Superoxide Dismutase Activity and Superoxide Dismutase-1 Gene Methylation in Normal and Tumoral Human Breast Tissues Martha S. Bianchi, N stor O. Bianchi, and Alejandro D. Bolz in
ABSTRACT: The superoxide dismutase [SOD) activities in normal and tumor breast tissues from 14 human females were determined by the epinephrine autaxidation assay. SOD levels showed a marked interindividual variability in normal and malignant cells. However, each donor had a higher SOD activity in cancer than in normal tissue samples. In three cases in which M n - and CuZnSOD activities were determined, it was found that tumoral increases in SOD were due to increases in both enzymatic forms. Therefore, it seems reasonable to assume a similar situation for all cases in our series. The level of DNA methylation in the SOD-1 gene was assessed in the first four donors. The four cases exhibited full methylation of SOD-1 genes corresponding to normal as well as to cancer cells. It is concluded that the variability in CuZnSOD activities is not related with the state of methylation of the SOD-1 gene. MspI restriction fragment polymorphisms between DNA samples from normal and malignant cells were detected in the four DNA donors. This phenomenon may be due to point mutations changing the frequency of Mspl sites or to methylation of the external C in CCGG sequences.
INTRODUCTION
MATERIALS AND METHODS
It has been reported that superoxide dismutase (SOD) (EC 1.15.1.1) activity in human blood shows a marked interindividual variation and that the sensitivity of lymphocyte chromosomes to ionizing radiation and b]eomycin is inversely correlated with SOD levels [1, 2]. We show here that SOD activities in normal and tumor breast tissues also exhibit a marked interindividual variation. Moreover, all donors in our series had higher total SOD activity in malignant than in normal cells. Determination of M n - and CuZnSOD levels strongly suggests that total SOD activity increases in tumor tissues results from increases of both isoenzymes. The causes responsible for the intertissue variations in SOD levels are not known. Hence, it is possible to speculate that SOD activity variations may result from changes in deoxyribonucleic acid (DNA) methylation that influence the rate of expression of the CuZn-superoxide dismutase-1 (SOD-l) gene. We explored this possibility in normal and tumor mammary tissues. Our results show no correlation between the methylation state of the SOD-1 gene and the cellular activity of SOD.
Normal and cancer breast tissue samples were obtained from patients undergoing therapeutic mastectomy. Samples were kept in dry ice until processing. Granulocytes and mononuclear leukocytes were separated from two blood samples by Ficoll fractionation [3]. High molecular weight DNA from solid tissue samples and from blood fractions were obtained as described elsewhere [4]. The use of human samples was authorized by the Committee of Bioethics of the "Hospital Interzonal General de Agudos Gral. San Martin." DNA samples (5/~g) were digested with MspI or Hpall (GIBCO-BRL). The amount of enzyme used was 7-8 U/t~g of DNA. Electrophoresis, blotting, and filter hybridizations were performed as previously indicated [51. The highest stringency washing of filters was with 0.1 x SSC, 0.1 × SDS for 20 min at 60°C. Autoradiograms were exposed for 48 hours at - 70°C with enhancer screens. The probe used is a SOD-1 cDNA fragment subcloned in SP64 (pSP64SOD), kindly provided by Y. Groner. The characterization of this probe has been detailed elsewhere [6, 7].
Southern Blotting
Enzyme Assay
From IMBICE, La Plata, Argentina. Address r e p r i n t requests to: Dr. Martha S. Bianchi, IMBICE 526 e~ 10 y 11, C. C. 403, 1900 La Plata, Argentina. Received M a r c h 25, 1991 -~ A u g u s t 9, 1991
Frozen tissue samples were finely minced, washed several times in cold phosphate buffer solution (pH 7) and homogenized in a Waring Blender during four periods of 15 seconds each with pauses of 10 seconds. Thereafter, Triton X100 (0.1%) was added, and mixtures were further lysed on crushed ice with l-rain sonications in bursts of 10 seconds
26 Cancer Genetics Cytogenet 59:26-29 (1992) O165-4608/92,'S05.00
':~'; 1992 Elsevier 5(:ience Publishing Co. Inc. 655 Avenue of the Americas, New York, NY 10010
Superoxide Dismutase and Gene Methylation
27
(W-225 R Sonicator, Heat Systems, Ultrasonic). The homogenates were centrifuged at 20,000 × g for 30 minutes at 4°C to remove cell particles. Enzyme activity determinations were made in the supernatant. Total SOD activity was d e t e r m i n e d following the inhibition of e p in e p hr i n e autoxidation to a d r e n o c h r o m e [8]. M n - and CuZnSOD activities were d e t e r m i n e d according to Misra [9]. Cyanide was used to inhibit CuZnSOD during the e p i n e p h r in e assay. The activity obtained was considered to represent MnSOD levels. Differences between total and MnSOD activities were assumed to represent the CuZnSOD activity. One unit of SOD activity was defined as the amount of e n z y m e inducing a 50% inhibition in the rate of epinephrine autoxidation. Total SOD, C u Z n - , and MnSOD activities were expressed as units per milligram of protein.
Table 2
RESULTS
while Mspl is not influenced by the state of methylation of the internal C [10]. DNA samples from normal and cancer tissues were digested with MspI or HpaII, electrophoresed, blotted, and hybridized with the pSP64-SOD probe. At low stringency washes, this probe hybridizes with the SOD-1 gene and with pseudogenes having partial homology with the SOD1 gene. On the other hand, at the stringency employed in this report, the pSP64-SOD probe reacts only with the SOD-1 gene [7]. MspI digestions produced several DNA fragments that point out the presence of several MspI sites along the SOD-1 gene domain. Some of these fragments appeared in DNAs from normal but not in DNAs from cancer cells, or vice versa (Fig. 1). HpalI generated high molecular weight smears characteristic of insufficient DNA digestions (Fig. 1). This finding indicates methylation of the internal cytosines in all HpaII sites along the SOD-1 domain. Mo n o n u cl ear leukocytes cultured in vitro in the presence of p h y t o h e m a g g l u t i n i n (PHA) start DNA synthesis and mitosis. This process, w h i c h resembles blastic transformation [11], usually induces an overexpression of most housekeeping genes. Because SOD-1 is a housekeeping gene [12], we analyzed its state of methylation in granulocytes, resting m o n o n u c l e a r granulocytes, and 72-hour PHA-stimulated m o n o n u c l e a r granulocytes. MspI and HpaII digestions of DNA samples from these three varieties of cells showed that SOD-1 genes were fully methylated regardless of the resting or active state of the cell population (Fig. 2).
The SOD activity in normal and tumor breast tissues is indicated in Table 1. It is clear that normal as well as malignant tissues show a marked interindividual variability in total SOD levels. Moreover, in each case, the activity of total SOD in tumor tissue is higher than in normal cells. M n - and CuZnSOD activities were determined in the last three cases of our series (Table 2). In the three individuals, it was observed that the increases in SOD activity in tumor tissues were due to increases in M n - and CuZnSOD isoenzymes. Therefore, it seems reasonable to assume that both enzymatic forms were also increased in all the breast tumors analysed. The first four donors were selected for analysis of SOD1 gene methylation. MspI and HpalI are isoschizomer enzymes that both recognize the sequence CCGG. However, HpaII is unable to cleave w h e n the inner C is methylated,
Table 1 Donor
SOD a activity in normal and tumor breast tissues b Normal
Tumor
Diagnosis Duct carcinoma Duct carcinoma Duct carcinoma Duct carcinoma Lobular carcinoma Duct carcinoma Duct carcinoma Tubular and papillary carcinoma Duct carcinoma Apocrine carcinoma Medullary carcinoma Duct carcinoma Duct carcinoma Duct carcinoma
A B C D E F G H
2.82 16.83 1.48 1.04 0.29 1.68 2.20 2.41
-+ 0.16 c -+ 0.25 d 4- 0.22 -+ 0.12 -* 0.03 - 0.05 -+- 0 . 0 2 +- 0.15
17.17 -* 0.42 21.81 *- 0.10 7.97 -*- 0.45 4.67 +- 0.06 6.08 -+ 0.05 3.16 -+ 0.19 27.60 + 0.62 3.99 - 0.23
I J K L M N
1.71 2.30 0.87 1.32 1.08 0.94
-+ 0.21 -+ 0.08 + 0.02 -+ 0.08 -+ 0.03 -+ 0.05
5.67 - 0.17 12.38 - 0.14 32.23 +- 0.43 5.19 +- 0.09 6.13 + 0.19 10.77 -+ 0.36
° Superoxide dismutase (SODI. bSOD activity is in U/mg of protein. Three independent SOD dosages were per/ormed in each case. ' Mean -+ SD. a Histopathologicanalysis showed that control tissue had infiltration with cancer cells.
MnSOD ° and CuZnSOD b activities in normal and tumor breast tissues c
Donor L M N
MnSOD CuZnSOD MnSOD CuZnSOD MnSOD CuZnSOD
Normal
Tumor
0.25 -+ 0.04'~ 1.07 -+ 0.02 0.47 -+ 0.03 0.61 -+ 0.00 0.19 4- 0.04 0.75 -+ 0.01
1.31 -+ 0.11 3.88 -+ 0.02 2.64 -+ 0.09 3.49 -+ 0.10 2.84 -+ 0.07 7.93 -+ 0.29
Diagnosis Duct carcinoma Duct carcinoma Duct carcinoma
Manganese-containing superoxide dismutase. ~:Copper- and zinc-containingsuperoxide dismutase. ': SOD activity is in U/mg of protein. Three independent SOD dosages were performed in each case. '~Mean -+ SD.
DISCUSSION We report here that t u m o r m a m m a r y tissues have higher total SOD activity and higher M n - and CuZnSOD activities than do their normal tissular counterparts. Similar findings have also been described in several other instances. In SV40-transformed WI-38 cells, SOD levels were somewhat higher than in normal WI-38 cells [13]. Moreover, total SOD activity in myelocytic, monocytic, and l y m p h o cy t i c leukemia cells was increased in regard to that observed in mature normal blood cells [14]. It is also worth mentioning that marked interindividual variations in the SOD content of 46 h u m a n tumors, i n cl u d i n g 10 breast ma-
28
M . S . Blanc:hi et al.
1
2
3
J q tp 4.4- i 2.32
L •
•
4
e,
•
o
•
•
~
~'
1R2 3 415t Re 7t819RlO 11t121131415--t16 A B C O Figure 1 Normal (n) and tumor (t) DNA samples digested with MspI (odd numbers) or Hpall (even numbers) and hybridized with the pSP64-SOD probe. A-D correspond to A-D cases in Table 1. Asterisks indicate bands that show in normal but not in tumor DNA (cases A-C) or in tumor but not normal DNA (case D). Size markers in kb are indicated in the left side of the figure. The weak smear in Hpoll digestions is due to deficient blotting of high molecular weight undigested DNA fragments.
lignancies, have been reported [15], yet, as no SOD determinations were performed in control tissues, no conclusion regarding SOD activities in cancer versus normal cells was available in this article. In the only instance in which M n - and CuZnSOD activities in normal and cancer tissues were measured, it was found that the levels of both isoenzymes showed a marked interindividual variation with no consistent differential pattern of enzyme activity in normal versus cancer tissues [16]. However, the conclusions of this article are probably made more difficult by the fact that normal and cancer samples derived from different individuals and from an ample variety of organs not always comparable. It is generally accepted that variations in the level of DNA methylation influence gene expression [see review in 17]. Bird [18, 19] has found "methylation free islands" in the genome of vertebrates. These islands occur in the 5' region of most housekeeping genes, are several h u n d r e d base pairs long, have a relative G + C richness over 50% as compared with the 40% exhibited by bulk DNA, and have GpC/CpG ratios close to I as opposed to ratios 5 - 1 0 shown by bulk DNA. Demethylation of CpG-rich islands usually correlates well with high expression of associated genes. Conversely, facultative methylation of the islands may produce underexpression or even s h u t d o w n of neighbor genes 1191. Analysis of Table 3 indicates that exon 1 of the SOD-1
-
4
Q
Figure 2 DNA from granulocytes (1), resting mononuclear leukocytes (2), or PHA-stimulated mononuclear leukocytes (3}. MspI(M) or Hpall(H) digestions. The asterisk points out an Mspl band that shows in one cell fraction and that is absent or remarkably weak in the other two cell fractions. gene and a 375 basepair long sequence located upstream of the exon 1 are CpG-rich islands. The pSP64-SOD probe used in this report hybridizes with the five exons and with sequences up- and downstream the SOD-1 gene [6, 7]. Therefore, it is logical to assume that some of the MspI Southern bands shown in Figures 1 and 2 result from fragments originated in the SOD-1 CpG rich islands. Hpall failed to cleave their canonical recognition sites in SOD-1 genes from normal and malignant tissues and from resting and PHA-stimulated blood cells. Hence, it can be conc:luded that the h u m a n SOD-1 gene is fully methylated in Table 3
Percentage of C - G and GpC/CpG ratios in the exons and 5' and 3' sequences of the h u m a n SOD-1U Gene b
Region
G + C (%)
GpC/CpG
Extent in bpc
5' sequence Exon I Exon II Exon llI Exon IV Exon V 3' sequence
61 60 47.7 51 50.4 44.7 31.9
1.2 2.3 4 4 3.5 4 15
375 69 96 72 117 105 531
" Superoxide dismutase-1. t' Table figureswere estimated from a partial gene sequencing[7]. ': Base pairs.
Superoxide Dismutase and Gene Methylation
breast and blood cells, and that interindividual and intertissue variations in SOD activity are not related to variations in SOD-1 gene methylation. Eukaryotic cells contain two distinct forms of SOD: a mitochondrial M n - c o n t a i n i n g (MnSOD) enzyme and a cytoplasmic Cu-Zn-containing (CuZnSOD) enzyme. The human SOD-1 gene located in the chromosome 21 codes for the CuZnSOD [20], while the SOD-2 gene in h u m a n chromosome 6 codes for the mitochondrial MnSOD [20]. Our results strongly suggest that both enzyme forms are responsible for SOD activity increases in tumoral tissues. Therefore, the lack of correlation between the state of SOD-1 gene methylation and the cellular SOD activity level seems to point out a lack of correlation between these two phenomena. A variability in the MspI b a n d i n g pattern between normal and tumor tissues was found in the four cases analyzed with the Southern blotting technique (Fig. 1). Malignant cell populations are of monoc:lonal origin. Accordingly, a point mutation producing the disappearance of a MspI site, or generating a new MspI recognition sequence from a cryptic site may give rise to somatic restriction fragment length polymorphisms (RFLP). Moreover, an alternative hypothesis can also explain the MspI RFLP observed. Methylation of the internal C in MspI recognition sequenc:es (CCGGI does not prevent DNA restriction by the enzyme. Conversely, when the external C is methylated, Mspl cannot cleave its recognition site [10]. DNA methylation in vertebrates preferentially occurs in the C of CpG dinucleotides [17]. Yet, methylation of the external C in CCGG sequences, although infrequent, has been demonstrated to exist in the DNA of h u m a n lymphocytes [21]. Hence, the Msp! RFLPs observed in our donors can be also explained by assuming a variable state of methylation in the external C of the sites delimiting the MspI polymorphic fragments. All cases in our series showed higher SOD activities in tumor than in normal breast tissues. On the other hand, the pattern of Mspl RFLP was not constant (donors A - C showed more MspI fragments in normal than in tumor DNA samples, while the situation was opposite for donor D) (Fig. 1). Accordingly, it seems unlikely that variations in the frequency or in the state of methylation of the external C in MspI sites may play a role in the intertissue variability in SOD content.
29
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12. 13.
14.
15.
16. We thank tile Surgery, Pathology, and Radiolngy Services of the "Hospital Interzonal General de Agudos Gral. San Martin," and we are grateful for the tecbnic:al assistance of Drs. R. Pascuale. R. Ifiigo, I,. Castelletto, and Mr. J. Padr6n. This work was supported by grants from CONICET. CIC, the Sigrid luselius Foundation. the Academy of Finland, the Folkh/ilsan Institute of Genetics, the Finnish Cancer Society, the Fundaci6n Antorchas and the Academy of Sciences of the Thircl World. A. D. Bolz&~ is a Fellow from the ('ONI('ET of Argentina.
17. 18. 19. 20.
REFERENCES 1. Lipecka K, Grabowska, B, Daniszewska K. Domanski T, Cisowska B (1984]: Correlation between the superoxide dis-
21.
mutase (SOD) activity in lymphocytes and the yield of radiation-induced chromosome aberrations. Studia Biophys 100:211-217. Larramendy ML, Bianchi MS, Padr6n J (1989): Correlation between the anti-oxidant enzyme activities of blood fractions and the yield of bleomycin-induced chromosome damage. Murat Res 214:129-136. B6yum A (1968): Isolation of leukocytes from human blood. Further observations. Stand J Clin Lab Invest 21(suppl 97):31-50. Mathew CGP (1984): The isolation of high molecular weight eukaryotic DNA. In: Methods in Molecular Biology, Nucleic Acids, Vol. 2, Walker, JM, ed. Clifton, NJ: The Humana Press, pp. 31-34. Bianchi NO, Peltom/iki P, Bianchi MS, Knuutila S, de la Chapelle A (1988): Demethylation of two specific DNA sequences in expressed human immunoglobulin light kappa constant genes. Somat Cell Mol Genet 14:13-20. Sherman L. Levanon D, l,ieman-Hurwitz ], Dafni N, Groner Y (1984): Human Cu/Zn superoxide dismutase gene: molecular characterization of its two mRNA species. Nucleic Acids Res 12:9349-9365. Levanon D, Lieman-Hurwitz J, Dafni N, et al. (1985): Architecture and anatomy of the chromosomal locus in human chromosome 21 encoding the Cu/Zn superoxide dismutase. EMBO J 4:77-84. Misra HP, Fridovich I (1972): The role of superoxide anion in the autoxidation of epinephrine and a simple assay for Superoxide Dismutase. ] Biol Chem 247:3170-3175. Misra HP (1985): Adrenochrome assay. In: Handbook of Methods for Oxygen Radical Research: Quantitation of Superoxide Dismutase, Greenwald RA, ed. CRC Press Inc., Boca Raton, F1, pp. 234-241. Nelson M. McClelland M (1989): Effect of site-specific methylation on DNA modification methyltransferases and restriction endonucleases. Nucleic Acids Res 17(suppl):R389-R415. Zackai EH, Mellman W] (1974): Human peripheral blood leukocyte cultures. In: Human Chromosome Methodology, Yunis JJ, ed. New York: Academic Press Inc., pp. 96-123. Holmquist GP (1989): Evolution of chromosome bands: molecular ec:olc~gyof noncoding DNA. J Mol Evol 28:469-486. Yamanaka NY, Deamer D (1974): Superoxide dismutase activity in WI-38 cell cultures. Effects of age, trypsinization, and SV-40 transformation. Physiol Chem Phys 6:95-106. Yamanaka NY, Ota K, tJtsumi K (1978): Changes in superoxide dismutase activities during development, aging and transformation. In: Biochemical and Medical Aspects of Active Oxygen, Hayaishi O, Asada K. eds. University Park Press. pp. 183-190. Sykes JA, McCormack Jr FX, O'Brien TJ (1978): A preliminary study of the superoxide dismutase content of some human tumors. Cancer Res 38:2759-2762. Westman NG, Marklund SL (1981): Copper- and zinc-containing superoxide dismutase and manganese-containing superoxide dismutase in human tissues and human malignant tumors. Cancer Res 41:2962-2966. Cooper DN (1983): Eukaryotic DNA methylation. Hum Genet 64:315-333. Bird AP (1986): CpG-rich islands and the function of DNA metbylation. Nature 321:209-213. Bird AP (1987): CpG islands as gene markers in the vertebrate nucleus. Trends Genet 3:342-347. Howard Hughes Medical Institute (1986): New Haven Human Gene Mapping Library. No. 1, HGM8, pp 1-76. Bianchi NO, Vidal-Rioja L, Cleaver JE (1986): Direct visualization of the sites of DNA methylation in human and mosquito chromosomes. Cbromosoma 94:362-366.
ABSTRACT: The superoxide dismutase [SOD) activities in normal and tumor breast tissues from 14 human females were determined by the epinephrine autaxidation assay. SOD levels showed a marked interindividual variability in normal and malignant cells. However, each donor had a higher SOD activity in cancer than in normal tissue samples. In three cases in which M n - and CuZnSOD activities were determined, it was found that tumoral increases in SOD were due to increases in both enzymatic forms. Therefore, it seems reasonable to assume a similar situation for all cases in our series. The level of DNA methylation in the SOD-1 gene was assessed in the first four donors. The four cases exhibited full methylation of SOD-1 genes corresponding to normal as well as to cancer cells. It is concluded that the variability in CuZnSOD activities is not related with the state of methylation of the SOD-1 gene. MspI restriction fragment polymorphisms between DNA samples from normal and malignant cells were detected in the four DNA donors. This phenomenon may be due to point mutations changing the frequency of Mspl sites or to methylation of the external C in CCGG sequences.
INTRODUCTION
MATERIALS AND METHODS
It has been reported that superoxide dismutase (SOD) (EC 1.15.1.1) activity in human blood shows a marked interindividual variation and that the sensitivity of lymphocyte chromosomes to ionizing radiation and b]eomycin is inversely correlated with SOD levels [1, 2]. We show here that SOD activities in normal and tumor breast tissues also exhibit a marked interindividual variation. Moreover, all donors in our series had higher total SOD activity in malignant than in normal cells. Determination of M n - and CuZnSOD levels strongly suggests that total SOD activity increases in tumor tissues results from increases of both isoenzymes. The causes responsible for the intertissue variations in SOD levels are not known. Hence, it is possible to speculate that SOD activity variations may result from changes in deoxyribonucleic acid (DNA) methylation that influence the rate of expression of the CuZn-superoxide dismutase-1 (SOD-l) gene. We explored this possibility in normal and tumor mammary tissues. Our results show no correlation between the methylation state of the SOD-1 gene and the cellular activity of SOD.
Normal and cancer breast tissue samples were obtained from patients undergoing therapeutic mastectomy. Samples were kept in dry ice until processing. Granulocytes and mononuclear leukocytes were separated from two blood samples by Ficoll fractionation [3]. High molecular weight DNA from solid tissue samples and from blood fractions were obtained as described elsewhere [4]. The use of human samples was authorized by the Committee of Bioethics of the "Hospital Interzonal General de Agudos Gral. San Martin." DNA samples (5/~g) were digested with MspI or Hpall (GIBCO-BRL). The amount of enzyme used was 7-8 U/t~g of DNA. Electrophoresis, blotting, and filter hybridizations were performed as previously indicated [51. The highest stringency washing of filters was with 0.1 x SSC, 0.1 × SDS for 20 min at 60°C. Autoradiograms were exposed for 48 hours at - 70°C with enhancer screens. The probe used is a SOD-1 cDNA fragment subcloned in SP64 (pSP64SOD), kindly provided by Y. Groner. The characterization of this probe has been detailed elsewhere [6, 7].
Southern Blotting
Enzyme Assay
From IMBICE, La Plata, Argentina. Address r e p r i n t requests to: Dr. Martha S. Bianchi, IMBICE 526 e~ 10 y 11, C. C. 403, 1900 La Plata, Argentina. Received M a r c h 25, 1991 -~ A u g u s t 9, 1991
Frozen tissue samples were finely minced, washed several times in cold phosphate buffer solution (pH 7) and homogenized in a Waring Blender during four periods of 15 seconds each with pauses of 10 seconds. Thereafter, Triton X100 (0.1%) was added, and mixtures were further lysed on crushed ice with l-rain sonications in bursts of 10 seconds
26 Cancer Genetics Cytogenet 59:26-29 (1992) O165-4608/92,'S05.00
':~'; 1992 Elsevier 5(:ience Publishing Co. Inc. 655 Avenue of the Americas, New York, NY 10010
Superoxide Dismutase and Gene Methylation
27
(W-225 R Sonicator, Heat Systems, Ultrasonic). The homogenates were centrifuged at 20,000 × g for 30 minutes at 4°C to remove cell particles. Enzyme activity determinations were made in the supernatant. Total SOD activity was d e t e r m i n e d following the inhibition of e p in e p hr i n e autoxidation to a d r e n o c h r o m e [8]. M n - and CuZnSOD activities were d e t e r m i n e d according to Misra [9]. Cyanide was used to inhibit CuZnSOD during the e p i n e p h r in e assay. The activity obtained was considered to represent MnSOD levels. Differences between total and MnSOD activities were assumed to represent the CuZnSOD activity. One unit of SOD activity was defined as the amount of e n z y m e inducing a 50% inhibition in the rate of epinephrine autoxidation. Total SOD, C u Z n - , and MnSOD activities were expressed as units per milligram of protein.
Table 2
RESULTS
while Mspl is not influenced by the state of methylation of the internal C [10]. DNA samples from normal and cancer tissues were digested with MspI or HpaII, electrophoresed, blotted, and hybridized with the pSP64-SOD probe. At low stringency washes, this probe hybridizes with the SOD-1 gene and with pseudogenes having partial homology with the SOD1 gene. On the other hand, at the stringency employed in this report, the pSP64-SOD probe reacts only with the SOD-1 gene [7]. MspI digestions produced several DNA fragments that point out the presence of several MspI sites along the SOD-1 gene domain. Some of these fragments appeared in DNAs from normal but not in DNAs from cancer cells, or vice versa (Fig. 1). HpalI generated high molecular weight smears characteristic of insufficient DNA digestions (Fig. 1). This finding indicates methylation of the internal cytosines in all HpaII sites along the SOD-1 domain. Mo n o n u cl ear leukocytes cultured in vitro in the presence of p h y t o h e m a g g l u t i n i n (PHA) start DNA synthesis and mitosis. This process, w h i c h resembles blastic transformation [11], usually induces an overexpression of most housekeeping genes. Because SOD-1 is a housekeeping gene [12], we analyzed its state of methylation in granulocytes, resting m o n o n u c l e a r granulocytes, and 72-hour PHA-stimulated m o n o n u c l e a r granulocytes. MspI and HpaII digestions of DNA samples from these three varieties of cells showed that SOD-1 genes were fully methylated regardless of the resting or active state of the cell population (Fig. 2).
The SOD activity in normal and tumor breast tissues is indicated in Table 1. It is clear that normal as well as malignant tissues show a marked interindividual variability in total SOD levels. Moreover, in each case, the activity of total SOD in tumor tissue is higher than in normal cells. M n - and CuZnSOD activities were determined in the last three cases of our series (Table 2). In the three individuals, it was observed that the increases in SOD activity in tumor tissues were due to increases in M n - and CuZnSOD isoenzymes. Therefore, it seems reasonable to assume that both enzymatic forms were also increased in all the breast tumors analysed. The first four donors were selected for analysis of SOD1 gene methylation. MspI and HpalI are isoschizomer enzymes that both recognize the sequence CCGG. However, HpaII is unable to cleave w h e n the inner C is methylated,
Table 1 Donor
SOD a activity in normal and tumor breast tissues b Normal
Tumor
Diagnosis Duct carcinoma Duct carcinoma Duct carcinoma Duct carcinoma Lobular carcinoma Duct carcinoma Duct carcinoma Tubular and papillary carcinoma Duct carcinoma Apocrine carcinoma Medullary carcinoma Duct carcinoma Duct carcinoma Duct carcinoma
A B C D E F G H
2.82 16.83 1.48 1.04 0.29 1.68 2.20 2.41
-+ 0.16 c -+ 0.25 d 4- 0.22 -+ 0.12 -* 0.03 - 0.05 -+- 0 . 0 2 +- 0.15
17.17 -* 0.42 21.81 *- 0.10 7.97 -*- 0.45 4.67 +- 0.06 6.08 -+ 0.05 3.16 -+ 0.19 27.60 + 0.62 3.99 - 0.23
I J K L M N
1.71 2.30 0.87 1.32 1.08 0.94
-+ 0.21 -+ 0.08 + 0.02 -+ 0.08 -+ 0.03 -+ 0.05
5.67 - 0.17 12.38 - 0.14 32.23 +- 0.43 5.19 +- 0.09 6.13 + 0.19 10.77 -+ 0.36
° Superoxide dismutase (SODI. bSOD activity is in U/mg of protein. Three independent SOD dosages were per/ormed in each case. ' Mean -+ SD. a Histopathologicanalysis showed that control tissue had infiltration with cancer cells.
MnSOD ° and CuZnSOD b activities in normal and tumor breast tissues c
Donor L M N
MnSOD CuZnSOD MnSOD CuZnSOD MnSOD CuZnSOD
Normal
Tumor
0.25 -+ 0.04'~ 1.07 -+ 0.02 0.47 -+ 0.03 0.61 -+ 0.00 0.19 4- 0.04 0.75 -+ 0.01
1.31 -+ 0.11 3.88 -+ 0.02 2.64 -+ 0.09 3.49 -+ 0.10 2.84 -+ 0.07 7.93 -+ 0.29
Diagnosis Duct carcinoma Duct carcinoma Duct carcinoma
Manganese-containing superoxide dismutase. ~:Copper- and zinc-containingsuperoxide dismutase. ': SOD activity is in U/mg of protein. Three independent SOD dosages were performed in each case. '~Mean -+ SD.
DISCUSSION We report here that t u m o r m a m m a r y tissues have higher total SOD activity and higher M n - and CuZnSOD activities than do their normal tissular counterparts. Similar findings have also been described in several other instances. In SV40-transformed WI-38 cells, SOD levels were somewhat higher than in normal WI-38 cells [13]. Moreover, total SOD activity in myelocytic, monocytic, and l y m p h o cy t i c leukemia cells was increased in regard to that observed in mature normal blood cells [14]. It is also worth mentioning that marked interindividual variations in the SOD content of 46 h u m a n tumors, i n cl u d i n g 10 breast ma-
28
M . S . Blanc:hi et al.
1
2
3
J q tp 4.4- i 2.32
L •
•
4
e,
•
o
•
•
~
~'
1R2 3 415t Re 7t819RlO 11t121131415--t16 A B C O Figure 1 Normal (n) and tumor (t) DNA samples digested with MspI (odd numbers) or Hpall (even numbers) and hybridized with the pSP64-SOD probe. A-D correspond to A-D cases in Table 1. Asterisks indicate bands that show in normal but not in tumor DNA (cases A-C) or in tumor but not normal DNA (case D). Size markers in kb are indicated in the left side of the figure. The weak smear in Hpoll digestions is due to deficient blotting of high molecular weight undigested DNA fragments.
lignancies, have been reported [15], yet, as no SOD determinations were performed in control tissues, no conclusion regarding SOD activities in cancer versus normal cells was available in this article. In the only instance in which M n - and CuZnSOD activities in normal and cancer tissues were measured, it was found that the levels of both isoenzymes showed a marked interindividual variation with no consistent differential pattern of enzyme activity in normal versus cancer tissues [16]. However, the conclusions of this article are probably made more difficult by the fact that normal and cancer samples derived from different individuals and from an ample variety of organs not always comparable. It is generally accepted that variations in the level of DNA methylation influence gene expression [see review in 17]. Bird [18, 19] has found "methylation free islands" in the genome of vertebrates. These islands occur in the 5' region of most housekeeping genes, are several h u n d r e d base pairs long, have a relative G + C richness over 50% as compared with the 40% exhibited by bulk DNA, and have GpC/CpG ratios close to I as opposed to ratios 5 - 1 0 shown by bulk DNA. Demethylation of CpG-rich islands usually correlates well with high expression of associated genes. Conversely, facultative methylation of the islands may produce underexpression or even s h u t d o w n of neighbor genes 1191. Analysis of Table 3 indicates that exon 1 of the SOD-1
-
4
Q
Figure 2 DNA from granulocytes (1), resting mononuclear leukocytes (2), or PHA-stimulated mononuclear leukocytes (3}. MspI(M) or Hpall(H) digestions. The asterisk points out an Mspl band that shows in one cell fraction and that is absent or remarkably weak in the other two cell fractions. gene and a 375 basepair long sequence located upstream of the exon 1 are CpG-rich islands. The pSP64-SOD probe used in this report hybridizes with the five exons and with sequences up- and downstream the SOD-1 gene [6, 7]. Therefore, it is logical to assume that some of the MspI Southern bands shown in Figures 1 and 2 result from fragments originated in the SOD-1 CpG rich islands. Hpall failed to cleave their canonical recognition sites in SOD-1 genes from normal and malignant tissues and from resting and PHA-stimulated blood cells. Hence, it can be conc:luded that the h u m a n SOD-1 gene is fully methylated in Table 3
Percentage of C - G and GpC/CpG ratios in the exons and 5' and 3' sequences of the h u m a n SOD-1U Gene b
Region
G + C (%)
GpC/CpG
Extent in bpc
5' sequence Exon I Exon II Exon llI Exon IV Exon V 3' sequence
61 60 47.7 51 50.4 44.7 31.9
1.2 2.3 4 4 3.5 4 15
375 69 96 72 117 105 531
" Superoxide dismutase-1. t' Table figureswere estimated from a partial gene sequencing[7]. ': Base pairs.
Superoxide Dismutase and Gene Methylation
breast and blood cells, and that interindividual and intertissue variations in SOD activity are not related to variations in SOD-1 gene methylation. Eukaryotic cells contain two distinct forms of SOD: a mitochondrial M n - c o n t a i n i n g (MnSOD) enzyme and a cytoplasmic Cu-Zn-containing (CuZnSOD) enzyme. The human SOD-1 gene located in the chromosome 21 codes for the CuZnSOD [20], while the SOD-2 gene in h u m a n chromosome 6 codes for the mitochondrial MnSOD [20]. Our results strongly suggest that both enzyme forms are responsible for SOD activity increases in tumoral tissues. Therefore, the lack of correlation between the state of SOD-1 gene methylation and the cellular SOD activity level seems to point out a lack of correlation between these two phenomena. A variability in the MspI b a n d i n g pattern between normal and tumor tissues was found in the four cases analyzed with the Southern blotting technique (Fig. 1). Malignant cell populations are of monoc:lonal origin. Accordingly, a point mutation producing the disappearance of a MspI site, or generating a new MspI recognition sequence from a cryptic site may give rise to somatic restriction fragment length polymorphisms (RFLP). Moreover, an alternative hypothesis can also explain the MspI RFLP observed. Methylation of the internal C in MspI recognition sequenc:es (CCGGI does not prevent DNA restriction by the enzyme. Conversely, when the external C is methylated, Mspl cannot cleave its recognition site [10]. DNA methylation in vertebrates preferentially occurs in the C of CpG dinucleotides [17]. Yet, methylation of the external C in CCGG sequences, although infrequent, has been demonstrated to exist in the DNA of h u m a n lymphocytes [21]. Hence, the Msp! RFLPs observed in our donors can be also explained by assuming a variable state of methylation in the external C of the sites delimiting the MspI polymorphic fragments. All cases in our series showed higher SOD activities in tumor than in normal breast tissues. On the other hand, the pattern of Mspl RFLP was not constant (donors A - C showed more MspI fragments in normal than in tumor DNA samples, while the situation was opposite for donor D) (Fig. 1). Accordingly, it seems unlikely that variations in the frequency or in the state of methylation of the external C in MspI sites may play a role in the intertissue variability in SOD content.
29
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12. 13.
14.
15.
16. We thank tile Surgery, Pathology, and Radiolngy Services of the "Hospital Interzonal General de Agudos Gral. San Martin," and we are grateful for the tecbnic:al assistance of Drs. R. Pascuale. R. Ifiigo, I,. Castelletto, and Mr. J. Padr6n. This work was supported by grants from CONICET. CIC, the Sigrid luselius Foundation. the Academy of Finland, the Folkh/ilsan Institute of Genetics, the Finnish Cancer Society, the Fundaci6n Antorchas and the Academy of Sciences of the Thircl World. A. D. Bolz&~ is a Fellow from the ('ONI('ET of Argentina.
17. 18. 19. 20.
REFERENCES 1. Lipecka K, Grabowska, B, Daniszewska K. Domanski T, Cisowska B (1984]: Correlation between the superoxide dis-
21.
mutase (SOD) activity in lymphocytes and the yield of radiation-induced chromosome aberrations. Studia Biophys 100:211-217. Larramendy ML, Bianchi MS, Padr6n J (1989): Correlation between the anti-oxidant enzyme activities of blood fractions and the yield of bleomycin-induced chromosome damage. Murat Res 214:129-136. B6yum A (1968): Isolation of leukocytes from human blood. Further observations. Stand J Clin Lab Invest 21(suppl 97):31-50. Mathew CGP (1984): The isolation of high molecular weight eukaryotic DNA. In: Methods in Molecular Biology, Nucleic Acids, Vol. 2, Walker, JM, ed. Clifton, NJ: The Humana Press, pp. 31-34. Bianchi NO, Peltom/iki P, Bianchi MS, Knuutila S, de la Chapelle A (1988): Demethylation of two specific DNA sequences in expressed human immunoglobulin light kappa constant genes. Somat Cell Mol Genet 14:13-20. Sherman L. Levanon D, l,ieman-Hurwitz ], Dafni N, Groner Y (1984): Human Cu/Zn superoxide dismutase gene: molecular characterization of its two mRNA species. Nucleic Acids Res 12:9349-9365. Levanon D, Lieman-Hurwitz J, Dafni N, et al. (1985): Architecture and anatomy of the chromosomal locus in human chromosome 21 encoding the Cu/Zn superoxide dismutase. EMBO J 4:77-84. Misra HP, Fridovich I (1972): The role of superoxide anion in the autoxidation of epinephrine and a simple assay for Superoxide Dismutase. ] Biol Chem 247:3170-3175. Misra HP (1985): Adrenochrome assay. In: Handbook of Methods for Oxygen Radical Research: Quantitation of Superoxide Dismutase, Greenwald RA, ed. CRC Press Inc., Boca Raton, F1, pp. 234-241. Nelson M. McClelland M (1989): Effect of site-specific methylation on DNA modification methyltransferases and restriction endonucleases. Nucleic Acids Res 17(suppl):R389-R415. Zackai EH, Mellman W] (1974): Human peripheral blood leukocyte cultures. In: Human Chromosome Methodology, Yunis JJ, ed. New York: Academic Press Inc., pp. 96-123. Holmquist GP (1989): Evolution of chromosome bands: molecular ec:olc~gyof noncoding DNA. J Mol Evol 28:469-486. Yamanaka NY, Deamer D (1974): Superoxide dismutase activity in WI-38 cell cultures. Effects of age, trypsinization, and SV-40 transformation. Physiol Chem Phys 6:95-106. Yamanaka NY, Ota K, tJtsumi K (1978): Changes in superoxide dismutase activities during development, aging and transformation. In: Biochemical and Medical Aspects of Active Oxygen, Hayaishi O, Asada K. eds. University Park Press. pp. 183-190. Sykes JA, McCormack Jr FX, O'Brien TJ (1978): A preliminary study of the superoxide dismutase content of some human tumors. Cancer Res 38:2759-2762. Westman NG, Marklund SL (1981): Copper- and zinc-containing superoxide dismutase and manganese-containing superoxide dismutase in human tissues and human malignant tumors. Cancer Res 41:2962-2966. Cooper DN (1983): Eukaryotic DNA methylation. Hum Genet 64:315-333. Bird AP (1986): CpG-rich islands and the function of DNA metbylation. Nature 321:209-213. Bird AP (1987): CpG islands as gene markers in the vertebrate nucleus. Trends Genet 3:342-347. Howard Hughes Medical Institute (1986): New Haven Human Gene Mapping Library. No. 1, HGM8, pp 1-76. Bianchi NO, Vidal-Rioja L, Cleaver JE (1986): Direct visualization of the sites of DNA methylation in human and mosquito chromosomes. Cbromosoma 94:362-366.
