Harvey ras Results in a Higher Frequency of ... - Cancer Research

[CANCER RESEARCH 58, 5097-5104.

November 15, 1998]

Harvey ras Results in a Higher Frequency of Mammary Carcinomas than Kirsten ras after Direct Retro viral Transfer into the Rat Mammary Gland1 Todd A. Thompson, Kwanghee Kim, and Michael N. Gould2 Department of Human Oncolog\ [T. A. T., K. K., M. N. G. I. Department of Oncology, McArdle Laboratory for Cancer Research ¡M.N. G.], and Environmental [T. A. T., M. N. G. I, University of Wisconsin-Madison, Madison, Wisconsin 53792

Toxicology Center

observed in rat colon carcinogenesis (7). Selective ras gene activation is also associated with many human cancers. For example, activation Exclusive activation of either the Harvey-, Kirsten-, or N-ros gene is of the Harvey raÃ-gene is observed in human bladder cancer (8); often found in human and rodent cancers, although the mechanisms responsible for tissue-specific ras gene activation are poorly understood. Kirsten ras gene activation is found in human colon cancer (9, 10) and pancreatic cancer (11); and N-ras gene activation is found in acute In this study, the contribution of ras gene expression and Ras protein activity to the tissue-specificity of ras gene activation was investigated myeloid leukemia (12). The mechanisms responsible for tissueusing the rat mammary carcinogenesis model where ras activation, when specific activation of the raÃ-gene family members are poorly under it occurs, is exclusively in the Harvey ras gene. Differential ras gene stood. Therefore, a more complete evaluation of the differences in expression was examined in mammary tissue from virgin, pregnant, and biological function of the ras family members would aid in under lactating rats. Harvey ras expression was 1.5-2-fold higher than Kirsten standing their role in cancer development. ra.vor ¡Vra.v at each adult stage of development, with the highest ras levels Expression of the independent members of the ras family of genes expressed during pregnancy. The modest difference in total inRN'Aex follow both a qualitative and quantitative tissue-dependent pattern. pression found between the independent members of the ras gene family For example, Leon et al. (13) found that in mice, the Harvey-, is unlikely to fully account for the exclusive tissue-specificity of Harvey ra.v Kirsten-, and N-ra.v genes are expressed in all tissues, but the relative activation observed in rat mammary carcinogenesis. Thus, the role of Ras levels of each form of raÃ-varies in a tissue-dependent manner and the protein specificity was studied by infecting the mammary gland of virgin total levels of raÃ-expressed differs between tissues. In the rat mam rats in situ with replication-defective retroviral vectors expressing either the activated or wild-type forms of Harvey- or Kirsten-ras. A mary gland, differential expression of the Harvey raÃ-gene has been 7-14-fold higher number of mammary carcinomas was observed after reported to vary depending on the differentiation status of the gland infection with vectors expressing the G35 to A activated Harvey ras (14). Interestingly, in carcinogenesis studies where raÃ-is found acti gene product compared with those expressing G35 to A activated vated, the ra.s gene family member found mutated frequently corre Kirsten ras. Mammary carcinomas also developed from infusion of lates with the form of raÃ-predominantly expressed within the tissue vectors expressing wild-type Harvey ras, but not wild-type Kirsten ras. of origin (15). Thus, tissue-dependent expression of the raÃ-genes has These data suggest the importance of the Ras protein itself in deter been hypothesized to play a role in the organ-specificity associated mining the specificity of the highly homologous Ras family members in with raÃ-gene activation. organ-specific carcinogenesis. Although the four members of the Ras proteins are highly homol ogous, their COOH termini vary, which could allow for distinct functional activities (1). For example, in vitro analyses have demon INTRODUCTION strated differences in the efficiency of posttranslational modifications The Ras family of proteins consists of four primary members: that occur at the COOH terminus for Harvey Ras and Kirsten Ras (16, Harvey Ras, Kirsten Ras 2A, Kirsten Ras 2B, and N-Ras. Ras proteins 17). Also, distinctions between the raÃ-genes are evident in knockout studies of mice, where deletion of the N-rai gene is without apparent are highly homologous, evolutionarily conserved Mr 21,000 proteins that are bound to guanosine 5'-triphosphate in their active form and consequence (18), whereas Kirsten raÃ-knockouts are developmenguanosine 5'-diphosphate in the inactive state, which function as tally lethal (19). Thus, differences in the roles of the independent ra.v intermediates in signal transduction (reviewed in Ref. l). Mutations gene family members at the protein level could account for the tissue-dependent functions of ra.v in carcinogenesis. that result in the activation of the Ras proteins are found in many different cancers, supporting an active role of raÃ-in cancer develop Retroviral gene transfer has been used extensively for the introduc ment (reviewed in Ref. 2). Interestingly, within cancers of a particular tion of novel genes to many different tissues, including the rat mam tissue or organ, these activating events are often restricted to a single mary gland (20). Infection of mammary parenchyma with retroviral ras gene family member (3). vectors expressing the viral Harvey ra.v gene produces mammary Tissue-specific ras gene activation is frequently observed in rodent carcinomas that are similar in morphology and temporal development carcinogenesis models (3). For example, the Harvey ra.v gene, but not to chemically induced mammary carcinomas (21). In this study, a the Kirsten ras gene, was found mutated in mammary carcinomas retroviral mammary ductal infusion methodology was used to intro from rats exposed to NMU3 during sexual development (4-6). In duce Harvey raÃ-and Kirsten ra.v genes to the rat mammary paren contrast, NMU-induced mutations of the Kirsten ras gene have been chyma under identical promoters, allowing the determination of the relative potency of the Harvey ra.v and Kirsten ra.v gene products in rat Received 5/20/98; accepted 9/17/98. mammary carcinogenesis. ABSTRACT

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. 1 Supported by NIH Grants CA77527 and CA44387 and Predocloral Fellowship DAMD17-94-J-4I04 (to T. A. T.) from the United States Army Medical Research Mate rial Command. 2 To whom requests for reprints should be addressed, at University of Wisconsin-

Cloning, Mutagenesis, and Sequencing. Wild-type rat Harvey ras l cDNA and rat Kirsten ras 2B cDNA were PCR-amplified from a Sprague

Madison, Department of Oncology, CSC K4/334, 600 Highland Avenue, Madison, WI 53792. 'The abbreviations used are: NMU. /V-methyl-A'-nitrosourea; LTR. long terminal repeat: GAPDH, glyceraldehyde-3-phosphate dehydrogenase: TPA. 12-O-tetradeconylphorbol-13-acetate: CPU, colony-forming unit.

Dawley rat brain cDNA library (Stratagene, La Jolla, CA) and subcloned using a TA cloning kit (Invitrogen, Carlsbad, CA). The 5' primer (atgacagaatacaagcttgtggtgg) and 3' primer (tcaggacagcacacacttgcagc) used to amplify rat Harvey ras spanned the entire coding region. The 5' primer (atacaagcttgtggtagttggagct-

5097

MATERIALS

AND METHODS

raÃ- TRANSFER

INTO THE MAMMARY

Imager (Molecular Dynamics, Sunnyvale, CA), and analyzed using ImageQuant software (Molecular Dynamics). The probes used included a 227-bp Harvey ras probe spanning codon 75-codon 150 (exon 3), a 174-bp Kirsten ra.v probe encompassing codon 74-codon 132, a 295-bp N-ra.v probe from codon 45-codon 144, and a 122-bp rat /3-actin probe. Retroviral Vector Construction and Production. The wild-type and codon 12-activated forms of the Harvey ras and Kirsten ra.v gene were

pJK backbone

ir rlrexonl

Hrexon2

HrcxolO

llr excm4

H-TO.Vconstructs

JRHrasG

35

subcloned into the BamHl and Sail restriction sites of pJR (Fig. 1), producing the plasmids used to generate the replication-defective retroviral vectors used

JRHrasA

in this study. The construction of JRHrasV and JRgal were described previ ously (21) and used in these studies as positive and negative retroviral infusion controls, respectively. For the purpose of this study. JRras (21 ) was redesig-

JRHrasV

nated JRHrasV to avoid confusion with the other vectors used in this study. Each retroviral vector plasmid was independently transfected into the ecotropic packaging cell line ^f-CRE (24) using Lipofectin (Life Technologies, Inc.,

175

Kr «on1 llr «unll Kr exon 2 Kr «

«on4B

GLAND

Gaithersburg, MD). Retroviral vectors from these cells were used to infect the amphotropic packaging cell line PA317 (25) to produce replication-defective

K-rai constructs

retroviral vectors for rat mammary gland infusions. Infected PA317 clones were selected using resistance to G418 (Life Technologies. Inc.). High-liter producing clones were expanded in 162 cm2 cell culture flasks and grown at

JRKrasG JRKrasA

Virgin

JRKrasV 35

Pregnant

Lactating

N-TO.Ç

175 ll-calaclosidase

Cont nil construct

JRgal

Harvey-raj

Fig. 1. Relroviral vector constructs. The pJR backbone (21 ) was used for construction of all TÕÕ.V expression vectors as described in "Materials and Methods." JRHrasG contains the wild-type Harvey ru\ coding region; JRHrasA contains the G35 to A-activated form (if the Harvey ms gene; JRKrasG contains the first 84 bp of the wild-type Harvey ras gene followed by the coding region for the Kirsten ras 2B gene; JRKrasA contains the first 84 hp of the G35 to A mutated form of the Harvey ras gene followed by the coding region for the Kirsten ras 2B gene; JRKrasV contains the viral Kirsten ras gene (containing exon 4A ). M7*, indicates the region of the retroviral packaging signal; TAG. specifies a mutation in the ppoO^'1*initiation codon; A. denotes the location of activating point mutations found

Kirsten-rov

in the ras genes. Constructs are not drawn to scale.

•¿M

ß-actin ggtggcg) used to clone Kirsten ra.v 2B contained a mutation that created a Hini3\\\ restriction site analogous to that found in the Harvey ra.v gene to assist in further subcloning without altering the resultant amino acid sequence. The 3' primer sequence (tcacatgactatacaccttg) was derived from the viral Kirsten ra.v region corresponding The rat Kirsten ra.v 2B pSP73 vector (Promega, viral Harvey rus gene.

B

to the termination codon of Kirsten ra.v exon 4B (22). sequence was subcloned into the WmdIII site of an Madison. WI| containing upstream sequence of the Site-directed mutagenesis of the rat Kirsten ras se

0.4 -0.3

quence was performed by replacing a //jndlll to DralH fragment with an oligonucleotide containing an activating G35 to A mutation in codon 12 of Harvey rax. resulting in the substitution of glutamic acid for glycine. The viral Kirsten ra.v sequence was subcloned as a C.v/?45I to Stul fragment from the pKSma vector (American Type Culture Collection, Manassas. VA) and placed in the Clti\ to Smal site of pBluescript II SK+ (Stratagene). ras clones were sequenced using either Sequenase (United States Biochemical Corp., Cleve land, OH) or AmpliTaq FS (Applied Biosystems, Foster City, CA). All subcloning procedures used in these studies were standard protocols (23). ras Gene Expression. The RPA II RNase protection assay kit (Ambion, Austin, TX) was used to determine the levels of Harvey-, Kirsten-, and N-ra.v expression in 50 day-old virgin. 15-day pregnant, and 21-day lactating WistarFurth and Copenhagen rat mammary glands. No difference was observed in ras expression from Wistar-Furth and Copenhagen rat mammary glands, therefore, data from these two strains were pooled for analysis. Radiolabeled-RNA probes were synthesized with the T7 Maxiscript kit (Ambion) and [a-'2P|UTP (New England Nuclear, Boston, MA). The full-length probes were purified, hybridized with 10 /xg of total RNA from tissue samples isolated using RNAzolB (Tel-Test, Friendswood, TX), and digested with RNase after the manufacturer's instructions. The protected fragments were resolved on a 5% sequencing gel, and exposed to a phosphor screen, scanned using a Phosphor-

Harvey-roÃ--r* 1 Klrsten-rasT*As!i-L V//A N-ru.vi|X!¿*i^B111 m^

-0.2

-0.1

-

n n -1 Virgin

Pregnant

Lactating

Fig. 2. Analysis of Harvey-, Kirsten-, and N-m.v mRNA expression in the mammary gland of virgin, pregnant, and lactating rats. A, RNase proteelion assay for N-rav, Harvey ras. Kirsten ras, and ß-actinmRNA expression as described in "Materials and Methods." B, histogram showing results from the RNase protection assay normalized to ß-actin expression. *, Harvey ras mRNA levels were significantly higher than Kirsten ras and N-ras levels in the virgin, pregnant, and lactating mammary gland (P < 0.05). Columns, means of two experiments done in triplicate; bars, SE.

5098

ras TRANSFER

INTO THE MAMMARY

GLAND

injected s.c. for 2 consecutive days before and 4 h before viral infusion. Viral preparations were thawed on ice, and then polybrene (Sigma Chemical Co.) was added to a final concentration of 80 fig/ml. Either 3 mg/ml indigo carmine (Sigma Chemical Co.) or 0.5 mg/ml fast green (Fluka, Ronkonkoma. NY) was added to visually assess the completeness of viral infusion into the entire mammary ductal structure. Under ether anesthesia, the thoracic, abdominal, and inguinal regions of each rat were shaved to expose the 12 mammary teats. Teats were individually clipped to allow the infusion of 12 ¡uor 20 ^1 of viral preparation into the central mammary lacteal using a blunt-ended 27-gauge needle. Occasionally,