones but CHSV3, a single crisis, characterized by high cell lethality and chromosome instability, occurred around passage 20. The surviving cells had acquired several clonal chromosome rearrangements. For CHSV3, a second marked crisis occurred around passage 30, and most clonal chromosome rearrangements were acquired afterward. SOD activities and proteins
SOD activities, measured at 4 passages at least for each clone, were compared with the activity in control cells. Except for CHSV3. SOD2 activity was always decreased after the first passage studied and remained low at later passages (Table I, Fig. 1). In clone CHSV3, SOD2 activity was increased at the
first 2 passages studied (passages 12 and 29), and decreased in later passages (Table I, Fig. 1). For all clones, S O D l activity was decreased at the first passage studied, but increased in later passages, frequently reaching higher values than those of control cells (Table I). This resulted in a strong decrease of the SOD2/SOD1 ratio with passages. SOD2 and SOD1 immunoreactive proteins were quantified for one passage at least of each clone and for control fibroblasts. In all cases, the amount of immunoreactive protein was correlated with the corresponding enzyme activity (Figs. 2, 3). SOD2 mRNA The I-kb and 4-kb bands, characteristic for SOD2 mRNA were observed at various intensities depending on passages (Figs. 1,4). The 1-kb band, which is assumed to give rise to the active SOD2 protein, followed a similar evolution in all clones: increased at early passages (p12-p29 for CHSV3, p14 for CHSV4, p17-p20 for DHSV2 and p9-p32 for DHSV4) and decreased at later passages. For the 4-kb band, a similar evolution was observed with variations of a lower magnitude. Correlations between cytogenetic, eniymologic and mRNA data When the 4 clones were considered, SOD2 activities were significantly correlated with the relative numbers of 6q arms (Fig. 5 ) . In spite of this good correlation, the chronology of 6q arm loss and decrease of SOD2 activity was not exactly superimposable, since the decrease of activity always preceded chromosome loss. The relationship between the amount of SOD2 mRNA and the number of 6q arms was less simple: the early increase of mRNA was not accompanied by a change in the number of 6q arms, whereas the decrease of mRNA paralleled the loss of 6q arms at late passages. The difference between the amount of SOD2 mRNA and SOD2 activity was still more marked at early passages, since SOD2 activity was low, whereas the amount of mRNA was high. At late passages, both the enzyme activity and the amount of mRNA were decreased. No correlation was found between the variations of S O D l activity and the relative numbers of chromosome 21 at early passages. A t late passages, the 2 parameters were similar or higher than in control cells. DISCUSSION
We have recently shown that, in SV40-transformed human fibroblast cell lines, a correlation exists between low SOD2 activity and the loss of the chromosome 6q arm where the gene SOD2 is mapped. This low enzyme activity was also related to low mRNA content. As these particularities may be related to a growth advantage of transformed cells, we have proposed that SOD2 gene could correspond to a new type of tumorsuppressor gene (Bravard et al., 1992). The study was performed on established cell lines which prevented us from reconstructing the sequence of the events leading to SOD2
TABLE 1 - SOD2 AND SODl ACTIVITIES IN UiMG PROTEIN AND RELATIVE NUMBERS OF CHROMOSOME 6q ARMS AND FHROMOSOMES 21 AT VARIOUS PASSAGES OF CLONES CHSV AND DHSV AND IN THEIR RESPECTIVE CONTROL CELLS ~~~
Control
pl?
p29
SOD2 activity 31 40 55.5 Relative number 1 1.01 1.16 of 6q 118.2 58.4 94.5 SODl activity Relative number 1 1.12 0.85 of 21 SODZ/SODl 0.26 0.68 0.59 activity 'p = passage; nd = not done.
CHSV3 p53 p65
CHSV4 pl00
p14
p28
p50
p63
pll2
Control
DHSV2 pY
p34
p44
DHSV4 ps3
pY
p32
p44
p53
22.4 24.8 nd 16.6 9.0 12.3 7.2 3.6 100.5 28.9 14.5 15.5 15.1 45.4 25.6 18.0 5.8 1.09 0.89 0.78 0.90 0.60 0.60 0.50 0.62 1 1.05 0.62 0.70 0.48 1 0.85 0.75 0.61 86.7 127.4 nd 58.4 122.7 137.9 121.2 92.9 83.9 77.7 91.7 135.1 119.5 41.5 53.8 61.5 78.1 0.85 1.25 1.22 1.07 1.07 0.96 0.99 1.14 1 1.01 1.34 1.45 0.78 1 1.01 0.90 0.94 0.26 0.19 nd 0.28 0.07 0.09 0.06 0.04
1.20 0.37 0.16 0.11 0.13 1.10 0.48 0.29 0.07
799
SOD ALTERATIONS IN SV40-INFECTED FIBROBLASTS
DHSV2 120
-
7
mRNA
100
CHSV3 6o 50
1
+
-
+
mRNA
80 60
40
40
30
20
20 0
C
10
p9
p34
p44
p53
assages
n C
p12
p29
p53
p65
Passages
DHSV4 120 100
1
+
+
-
-
p9
p32
p44
p53
mRNA
80 40 60
30 40
20
20
10
0
0 C
p14
p28
p50
p63
p112
C
passages
FIGURE 1 - SOD2 activities in Uimg protein (black bars) and relative numbers of 6q arms (hatched bars; multiplied by 50 for clones CHSV and by 100 for clones DHSV to use the same scale as that of SOD2 activity) at passages of the different clones and in their respective control cells (C). Quantitative variations of SOD2 mRNA are indicated by + or - for increase or decrease in comparison with
control cells.
disregulation. This is why we investigated, on recently transfected cells, the variations of chromosome number, of the amounts of SOD mRNA and protein, and of enzyme activity. The study enables us to propose a scheme with 2 major phases. For all clones except CHSV3, a first phase occurred before the crisis, which took place around passage 20. This phase was characterized by high chromosome instability, shown by the occurrence of jumping and clonal translocations (Hoffschir et al., 1992), high SOD2-mRNA content, a small amount of SOD2 protein, and low SOD2 activity. Chromosome instability may be regarded as a direct consequence of high mutagenesis also resulting in gene mutations. This might affect at least one allele of SOD2 gene which was transcribed but not efficiently translated, or result in increased protein degradation, as suggested by the high mRNA and low amounts of protein. A second phase took place after the crisis. All the cells of a given clone, having acquired several chromosomal rearrangements in common, were obviously of a monoclonal origin (Hoffschir et al.. 1992). Some of these rearrangements led to the deletion of a 6q arm and thus to hemizygozity for SOD2 gene. This might result in the loss of the mutated rather than of the normal allele, as suggested by the decrease of SOD2 mRNA, whereas SOD2 activity and protein remain at a level that is low, but comparable with that of the pre-crisis period. At this stage, SOD2 status became similar to that described in established SV40-transformed human fibroblast cell lines (Yamanaka and Deamer, 1974; Marklund et al., 1982; Marlhens ef al., 1985; Oberley ef al., 1989; Bravard et al., 1992). For the
CHSV3 clone, the same events occurred but with a delay apparently related to the occurrence of 2 crises: one around passage 20, and another around passage 30. The alterations of SOD2 characterizing the second phase were observed only after the second crisis. Modifications affecting S O D l protein and activity were similar to those of SOD2 during the early passages, with low protein content and low enzyme activity. However, the postcrisis phase was obviously different, with an increase both of protein and of activity reaching levels similar to those of control cells. As far as we could correctly estimate it, the number of copies of chromosome 21 was not changed. These results on S O D l in the post-crisis period are similar to those obtained in established SV40-transformed human fibroblast cell lines (Yamanaka and Deamer, 1974; Marklund et al., 1982; Bravard et al., 1992). Finally, the decrease of SOD2 activity, which is frequently observed in tumors (Galeotti et al., 1989; Sun, 1990; Kwee ef al., 1991), appears to be an early event in the process of SV40-transformation of human fibroblasts. Previous data suggested that variations OP SOD2 activity are related to cell proliferation and differentiation processes, low SOD2 activity being generally observed in proliferative cells and increased SOD2 activity in differentiating cells (Sohal et al., 1986; Oberley and Oberley, 1988; Allen, 1991). Antibodies directed against SOD were able to promote the proliferation of human fibroblasts (Michiels et al., 1988), and SOD-like compounds or
800
ERAVARD E T A L .
I2O I00
a
1
80 60 40
20 0
200
b 150
I00
50
0
FIGURE2 - SOD activities (Uimg protein; black bars) and amounts of SOD immunoreactive proteins (arbitrary units; hatched bars) at passages of clones during transformation and in their respective control cells (C). (a) SOD2; (h) SOD1.
FIGURE 3 -Western blots for SOD proteins. (a) SOD2. CHSV3p29; 2, CHSV3p53; 3, CHSV control: 4, CHSV4p14; DHSV control; 6, DHSV2p9; 7, DHSV4p9. (b) SOD1. CHSV3p12; 2, CHSV3p53; 3, CHSV control; 4, CHSV4p14; CHSV4p50; 6, DHSV control; 7, DHSV2p53; 8, DHSV4p9.
1, 5, 1, 5,
liposornic SOD were successfully used to reduce the growth and increase the differentiation of tumor cells (Beckman et aL, 1989; Allen, 1991). Our results also suggest that the decrease of SOD2 correlates with cell proliferation and immortalization as suggested by Oberley and Oberley (1988). These character-
FIGURE 4 - Northern blot for SOD2 ( I - and 4-kb bands; only the 1-kb mRNA is thought to give rise to active SOD2) from normal DHS fibroblasts (C) and various passages of DHSV2 and DHSV4 clones. The comparison of the intensity of the 1-kb band with that of methylene blue (not shown) was performed to give the indications (+ or -) of Figure 1.
istics being shared with tumor-suppressor genes, we propose that the SOD2 gene has a tumor-suppressor function. In view of this hypothesis, it is interesting to compare SOD2 with the 2 major suppressor genes, RB1 and P53, whose mechanisms of action are being elucidated (Geiser and Stanbridge, 1989; Bookstein and Lee, 1991; Boylan and Zarbl, 1991). Complete inactivation of the RB1 gene seems to be required for the development of retinoblastoma, while in other tumors, such as
SOD ALTERATIONS IN SV40-INFECTED FIBROBLASTS
60 h
.-> .-
I
m
-
50
N
30 -
v,
20-
0 0
0
40 -
0
rz0.83 0 0 .
10
-i
'.,
0
0,4
0
0,6 0.8 1,o Relative number of 6q
e 1,2
FIGURE 5 - Correlation between SOD2 activities (U/mg protein) and the relative numbers of 6q arms (data from Table I) in passages of CHSV (black circles) and DHSV clones (open circles).
801
breast, lung, prostate and bladder carcinomas, RB alterations leading t o loss of the normal protein or to its mutation appear as frequent non-random events. T h e P53 gene also undergoes frequent alterations in human tumors. Except for the LiFraumeni syndrome, most of these alterations occur during tumor progression, leading t o the formation of a mutated P53 protein. Deletions of the 17p arm, where the P53 gene is mapped, may also occur, suppressing the normal allele. SOD2 disregulation may correspond to a different mechanism; since the activity of the protein is never totally lost, its decrease seems related t o a gene dosage effect, a t least after the crisis. T h e tumor-suppressor activity of RBI and P53 protein may be related to transcriptional control of genes involved in cell proliferation through their DNA-binding properties. For SOD2, a hypothesis based on the known biological effects of free radicals might explain its anti-oncogenic properties: the decrease of SOD2 activity may induce a pro-oxidant state, a condition frequently related to the initiation and promotion phases of carcinogenesis (Oberley and Oberley, 1988; Sun, 1990; Cerutti and Trump, 1991).
REFERENCES
GENEMAPPING 10: Tenth International Workshop on Human ALLEN,R.G., Oxygen-reactive species and antioxidant responses HUMAN during development: the metabolic paradox of cellular differenciation. Gene Mapping. Cytogenet. Cell Genet., 51 (1989). Proc. Soc. exp. Biol. Med., 196,117-129 (1991). O.R., Lowered superoxide KWEE,J.K., MITIDlERI, E. and AFFONSO, I., Superoxide dismutase: improved dismutase in highly metastatic B16 melanoma cells. Cancer Lett., 57, BEAUCHAMP, C. and FRIDOVICH, assay and an assay applicable to acrylamide gels. Anal. Biochem., 44, 199-202 (1991). 276-287 (1971). S.L., WESTMAN, N.G., LUNDGREN, E. and Roos, G., MARKLUND, BECKMAN, B.S., BALIN,A.K. and ALLEN,R.G., Superoxide dismutase Copper-and-zinc-containing superoxide dismutase, catalase, and glutainduces differentiation of friend erythroleukemia cells. J. CeN Physiol., thione peroxidase in normal and neoplastic human cell lines and 139,370-376 (1989). normal human tissues. Cancer Res., 42,1955-1961 (1982). BOOKSTEIN, R. and LEE,W., Molecular genetics of the retinoblastoma MARLHENS, F., NICOLE,A. and SINET,P.M., Lowered level of translatsuppressor gene. Crit. Rev. Oncogenesis, 2,211-227 (1991). able messenger RNAs for manganese superoxide dismutase in human BOYLAN, M.O. and ZARBL,H., Transformation effector and suppres- fibroblasts transformed by SV40. Biochem. Biophys. Res. Commun., sor genes. J. Cell. Biochem., 46,199-205 (1991). 129,300-305 (1985). BRAVARD, A,, SABATIER. L., HOFFSCHIR, F., RICOUL, M., LUCCIONI, C. MICHIELS, C., RAES,M., ZACHARY, M.D., DELAIVE, E. and REMACLE, and DUTRILLAUX, B., SOD2: a new type of tumor-suppressor gene? J., Microinjection of antibodies against superoxide dismutase and Int. J. Cancer, 52, 1-5 (1992). glutathione peroxidase. Exp. Cell Res., 179,581-589 (1988). CERUTTI, P.A. and TRUMP, B.F., Inflammation and oxidative stress in OBERLEY,L.W., MCCORMICK, E. and STM.L., SIERRA-RIVERA, carcinogenesis. Cancer Cells, 3, 2-7 (1991). CLAIR,D.K., Manganese superoxide dismutase in normal and transDUTRILLAUX, B. and COUTURIER, J., La pratiqrie de lhnalyse chromo- formed human embryonic lung fibroblasts. Free Radic. Biol. Med., 6, somique, Masson, Paris (1981). 379-384 (1989). S. and DE LEO, M.E., OBERLEY, GALEOITI.T., WOHLRAB,H., BORRELLO, L.W. and OBERLEY, T.D., Role of antioxidant enzymes in Messenger RNA for manganese and copper-zinc superoxide dismuta- cell immortalization and transformation. Mol. cell. Biochem., 84, ses in hepatomas: correlation with degree of differentiation. Biochem. 147-153 (1988). Biophys. Res. Commun., 165,581-589 (1989). SANDBERG, A.A., The chromosomes in human cancer and leukemia, 2nd GEISER,A.G. and STANBRIDGE, E.J., A review of the evidence for ed., Elsevier, New York (1990). tumor suppressor genes. Crit. Rev. Oncogenesis, 1,261-276 (1989). R.S., ALLEN,R.G. and NATIONS, C., Oxygen free radicals play SOHAL, HOFFSCHIR, F., RICOUL.M. and DUTRILLAUX, B., SV40-transformed a role in cellular differentiation: an hypothesis. J. free Radic, Biol. Med., human fibroblasts exhibit a characteristic chromosomal pattern. Cyto- 2, 175-181 (1986). genet. Cell Genet., 49,264-268 (1988). HOFFSCHIR, F., RICOUL,M., LEMIEUX, N., ESTRADE, S., CASSINGENA,SUN,Y., Free radicals, anti-oxidant enzymes, and carcinogenesis. Free R. and DUTRILLAUX, B., Jumping translocations originate clonal Radic. Biol. Med., 8,583-599 (1990). rearrangements in SV40-transformed human fibroblasts. Int. J. Cancer, YAMANAKA, N. and DEAMER,D., Effects of age, trypsinisation and 53,l-7 (1992). SV40 transformation. Chem. Phys., 6,95-106 (1974).
(m '
Publication of the lniernational Union Against Cancer Publication de I'Union lniernationaleConire le Cancer
*e5
EARLY SUPEROXIDE DISMUTASE ALTERATIONS DURING SV40-TRANSFORMATION OF HUMAN FIBROBLASTS A. BRAVARD',F. HOFFSCHIR',L. SABATIER', M. RICOUL',A. PINTON',R. CASSINGENA~, S. ESTRADE~, C. LUCCIONI' and
B. D U T R I L L A U X ' . ~ ~ ~ 'CEAIDSVIDPTEILCG, B P 6 92265 Fonteriay aux Roses Cedex; W P R 42, IRSC, BP 8, 94802 Villejuf Cedex and 3URA 620, CNRS, Institiit Curie, Section Biologie, 26, nce d'Ulm, 75231 Paris Cedex 05, France. The expression of superoxide dismutases (SOD) I and 2 was studied in 4 clones of human fibroblasts after their infection by simian virus 40 (SV40). in parallel with the alterations of chromosomes 21 and chromosome 6q arms, carrying the genes that encode for SOD1 and SOD2 respectively. For all clones, a similar scheme with 2 main phases was observed for both chromosome and SOD variations. The first phase, defined as the pre-crisis phase, was characterizedby chromosomal instability, but maintenance of normal numbers of chromosome 6q arms and chromosomes 2 I. The level of SOD2 mRNA was high, while SODZ activity and immunoreactive protein were low. SOD I protein and activity were decreased. In the second phase, defined as the post-crisis phase, the accumulation of clonal chromosomal rearrangements led to the loss of 6q arms, while the number of chromosomes 21 remained normal. SOD2 mRNA level was decreased and SOD2 immunoreactive protein and activity remained low. SOD I protein and activity increased with passages, reaching values similar to those of control cells at late passages. As in established SV40-transformed human fibroblast cell lines, good correlation was found between SOD2 activity and the relative number of 6q arms. These results allow us to reconstruct the sequence of events leadingto the decrease of SODZ, a possible tumor-suppressor gene, during the process of SV40-transformation of human fibroblasts.
o 1992 Wiley-Liss, Inc.
The frequent deletions occurring in human cancers suggest the existence of tumor-suppressor genes. The deletion of the long arm of chromosome 6 (6q) is frequently observed in a number of malignancies such as lymphomas, melanomas, glioblastomas, and breast and ovarian carcinomas, among others (Sandberg, 1990), as well as in SV40 (simian virus 40)-transformed human cells (Hoffschir et al., 1988). This chromosome segment carries the gene encoding for the enzyme superoxide dismutase 2 (SOD2; E.C.1.15.1.1) located on band 6q21 (HGM 10,1989), which catalyzes the dismutation of supcroxide radicals. It has been repeatedly proposed that free radicals of oxygen are implicated in the initiation/promotion phases of carcinogenesis as well as in cellular differentiation and proliferation process (for review, see Oberley and OberIcy, 1988; Sun, 1990; Allen, 1991; Cerutti and Trump, 1991). A decrease of SOD2 expression has been observed in a great number of tumors (Galeotti et al., 1989; Sun, 1990; Kwee et al., 1991) and in SV4O-transformed human fibroblasts (Yamanaka and Deamer, 1974; Marklund et al., 1982; Marlhens et al., 1985; Oberley et af., 1989). In our model, we reported the decrease of SOD2 activity as well correlated with the loss of chromosome 6q (Bravard et af., 1992). These data led us to suggest that SOD2 might be a new type of tumor-suppressor gene. In the prcsent study, we tried to reconstruct the sequence of events resulting in chromosome 6q arm losses and decrease of SOD2 activity by studying these parameters during the early steps of human fibroblast transformation by SV40. SOD2 activity and loss of the 6q arm were also compared with SODl activity and chromosome 21 on which this enzyme is mapped (HGM10, 1989). MATERIAL AND METHODS
Cells Four clones of human fibroblasts were analyzed at different passages (ranging from 9 to 112) after infection by SV40. Two
clones, CHSV3 and CHSV4, were derived from cornea and 2, DHSV2 and DHSV4, from dermis. The corresponding normal non-infected cells in culture were used as controls. They were primary fibroblast cultures which became senescent around passage 20. Cells were grown in DMEM-F12 medium with 10% FCS.
Cytogenetic analysis Metaphases were spread and R-banded according to our usual methods (Dutrillaux and Couturier, 1981). For each passage, at least 10 karyotypes were established. After tentative identification of the rearranged chromosomes, we tried to determine the location of the breakpoints and counted the number of copies of each chromosome or chromosome segment. This allowed us to estimate the average number of copies of each chromosome per passage and in particular to count the relative number of 6q arms: the average number of chromosome 6q was divided by the average number of chromosomes and multiplied by 23 (normal haploid number), the balanced status being 1. Identical analysis was done for chromosome 21. SOD activities and proteins Cells were collected by trypsinization in the exponential growing phase, 72 hr after feeding, washed twice in PBS by centrifugation and stored in liquid nitrogen until extraction. Extraction was performed by successive freezing and thawing in 10 mM Tris HCI, p H 7.5, 0.1% Triton X-100 and 200 mM sucrose. Enzymatic and Western-blotting assays were performed on 20,000 g supernatant. S O D assay was adapted from the method of Beauchamp and Fridovich (1971). It was carried out at 30°C in 1 ml of 20 mM sodium carbonate buffer, p H 10.0, containing 0.1 mM DTPA, 0.2 mM xanthine, 12 FM nitro blue tetrazolium (NBT) and 1.95 munits xanthine oxidase. SOD2 activity was determined as the remaining SOD activity after addition of 5 mM KCN. The reaction was followed at 560 nm with a spectrophotometer. One unit of SOD activity was defined as the amount of enzyme inhibiting the rate of NBT reduction by 50%. Protein content was determined using a Biorad kit with albumin as standard. Antibodies against E. coli MnSOD (SOD2) and human CuZnSOD (SOD1) (Sigma, St Louis, MO) were raised in rabbit by serial injections of 1 mg purified protein. Total sera were used for immunoblotting. After pre-treatment with SDS and P-mercaptoethanol at 100°C for 4 min, electrophoresis of 10 to 20 pg of homogenates were performed in a 4% stacking, 15% separating polyacrylamide gel in 25 mM Tris, 192 mM glycine, 0.1% SDS buffer, p H 8.3. After electrotransfer of proteins, nitrocellulose sheets were saturated with dried milk and incubated for 1 hr with diluted anti-SOD antibodies. The specific immunoreactive S O D l or SOD2 band was revealed using an Amersham blotting-detection kit for rabbit bound ?To whom correspondence and reprint requests should be addressed. at the Institut Curie.
Received: April 30,1992 and in revised form June 27,1992.
798
B R A V A R D ET AL.
antibodies (RPN.23). For quantification, the density of each band was measured with a laser densitometer 2202 Ultrascan LKB (Bromma, Sweden).
Northern blot The technique and references have been described in Bravard et al. (1992). Briefly, total cellular RNA was prepared either from cell pellets after trypsinization or directly from cells in culture flasks. The purified R N A was quantified spectrophotometrically; 20 p,g per lane were size-fractionated by formaldehyde agarose (1.5%) gel electrophoresis and transferred to nylon. The filters were stained with methylene blue to confirm R N A integrity. For detection of R N A sequences, a random-primed cDNA SOD2 probe was utilized. Random priming was performed using a BRL kit (Paris, France). The levels of 4-kb and 1-kb SOD2 mRNA were estimated by comparison with methylene blue staining. RESULTS
Cytogenetic analysis The cytogenetic evolution of the 4 clones described by Hoffschir ef al. (1992), may be summarized as follows. Clone CHSV3: progressive decrease of chromosome number to about 35 (passage 36), followed by endoreduplication (between passages 36 and 48) and further chromosome losses. No deletion of 6q arm nor loss of chromosomes 21 were observed, all variations being related to the changes of ploidy of the metaphases. Clone CHSV4: progressive decrease of chromosome number to about 37 (passage 44), without polyploidization. One chromosome 6 was missing after passage 21. Chromosomes 21 remained unchanged. Clone DHSV2: progressive decrease of chromosome number to about 39 (passage 18), followed by endoreduplication and further chromosome losses. A deletion of 6q arm occured between passages 18 and 32. Chromosomes 21, not lost, increased in relative number. Clone DHSV4: progressive decrease of chromosome number to about 42 (passage 44). The occurrence of various rearrangements of chromosome 6 resulted in its frequent loss at late passages. Chromosomes 21 remained stable. The relative numbers of chromosome 6q arms and of chromosomes 21 at the passages studied for enzyme activities are given in Table I. For all clones but CHSV3, a single crisis, characterized by high cell lethality and chromosome instability, occurred around passage 20. The surviving cells had acquired several clonal chromosome rearrangements. For CHSV3, a second marked crisis occurred around passage 30, and most clonal chromosome rearrangements were acquired afterward. SOD activities and proteins
SOD activities, measured at 4 passages at least for each clone, were compared with the activity in control cells. Except for CHSV3. SOD2 activity was always decreased after the first passage studied and remained low at later passages (Table I, Fig. 1). In clone CHSV3, SOD2 activity was increased at the
first 2 passages studied (passages 12 and 29), and decreased in later passages (Table I, Fig. 1). For all clones, S O D l activity was decreased at the first passage studied, but increased in later passages, frequently reaching higher values than those of control cells (Table I). This resulted in a strong decrease of the SOD2/SOD1 ratio with passages. SOD2 and SOD1 immunoreactive proteins were quantified for one passage at least of each clone and for control fibroblasts. In all cases, the amount of immunoreactive protein was correlated with the corresponding enzyme activity (Figs. 2, 3). SOD2 mRNA The I-kb and 4-kb bands, characteristic for SOD2 mRNA were observed at various intensities depending on passages (Figs. 1,4). The 1-kb band, which is assumed to give rise to the active SOD2 protein, followed a similar evolution in all clones: increased at early passages (p12-p29 for CHSV3, p14 for CHSV4, p17-p20 for DHSV2 and p9-p32 for DHSV4) and decreased at later passages. For the 4-kb band, a similar evolution was observed with variations of a lower magnitude. Correlations between cytogenetic, eniymologic and mRNA data When the 4 clones were considered, SOD2 activities were significantly correlated with the relative numbers of 6q arms (Fig. 5 ) . In spite of this good correlation, the chronology of 6q arm loss and decrease of SOD2 activity was not exactly superimposable, since the decrease of activity always preceded chromosome loss. The relationship between the amount of SOD2 mRNA and the number of 6q arms was less simple: the early increase of mRNA was not accompanied by a change in the number of 6q arms, whereas the decrease of mRNA paralleled the loss of 6q arms at late passages. The difference between the amount of SOD2 mRNA and SOD2 activity was still more marked at early passages, since SOD2 activity was low, whereas the amount of mRNA was high. At late passages, both the enzyme activity and the amount of mRNA were decreased. No correlation was found between the variations of S O D l activity and the relative numbers of chromosome 21 at early passages. A t late passages, the 2 parameters were similar or higher than in control cells. DISCUSSION
We have recently shown that, in SV40-transformed human fibroblast cell lines, a correlation exists between low SOD2 activity and the loss of the chromosome 6q arm where the gene SOD2 is mapped. This low enzyme activity was also related to low mRNA content. As these particularities may be related to a growth advantage of transformed cells, we have proposed that SOD2 gene could correspond to a new type of tumorsuppressor gene (Bravard et al., 1992). The study was performed on established cell lines which prevented us from reconstructing the sequence of the events leading to SOD2
TABLE 1 - SOD2 AND SODl ACTIVITIES IN UiMG PROTEIN AND RELATIVE NUMBERS OF CHROMOSOME 6q ARMS AND FHROMOSOMES 21 AT VARIOUS PASSAGES OF CLONES CHSV AND DHSV AND IN THEIR RESPECTIVE CONTROL CELLS ~~~
Control
pl?
p29
SOD2 activity 31 40 55.5 Relative number 1 1.01 1.16 of 6q 118.2 58.4 94.5 SODl activity Relative number 1 1.12 0.85 of 21 SODZ/SODl 0.26 0.68 0.59 activity 'p = passage; nd = not done.
CHSV3 p53 p65
CHSV4 pl00
p14
p28
p50
p63
pll2
Control
DHSV2 pY
p34
p44
DHSV4 ps3
pY
p32
p44
p53
22.4 24.8 nd 16.6 9.0 12.3 7.2 3.6 100.5 28.9 14.5 15.5 15.1 45.4 25.6 18.0 5.8 1.09 0.89 0.78 0.90 0.60 0.60 0.50 0.62 1 1.05 0.62 0.70 0.48 1 0.85 0.75 0.61 86.7 127.4 nd 58.4 122.7 137.9 121.2 92.9 83.9 77.7 91.7 135.1 119.5 41.5 53.8 61.5 78.1 0.85 1.25 1.22 1.07 1.07 0.96 0.99 1.14 1 1.01 1.34 1.45 0.78 1 1.01 0.90 0.94 0.26 0.19 nd 0.28 0.07 0.09 0.06 0.04
1.20 0.37 0.16 0.11 0.13 1.10 0.48 0.29 0.07
799
SOD ALTERATIONS IN SV40-INFECTED FIBROBLASTS
DHSV2 120
-
7
mRNA
100
CHSV3 6o 50
1
+
-
+
mRNA
80 60
40
40
30
20
20 0
C
10
p9
p34
p44
p53
assages
n C
p12
p29
p53
p65
Passages
DHSV4 120 100
1
+
+
-
-
p9
p32
p44
p53
mRNA
80 40 60
30 40
20
20
10
0
0 C
p14
p28
p50
p63
p112
C
passages
FIGURE 1 - SOD2 activities in Uimg protein (black bars) and relative numbers of 6q arms (hatched bars; multiplied by 50 for clones CHSV and by 100 for clones DHSV to use the same scale as that of SOD2 activity) at passages of the different clones and in their respective control cells (C). Quantitative variations of SOD2 mRNA are indicated by + or - for increase or decrease in comparison with
control cells.
disregulation. This is why we investigated, on recently transfected cells, the variations of chromosome number, of the amounts of SOD mRNA and protein, and of enzyme activity. The study enables us to propose a scheme with 2 major phases. For all clones except CHSV3, a first phase occurred before the crisis, which took place around passage 20. This phase was characterized by high chromosome instability, shown by the occurrence of jumping and clonal translocations (Hoffschir et al., 1992), high SOD2-mRNA content, a small amount of SOD2 protein, and low SOD2 activity. Chromosome instability may be regarded as a direct consequence of high mutagenesis also resulting in gene mutations. This might affect at least one allele of SOD2 gene which was transcribed but not efficiently translated, or result in increased protein degradation, as suggested by the high mRNA and low amounts of protein. A second phase took place after the crisis. All the cells of a given clone, having acquired several chromosomal rearrangements in common, were obviously of a monoclonal origin (Hoffschir et al.. 1992). Some of these rearrangements led to the deletion of a 6q arm and thus to hemizygozity for SOD2 gene. This might result in the loss of the mutated rather than of the normal allele, as suggested by the decrease of SOD2 mRNA, whereas SOD2 activity and protein remain at a level that is low, but comparable with that of the pre-crisis period. At this stage, SOD2 status became similar to that described in established SV40-transformed human fibroblast cell lines (Yamanaka and Deamer, 1974; Marklund et al., 1982; Marlhens ef al., 1985; Oberley ef al., 1989; Bravard et al., 1992). For the
CHSV3 clone, the same events occurred but with a delay apparently related to the occurrence of 2 crises: one around passage 20, and another around passage 30. The alterations of SOD2 characterizing the second phase were observed only after the second crisis. Modifications affecting S O D l protein and activity were similar to those of SOD2 during the early passages, with low protein content and low enzyme activity. However, the postcrisis phase was obviously different, with an increase both of protein and of activity reaching levels similar to those of control cells. As far as we could correctly estimate it, the number of copies of chromosome 21 was not changed. These results on S O D l in the post-crisis period are similar to those obtained in established SV40-transformed human fibroblast cell lines (Yamanaka and Deamer, 1974; Marklund et al., 1982; Bravard et al., 1992). Finally, the decrease of SOD2 activity, which is frequently observed in tumors (Galeotti et al., 1989; Sun, 1990; Kwee ef al., 1991), appears to be an early event in the process of SV40-transformation of human fibroblasts. Previous data suggested that variations OP SOD2 activity are related to cell proliferation and differentiation processes, low SOD2 activity being generally observed in proliferative cells and increased SOD2 activity in differentiating cells (Sohal et al., 1986; Oberley and Oberley, 1988; Allen, 1991). Antibodies directed against SOD were able to promote the proliferation of human fibroblasts (Michiels et al., 1988), and SOD-like compounds or
800
ERAVARD E T A L .
I2O I00
a
1
80 60 40
20 0
200
b 150
I00
50
0
FIGURE2 - SOD activities (Uimg protein; black bars) and amounts of SOD immunoreactive proteins (arbitrary units; hatched bars) at passages of clones during transformation and in their respective control cells (C). (a) SOD2; (h) SOD1.
FIGURE 3 -Western blots for SOD proteins. (a) SOD2. CHSV3p29; 2, CHSV3p53; 3, CHSV control: 4, CHSV4p14; DHSV control; 6, DHSV2p9; 7, DHSV4p9. (b) SOD1. CHSV3p12; 2, CHSV3p53; 3, CHSV control; 4, CHSV4p14; CHSV4p50; 6, DHSV control; 7, DHSV2p53; 8, DHSV4p9.
1, 5, 1, 5,
liposornic SOD were successfully used to reduce the growth and increase the differentiation of tumor cells (Beckman et aL, 1989; Allen, 1991). Our results also suggest that the decrease of SOD2 correlates with cell proliferation and immortalization as suggested by Oberley and Oberley (1988). These character-
FIGURE 4 - Northern blot for SOD2 ( I - and 4-kb bands; only the 1-kb mRNA is thought to give rise to active SOD2) from normal DHS fibroblasts (C) and various passages of DHSV2 and DHSV4 clones. The comparison of the intensity of the 1-kb band with that of methylene blue (not shown) was performed to give the indications (+ or -) of Figure 1.
istics being shared with tumor-suppressor genes, we propose that the SOD2 gene has a tumor-suppressor function. In view of this hypothesis, it is interesting to compare SOD2 with the 2 major suppressor genes, RB1 and P53, whose mechanisms of action are being elucidated (Geiser and Stanbridge, 1989; Bookstein and Lee, 1991; Boylan and Zarbl, 1991). Complete inactivation of the RB1 gene seems to be required for the development of retinoblastoma, while in other tumors, such as
SOD ALTERATIONS IN SV40-INFECTED FIBROBLASTS
60 h
.-> .-
I
m
-
50
N
30 -
v,
20-
0 0
0
40 -
0
rz0.83 0 0 .
10
-i
'.,
0
0,4
0
0,6 0.8 1,o Relative number of 6q
e 1,2
FIGURE 5 - Correlation between SOD2 activities (U/mg protein) and the relative numbers of 6q arms (data from Table I) in passages of CHSV (black circles) and DHSV clones (open circles).
801
breast, lung, prostate and bladder carcinomas, RB alterations leading t o loss of the normal protein or to its mutation appear as frequent non-random events. T h e P53 gene also undergoes frequent alterations in human tumors. Except for the LiFraumeni syndrome, most of these alterations occur during tumor progression, leading t o the formation of a mutated P53 protein. Deletions of the 17p arm, where the P53 gene is mapped, may also occur, suppressing the normal allele. SOD2 disregulation may correspond to a different mechanism; since the activity of the protein is never totally lost, its decrease seems related t o a gene dosage effect, a t least after the crisis. T h e tumor-suppressor activity of RBI and P53 protein may be related to transcriptional control of genes involved in cell proliferation through their DNA-binding properties. For SOD2, a hypothesis based on the known biological effects of free radicals might explain its anti-oncogenic properties: the decrease of SOD2 activity may induce a pro-oxidant state, a condition frequently related to the initiation and promotion phases of carcinogenesis (Oberley and Oberley, 1988; Sun, 1990; Cerutti and Trump, 1991).
REFERENCES
GENEMAPPING 10: Tenth International Workshop on Human ALLEN,R.G., Oxygen-reactive species and antioxidant responses HUMAN during development: the metabolic paradox of cellular differenciation. Gene Mapping. Cytogenet. Cell Genet., 51 (1989). Proc. Soc. exp. Biol. Med., 196,117-129 (1991). O.R., Lowered superoxide KWEE,J.K., MITIDlERI, E. and AFFONSO, I., Superoxide dismutase: improved dismutase in highly metastatic B16 melanoma cells. Cancer Lett., 57, BEAUCHAMP, C. and FRIDOVICH, assay and an assay applicable to acrylamide gels. Anal. Biochem., 44, 199-202 (1991). 276-287 (1971). S.L., WESTMAN, N.G., LUNDGREN, E. and Roos, G., MARKLUND, BECKMAN, B.S., BALIN,A.K. and ALLEN,R.G., Superoxide dismutase Copper-and-zinc-containing superoxide dismutase, catalase, and glutainduces differentiation of friend erythroleukemia cells. J. CeN Physiol., thione peroxidase in normal and neoplastic human cell lines and 139,370-376 (1989). normal human tissues. Cancer Res., 42,1955-1961 (1982). BOOKSTEIN, R. and LEE,W., Molecular genetics of the retinoblastoma MARLHENS, F., NICOLE,A. and SINET,P.M., Lowered level of translatsuppressor gene. Crit. Rev. Oncogenesis, 2,211-227 (1991). able messenger RNAs for manganese superoxide dismutase in human BOYLAN, M.O. and ZARBL,H., Transformation effector and suppres- fibroblasts transformed by SV40. Biochem. Biophys. Res. Commun., sor genes. J. Cell. Biochem., 46,199-205 (1991). 129,300-305 (1985). BRAVARD, A,, SABATIER. L., HOFFSCHIR, F., RICOUL, M., LUCCIONI, C. MICHIELS, C., RAES,M., ZACHARY, M.D., DELAIVE, E. and REMACLE, and DUTRILLAUX, B., SOD2: a new type of tumor-suppressor gene? J., Microinjection of antibodies against superoxide dismutase and Int. J. Cancer, 52, 1-5 (1992). glutathione peroxidase. Exp. Cell Res., 179,581-589 (1988). CERUTTI, P.A. and TRUMP, B.F., Inflammation and oxidative stress in OBERLEY,L.W., MCCORMICK, E. and STM.L., SIERRA-RIVERA, carcinogenesis. Cancer Cells, 3, 2-7 (1991). CLAIR,D.K., Manganese superoxide dismutase in normal and transDUTRILLAUX, B. and COUTURIER, J., La pratiqrie de lhnalyse chromo- formed human embryonic lung fibroblasts. Free Radic. Biol. Med., 6, somique, Masson, Paris (1981). 379-384 (1989). S. and DE LEO, M.E., OBERLEY, GALEOITI.T., WOHLRAB,H., BORRELLO, L.W. and OBERLEY, T.D., Role of antioxidant enzymes in Messenger RNA for manganese and copper-zinc superoxide dismuta- cell immortalization and transformation. Mol. cell. Biochem., 84, ses in hepatomas: correlation with degree of differentiation. Biochem. 147-153 (1988). Biophys. Res. Commun., 165,581-589 (1989). SANDBERG, A.A., The chromosomes in human cancer and leukemia, 2nd GEISER,A.G. and STANBRIDGE, E.J., A review of the evidence for ed., Elsevier, New York (1990). tumor suppressor genes. Crit. Rev. Oncogenesis, 1,261-276 (1989). R.S., ALLEN,R.G. and NATIONS, C., Oxygen free radicals play SOHAL, HOFFSCHIR, F., RICOUL.M. and DUTRILLAUX, B., SV40-transformed a role in cellular differentiation: an hypothesis. J. free Radic, Biol. Med., human fibroblasts exhibit a characteristic chromosomal pattern. Cyto- 2, 175-181 (1986). genet. Cell Genet., 49,264-268 (1988). HOFFSCHIR, F., RICOUL,M., LEMIEUX, N., ESTRADE, S., CASSINGENA,SUN,Y., Free radicals, anti-oxidant enzymes, and carcinogenesis. Free R. and DUTRILLAUX, B., Jumping translocations originate clonal Radic. Biol. Med., 8,583-599 (1990). rearrangements in SV40-transformed human fibroblasts. Int. J. Cancer, YAMANAKA, N. and DEAMER,D., Effects of age, trypsinisation and 53,l-7 (1992). SV40 transformation. Chem. Phys., 6,95-106 (1974).
