Immunohistochemical Determination of Tumor-Associated Antigens in Oral Epithelial Dysplasia and Squamous Cell Carcinoma

B. KH. Moghadam DDS DScD,* I. Yazdi DMD FICD,** C. Cobb DDS MS PhD,^H

R.E. Shultz DDS,^I B.L. Ferguson DDS^I

Departments of *Diagnostic Sciences, ^HOral Biology and Periodontics, and ^IOral Surgery, University of Missouri-Kansas City, School of Dentistry, Kansas City, Missouri, U.S.A., and **Department of Oral Pathology, Tehran University of Medical Sciences, Tehran, Iran

• Background-Malignant cell-lines derived from oral squamous cell carcinoma (SCC) express messenger RNA for transforming growth factor alpha (TGF-a ) and parathyroid hormone-like proteins (PTHLP) in culture. The purpose of the present study was to determine the precise immunohistolocalization and tissue distribution of these bioactive peptides in paraffin-embedded sections of normal and diseased oral mucosa and to relate them to their cell-lines occurrences reported in the literature.
• Methods-Formalin fixed paraffin embedded sections of twenty-two primary oral squamous cell carcinoma, 18 benign and dysplastic oral epithelial lesions and 15 specimens of normal buccal mucosa were included in the study. Expression of TGF-a and PTHLP was determined immunohistochemically.
• Results-A detectable level of TGF-a and PTHLP was found in normal and benign oral mucosa. A higher level of expression was demonstrated in dysplastic oral lesions and the greatest staining level was found in oral SCCs. Malignant epithelial cells at the peripheral margins of the invading tumors demonstrated strong positive staining for both antigens.
• Conclusions-The in situ expression of TGF-a and PTHLP by normal and diseased oral mucosa was detectable immunohistochemically in formalin fixed paraffin embedded sections of oral mucosa and their occurrences were not altered by fixation methods used to process these tissues. Secondly, the expression of TGF-a and PTHLP was a reflection of dysplasia and loss of differentiation in oral epithelial cells. Thus, the immunohistochemical determination of these bioactive peptides may be valuable as an adjunct to conventional histology for detection of potential malignancies in histological examination of oral mucosal diseases.

Introduction

mmunohistochemical detection of rapidly growing cells is valuable for assessment of malignant and premalignant conditions in addition to the clinical and pathological criteria. Transforming growth factor alpha (TGF-a ) is a polypeptide found in various malignant cell-lines and is implicated as an autocrine growth factor for induction and maintenance of malignancy.¹ TGF-a messenger RNA has been detected in embryonic cells, in normal adult cells and in cell-lines derived from many solid tumors including oral squamous cell carcinoma.^2-6 TGF-a also induces bone resorption in vitro and is suggested as a cause of malignancy associated hypercalcemia in humans.^6,7

Parathyroid hormone-like proteins (PTHLP) are other secretory components of malignant cells that can induce bone resorption and hypercalcemia.^8-11 These bioactive peptides have been detected in both normal and malignant keratinocytes and are potential markers for detection of invasive malignant cells.^8-18

Despite the numerous studies on characteristics of TGF-a and PTHLP in human cell-lines, their immunohistochemical profile in malignant oral tissues has seldom been reported.³ Thus the purpose of the present study was to determine the tissue distribution and the precise immunolocalization of these antigens in paraffin-embedded sections of oral malignant and dysplastic lesions, and to relate them to their cell-line occurrences as reported n the literature. Normal oral mucosa and benign mucosal lesions were included for comparison.

Materials and Methods

Tissue specimens:

Twenty-two samples of primary oral squamous cell carcinoma and eighteen specimens of hyperplastic/dysplastic oral lesions were included in the study. The SCCs were graded as well differentiated (n=14) moderately differentiated (n=6) and poorly differentiated (n=2). The hyperplastic/dysplastic lesions were graded as benign acanthosis/hyperkeratosis (n=9), mild to moderate dysplasia (n=7), severe dysplasia (carcinoma in situ) (n=2) and papilloma (n=1). Fifteen samples of normal buccal mucosa were included for comparison.

Immunohistochemistry:

Five micron sequential sections from each paraffin-embedded blocks were deparaffinized by heating for 24 hours at 56? C and subsequently immersing in xylenes and graded alcohol changes two times each at 10 minute intervals. The sections were then treated with 0.1% hydrogen peroxide to block endogenous peroxidase activity. Following a ten minute wash in distilled water, the sections were incubated for twenty minutes with 0.05% saponin (Sigma) and then with protein blocking sera (10% normal goat serum) to reduce non-specific reactions. The initial step of immunohistochemical staining involved incubation of tissues with 1:10 dilution of each primary monoclonal antibody overnight. The primary antibodies used were anti-TGF-a (Ab-2), and anti-PTHLP (Ab-1), obtained from Oncogens Science, Inc, San Diego, CA.

The next step involved incubation of the sections, first with multilink biotinated secondary antibodies, and then with peroxidase conjugated streptavidin label (Bio Genex super sensitive multilink kit, Sam Ramon, CA.). Between each step, the sections were washed for five minutes in two changes of phosphate-buffered saline (PBS; pH 7.2). Staining was accomplished by incubating the tissues with AEC chromogen for 5-10 minutes. Following a five minute wash in distilled water, the sections were counterstained with hematoxylin and mounted for microscopic examination.

Controls:

Parallel control sections were processed with each batch of tissues. Positive controls were sections of normal skin and breast cancer, which are known to express TGF-a ^1,2 and PTHLP.^12,14 To control the specific cytoplasmic binding of the primary antibodies, parallel sections of malignant tissues were treated with anti p₅₃ antibodies (Ab-6, Oncogene Science Inc. San Diego, CA.) with the same concentration as TGF-a and PTHLP. These antibodies bind to the p₅₃ proteins and demonstrate strong nuclear staining in malignant cells.¹⁷ Negative controls were non-immune sera and PBS in place of primary antibodies.

The intensity of the staining was assessed by using a semiquantitative scoring: no staining (-), barely staining (?), weak staining (+), moderate staining (+2), strong staining (+3), and highly strong staining (+4).

Results

Normal tissues:

All samples of normal oral mucosa (n=15) demonstrated a detectable level of TGF-a and PTHLP. The staining intensity was relatively weak (+) but consistent for both antigens and was found in all epithelial cells in various anatomical zones. Increased staining intensity (+2) of basal cells and epithelial components of minor salivary glands was observed in many sections.

The staining pattern of normal skin taken from the palm of the hand (positive control) was similar to that of normal mucosa but had a greater intensity (+2). All cutaneous keratinocytes stained for both antigens throughout the epidermis.

Benign and dysplastic mucosal lesions:

Seven of the nine benign mucosal lesions without dysplasia had only detectable levels of TGF-a and PTHLP (+) and two samples had a relatively stronger pattern of staining (+2). Both of these lesions stained positively in the most superficial and deep epithelial cell layers for TGF-a and PTHLP. One sample was a gingival lesions in the vicinity of tooth #32 with hyperkeratosis and mild hyperplasia. The other specimens was hyperkeratotic lesion of buccal mucosa with acanthosis. The gingival lesion demonstrated positive staining for PTHLP in nuclear remnants in the keratin layer, unstained keratin layer and staining of a cellular band beneath the keratin layer.

All dysplastic oral mucosal lesions (n=7) with mild to moderate dysplasia had positive focal staining for TGF-a and PTHLP with moderate intensity (+2) in distinct regions of tissues with dysplasia. There was an increase in the distribution and intensity of staining (+3) in diaplastic epithelial cells in two mucosal lesions with severe dysplasia categorized as carcinoma in situ bordering on early SCC (Fig 1). The mucosa adjacent to a section of SCC demonstrated a moderate positive reaction (+2) throughout the epithelium for both antigens.

Oral malignant lesions:

All primary oral SCC samples demonstrated strong positive staining (+4) for TGF-a (Fig 2) and PTHLP (Fig 3) in malignant cells across the entire specimen. The reaction was particularly strong at the margin of invading cords and nests of tumor cells and in highly anaplastic regions of the same tumor with different degree of differentiation. Individual malignant cells in the group of well-differentiated SCCs (n=14) with numerous keratin pearls stained strongly positive but the keratin pearls of varying sizes remained relatively unstained. (Fig 4) The staining pattern of TGF-a and PTHLP was cytoplasmic in all samples examined in this study. Control sections treated with anti-p53 antibodies had a selective nuclear staining pattern in 19./22 oral SCC sections examined. These sections had a variable number of p53 positive cells. The positive cells were either present throughout the tissues or were scattered as focal positive areas among negative cells (Fig 5). The intensity of nuclear staining varied from faint to quite strong reaction.

Staining was not demonstrated in any negative control sections treated with non-immune sera or with PBS in place of primary antibodies.

Discussion

Previous studies have shown that cell-lines derived from both normal and malignant human keratinocytes synthesize TGF-a and PTHLP in culture.^1-5,12-16

This current study determined the in situ expression of these bioactive peptides and their tissue distribution in paraffin embedded sections of normal, benign and malignant oral mucosa.

The existence of TGF-a in each tissue specimen was evident by staining of the epithelial cells within different anatomical zones. The staining intensity of normal tissues and of benign mucosal lesions was less than that of severe dysplastic lesions. The malignant epithelial lesions had the highest expression level of TGF-a localized at the peripheral margins of the invading tumor cords and in poorly differentiated malignant cells. These results are primarily in agreement with the reports of TGF-a MRNA occurrence in normal and malignant kertinocytes cell-lines in culture^1-5 and illustrate the endogenous expression and tissue distribution of TGF-a in oral mucosal specimens. Secondly, the current results confirm that oral malignant epithelial cells commonly express a high level of TGF-a at the marginal border of tumor which could accelerate proliferation and invasion of malignant cells. In addition, results of these immunohistologic investigation extend the use of paraffin-embedded tissues in detection of TGF-a positive cells in various mucosal lesions. Over expression of TGF-a in current study was related to increasing dysplasia and loss of differentiation in oral epithelial cells.

The in situ expression of PTHLP in oral mucosal samples was demonstrated by the widespread and strong staining of malignant cells. While a detectable level of PTHLP was found in normal oral mucosa, a high level and diffuse staining was observed in individual malignant cells and in cells arranged in groups and clusters making up the bulk of the tumor. These immunohistochemical determinations are consistent with the occurrence of the bioactive peptide in cell-lines of normal and malignant human keratinocytes^8,14,15 and indicate that PTHLP is highly over expressed in oral malignant conditions. The literature suggest that the wide distribution and the endogenous synthesis of PTHLP by keratinocytes imply an important functional role for the peptide in epithelial cell physiology.^14,15 Some recent studies have shown that the peptide stimulates calcium transportation and keratinocyte differentiation in culture.¹⁸ In addition, PTHLP has the ability to cause local bone resorption and may contribute to tumor expansion by eroding of bone and allowing the metastasis to grow.^8-11 Results of the current study imply that the local expression of PTHLP by normal and malignant oral mucosal cells is detectable immunohistochemically in paraffin-embedded sections and enhanced expression of PTHLP is a reflection of malignancy. Since the peptide is over expressed in malignant conditions, the assessment of PTHLP proteins may be a useful additional confirmatory marker for detection of malignant cells within surgical specimens.

In conclusion, the apparent immunohistochemical determination of TGF-a and PTHLP synthesis in oral mucosal lesions and their increased expression in malignant condition indicates that these markers have a potential diagnostic value in histological examination. In addition, the stability of these antigens is not affected by formalin fixation and the methods used to process the paraffin-embedded tissues.

Acknowledgement

We thank Ann Marie Corry, Harry W. Mueller for help in providing references and checking citations, Lisa Cabra for her word processing preparation of this manuscripts, David A. Coats for preparation of the illustrations and Azarakhsh Mokri for assistance in checking the finalized manuscript.

References

AIM Home|Table of Contents