Expression of beta-2-microglobulin by normal, benign and

malignant Oral Epithelia

A.H.M. Shabana, BDS, MSc, PhD*
Laboratoire de Biologie-Odontologie, Instltut des Cordeliers, Escalier E, 2eme etage,
15 ruede L'Ecole demedecine, 75006, Paris, France.

Beta-2-microglobulin is a cell surface protein that plays an important role in cell-cell recognition and interactions. The cell expression of this protein was investigated on both normal and pathological oral epithelia using immunortistochemical methods. In normal and benign reactive epithelia, beta-2 microglobulin was detected at the cell surface. In contrast, carcinoma cells and the associated non-malignant surface epithelial cells showed aberrant expression of beta-2-microglobulin in the form of cytoplasmic localization and depletion at the cell membrane. This localization indicated that these cells maintained the synthesis of the protein. The depletion from the cell surface might be due to masking by receptors such as viruses or autoantibodies. The absence of detectable surface beta-2- microglobulin might be a mechanism by which the malignant cell escapes the immune surveillance.

In 1968, Berggard and Beam1 discovered a small molecular weight protein (MW = 11,815) during their investigations on macromolecules present in urine. They called it beta-2-microglobulin (B2M). This protein is a single polypeptide chain made of 18 amino acids and devoid of carbohydrates.2 The interest in B2M began when a significant homology between it and the constant domains of IgG light chain was demonstrated.3-4 However, no cross-reaction with the precipitating antibodies has been found. Although B2M is encoded in a different chromosome from that of HLA antigens, it appears to be necessary for the synthesis and expression of class I antigens.5-6 Recent studies have shown that B2M is important for the conformation of the V2 domains of the histocompatibility antigens, suggesting a roie of B2M in the folding of class l ligand binding site.7-8 Further, in vitro studies showed that class I molecules synthesized in the absence of B2M have
aberrant conformations and are non-functional.9 These findings indicate that B2M plays a role in cell recognition and interactions.
B2M has been found in all nucleated cells that carry HLA antigens. Accordingly, it was not found in mature red blood cells and trophoblastic cells of the placenta.10 This protein was also detected in almost all body fluids such as urine, serum, saliva, cerebrospinal, synovial, amniotic, colostral and seminal fluids.1,11-12 Elevation and depression of serum levels of B2M have been reported in several diseased states.13-17
Immunohistochemical studies have shown loss of B2M in tissue sections of basal cell carcinoma,T8-19 partial loss in epidermal carcinomas, and pre-malignant lesions.19-21 In oral squamous cell car­ cinoma patients, Scully22 reported significant increase of serum level of B2M in 26 patients com­ pared with normal controls. He found no signifi­ cant difference between leukoplakia patients and normal controls. In a short communication, Scully et al23 showed partial loss of B2M in oral squamous cell carcinoma by using the immunohistochemical method. The present paper describes the expres­ sion of B2M in oral squamous cell carcinoma and compares it with that in benign and normal oral epithelia.

In the present study, 67 specimens from the oral mucosa were examined for the in-situ expression of B2M. Fifty-eight were selected from the files of the Pathology Department, Eastman Dental Hospital, London, representing a range of benign and malig­ nant oral epithelial lesions. Nine other specimens were obtained from clinically healthy oral mucosae of patients undergoing minor oral surgery. All specimens were fixed in formol saline and routinely processed to paraffin wax blocks. In each case, the diagnosis was confirmed by examining new sec­ tions cut from the tissue blocks used for the immunohistochemical study. The benign lesions comprised a variety of reactive and inflammatory lesions and simple traumatic ulcers as shown in Table 1. The twenty-eight carcinomas were graded for histodifferentiation according to the WHO classification of tumors.24
The antibodies used in the immunohistochemical methods23 were rabbit anti-human B2M (DAKO) and FITC-labelled goat anti-rabbit anti­ bodies.* In the peroxidase-anti-peroxidase technique, the first antibody was detected with a goat anti-rabbit IgG, and the reaction site was detected by a rabbit anti-peroxidase complex (DAKO). Normal rabbit serum (DAKO) was used to replace the primary specific antibody in control sections, and to block non-specific binding sites in the test sections.Serial sections 5 pun thick were incubated over­night in a hot-air oven at 37°C. The sections were then dewaxed, rehydrated, washed, and incu­bated with 0.5% trypsin in PBS, pH7.6, at 37°C for 15 minutes. After washing, the sections were then processed for either the immunoperoxidase or the immunofluorescence technique using methods described previously.25
In the immunofluorescence technique, the sections were incubated with the primary antibody, washed and then incubated in goat anti-rabbit FITC labelled antibody diluted 1/25 in PBS containing 10% bovine serum albumin at 37°C. After washing, the sections were mounted with aqueous moun-tant, covered and examined. Positive reaction in epithelial cells was then arbitrarily quantified as homogeneous (+ + +) where B2M was detected in a diffuse distribution; focal (++) when distributed in at least 50% of the epithelium; weak (+) when labelling was 10 to < 50% of the epithelium and negative (-) when < 10% labelling was noticed.
Non-parametric Chi-square test was used in this study to compare differences between groups and significant values were accepted where the proba­ bility of no difference (P) was less than 5%.

In the normal epithelia, a consistent distribution for B2M was found in the different specimens. The epithelium was labelled in the basal and spinous cell layers. The labelling was stronger in the deeper cell layers than in the more superficial ones. The granular and keratinized layers showed weak to negative reactions. This distribution was seen in both immunofluorescence and immunoperoxidase techniques. The connective tissue cells reacted weakly as compared with the stronger epithelial reaction [Fig. 1]. The omission of the first antibody resulted in negative reaction in both the epithelial and connective tissue cells.
The results of the immunohistochemical localiza­ tion of B2M in the pathological tissues are presented in Table 2. In benign lesions, epithelial labelling varied from one specimen to another.  In some lesions the reaction was homogeneous, in others focal distribution was found.
In the fibroepitheiial polyps, homogeneous reac­ tion predominated in 11 of 18 specimens showing principally the same reaction as in the normal epithelium. Focal reaction was found in six cases and weak reaction in one case.
In the fibrous epulides, B2M was homogene­ously distributed in 2 out of 6 cases. Focal reaction was registered in 4 cases that were ulcerated. This depletion in B2M expression in ulcerated lesions was also seen in the 4 lesions of non-specific ulcers. Often the focal depletion was found at the epithe­ lial edges of the ulcerated area. The epithelium distant from the ulcer was homogeneous in 3 cases and focal in one. In the pyogenic granuloma and the denture induced granuloma, focal reactions were registered. These 2 lesions were also ulcerated; however, the depletion in B2M expression was not restricted to the ulcerated area. No significant correlation was found between the expression of B2M and either age, sex or site using the non-parametric Chi-square test with Yates correction when necessary.
Oral squamous cell carcinoma showed signifi­ cant reduction in the expression of B2M. Out of the 28 lesions, 3 were homogeneous, 7 focal, 6 weak and 12 negative. In the 3 patients from whom 2 tumors were investigated, 2 cases had identical dis­ tribution in the 2 specimens. In the third patient, 1 specimen showed weak reaction and the other focal reaction. This patient was grouped according to the specimen showing lower value for B2M expression grade.
Comparing the expression of B2M by benign and malignant epithelia, the difference was highly significant (P < 0.001). In squamous cell carcinoma, B2M was also detected in the cell cytoplasm by both immunofluorescence and immunoperoxrdase techniques. This cytoplasmic staining was not detected either in the normal group or in the group of benign lesions. Cytoplasmic staining was also detected in the surface oral epithelium of car­ cinoma specimens which happened to be present in 4 cases [Fig. 2],
In squamous cell carcinoma, although no correlation between B2M expression and sex or age of the patients could be found, certain differences were found in association with lymph node involvement, tumor size and tumor differentiation. Depletion of B2M was more often found in cases with lymph node involvement. Negative reaction was found in 8 out of 12 lesions (67%) from patients whose lymph nodes were involved, while the cor­ responding figure for the 16 lesions from patients with no lymph node involvement was 11 (59%). In well-differentiated carcinomas (Grade I), 6 out of 10 lesions showed homogeneous to focal reaction [Figs. 3,4]. This reaction was found in 2 out of 11 grade II carcinomas [Figs. 5,6] and in 2 out of 7 grade III carcinomas [Figs. 7,8].

B2M has been reported as a cell surface product of all nucleated cells. Cell culture studies have shown that lymphocytes, macrophages, and endo­thelial ceils produce B2M.26-27 Governa and Biguzzi28 reported the expression of B2M by these cells in vivo, using immunofluorescence technique. In the present study this was confirmed by both immunohistochemrcal and immunofluorescence techniques. These cells were more prominently labelled in sections from oral squamous cell car­ cinoma, which might suggest activation of the immune response in carcinoma. In the normal oral epithelium, the distribution of the protein was homogeneous in basal, spinous and to a lower extent in the granular cell layer. The present results showed that B2M was detectable in all benign lesions and weak reaction was recorded in only one case (3.3%). Focal distribution was more often found in ulcerated lesions. This depletion may be explained by cell migration at the edges of the ulcers.
In the present study, the difference in expression of B2M by malignant and benign cells was highly significant. Moreover, 64% of the carcinomas showed near complete loss of the protein from the cell membrane; such reaction was found in only one benign lesion and none of the normal specimens. The malignant cells also showed cytoplasmic reaction specific for B2M in some cases. Cytoplasmic localization was previously reported in a study of breast tumors.29 However, they found the fluorescence of the surface stronger than that of the cytoplasm. In the present study, cytoplasmic stain was often as strong as the cell surface stain by both immunofluorescence and immunoperoxydase techniques [Figs. 7,8]. Furthermore, cytoplasmic stain was also detected in non-malignant cells, in the oral epithelium adjacent to invasion. The detec­ tion of B2M intracellularly may be a result of accumulation of the protein and/or increase in its synthesis. The cytoplasmic localization in malignant and adjacent non-invasive oral epithelia suggests that these epithelia were subjected to common biological events, and might be a useful marker for detection of pre-invasive epithelium. It will be interesting to find out if the cytoplasmic localization of B2M can be found in oral premalig-nant epithelium especially since one of the previ­ ous morphometric studies on leukoplakia reported increase in the size of cells and nuclei of both inva­sive and pre-invasive basal cells.30
The loss of the antigen from cell surface may be explained by the presence of autoantibodies.31 In the carcinoma group, the expression of B2M is cor­ related with tumor differentiation. In grade I car­ cinomas, 60% had homogeneous or focal reaction whereas in grades II and III, only 22% had similar reaction. These results suggest that depletion of B2M correlates with poor prognosis of oral car­ cinoma.
In the present study, B2M expression was more marked in small tumors than in larger ones, which provides further support to the findings of Wenner-berg et al.32 As the histological grading of tumor dif­ ferentiation is rather subjective and requires certain experience, the expression of B2M may prove to be useful as a complementary method in the assess­ ment of tumor differentiation and growth. The biological role of B2M in carcinoma and precan­ cerous states has to be further elucidated. Nevertheless, the depletion of B2M or its altered expression in the carcinoma cell may play a role in the malignant cell escape from the immune mechanism and consequently suggests poor prog­ nosis.

 
The author is grateful for the valuable discussions with Professor I.R.H. Kramer and Dr. L. Ivanui, Eastman Den­ tal Hospital, London and the secretarial work of Mme F. Poiraud. This work was supported financially by the Ministry of Higher Education, Arab Republic of Egypt.