Differential Expression of ras Signal Transduction Mediators in Verrucous and Squamous Cell Carcinomas of the Upper Aerodigestive Tract

Friedell and Rosenthal1 first recognizably described oral verrucous carcinoma in 1941 in their study of 8 patients who developed cancer of the mouth; chewing tobacco was the presumed etiology in these cases. The tumors they described had a “papillary verrucoid character which was covered and surrounded by areas of leukoplakia.” Furthermore, they noted, “extensive infiltration into the submucosal tissues was not present.” However, it was not until Ackerman2 more fully described the pathologic features of verrucous carcinoma in 1948 that we gained our best understanding of the gross and histologic features of the tumor. Ackerman's original observations of the pathologic features of verrucous carcinoma are complete and little can be improved upon his morphologic description.

In subsequent years, additional information concerning molecular mechanisms of verrucous carcinoma causation have been described. These include studies involving bcl-2, p16, Rb, p53, cyclin D1, and human papillomavirus, as well as other mechanisms.3–6 In addition, H-ras mutations have been described in verrucous carcinomas. In a study by Anderson et al,7 the incidence of ras mutation and human papillomavirus DNA detection was twice as high in verrucous cancer as compared to oral squamous cell carcinoma.7 Previous studies have also shown that the Ras gene product showed differences of expression in verrucous carcinoma versus squamous cell carcinoma.8,9 In the study by Freer et al,8 normal mucosa and verrucous carcinoma showed significant expression of p21 Ras protein, but a marked decrease in the detectable amounts of this protein in more undifferentiated cells. More recently, we learned more of the complexities of how ras is a key link in extracellular stimuli transduction in the cell. Key to transferring extracellular signals to the nucleus are a series of kinase cascades, including Raf and mitogen-activated protein kinases (MAPK). Raf, the cellular homologue of the viral oncogene v-raf, is a cytoplasmic protein with serine/threonine kinase activity. The binding of ras to raf causes raf to phosphorylate and thereby activate MAPK (also known as MAPK/ERK kinases 1 and 2), which in turn activates the next set of downstream kinases, ERK-1 (extracellular receptor-stimulated kinase 1) and ERK-2 (extracellular receptor-stimulated kinase 2).10–13 Once activated, ERKs can move into the nucleus to phosphorylate and activate transcription factors such as Elk-1. Elk-1 and other factors are further involved ultimately with gene expression. The mediators of this signal transduction are highly conserved in nature, suggesting their importance. In addition, many of these signal mediators that lie downstream of ras in a series of kinase cascades are linked to many cellular functions other than external signal transduction. In a recent article by Mishima et al,14 the authors described the expression of extracellular signal-regulated kinases ERK-1 and ERK-2 in oral squamous carcinomas. In this study, we observed the expression of Raf, ERK-1, and ERK-2 in verrucous carcinomas and squamous carcinomas of the upper aerodigestive tract.

Seventeen verrucous carcinomas and 10 squamous carcinomas were studied. The sites of the verrucous carcinomas were gingiva (7 patients), hard palate (4 patients), floor of mouth (2 patients), tongue (2 patients), oral pharynx (1 patient), and buccal mucosa (1 patient). The patients ranged in age from 42 to 80 years. Ten patients were men and 7 were women. The sites of the squamous cell carcinomas were the buccal mucosa (3 patients), larynx (3 patients), tongue (2 patients), oral pharynx (1 patient), and gingiva (1 patient). Five tumors were well differentiated and 5 were moderately differentiated. The patients' ages ranged from 55 to 87 years. Seven patients were men and 3 were women.

Four-micrometer sections of each tumor were cut onto coated slides for immunohistochemical analysis with antibodies to ERK-1, ERK-2, and Raf-1. An immunoperoxidase reaction for ERK-1 and ERK-2 (ERK-1 [K23] and ERK-2 [C14], Santa Cruz Biotechnology, Santa Cruz, Calif) was performed on each case using an avidin-biotin complex method. ERK-1 was directed against an epitope within subdomain XI of ERK-1–encoded MAP kinase p44 of rat origin. ERK-2 was directed against an epitope mapping at the carboxy terminus of ERK-2–encoded MAP kinase p42 of rat origin. There is some cross-reactivity of ERK-1 and ERK-2. Sections were deparaffinized in xylenes, rehydrated in graded alcohols, and rinsed. No proteolytic digestion or antigen unmasking was used. Endogenous peroxidase activity was quenched using hydrogen peroxide in distilled water (1.88%). Nonspecific background staining was prevented by application of blocking agent (BioGenex Power Block; BioGenex, San Ramon, Calif) for 15 minutes at room temperature. Sections were covered with ERK-1 or ERK-2 primary antisera used at a dilution of 1:300, incubated overnight at 4°C, and rinsed. Sections were then covered with DAKO LINK (Dako Corporation, Carpinteria, Calif) for 20 minutes at room temperature, rinsed and covered with DAKO LSAB2 (Dako) label for 20 minutes at room temperature, and rinsed in phosphate–buffered saline. Sections were then incubated with diaminobenzidine for 6 minutes to demonstrate the signal of the primary antibody. A counterstain of 10% Harris hematoxylin for 1 minute was used. Negative control slides were prepared by substituting normal rabbit serum. The immunoperoxidase reaction for Raf-1 (Raf-1 [E-10] Santa Cruz Biotechnology) was carried out in a similar manner, except that antigen unmasking was used and accomplished with citrate buffer in a microwave oven for two 5-minute cycles on full power (500 W). The anti-Raf-1 is directed against an epitope mapping at the carboxy terminus of Raf-1 p74 of human origin. Slides were allowed to cool to room temperature for 20 minutes. The slides were then incubated in the secondary antibody (biotinylated anti-mouse immunoglobulin; 1:200) for 1 hour at room temperature, rinsed, and covered with Vector-ABC Kit (Vector Laboratories, Burlingame, Calif) for 1 hour at room temperature.

Intensity of staining levels was categorized as 0 to 3+ (negative to intense). The staining was observed also for localization. The staining intensity was semiquantitatively scored as follows: 0, no staining; 1, weak staining but clearly observable; 2, moderate staining; and 3, intense staining that obscured cellular detail in part.

Statistical calculations were performed using Stata Statistical Software: Release 6.0 (Stata Corporation, College Station, Tex).

Normal epithelium exhibited weak cytoplasmic (0–1+) staining with the anti-ERK-1 and anti-ERK-2 antibodies. Raf-1 was also minimally to weakly (0–1+) expressed.

The staining pattern with Raf-1 showed significant differences between verrucous carcinoma and the squamous carcinomas (Table 1). In every case of verrucous carcinoma there was moderate to intense (2–3+) cytoplasmic staining of the most basally situated epithelial cells, with markedly diminished to absent staining of the overlying neoplastic epithelial cells (Figure 1). In the squamous carcinomas, anti-Raf-1 stained all cells (Figure 2), but the intensity was weak except in 1 case. In that case, the squamous cell carcinoma arose within a preexisting verrucous carcinoma. This tumor showed strong anti-Raf-1 staining in the invasive squamous cell carcinoma portions, as well as in the surrounding verrucous carcinoma component. The differences in staining of Raf-1 at the basal layers in the squamous cell carcinomas and verrucous carcinomas was significant (P < .001; Table 1).

In 16 of 17 verrucous carcinomas, the most basally situated layers of epithelial cells minimally stained with ERK-1 and ERK-2 (Figure 3). This was in contrast to the staining pattern of moderately differentiated squamous cell carcinoma, in which the majority of tumor cells (8/10), even the ones most basally situated (situated next to stroma), stained with ERK-1 and ERK-2 (Figure 4). The differences in staining of ERK-1 and ERK-2 at the basal layer between squamous cell carcinoma and verrucous carcinoma were significant (P < .001; Table 2).

In the last decade, numerous studies have begun to elucidate the critical role of ras-mediated signal transduction. ras is a meeting or convergence point for cellular signaling pathways.15 The ras-mediated, extracellular stimulus–mediated pathway begins when an external factor, such as a growth factor, binds to its cell membrane receptor. In the case of epidermal growth factor receptor, epidermal growth factor stimulates its receptor to autophosphorylate tyrosine sites in the cytoplasmic portion of the receptor.16 This in turn stimulates activation of guanine nucleotide exchange factors, which promote the formation of the active guanosine triphosphate (GTP)-bound form of Ras. To regulate this action, GAPs (GTP-ase activating proteins) stimulate the intrinsic GTP hydrolytic activity of Ras to form the inactive guanosine diphosphate–bound form of Ras.10,11,15,17,18 When Ras is in its active GTP-binding state, it is able to bind Raf.19,20 

The current study demonstrates differences in the staining patterns of some of these kinases of extracellular stimuli mediation. When examining for Raf-1 expression, the most basally oriented cell layers of verrucous carcinomas stained strongly with the antibody, and the more superficial layers did not stain or stained very weakly. The squamous carcinoma basal layers only occasionally stained with Raf-1, but the staining was never more than weak (0–1+). We also found marked differences in the localization of ERK-1 and ERK-2 expression between verrucous carcinoma and squamous cell carcinoma. In squamous carcinomas, there was generally uniform staining for ERK-1 and ERK-2 throughout the tumor. In contrast, verrucous carcinoma almost never showed staining of its most basal layer cells with ERK-1 and ERK-2.

It is unclear why there is such exquisite localization of the staining. Possibly there is a block in synthesis of ERK-1 and ERK-2 in the proliferating cells of verrucous carcinoma with a resultant accumulation of Raf-1, a protein that is situated further upstream in the cascade. It must also be realized that immunohistochemical studies only can localize cross-reacting antigen and do not necessarily indicate strength of enzyme activity.

It is possible that the staining of the suprabasalar portions of the epithelium is related to maturation of the cells. However, in none of the control tissues did there ever appear the level of intensity of Raf-1 staining as seen in the basal layers of the verrucous carcinomas. Mishima et al reported that there was staining of the most superficial epithelial cells in normal gingiva by ERK-1 and ERK-2.14 They felt this might be due to cellular concentration of proteins as the cells mature and lose glycogen. The photographs demonstrated by Mishima et al show much less intense or widespread staining as compared to our cases of verrucous carcinoma and resemble the staining pattern we observed in normal gingiva.14 In addition, if the increase in staining intensity was due to loss of glycogen or concentration of cellular proteins, we should not have found increased levels of Raf-1 expression in the basal layers with minimal to no expression in the superficial epithelial levels in the verrucous carcinomas.

Differences in localization of growth control molecules in verrucous carcinoma compared to squamous cell carcinoma have been reported previously. Anderson et al reported that verrucous carcinomas had transforming growth factor-β receptors in a membranous pattern, whereas the expression of the receptor in squamous cell carcinoma was predominantly cytoplasmic.21 These authors concluded that the membranous location of the receptor in verrucous carcinoma exposes it to the growth inhibitory control of transforming growth factor-β and may explain why verrucous carcinomas are less aggressive clinically.

Although verrucous carcinoma does invade, it does so by a pushing, downward proliferation and not in the sharply infiltrative manner of other squamous carcinomas. In an excellent review, Crowe and Shuler22 described the regulation of tumor cell extracellular matrix invasion and its dependence on the kinases of the MAPK family. The MAPK-driven activation of transcription factors results not only in cell division, but also in the transcription of factors necessary in regulating genes involved in extracellular matrix degradation and cell migration. Thus, differences in the site of localization or relative amount of MAPK-related kinase, such as low levels of ERK-1 and ERK-2 at the proliferating front of the tumor could be related to differences in the manner of invasion of these 2 tumors.

We did not investigate in this study whether any of the tumors carried ras mutations. H-ras mutations have been reported in both verrucous carcinomas and squamous cell carcinomas, with a higher incidence in the former.7 Regardless of whether ras mutations exist in any specific tumor, the extracellular signaling pathways appear to be potential candidates for tumor regulation.

It is possible that the differences of staining patterns seen with Raf-1 and ERK-1/ERK-2 could be used to help separate verrucous carcinoma and well-differentiated squamous cell carcinoma on small biopsies. The separation of verrucous carcinoma and verrucous hyperplasia would be of interest. This study did not address this issue; presumably, it would appear that in the areas of verrucous hyperplasia in our cases the expression pattern was similar between verrucous carcinoma and verrucous hyperplasia.

Additional studies will need to be carried out to determine the practical utility of these findings, and further investigation of the expression of these multifunction cell control molecules in verrucous carcinoma, squamous cell carcinoma, and other squamous lesions of the upper airway is needed.

This work was supported in part by the Frederic W. Stamler Professorship to Dr Robinson.