Immunological Methods for the Diagnosis of Oral Mucosal Diseases

S. Sun; B. Zhong; W. Li; X. Jin; Y. Yao; J. Wang; J. Liu; H. Dan; Q. Chen; X. Zeng


The British Journal of Dermatology. 2019;181(1):23-36. 

In This Article

Detection Results of Immunological Methods

Pemphigus Vulgaris

Pemphigus is a serious autoimmune blistering disease of skin and mucous membranes. The most common and serious type of pemphigus is PV with intraepithelial flaccid blisters, which usually presents with the oral mucosa first affected or lesions limited to the oral mucosa.

DIF is the gold standard in the diagnosis of PV, which reveals intercellular deposition of IgG and/or C3 or in combination with IgA and IgM in the epidermis.

The most common assays used to detect PV antibodies are IIF and IB.[27] Substrates such as monkey oesophagus and lip, normal human skin, guinea pig oesophagus, rabbit oesophagus and human foreskin have been reported in IIF. The diagnostic criteria for pemphigus published in European guidelines included IIF on monkey oesophagus or human skin.[58] A study proposed that using human cervix as the substrate for PV can achieve high sensitivity (90·5%) and specificity (96·2%).[28] IgG4, the Dsg-specific autoantibodies predominant in PV,[59] enlightened us that IgG4-targeted therapy may be an option.

The anti-Dsg3 and anti-Dsg1 antibodies are related to oral mucosal lesions and skin lesions, respectively. Furthermore, the anti-Dsg1 antibodies may indicate poor prognosis. In addition, anti-E-cadherin antibodies at low levels[60] and other antibodies in patients with PV have been found (Table 1 shows autoantigens).[61–65] Dsg3 and Dsg1 are the major autoantigens in patients with PV. Intraepidermal blisters are formed because the conformation and stability of autoantigens are destroyed by the specific binding of autoantibodies and antigens.

ELISA tests usually show that serum samples of patients with PV are positive for Dsg3, with or without reactivity against Dsg1.[66] A study indicated that using ELISA and DIF with results of positive anti-Dsg3 autoantibodies and positive DIF can be predictors of relapse for patients during clinical remission.[67] However, it is only applicable to patients with detectable antibodies against Dsg; we also discuss below non-Dsg antigens in PV. Studies have clearly shown that anti-Dsg1 levels can be more indicative of PV activity than can anti-Dsg3 levels.[68,69] Furthermore, ELISA is better than IIF for monitoring disease activity and response to treatment. Initial intercellular antibody titre is a useful predictor for time of disease remission, and initial Dsg3 antibody levels may predict total disease duration in PV.[70] Although ELISA is highly sensitive, expensive ELISA systems have led to increased use of immunofluorescence in laboratories.

Novel targets of PV can be identified by protein microarrays. A few studies showed several new (non-Dsg) antigens with high specificity to PV, such as desmocollins 1 and 3, mitochondrial proteins, human leucocyte antigen molecules, hSPCA1 (the calcium-transporting ATPase type 2C member 1, encoded by ATP2C1), several muscarinic acetylcholine receptors (mAChR1, mAChR2, mAChR4, mAChR5 and pemphaxin), thyroid peroxidase, CD2, CD31, CD33, CD36 molecules, etc.[53–55,71,72] However, the pathogenicity of these autoantibodies remains to be confirmed. Another study demonstrated that autoantibody reactivity was associated with disease activity and morphology with varied relevant antigens in different patients with PV.[56]

Clinical manifestations and histopathology may lead to misdiagnosis. A 49-year-old Chinese man with a 10-month history of recurrent oral ulcers presented clinically with multiple small erosions with yellow floors (Figure 3a, b). The Nikolsky sign was negative. Histopathology of normal-appearing gingiva revealed normal tissue. Doctors initially suspected the disease was recurrent aphthous ulcer. However, DIF with intercellular deposition of IgG and C3 in the epidermis resulted in a diagnosis of PV (Figure 3c, d).

Figure 3.

Clinical features and direct immunofluorescence patterns for the 49-year-old man (pemphigus vulgaris). (a, b) Multiple small erosions with yellow floors on the buccal mucosa and tongue (histopathology of normal-appearing mucosa revealed normal tissue). (c, d) Direct immunofluorescence showed intercellular deposition of IgG and C3 in the epidermis.

Paraneoplastic Pemphigus

PNP, a rare autoimmune mucocutaneous blistering disease, is commonly associated with neoplasms, and has a poor prognosis.[73] The main clinical manifestations of PNP are mucosal erosions and polymorphous cutaneous lesions. For occult tumours, which may not be found by radiological examination, attention should be paid to other indicators, such as detection results of immunological methods.

PNP's antigens mainly include proteins of the plakin family (envoplakin and periplakin were the most commonly found), alpha-2-macroglobulin-like protein 1 (A2ML1), desmocollins, Dsg 3, etc. (Table 1).[74–76] In recent years, new autoantigens including laminin 332, BP180, epiplakin and protein P200 have been identified.[77–80] To detect anti-envoplakin and anti-periplakin antibodies, using full-length recombinant proteins in IP and ELISA is highly effective.[81] Nagata et al. found autoantibodies in a majority of patients with PNP showed a strong reactivity with multiple recombinant proteins of envoplakin and periplakin, but few cases for the C-terminal homologous domain of periplakin (amino acid residues 1646–1756), which was detected by IB.[82] Another study demonstrated that envoplakin residues 1–141 and residues 1684–1784 were the main antigenic epitopes of PNP proven by ELISAs, and envoplakin (residues 213–377) might be relevant to bronchiolitis obliterans or lichen planus-like lesions.[45] Also, anti-epiplakin antibodies may have a significant correlation with bronchiolitis obliterans and mortality.[79] After p170 antigen was identified as A2ML1 (a broad-range protease inhibitor),[83] Numata et al.[76] demonstrated that anti-A2ML1 antibody, which can be detected by IP with high sensitivity, was associated with early onset and absence of ocular involvement, and that its possible pathogenic role may be to decrease adhesion of normal human keratinocytes by activating plasmin via inhibition of A2ML1.

DIF showed the intercellular deposition of IgG and C3 and linear/granular deposition of immunoglobulins and/or C3 along the BMZ in most cases. DIF may not confirm the diagnosis of PNP, as the proportion of both epidermal cell-surface and BMZ deposition (the major characteristic of PNP) in the DIF patterns of PNP is not high.[75,84] Therefore, IIF and biochemical tests should be considered.

IIF on monkey oesophagus makes it difficult to differentiate PNP from PV. In contrast with PV, the autoantibodies of PNP can bind to urinary bladder epithelium (the strongest and most consistent binding being due to its high expression of plakins), respiratory epithelium, small-bowel epithelium, etc.[73] The diagnosis of PNP is based on the convenient and inexpensive method of IIF on rat bladder epithelium, which is more sensitive than DIF.[75] For highly suspected PNP with negative results of IIF on rat bladder epithelium, IB and/or IP, with high sensitivity and specificity, should be applied.[29]

Moreover, a new ELISA using a recombinant 56-kDa N-terminal fragment of envoplakin and a novel commercially available ELISA kit (EUROIMMUN US Inc., Mountain Lakes, NJ, U.S.A.) showed high diagnostic specificity for PNP.[44,46]

Lim et al.[85] reported on a 56-year-old woman, initially diagnosed with lichen planus based on the clinical manifestations and histopathology. Immunological methods were performed when the patient showed refractory lesions and rapid weight loss. The IIF on human skin and rat bladder demonstrated epidermal cell-surface and BMZ deposition, respectively. ELISAs were positive for Dsg1 and desmocollins 1, 2 and 3. IB assays for envoplakin and periplakin and laminin γ1 were also positive. A malignant thymoma was subsequently found. Immunological techniques eventually helped confirm the diagnosis of PNP.

Remarkably, some cases may be better explained as the neoplasia-induced and cell-mediated immune response rather than humoral response to autoantigens. Cummins et al.[86] found that four patients with lymphocytic neoplasms and clinical features of the lichenoid variant of PNP had no detectable autoantibodies, by DIF, IIF or IP.

Mucous Membrane Pemphigoid

MMP is a chronic inflammatory, immunobullous subepithelial blistering disease characterized by predominant involvement of mucous membranes and autoantibodies against structural proteins of the BMZ.[87]

Autoantibodies of MMP include BP180 (the principal antigen), laminin 332 (laminin 5) and α6β4-integrin; others have been identified (Table 1). The reactivity of autoantibodies against BP180 showed association with localization of MMP lesions and inverse correlation with disease severity score.[88] The C-terminal domain of BP180 is predominantly recognized, followed by NC16A epitopes on BP180. Amniotic membrane IB to detect BP180 is sensitive for MMP compared with common ELISA.[89] The full-length BP180 ELISA is preferable, as the common ELISA is only for NC16A, which cannot detect anti-C-terminal autoantibodies.[90] Laminin 332 is a heterotrimer in which the α3 chain is mostly targeted (86·4%).[47,91] Mucosal involvement (especially the mouth and eye) was commonly found in patients with MMP with anti-laminin 332 autoantibodies, and was usually more severe than with skin involvement alone. IgG autoantibodies against laminin 332 are significantly associated with more severe MMP, such as MMP with pharyngolaryngeal and oropharyngolaryngeal involvement, which may be life-threatening.[47,92,93] Laminin 332 may have a correlation with tumours; patients with MMP with autoantibodies against laminin 332 showed a high mortality and only a minority go into remission.[94] For detecting anti-laminin 332 antibodies, the most sensitive method is IP; nevertheless, IB of human keratinocyte extracellular matrix is a practical alternative.[95,96] The α6 integrin antibodies were detected in the serum of MMP with oral lesions throughout the clinical course, and β4 integrin is a specific autoantigen in pure ocular MMP.[97,98]

DIF showed linear deposition of IgG and/or C3 along the BMZ. A few cases with IgA dominant in MMP need biochemical tests for differentiating MMP from LAD. Biopsies should be taken from perilesional oral mucosa or skin, which can obtain accurate DIF results;[99,100] repeated biopsies can increase diagnostic sensitivity.[20] Low diagnostic sensitivity of routine IIF[89] can be improved by a novel technique that uses recombinant laminin 332, or IIF on SSS, and which is characterized by IgG bound to epidermal and/or dermal sides of SSS. The presence of both IgA and IgG, compared with IgG alone, is usually associated with the serious condition of MMP.

Up to the present, protein microarrays have been used to evaluate the global immune and angiogenic profile of tear proteins in ocular MMP,[101] and this may also be useful in the diagnosis of oral MMP.

The following two cases from our clinic of oral medicine showed atypical clinical manifestations or histopathology of MMP, which were eventually confirmed by immunological methods. A 72-year-old Chinese woman without any particular past and allergic histories presented with recurrence of desquamation on the soft palate for 2 months. Intraoral examination revealed a similar clinical picture of traumatic ulceration and a negative result of the Nikolsky sign. The soft palate showed several erosions with yellow pseudomembrane and the right portion of the palate showed large areas of hyperaemia and redness with white streaks on the periphery (Figure 4a). However, the patient denied history of trauma. Histopathology of the oral mucosa revealed local subepithelial cleavage (Figure 4b). DIF showed linear deposition of IgG and C3 along the BMZ (Figure 4c, d). Moreover, ELISA for BP180 NC16A also showed a positive result. Based on these findings, the patient was diagnosed with oral MMP. The second patient, a 73-year-old man, complained of painful erosions for 11 months. He had extensive erythematous lesions on the gingiva with scattered erosions, as well as extensive erosions on the right buccal mucosa and palatal mucosa (Figure 5a–c). Nikolsky sign was negative. Histopathology showed no subepithelial blister formation (Figure 5d). Dsg1, Dsg3 and BP180 NC16A ELISAs also showed negative results. However, the patient was diagnosed with oral MMP when DIF revealed linear IgG, C3 and IgA deposition along the BMZ (Figure 5e–g).

Figure 4.

Clinical and histopathological features and direct immunofluorescence patterns for the 72-year-old woman (mucous membrane pemphigoid). (a) Large areas of hyperaemia and redness with white streaks on the periphery and several erosions with yellow pseudomembrane appeared on the soft palate. (b) Histopathology revealed local subepithelial cleavage. (c, d) Direct immunofluorescence demonstrated linear deposition of IgG and C3 along the basement membrane zone.

Figure 5.

Clinical and histopathological features and direct immunofluorescence patterns for the 73-year-old man (mucous membrane pemphigoid). (a–c) Extensive erythematous lesions on the gingiva with scattered erosions, as well as extensive erosions on the right buccal mucosa and palatal mucosa. (d) Histopathology revealed no subepithelial blister formation. (e–g) Direct immunofluorescence showed linear IgG, C3 and IgA deposition along the basement membrane zone.

Linear IgA Disease

LAD is a rare acquired subepidermal autoimmune blistering disorder that was proposed by Jablonska and Chorzelski in 1979[102] based on linear deposition of IgA at the BMZ.

So far many autoantigens of LAD have been detected, including proteins of 97, 120, 100, 130, 145–290, 160, 180, 190, 200, 210, 220, 230, 250, 255, 285 and 290 kDa. LAD-1 (120-kDa, shed ectodomain of BP180) and LAD97 (97-kDa, N-terminal portion of LAD-1) are the major antigens, followed by others (Table 1).[103–109] The NC16A domain of BP180 can be targeted in 20% of cases.[110] In addition to IB, a new ELISA system for detecting IgA specific to BP180 showed high sensitivity (Table 3).[49]

DIF, the diagnostic gold standard for LAD, demonstrated linear deposition of IgA (occasionally with IgG and C3) along the dermoepidermal junction, and the antibodies are universally IgA1 subclass.[111] IIF with monkey oesophagus or human epidermal substrate showed low titres of circulating IgA antibodies, and not a high proportion was positive on IIF. When SSS is used in IIF, it can increase the total proportion of positive IIF results and titres of antibody binding,[23,112] and epidermal binding of SSS is more frequent than dermal binding or dermal and epidermal binding. Substrate from skin of inherited epidermolysis bullosa was considered useful in IIF to identify autoantigens for LAD.[113]

Lichen Planus Pemphigoides

LPP is a mucocutaneous autoimmune blistering disease considered by most scholars to be an independent disease. The immunofluorescent findings have considerable value for diagnosing LPP.

DIF from lesions and normal tissues have revealed deposition of IgG, IgM, C3, fibrin or fibrinogen in a linear/granular pattern along the BMZ, and cytoid bodies (CBs) may exist in some cases. Circulating IgG antibodies can be detected at high titres using IIF, and IgG showed epidermal deposition by IIF on SSS. Diagnostic sensitivity needs to be further studied due to limited reports of LPP. IB with epidermal extracts showed the main antigen epitope was the BP180 NC16A domain. The BP180 C-terminal domain and other antigens were also reported (Table 1),[114–117] and the 200-kDa antigen needs molecular recognition.

Oral Lichen Planus

OLP, the most common immune-mediated disorder with oral mucosal involvement (51–76% of lichen planus affects the oral mucosa),[118] presents as white/grey striae or plaques.

The clinical characteristics are usually used to make the initial diagnosis, and immunological methods can differentiate OLP from AIBDs, DLE and other patchy strial diseases. Some scholars have suggested that immunofluorescence is more reliable compared with histopathology.[119] Using IIF, no circulating BMZ antibodies have been detected in OLP, and this can be a differential diagnostic point between LPP and bullous OLP. When atypical clinical manifestations such as erosive OLP presenting as desquamative gingivitis occur, the disease can be diagnosed by DIF and biochemical tests.

Shaggy or granular deposition of fibrin, fibrinogen, immunoglobulins and C3 (C3 may be associated with severe OLP[119–121]) along the BMZ with or without IgM-positive CBs[122,123] have been detected by DIF from lesional biopsies. Some cases showed fibrin/fibrinogen of the small blood vessels.[122,124] The CBs, which can be a diagnostic clue for OLP, may also be found in LPP, lupus erythematosus and BP.

Using IIF on HEp-2, OLP had a significantly higher frequency of serum ANA (23–28%) than the healthy control group (6%), and the higher erosive level of OLP was correlated with a higher frequency of ANA.[125–127] Some cases reported that Dsg3 and/or Dsg1 autoantibodies have been detected by ELISA in erosive OLP with negative results from DIF and IIF. However, one of these cases[128] found that the Ca2+-independent linear epitope of Dsg3, which had been considered to be nonpathogenic, was targeted by Dsg3 autoantibodies. Autoantibodies against Dsg in erosive OLP may be induced by severe damage of oral keratinocytes and subsequent release of Dsg proteins.[128–130] Thus, immunological methods are critical for the definite diagnosis of this type of OLP.

A retrospective study[131] showed that 65 cases without histopathological features of OLP were found to have DIF patterns of OLP. Three of the cases were dysplasia, carcinoma in situand squamous cell carcinoma, and the other 62 patients were finally diagnosed with OLP. Therefore, DIF is of great value, especially when the atypical histopathological features of OLP appear. Combinations of clinical findings, histopathology and DIF can help to achieve the final diagnosis.

Discoid Lupus Erythematosus

DLE is a chronic, mucocutaneous disorder that typically involves sun-exposed areas, and manifests as persistent erythema and discoid plaques. The oral mucosa is involved in approximately 20% of cases, with the lower lip being the most common site.

DIF revealed a characteristic 'lupus band' at the BMZ[132,133] in which IgG and IgM were the most common immunoreactants, followed by IgA/C3/fibrinogen.[119,132,133] The band presented as various depositions, including granular, homogeneous and reticular patterns. Sometimes, deposition of CBs (IgM, IgG), small blood vessel walls and epidermal nuclei may appear in DLE. The lupus band contributes to diagnosis, prognosis and evaluation of therapeutic effects. Biopsies should be taken from lesions, as biopsies from normal tissues are often negative, especially from nonsun-exposed areas.[134,135] When a positive result from a biopsy of normal tissues appears, it indicates that DLE may develop to SLE. DIF is usually positive for lesions with a duration of more than 6 weeks and lesions from the head, neck and upper extremities rather than the trunk.

Detection of ANA titre by IIF can assess the severity of DLE. One of the risk factors for DLE developing to SLE is a high ANA titre,[136–138] which reminds us to adjust treatment plans for preventing SLE development.