Periodontal Pathogens as Possible Etiologic Agents in Benign Mucous Membrane Pemphigoid

Dr. Allen Samuel Honigman

INTRODUCTION

The differential diagnosis on intraoral lesions that occur as red ulcerative vesiculobullous lesions include bullous pemphigoid, lichen planus, all forms of pemphigus, discoid and systemic lupus erythematoeus and benign mucous membrane pemphigoid. A definitive diagnosis is made from immunofluorescence staining characteristics of the sample. Bullous pemphigoid and benign mucous membrane pemphigoid are very closely related in their staining characteristics. However, bullous pemphigoid and benign mucous membrane pemphigoid are considered to be separate clinical entities based on their tissue involvements.

Bullous pemphigoid (BP), occurs mainly in the elderly, with the majority being over 60 years of age with no sex predilection. Clinically, the oral lesions of BP may present as transient and localized and more importantly, non-scarring. Large bullae on normal or inflamed skin or mucosa is a distinguishing clinical feature, with a negative 'Nikolsky sign'. There is little to no ocular involvement in BP (1,56,60). Histopathologically, acantholysis is absent, and vesicles are subepithelial and contain inflammatory inflitrate (1). Examination by electron microscopy shows that the layer. This is in sharp contrast to benign mucous membrane pemphigoid histological characteristics. The basement membrane may also be discontinuous. Direct immunofluorescence reveal IgG and C3 deposits along the basement membrane zone (BMZ) (1). The antibodies present are anti-BMZ and the patterns are usually 10096 diagnostic. Indirect immunofluorescence of those witn BP, show that anti-BMZ antibodies are present in the serum. When tested with anti-BMZ antibodies on normal skin biopsies, a linear deposition along the BMZ is observed. There is no correlation between antibody titer and disease activity. The current consensus of opinion is that BP is possibly an autoimmune disease, with an antigen in the BMZ the target of the autoantibodies. The antigen had been identified as a 2 20-240 kD polypeptide consitituent, bullous pemphigoid antigen (BPA) possibly associated with the hemidesmosomes (35,37,57). Immunoprecipitation studies have shown that bullous pemphigoid antibodies are bound to a protein of 230kD (45). The partial sequencing and cloning of this protein has been accomplished and more insight into this disease will be known when this has been completed. Research by Westgate et. al. suggest that the protein may be part of the linkage between the basement membrane and the connective tissue (57). The exact function of the antigen is not known, however, it may play a role in cell substrate attachment during epidermal migration and in keratinocyte differentiation (26). Studies, in-vitro and in-vivo,have demonstrated the pathogenicity demonstrated in an in-vitro model, that separation of the BMZ was a result of pemphigoid antibodies, leukocytes and complement (15). The antibodies are deposited, for the most part, in the lamina lucida, when examined in-vivo. Although the pathogenicity is not known, it is felt that binding of the autoantibodies to the antigens in the hemidesmosomes, may account for the disease process. The disease itself, is usually self-limiting.

Benign mucous membane pemphigoid (BMMP) or cicatricial pemphigoid (CP), first described by Wichmanns in 1793 (59), is a vesiculo-bullous mucocutaneous disease that primarily affects the oral and ocular mucosa. It is seen primarily in adults between 40-70 years of age. Unlike bullous pemphigoid, females seem to affected more than males at a 2:1 ratio. Intraorally, lesions are seen in dentulous regions. The clinical manifestations of BMMP, after initial bullae formation, include degeneration of the epithelium, which leads to desquamation and leaves the affected surface area raw, red and eroded. A positive Nikolsky's sign is characteristic of the mucosa around the lesion (36). Intraoral erythema may persist several weeks to months after healing of the original lesions. Scarring intraorally is rarely seen, but may be observed in the soft palate, uvula, tonsillar pillars and buccal mucosa. Ocular involvement is seen in approximately 80111f variant, has no mucosal lesions, but has scarring blisters in the head and neck area (1,29,31,33,42,43,54,56,60).

The histopathology of 8MMP lesions demonstrate subepidermal vesicles and bullae and no acantholysis. There is a separation of the basement membrane from the epithelium, with some portions of the basement membrane attached to the epithelium occasionally found. Examination by electron microscopy displays a separation of the basal lamina from the epithelium. A definitive diagnosis is made via direct and indirect immunofluorescence staining (50). These staining procedures show a uniform, continuous deposition of the marker between the underlying connective tissue and the basal epithelium. Direct immunofluorescence demonstrates IgG and complement components along the BMZ. Bernard et. al, demonstrated that the CP target-antigen is a 180 kD protein (3), while Niimi et. al determined that the antigen is a 160 kD protein present in the BMZ. Both studies showed that the 230 kD BP antigen showed some cross-reactivity with the CP sera studied (40). Bernard et. al. (1990) concluded that, although the target antigen that is involved in CP is similiar to the BP antigen, there is an unnatural expression in lamina lucida of the dermal epidermal junction in epithelium (4) (Fig 1). This may account for the scamog that is characteristic of this disease.

BMMP is assumed to be an autoimmune disease. Treatment for the disease is determined by the organs affected. Treaments include topical and oral steroid thereapy and, in some cases, dapsone (30,55). If there is an increase in severity, immunosuppressives may be indicated. Surgical removal of the affected area, antibiotic treatement and chlorhexidine rinses, can be of help to control the oral manifestation of the disease. Prognosis is better for the oral lesions than for ocular lesions.

The etiology of BMMP is still unknown. Initiation of the lesions has been attributed to stress, hormonal imbalances, local irritants, viral, bacterial infections and autoimmune disorders. An immunogenetic association has been suggested as a possible indicator to predispostion for the disease development (39). Interestingly enough, BMMP is limited to tooth-bearing areas, and * is rarely observed in edentulous regions. Anecdotal evidence relates the use of Peridex (chlorhexidine gluconate) as an alternate treatment of BMMP, with the lesions resolving in similar time periods as those on steroidal therapy. This, on the surface, may indicate a bacterial role in the pathogenesis of BMMP. Certainly, the inflammatory and immune response cells seen microscopically in samples indicate that more than an autoimmune response is responsible for the lesions. Possible mechanisms proposed here are:

1. direct bacterial invasion of the mucosa, similar to spirochetes in ANUG.
2. induction of an autoimmune response by a bacterial antigen acting as a hapten after phagocytosis and subsequent presentation by macrophage
   * Personal communique from Dr. E B. Kenney
3. a combination of both of the above.

The role of bacteria in the progression of periodontal disease has been well documented. The presence of the bacteria in the dental plaque is not the only factor that contributes to the progession of periodontal disease. Bacterial cell wall components, such as endotoxin, or lipopolysaccharides (LPS), present in gram negative bacteria, can contribute to the overall inflammation or by acting as an immunogenic stimulator. The bacteria may also release chemotactic factors that cause neutrophilic granulocytes to invade the area, thereby causing an increase in the release of host inflammatory products and the resulting inflammation and progression of the disease. Patients affected by localized juvenile periodontitis have a defect in their neutrophils that do not allow them to respond to these chemotactic factors, allowing the disease to progress at an accelerated rate (19). Research has shown that specific bacteria are intimately associated with various periodontal disease states. The gram negative species, Fusobacterium nucleatum, and Treponema spp. are associated with gingivitis. Cagnocvtophaga ochracea. spirochetes, Prevotella intermedia and Actinomyces spp. are observed in chronic gingivitis. The periodontopathogens, Porphyromonas gingivalis, Eikenella corroders. and Campvlobacter rectus, among others, are just part of the complex microbiota associated with adult periodontitis (6).

Periodontitis cases exist where the bacteria is specific and distinct and can be used for monitoring the progress of the disease. Actinobacillus actinomvcetecomitans is the most closely associated bacteria with localized juvenile periodontitis (LJP), although P. intermedia, C. ochracea and E. corroders, are also seen in this disease (6). Cultures taken from rapidly progressive periodontitis cases have demonstrated the presence of C. rectus, E. corroders, P . intermedia, P . gingivalis and A. actinomvcetecomitans (19). Biopsy samples of affected gingiva have demonstrated that bacteria is present within the periodontium (47).

Bacterial invasion into the periodontium has been shown to occur (47). Saglie (1977) found bacteria in the junctional epithelium in the most apical subgingival plaque through electron microscopy (49). Bacteria have been shown to be present in the vicinity of alveolar bone in advanced human periodontitis. In the apical wall of periodontal pockets, bacterial invasion was found to occur. In 1982, Saglie showed the bacteria, in the epithelial wall of deep periodontal pockets, showed a definite pattern of penetration, and not a random arrangement, as might be seen if they were introduced via surgery or sample processing (48). In 1982, it was shown Actinobacillus actinomvcetecomitans was within the gingival tissues in localized juvenile periodontitis (LJP), and there was an interaction between bacteria and leukocytes at the pocket epithelium surface and within the connective tissue (6). In 1983 Carranza et. al., showed that bacteria could be demonstrated within PMNs in the epithelium and the connective tissue, which gave further evidence to bacterial invasion and immune cell interactions in the gingiva (7).

Bacteria, upon being phagocytized by macrophages, are either completely degraded or partially degraded, with the proteins being presented for recognition by lymphocytes, on the macrophage or polymorhonuclear leukocyte cell surface. Tlymphocytes recognize small antigen fragments, while B lymphocytes may only recognize larger, intact fragments. Autoimmune disease occurs when the body's immune systems begins to react against its own tissues and result in pathological damage. It is possible that in autoimmune diseases, a bacterial antigen could be cross-reactive and bind to a T-lymphocyte. This in turn could induce and auto-reactive B-lymphocyte. This is similiar to Type II hypersensitivity, where the antibody is formed against a normal structures, such as a connective tissue component in pemphigus and pemphigoid (32,36). Another mechanism, can involve polyclonal stimulators, such as the Epstein Barr virus or lipolysaccharides, a constituent of periodontal pathogens, that may directly stimulate the B-lymphocytes. These both result in failure of immunological tolerance. In general, these reactions are directly or indirectly cytotoxic and entail antibody generation against cell surface or connective tissue antigens. Complement cascade activation, the classical pathway, by the formation of the immune complexes, is necessary for these cytotoxic event. Lysis and, or, tissue damage is mediated directly by complement, producing a ³membrane attack complex" or indirectly via the opsonization or chemotactic attraction of the phagocytic cells. Recent research has shown that short chains of amino acids, as small as 6 amino acid residues, can sensitize Tcells and cause an autoimmune disease (16). Many viral and bacterial proteins may actually share these short chain runs of amino acids with host proteins, which is known as molecular mimicry.

The use of labelled immunoassays such as the immunofluorescence assay (IFA) (14,17), the radioimmunoassay (RIA) and the enzyme linked immunosorbent assay (ELISA) have been used for the detection of antigens and antibodies in diseases in which antigen-antibody complexes are formed. The IFA utilizes a fluorescent label, such as fluorescein, that is conjugated to either an immunoglobulin or anti-species immunoglobulin for direct or indirect procedures respectively. After attachment is achieved, the cells are observed by fluorescent microscope or counted by flouroescence activated cell sorter (FACS). The RIA and ELISA are both solid phase immunoassays. Both utilized a labelled antispecies conjugate. However, the RIA's conjugate is labelled with a radioactive label, such as 1251 and then counted by a scintillation or gamma counter (53). The ELISA¹s conjugate is labelled with an enzyme, such as alkaline phosphatase, and requires the appropriate substrate to produce a color change, that is then read by a spectrophotometer that is set at the appropriate wavelength (12,13). Both are sensitive and quantitative assays. Use of avidinbiotin conjugates offers increased sensitivity (27). Titers for both the RIA and ELISA are determined by the last count or absorbance that is greater than the negative or normal absorbance. The similiarty between these assays is that an antispecies immunoglobulin that is conjugated to a label, which can be fluorescent, radioactive or an enzyme. The IFA is a labor intensive assay, while the RIA utilizes hazardous radioactive materials. The use of an enzyme conjugate gives the ELISA an advantage over the RIA, in this respect (28).

In recent years, a dot immunosorbent assay has been developed for the detection of certain diseases and allergies and for the identification and testing of monoclonal antibodies (5,16,23). Mealey et. al. used this assay to look at periodontal pathogen antibody levels in patients with periodontal disease and developed optimal parameters for the procedure (34). The procedure is carried out on a nitrocellulose membrane and requires on 1-2 uL of antigen, as compared with 200 uL of antigen in the conventional ELISA system. Essentially based on the ELISA, with serum and enzyme labelled conjugate additions, the DOTELISA relies on the formation of an insoluble product to form a characteristic dot of specific periodontal pathogens as possible etiologic factors of BMMP via light and electron microscopic biopsy sample examination and by DOT-ELISA assay to determine if these bacterial antigens are possible autoimmune inducers.

MATERIALS AND METHODS

Top

Biopsy samples of patients with diagnosed BMMP, were subjected to microscopic examination of their pathology slides to be certain that there was a definite, and clear split of the epithelium from the connective tissue layer in the BMZ. In addition, the area adjacent to the split, had to demonstrate normal structural integrity, i.e. no split. This was done until samples from ten patients were obtained. The biopsy tissue, embedded in paraffin, was then sectioned on a microtome and prepared for staining for light microscopic evaluation. In addition, 3 sections per sample were also cut placed onto round microcoverslips, in preparation for electon microscopic evaluation.

The staining procedures selected were: a standard hemotoxylin-eosin stain for examination of tissue structures and a Brown and Brenn stain for identifciation of gram positive and gram negative organisms. By examination under a light microscope, bacteria in the area of the split or embedded in the tissues, would lend evidence to support the hypothesis of a bacterial role in BMMP. The slides were examined at lOOX, 400X and lOOOX (oil immersion) magnifications, with photography done at each of the magnifications to show the various features of the sample for later examination.

The tissue samples were prepared for electron microscopic evaluation. Samples were deparaffinized and then dried with Peldri II (Ted Pella Inc. Cat. No. 1240, Redding, CA), which is a proprietary fluorocarbon used as a sublimation dehydrant, as an alternative to critical point drying. After drying of the samples was completed, the microcoverslips containing the samples were attached to EM platforms with a silver conducting paint (Pelco, Tustin, CA) and allowed to dry and kept in dessicator until ready for use. The specimens were then sputter coated with a gold palladium layer in a Hummer I apparatus, then examined under an ETEC Autoscan scanning electron microscope. SEM photomicrographs, at various magnifications, were taken for further evaluation.

Human serum samples were obtained from patients with diagnosed benign mucous membrane pemphigoid by UCLA Pathology Department. A total of 3 sera were tested. Normal sera,selected at random, was kindly provided by the UCLA Pathology Department, which was obtained during a yearly exercise in phlebotomy of the second year dental school class (Class of 1996). These normal sera, were used at negative controls, as the average age, 25 years, and clinical inspection of the students precluded them to have BMMP. In addition, normal sera samples were unlabelled, so that the identity of the student could not be detemined. All specimens were stored at -20 degrees C until used.

There were nine periodontitis associated pathogens tested for this research. The samples were outer membrane protein (OMP) preparations of the selected bacteria. The preparations were obtained from pure bacterial cultures and subjected to a French Press protocol to isolate the OMPs. Specimens were stored at -20 degrees C until used. The bacteria samples were as follows: Prevotella intermedia, Porphvromonas gingivalis. Candida albicans, Actinobacillus actinomvcetecomitans, Treponema denticola, Campvlobacter rectus, Fusobacterium nucleatum, Eikenella corroders, and Capnocvtophaga ochracea. These were kindly prepared and provided by Dr. Thomas Bramanti (UCSF Dept. of Stomatology).

A modification of the Dot-ELISA procedure as described by Honigman (24) and Mealey (34) was employed to aetect antibody in the serum samples. The antigens used were the nine periodontopathic organisms previously described. Laminin and Type IV collagen were used to represent the major components of the mammallian basement membrane. The following reagents were supplied in the Immunoblot³ Assay Kit form (BioRad 1706465 Hercules CA): TBS, Tween-20, color development buffer system, and alkaline phosphatase conjugated goat anti-human IgG (H+L) and prepared according to supplied instructions.

The nitrocellulose membranes (No. 162-0117, 0.45 um; Biorad Richmond, CA) were placed into a dish with Tris-buffered saline (TBS) for 5 minutes, then placed onto a Whatman #5 chromatography paper to remove excess TBS. The paper was then placed into a Bio-Dot 96 well manifold (Bio-DOT apparatus 84BR20063; Richmond, CA) (Fig. 2) and secured into place. Due to the high protein binding capacity of the nitrocellulose, the membrane was handled carefully with forceps. Antigen was then carefully dotted into each well (duplicate samples) in 1-2 uL volumes and then allowed to dry for 1 hour at room temperature. After drying, the non-specific binding sites of the nitrocellulose were blocked by adding 100 uL of 5bovine serum albumin (BSA in TBS) (Albumin, bovine, 98-99albumin; Sigma Chemical Co.; St. Louis, MO) at room temperature for 1 hour on a shaking platform. After incubation, the blocking solution was aspirated off and the wells were washed once with TTBS (TBS + 0.05Tween) and gently agitated for 10 minutes, and then aspirated off again. Sera, diluted 1:50, were prepared in 1BSA-TTBS, was added ( 100 uL) to the appropriate wells. This incubated with gentle agitation for 1 hour at room temperature. The sera were then aspirated off and the membrane washed 2 times with TTBS for 5 minutes/wash at room temperature. Alkaline phosphatase-conjugated goat antihuman (GAH) antibodies, optimally diluted to 1:3000, in 1BSATBS was added ( 100 uL) to each well, then incubated for 2 hours at room temperature with gentle agitation. The conjugate solution was then aspirated off and the wells were again washed 2 times in TTBS for 5 minutes/wash. The membrane was then removed from the manifold and washed for 5 minutes in TBS, so that residual Tween-20 from the membrane surface. After alkaline phosphatase developer preparation, the membrane was immersed into a dish with the developer. If a precipitate formed quickly, then the solution was aspirated off and fresh developer added. After approximately 20-30 minutes, or when the background started to develop, the membrane was immersed into double distilled water for 10 minutes to stop the reaction. Sera causing the development of a well-defied purple dot on the membrane was relatively quickly were considered strongly positive (+++). Samples that gave a purple color, but not as intense were graded as moderate (++) and mild (+). Samples with a color intensity between positive and negative were considered weakly positive (+/-). Appropriate controls for positive and negative results were run in each set of assays. The total time for the process was approximately 7 hours (Fig 3), (Fig 4)

The nine periodontal pathogen antigens were used in the DOT-ELISA to detect antibody in sera from 3 patients with diagnosed BMMP and the 3 normal samples . Bullous pemphigoid serum (No. BP116 The Binding Site, Birmingham England), which is a pooled sample, was used as a control, as this disease is well characterized.

The nine periodontal pathogen antigens were employed in the DOT-ELISA for cross reactivity assays of antibodies to laminin and type IV collagen. The three BMMP sera, normal sera and bullous pemphigoid serum, were cross-reacted with laminin (No. L2020 Sigma Cell Culture, Sigma Chemical Co. St. Louis, MO) and Type IV collagen (No. C0543 Sigma Cell Culture Sigma Chemical Co. St. Louis, MO) antigens to determine if cross reactive antibodies exist. Both the laminin and Type IV collagen were obtained from the basement membrane of an Engelbreth Holm Swarm mouse sarcoma. Specific mouse monoclonal antibodies to laminin, (No. L8271 Sigma humunoChemicals, Sigma Chemical Co. St. Louis, MO) and Type IV collagen (No. C1926 Sigma ImmunoChemicals, Sigma Chemical Co. St. Louis, MO) were substituted for human sera in cross-reactivity studies, with an alkaime-pnospuatase conjugated goat antl-mouse (uAM) lgo (H+L) (No. 170-6520, Bio-Rad Hercules, Ca) used as the secondary antibody.

RESULTS

Top

Selected cases were examined using light microscopy. The hemotoxylin-eosin staining showed the distinct split between the connective tissue layer and the epithelial layer at the basement membrane zone. The split could be followed to a point of normal epithelial cell attachment. At higher magnifications, red blood cells in the split area were observed, as well as infiltration of inflammatory cells in the region of the split. There was no acantholysis observed in the epithelial layer. Occassionally, minute dark staining elements were found in the region where the transition to a normal structure occurred. When examined at 1000X (oil immersion), it was still not possible to accurately determine if these elements were bacteria, bacterial remnants, cell debris, or contamination from the staining process. No bacteria were observed intracellularly or within the connective tissue of epithelial layers (Fig 5), (Fig 6) .

The Brown and Brenn staining technique was utilized to identify Gram positive and Gram negative organisms. Once again, at high magnifications, there were no discernible bacteria observed, although some dark staining elements could be made out at the same position, seen on the hemotoxylin-eosin stained sections. (Fig 7), (Fig 8)

The biopsy samples were systematically searched for evidence of bacterial penetration into the epithelial and connective tissue layers. The connective tissue and epithelial layers were easily distinguishable, with the layers distal to the split transition demonstrating a more uniform and organized appearance than the area mesial to the split transition. Red blood cells were seen where the split was occurring. In the region where the split blended with normal epithelium, there were some particles seen which appeared as if they could bebacteria, bacterial remnants, cell debris, or contamination from processing. However, no obvious bacteria could be identified, even at magnifications up to 5000X. It was difficult to characterize the inflamatory cells with scanning electron microscope (SEM) (Fig 9), (Fig 10), (Fig 11).

The serum baseline evaluations of the BMMP positive, bullous pemphigoid and normal sera are summarized in Table ( 1) and (2) and in (Fig 12) and (Fig 13) respectively. The patterns of positive dots to the periodontal pathogens were varied among each of the sera tested. In the BMMP and BP sera, Prevotella intermedia and Candida albicans gave the most reactive (+++) results, moderate (++) results from Cagnocytophaga ochracea and Campylobacter rectus. The other bacteria gave mild (+) to no reaction (-). The sera from the normal patients showed a strong positive (+++) reaction with Fusobacterium nucleatum and Candida albicans. Prevotella intemedia gave a moderate (++) reaction. The remaining bacteria, again, varied in intensity from mild (+) to negative (-). Controls that were run for each for the assays had the predicted results and confirmed the assay and reagents worked properly.

The cross-reactivity assay results are seen in (Fig 14) and summarized in Table (3). A strong positive (+++) result was seen in BMMP serum from patient 1 when reacted with the laminin and Type IV collagen. The bullous pemphigoid serum reacted strongly to the Type IV collagen, but only mildy (+) to the laminin. BMMP sera from patients 2 and 3 only gave mild(+) to weak (+/-) reactions to the laminin and Type IV collagen. The control sera from the normal patients showed moderate (++) to mild (+) reactivity with the Type IV collagen, and a weak (+/-) reaction with laminin. When the anti-laminin and anti-Type IV collagen antibodies were reacted against the periodontal pathogen antigens, strong positive results were seen with the following bacteria: Candida albicans. Cagnocvtophaga ochracea and Fusobacterium nucleatum and A. actinomvcetecomitams. Mild (+) to moderate (++) results were observed with the remaining bacteria. There was no cross-reactivity with laminin vs. anti-Type IV collagen antibody (well F9) and Type IV collagen vs. antilaminin antibody (well H9), which was as predicted and acted as a control for the monoclonal's specificity. Controls run during the assay gave predicted results and confirmed the assay and reagents worked properly.

PERIODONTICS INFORMATION CENTER