Strong synergy between mutant ras and HPV16 E6/E7 in the development of primary tumors

Abstract

E6/E7 oncogenes of high-risk human papilloma virus (HPV) subtypes are essential for the development of certain types of cancers. However, these oncogenes are insufficient to transform normal cells into an immortalized or malignant state. Mutant Ha-ras cooperates with E6/E7 of HPV subtype 16 in transformation of cells in vitro and may contribute to some HPV-associated cancers in humans. This study investigates whether HPV16 E6/E7 and v-Ha-ras synergize in vivo. FVB/n mice transgenic for v-Ha-ras gene (R+) were crossed with transgenic C57BL/6 mice that harbor E6/E7 of HPV16 (E+). Beginning at about 3 months of age, the bitransgenic E+R+(C57BL/6 × FVB/n) F1 mice developed mouth, eye and ear tumors. By 6 months, the prevalence of these types of mouth, eye and ear tumors was 100, 71 and 79% respectively in the E+R+ mice. Most tumors grew progressively until the mice had to be killed. The median times for the appearance of the first mouth, eye and ear tumor were 3.6, 4.3 and 4.2 months, respectively. For the two singly transgenic groups of mice, the prevalence of mouth, eye and ear tumors was 0, 0 and 6% (ER+) and 0, 0 and 0% (E+R), respectively, and the median time to first tumor was greater than 12 months for singly transgenic mice (ER+, E+R). Thus, a remarkable synergy occurred between the v-Ha-ras and HPV16 E6/E7 oncogenes in the development of primary tumors in mice.

Introduction

The E6/E7 oncogenes of certain human papilloma viruses (HPV) such as HPV16 and HPV18 appear to be essential causative factors for the majority of cervical cancers (Campo, 1998; zur Hausen, 2000) and for a significant fraction of neoplasms in humans at specific other sites, particularly anogenital, head-neck (oropharyngeal) (Gillison et al., 2000; Klussmann et al., 2001; Ringstrom et al., 2002), eye (conjunctival) (Scott et al., 2002) and ear (middle ear) (Bergmann et al., 1994; Jin et al., 1997; Tsai et al., 1997). These types of cancers are typically preceded by clinically detectable premalignant, that is, noninvasive lesions that usually arise in epithelial transition zones where cell types are changing. The reasons for the restriction of HPV16-associated tumor development to certain anatomic sites are not clear, but epithelium-specific transcriptional activation of the HPV16 oncogenes (Cripe et al., 1987; Gloss et al., 1987) is likely to play a critical role. More people are infected with HPV16 and HPV18 than develop premalignant lesions and most individuals clear initial infections (IARC, 1995a). Nevertheless, a small percentage of individuals shows viral persistence and develops preneoplastic lesions, which may progress to cancer, usually after many years (Campo, 1998). Therefore, if premalignant lesions are identified and removed early, a significant proportion of HPV-associated cancers can be prevented (Janicek and Averette, 2001). In addition, identifying subjects whose lesions are likely to take longer to progress may not only be cost-saving, but also reduce morbidity from unnecessary procedures (Schlecht et al., 2003).

Although E6/E7 may be necessary for developing certain carcinomas, the low percentage of individuals developing carcinomas even with persistent HPV infection suggests that these genes are not sufficient for carcinogenesis. Therefore, researchers have sought for other genes that could synergize with E6/E7 in producing carcinomas. However, these efforts have been inconclusive. For example, researchers have failed to find a strong association between specific p53 polymorphisms and an increased risk of cervical neoplasia (Storey et al., 1998; Klug et al., 2001). Observations about a possible synergistic role for ras have also been inconclusive. An early study found mutant Ha-ras genes in a significant fraction of late-stage cervical cancers (Riou et al., 1988). However, subsequent studies by other groups failed to confirm these findings (Willis et al., 1993), and this led the International Agency for Carcinogenic Risks (IARC, 1995b) to suggest in 1995 that ras mutations were not an important step in the development of cervical cancers. Nevertheless, more recent histochemical, serological and sequencing analyses suggested that the ras oncogene may indeed contribute to a small (approximately 20%) but significant fraction of HPV-associated neoplasia (Lee et al., 1996; Dokianakis et al., 1999; Golijow et al., 1999; Alonio et al., 2000; Mouron et al., 2000; Pedroza-Saavedra et al., 2000; Stenzel et al., 2001; Prokopakis et al., 2002). Experimentally, two papillomas isolated from an HPV18 E6/E7 transgenic mouse line were found to have acquired activating somatic point mutations in the Ha-ras gene (Greenhalgh et al., 1994). In vitro, mutant ras synergizes with high-risk HPV types 16 and 18 E7 genes in transforming primary rodent cells in culture (Crook et al., 1988; Phelps et al., 1988). Transplantable HPV16 E6/E7 expressing tumor models used cells that required co-transfection with HPV16 E6/E7 and mutant ras (Feltkamp et al., 1993; Lin et al., 1996; Eiben et al., 2002). Finally, a recent study of tumor cell metabolism in vitro showed that ras expression in primary cells leads to a metabolic inhibition of proliferation that is reversed by the additional expression of HPV16 E7 (Mazurek et al., 2001). Thus, mutant ras may cooperate with HPV oncogenes, and this cooperation may be important for a subgroup of HPV-induced cancers.

The above conflicting and inconclusive evidence on ras could be explained by assuming that ras enhances tumorigenesis may be able to help exceed a threshold in combination with another oncogene. However, we do not know of experimental evidence showing that the two oncogenes can indeed synergize in vivo, and if they do synergize, at what stage of tumorigenesis synergy may occur, since in other tumor systems ras activation may be an early and/or a late event in the multistep process of carcinogenesis. Experimental evidence in mice has also suggested variable, strain-dependent susceptibility to cancer induced by HPV or its oncogenes. For example, C57BL/6 mice (Coussens et al., 1996) are highly resistant to HPV-mediated carcinogenesis. In contrast, carcinomas can be induced in FVB/n mice by E6 alone (Song et al., 1999; Riley et al., 2003) even though E7 as well as E6 appear to be regularly required for carcinogenesis in humans since theses two oncogenes are regularly coexpressed in human HPV disease (zur Hausen, 2002). Humans are F1 hybrids genetically, and appear to be more sensitive to HPV16-induced carcinogenesis than C57BL/6 mice, but less so than FVB/n mice. Accordingly, we sought an experimental system that would reflect this strain-dependent variability, and in turn would reflect the clinical findings that HPV infection in humans is usually not sufficient to produce tumors even in the persistence of HPV. In this paper, we test the hypothesis that a potent oncogene, mutant ras (mras) can synergize with the HPV genes E6/E7. We will present evidence from double and single transgenic (FVB/n × C57BL/6) F1 animals that mutant ras (mras) and HPV16 E6/E7 strongly synergize in vivo to induce premalignant neoplasms at specific anatomic sites.

Results

Higher tumor incidence and shorter tumor latency in mras: E6/E7 bitransgenic mice

In order to generate double transgenic (E+/R+) animals, we crossed male C57BL/6 mice hemizygous for the keratin 14-promoted E6/E7 (E+) gene of HPV16 (Melero et al., 1997) with female FVB/n mice that are homozygous for the oncogenic Arg12 mras gene (R+) (Leder et al., 1990). Expression of the v-Ha-ras oncogene in epidermal cells has been shown to render mice susceptible to papilloma development following a ‘second event’ such as chemical promotion or a mild wounding (Bailleul et al., 1990; Leder et al., 1990). Similarly, expression of the HPV16 E6/E7 oncogenes in epithelial cells can render mice susceptible to the development of squamous cell neoplasms, but the penetrance of neoplasms varies greatly depending upon other genetic predispositions of the mouse strain used (Drinkwater and Bennett, 1991; Lambert et al., 1993; Arbeit et al., 1994; Coussens et al., 1996; Melero et al., 1997). The transgenic C57BL/6 lineage we used does not develop tumors even though E6/E7 is expressed at detectable levels (Melero et al., 1997). Thus, neither of the parental mouse lines develop tumors ‘spontaneously’ (Leder et al., 1990; Melero et al., 1997). The (FVB/n × C57BL/6) F1 litters were screened in a blinded manner independently by DNA blotting and genomic PCR analysis for the presence or absence of the E6/E7 gene (E+) and/or the Arg12 mras gene (R+). There was full agreement between the DNA blot and PCR results. 19 E+R+, 19 ER+ and 25 E+R mice were identified and utilized. All mice were observed for 12 months or until death. The Kaplan–Meier overall survival curves (Figure 1, left) show a median survival of 8.4 months for the E+R+, 10.3 months for the ER+ group and greater than 1 year for the E+R group. Differences in survival among the three groups (P<0.0001) or in pairwise comparison (P<0.0001) were statistically significant.

Figure 1
figure1

Comparison of the overall survival curves (left) and times-to-first tumor of any type (right) of mice carrying both the E6/E7 and mutant ras oncogenes (E+R+) or only one of the oncogenes (E+R or ER+). The Kaplan–Meier plot shows that the E+R+ mice carrying both oncogenes have a shorter median survival and a much shorter median time to first tumor than ER+ or E+R mice. The E+R+ and ER+ group have 19 animals each; the E+R group has 25 animals. The vertical tic marks on the curves indicate censors due to death from tumors other than that plotted. A total of 25 mice in the E+R group remain alive after 1 year

Similarly, Kaplan–Meier curves for time to first tumor showed statistically highly significant differences between each of the groups (Figure 1, right). The median time to first tumor development was 3.3 months for the E+R+ group, 7.2 months for the ER+ group and greater than 12 months for the E+R group (P<0.0001 in the log-rank test and in pairwise comparisons). (C57BL/6 × FVB/n) F1 mice carrying only the E6/E7 gene (E+R) failed to develop any tumors during the observation period, showing that the mutant ras gene was required.

Tumors of the mouth, eye and ear arise more frequently and earlier in mras transgenic mice also carrying the E6/E7 oncogene

The first remarkable difference among the three groups was the rapid development of mouth tumors beginning at 3 months of age. At the end of 6 months, 100% of E+R+ mice developed mouth tumors while ER+ or E+R mice had none (Figure 2). The median time to the first mouth tumor for the E+R+ group was 3.6 months, while the median time for ER+ mice was greater than 12 months. None of the E+R mice developed mouth tumors (P<0.0001 for differences among the three groups using the log-rank test, and P<0.005 for all pairwise comparisons). Table 1 shows the prevalence of mouth, eye and ear tumors at 6 and 9 months of age among the three groups. For those mice that developed mouth tumors during the first year of life, the average number of mouth tumors per mouse was 6.5 in E+R+ mice, with only two tumors per mouse in ER+ mice (P=0.02, Wilcoxon rank-sum test). Examples of the macroscopic and microscopic appearance of these tumors are shown in Figure 3. The tumors developed at the mucocutaneous junction of the mouth. Predominantly the angles of the mouth were affected (Figure 3a), but additional tumors developed at other sites of the lips.

Figure 2
figure2

Tumors of the mucocutaneous junctions of the lips, eyes and external ear canal develop much faster in E+R+ mice than in the single oncogene transgenic mice (P<0.0001). The vertical tic marks on the curves indicate censors due to death from tumors other than that plotted. In total, 25 mice in the E+R group remain alive after 1 year

Table 1 Prevalence of mouth, eye and ear tumors and numbers of mice alive in the three genetic groups at times specified
Figure 3
figure3

Macroscopic and histologic appearance of mouth, eye and ear tumors in E+R+ positive mice. Pictures were taken from mice at 5–6 months of age. (a) Note that the angles of the snouts are primarily affected by exuberant, papillary tumors. Morphologically similar tumors developed on the eyelids (f) and in the ear canals (j) at the squamocolumnar junctions (i,m). Histolologically, all the tumors are dysplastic squamous papillomas (b,c,g,k). Even though the squamous epithelium exhibits maturation, there are histologic features of dysplasia in the form of disturbed maturation, broadening of the basal layer with increased nuclear/cytoplasmic ratios and extension of mitotic activity to the suprabasal layers (d,h,l,m). The dysplastic nature of the lesion is also reflected by the BrdU nuclear incorporation to various layers of the epithelium (e) in contrast to basal layer restriction in nondysplastic settings. The human equivalent of such tumors would be regarded as squamous papillomas with intraepithelial neoplasia

Tumor development at the mucocutaneous junctions of the eyelid was also frequent (Figure 3f). The median time to the first eyelid tumor was 4.3 months in the E+R+ group, greater than 1 year for the ER+ group; none of E+R mice developed eyelid tumors during the observation period (Figure 2). The three groups were statistically significantly different (P<0.0001, log-rank test and P<0.005 for all pairwise comparisons). Among mice that developed eyelid tumors, the median number of eyelid tumors was 5 in the E+R+ group vs 3 in the ER+ group (P=0.10 Wilcoxon rank test).

Finally, E+R+ mice developed tumors in the external ear canal (Figure 3j) with the median time to first tumor being 4.2 months compared to greater than a year in the ER+ mice; again none of the 25 E+R mice developed ear tumors within the 1 year observation period (Figure 2). Pairwise comparisons indicated that the E+R+ group was statistically significantly different from the other two groups (P<0.0001). Among mice that developed ear tumors, the average number of ear tumors per mouse was 2 in E+R+ mice compared to 1 in ER+ mice (P=0.22 Wilcoxon rank-sum test).

There were no statistically significant differences between the E+R+ and the ER+ groups in terms of ‘time to first tumor’ or number of tumors originating from the nostrils (3/19 mice in the E+R+ and 2/19 in the ER+ group) or originating from the vulva or anus (6/19 in the E+R+ and 3/19 in the ER+ group).

The incidence of tumors arising from sites other than mucocutaneous junctions is not increased in E+R+mice

Mice expressing the mras oncogene as a transgene in skin have been shown to have an increased susceptibility to developing dorsal skin tumors after surgical wounding or after local application of chemical promoters to dorsal skin. Even in the absence of intentional promotion, these mice develop skin tumors at a low frequency at sites subject to biting and scratching, particularly at the base of the tail, or behind the ears and dorsal neck. Similarly, the development of jaw tumors (odontoblastomas) occurred in the absence of intentional promotion and was not altered by application of chemical promoters to dorsal skin. In agreement with these observations, Figure 4 (left and middle panel) shows that the latency of tumorigenesis along the dorsum of the animal was not statistically different between E+R+ and ER+ mice (P=0.99 and 0.26 respectively). These data demonstrate that the HPV16 E6/E7 oncogene does not display any synergistic effect with the mutant ras oncogene (v-Ha-ras) at these sites. The average number of dorsal skin tumors per mouse in the E+R+ mice was similar to the ER+ mice (2 vs 1 for head and neck and, 2 vs 3.5 and for ‘back to base of tail tumors’). Finally, as would be expected, the latency of jaw tumors that develop from sites where the K14-promoter-driven E6/E7 gene is not expressed was also not different in any detectable way between E+R+ and ER+ mice (P=0.73) (Figure 4, right panel).

Figure 4
figure4

The presence of the E6/E7 oncogene fails to synergize with the mutant ras gene in causing an increased incidence of tumors of the dorsal skin or the jaw which are typical sites of tumor development in mutant ras only (ER+) mice. (The horizontal line at 100% corresponds to the E+R group). The vertical tic marks on the curves indicate censors due to death from tumors other than that plotted. In total, 25 mice in the E+R group remain alive after 1 year

E+R+ mice develop a mixture of malignant and premalignant lesions

Tumors arising from the lip, eyelids, anus, vulva or skin of the back had macroscopic characteristics of premalignant or malignant lesions. Nonmalignant tumors were very well circumscribed symmetrical, often pedunculated and, when broad based, well demarcated and nonulcerated. Histologically these tumors simulated condylomas with mild dysplasia observed in humans. Analysis of BrdU incorporation into the neoplastic epithelia revealed persistence of DNA in differentiating keratinocytes (Figure 3e), indicating a profound failure to control or abolish DNA synthesis in the differentiating keratinocytes. Furthermore, these tumors persisted once they had appeared. However, fragments of these tumors did not adapt as cell lines to culture and did not produce tumors within 6 weeks when transplanted into immunodeficient (athymic nude) mice, confirming that these tumors were not frankly malignant. By contrast, malignant lesions typically showed heaped ill-defined reddened margins, irregular shape and central ulcerations or necrosis. Three tumors, one vulvar and two ear tumors, had this malignant appearance macroscopically. All of them readily adapted to growth in culture. Transplantation of these three tumors into athymic nude mice also produced tumors, thereby confirming the malignant state. Histopathologically, the transplanted tumors had the appearance of spindle cell carcinomas. Of particular interest was the histopathological analysis of the ear tumors that appeared to arise from the ventral surface of the ear extending into and often filling the external ear canal (Figure 3k). In total, 50 μm step sectioning of the temporal bone of two of these ear tumors revealed that both of them extended into the inner ear (Figure 3k) and resembled verrucous carcinoma observed in the middle ear of humans (Figure 3k and l). Transition zones of columnar epithelium and squamous neoplastic epithelium were frequently observed in the middle ear (Figure 3m). Several of the mice with ear tumors showed circling gait and tilting of the head consistent with inner ear damage.

Simultaneous expression of the two oncogenes in tumors arising in E+R+ mice

To assess gene expression in vivo, we isolated RNA from snap-frozen primary tumors of E+R+ mice to determine whether the two oncogenes were indeed both expressed at time of tumor growth in vivo. mRNA was reverse transcribed to generate cDNA that was incubated with primers specific for message made by the mutant ras transgene or the E7 oncogene. Figure 5 shows simultaneous expression of the two oncogenes in premalignant and in malignant primary tumors. Furthermore, Figure 5 shows that RNA message was analysed since no bands were identified without the use of reverse transcriptase (RT). Analysis of the skin in nontumor areas was consistent with previous studies that had analysed the expression of E6/E7 and mutant ras in the parental mice carrying the oncogenes separately; that is, normal skin of E+R+ F1 mice failed to express mras message, while message of the HPV16 oncogenes was detected. Similarly, the mras transgene was not constitutively expressed in uninvolved skin but only in neoplastic lesions of ER+ (Hansen and Tennant, 1994), while the E+R mice constitutively expressed the HPV 16 oncogenes even in uninvolved epidermis (Melero et al., 1997).

Figure 5
figure5

Primary ear and mouth tumors of E+R+ mice express the E7 as well as the mutant ras gene. RNA prepared from the tumor fragments were reverse transcribed. A 292 base pair band developed with E7-specific primers and the 214 base pair band were generated by primers specific for spliced transgenic mutant ras RNA message. PCR reaction of control material lacking RT failed to generate these bands indicating messenger RNA was needed for detection. The figure shows results obtained with two primary ear tumors (758R and 758L) and one primary mouth tumor (925)

Discussion

This study finds that HPV16 E6/E7 and mutant Arg12 ras can strongly synergize in vivo to reduce latency and increase primary tumor incidence. The synergy was restricted to anatomic sites with epithelial transition zones classically affected by HPV16 E6/E7 oncogene expression. There was no synergy at sites unknown to be affected by HPV16 E6/E7 (ie, dorsal skin and jaw). The keratin 14 promoter-driven HPV16 E6/E7 transgenic C57BL/6 line we used did not develop tumors even though E6/E7 is expressed at detectable levels (Melero et al., 1997) consistent with the known resistance of this mouse strain to various types of carcinogenesis and tumor promotion (Drinkwater and Bennett, 1991). By using the hemizygous genetic background of the FVB/n mice we increased susceptibility because homozygous FVB/n are known to be susceptible to E6/E7 tumorigenesis (Coussens et al., 1996). But unlike what is observed in homozygous FVB/n mice, the E6/E7 transgene alone was not sufficient to cause spontaneous tumor development in the F1 mice unless the mutant ras oncogene was also present. By analogy (C57BL/6 × mutant Ha-ras transgenic FVB/n), F1 mice are less sensitive to tumor development after promotion than mutant Ha-ras transgenic homozygous FVB/n mice (Siegel et al., 2000). Therefore, in order to test most compellingly the hypothesis that oncogenes synergize in vivo, we intentionally designed our experimental system to use F1 hybrids between C57BL/6 and FVB/n strains. This bypassed the possibility that a co-tumorigenic effect of mras might have been silenced by the known resistance of homozygous C57BL/6 mice to HPV-induced carcinogenesis. Also, the model bypassed the possibility that the extreme sensitivity of homozygous FVB/N mice to the HPV16 E6 oncogene or the Ha-ras transgene might mask a possible synergistic effect of the two oncogenes such as mras and E6/E7. Indeed, these prospects were confirmed by our experimental data showing that when the two strains were crossed, both single transgenic strains (both E/R+ and E+/R) develop only few and late tumors. In striking contrast, tumors were observed in virtually all double transgenic mice (E+R+) in the hybrid background (C57BL/6 × FVB/n). The synergy that we observed may therefore be the result of having two oncogenes expressed in a genetic setting of F1 hybrids between a very sensitive and a very resistant mouse strain so that neither oncogene could readily penetrate unless aided by the other. Regression of individual neoplasms could not be documented and in three cases, the lesions progressed to macroscopically discernable malignant disease within a few months. In most mice, prominence of mouth and jaw tumors prevented proper feeding and required killing of the animals. Thus, other lesions may not have had sufficient time to progress to cancer. Notably, none of the single transgenic animals developed malignant lesions, which underscores the strength of the synergism between these two oncogenes.

Recent findings analysing human cervical specimens are consistent with our conclusion that mras gene expression aids in the development of high-risk HPV lesions in vivo. For example, a recent clinical study of noncancerous cytological cervical specimens positive for high- or low-risk HPV types found that the frequency of mutant ras was more than 10-fold greater in cervical specimens positive for high-risk (HPV16 or HHPV18) as compared to the low-risk (HPV6) subtypes (Golijow et al., 1999). In addition, a recent study of premalignant cervical lesions by Alonio et al. (2003) found progression to cancer was less than 2 years when mutated Ha-ras was detected, while none of the cases without progression for 6–10 year showed mutated ras. Together, these studies and our work support the idea that simultaneous presence of ras mutations with high-risk HPV subtypes identifies an irreversible premalignant stage and when lesion should be surgically removed. We did not search for additional mutations that may have occurred in the lesions we observed and we surmise that additional factors are required for tumor development and progression to cancer. This is also demonstrated by the incomplete penetrance of the tumorigenesis (except for mouth tumors) observed in the E+R+ mice.

The tumors that developed in the double transgenic mice showed a striking anatomic distribution to skin creases of the mouth, ear and eye. Even in highly susceptible homozygous FVB/n mice, induction of cervical carcinogenesis requires long-term treatment with exogenous estrogen and therefore would not be expected in our mice without hormone treatment. While we do not have a full explanation of the observed site pattern of tumor development, data in the literature are suggestive. Full thickness wounding has been identified and highly studied as a promotional stimulus for skin carcinogenesis (Bailleul et al., 1990; Leder et al., 1990), and primary tumor development may be enhanced by inflammation regardless of whether it is caused by wounding (Bailleul et al., 1990; Leder et al., 1990), infections (Daniel et al., 2003, for a review, see, Balkwill and Mantovani, 2001) or immune responses to the oncogene (Siegel et al., 2000). While the infection with the human papilloma virus is thought to occur at sites of superficial epithelial wounding, it is not clear whether such wounding also promotes HPV-mediated tumor development once infection has occurred. Gnawing on the very firm pelleted laboratory chow may have caused superficial injuries of the lips and facial irritations due to the feed dust generated by the gnawing and entering into the openings of eyes and ears. Certainly the angles of the mouth that are points of physical stress had predilection for mouth tumor development. In addition, members of the AP-1 pathways such as cfos (Saez et al., 1995) and certain metallo-proteinases (Coussens et al., 2000), which are activated during inflammation and wound repair, cooperate with mutant ras and appear to be needed but are not sufficient for progression of the papillomas to invasive carcinomas.

Additionally, our mouse model provides a defined genetic background in which the enhancing and preventive influences of the immune system on primary tumor development can be explored. The premalignant and malignant lesions that develop in the (C57BL/6 × FVB/n) F1 mice express MHC Class I Db, the molecule that restricts the RAHYNIVTF epitope of HPV16 E7 (Feltkamp et al., 1995). In addition, since HLA-A2.1 transgenic C57BL/6 mice are available, our primary tumor model can be adjusted immediately to express the human MHC Class I HLA-A2.1 molecule presenting the known HPV16 E6/E7 epitopes (Ressing et al., 1995). T cells specific for these epitopes can be induced that protect mice in tumor transplantation models against E7-transfected mras expressing tumor cells (Feltkamp et al., 1995; Eiben et al., 2002). Nevertheless, for primary HPV16-associated tumor development, we have yet to find conditions that provide effective protection against cancer development in humans and mice. Together, the model we describe may help development of immunological, molecular or biochemical means for preventing the development of premalignant lesions or ways to destroy irreversible premalignant lesions once they have appeared (Siegel et al., 2000; Girardi et al., 2001). In addition, our results should further stimulate diagnostic screening for mutant ras in clinical specimens positive for high-risk HPV, since our findings support the notion that simultaneous expression of the two oncogenes identifies an irreversible premalignant stage that must be removed surgically.

Materials and methods

Mice

C57BL/6 mice (Charles River) made transgenic for the keratin-14 promoter driven E6/E7 genes of HPV16 (Melero et al., 1997) were carried as heterozygotes in the C57BL/6 (Charles River) background. Previous studies have shown that this lineage of mice whether homozygous or heterozygous for the transgene develops hypertrophy keratinized epithelia but no cutaneous neoplasms. FVB/n mice (Taconic) also referred elsewhere to as Tg.AC (Leder et al., 1990) were transgenic for the zeta-globin promoter-driven v-Ha-ras gene (containing activating glycine-to-arginine mutation at codon 12, and an alanine-to-threonine mutation at codon 59) (Leder et al., 1990). Wild-type FVB/n mice were also obtained from Taconic Farms. Despite the use of the zeta-globin promoter, the v-Ha-ras transgenic mice express detectable levels of the transgene in keratinocytes, apparently due to anomalous expression attributed to the site and the manner of genomic integration of the transgene (Leder et al., 1990). Expression increases even further in papillomas and squamous cell carcinomas developing in these mice (Leder et al., 1990). Heterozygous E6/E7 transgenic mice (E+) were mated with homozygous mras transgenic FVB/n mice (R+) or wild-type FVB/n mice to obtain female E+R+, ER+ and E+R (FVB/n × C57BL/6) F1 mice carrying one allele of one or both oncogenes. Mice were inspected for tumors at least twice each week. Complete autopsies were performed and abnormalities were also evaluated microscopically. The three orthogonal caliper measurements (a,b,c) were made and the tumor volume was calculated by the formula π abc/6. All mice were maintained in a specific pathogen-free barrier facility at The University of Chicago. Only female mice were used in the experiments to reduce skin injury due to fighting.

Derivation of cell lines

Some of the tumors were excised under sterile conditions and small minced fragments were placed in vitro for establishing a tumor cell line and transplanted into syngeneic nude and normal mice to determine growth behavior in vivo and stored in liquid nitrogen for further reference.

Tumor transplantation

For tumor challenge an equal number (three to six) of 1 mm3 tumor fragments were implanted subcutaneously with a trocar into nude and normal mice. The nude mice were always injected last to verify that the tumor tissue remained viable until the end of the transplantation procedure.

Genotyping of mice

The genotype of the parents and F1 litters was determined by genomic PCR sequencing at The University of Chicago and by DNA blotting at the NIEHS using coded samples. For E6/E7 genomic analysis tail DNA was amplified by PCR with primers (sense, 5′-IndexTermGCATGGAGATACACCTACATTG-3′; antisense 5′-IndexTermTGGTTTCTGAGAACAGATGG-3) for the presence or absence of the 291 base-pair HPV16 E7 gene sequence. In addition, DNA blotting using the 291 base pair HPV16 E7 gene fragment generated from an HPV16 E7 gene vector was used for DNA blot analysis, which confirmed the PCR analysis. The presence of the intact mutant ras gene needed for the responder genotype was determined by the presence of a 2000-bp band (Kantz et al., 1999); nonresponders that occur with a frequency of <2.0% per generation can be recognized by various size deletions in this band (Thompson et al., 1998).

Analysis of mRNA expression

Tumor fragments were snap frozen in liquid nitrogen, the tissue was pulverized, and RNA was isolated using the Quiagen RNA kit. In all, 4 μg of RNA, showing a ratio of the optical densities at 260 and 280 nm of greater than 1.8, was treated with 4 U of DNase I (Invitrogen), and 1 μg of RNA was then reverse transcribed, using the Advantage-RT for PCR kit (Clontech) containing oligo-dT primers. In total, 20 μl of the complementary (cDNA) preparation were used for the RT–PCR reaction in 50 μl final volume containing 200 μ M of each of the dNTPs, 1.5 μ M MgCl2, 1 U Taq DNA polymerase (Promega) and 0.4 μ M of each of the ras and E7 primers. In total, 25 μl of the above RT–PCR was loaded on a 1.5% TAE gel and fragments were visualized by ethidium bromide staining. The HPV 16 E7 primers (sense, 5′-IndexTermGCATGGAGATACACCTACATTG-3′; antisense, 5′-IndexTermTGGTTTCTGAGAACAGATGG-3). Spanned nucleotides so that a 292 base pair fragment was amplified. The ras primers (sense, 5′-IndexTermTTGGACAAACTACCTACAG-3′; anti-sense, 5′IndexTermAATTCTGAAGGAAAGCTCC-3′) targeted the 3′UTR of the SV40 sequence and surrounded a 65 base pair fragment of the sequence. The genomic fragment gives a 279 bp fragment while the RNA, since it is spliced, will give a 214 bp fragment. In some experiments, the identities of the E7 and ras fragments were confirmed by Southern blotting. Furthermore, PCR was also performed using only the E7 specific or only the ras transgene specific primers to ensure further the identity of the fragments observed.

Statistical analysis

The time to first tumor (any tumor or particular type of tumor) was analysed and presented using the method of Kaplan and Meier (1958) and the risks of developing tumor were compared among the three groups using the censored rank order (log rank) test derived by Breslow (1970). Among mice that developed tumors, nonparametric Wilcoxon rank-sum tests were used to compare the maximum number of tumors between E+R+ and ER+ groups.

References

  1. Alonio LV, Dalbert D, Picconi MA, Cervantes Vazquez G, Garcia Carranca A, Distefano AL, Mural J, Bartt O, Bazan G and Teyssie AR . (2000). Medicina, 60, 895–901.

  2. Alonio LV, Picconi MA, Dalbert D, Mural J, Bartt O, Bazan G, Dominguez M and Teyssie AR . (2003). J. Clin. Virol., 27, 263–269.

  3. Arbeit JM, Munger K, Howley PM and Hanahan D . (1994). J. Virol., 68, 4358–4368.

  4. Bailleul B, Surani MA, White S, Barton SC, Brown K, Blessing M, Jorcano J and Balmain A . (1990). Cell, 62, 697–708.

  5. Balkwill F and Mantovani A . (2001). Lancet, 357, 539–545.

  6. Bergmann K, Hoppe F, He Y, Helms J, Muller-Hermelink HK, Stremlau A and de Villiers EM . (1994). Int. J. Cancer, 59, 463–466.

  7. Breslow N . (1970). Biometrika, 57, 579.

  8. Campo MS . (1998). Trends Microbiol., 6, 424–426.

  9. Coussens LM, Hanahan D and Arbeit JM . (1996). Am. J. Pathol., 149, 1899–1917.

  10. Coussens LM, Tinkle CL, Hanahan D and Werb Z . (2000). Cell, 103, 481–490.

  11. Cripe TP, Haugen TH, Turk JP, Tabatabai F, Schmid III PG, Durst M, Gissmann L, Roman A and Turek LP . (1987). EMBO J., 6, 3745–3753.

  12. Crook T, Storey A, Almond N, Osborn K and Crawford L . (1988). Proc. Natl. Acad. Sci. USA, 85, 8820–8824.

  13. Daniel D, Meyer-Morse N, Bergsland EK, Dehne K, Coussens LM and Hanahan D . (2003). J. Exp. Med., 197, 1017–1028.

  14. Dokianakis DN, Papaefthimiou M, Tsiveleka A and Spandidos DA . (1999). Oncol. Rep., 6, 1327–1331.

  15. Drinkwater NR and Bennett LM . (1991). Prog. Exp. Tumor Res., 33, 1–20.

  16. Eiben GL, Velders MP, Schreiber H, Cassetti MC, Pullen JK, Smith L and Kast WM . (2002). Cancer Res., 62, 5792–5799.

  17. Feltkamp MC, Smits HL, Vierboom MP, Minnaar RP, de Jongh BM, Drijfhout JW, ter Schegget J, Melief CJ and Kast WM . (1993). Eur. J. Immunol., 23, 2242–2249.

  18. Feltkamp MC, Vreugdenhil GR, Vierboom MP, Ras E, van der Burg SH, ter Schegget J, Melief CJ and Kast WM . (1995). Eur. J. Immunol., 25, 2638–2642.

  19. Gillison EW, Powell DJ, McConkey C and Spychal RT . (2000). Br. J. Surg., 87, 362–373.

  20. Girardi M, Oppenheim DE, Steele CR, Lewis JM, Glusac E, Filler R, Hobby P, Sutton B, Tigelaar RE and Hayday AC . (2001). Science, 294, 605–609.

  21. Gloss B, Bernard HU, Seedorf K and Klock G . (1987). EMBO J., 6, 3735–3743.

  22. Golijow CD, Mouron SA, Gomez MA and Dulout FN . (1999). Gynecol. Oncol., 75, 108–112.

  23. Greenhalgh DA, Wang XJ, Rothnagel JA, Eckhardt JN, Quintanilla MI, Barber JL, Bundman DS, Longley MA, Schlegel R and Roop DR . (1994). Cell Growth Differ., 5, 667–675.

  24. Hansen LA and Tennant R . (1994). Mol. Carcinog., 9, 143–154.

  25. IARC (International Agency for Research on Cancer) (1995a). Monographs on the evaluation of carcinogenic risks to humans. Human papilloma virus. Vol 64, p. 65.

  26. IARC. (International Agency for Research on Cancer) (1995b). Monographs on the evaluation of carcinogenic risks to humans. Human papilloma virus. Vol 64, p. 245.

  27. Janicek MF and Averette HE . (2001). CA Cancer J. Clin., 51, 92–114.

  28. Jin YT, Tsai ST, Li C, Chang KC, Yan JJ, Chao WY, Eng HL, Chou TY, Wu TC and Su IJ . (1997). Am. J. Pathol., 150, 1327–1333.

  29. Kantz DC, Lacks GD and Cannon RE . (1999). Biotechniques, 27, 278–280.

  30. Kaplan E and Meier P . (1958). J. Am. Stat. Assoc., 53, 457–481.

  31. Klug SJ, Wilmotte R, Santos C, Almonte M, Herrero R, Guerrero I, Caceres E, Peixoto-Guimaraes D, Lenoir G, Hainaut P, Walboomers JM and Munoz N . (2001). Cancer Epidemiol. Biomarkers Prev., 10, 1009–1012.

  32. Klussmann JP, Weissenborn SJ, Wieland U, Dries V, Kolligs J, Jungehuelsing M, Eckel HE, Dienes HP, Pfister HJ and Fuchs PG . (2001). Cancer, 92, 2875–2884.

  33. Lambert PF, Pan H, Pitot HC, Liem A, Jackson M and Griep AE . (1993). Proc. Natl. Acad. Sci. USA, 90, 5583–5587.

  34. Leder A, Kuo A, Cardiff RD, Sinn E and Leder P . (1990). Proc. Natl. Acad. Sci. USA, 87, 9178–9182.

  35. Lee JH, Lee SK, Yang MH, Ahmed MM, Mohiuddin M and Lee EY . (1996). Gynecol. Oncol., 62, 49–54.

  36. Lin XH, Dahlin-Huppe K and Stallcup WB . (1996). J. Cell. Biochem., 63, 463–477.

  37. Mazurek S, Zwerschke W, Jansen-Durr P and Eigenbrodt E . (2001). Oncogene, 20, 6891–6898.

  38. Melero I, Singhal MC, McGowan P, Haugen HS, Blake J, Hellstrom KE, Yang G, Clegg CH and Chen L . (1997). J. Virol., 71, 3998–4004.

  39. Mouron SA, Abba MC, Guerci A, Gomez MA, Dulout FN and Golijow CD . (2000). Mutat. Res., 469, 127–134.

  40. Pedroza-Saavedra A, Cruz A, Esquivel F, De La Torre F, Berumen J, Gariglio P and Gutierrez L . (2000). Arch. Virol., 145, 603–623.

  41. Phelps WC, Yee CL, Munger K and Howley PM . (1988). Cell, 53, 539–547.

  42. Prokopakis P, Sourvinos G, Koumantaki Y, Koumantakis E and Spandidos DA . (2002). Oncol. Rep., 9, 129–133.

  43. Ressing ME, Sette A, Brandt RM, Ruppert J, Wentworth PA, Hartman M, Oseroff C, Grey HM, Melief CJ and Kast WM . (1995). J. Immunol., 154, 5934–5943.

  44. Riley RR, Duensing S, Brake T, Munger K, Lambert PF and Arbeit JM . (2003). Cancer Res., 63, 4862–4871.

  45. Ringstrom E, Peters E, Hasegawa M, Posner M, Liu M and Kelsey KT . (2002). Clin. Cancer Res., 8, 3187–3192.

  46. Riou G, Barrois M, Sheng ZM, Duvillard P and Lhomme C . (1988). Oncogene, 3, 329–333.

  47. Saez E, Rutberg SE, Mueller E, Oppenheim H, Smoluk J, Yuspa SH and Spiegelman BM . (1995). Cell, 82, 721–732.

  48. Schlecht NF, Platt RW, Duarte-Franco E, Costa MC, Sobrinho JP, Prado JC, Ferenczy A, Rohan TE, Villa LL and Franco EL . (2003). J. Natl. Cancer Inst., 95, 1336–1343.

  49. Scott IU, Karp CL and Nuovo GJ . (2002). Ophthalmology, 109, 542–547.

  50. Siegel CT, Schreiber K, Meredith SC, Beck-Engeser GB, Lancki DW, Lazarski CA, Fu Y, Rowley DA and Schreiber H . (2000). J. Exp. Med., 191, 1945–1956.

  51. Song S, Pitot HC and Lambert PF . (1999). J. Virol., 73, 5887–5893.

  52. Stenzel A, Semczuk A, Rozynskal K, Jakowicki J and Wojcierowski J . (2001). Pathol. Res. Pract., 197, 597–603.

  53. Storey A, Thomas M, Kalita A, Harwood C, Gardiol D, Mantovani F, Breuer J, Leigh IM, Matlashewski G and Banks L . (1998). Nature, 393, 229–234.

  54. Thompson KL, Rosenzweig BA and Sistare FD . (1998). Toxicol. Pathol., 26, 548–555.

  55. Tsai ST, Li C, Jin YT, Chao WY and Su IJ . (1997). Int. J. Cancer, 71, 208–212.

  56. Willis G, Jennings B, Ball RY, New NE and Gibson I . (1993). Gynecol. Oncol., 49, 359–364.

  57. zur Hausen H . (2000). J. Natl. Cancer Inst., 92, 690–698.

  58. zur Hausen H . (2002). Nat. Rev. Cancer, 2, 342–350.

Download references

Acknowledgements

We thank Donald A Rowley, Michael T Spiotto and Michele Carbone for their critical suggestions during the course of the study and for reviewing the manuscript. We also thank Gordon Bowie for excellent photographical work, Christine Langan and Keith Voogd for excellent technical support and Ruthie Cornelius for help with preparing the manuscript. This work was supported by National Institute of Health Grants P01-CA97296, R01-CA22677 and R01-CA37516, and a grant by a pre-Clinical grant from the Cancer Research Institute.

Author information

Affiliations

Authors

Corresponding author

Correspondence to Hans Schreiber.

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Schreiber, K., Cannon, R., Karrison, T. et al. Strong synergy between mutant ras and HPV16 E6/E7 in the development of primary tumors. Oncogene 23, 3972–3979 (2004). https://doi.org/10.1038/sj.onc.1207507

Download citation

Keywords

  • HPV 16
  • human papilloma virus 16
  • E6/E7
  • vHa ras
  • transgenic mice
  • synergy
  • primary tumors

Further reading

Search