Mammalian drug resistant mutants with multiple gene amplifications: Genes encoding the M1 component of ribonucleotide reductase, the M2 component of ribonucleotide reductase, ornithine decarboxylase, p5-8, the H-subunit of ferritin and the L-subunit of ferritin

165

Biochimica et Biophysica Acta, 1087 (1990) 165-172 Elsevier BBAEXP 92171

Mammalian drug resistant mutants with multiple gene amplifications: genes encoding the M1 component of ribonucleotide reductase, the M2 component of ribonucleotide reductase, ornithine decarboxylase, p5-8, the H-subunit of ferritin and the L-subunit of ferritin Robert A.R. Hurta and Jim A. Wright Departments of Biochemistry and Microbiology, Department of Internal Medicine and Manitoba Institute of Cell Biology, University of Manitoba, Winnipeg Manitoba (Canada) (Received 13 March 1990)

Key words: Gene amplification; Drug resistance; Ribonucleotide reductase regulation; Hydroxyurea resistance

Hydroxyurea was used to select two very highly drug resistant cell lines, designated Ha-15 and HR-30. Both drug resistant lines contained elevated levels of ribonucleotide reductase activity. Northern and Southern blot analysis indicated that the two drug resistant lines contained increased levels of mRNA for the two components, M I and M2, of ribonucleotide reductase, and M I and M2 gene amplifications. Alterations in M1 and M2 protein levels were also evident in Western blot analysis. Further studies of Ha-15 and Ha-30 cells by Northern and Southern blot analysis showed that the drug resistant cell lines had elevated levels of ornithine decarboxylase mRNA and p5-8 mRNA, as well as increased ornithine decarboxylase and p5-8 gene copy numbers, respectively. Furthermore, characterization of Hn-15 and Ha-30 drug-resistant cell lines revealed increased mRNA levels for both H- and L-ferritin. Both cell lines exhibited by Southern blot analysis, amplification of the H- and L-ferritin genes. Increases in the cellular levels of H- and L-ferritin subunit proteins were also observed in both Ha-15 and Ha-30 cells, by Western blot analysis. This is the first description of mutant cell lines containing this complex combination of modified gene expressions and gene amplifications. The alterations exhibited by these lines confirm and extend present models of hydroxyurea resistance, are in agreement with and help substantiate models of ribonucleotide reductase regulation and provide interesting links between the expressions of several cellular activities important in proliferation.

Introduction

Hydroxyurea is a specific inhibitor of DNA synthesis and arrests cell proliferation at S-phase [1,2], making the drug a useful tool in cell synchrony investigations [3]. The drug enters animal cells by a diffusion process [4] and has been used clinically in the treatment of a wide variety of solid tumors, as well as acute and chronic leukemias [5,6]. In addition, it has shown promise as a radiation potentiator [7], as a myelosuppressive agent in treating polycythemia vera [8] and in controlling the proliferation of psoriasis [9]. The major target

Correspondence: J.A. Wright, Manitoba Institute of Cell Biology, University of Manitoba, 100 Olivia Street, Winnipeg, Manitoba, Canada, R3E 0V9.

for hydroxyurea is the highly regulated enzyme ribonucleotide reductase, which is responsible for the de novo conversion of ribonudeotides to deoxyribonucleotides [1,2], required for the synthesis of DNA. The reaction is rate-limiting for DNA synthesis, and therefore, the enzyme plays an important role in the regulation of cell division [1,2]. In mammalian cells, this enzyme contains two dissimilar components often called M1 and M2 [1,2,10]. Protein M1 is a dimer with a molecular weight of 170 000 and possesses substrate and effector binding sites [11]. Protein M2 is a dimer with a molecular weight of 88 000 and contains non-heme iron and a unique tyrosyl-free radical needed for activity [12,13]. The mode of action of hydroxyurea has been shown both in vivo and in vitro to involve destruction of the tyrosyl-free radical within the M2 protein [13,14], through destabilization of the iron center [15].

0167-4889/90/$03.50 © 1990 Elsevier Science Publishers B.V. (Biomedical Division)

166 Hydroxyurea has been very useful as a selective agent in cell culture for the isolation of drug resistant cell lines with alterations in ribonucleotide reductase [1,2]. Drug resistance is a major clinical problem in the treatment of cancer [16], and the biochemical and molecular processes that lead to the development of cell populations exhibiting drug-resistant properties are of obvious interest [16-18]. In previous work, we have shown that an early and critical event in the establishment of a hydroxyurea resigtant phenotype involves changes in expression of the M2 subunit of ribonucleotide reductase, frequently through M2 gene amplification [1,2,19-21]. Higher drug resistance can be accompanied by alterations in the expression of the M1 gene, with [2,22,23] or without [1,2,13] an increase in M1 gene copy number. In addition, various drug resistant cell lines have been observed to contain elevations in expression of ornithine decarboxylase [24,25], a highly regulated activity in the polyamine biosynthetic pathway required to maintain cell growth [26], in the p5-8 gene whose function is presently unknown [24,25], and more recently in the expression of the genes encoding the heavy (H) a n d / o r light (L) subunits of the iron storage protein, ferritin [15]. In this report, we detail the molecular properties of two very highly hydroxyurea resistant cell lines, each of which exhibit clear alterations in expression of all six genes noted above (M1, M2, ornithine decarboxylase, p5-8, H-ferritin and L-ferritin), due at least in part, to the amplification of their respective genes. To the best of our knowledge, this is the first report to describe mutant cell lines containing this complex combination of altered gene expressions and gene amplifications. Materials and Methods Cell lines and culture conditions Cells were routinely cultured at 37°C on plastic tissue culture plates (Lux Scientific) in a-minimal essential medium (Flow Laboratories) supplemented with 10% fetal calf serum (Gibco) and antibiotics [19]. The procedure used to isolate mouse cell lines with increasing drug resistance characteristics have been described [27]. Starting with a wild-type population of mouse L cells, the following hydroxyurea concentrations were used in the selections: 0.35 raM; 1.3 mM; 1.5 raM; 2.0 mM; 3.0 mM; 4.0 mM; 5.0 mM; 15.0 mM (HR-15) and 30 mM (HR-30). Cells were frozen in the presence of growth medium containing 5% dimethyl sulfoxide at each selection step and those selected between 0.35 and 5.0 mM drug have been characterized in detail [21]. Nucleic acid analysis Genomic D N A was prepared from logarithmically growing cells by phenol-chloroform extraction [29]. For Southern blot analysis, 20 ~g of genomic D N A was

digested to completion with EcoRl or Hindlll restriction endonucleases, followed by fractionation on 0,75% agarose gels and subsequent transfer to nitrocellulose membranes. Total cellular R N A was isolated from logarithmically growing cells by the guanidium-cesium chloride method [30], and 20 ~g of total cellular R N A was electrophoresed through 1% formaldehyde-agarose gels followed by transfer to Nytran nylon membranes (Schleicher and Schuell). All blots were prehybridized at 42°C for 3 h in 50% v / v formamide, 5 x SSC (1 x SSC is 0.15 M NaC1, 15 mM sodium citrate, pH 7.0), 7.5 x Denhardt's solution (1 x Denhardt's solution contains 20 mg each of Ficoll, polyvinyl pyrrolidone and bovine serum albumin in 100 ml water), 50 mM sodium phosphate (pH 7.0), 0.1% sodium dodecyl sulfate (SDS), 5% dextran sulfate and 300 ~ g / m l denatured salmon sperm DNA. Hybridization was performed in the same solution for an additional t8 h with (1 2)- 106 c p m / m l of a 32p-labelled fragment containing the cDNA of interest. The /tactin probe was a 1.2 kilobase Bgll-BglI fragment from the rat /~-actin eDNA [21]. The transferrin receptor probe (kindly provided by Dr. F. Ruddle, Dept. of Biology, Yale University) was expression vector pcD-TR1 containing full length human transferrin receptor c D N A [31]. A NcoI (Boehringer-Mannheim) generated fragment containing the cDNA clone 65 (M1 protein) or the PstI (Boehringer-Mannheim) fragment of clone 10 (M2 protein) [32] were used to examine ribonucleotide reductase. The ornithine decarboxylase probe (kindly provided by Dr. A.E. Pegg, Milton S. Hershey Medical Center), was cDNA from pODC 934 [25-33] and the p5-8 probe (kindly provided by Dr. W.H. Lewis, University of Toronto) was cDNA p5-8 [24]. The heavy (H) and light (L) chain ferritin probes were provided by Dr. H. Munro (Dept. of Nutrition and Food Sciences, Massachusetts Institute of Technology). The H subunit cDNA was isolated from rat heart [34] and the L subunit cDNA was isolated from rat liver [35]. Plasmids and c D N A fragments were ~2p-labelled by nick translation and random primer labelling reactions, respectively. Blots were washed twice for 30 rain each at 25°C in 2 x SSC (1 x SSC = 0.15 M NaC1 and 15 mM sodium citrate, pH 7.0) and 0.1% SDS, then two times for 30 rain each at 60-65°C in 0.2 x SSC and 0.1% SDS. Autoradiography was performed at - 7 0 ° C with X-Omat A.R. film and Cronex lighting plus intensifying screens. Hybridized probe was stripped from the blots by washing for 60 min in 50% formamide in 0.1 x SSC at 7 0 ° C and then rinsed for 1-2 h in 0.1 × SSC. Loading was determined by probing with fi-actin c D N A [21]. Densitometric analysis on appropriate autoradiograms, in which film exposure to 32p-cDNA resulted in hybridization band intensities appropriate for direct comparisons between cell lines (not necessarily the exposures presented), was performed using a Beckman DU-8 gel scanning spectrophotometer.

167 Western blot analysis Anti-M1 mouse monoclonal antibody, AD203 [38], or anti-M2 rat monoclonal antibody, JB4 [13], was used in Western blot analysis as previously described [13,21]. Western blot analysis for H and L ferritin subunits was carried out with the following modifications [15]. Following cell extract preparations, total ,cell extract protein content was determined and then a given amount of protein was analyzed on an SDS 10%-17.5% gradient pore polyacrylarnide gel. After protein transfer and blocking, membranes were incubated with a rabbit polyclonal anti-human ferritin IgG (Boehringer-Mannheim) which recognizes both mouse and human ferritins [42]. Goat anti-rabbit IgG conjugated with alkaline phosphatase (Sigma) was used for H and L subunit detection. Results

Hydroxyurea sensitivity, elevation in ribonucleotide reductase activity and alterations in the M1 and M2 components Two hydroxyurea resistant cell lines designated H a_ 15 and HR-30 were selected in the presence of increasing concentrations of hydroxyurea as outlined in Materials and Methods, and in colony-forming assays they were found to be 68- and 103-fold, respectively, less sensitive than wild type cells to the cytotoxic effects of the selective agent (Table I). Also, compared to wild type enzyme levels, HR-15 and HR-30 cells contained elevated levels of ribonucleotide reductase activity. There were approx. 4- and 23-fold increases in the enzyme activities for HR-15 and HR-30 cells, respectively, over the wild type levels (Table I). Northern blot analysis, using M1 or M2 specific cDNA as hybridization probes showed obvious elevations of M1 and M2 message levels in the drug resistant lines relative to the wild type situation. Densitometric

measurements indicated 7- and 10-fold increases in M1 m R N A concentrations in HR-15 and HR-30 cells, respectively and indicated very large elevations in M2 mRNA, of 125- and 150-fold above the wild type levels in HR-15 and HR-30 cell lines, respectively (Table I). Southern blot analysis showed that both M1 and M2 genes were amplified in HR-15 and HR-30 cell lines. M1 gene copy number was increased by approx. 3- to 4-fold in the drug resistant fines, and the M2 gene was amplified by about 12- to 14-fold in drug resistant cells when compared to parental cells (Table I). The levels of M1 and M2 proteins were determined by Western blot analysis, and the results are provided in Table II. Densitometric measurements indicated a 3- to 5-fold increase in M1 protein level and a 30- to 40-fold increase in M2 protein level in HR-15 cells when compared to wild type cells. However, we have recently shown that cells cultured in the presence of hydroxyurea can significantly elevate the levels of ribonucleotide reductase components [13,43]. When HR-15 cells were grown in the presence of 15 mM hydroxyurea (the highest drug concentration used in its selection) for a week, large elevations in both M1 (about 10- to 15-fold) and M2 (about 80- to 90-fold) proteins were observed, when compared to the wild type and the HR-15 cells cultured in the absence of drug. M1 and M2 protein levels were also determined in HR-30 cells. The results of these experiments showed large increases of about 15- to 25-fold and at least 100-fold for M1 and M2 proteins, respectively, in the HR-30 cells compared to parental wild type cells. Alterations in ornithine decarboxylase and p5-8 message levels and gene copies The relative levels of omithine decarboxylase message in wild type, HR-15 and HR-30 cells were exarnined by Northern blot analysis, using ornithine decarboxylase specific cDNA. As reported in other

TABLE I Results from drug sensitivity, ribonucleotide reductase activity and M1 or 342 nucleic acid hybridization experiments e Cells lines

Dlo values (mM) a

CDP reductase activity b

Relative M1 RNA hybridization c

Relative M2 RNA hybridization ~

Relative M1 DNA hybridization d

Relative M2 DNA hybridization d

Wild type HR-15 Ha-30

0,15 10.20 15.50

0.98 3.79 22.60

1.0 7.0 10.0

1.0 125 150

1.0 3.0 4.0

1.0 12.0 14.0

The Dr0 values are the concentrations of hydroxyurea that reduce relative colony-forming efficiency to 105~, and was determined as described [21, 23, 281. b Enzyme was prepared and activity was determined as previously described [20,21,27,37]; activity is expressed as nmol CDP reduced per h per mg protein. a

c Determined from densitometric scanning of appropriate autoradiograms of R N A from drug-resistant and wild-type cells and expressed relative to the wild-type result. d Determined from densitometric measurements of appropriate autoradiograms of the most prominent band hybridizing with M1 or M2 eDNA from drug-resistant cells, relative to measurements of the most prominent band hybridizing with D N A from wild-type cells. Further infomaation can be found in Ref. 23.

168 Table I1 Results from Western blot analyses and nucleic acid hybridization experiments involving ornithine decarboxylase or p5-8

Cell lines

Relative M1 RelativeM2 Relative M1 Relative M2 RelativeODC Relativep5-8 RelativeODC Relativep5-8 protein protein protein level, protein level, RNA RNA DNA DNA levels a levels b drug challenge a.b drug challenge a.b hybridization d hybridization d hybridization ~ hybridization c

Wild type 1.0 HR-15 30-40 HR-30 100

1.0 3-5 15-20

1.0 80-90 n.d. ~

1.0 10-15 n.d. ~

1.0 9.0 13.0

1.0 3.0 4.0

1.0 5.0 7.0

1.0 2-3 3.0

a Determined from densitometric scanning of data from Western blots, using MI or M2 specific monoclonal antibodies and total protein from wild-type and hydroxyurea-resistant cells. Results are expressed relative to the wild-type situation. b HR15 cells were grown in the presence of 15 mM hydroxyurea; the values shown were determined from densitometric evaluation of appropriate Western blots and expressed relative to the wild-type situation (grown in the absence of hydroxyurea). c n.d. = not determined. d Determined from densitometric scanning of appropriate Northern blot autoradiograms of RNA from wild-type or drug-resistant cells and expressed relative to the wild-type result. Experiments were performed with ornithine decarboxylase (ODC) or p5-8 cDNA. e Determined from densitometric measurements of appropriate Southern blot autoradiograms of the most prominent hybridization bands and expressed relative to the wild-type result. Experiments were performed with ornithine decarboxylase (ODC) or p5-8 cI)NA

studies with m a m m a l i a n cells [24,25,44], several hybridizing bands were observed in the three cell lines, which were increased in intensity in the h y d r o x y u r e a resistant cells when compared to the wild type line, to approx. 9- and 13-fold in Hr~-15 and HR-30 cells, respectively (Table II). M a n y hybridization bands were observed in Southern blot analysis of D N A from wild type, HR-15 and HR-30 cell lines, in keeping with previous findings that more than one sequence homologous to the ornithine decarboxylase gene is present in m a m m a l i a n cells [24,25,44]. Amplification of the ornithine decarboxylase gene was clearly evident and densitometric estimates showed about 5-fold and 7-fold increases in ornithine decarboxylase gene copy numbers in HR-15 and HR-30 cells, respectively (Table II). N o r t h e r n blot analysis of p5-8 message levels indicated increases of approx. 3-fold and 4-fold in Ha-15 and HR-30 cells relative to wild type cells, respectively (Table II). Furthermore, densitometric scans of hybridizing bands in Southern blots provided estimates of 2.5-3-fold amplification of the p5-8 gene in HR-15 and H R-30 cells relative to parental cells, respectively (Table II). H and L ferritin message levels a n d gene copies

c D N A s specific for heavy (H) and light (L) subunits of rat ferritin were used as specific hybridization probes in N o r t h e r n and Southern blot analyses of message and relative gene copy numbers in wild type, HR-15 and HR-30 cells. The N o r t h e r n blot shown in Fig. 1A indicated that the H-chain ferritin c D N A hybridizes to an m R N A species of about 1.1 to 1.2 kb, which is elevated in drug-resistant cells. Densitometric analysis estimated a 4-fold and a 6-fold increase in H-ferritin m R N A levels in HR-15 and HR-30 cells, respectively. The Lchain ferritin c D N A was also found to hybridize to a 1.1 to 1.2 kb m R N A species in the three cell lines (Fig.

1B). Densitometric analysis indicated approx, a 6- and a 8-fold increase in L-ferritin m R N A in HR-15 and H~-30 cells, respectively. T o evaluate if the elevated levels of ferritin H and L m R N A s in variant cells were due to gene amplification, Southern blot analyses were performed (Fig. 2A and B; Fig. 3A and B). At the wash stringency used (0.2 × SSC and 0.1% SDS at 60°C) it has been shown that the H- and L-ferritin c D N A s do not cross-hybridize [45]. In agreement with other studies [15], complex banding patterns with multiple hybridization sequences were detected, some of which were likely due to the presence of ferritin pseudogenes in mammalian cells [46]. Evidence for amplification of both H-chain and L-chain ferritin genes was obtained with HR-15 and HR-30 cells, when the intensities of bands were c o m p a r e d to the wild type situation. Densitometric analysis of amplified hybridizing bands indicated about 3- and 4-fold increases in H-ferritin gene copies in HR-15 and HR-30 cell lines, respectively. Similarly, approx. 4- and 5-fold amplifications of the L-ferritin genes were found in HR-15 and HR-30 cell lines when compared to the wild type cells. N o evidence of H- or L-ferritin gene rearrangements were detected. Like ferritin, the expression of the transferrin receptor in proliferating cells is regulated by iron availability [46-48]. T o determine if, as with H- and L-ferritin, selection for resistance to h y d r o x y u r e a leads to changes in transferrin receptor m R N A levels, a h u m a n transferrin receptor c D N A [31], was used as a specific molecular p r o b e for N o r t h e r n blot analysis. The transferrin receptor c D N A detected one prominent m R N A species with an a p p r o x i m a t e size of 5 kb in b o t h the wild-type and the drug-resistant cell lines, HR-15 and HR-30. However, no significant changes in the levels of transferrin receptor m R N A were found when wild type cells were c o m p a r e d to drug-resistant cell lines (data not shown).

169

-18s

a

b

c

a

A

b

c

B

Fig. 1. (A): Northern blot of H-ferritin mRNA in (a) wild type, (b) H•-15 and (c) HR-30 cell lines. (B): Northern blot of L-ferritin mRNA in (a) wild-type, (b) HR-15 and (c) Ha-30 cell lines. 20/xg of total cellular R N A isolated from wild-type and hydroxyurea resistant cell lines were run on 1% agarose formaldehyde gels and hybridized with H- or L-ferritin c D N A as indicated in Materials and Methods. The positions of 28 S and 18 S rRNA are shown.

H-ferritin and L-ferritin subunit protein levels - 2 3.1

a

i;:ii

i

b

c

A

if!fill ~

a

b

c

To determine if the elevation in ferritin H - m R N A and L - m R N A present in HR-15 and HR-30 cells results in a corresponding increase in H- and L-subunit proteins, a polyclonal antiferritin IgG was used to perform a Western blot analysis. This anti-human ferritin antibody recognizes both mouse and human H- and L-ferritin subunits [42]. Fig. 4 shows that two characteristic ferritin bands with molecular masses of 24.5 and 18 kDa were observed. Previous studies have concluded that the 24.5 k D a band corresponds functionally to the L-subunit and the 18 kDa band corresponds to the H-subunit of ferritin [49,50]. Clear increases in H- and L-ferritin subunit proteins were observed in the two variant lines compared to parental wild type cells. Densitometric analyses of the 18 kDa band revealed approx. 2.5-3 and 3-4-fold increases in the H-subunit protein of ferritin in the HR-15 and HR-30 variants, respectively, relative to their wild-type line. Densitometric analysis of the 24.5 kDa band revealed approx. 2-fold and 2.5-3-fold increases in the L-subunit protein of ferritin in the HR-15 and Ha-30 variants, respectively, relative to their wild-type line.

B

Fig. 2. Southern blot of H-ferritin genes in genomic D N A of the wild type and drug resistant cell lines. 20 /xg of high-molecular-weight D N A were digested to completion with Eeo R1 (A) or Hind III (B). Blots were hybridized with a 32p-labelled H-ferritin e D N A as described in Materials and Methods. (A): (a) wild type, (b) HR-15 and (c) HR-30; (B): (a) wild type, (b) HR-15 and (c) HR-30.

Discussion It has been well documented that the specific site of action for hydroxyurea is at the M2 subunit of ribonucleotide reductase [1,2,15]. The present study shows that

170 even though the mode of action of hydroxyurea is quite specific, a relatively large number of gene alterations can accompany conversion of a drug-sensitive population of cells to a drug-resistant population. These changes not only involve alterations in the expression and copy number of the M2 gene of ribonucleotide reductase, but also the M1 gene of the reductase, and the genes encoding ornithine decarboxylase, p5-8, the H-subunit of ferritin and the L-subunit of ferritin. Indeed, this may be a general principle for drug-resistance mechanisms in mammalian cells. Changes involving the expression of a gene encoding the target protein for a drug, will likely necessitate alterations in the expression of other genes, which passively or directly play a role in the drug resistance mechanism. Therefore, it is important to comment on which changes described in the present investigation are required directly for drug resistance and which are indirectly or passively involved. The changes, in ribonucleotide reductase activity, M2 m R N A and relative M2 gene copy number are in keeping with the finding that hydroxyurea destroys the tyrosyl free radical of the M2 component by destabilizing the non-heme iron centre in the protein [15], and in

13 7 5 3

Kd a

b

c

Fig. 4. Western blot analysis of H- and L-ferritin in wild-type and drug-resistant variants. Electrophoresis, transfer to nitrocellulose and immunostaining were performed as described in Materials and Methods. Lanes are: (a) HR-30, (b) HR-15 and (c) wild type cells. 50/~g cell extract protein was loaded in each lane. The apparent molecular mass (kDa) of the two ferritin subunits is indicated on the right. The molecular weight marker standards used were prestamed SDS-PAGE standards for Western blotting of low-molecular-weight ranges (17130 kDa) (Bio-Rad).

-23.1 - 9,4

-6.7 -4.3

-2.3 -2.0

Kb

a

b

A

c

a

b

c

B

Fig. 3. Southern blot analysis of L-ferritin genes in genomic D N A of the wild type and drug-resistant cell lines. 20 #g of high-molecularweight D N A were digested to completion with Eco R1 (A) or Hind III (B). Blots were hybridized with a 32p-labelled L-ferritin cDNA as described in Materials and Methods. The lanes are (A): (a) wild type, (b) HR-15 and (c) HR-30; and (B): (a) wild type, (b) HR-15 and (c) HR-30-

the observations that the level of the M2 protein is normally rate-limiting for ribonucleotide reductase activity and D N A synthesis [1,2,19,21,43]. Therefore, modifications in M2 gene expression leading to increased M2 protein play an important direct role in determining resistance to hydroxyurea [19-21,43], and the molecular changes described for HR-15 and HR-30 cells support this model. However, it was also clear from the studies with HR-15 and HR-30 cells, that at very high drug resistance, modest changes in M1 protein levels were required, which was accomphshed in these cell lines by increasing M1 message through M1 gene amplification. This suggests that at very high drug concentrations increased M2 gene expression is not sufficient on its own to allow cell proliferation, because in this unique situation, the M1 component has become limiting for enzyme activity and DNA synthesis. Therefore, modifications in M1 gene expression become necessary for survival. The findings that the increases in M1 and M2 message levels were higher than the increases in gene copies indicate that other mechanisms [13,21], in addition to gene amplification are working to regulate the increased message levels in these drug-resistant cells. Furthermore, the dramatic elevation in M1 and M2 protein levels in Hn-15 cells cultured in the presence of hydroxyurea (Table I) is in agreement with our previous observations that this drug can modulate ribonucleotide reductase expression post-transcriptionally by modifying the half-life and biosynthetic rates of these two proteins [13,43].

171 Recent studies with hydroxyurea resistant lines have suggested that the genes for omithine decarboxylase and p5-8 are part of a single amplicon linked to the M2 gene of ribonucleotide reductase [24,25]. Gene coamplification may play an important role in drug resistance and can provide information relevant to mechanisms of gene amplification [51,52]. Our studies clearly support this idea, since elevations in message levels and gene copies were observed for these three genes in both drug-resistant lines. Although ornithine decarboxylase and the product of the p5-8 gene probably play passive roles through co-amplification, as opposed to direct roles in establishing the hydroxyurea resistant phenotype in mammalian cells, further investigations are required. For example, interesting similarities exist between the M2 and omithine decarboxylase proteins and their regulation to warrant further studies. Both the ornithine decarboxylase and M2 proteins are highly controlled, rate-limiting steps in biosynthetic pathways important in cellular proliferation. Regulation of the M2 protein varies significantly with growth conditions, is limiting for ribonucleotide reductase enzyme activity and subsequently DNA synthesis [1,2,19,21,43]. M2 has been suggested to be a possible candidate belonging to a small group of proteins directly involved in regulation of cell cycle progression [53]. On the other hand, ornithine decarboxylase is a key enzyme in the biosynthesis of polyamines [26], and these compounds have been implicated in the control of mammalian cell proliferation [26,54]. Molecular genetic studies suggest that regulatory mechanisms controlling M2 and ornithine deearboxylase during cell growth have some interesting similarities [25]. These observations suggest that further studies into the regulation of these co-amplified genes in hydroxyurea resistant variants, particularly ornithine decarboxylase, would be worthwhile. Although iron is an essential constituent of living cells, under physiological conditions, iron is so insoluble that special proteins, the ferritins, are required to maintain it in soluble and readily available form [46]. Since we have recently shown that hydroxyurea inactivation of protein M2 destabilizes the iron centre leading to iron release [15], we have examined the relationship between hydroxyurea resistance and ferritin gene expression. Very recently, we have observed elevated levels of H-ferritin a n d / o r L-ferritin message in a variety of variant cell lines, without changes in the corresponding ferritin gene copy numbers [15]. The results with the HR-15 and HR-30 lines are in keeping with these studies, and also show for the first time that increased ferritin gene expression can occur through a mechanism of gene amplification. Interestingly, selection for hydroxyurea resistance has led to the first description of somatic cell mutants altered in ferritin gene expression. The ferritins are highly controlled at the levels of transcription, translation and post-translation [15,45-47]

and the hydroxyurea resistant lines provide an unusual opportunity to investigate regulation of these important proteins further. We propose that the changes in ferritin gene expression described in this report are required in establishing resistance to high levels of hydroxyurea. Recent studies with revertants of drug-resistance in mouse and hamster cells further support this point (data not shown). During selection in hydroxyurea, destabilization of the M2 iron centre would lead to release of iron which is toxic for cells [15]. Ferritin would play a role in the detoxification process [46], and therefore, it would be advantageous for cells with the high levels of M2 protein required for resistance ~to hydroxyurea to also contain elevated levels of ferritin proteins, as was observed with the HR-15 and HR-30 cell lines. Even though protein M2 is the primary target for hydroxyurea, we suggest that changes in ferritin expression occur simultaneously with changes in M2 expression to achieve resistance to hydroxyurea. Additional evidence supporting this proposal is available [15]. Furthermore, it has been demonstrated that in the presence of a reductant, ferritin can release Fe 2+, which can facilitate the generation of free radicals [55]. Therefore, it is tempting to suggest that ferritin may also play an important role, perhaps indirectly, in providing the iron needed for generation of the tyrosyl free radical during the biosynthesis of a functional M2 protein. If this is correct, it Would provide a novel mechanism for the involvement of ferritin in the regulation of ribonucleotide reduction and therefore D N A synthesis. Clearly, this link between altered expression of ribonucleotide reductase and ferritin genes merits further investigation. In summary, we describe the novel properties of two highly hydroxyurea resistant cell lines altered in gene expression and gene copy numbers of several interesthag, highly controlled cellular activities, which play key roles in the regulation of biosynthetic pathways clearly important for cell proliferation. The biochemical and molecular changes exhibited by these variant cell lines confirm and extend present models of hydroxyurea resistance, are in keeping with and help substantiate models of ribonucleotide reductase regulation and establish interesting relationships between the expressions of several activities important in cell growth. Also, the work described in this report emphasizes the potential of using highly hydroxyurea resistant lines, with these unique properties, for investigations into key regulatory processes.

Acknowledgements Financial support for this work was provided by the National Cancer Institute of Canada and the Natural Sciences and Engineering Research Council of Canada (to J.A.W.). We thank the Sellers Foundation, Depart-

172

ment of Medicine, University of Manitoba for a postdoctoral fellowship for R.A.R.H. Also, we acknowledge and thank Arthur Chan for excellent technical advice. J.A.W. is a Senior Research Scientist of the National Cancer Institute of Canada. References 1 Wright, J.A. (1989) Int. Encyclopedia of Pharmacol. and Therapeut. 128, 89-111. 2 Wright, J.A., McClarty, G.A., Lewis, W.N. and Srinivason, P,R, (1989) Drug Resistance in Mammalian Cells, Vol. 1, Antimetabotite and cytotoxic analogues, CRC Press Inc., Boca Raton, pp 15-27. 3 Ashihara, J. and Baserga, R. (1979) Methods Enzymol. 58, 259261. 4 Tagger, A.Y., Boux, J. and Wright, J.A. (1987) Biochem. Cell Biol. 65, 925-929. 5 Engstrom, P.F., Maclntyre, J.M., Mittelman, A. and Klaassen, D.J. (1984) Am. J. Clin. Oncol. 7, 313-318. 6 Bolin, R.W., Robinson, W.A., Sutherland, J. and Hamman, R.F. (1982) Cancer 50, 1683-1686. 7 Piver, M.S., Barlow, J.J., Vongtania, V. and Blumenson, L. (1983) Am. J. Obstet. Gynecol. 147, 803-808. 8 Donovan, P.B., Kaplan, M.E., Goldberg, J.D., Tatarsky, I., Najean, Y., Stilberstein, E.B., Knospe, W.H., Lazlo, J., Mack, K., Berk, P.D. and Wasserrnan, L.R. (1984) Am. J. Hematol. 17, 329-334. 9 McDonald, C.J. (1981) Pharmacol. Ther. 14, 1-24. 10 Hopper, S. (1972) J. Biol. Chem. 247, 3336-3340. 11 Thelander, L., Eriksson, S. and Akerman, M. (1980) J. Biol. Chem. 255, 7426-7432. 12 Thelander, M., Graslund, A. and Thelander, L. (1985) J. Biol. Chem. 260, 2737-2741. 13 McClarty, G.A., Chan, A.K., Engstrom, Y., Wright, J.A. and Thelander, L. (1987) Biochemistry 26, 8004-8011. 14 Graslund, A., Ehrenberg, A. and Thelander, L. (1982) J. Biol. Chem. 257, 5711-5715. 15 McClarty, G.A., Chan, A.K., Choy, B.K. and Wright, J.A, (1990) J. Biol. Chem. 265, 7539-7547. 16 Stark, G.R. (1986) Cancer Surv. 5, 1-23. 17 Stark, G.R., Debatisse, M., Giulotto, E. and Wahl, G.M. (1989) Cell 57, 901-908. 18 Schimke, R.T. (1988) J. Biol. Chem. 263, 5989-5992. 19 Wright, J.A., Alam, T.G., McClarty, G.A., Tagger, A.Y. and Thelander, L. (1987) Somat. Cell Mol. Genet. 13, 155-165. 20 Tagger, A.Y. and Wright, J.A. (1988) Int. J. Cancer 42, 760-766. 21 Choy, B.K., McClarty, G.A., Chan, A.K., Thelander, L. and Wright, J.A. (1988) Cancer Res. 48, 2029-1035. 22 Cocking, J.M., Tonin, P.N., Stokoe, N.M., Wensing, E.J., Lewis, W.H. and Srinivason, P.R. (1987) Somat. Cell Mol. Genet. 13, 221-233. 23 Hurta, R.A.R. and Wright, J.A. (1990) Biochem. Biophys. Res. Commun. 167, 258-264.

24 Srinivasan, P.R., Tonin, P.R., Wensing, E.J. and Lewis, W.H. (1987) J. Biol. Chem. 262, 12871-12878. 25 McClarty, G.A., Tonin, P.R., Srinivasan, P.R. and Wright, J.A. (1988) Biochem. Biophys. Res. Commun. 154, 975-981. 26 Pegg, A.E. (1988) Cancer Res. 48, 759-774. 27 McClarty, G.A., Chan, A. and Wright, J.A. (1986) Somat. Cell Mol. Genet. 12, 121-131. 28 Hards, R.G. and Wright, J.A. (1981) J. Cell. Physiol. 106, 309-319. 29 Bin, N. and Stafford, D.W. (1976) Nucleic Acids Res. 3. 23032308. 30 Chirgwin, J.M., Przybyla, A.E., MacDonald, R.J. and Rutter, W.J. (1979) Biochemistry 18, 5294-5299. 31 Kuhn, L.C., McClelland, A. and Ruddle, F.H. (1984) Cell 39, 267 274. 32 Thelander, L. and Berg, P. (1986) Mol. Cell. Biol. 6, 3433 3442. 33 Madhubala, R., Secrist, J.A., Ill and Pegg, A.E. (1987) Fed. Proc. 46, 2254. 34 Murray, M.T., While, K. and Munro, H.N. (1987) Proc. Natl. Acad. Sci. USA 84, 7438-7442. 35 Leibold, E.A., Aziz, N., Brown, A.J.P. and Munro, H.N. (1984) J_ Biol. Chem. 259, 4327-4334. 36 Steeper, J.P. and Steuart, C.D. (1970) Anal. Biochem. 34, 123 130. 37 Choy, B.K., McClarty, G.A. and Wright, J.A. (1989) Biochem. Biophys. Res. Commun. 162, 1417 1424. 38 Engstrom, Y., Rozell, B., Hansson, H.A., Stenne, S. and Thelander, L. (1984) EMBO J. 3, 863-867. 39 Laemmli, W.K. (1970) Nature 227, 680-685. 40 Towbin, H., Stachlin, T. and Gordon, J. (1979) Proc. Natk Acad. Sci. USA 76, 4350-4354. 41 Blake, M.S., Johnston, K.H., Russel-Jones, G.J. and Gotschliech, E.C.A. (1984) Anal. Biochem. 136, 175-170. 42 Rouault, T.A., Hentze, M.W., Dancis, A., Caughman, W., liarford, J.B. and Klausner, R.D. (1987) Proc, Natl. Acad. Sci. USA 84, 6335-6339. 43 McClarty, G.A., Chan, A.K., Choy, B.K., Thelander, 1.. and Wright, J.A, (1988) Biochemistry 27, 7524-7531. 44 Feinberg, A.P. and Volgelstein, B. (1983) Anal. Biochem. 132, 6-13. 45 Aziz, N. and Munro, H.N. (1986) Nucleic Acids Res. 14, 915 927. 46 Theil, E.C. (1987) Annu. Rev. Biochem. 56, 289 315. 47 Bomford, A.B. and Munro, H.N. (1985) Hepatology 5, 870-875. 48 Casey, J.L., DiJeso, Rao, K., Klausner, R.D. and Harford, J.B. (1988) Proc. Natl. Acad. Sci. USA, 85, 1787-1791. 49 Beaumont, C., Jain, S.K., Bogard, M., Mordmann, Y. and Drysdale, J. (1987) J. Biol. chem. 262, 10619-10623. 50 Beaumont, C., Dugast, I., Renaudie, F., Souroujom M. and Grandchamp, B. (1989) J. Biol. Chem. 264, 7498-7504. 51 Schimke, R.T. (1984) Cancer Res. 44, 1735-1742. 52 Debatisse, M., Hyrien, O., Petit-Koskas, E., De Saint-Vincent, B. and Buttin, G. (1986) Mol. Cell. Biol. 6, 1776-1781. 53 Eriksson, S. and Martin, D.W., Jr. (1981) J. Biol. Chem. 256, 9436-9440. 54 Pohjanpelto, P., Holtta, E. and James, O.A. (1985) Mol. Cell. Biol. 5, 1385-1390. 55 Thomas, C.E. and Aust, S. (1986) J. Biol. Chem. 261, 13064-13070.